Muscle & Nervelib.ajaums.ac.ir/booklist/904492.pdf · BASIC ANATOMY AND PHYSIOLOGY OF THE SYMPATHETIC SYSTEM ... Sympathetic Innervation of the Skin. Compared to muscle, the sympathetic

Muscle & Nerve

Volume 36 Issue 5 , Pages 595 - 725 (November 2007) Invited Reviews Sympathetic neural control of integrated cardiovascular function: Insights from measurement of human sympathetic nerve activity (p 595-614) B. Gunnar Wallin, Nisha Charkoudian Published Online: Jul 10 2007 3:46PM DOI: 10.1002/mus.20831

Clinical and immunological spectrum of the Miller Fisher syndrome (p 615-627) Y. L. Lo Published Online: Jul 26 2007 2:15PM DOI: 10.1002/mus.20835

Main Articles Multiple measures of axonal excitability in peripheral sensory nerves: An in vivo rat model (p 628-636) Annette George, Hugh Bostock Published Online: Jul 24 2007 1:35PM DOI: 10.1002/mus.20851

Effect of age on adrenergic and vagal baroreflex sensitivity in normal subjects (p 637-642) Chih-Cheng Huang, Paola Sandroni, David M. Sletten, Stephen D. Weigand, Phillip A. Low Published Online: Jul 24 2007 1:36PM DOI: 10.1002/mus.20853

Comparative efficacy of repetitive nerve stimulation, exercise, and cold in differentiating myotonic disorders (p 643-650) Patrik Michel, Damien Sternberg, Pierre-Yves Jeannet, Murielle Dunand, Francine Thonney, Wolfram Kress, Bertrand Fontaine, Emmanuel Fournier, Thierry Kuntzer Published Online: Jul 24 2007 1:37PM DOI: 10.1002/mus.20856

Frequency of seronegativity in adult-acquired generalized myasthenia gravis (p 651-658) Koon Ho Chan, Daniel H. Lachance, C. Michel Harper, Vanda A. Lennon Published Online: Jul 24 2007 1:36PM DOI: 10.1002/mus.20854

Lateral femoral cutaneous neuropathy and its surgical treatment: A report of 167 cases (p 659-663) Igor Benezis, Benoit Boutaud, Jerome Leclerc, Thierry Fabre, Alain Durandeau Published Online: Jul 26 2007 2:19PM DOI: 10.1002/mus.20868

Can end-to-side neurorrhaphy bridge large defects? An experimental study in rats (p 664-671) Marios G. Lykissas, Anastasios V. Korompilias, Anna K. Batistatou, Gregory I. Mitsionis, Alexandros E. Beris Published Online: Jul 27 2007 1:47PM DOI: 10.1002/mus.20861

Isotonic fatigue in laminin 2-deficient dy/dy dystrophic mouse diaphragm (p 672-678) Jennifer Pollarine, Michelle Moyer, Erik Van Lunteren Published Online: Jul 27 2007 1:46PM DOI: 10.1002/mus.20860

Altered expression of PGK1 in a family with phosphoglycerate kinase deficiency (p 679-684) Eva K. Svaasand, Jan Aasly, Veslemøy Malm Landsem, Helge Klungland Published Online: Jul 27 2007 1:45PM DOI: 10.1002/mus.20859

Ubiquitin-ligase and deubiquitinating gene expression in stretched rat skeletal muscle (p 685-693) Antonio Garcia Soares, Marcelo Saldanha Aoki, Elen Haruka Miyabara, Camila Valentim DeLuca, Hélcio Yogi Ono, Marcelo Damário Gomes, Anselmo Sigari Moriscot Published Online: Jul 26 2007 2:16PM DOI: 10.1002/mus.20866

Multijoint reflexes of the stroke arm: Neural coupling of the elbow and shoulder (p 694-703) Samir G. Sangani, Andrew J. Starsky, John R. Mcguire, Brian D. Schmit Published Online: Jul 12 2007 2:54PM DOI: 10.1002/mus.20852

Short Reports Analysis of a genetic defect in the TATA box of the SOD1 gene in a patient with familial amyotrophic lateral sclerosis (p 704-707) Stephan Niemann, Wendy J. Broom, Robert H. Brown Jr. Published Online: Jul 18 2007 11:10AM DOI: 10.1002/mus.20855 Episodic hypoxia exacerbates respiratory muscle dysfunction in DMDmdx mice (p 708-710) Gaspar A. Farkas, Kathleen M. Mccormick, Luc E. Gosselin Published Online: Jul 24 2007 1:38PM DOI: 10.1002/mus.20858

Pain and soreness associated with a percutaneous electrical stimulation muscle cramping protocol (p 711-714) Kevin C. Miller, Kenneth L. Knight Published Online: Jul 24 2007 1:37PM DOI: 10.1002/mus.20857

Cases of the Month Intraneural perineurioma of the radial nerve visualized by 3.0 Tesla MRI (p 715-720) Doris Nguyen, P. James Dyck, Jasper R. Daube Published Online: Apr 30 2007 3:01PM DOI: 10.1002/mus.20795 Myositis with sensory neuronopathy (p 721-725) Marcondes C. França Jr., Andréia V. Faria, Luciano S. Queiroz, Anamarli Nucci Published Online: Apr 27 2007 11:53AM DOI: 10.1002/mus.20783

INVITED REVIEW ABSTRACT: Sympathetic neural control of cardiovascular function is es-sential for normal regulation of blood pressure and tissue perfusion. In thepresent review we discuss sympathetic neural mechanisms in human car-diovascular physiology and pathophysiology, with a focus on evidence fromdirect recordings of sympathetic nerve activity using microneurography.Measurements of sympathetic nerve activity to skeletal muscle have pro-vided extensive information regarding reflex control of blood pressure andblood flow in conditions ranging from rest to postural changes, exercise, andmental stress in populations ranging from healthy controls to patients withhypertension and heart failure. Measurements of skin sympathetic nerveactivity have also provided important insights into neural control, but areoften more difficult to interpret since the activity contains several types ofnerve impulses with different functions. Although most studies have focusedon group mean differences, we provide evidence that individual variability insympathetic nerve activity is important to the ultimate understanding of theseintegrated physiological mechanisms.

Muscle Nerve 36: 595–614, 2007

SYMPATHETIC NEURAL CONTROL OFINTEGRATED CARDIOVASCULAR FUNCTION:INSIGHTS FROM MEASUREMENTOF HUMAN SYMPATHETIC NERVE ACTIVITY

B. GUNNAR WALLIN, MD,1 and NISHA CHARKOUDIAN, PhD2

1 Institute of Neuroscience and Physiology, Sahlgrenska Academy at Goteborg University,S-413 45 Goteborg, Sweden

2 Department of Physiology & Biomedical Engineering, Mayo Clinic College of Medicine,Rochester, Minnesota, USA

Accepted 25 April 2007

In an era when the importance of integrative systemsphysiology is reemerging into the spotlight of bio-medical science, the sympathetic nervous system canbe viewed as the ultimate integrator of systems phys-iology in control of cardiovascular function. It isliterally impossible to consider systemic control ofthe cardiovascular system without integrating theroles and prominence of sympathetic neural mech-anisms into any proposed scheme. Indeed, one ofthe most exciting aspects of measuring sympatheticneural activity is the ability of the investigators to seeintegrative physiology “in action” every time they doan experiment.

Our goal in the present review is to synthesizeevidence regarding the physiological and pathophys-

iological roles of sympathetic nerve activity that arecentral to the understanding of integrated humancardiovascular function. We will primarily focus onwork involving microneurographic measurements ofsympathetic nerve activity, but also on evidence fromnorepinephrine spillover studies and other comple-mentary techniques, where appropriate. For moredetailed explanation of microneurographic method-ology, or for more in-depth discussion of studies notcovered here, the reader is referred to several otherreviews on related topics.50,117,122,201

BASIC ANATOMY AND PHYSIOLOGY OF THESYMPATHETIC SYSTEM

The sympathetic neural innervation of the heartand peripheral circulation originates primarilyfrom the intermediolateral cell column of the spi-nal cord. Sympathetic preganglionic neurons havecell bodies in the thoracic and upper lumbar re-gions. The short preganglionic fibers synapse atparavertebral or visceral (prevertebral) ganglia.The postsynaptic neurons have longer axons that

Abbreviations: LBNP, lower body negative pressure; MSNA, muscle sym-pathetic nerve activity; SSNA, skin sympathetic nerve activityKey words: blood pressure; cardiovascular; circulation; hypertension; sym-pathetic nervous systemCorrespondence to: B. G. Wallin; e-mail: [email protected]

© 2007 Wiley Periodicals, Inc.Published online 10 July 2007 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/mus.20831

Sympathetic Nerve Activity MUSCLE & NERVE November 2007 595

extend to target organs such as heart, blood ves-sels, and sweat glands. In humans, most vascularsympathetic nerves cause vasoconstriction; theirprimary transmitter is norepinephrine but, in ad-dition, they release several cotransmitters (e.g.,neuropeptide Y).

The relationship between pre- and postgangli-onic impulse activity is complex.127 There are manymore postganglionic than preganglionic neurons; inhumans the ratio may be as high as 200:1. It ispredicted that in human paravertebral ganglia a sin-gle preganglionic neuron may synapse with around4,000 postganglionic neurons, and there are at least20 preganglionic inputs converging on each post-ganglionic cell body. Usually, however, only one (orvery few) of these inputs forms a “strong” synapse onthe postganglionic neuron, i.e., has a high safetyfactor for transmission of impulses. The likelihoodof synaptic transmission due to summation of “weak”preganglionic inputs is low, at least in paravertebralganglia.

For a long time it was believed that the sympa-thetic system was undifferentiated so that thestrength of sympathetic activity varied in parallel innerves to different tissues. With the introduction ofmicroneurographic recordings of human skin andmuscle sympathetic activity, it immediately becameclear that this concept was erroneous, and today it isgenerally accepted that the sympathetic system ishighly differentiated, and that each sympathetic sub-division is governed by its own specific reflexes.

Sympathetic Innervation of Skeletal Muscles. Be-cause the skeletal muscle circulation makes up alarge proportion of cardiac output, both at rest andduring physical activity, the neural control of thiscirculation is fundamental to systemic hemodynam-ics. Human muscle sympathetic nerve activity(MSNA) consists only of vasoconstrictor impulses,the outflow of which is modulated from central ner-vous sites and from a large number of peripheralreceptor populations, the most important of whichare listed in Table 1.

Data on resting MSNA and responses to varioustypes of perturbations have provided substantialmechanistic information about baroreflex control ofblood pressure, about intraindividual sympatheticresponsiveness, and about changes with aging anddisease. Recently, the study of interindividual vari-ability in MSNA and its relationship to other aspectsof hemodynamic control has also given importantinsight into blood pressure regulation and the bal-ance of factors that keeps blood pressure normal inspite of substantial differences among individuals.

Sympathetic Innervation of the Skin. Compared tomuscle, the sympathetic innervation of the skin inhumans is more complex, since cutaneous sympa-thetic nerves include four different fiber types: vaso-constrictor, vasodilator, sudomotor, and pilomotor.The fibers are mainly involved in thermoregulationbut can also be activated from other peripheral re-ceptor stations and the central nervous system (Ta-ble 1). Cutaneous sympathetic vasoconstrictornerves are tonically active in thermoneutral environ-ments,14 and changes in the activity of these nervesare responsible for the minor variations in skinblood flow that occur during normal daily activitiesin the absence of significant hyperthermia. Sympa-thetic vasodilator nerves do not exhibit resting tone,and are only activated during increases in body coretemperature. Once activated, however, vasodilatorimpulses are responsible for 80%–90% of the largeincreases in skin blood flow seen in conditions ofhyperthermia. Sympathetic vasodilator fibers act viaa mechanism that involves cholinergic cotransmis-sion,106 although acetylcholine itself is not the mainmediator, since atropine does not block active vaso-dilation in the skin.106,112 Such observations havebrought into question whether vasodilator and sudo-motor nerves are in fact one nerve type; this issueremains unresolved. Sympathetic sudomotor nervesare cholinergic, and an increase in their activityduring hyperthermia causes sweat release. Sympa-thetic pilomotor nerves are the least well understoodand so far no recordings have been made fromhuman pilomotor fibers. In the second part of thisreview we will discuss the challenges to interpreta-tion of skin sympathetic nerve activity (SSNA) that

Table 1. Receptors and simple maneuvers that influencesympathetic nerve traffic.

SSNA MSNA

Temperature receptors inCNS and skin

Arousal and stressRespirationCardiopulmonary receptorsSleepPain

Arterial baroreceptorsCardiopulmonary receptorsSystemic chemoreceptorsIntramuscular mechano- and

metaboreceptorsRespiration/apneaVestibular receptorsLaryngeal receptorsStretch receptors in the

urinary bladderTemperature receptors in

CNS and skinPainArousal and stressSleep

SSNA, skin sympathetic nerve activity; MSNA, muscle sympathetic nerveactivity; CNS, central nervous system.

596 Sympathetic Nerve Activity MUSCLE & NERVE November 2007

arise from the simultaneous existence of these foursympathetic nerve types in the skin.

MEASUREMENT OF SYMPATHETIC NEURAL ACTIVITYIN HUMANS

Microneurography. In the 1960s, Hagbarth andVallbo201 developed the microneurographic tech-nique for direct measurement of action potentials inmyelinated nerve fibers of awake human subjects.The method soon proved to be useful also for re-cordings from unmyelinated sympathetic nerve fi-bers.76 The technique involves the insertion of atungsten microelectrode with a tip of a few micronsinto a suitable peripheral nerve, and for sympatheticrecordings the peroneal nerve (innervating thelower leg) is a common choice. Most sympatheticrecordings are multifiber recordings (multiunit ac-tivity) but impulses in single sympathetic fibers (sin-gle unit activity) can also be monitored.117

A typical characteristic of sympathetic nerve fibersis that they display spontaneous activity and, since ac-tivity in neighboring fibers is usually synchronized, sym-pathetic multiunit activity occurs as “bursts” of impulsesseparated by silent periods. A typical multiunit record-ing of muscle sympathetic nerve activity is shown inFigure 1. To interpret such records, it is necessary totake into account a time delay between the neurogramand tracings of cardiovascular variables. This time delayis brought about mainly by the slow conduction veloc-ity of the postganglionic sympathetic impulses (seebelow). The bursts occur in different temporal patternsin skin and muscle nerve branches; responses inducedby certain maneuvers also differ. These characteristicsprovide sufficient information for reliable differentia-

tion between multiunit MSNA and SSNA. Identifica-tion of single fiber activity is more difficult, and this isespecially so for sympathetic nerve fibers innervatingthe skin.

When recording from only one type of nervefiber (e.g., muscle vasoconstrictor fibers) the multi-unit activity (displayed in an “integrated” neurogramwith a time constant of 0.1 s, cf. Fig. 1) providesuseful quantitative information on the averagestrength of activity. In contrast, information frommultiunit activity containing more than one type ofimpulses (e.g., from both muscle and skin sympa-thetic fibers) is difficult to interpret since the actionpotentials from the different fibers cannot be sepa-rated in the neurogram.

In a constant electrode site the strength of mul-tiunit activity is quantified by counting the numberof bursts and their areas (or amplitudes) in theintegrated neurogram (no. of bursts � average burstarea � total MSNA). An important limitation, how-ever, is that burst area/amplitude can only be usedin an unchanged electrode site. If the electrodemoves during the recording or interindividual com-parisons are required, only the number of bursts canbe used. Single unit recordings are analyzed in theoriginal neurogram and provide information aboutimpulse frequency and discharge characteristics inan individual fiber, and how such characteristics varybetween fibers. Single unit discharge frequency canbe given as average frequency over a longer time oras instantaneous frequency (i.e., frequency calcu-lated from the interval between two succeedingspikes). In addition, it may be useful to relate firingto the number of cardiac intervals by calculating theprobability of firing for a unit (� % cardiac intervalsassociated with spikes) and the probability of multi-ple spikes in a cardiac interval (in % of all cardiacintervals with spikes). Methodological details areprovided elsewhere.65,117,200,207

Measurements of Noradrenaline Spillover. In addi-tion to microneurography, measurements of nor-adrenaline spillover from sympathetic nerves (pio-neered by Murray Esler and colleagues in Mel-bourne50,53) have provided important new informationon sympathetic activity. The rationale is that each sym-pathetic nerve impulse releases a certain amount ofnoradrenaline from the nerve endings, a small fractionof which “spills over” into the circulation. The tech-nique allows assessment both of whole-body noradren-aline spillover and of regional spillover from visceralnerves that are inaccessible to microneurography. Spill-over measurements do not provide the same time res-olution as microneurography but give good indirect

FIGURE 1. Typical “integrated” (mean voltage) record of multiunitmuscle sympathetic nerve activity (MSNA) with simultaneouselectrocardiogram (ECG) and arterial pressure (AP) tracing froma healthy human subject, showing bursting pattern and relation-ship of MSNA to the cardiac cycle. Solid arrows show the rela-tionship between a given cardiac cycle and the correspondingburst of MSNA. Baroreflex latency (shown by the dashed arrow)is usually calculated from the R wave associated with the systolicpulse wave to the peak of the MSNA burst (the peak taken as thestart of inhibition).


estimates of the average strength of sympathetic activityover a couple of minutes.

MUSCLE SYMPATHETIC NERVE ACTIVITY

Physiology. Resting Activity. In microneurographicmultiunit recordings the amount of spontaneousresting activity is similar in arm and leg nerves, butthere are large interindividual differences, and burstincidence may vary from less than 5 up to close to100 bursts per 100 heartbeats.25,58,192 In a given sub-ject the number of bursts is reproducible over a longtime, which makes it possible to monitor long-termchanges in MSNA, both in the context of diseasesand therapeutic interventions. However, since foodintake increases resting activity,57 reliable data areobtained only if no food intake is allowed for at least2 h prior to the recording. Since both subjects inpairs of homozygotic twins have similar MSNA levels,the interindividual differences are likely to be ofgenetic origin.211 At the single fiber level, restingfiring frequency is �0.4 Hz, probability of firing�30%, and probability of multiple spikes �30%.117

A high multiunit burst incidence at rest is due to ahigher number of active sympathetic fiber; firingfrequencies are similar in subjects with high and lowburst incidence.120

When measured simultaneously, MSNA and nor-adrenaline spillover in the heart210 or the kidney213

in subjects at rest have shown significant positivecorrelations. Thus, it appears that the interindi-vidual differences in resting sympathetic traffic aresimilar in nerves to these three hemodynamicallyimportant tissues. Similar results were obtained inrecent studies in rabbits in which muscle, cardiac,and renal sympathetic activities were recorded simul-taneously.100,101 This parallelism is probably a mainexplanation of why the plasma concentrations ofnoradrenaline in forearm venous blood correlatewith MSNA at rest: the noradrenaline concentrationreflects spillover, not only from muscle nerves butalso from nerves to other tissues that have similarinterindividual differences in sympathetic activity.

Dynamic relationship to blood pressure. The moststriking and best-known attribute of MSNA is itsclose, dynamic relationship to blood pressure and itsinvolvement in blood pressure regulation by way ofthe arterial baroreflex. Within a given individual,short-lasting spontaneous variations in blood pres-sure cause marked opposing changes in MSNA,which act to reverse or “buffer” the changes in pres-sure. It is this negative feedback mechanism thatinduces the characteristic cardiac rhythmicity andthe inverse relationship between variations of pres-

sure and nerve traffic.191,201,206,207 Accordingly, if theafferent activity from the baroreceptors is preventedfrom reaching the sympathetic preganglionic neu-rons, both these characteristics are eliminated.61,188

When studying arterial baroreflex effects on MSNA,it is necessary to compensate for the delay (latency)in the baroreflex arc. Usually this latency is calcu-lated as the time from the R-wave of the electrocar-diogram and the peak of the appropriate sympa-thetic burst (the peak is interpreted as the start ofthe sympathetic inhibition brought about by the sys-tolic pressure wave).59,191 The details of arterialbaroreflex control of MSNA are still incompletelyunderstood, but there is evidence suggesting that themechanism controlling MSNA burst occurrence dif-fers from that controlling burst strength.108

It is not fully clarified how and from whichbaroreceptors the afferent beat-to-beat informationis conveyed to the brainstem. The pressure parame-ter that correlates best to variations of MSNA is theend diastolic blood pressure.191 However, since aburst starts to occur before the end diastolic bloodpressure is reached, some other factor, closely re-lated to end diastolic blood pressure, is likely to beprimary. Our recent finding of a systematic relation-ship between cardiac output and MSNA25 suggeststhat stroke volume or cardiac interval are key param-eters and, if so, both arterial and cardiopulmonaryreceptors may be involved.

The intraindividual relationship between MSNAand arterial pressure can be modified by a number offactors, including, but not limited to, age, posture,hypoxia, hydration, exercise, female reproductive hor-mones, and arousal.23,43,79,129,143,191 To characterizesuch modifications it is useful to evaluate the sensitivity,or responsiveness, of the arterial baroreflex. The mostcommon approach is the so-called “modified Oxford”technique,23,46,160 which involves intravenous injectionsof sequential boluses of nitroprusside and phenyl-ephrine. The vasoactive drugs induce changes ofblood pressure that are counteracted by oppositechanges of sympathetic activity. Less common meth-ods are to use steady-state infusions of the vasoactivesubstances96 or to determine baroreflex sensitivityfrom spontaneous blood pressure variations.108 Thespontaneous variations of resting blood pressure andMSNA can also be quantified in a “threshold vari-ability diagram” that defines the mean baroreflexsetpoint and its variability.108 Another approach in-volves directly changing transmural pressure at thecarotid sinus using neck pressure and suction. Theusefulness of this approach is limited by the simulta-neous counteracting influence of aortic baroreceptors.


Chronic resting levels of blood pressure: long-term inter-actions between MSNA and blood pressure. At rest, thenumber of sympathetic bursts in MSNA may vary10-fold or more among normal healthy sub-jects.25,26,179,191 Variables such as age, sex, body massindex, and growth hormone activity have beenfound to influence or correlate with resting MSNA,but even if these factors are taken into account, thelarge interindividual variability remains. In contrastto the close inverse intraindividual relationship be-tween MSNA and blood pressure, the interindividualrelationship between mean levels of MSNA andblood pressure at rest is weak216 or absent.25,26,179,191

Recently, Narkiewicz et al.139 showed that the corre-lation was completely absent in healthy subjects be-low the age of 40 years, whereas in older subjectsthere was a significant positive relationship betweenMSNA and blood pressure (discussed later). Thus,young people may have very high or very low restinglevels of MSNA and still have similar blood pressures.This is particularly perplexing in light of the fact thatmeasurement of the number of bursts in MSNA(bursts/min or bursts/100 heartbeats) is extremelyreproducible in a given subject58 over months andeven years. So the lack of relationship betweenMSNA and blood pressure among individuals is notdue to a day-to-day variability of resting sympathetictraffic.

A possible explanation for this lack of relation-ship between MSNA and blood pressure in youngsubjects might be an inverse relationship betweenresting levels of MSNA and resting levels of sympa-thetic nerve activity to other vascular beds, whichwould “balance out” the variable vasoconstrictor in-fluence of MSNA, such that net effects on bloodpressure would be minimal. There is, however, noexperimental support for this alternative: restingMSNA was found to correlate well with norepineph-rine spillover to the heart and the kidney, as well aswith whole-body norepinephrine spillover.210,213

These findings suggested that, at rest, MSNA is agood index of whole-body sympathetic vasoconstric-tor activity, and did not solve the mystery as to why(or how) MSNA does not correlate well with restingblood pressure.

We recently undertook a series of studies to clar-ify why MSNA at rest is not related with resting bloodpressure. The two main contributors to mean arte-rial pressure are cardiac output and total peripheralresistance and we hypothesized that the variability inMSNA at rest was balanced by a reciprocal variabilityin resting cardiac output. That is, if MSNA is a goodindicator of whole-body sympathetic nerve activity atrest and is therefore important for determining total

peripheral resistance, then if people with highMSNA have low cardiac output, or vice versa, the twomain contributors to blood pressure would balanceeach other, resulting in minimal interindividual dif-ferences in pressure. The results showed that therewas indeed an inverse relationship between restingMSNA and resting cardiac output (Fig. 2). There wasalso a strong positive correlation between restinglevels of MSNA and total peripheral resistance, sup-porting the idea that MSNA is a major contributor tototal peripheral resistance at rest. Furthermore, ourdata suggested inhibitory influences of both cardiacoutput and stroke volume on baroreflex control ofMSNA which might provide the mechanistic basis forour findings.25

In a second series of experiments, we tested thehypothesis that variability in vascular adrenergic re-sponsiveness also contributes to the integrated bal-ance of factors that keeps blood pressure normal inspite of wide interindividual variability in sympa-thetic nerve activity.26 We found an inverse relation-ship between resting MSNA and forearm vascularresponsiveness to norepinephrine and tyramine.This was particularly striking when responses to tyra-mine were quantified based on arterio-venous differ-ences in norepinephrine: individuals with lowerMSNA had greater vasoconstrictor responses for agiven amount of local norepinephrine release (Fig.3). These findings suggest that vascular adrenergicresponsiveness is downregulated in proportion tothe level of resting MSNA.26

FIGURE 2. Inverse relationship between MSNA at rest (ex-pressed as bursts/100 heart beats) and cardiac output (CO)among young healthy male subjects. We hypothesize that thisinverse relationship helps to balance the potential pressor effectsof higher MSNA, contributing to the lack of relationship betweenMSNA and blood pressure in young healthy individuals. FromCharkoudian et al., 2005; reprinted with permission.


There has been some suggestion that tonic vari-ability in the contribution of nitric oxide to restingvascular tone might also contribute to the normalregulation of blood pressure in individuals withwidely varying MSNA.179 In a recent study, we foundthat systemic inhibition of nitric oxide synthesis withL-NMMA caused greater increases in arterial pres-sure in individuals with higher resting MSNA. How-ever, we could not support the idea that this was dueto a greater contribution of nitric oxide to vasodila-tion in these individuals; changes in total peripheralresistance were similar between low and high MSNAsubjects.24 A possible explanation might be that thatthe role of nitric oxide in buffering the variability inresting sympathetic nerve activity is quantitativelymuch less important than that of variability in car-diac output itself.

Modifiers of MSNA. Since MSNA displays cardiacrhythmicity, the maximum outflow of MSNA burstsfrom the central nervous system is one burst forevery cardiac cycle. Although a change of the num-ber of bursts/min (burst frequency) or a change ofburst strength are likely to induce changes of musclevasoconstriction, it is unclear whether the relation-ships are linear. One difficulty is that a change ofburst frequency may be due to a change of thenumber of bursts/100 heart beats (burst incidence)or to a change of heart rate. Since burst duration isprolonged in cardiac intervals of long dura-tion,209,215 it is unclear whether the two ways ofchanging burst frequency will lead to identicalchanges of the number of vasoconstrictor impulsesand the degree of vasoconstriction. A second uncer-

tainty is that changes of single fiber firing frequen-cies may be due to a change in the probability offiring or a change in the probability of multiplefiring.117,137 Multiple spikes in a single heartbeat(burst) increase the irregularity of firing and (for agiven number of spikes) this may increase the de-gree of vasoconstriction.144

Short-Term Modifiers. Posture. Movement fromsupine to seated to standing posture results in pro-gressive increases in sympathetic neural activity.15

These increases are part of normal baroreflex re-sponses aimed at the maintenance of cerebral per-fusion pressure during gravity-induced decreases invenous return. Usually, assumption of the uprightposture does not lead to significant changes in arte-rial pressure; this is particularly true if the individualcan move his or her legs, in which case the pumpingaction of leg muscles minimizes the decrease of ve-nous return. The increase in MSNA between lyingand sitting is a combined effect of unloading of thearterial baroreceptors in the carotid sinuses and un-loading of central volume receptors (cardiopulmo-nary receptors), whereas the increase between sittingand standing is thought to be predominantly a car-diopulmonary reflex effect.

Experimentally, the sympathetic neural responsesto posture/orthostasis are often evaluated using eitherhead-up tilt or lower body negative pressure (LBNP),with simultaneous recording of MSNA. LBNP results inpooling of blood in the lower extremities and can beapplied in a graded fashion to quantify baroreflexresponsiveness. Both head-up tilting121 and LBNP190

result in increases in MSNA that are dependent onboth the severity (degree of tilt or amount of negativepressure) and duration of the perturbation. Recentdata show that arterial baroreflex sensitivity increasesduring head-up tilt163 and, furthermore, during sus-tained upright posture, there is a continuous increaseof MSNA that is related to a progressive decrease instroke volume.63

There has been some debate regarding the reflexinfluences of cardiopulmonary baroreceptors onsympathetic control of the circulation. In short, thecontroversy relates to the interpretation of data fromstudies in which sympathetic reflex responses oc-curred in the absence of obvious changes in arterialpressure. Thus, when low-level LBNP was applied,peripheral vasoconstriction occurred prior to detect-able changes in arterial pressure, suggesting involve-ment of cardiopulmonary receptors sensing changesin central blood volume.95 Later studies have found,however, that even if mean blood pressure does notchange, there may be changes of pulse pressure orstroke volume,113,198 raising the possibility that defor-

FIGURE 3. Relationship between resting MSNA and vascularadrenergic responsiveness (FBF � forearm blood flow) to threedoses of tyramine, shown as a function of the arterio-venous (AV)difference in norepinephrine (NA) at each dose. Note that foreach level of endogenous norepinephrine, the high MSNA grouphad less vasoconstriction than the group with low MSNA. FromCharkoudian et al., 2006; reprinted with permission.


mation of arterial pressure receptors nevertheless isresponsible for the response.

Another important factor controlling MSNA withchanges in posture is the vestibulosympathetic re-flex. This reflex has been studied extensively by Rayand colleagues using a head-down rotation modelduring simultaneous measurement of MSNA. In-creases in MSNA, but not of SSNA, occur duringhead-down rotation in the prone position.153,156,176

These reflex increases appear to be mediated viaengagement of the otolith organs. For further reviewof this topic, the reader is referred to Ray and Mona-han156 and Ray and Carter.153

Breathing. Respiration has complex effects onMSNA. An early observation was that bursts are morelikely to occur during expiration than during inspi-ration.76 Subsequently, Eckberg et al.47 demon-strated that MSNA is highest at end-expiration andlowest at end-inspiration during normal breathing inhealthy subjects. Seals et al.171 extended these find-ings and reported that the modulatory effect of nor-mal tidal breathing is enhanced during deep, low-frequency breathing, and that both the starting lungvolume and the rate of change of lung volume influ-ence the extent to which MSNA changes in a givenbreath. A later study from the same group reportedthat approximately 70% of MSNA in a given breathoccurs when the lung is at lower volumes—either inthe first half of inspiration or in the latter half ofexpiration.172

Part of the within-breath influence is thought tobe related to changes in intrathoracic pressure, re-sulting from changes in pulmonary transmural pres-sure, which alter venous return and therefore influ-ence baroreceptor afferent firing. Additionally, datafrom lung transplant patients indicate that impor-tant inhibitory influences of lung inflation comefrom pulmonary vagal afferents.107,172 Furthermore,respiratory blood pressure variations may influenceMSNA via arterial baroreflex mechanisms. Finally,static increases in lung volume cause sustained in-creases in MSNA, thought to be due to unloading ofcardiopulmonary baroreceptors and not to changesin arterial pressure.119

MSNA also responds to chemoreceptor activa-tion80,140,185; thus, any changes in breathing patternwhich alter arterial partial pressure of oxygen orcarbon dioxide will substantially alter firing of mus-cle sympathetic nerves. Hypercapnia is a strong stim-ulus for sympathoexcitation. For example, Somers etal.185 showed that although both isocapnic hypoxiaand hyperoxic hypercapnia caused increases inMSNA, the increases were twice as great in the hy-percapnic condition. Halliwill et al.80 found that iso-

capnic hypoxia increased resting MSNA and bloodpressure, and reset baroreflex control of MSNA tohigher blood pressures, although sensitivity ofbaroreflex control was not altered.

Most respiratory studies relate to dynamic mod-ulation of MSNA, but recently it was shown thathealthy subjects with high respiratory rates havehigher long-term levels of resting MSNA than sub-jects with low respiratory rates.142 Another interest-ing observation is that 4 weeks stay at an altitude of5,250 m leads to a marked increase of resting MSNA,an effect which is still present 3 days after return tosea level.81 Along the same lines, even short, re-peated periods of hypoxia have been found to leadto prolonged increases of resting MSNA.34,116 Theunderlying mechanisms are unclear.

Exercise. In addition to mechanoreceptors di-rectly involved in the muscle contractions, skeletalmuscles also contain mechano- and chemoreceptorsof importance for blood flow and blood pressureregulation. The chemoreceptors are activated by in-tramuscular acidosis and accumulation of severaltypes of metabolites, and the mechanoreceptors bymechanical deformation. In addition, if the intra-muscular temperature increases during exercise, thismay also augment the reflex increase of sympatheticactivity, presumably by a sensitization of the intra-muscular afferent nerve endings.154 The action po-tentials are conveyed in group III and IV muscleafferents to the spinal cord and the brainstem, wherecentral sites of integration of cardiovascular effectsduring exercise are localized. Details are providedelsewhere.37,102,178,182

Exercise is associated with redistribution of bloodflow and increases in cardiac output, vascular resis-tance, and blood pressure and, depending on inten-sity and type of exercise, reflex mechanisms andcentral command contribute to different degrees.During isometric hand muscle contractions there isa successive increase of MSNA,123 which starts after30–60 s and is due primarily to activation of intra-muscular chemo- and mechanoreceptors (the so-called exercise pressor reflex). In addition, animalstudies have demonstrated increases in sympatheticactivity in cardiac124 and renal203 nerves. Centralcommand (which refers to an activation of brain-stem cardiovascular centers occurring in parallelwith the activation of the motor centers), by contrast,causes only small increases of MSNA during sus-tained handgrip contractions. In contrast, the in-crease of MSNA during intermittent hand contrac-tions is to a large extent due to activation via centralcommand. This was shown by an elegant experimentin which the contribution of muscle chemoreceptors


was eliminated/minimized by blocking the muscularcontractions with curare.204

At one time it was thought that arterial barore-flex control of sympathetic nerve activity was atten-uated during exercise, but the finding of exercise-induced increases in both MSNA and bloodpressure123 confirmed previous indirect evidencethat, instead, baroreflex control of MSNA is reset toa higher operating range of blood pressures.Whether, in addition, the sensitivity of the barore-flex is altered during exercise is unclear. Using necksuction and neck pressure to alter carotid barorecep-tor firing during one-legged dynamic leg exercise,Keller et al.104 found an attenuation of the vascularresponse to MSNA in the exercising leg but no dif-ferences in overall carotid baroreceptor control ofblood pressure during exercise compared to restingconditions. By contrast, data obtained during iso-metric handgrip contractions by Ichinose et al.88–90

indicate that there is an increased sensitivity of arte-rial baroreflex control of MSNA.

In contrast to the acute effects of exercise, moststudies of resting MSNA or sympathetic responses tovarious perturbations have found little or no effectof exercise training.28,169,173,193

Mental stress. In conscious subjects, a sensorystimulus causing arousal is known to induce sympa-thetic activation and vasoconstriction in several vas-cular beds, the result being a transient blood pres-sure increase. In MSNA, however, such stimuliinhibit one or two sympathetic bursts43 which, pre-sumably, leads to an increase of muscle blood flow.This sympathoinhibition occurs in �50% of healthysubjects, and, in a given individual, the effect isreproducible over several months,44 suggesting thatit is a robust phenomenon, characteristic for theindividual. In syncope patients who are phobic toblood and injury the inhibition is exaggerated com-pared to that seen in nonphobic syncope patients orhealthy controls.45 This raises the possibility that the“normal” interindividual variability also is related tointerindividual differences in personality.

The sympathoinhibitory effect of sensory stimuliis most likely a central neural response, since thelatency to the inhibition is too short for a peripheralreflex effect. Since the inhibition occurs primarily ifthe stimulus is delivered during a time window cor-responding to the systolic pressure wave, the sensorystimulus probably potentiates the effect of the affer-ent baroreceptor discharge. Also in support of aninteraction between sensory and afferent barorecep-tor activity is the observation that after acute barore-ceptor deafferentation an arousal stimulus evokes aburst in MSNA,61 in contrast to the inhibition seen

when baroreceptors are intact.43 The sympathoinhi-bition in response to sensory stimuli has cardiovas-cular counterparts: in subjects without inhibition thestimulus induces a greater blood pressure increasethan in subjects with inhibition.44 The functionalsignificance of the response is most likely “prepare tofight or flee!”. Large increases of blood flow to theperipheral musculature are required during ener-getic movement159 and an early start of muscle vaso-dilatation may increase the chance of a successfulfight or flight.

Hemodynamic studies have shown that experi-mentally induced emotional stress leads to vasodila-tion in the forearm but not in the calf.6,11,39,78,163 Theunderlying mechanisms probably involve a combina-tion of sympathetic withdrawal and nonsympatheticnitric oxide–mediated vasodilation, the relative con-tributions of which may vary among individuals. Inthis context, microneurographic studies have pro-duced conflicting results. In one study in whichstress was induced by a few minutes of mental arith-metic or the Stroop color word conflict test, therewas a reduction of leg MSNA (peroneal nerve) overthe initial 30–60 s followed by an increase, thestrength of which was influenced by task difficultyand the subject’s emotional state.17 Halliwill et al.78

reported that sympathetic nerve activity to the fore-arm (radial nerve) also decreased during mentalstress. Those investigators also showed that substan-tial vasodilation remained during blockade of adren-ergic neurotransmission (bretylium plus phentol-amine) and during stellate ganglionic block. Earlierwork from Dietz et al.39 suggested that a major mech-anism for vasodilation during mental stress is localnitric oxide release.

In contrast to the above, increases in leg MSNA(peroneal nerve) during mental stress have beenobserved in other studies.4,21,83 In the study by Carteret al.21 the increase of activity was potentiated bysimultaneous stimulation of vestibular receptors byhead-down rotation. In a more recent study from thesame laboratory, however, leg MSNA did not in-crease during a 5-min period of mental arithmetic.20

Simultaneous recordings of MSNA in arm andleg nerves have also given conflicting results. In onestudy, arm and leg MSNA differed markedly with nochange in the arm and a successive increase in theleg,4 whereas in the other study,20 MSNA did notincrease in either nerve.

Several factors may contribute to the discrepan-cies among studies. From an experimental point ofview, emotional reactions often include a varyingdegree of muscle tension and body movements. As aresult, the change in MSNA may be a combined


effect of a primary autonomic reaction and a reac-tion related to the muscle contractions. The propor-tions of the two effects are likely to vary amongstudies. In addition, EMG or movement artifacts mayoccur in the neurogram resulting in a low yield ofreliable quantitative data from the microneuro-graphic recordings. Furthermore, the vasodilator re-sponse to mental stress is highly variable amongindividuals.39 Finally, from a functional point of view,there may well be clear interindividual differencesbetween the “responsiveness” to experimentally in-duced stress: a test that is stressful to some subjectsmay not be so to others. Such differences are likelyto result not only in differences in motor and auto-nomic reactions but also in differences in plasmaadrenaline concentrations, which is another factorknown to affect MSNA.151

Sleep. During normal, non–rapid eye movementsleep, blood pressure and heart rate decrease succes-sively, and there is a concomitant decrease in MSNA.Rapid eye movement sleep, however, is associatedwith a clear increase of sympathetic activi-ty.86,133,147,174 In addition, a K-complex occurringduring stage 2 sleep is associated with a distinct burstin MSNA followed by a transient blood pressurepeak.86,147,196 This is interesting because K-com-plexes are considered to be indicators of arousal. Inagreement with this, an auditory stimulus deliveredduring sleep evokes one or two strong MSNAbursts174 and a pressor response that is primarilycaused by increases in peripheral vascular resis-tance.133 This arousal-induced sympathetic excita-tion during sleep is in marked contrast to the arous-al-induced inhibition that occurs in the awakestate.43 The reason for the excitation during sleepmay be a transient weakening of arterial baroreflexinhibition. In the study of Morgan et al.,133 periodsof disordered breathing or apnea caused similaracute sympathoexcitation. It is unclear whetherthese effects of arousal during sleep contribute tothe chronic sympathoexcitation seen in conditionssuch as obstructive sleep apnea (discussed below).

Long-Term Modifiers. Aging. Normal humanaging is associated with progressive increases in rest-ing activity of sympathetic nerves to skeletal muscle,heart, and the splanchnic area (but not to the kid-ney).170 Fagius and Wallin58 estimated the increasein resting MSNA to be about 1 burst/min per year.The age-related increase occurs in both men andwomen; in fact, women seem to have more markedincreases in MSNA with aging,126,139 which may berelated to the fact that younger women have, onaverage, lower levels of resting activity thanmen.143,175 (As the number of bursts cannot increase

above 100 bursts/100 heart beats, starting at a lowernumber means a greater capacity to increase withage.) Since subjects above (but not below) age 40years show a significant correlation between restinglevels of MSNA and blood pressure139 (Fig. 4), itseems likely that the age-related increase of MSNAcontributes to the increase of blood pressure. How-ever, the age-related increase in sympathetic nerveactivity is not necessarily associated with hyperten-sion, perhaps because peripheral vascular respon-siveness to alpha-adrenergic stimulation is decreasedin older men.36,40 Additionally, there is an age-re-lated inverse relationship between MSNA and serumlevels of insulin-like growth factor I,194 the signifi-cance of which is unclear.

Altered baroreflex control mechanisms may beresponsible for the increase in sympathetic neuraloutflow seen in older humans. Initial studies (basedon the modified Oxford technique) indicated thatbaroreflex sensitivity was unchanged in older sub-jects.46,125 However, older subjects have reduced re-sponsiveness to vasoactive substances,40,97 which wasan important confounding factor in those studies.To minimize this problem, Jones et al.96 compared

FIGURE 4. Relationships between MSNA and mean blood pres-sure (MAP) in men and women of different ages. Note the lack ofrelationship between MAP and MSNA in both men and womenbelow the age of 40. In contrast, there is a positive relationshipbetween these two variables in both sexes over the age of 40.From Narkiewicz et al., 2005; reprinted with permission of theAmerican Heart Association.


increases of blood pressure induced by phenyleph-rine, before and after ganglionic blockade with tri-metaphan (i.e., with and without the presence ofsympathetic nerve activity). Using the difference inpressor responses as a measure of the degree ofbaroreflex inhibition of MSNA, they found clearevidence of reduced inhibition in older subjects.

Another possible mechanism underlying the age-related increase of MSNA is an increased centralsympathetic drive. There is experimental supportalso for this alternative: when comparing healthymen age 20–30 years and 60–75 years, subcorticalbrain noradrenaline spillover is almost three timeshigher in the older group.51 Why the central sympa-thetic drive increases is unclear but lifestyle-relatedstressors may contribute. Consistent with this idea,Timio et al.199 showed that nuns living in a secludedenvironment do not develop an age-related bloodpressure increase and live longer than a well-matched control group, living in an ordinary West-ern-style society in the same geographical area.

Sex and female reproductive hormones. In general,young women appear to have lower resting MSNAthan men,54,98,143,175 although some have found nodifference between sexes.64 Some variability may berelated to menstrual-cycle phase, as women in themidluteal phase, when estrogen and progesteroneare high, have higher resting MSNA compared tothe same women studied in the early follicularphase, when these hormones are low.129 The possi-bility that elevated reproductive hormones may in-crease MSNA in women is also supported by studiesshowing sympathetic hyperactivity during normalearly and late pregnancy.72

There is also evidence that baroreflex control ofsympathetic nerve activity is affected by reproductivehormone status during the menstrual cycle. Thus,the sensitivity of baroreflex control of MSNA wasgreater in the midluteal than early follicularphase.129 In contrast, intake of estrogen and proges-tin in the form of oral contraceptives was not asso-ciated with increased baroreflex sensitivity.130 Thedifference between exogenous and endogenous hor-mones may have been due to differences in relativebioactive concentrations of hormones.

Acute sympathoexcitation during static handgripexercise is weaker in women than men.54,98 To eval-uate whether estrogen might be responsible for thegender differences, Ettinger et al.55 comparedwomen in the early follicular (low estrogen and pro-gesterone) and preovulatory (high estrogen, lowprogesterone) phases of the menstrual cycle. Therewas no difference in baseline MSNA, but the preovu-latory phase was associated with attenuated sympa-

thoexcitatory responses. In contrast to the findingsin static exercise, the increase in MSNA during dy-namic handgrip exercise showed no gender differ-ence.54,98 Furthermore, nonexercise sympathoexcit-atory responses (e.g., evoked by cold pressor test ormental stress) have not been found to differ betweenthe sexes.54,98

Race. Although much remains unknown aboutinfluences of race on control of MSNA, availableevidence suggests some differences among racialgroups. Pima Indians were found to have lower rest-ing MSNA than age- and weight-matched Caucasiansubjects.187,197 This is in contrast to black men, whohave, on average, higher resting MSNA compared towhite men and women.1 Although in whites obesityis associated with an increase in MSNA, lean blackmen were shown to have levels of MSNA that weresimilar to obese black men and women, and to obesewhite men and women.1

Acute increases in MSNA during cold pressortests were greater in normotensive blacks thanwhites.16 The differences in resting MSNA, and inacute sympathoexcitation, may be relevant to thelikelihood of developing hypertension, which isknown to be lower in Pima Indians and higher inblacks compared to Caucasians. In this context, it isalso of interest that vascular responsiveness to sym-pathetic activity may be augmented in black com-pared to Caucasian subjects.155

Pathophysiology. Congestive Heart Failure. Con-gestive heart failure is associated with marked in-creases in resting MSNA.62,115 At the single-fiber levelthere is an increased firing frequency due to anincreased probability of firing, whereas the probabil-ity of multiple firing is normal.118 The degree ofincrease of multiunit activity is related to the level ofthe heart failure, and in severe cases a burst inci-dence of 100 bursts/100 heart beats (one burst foreach cardiac cycle) is not uncommon. Burst inci-dence, however, is not a sensitive indicator of theseverity of heart failure; very high values may occuralso in patients with mild or moderate failure. There-fore, new methods of analyzing the neurogram havebeen developed, which provide more sensitive mea-sures for interindividual comparisons in patientswith high burst incidences.146,195

Studies of norepinephrine spillover have indi-cated that sympathetic nerve activity to the heart andthe kidney are markedly elevated as well.82 The in-crease in SNA to the heart precedes the increases ofMSNA and renal noradrenaline spillover.162 Thus,the heart is exposed to an increased noradrenalineload for a longer time than other tissues and in-


creased cardiac noradrenaline spillover is a main riskfactor for patients with severe cardiac failure.103

Such findings have provided the mechanistic basisfor the utility of �-adrenergic blockade in these pa-tients, which decreases mortality by up to 30%.52

Several mechanisms may contribute to the highlevels of sympathetic activity. Initially, a reduced ar-terial baroreflex sensitivity was thought to be impor-tant, but current opinion holds that the main defectis impaired cardiopulmonary reflex regulation.62

Some patients with heart failure have sleep apnea,which may contribute to the increase of MSNA alsoin the awake state.186 Furthermore, there is evidencefor exaggerated sympathetic responses to rhythmicexercise in heart failure, presumably because of in-creased accumulation of metabolic byproducts (mus-cle hypothesis).178

Cardiac transplantation in heart failure patients re-sults in a rapid reduction of resting MSNA: nerve ac-tivity was shown to be decreased as soon as 1 monthafter surgery and remained at the same (lower) levelafter 12 months, regardless of whether the patientsdeveloped posttransplant hypertension.161

Hypertension. Essential hypertension. Most, butnot all, studies of sympathetic nerve activity in essen-tial hypertension indicate increased resting activityof sympathetic nerves innervating skeletal muscle(i.e., MSNA),3,69,165,220 heart, and kidneys.165

With regard to our recent findings of a balance offactors contributing to normal blood pressure regula-tion (i.e., MSNA, cardiac output, and adrenergic re-sponsiveness),25,26 it is likely that individuals with essen-tial hypertension are “out of balance” with regard toone or more of these factors. For example, restingMSNA exhibits a wide range in normotensive individ-uals, such that a high level of resting MSNA cannot bedefined as “pathological.” However, in a hypertensiveindividual the level of MSNA may be too high whenviewed in the context of cardiac output and vascularadrenergic responsiveness. Taking all elements of thecardiac output–MSNA balance into account may alsohelp to explain previously contradictory reports re-garding whether individuals with hypertension havehigher sympathetic nerve activity compared to healthycontrols.75,158,168

In patients with essential hypertension, the sym-pathetic nervous system appears to have an exagger-ated responsiveness to chemoreflex activation viahypoxia or hypercapnia.141,184 This was most strikingduring voluntary apnea, when the increase in MSNAin hypertensives was shown to be 12 times that seenin healthy controls.184 In borderline hypertensivepatients, increased MSNA responses to low-grade

LBNP have also been reported and interpreted as anincreased sensitivity of cardiopulmonary reflexes.158

Several mechanisms may contribute to the initi-ation of chronic hypertension, and genetic factorsare likely to be important. Available evidence, how-ever, is difficult to interpret. For example, offspringof hypertensive patients and normotensive controlswere found to have similar resting levels of MSNAbut the hypertensive offspring had stronger MSNAresponses to mental stress than the control group.145

However, the cold pressor test evoked weaker bloodpressure and MSNA responses in subjects with astrong family history of hypertension.114 Interest-ingly, some patients with essential hypertension havebeen reported to have neurovascular compression atthe rostral ventrolateral medulla and increased levelsof MSNA at rest.167 As neurovascular compression withincreased MSNA has also been found in normotensivesubjects,181 it remains unclear whether there is a patho-genetic link to essential hypertension.

Renovascular hypertension. Patients with renovas-cular hypertension due to renal artery stenosis haveincreased resting MSNA.131 Johansson et al.92 inves-tigated a large group of patients with microneurog-raphy and the noradrenaline spillover techniqueand found elevated levels of both MSNA and totalbody noradrenaline spillover in these patients.There is also a report of increased cardiac norepi-nephrine spillover in such patients.152

Increased sympathetic nerve activity, however, isnot limited to renal artery stenosis: patients withpolycystic kidney disease and hypertension have alsobeen found to have increased MSNA,111 regardlessof functional status of the kidney. Similarly, restingMSNA was increased in hypertensive patients withchronic renal failure with native kidneys but theincrease was not present in patients who had under-gone bilateral nephrectomy.27

The mechanism underlying the increased sympa-thetic drive in renal disease is likely to be related tothe renin-angiotensin system. Consistent with thisidea is the evidence that both angiotensin-convert-ing enzyme inhibitors and angiotensin II blockershave been found to reduce resting levels of bloodpressure and MSNA.110

Other forms of hypertension. Increases of MSNAhave been reported in hypertension during pregnan-cy,74,166 preeclampsia,73,166 and pulmonary artery hy-pertension,202 but not in primary hyperaldosteron-ism with hypertension.131

Obstructive Sleep Apnea. Obstructive sleep apneais associated with large elevations in resting MSNA.18,183

This is true both during sleep and while individuals areawake during the day, despite normal arterial oxygen


saturation and normocapnia during wakefulness.183

There is also evidence of impaired arterial baroreflexcontrol of sympathetic19 and vagal13 activities. Themechanism underlying the chronic sympathoexcita-tion may be related to the apnea-related episodes ofhypoxia and hypercapnia. Initial evidence came fromfindings of prolonged sympathoexcitation after hypox-ia,219 and recently Monahan et al.132 showed that re-peated end-expiratory apneas led to prolonged in-creases of MSNA that were due to a shift of thebaroreflex operating point to higher blood pressures.Interestingly, acute or chronic treatment with contin-uous positive airway pressure at night reverses both thesympathoexcitation138 and the hypertension217 associ-ated with obstructive sleep apnea. For a review of sym-pathetic mechanisms in obstructive sleep apnea, seeNarkiewicz and Somers.140

Studies of single unit sympathetic neural activityin obstructive sleep apnea patients have shown thatthe chronic sympathoexcitation in these patientsmight involve different mechanisms from those incongestive heart failure.49 Although both patientpopulations show similar increases in probability offiring of individual neurons (�50% compared to�30% in normal subjects), patients with obstructivesleep apnea also exhibit increased probability ofmultiple firing of a single neuron within an individ-ual cardiac cycle.

Syncope. Syncope is a sudden transient loss ofconsciousness and postural tone due to cerebral hypo-perfusion and may have multiple causes. An occa-sional, neurally mediated syncope (often called vasova-gal syncope) may be triggered in otherwise healthysubjects by different stressors (e.g., by stimulation ofsensory or visceral afferents, changes of posture, oremotional reactions) and is characterized by an abruptcessation of sympathetic nerve activity,180,212 causingmarked peripheral vasodilation.

Withdrawal of sympathetically mediated vasocon-striction is thought to be a main reason for theprofound systemic vasodilation that leads to hypo-tension and subsequent syncope, but it is probablynot the only mechanism.7 Another possibility is thatcirculating epinephrine contributes to vasodilationvia activation of vascular beta-adrenergic receptors.However, Dietz et al.38 showed that neither totaladrenergic blockade (� � �) nor L-NMMA (� nitricoxide synthase inhibition) blocked the large vasodi-lator response observed during syncope induced byprogressive LBNP. Their data are consistent with theidea that sympathetic withdrawal alone does not me-diate the entirety of the peripheral vasodilator re-sponse during syncope, but additional mechanismsremain unclear.

The mechanisms behind the abrupt withdrawalof sympathetic vasoconstrictor activity during syn-cope are also incompletely understood77,135 and in-terindividual differences have been reported, espe-cially during experimentally induced syncopalreactions.134 Patients with recurrent orthostatic syn-cope have been found to have reduced arterialbaroreflex sensitivity during head-up tilt.29 Thisagrees with a case report by Iwase et al.91 of alteredproperties of MSNA bursts (variable durations andbaroreflex latencies) prior to syncope. Similarly, Ka-miya et al.99 found evidence of altered spectral char-acteristics of the MSNA signal during the 60 s pre-ceding a syncopal event. In a recent study, patientswith recurrent syncope who were phobic to blood/injury were found to have prolonged MSNA inhibi-tions in response to surprising somatosensory stimu-li.45 The findings are compatible with altered centralnervous mechanisms but the specific explanation ofwhy a fully compensated circulation is transformedto a highly unstable system is still unknown.

Syncope with sudden withdrawal of MSNA hasalso been reported after carotid sinus massage30,180

and glossopharyngeal neuralgia with syncope.214 Inthese cases, a likely mechanism is that an exagger-ated afferent discharge from the arterial barorecep-tors causes vasodilatation, bradycardia, or asystole.Syncope may also occur in the rare condition ofdopamine beta-hydroxylase deficiency.109,157 Suchpatients suffer from severe orthostatic hypotensionbecause they lack the enzyme activity necessary forconversion of dopamine to noradrenaline. The re-sult is that total body noradrenaline spillover is verylow and neurally induced vasoconstriction virtuallynonexistent. Interestingly, they have high resting lev-els of MSNA and qualitatively normal responses toprovocations.109,157

Ischemic Heart Disease. The effects of ischemicheart disease and myocardial infarction on MSNAwere studied in a series of articles by Mary andcoworkers.66,67,85 Patients with stable coronary ar-tery disease had normal MSNA; however, afteruncomplicated acute myocardial infarction MSNAwas elevated for several months compared bothwith normal controls and patients with stable cor-onary artery disease.67 Patients with unstable an-gina had increased MSNA compared to healthycontrols but less so than after myocardial infarc-tion.66 Patients who had been hypertensive priorto the myocardial infarction had greater and morelong-lasting sympathetic hyperactivity than pa-tients who had been normotensive prior to theinfarction.85 Whether the level of MSNA during amyocardial infarction or the time course of reduc-


tion after infarction has prognostic value is un-clear.

Obesity and Diabetes. The relationships amongobesity, metabolic hormones, and sympathetic nerveactivity have attracted considerable interest in recentyears, from both physiological and clinical points ofview.35,56,136 Increases in body weight and associatedadipose tissue accumulation are associated with in-creases in resting MSNA in human subjects164; thischronic sympathoexcitation is thought to contributeto the increased risk for cardiovascular disease inobese individuals.2,70,71 Sympathoexcitation in obesesubjects likely also influences fat metabolism: stimu-lation of cutaneous sympathetic nerves was shown tostimulate lipolysis in a group of healthy women.42

Normotensive obese individuals have been shown tohave MSNA that is, on average, more than twice ashigh as normotensive lean individuals matched forage.70 This elevated MSNA was reduced by a 16-weekhypocaloric diet that resulted in weight loss in obeseindividuals.71 Similar findings were obtained in agroup of obese women with borderline hyperten-sion.5

The mechanisms for obesity-associated sympa-thoexcitation may involve blunted baroreflex-medi-ated sympathoinhibition in obese subjects, whichcould contribute to higher resting levels of activity.70

Interestingly, the influence of obesity appears to bespecific to the type of obesity: accumulation of vis-ceral fat, which carries a greater risk for develop-ment of cardiovascular disease, is also associated withgreater increases in total MSNA compared to indi-viduals with less abdominal visceral fat.2 This obser-vation may also provide further insight into mecha-nisms of obesity-mediated sympathoexcitation. Sinceleptin is released from fat cells, and the total adiposetissue mass was similar in the two groups, the totalleptin levels were similar between the two groups.Therefore, the higher MSNA seen in the group withvisceral obesity must have been due to some mech-anism other than leptin. The relationship betweensympathetic nerve activity and leptin is reviewed byEikelis and Esler.48

Obesity is associated with an increased risk ofmetabolic diseases, in particular type 2 diabetes.Type 2 diabetes is associated with significant in-creases in resting MSNA, which may contribute to oraugment the cardiovascular risk associated with con-ditions such as hypertension, which often occur inindividuals with type 2 diabetes.87 The mechanismsfor the changes in type 2 diabetes are unknown, butmay involve an influence of insulin to cause sympa-thoexcitation. For example, euglycemic hyperinsu-linemia was associated with progressive increases in

MSNA in healthy control subjects.8 In contrast, type1 diabetes is associated with a significant decrease inthe number of bursts, by about half,84 and the repro-ducibility of MSNA measurement appears to remainconsistent with that in nondiabetic control subjects.

Polyneuropathy is a common complication ofdiabetes. If a patient has diabetic polyneuropathy, amethodological difficulty is that failure to find sym-pathetic activity in microneurographic recordings ismuch more frequent than in healthy subjects.60 Pre-sumably, this is because the neuropathy leads to asuccessive reduction of conducting sympatheticnerve fibers.

SKIN SYMPATHETIC NERVE ACTIVITY

Physiology. Resting Activity. Although skin sympa-thetic nerve activity (SSNA) occurs in bursts, thesebursts have a much more variable duration, they mayhave more than one peak, and although they oftendisplay a respiratory rhythmicity, they are generallynot coordinated with the cardiac cycle (i.e., arterialbaroreflex influence is weak or absent). The variabil-ity makes it difficult both to define the start and endof some bursts and to evaluate the significance ofdifferences in number of bursts. Furthermore, un-like MSNA, which includes only vasoconstrictornerve fibers, SSNA can encompass any of four nervetypes: vasoconstrictor, vasodilator, sudomotor, andpilomotor. This makes it more challenging to inter-pret measurements of SSNA as they relate to down-stream effector function.

In this context, some investigators have quanti-fied differences among different types of skin sym-pathetic bursts. The results show that, on average,sudomotor bursts have higher conduction veloci-ties59 and shorter duration10 than vasoconstrictorbursts. However, the variability among bursts resultsin a large overlap in duration between the two bursttypes, such that this parameter cannot be used for areliable distinction between sudomotor and vasocon-strictor bursts.

The innervation of human skin by sympatheticvasodilator nerves has been recognized since the1930s, when Grant and Holling68 demonstrated thatnerve blockade prevented the large increases in skinblood flow seen during core hyperthermia. Micro-neurographic recordings, however, have providedlittle information on the character and control ofthe vasodilator impulses. Intraneural microstimula-tion has shown that efferent sympathetic activity caninduce cutaneous vasodilation,12 but whether theeffect was due to activity in sudomotor fibers or aseparate set of vasodilator fibers is unclear. Sugenoya


et al.189 studied the temporal relationship betweenSSNA bursts in the peroneal nerve and succeedingsigns of vasodilation and sweating on the dorsum ofthe foot in mildly heated subjects. They reportedthat 75% of all bursts contained both vasodilator andsudomotor impulses. In a similar way, bursts in non-heated subjects very often contain a mixture of su-domotor and vasoconstrictor impulses. Such find-ings reinforce the idea that quantitative data on therelationship between SSNA and blood flow are diffi-cult to interpret.

Some studies have attempted to compare restinglevels of SSNA between groups of subjects. The in-terpretation of such results is also difficult. SSNA iseasily activated by all kinds of sensory stimuli,arousal, and perceived stress and, therefore, record-ings have to be made in a relaxed environment inwhich all sensory stimuli are minimized. In addition,SSNA has a thermoregulatory function, which meansthat the strength and composition of the multiunitactivity is very sensitive to changes of ambient andinternal temperatures. Therefore, before resting lev-els of SSNA can be compared, subjects have to becompletely calm and fully adapted to a standardizedthermal setting, a precaution which rarely has beenachieved.

As a consequence of these difficulties, most sig-nificant advances in our understanding of neuralcontrol of skin blood flow have been obtained bynonneural measurements and pharmacological orother interventions (e.g., local bretylium treatmentto specifically block neurotransmission from norad-renergic nerve terminals).

Modifiers of SSNA. Body temperature. In hu-mans, the skin is a major effector organ for physio-logical thermoregulation: increases in skin bloodflow and sweating during hyperthermia representthe major avenue of heat dissipation, and decreasesin skin blood flow are important for decreased heatloss during body cooling. Thermoregulatory sweat-ing is mediated by sympathetic cholinergic innerva-tion of eccrine sweat glands in the skin. Vasoconstric-tion during body cooling and vasodilation duringbody heating are primarily neurally mediated events,with some role for local effects of temperature aswell.22,94 Thus, it makes sense that measurement ofSSNA shows increases in activity during whole-bodyheating, presumably reflecting increases in sudomo-tor or active vasodilator activity to the skin.10,218 Mea-surement of skin blood flow in combination withnerve blockade or local bretylium for blockade ofnoradrenergic neurotransmission have indicatedthat active cutaneous vasodilation is responsible for

as much as 80%–90% of the substantial cutaneousvasodilation observed during whole-body heating.93

SSNA also increases during body cooling, reflectingincreases in vasoconstrictor activity.10,32 Interestingly,the overall extent of increase in SSNA is often similarbetween whole-body heating and whole-body cooling.During body cooling, burst duration is longer, whichmay be due in part to slower conduction velocity invasoconstrictor than in sudomotor fibers.59 In addi-tion, the spectral characteristics of SSNA exhibit differ-ences between normothermia/cooling and whole-body heating conditions,32 which may reflectdifferential central modulation of activity in vasocon-strictor, sudomotor, and vasodilator fibers.

Simultaneous recordings of SSNA in arm and legnerves have shown that, in some nerves, there is amarked parallelism between arm and leg sympa-thetic discharges under certain thermal conditionsand differences in other nerves or thermal situa-tions.9,149

Posture and blood pressure. The challenges re-garding interpretation of SSNA are particularly evi-dent in the literature regarding SSNA responses toorthostatic and baroreflex stimuli. Measurements ofskin blood flow in both normothermic and hyper-thermic conditions have clearly shown reflex vaso-constriction in the skin during LBNP.31,33,105 In con-trast, the majority of studies have not found changesin SSNA during baroreflex stimuli including LBNP,tilt, and vasoactive drug boluses either in normother-mia or hyperthermia,33,205,218 although SSNA in onestudy was shown to decrease during both tilt andLBNP in hyperthermic subjects.41

During hyperthermia, a lack of change in SSNAmay be because of reciprocal changes in differentnerve types: decreases in vasodilator or sudomotorneural signals are offset by increases in vasoconstric-tor neural activity during LBNP. If this is the case, itremains somewhat confusing that bretylium treat-ment (which prevents noradrenergic neurotransmis-sion) does not alter the reflex cutaneous vasocon-striction that is seen during application of LBNPduring whole-body heating.31,105 This latter findingwould suggest that withdrawal of active vasodilationis primarily responsible for the cutaneous vasocon-striction.

One possibility is that there is some regionalspecificity to SSNA measurement with regard tobaroreflex influences: in all studies in which SSNAhas shown no change with baroreflex stimuli, SSNAwas measured in the leg, whereas skin blood flow isusually measured in the arm. In the only study inwhich SSNA was measured in the arm during mild tomoderate body heating combined with LBNP or


head-up tilt, a decrease in SSNA occurred duringboth maneuvers.41

Another way in which body position may affectSSNA relates to an influence of local pressure due tobody position when lying down. When applying me-chanical pressure on one side of the body, the con-tralateral/ipsilateral ratios of sweating and sudomo-tor SSNA are increased.148 The results suggested thatthe effects are due to a spinal sudomotor reflex.

Mental stress. All stimuli associated with emo-tional reactions co-activate skin vasoconstrictor andsudomotor nerve traffic; even mild stimuli, such as asurprising noise, induces a distinct burst containingboth types of impulses. In fact, startle stimuli areregularly used to test whether a putative nerve re-cording site contains SSNA. An individual’s thermalbias influences the relative proportion of the twotypes of impulses in arousal responses: vasoconstric-tor impulses dominate in a cold and sudomotorimpulses in a warm environment.10 Specific workregarding the influence of mental or emotionalstress on SSNA, however, is sparse.

Congestive heart failure. Unlike MSNA, sympa-thetic nerve activity to the skin does not appear to beelevated in patients with congestive heart failure,suggesting that this disease affects sympathetic neu-ral control differentially.69,128,177 Although SSNA atrest appears to be similar between congestive heartfailure patients and control subjects, the responsesof SSNA to metaboreflex stimulation appears to bealtered/augmented in this patient group.177 Silber etal.177 showed that increases in SSNA during rhythmichandgrip exercise were similar in congestive heartfailure and controls, but that the increases duringpost-handgrip circulatory arrest were augmented inthe patient group. This augmented responsivenesswas associated with augmented peak levels of H� andH2PO4

� in the congestive heart failure group duringpost-handgrip circulatory arrest, as assessed by nu-clear magnetic resonance spectroscopy. SSNA didnot increase during post-handgrip circulatory arrestin control subjects.177

OVERALL SUMMARY AND CONCLUSIONS

During the last decades knowledge about sympa-thetic neural function in humans has improveddramatically. Quantitative data from microneuro-graphic recordings of sympathetic traffic, primarilyin muscle nerves, and measurements of noradrena-line spillover from nerves to some visceral tissueshave provided novel information on physiologicalcontrol of the cardiovascular system under a varietyof conditions and ages. Studies of patients with dis-

eases such as hypertension and cardiac failure havecontributed to a better understanding of underlyingpathophysiological mechanisms. Thus far, most stud-ies have involved comparisons of groups of subjects,but recent results hint that, in the future, analysis ofthe variability in physiological characteristics amongindividuals may lead to further insight into centralnervous control of the circulation.25,208

Supported by Swedish Research Council Grant 12170 (to B.G.W.)and National Institutes of Health Grant HL73884 (to N.C.).

REFERENCES

1. Abate NI, Mansour YH, Tuncel M, Arbique D, Chavoshan B,Kizilbash A, et al. Overweight and sympathetic overactivity inblack Americans. Hypertension 2001;38:379–383.

2. Alvarez GE, Beske SD, Ballard TP, Davy KP. Sympatheticneural activation in visceral obesity. Circulation 2002;106:2533–2536.

3. Anderson EA, Sinkey CA, Lawton WJ, Mark AL. Elevatedsympathetic nerve activity in borderline hypertensive hu-mans. Evidence from direct intraneural recordings. Hyper-tension 1989;14:177–183.

4. Anderson EA, Wallin BG, Mark AL. Dissociation of sympa-thetic nerve activity in arm and leg muscle during mentalstress. Hypertension 1987;9:III114–119.

5. Andersson B, Elam M, Wallin BG, Bjorntorp P, AnderssonOK. Effect of energy-restricted diet on sympathetic musclenerve activity in obese women. Hypertension 1991;18:783–789.

6. Barcroft H, Brod J, Hejl BZ, Hirsjarvi EA, Kitchin AH. Themechanism of the vasodilatation in the forearm muscle dur-ing stress (mental arithmetic). Clin Sci 1960;19:577–586.

7. Barcroft H, Edholm OG. On the vasodilatation in humanskeletal muscle during post-haemorrhagic fainting. J Physiol(Lond) 1945;104:161–175.

8. Berne C, Fagius J, Pollare T, Hjemdahl P. The sympatheticresponse to euglycaemic hyperinsulinaemia. Evidence frommicroelectrode nerve recordings in healthy subjects. Diabe-tologia 1992;35:873–879.

9. Bini G, Hagbarth KE, Hynninen P, Wallin BG. Regionalsimilarities and differences in thermoregulatory vaso- andsudomotor tone. J Physiol (Lond) 1980;306:553–565.

10. Bini G, Hagbarth KE, Hynninen P, Wallin BG. Thermoreg-ulatory and rhythm-generating mechanisms governing thesudomotor and vasoconstrictor outflow in human cutaneousnerves. J Physiol (Lond) 1980;306:537–552.

11. Blair DA, Glover WE, Greenfield AD, Roddie IC. Excitationof cholinergic vasodilator nerves to human skeletal musclesduring emotional stress. J Physiol (Lond) 1959;148:633–647.

12. Blumberg H, Wallin BG. Direct evidence of neurally medi-ated vasodilatation in hairy skin of the human foot. J Physiol(Lond) 1987;382:105–121.

13. Bonsignore MR, Parati G, Insalaco G, Marrone O, Castigli-oni P, Romano S, et al. Continuous positive airway pressuretreatment improves baroreflex control of heart rate duringsleep in severe obstructive sleep apnea syndrome. Am JRespir Crit Care Med 2002;166:279–286.

14. Brengelmann GL, Savage MV. Temperature regulation inthe neutral zone. Ann N Y Acad Sci 1997;813:39–50.

15. Burke D, Sundlof G, Wallin G. Postural effects on musclenerve sympathetic activity in man. J Physiol (Lond) 1977;272:399–414.

16. Calhoun DA, Mutinga ML, Collins AS, Wyss JM, Oparil S.Normotensive blacks have heightened sympathetic responseto cold pressor test. Hypertension 1993;22:801–805.


17. Callister R, Suwarno NO, Seals DR. Sympathetic activity isinfluenced by task difficulty and stress perception duringmental challenge in humans. J Physiol (Lond) 1992;454:373–387.

18. Carlson JT, Hedner J, Elam M, Ejnell H, Sellgren J, WallinBG. Augmented resting sympathetic activity in awake pa-tients with obstructive sleep apnea. Chest 1993;103:1763–1768.

19. Carlson JT, Hedner JA, Sellgren J, Elam M, Wallin BG.Depressed baroreflex sensitivity in patients with obstructivesleep apnea. Am J Respir Crit Care Med 1996;154:1490–1496.

20. Carter JR, Kupiers NT, Ray CA. Neurovascular responses tomental stress. J Physiol (Lond) 2005;564:321–327.

21. Carter JR, Ray CA, Cooke WH. Vestibulosympathetic reflexduring mental stress. J Appl Physiol 2002;93:1260–1264.

22. Charkoudian N. Skin blood flow in adult human thermoreg-ulation: how it works, when it does not, and why. Mayo ClinProc 2003;78:603–612.

23. Charkoudian N, Halliwill JR, Morgan BJ, Eisenach JH, Joy-ner MJ. Influences of hydration on post-exercise cardiovas-cular control in humans. J Physiol (Lond) 2003;552:635–644.

24. Charkoudian N, Joyner MJ, Barnes SA, Johnson CP,Eisenach JH, Dietz NM, et al. Relationship between musclesympathetic nerve activity and systemic hemodynamics dur-ing nitric oxide synthase inhibition in humans. Am J PhysiolHeart Circ Physiol 2006;291:H1378–1383.

25. Charkoudian N, Joyner MJ, Johnson CP, Eisenach JH, DietzNM, Wallin BG. Balance between cardiac output and sym-pathetic nerve activity in resting humans: role in arterialpressure regulation. J Physiol (Lond) 2005;568:315–321.

26. Charkoudian N, Joyner MJ, Sokolnicki LA, Johnson CP,Eisenach JH, Dietz NM, et al. Vascular adrenergic respon-siveness is inversely related to tonic activity of sympatheticvasoconstrictor nerves in humans. J Physiol (Lond) 2006;572:821–827.

27. Converse RL Jr, Jacobsen TN, Toto RD, Jost CM, CosentinoF, Fouad-Tarazi F, et al. Sympathetic overactivity in patientswith chronic renal failure. N Engl J Med 1992;327:1912–1918.

28. Cooke WH, Reynolds BV, Yandl MG, Carter JR, TahvanainenKU, Kuusela TA. Effects of exercise training on cardiovagaland sympathetic responses to Valsalva’s maneuver. Med SciSports Exerc 2002;34:928–935.

29. Cooper VL, Hainsworth R. Effects of head-up tilting onbaroreceptor control in subjects with different tolerances toorthostatic stress. Clin Sci (Lond) 2002;103:221–226.

30. Costa F, Biaggioni I. Microneurographic evidence of suddensympathetic withdrawal in carotid sinus syncope; treatmentwith ergotamine. Chest 1994;106:617–620.

31. Crandall CG, Johnson JM, Kosiba WA, Kellogg DL Jr. Barore-ceptor control of the cutaneous active vasodilator system.J Appl Physiol 1996;81:2192–2198.

32. Cui J, Sathishkumar M, Wilson TE, Shibasaki M, Davis SL,Crandall CG. Spectral characteristics of skin sympatheticnerve activity in heat-stressed humans. Am J Physiol HeartCirc Physiol 2006;290:H1601–1609.

33. Cui J, Wilson TE, Crandall CG. Orthostatic challenge doesnot alter skin sympathetic nerve activity in heat-stressed hu-mans. Auton Neurosci 2004;116:54–61.

34. Cutler MJ, Swift NM, Keller DM, Wasmund WL, Smith ML.Hypoxia-mediated prolonged elevation of sympathetic nerveactivity after periods of intermittent hypoxic apnea. J ApplPhysiol 2004;96:754–761.

35. Davy KP. The global epidemic of obesity: are we becomingmore sympathetic? Curr Hypertens Rep 2004;6:241–246.

36. Davy KP, Seals DR, Tanaka H. Augmented cardiopulmonaryand integrative sympathetic baroreflexes but attenuated pe-ripheral vasoconstriction with age. Hypertension 1998;32:298–304.

37. Delp MD, O’Leary DS. Integrative control of the skeletalmuscle microcirculation in the maintenance of arterial pres-sure during exercise. J Appl Physiol 2004;97:1112–1118.

38. Dietz NM, Halliwill JR, Spielmann JM, Lawler LA, Pa-pouchado BG, Eickhoff TJ, et al. Sympathetic withdrawaland forearm vasodilation during vasovagal syncope in hu-mans. J Appl Physiol 1997;82:1785–1793.

39. Dietz NM, Rivera JM, Eggener SE, Fix RT, Warner DO,Joyner MJ. Nitric oxide contributes to the rise in forearmblood flow during mental stress in humans. J Physiol (Lond)1994;480:361–368.

40. Dinenno FA, Dietz NM, Joyner MJ. Aging and forearmpostjunctional alpha-adrenergic vasoconstriction in healthymen. Circulation 2002;106:1349–1354.

41. Dodt C, Gunnarsson T, Elam M, Karlsson T, Wallin BG.Central blood volume influences sympathetic sudomotornerve traffic in warm humans. Acta Physiol Scand 1995;155:41–51.

42. Dodt C, Lonnroth P, Fehm HL, Elam M. Intraneural stimu-lation elicits an increase in subcutaneous interstitial glycerollevels in humans. J Physiol (Lond) 1999;521:545–552.

43. Donadio V, Kallio M, Karlsson T, Nordin M, Wallin BG.Inhibition of human muscle sympathetic activity by sensorystimulation. J Physiol (Lond) 2002;544:285–292.

44. Donadio V, Karlsson T, Elam M, Wallin BG. Interindividualdifferences in sympathetic and effector responses to arousalin humans. J Physiol (Lond) 2002;544:293–302.

45. Donadio V, Liguori R, Elam M, Karlsson T, Montagna P,Cortelli P, et al. Exaggerated inhibition of sympathetic nerveactivity induced by arousal in phobic patients with vasovagalsyncope. Brain 2007 (to be published).

46. Ebert TJ, Morgan BJ, Barney JA, Denahan T, Smith JJ. Effectsof aging on baroreflex regulation of sympathetic activity inhumans. Am J Physiol 1992;263:H798–803.

47. Eckberg DL, Nerhed C, Wallin BG. Respiratory modulationof muscle sympathetic and vagal cardiac outflow in man.J Physiol (Lond) 1985;365:181–196.

48. Eikelis N, Esler M. The neurobiology of human obesity. ExpPhysiol 2005;90:673–682.

49. Elam M, McKenzie D, Macefield V. Mechanisms of sympa-thoexcitation: single-unit analysis of muscle vasoconstrictorneurons in awake OSAS subjects. J Appl Physiol 2002;93:297–303.

50. Esler M. The sympathetic system and hypertension. Am JHypertens 2000;13:99S–105S.

51. Esler M, Hastings J, Lambert G, Kaye D, Jennings G, SealsDR. The influence of aging on the human sympatheticnervous system and brain norepinephrine turnover. Am JPhysiol Regul Integr Comp Physiol 2002;282:R909–916.

52. Esler M, Lambert G, Brunner-La Rocca HP, Vaddadi G, KayeD. Sympathetic nerve activity and neurotransmitter releasein humans: translation from pathophysiology into clinicalpractice. Acta Physiol Scand 2003;177:275–284.

53. Esler MD, Jennings GL, Johns J, Burke F, Little PJ, LeonardP. Estimation of total renal, cardiac and splanchnic sympa-thetic nervous tone in essential hypertension from measure-ments of noradrenaline release. J Hypertens Suppl 1984;2:S123–125.

54. Ettinger SM, Silber DH, Collins BG, Gray KS, Sutliff G,Whisler SK, et al. Influences of gender on sympathetic nerveresponses to static exercise. J Appl Physiol 1996;80:245–251.

55. Ettinger SM, Silber DH, Gray KS, Smith MB, Yang QX,Kunselman AR, et al. Effects of the ovarian cycle on sympa-thetic neural outflow during static exercise. J Appl Physiol1998;85:2075–2081.

56. Fagius J. Sympathetic nerve activity in metabolic control—some basic concepts. Acta Physiol Scand 2003;177:337–343.

57. Fagius J, Berne C. Increase in muscle nerve sympatheticactivity in humans after food intake. Clin Sci (Lond) 1994;86:159–167.

58. Fagius J, Wallin BG. Long-term variability and reproducibil-ity of resting human muscle nerve sympathetic activity at


rest, as reassessed after a decade. Clin Auton Res 1993;3:201–205.

59. Fagius J, Wallin BG. Sympathetic reflex latencies and con-duction velocities in normal man. J Neurol Sci 1980;47:433–448.

60. Fagius J, Wallin BG. Sympathetic reflex latencies and con-duction velocities in patients with polyneuropathy. J NeurolSci 1980;47:449–461.

61. Fagius J, Wallin BG, Sundlof G, Nerhed C, Englesson S.Sympathetic outflow in man after anaesthesia of the glosso-pharyngeal and vagus nerves. Brain 1985;108:423–438.

62. Floras JS. Sympathetic activation in human heart failure:diverse mechanisms, therapeutic opportunities. Acta PhysiolScand 2003;177:391–398.

63. Fu Q, Shook RP, Okazaki K, Hastings JL, Shibata S, ConnerCL, et al. Vasomotor sympathetic neural control is main-tained during sustained upright posture in humans. J Physiol(Lond) 2006;577:679–687.

64. Fu Q, Witkowski S, Okazaki K, Levine BD. Effects of genderand hypovolemia on sympathetic neural responses to ortho-static stress. Am J Physiol Regul Integr Comp Physiol 2005;289:R109–116.

65. Gandevia SC, Hales JP. The methodology and scope of hu-man microneurography. J Neurosci Methods 1997;74:123–136.

66. Graham LN, Smith PA, Stoker JB, Mackintosh AF, Mary DA.Sympathetic neural hyperactivity and its normalization fol-lowing unstable angina and acute myocardial infarction.Clin Sci (Lond) 2004;106:605–611.

67. Graham LN, Smith PA, Stoker JB, Mackintosh AF, Mary DA.Time course of sympathetic neural hyperactivity after un-complicated acute myocardial infarction. Circulation 2002;106:793–797.

68. Grant RT, Holling HE. Further observations on the vascularresponses of the human limb to body warming: evidence forsympathetic vasodilator nerves in the normal subject. ClinSci 1938;3:273–285.

69. Grassi G, Colombo M, Seravalle G, Spaziani D, Mancia G.Dissociation between muscle and skin sympathetic nerveactivity in essential hypertension, obesity, and congestiveheart failure. Hypertension 1998;31:64–67.

70. Grassi G, Seravalle G, Cattaneo BM, Bolla GB, Lanfranchi A,Colombo M, et al. Sympathetic activation in obese normo-tensive subjects. Hypertension 1995;25:560–563.

71. Grassi G, Seravalle G, Colombo M, Bolla G, Cattaneo BM,Cavagnini F, et al. Body weight reduction, sympathetic nervetraffic, and arterial baroreflex in obese normotensive hu-mans. Circulation 1998;97:2037–2042.

72. Greenwood JP, Scott EM, Stoker JB, Walker JJ, Mary DA.Sympathetic neural mechanisms in normal and hypertensivepregnancy in humans. Circulation 2001;104:2200–2204.

73. Greenwood JP, Scott EM, Walker JJ, Stoker JB, Mary DA. Themagnitude of sympathetic hyperactivity in pregnancy-in-duced hypertension and preeclampsia. Am J Hypertens2003;16:194–199.

74. Greenwood JP, Stoker JB, Walker JJ, Mary DA. Sympatheticnerve discharge in normal pregnancy and pregnancy-in-duced hypertension. J Hypertens 1998;16:617–624.

75. Gudbjornsdottir S, Lonnroth P Sverrisdottir YB, Wallin BG,Elam M. Sympathetic nerve activity and insulin in obesenormotensive and hypertensive men. Hypertension 1996;27:276–280.

76. Hagbarth KE, Vallbo ÅB. Pulse and respiratory grouping ofsympathetic impulses in human muscle-nerves. Acta PhysiolScand 1968;74:96–108.

77. Hainsworth R. Pathophysiology of syncope. Clin Auton Res2004;14(Suppl 1):18–24.

78. Halliwill JR, Lawler LA, Eickhoff TJ, Dietz NM, Nauss LA,Joyner MJ. Forearm sympathetic withdrawal and vasodilata-tion during mental stress in humans. J Physiol (Lond) 1997;504:211–220.

79. Halliwill JR, Minson CT. Effect of hypoxia on arterial barore-flex control of heart rate and muscle sympathetic nerveactivity in humans. J Appl Physiol 2002;93:857–864.

80. Halliwill JR, Morgan BJ, Charkoudian N. Peripheral che-moreflex and baroreflex interactions in cardiovascular reg-ulation in humans. J Physiol (Lond) 2003;552:295–302.

81. Hansen J, Sander M. Sympathetic neural overactivity inhealthy humans after prolonged exposure to hypobaric hyp-oxia. J Physiol (Lond) 2003;546:921–929.

82. Hasking GJ, Esler MD, Jennings GL, Burton D, Johns JA,Korner PI. Norepinephrine spillover to plasma in patientswith congestive heart failure: evidence of increased overalland cardiorenal sympathetic nervous activity. Circulation1986;73:615–621.

83. Hjemdahl P, Fagius J, Freyschuss U, Wallin BG, Daleskog M,Bohlin G, et al. Muscle sympathetic activity and norepineph-rine release during mental challenge in humans. Am JPhysiol 1989;257:E654–664.

84. Hoffman RP, Sinkey CA, Anderson EA. Microneurographi-cally determined muscle sympathetic nerve activity levels arereproducible in insulin-dependent diabetes mellitus. J Dia-betes Complications 1998;12:307–310.

85. Hogarth AJ, Mackintosh AF, Mary DA. The sympathetic driveafter acute myocardial infarction in hypertensive patients.Am J Hypertens 2006;19:1070–1076.

86. Hornyak M, Cejnar M, Elam M, Matousek M, Wallin BG.Sympathetic muscle nerve activity during sleep in man. Brain1991;114:1281–1295.

87. Huggett RJ, Scott EM, Gilbey SG, Stoker JB, Mackintosh AF,Mary DA. Impact of type 2 diabetes mellitus on sympatheticneural mechanisms in hypertension. Circulation 2003;108:3097–3101.

88. Ichinose M, Saito M, Kondo N, Nishiyasu T. Time-depen-dent modulation of arterial baroreflex control of musclesympathetic nerve activity during isometric exercise in hu-mans. Am J Physiol Heart Circ Physiol 2006;290:H1419–1426.

89. Ichinose M, Saito M, Wada H, Kitano A, Kondo N, NishiyasuT. Modulation of arterial baroreflex control of muscle sym-pathetic nerve activity by muscle metaboreflex in humans.Am J Physiol Heart Circ Physiol 2004;286:H701–707.

90. Ichinose M, Saito M, Wada H, Kitano A, Kondo N, NishiyasuT. Modulation of arterial baroreflex dynamic response dur-ing muscle metaboreflex activation in humans. J Physiol(Lond) 2002;544:939–948.

91. Iwase S, Mano T, Kamiya A, Niimi Y, Fu Q, Suzumura A.Syncopal attack alters the burst properties of muscle sympa-thetic nerve activity in humans. Auton Neurosci 2002;95:141–145.

92. Johansson M, Elam M, Rundqvist B, Eisenhofer G, Herlitz H,Lambert G, et al. Increased sympathetic nerve activity inrenovascular hypertension. Circulation 1999;99:2537–2542.

93. Johnson JM, Brengelmann GL, Hales JR, Vanhoutte PM,Wenger CB. Regulation of the cutaneous circulation. FedProc 1986;45:2841–2850.

94. Johnson JM, Proppe DW. Cardiovascular adjustments to heatstress. In: Handbook of physiology. environmental physiol-ogy. Vol 1 Sect 4. New York: Oxford University Press; 1996. p215–244.

95. Johnson JM, Rowell LB, Niederberger M, Eisman MM. Hu-man splanchnic and forearm vasoconstrictor responses toreductions of right atrial and aortic pressures. Circ Res 1974;34:515–524.

96. Jones PP, Christou DD, Jordan J, Seals DR. Baroreflex buff-ering is reduced with age in healthy men. Circulation 2003;107:1770–1774.

97. Jones PP, Shapiro LF, Keisling GA, Jordan J, Shannon JR,Quaife RA, et al. Altered autonomic support of arterialblood pressure with age in healthy men. Circulation 2001;104:2424–2429.

98. Jones PP, Spraul M, Matt KS, Seals DR, Skinner JS, RavussinE. Gender does not influence sympathetic neural reactivity


to stress in healthy humans. Am J Physiol 1996;270:H350–357.

99. Kamiya A, Hayano J, Kawada T, Michikami D, Yamamoto K,Ariumi H, et al. Low-frequency oscillation of sympatheticnerve activity decreases during development of tilt-inducedsyncope preceding sympathetic withdrawal and bradycardia.Am J Physiol Heart Circ Physiol 2005;289:H1758–1769.

100. Kamiya A, Kawada T, Yamamoto K, Michikami D, Ariumi H,Miyamoto T, et al. Dynamic and static baroreflex control ofmuscle sympathetic nerve activity (SNA) parallels that ofrenal and cardiac SNA during physiological change in pres-sure. Am J Physiol Heart Circ Physiol 2005;289:H2641–2648.

101. Kamiya A, Kawada T, Yamamoto K, Michikami D, Ariumi H,Miyamoto T, et al. Muscle sympathetic nerve activity aver-aged over 1 minute parallels renal and cardiac sympatheticnerve activity in response to a forced baroreceptor pressurechange. Circulation 2005;112:384–386.

102. Kaufman MP, Hayes SG. The exercise pressor reflex. ClinAuton Res 2002;12:429–439.

103. Kaye DM, Lefkovits J, Jennings GL, Bergin P, Broughton A,Esler MD. Adverse consequences of high sympathetic ner-vous activity in the failing human heart. J Am Coll Cardiol1995;26:1257–1263.

104. Keller DM, Fadel PJ, Ogoh S, Brothers RM, Hawkins M,Olivencia-Yurvati A, et al. Carotid baroreflex control of legvasculature in exercising and non-exercising skeletal musclein humans. J Physiol (Lond) 2004;561:283–293.

105. Kellogg DL Jr, Johnson JM, Kosiba WA. Baroreflex control ofthe cutaneous active vasodilator system in humans. Circ Res1990;66:1420–1426.

106. Kellogg DL Jr, Pergola PE, Piest KL, Kosiba WA, CrandallCG, Grossmann M, et al. Cutaneous active vasodilation inhumans is mediated by cholinergic nerve cotransmission.Circ Res 1995;77:1222–1228.

107. Khayat RN, Przybylowski T, Meyer KC, Skatrud JB, MorganBJ. Role of sensory input from the lungs in control of musclesympathetic nerve activity during and after apnea in hu-mans. J Appl Physiol 2004;97:635–640.

108. Kienbaum P, Karlsson T, Sverrisdottir YB, Elam M, Wallin BG.Two sites for modulation of human sympathetic activity byarterial baroreceptors? J Physiol (Lond) 2001;531:861–869.

109. Kingwell BA, Thompson JM, Kaye DM, McPherson GA, Jen-nings GL, Esler MD. Heart rate spectral analysis, cardiacnorepinephrine spillover, and muscle sympathetic nerve ac-tivity during human sympathetic nervous activation and fail-ure. Circulation 1994;90:234–240.

110. Klein IH, Ligtenberg G, Oey PL, Koomans HA, BlankestijnPJ. Enalapril and losartan reduce sympathetic hyperactivityin patients with chronic renal failure. J Am Soc Nephrol2003;14:425–430.

111. Klein IH, Ligtenberg G, Oey PL, Koomans HA, BlankestijnPJ. Sympathetic activity is increased in polycystic kidney dis-ease and is associated with hypertension. J Am Soc Nephrol2001;12:2427–2433.

112. Kolka MA, Stephenson LA. Cutaneous blood flow and localsweating after systemic atropine administration. PflugersArch 1987;410:524–529.

113. Lacolley PJ, Pannier BM, Slama MA, Cuche JL, Hoeks AP,Laurent S, et al. Carotid arterial haemodynamics after milddegrees of lower-body negative pressure in man. Clin Sci(Lond) 1992;83:535–540.

114. Lambert EA, Schlaich MP. Reduced sympathoneural responsesto the cold pressor test in individuals with essential hyperten-sion and in those genetically predisposed to hypertension. Nosupport for the “pressor reactor” hypothesis of hypertensiondevelopment. Am J Hypertens 2004;17:863–868.

115. Leimbach WN Jr, Wallin BG, Victor RG, Aylward PE, SundlofG, Mark AL. Direct evidence from intraneural recordings forincreased central sympathetic outflow in patients with heartfailure. Circulation 1986;73:913–919.

116. Leuenberger UA, Brubaker D, Quraishi S, Hogeman CS,Imadojemu VA, Gray KS. Effects of intermittent hypoxia on

sympathetic activity and blood pressure in humans. AutonNeurosci 2005;121:87–93.

117. Macefield VG, Elam M, Wallin BG. Firing properties ofsingle postganglionic sympathetic neurones recorded inawake human subjects. Auton Neurosci 2002;95:146–159.

118. Macefield VG, Rundqvist B, Sverrisdottir YB, Wallin BG,Elam M. Firing properties of single muscle vasoconstrictorneurons in the sympathoexcitation associated with conges-tive heart failure. Circulation 1999;100:1708–1713.

119. Macefield VG, Wallin BG. Effects of static lung inflation onsympathetic activity in human muscle nerves at rest andduring asphyxia. J Auton Nerv Syst 1995;53:148–156.

120. Macefield VG, Wallin BG. Firing properties of single vasocon-strictor neurones in human subjects with high levels of musclesympathetic activity. J Physiol (Lond) 1999;516:293–301.

121. Mano T. Microneurographic research on sympathetic nerveresponses to environmental stimuli in humans. Jpn J Physiol1998;48:99–114.

122. Mano T, Iwase S, Toma S. Microneurography as a tool inclinical neurophysiology to investigate peripheral neuraltraffic in humans. Clin Neurophysiol 2006;117:2357–2384.

123. Mark AL, Victor RG, Nerhed C, Wallin BG. Microneuro-graphic studies of the mechanisms of sympathetic nerveresponses to static exercise in humans. Circ Res 1985;57:461–469.

124. Matsukawa K, Wall PT, Wilson LB, Mitchell JH. Reflex stimu-lation of cardiac sympathetic nerve activity during static musclecontraction in cats. Am J Physiol 1994;267:H821–827.

125. Matsukawa T, Sugiyama Y, Mano T. Age-related changes inbaroreflex control of heart rate and sympathetic nerve ac-tivity in healthy humans. J Auton Nerv Syst 1996;60:209–212.

126. Matsukawa T, Sugiyama Y, Watanabe T, Kobayashi F, ManoT. Gender difference in age-related changes in muscle sym-pathetic nerve activity in healthy subjects. Am J Physiol 1998;275:R1600–1604.

127. McLachlan EM. Transmission of signals through sympa-thetic ganglia—modulation, integration or simply distribu-tion? Acta Physiol Scand 2003;177:227–235.

128. Middlekauff HR, Hamilton MA, Stevenson LW, Mark AL.Independent control of skin and muscle sympathetic nerveactivity in patients with heart failure. Circulation 1994;90:1794–1798.

129. Minson CT, Halliwill JR, Young TM, Joyner MJ. Influence ofthe menstrual cycle on sympathetic activity, baroreflex sen-sitivity, and vascular transduction in young women. Circula-tion 2000;101:862–868.

130. Minson CT, Halliwill JR, Young TM, Joyner MJ. Sympatheticactivity and baroreflex sensitivity in young women takingoral contraceptives. Circulation 2000;102:1473–1476.

131. Miyajima E, Yamada Y, Yoshida Y, Matsukawa T, Shionoiri H,Tochikubo O, et al. Muscle sympathetic nerve activity inrenovascular hypertension and primary aldosteronism. Hy-pertension 1991;17:1057–1062.

132. Monahan KD, Leuenberger UA, Ray CA. Effect of repetitivehypoxic apnoeas on baroreflex function in humans.J Physiol (Lond) 2006;574:605–613.

133. Morgan BJ, Crabtree DC, Puleo DS, Badr MS, Toiber F,Skatrud JB. Neurocirculatory consequences of abruptchange in sleep state in humans. J Appl Physiol 1996;80:1627–1636.

134. Mosqueda-Garcia R, Furlan R, Fernandez-Violante R, DesaiT, Snell M, Jarai Z, et al. Sympathetic and baroreceptorreflex function in neurally mediated syncope evoked by tilt.J Clin Invest 1997;99:2736–2744.

135. Mosqueda-Garcia R, Furlan R, Tank J, Fernandez-Violante R.The elusive pathophysiology of neurally mediated syncope.Circulation 2000;102:2898–2906.

136. Muntzel MS, Anderson EA, Johnson AK, Mark AL. Mecha-nisms of insulin action on sympathetic nerve activity. ClinExp Hypertens 1995;17:39–50.

137. Murai H, Takata S, Maruyama M, Nakano M, Kobayashi D,Otowa K, et al. The activity of a single muscle sympathetic


vasoconstrictor nerve unit is affected by physiological stress inhumans. Am J Physiol Heart Circ Physiol 2006;290:H853–860.

138. Narkiewicz K, Kato M, Phillips BG, Pesek CA, Davison DE,Somers VK. Nocturnal continuous positive airway pressuredecreases daytime sympathetic traffic in obstructive sleepapnea. Circulation 1999;100:2332–2335.

139. Narkiewicz K, Phillips BG, Kato M, Hering D, BieniaszewskiL, Somers VK. Gender-selective interaction between aging,blood pressure, and sympathetic nerve activity. Hyperten-sion 2005;45:522–525.

140. Narkiewicz K, Somers VK. Sympathetic nerve activity in obstruc-tive sleep apnoea. Acta Physiol Scand 2003;177:385–390.

141. Narkiewicz K, Somers VK. The sympathetic nervous systemand obstructive sleep apnea: implications for hypertension.J Hypertens 1997;15:1613–1619.

142. Narkiewicz K, van de Borne P, Montano N, Hering D, KaraT, Somers VK. Sympathetic neural outflow and chemoreflexsensitivity are related to spontaneous breathing rate in nor-mal men. Hypertension 2006;47:51–55.

143. Ng AV, Callister R, Johnson DG, Seals DR. Age and genderinfluence muscle sympathetic nerve activity at rest in healthyhumans. Hypertension 1993;21:498–503.

144. Nilsson H, Ljung B, Sjoblom N, Wallin BG. The influence ofthe sympathetic impulse pattern on contractile responses ofrat mesenteric arteries and veins. Acta Physiol Scand 1985;123:303–309.

145. Noll G, Wenzel RR, Schneider M, Oesch V, Binggeli C, ShawS, et al. Increased activation of sympathetic nervous systemand endothelin by mental stress in normotensive offspringof hypertensive parents. Circulation 1996;93:866–869.

146. Oda Y, Asanoi H, Ueno H, Yamada K, Joho S, Kameyama T,et al. Pulse-synchronous sympathetic burst power as a newindex of sympathoexcitation in patients with heart failure.Am J Physiol Heart Circ Physiol 2004;287:H1821–1827.

147. Okada H, Iwase S, Mano T, Sugiyama Y, Watanabe T.Changes in muscle sympathetic nerve activity during sleep inhumans. Neurology 1991;41:1961–1966.

148. Okagawa T, Sugenoya J, Iwase S, Mano T, Suzumura A,Matsumoto T, et al. Occurrence of the spinal reflex due toskin pressure in sudomotor and cutaneous vasoconstrictornerve system of humans. Auton Neurosci 2003;105:62–70.

149. Okamoto T, Iwase S, Sugenoya J, Mano T, Sugiyama Y,Yamamoto K. Different thermal dependency of cutaneoussympathetic outflow to glabrous and hairy skin in humans.Eur J Appl Physiol Occup Physiol 1994;68:460–464.

150. O’Leary DD, Kimmerly DS, Cechetto AD, Shoemaker JK. Dif-ferential effect of head-up tilt on cardiovagal and sympatheticbaroreflex sensitivity in humans. Exp Physiol 2003;88:769–774.

151. Persson B, Andersson OK, Hjemdahl P, Wysocki M, AgerwallS, Wallin G. Adrenaline infusion in man increases musclesympathetic nerve activity and noradrenaline overflow toplasma. J Hypertens 1989;7:747–756.

152. Petersson MJ, Rundqvist B, Johansson M, Eisenhofer G, Lam-bert G, Herlitz H, et al. Increased cardiac sympathetic drive inrenovascular hypertension. J Hypertens 2002;20:1181–1187.

153. Ray CA, Carter JR. Vestibular activation of sympathetic nerveactivity. Acta Physiol Scand 2003;177:313–319.

154. Ray CA, Gracey KH. Augmentation of exercise-induced mus-cle sympathetic nerve activity during muscle heating. J ApplPhysiol 1997;82:1719–1725.

155. Ray CA, Monahan KD. Sympathetic vascular transduction isaugmented in young normotensive blacks. J Appl Physiol2002;92:651–656.

156. Ray CA, Monahan KD. The vestibulosympathetic reflex inhumans: neural interactions between cardiovascular re-flexes. Clin Exp Pharmacol Physiol 2002;29:98–102.

157. Rea RF, Biaggioni I, Robertson RM, Haile V, Robertson D.Reflex control of sympathetic nerve activity in dopaminebeta-hydroxylase deficiency. Hypertension 1990;15:107–112.

158. Rea RF, Hamdan M. Baroreflex control of muscle sympa-thetic nerve activity in borderline hypertension. Circulation1990;82:856–862.

159. Rowell LB. Ideas about control of skeletal and cardiac mus-cle blood flow (1876-2003): cycles of revision and new vision.J Appl Physiol 2004;97:384–392.

160. Rudas L, Crossman AA, Morillo CA, Halliwill JR, Tah-vanainen KU, Kuusela TA, et al. Human sympathetic andvagal baroreflex responses to sequential nitroprusside andphenylephrine. Am J Physiol 1999;276:H1691–1698.

161. Rundqvist B, Casale R, Bergmann-Sverrisdottir Y, Friberg P,Mortara A, Elam M. Rapid fall in sympathetic nerve hyper-activity in patients with heart failure after cardiac transplan-tation. J Card Fail 1997;3:21–26.

162. Rundqvist B, Elam M, Bergmann-Sverrisdottir Y, EisenhoferG, Friberg P. Increased cardiac adrenergic drive precedesgeneralized sympathetic activation in human heart failure.Circulation 1997;95:169–175.

163. Rusch NJ, Shepherd JT, Webb RC, Vanhoutte PM. Differentbehavior of the resistance vessels of the human calf andforearm during contralateral isometric exercise, mentalstress, and abnormal respiratory movements. Circ Res 1981;48:I118–130.

164. Scherrer U, Randin D, Tappy L, Vollenweider P, Jequier E,Nicod P. Body fat and sympathetic nerve activity in healthysubjects. Circulation 1994;89:2634–2640.

165. Schlaich MP, Lambert E, Kaye DM, Krozowski Z, Campbell DJ,Lambert G, et al. Sympathetic augmentation in hypertension:role of nerve firing, norepinephrine reuptake, and Angiotensinneuromodulation. Hypertension 2004;43:169–175.

166. Schobel HP, Fischer T, Heuszer K, Geiger H, Schmieder RE.Preeclampsia — a state of sympathetic overactivity. N EnglJ Med 1996;335:1480–1485.

167. Schobel HP, Frank H, Naraghi R, Geiger H, Titz E, HeusserK. Hypertension in patients with neurovascular compressionis associated with increased central sympathetic outflow.J Am Soc Nephrol 2002;13:35–41.

168. Schobel HP, Heusser K, Schmieder RE, Veelken R, FischerT, Luft FC. Evidence against elevated sympathetic vasocon-strictor activity in borderline hypertension. J Am Soc Neph-rol 1998;9:1581–1587.

169. Seals DR. Sympathetic neural adjustments to stress in phys-ically trained and untrained humans. Hypertension 1991;17:36–43.

170. Seals DR, Esler MD. Human ageing and the sympathoadre-nal system. J Physiol (Lond) 2000;528:407–417.

171. Seals DR, Suwarno NO, Dempsey JA. Influence of lungvolume on sympathetic nerve discharge in normal humans.Circ Res 1990;67:130–141.

172. Seals DR, Suwarno NO, Joyner MJ, Iber C, Copeland JG,Dempsey JA. Respiratory modulation of muscle sympatheticnerve activity in intact and lung denervated humans. CircRes 1993;72:440–454.

173. Sheldahl LM, Ebert TJ, Cox B, Tristani FE. Effect of aerobictraining on baroreflex regulation of cardiac and sympatheticfunction. J Appl Physiol 1994;76:158–165.

174. Shimizu T, Takahashi Y, Suzuki K, Kogawa S, Tashiro T,Takahasi K, et al. Muscle nerve sympathetic activity duringsleep and its change with arousal response. J Sleep Res1992;1:178–185.

175. Shoemaker JK, Hogeman CS, Khan M, Kimmerly DS, Sino-way LI. Gender affects sympathetic and hemodynamic re-sponse to postural stress. Am J Physiol Heart Circ Physiol2001;281:H2028–2035.

176. Shortt TL, Ray CA. Sympathetic and vascular responses tohead-down neck flexion in humans. Am J Physiol 1997;272:H1780–1784.

177. Silber DH, Sutliff G, Yang QX, Smith MB, Sinoway LI,Leuenberger UA. Altered mechanisms of sympathetic acti-vation during rhythmic forearm exercise in heart failure.J Appl Physiol 1998;84:1551–1559.

178. Sinoway LI, Li J. A perspective on the muscle reflex: implica-tions for congestive heart failure. J Appl Physiol 2005;99:5–22.

179. Skarphedinsson JO, Elam M, Jungersten L, Wallin BG. Sym-pathetic nerve traffic correlates with the release of nitric


oxide in humans: implications for blood pressure control.J Physiol (Lond) 1997;501:671–675.

180. Smith ML, Ellenbogen KA, Eckberg DL. Sympathoinhibitionand hypotension in carotid sinus hypersensitivity. Clin AutonRes 1992;2:389–392.

181. Smith PA, Meaney JF, Graham LN, Stoker JB, MackintoshAF, Mary DA, et al. Relationship of neurovascular compres-sion to central sympathetic discharge and essential hyper-tension. J Am Coll Cardiol 2004;43:1453–1458.

182. Smith SA, Mitchell JH, Garry MG. The mammalian exercisepressor reflex in health and disease. Exp Physiol 2006;91:89–102.

183. Somers VK, Dyken ME, Clary MP, Abboud FM. Sympatheticneural mechanisms in obstructive sleep apnea. J Clin Invest1995;96:1897–1904.

184. Somers VK, Mark AL, Abboud FM. Potentiation of sympa-thetic nerve responses to hypoxia in borderline hypertensivesubjects. Hypertension 1988;11:608–612.

185. Somers VK, Mark AL, Zavala DC, Abboud FM. Contrastingeffects of hypoxia and hypercapnia on ventilation and sympa-thetic activity in humans. J Appl Physiol 1989;67:2101–2106.

186. Spaak J, Egri ZJ, Kubo T, Yu E, Ando S, Kaneko Y, et al.Muscle sympathetic nerve activity during wakefulness inheart failure patients with and without sleep apnea. Hyper-tension 2005;46:1327–1332.

187. Spraul M, Ravussin E, Fontvieille AM, Rising R, Larson DE,Anderson EA. Reduced sympathetic nervous activity. A po-tential mechanism predisposing to body weight gain. J ClinInvest 1993;92:1730–1735.

188. Stjernberg L, Blumberg H, Wallin BG. Sympathetic activityin man after spinal cord injury. Outflow to muscle below thelesion. Brain 1986;109:695–715.

189. Sugenoya J, Iwase S, Mano T, Sugiyama Y, Ogawa T, Nish-iyama T, et al. Vasodilator component in sympathetic nerveactivity destined for the skin of the dorsal foot of mildlyheated humans. J Physiol (Lond) 1998;507:603–610.

190. Sundlof G, Wallin BG. Effect of lower body negative pressureon human muscle nerve sympathetic activity. J Physiol(Lond) 1978;278:525–532.

191. Sundlof G, Wallin BG. Human muscle nerve sympatheticactivity at rest. Relationship to blood pressure and age.J Physiol (Lond) 1978;274:621–637.

192. Sundlof G, Wallin BG. The variability of muscle nerve sym-pathetic activity in resting recumbent man. J Physiol(Lond)1977;272:383–397.

193. Svedenhag J, Wallin BG, Sundlof G, Henriksson J. Skeletalmuscle sympathetic activity at rest in trained and untrainedsubjects. Acta Physiol Scand 1984;120:499–504.

194. Sverrisdottir YB, Johannsson G, Jungersten L, Wallin BG,Elam M. Is the somatotropic axis related to sympatheticnerve activity in healthy ageing men? J Hypertens 2001;19:2019–2024.

195. Sverrisdottir YB, Rundqvist B, Elam M. Relative burst ampli-tude in human muscle sympathetic nerve activity: a sensitiveindicator of altered sympathetic traffic. Clin Auton Res 1998;8:95–100.

196. Tank J, Diedrich A, Hale N, Niaz FE, Furlan R, RobertsonRM, et al. Relationship between blood pressure, sleep K-complexes, and muscle sympathetic nerve activity in hu-mans. Am J Physiol Regul Integr Comp Physiol 2003;285:R208–214.

197. Tataranni PA, Cizza G, Snitker S, Gucciardo F, Lotsikas A,Chrousos GP, et al. Hypothalamic-pituitary-adrenal axis andsympathetic nervous system activities in Pima Indians andCaucasians. Metabolism 1999;48:395–399.

198. Taylor JA, Halliwill JR, Brown TE, Hayano J, Eckberg DL.Non-hypotensive hypovolaemia reduces ascending aortic di-mensions in humans. J Physiol (Lond) 1995;483:289–298.

199. Timio M, Lippi G, Venanzi S, Gentili S, Quintaliani G,Verdura C, et al. Blood pressure trend and cardiovascular

events in nuns in a secluded order: a 30-year follow-up study.Blood Press 1997;6:81–87.

200. Vallbo ÅB, Hagbarth KE, Torebjork HE, Wallin BG. Somato-sensory, proprioceptive, and sympathetic activity in humanperipheral nerves. Physiol Rev 1979;59:919–957.

201. Vallbo ÅB, Hagbarth KE, Wallin BG. Microneurography:how the technique developed and its role in the investiga-tion of the sympathetic nervous system. J Appl Physiol 2004;96:1262–1269.

202. Velez-Roa S, Ciarka A, Najem B, Vachiery JL, Naeije R, vande Borne P. Increased sympathetic nerve activity in pulmo-nary artery hypertension. Circulation 2004;110:1308–1312.

203. Victor RG, Rotto DM, Pryor SL, Kaufman MP. Stimulation ofrenal sympathetic activity by static contraction: evidence formechanoreceptor-induced reflexes from skeletal muscle.Circ Res 1989;64:592–599.

204. Victor RG, Secher NH, Lyson T, Mitchell JH. Central com-mand increases muscle sympathetic nerve activity duringintense intermittent isometric exercise in humans. Circ Res1995;76:127–131.

205. Vissing SF, Secher NH, Victor RG. Mechanisms of cutaneousvasoconstriction during upright posture. Acta Physiol Scand1997;159:131–138.

206. Wallin BG. Human sympathetic nerve activity and bloodpressure regulation. Clin Exp Hypertens A 1989;11(Suppl1):91–101.

207. Wallin BG. Microneurographic assessment of sympatheticnerve traffic. In: Advances in clinical neurophysiology. SupplClin Neurophysiol 2004;57:345–351.

208. Wallin BG. Regulation of sympathetic nerve traffic to skeletalmuscle in resting humans. Clin Auton Res 2006;16:262–269.

209. Wallin BG, Delius W, Sundlof G. Human muscle nerve sym-pathetic activity in cardiac arrhythmias. Scand J Clin LabInvest 1974;34:293–300.

210. Wallin BG, Esler M, Dorward P, Eisenhofer G, Ferrier C,Westerman R, et al. Simultaneous measurements of cardiacnoradrenaline spillover and sympathetic outflow to skeletalmuscle in humans. J Physiol (Lond) 1992;453:45–58.

211. Wallin BG, Kunimoto MM, Sellgren J. Possible genetic influ-ence on the strength of human muscle nerve sympatheticactivity at rest. Hypertension 1993;22:282–284.

212. Wallin BG, Sundlof G. Sympathetic outflow to muscles dur-ing vasovagal syncope. J Auton Nerv Syst 1982;6:287–291.

213. Wallin BG, Thompson JM, Jennings GL, Esler MD. Renalnoradrenaline spillover correlates with muscle sympatheticactivity in humans. J Physiol (Lond) 1996;491:881–887.

214. Wallin BG, Westerberg CE, Sundlof G. Syncope induced byglossopharyngeal neuralgia: sympathetic outflow to muscle.Neurology 1984;34:522–524.

215. Welch WJ, Smith ML, Rea RF, Bauernfeind RA, Eckberg DL.Enhancement of sympathetic nerve activity by single prematureventricular beats in humans. J Am Coll Cardiol 1989;13:69–75.

216. Weyer C, Pratley RE, Snitker S, Spraul M, Ravussin E,Tataranni PA. Ethnic differences in insulinemia and sympa-thetic tone as links between obesity and blood pressure.Hypertension 2000;36:531–537.

217. Wilcox I, Grunstein RR, Hedner JA, Doyle J, Collins FL,Fletcher PJ, et al. Effect of nasal continuous positive airwaypressure during sleep on 24-hour blood pressure in obstruc-tive sleep apnea. Sleep 1993;16:539–544.

218. Wilson TE, Cui J, Crandall CG. Absence of arterial barore-flex modulation of skin sympathetic activity and sweat rateduring whole-body heating in humans. J Physiol (Lond)2001;536:615–623.

219. Xie A, Skatrud JB, Puleo DS, Morgan BJ. Exposure to hyp-oxia produces long-lasting sympathetic activation in hu-mans. J Appl Physiol 2001;91:1555–1562.

220. Yamada Y, Miyajima E, Tochikubo O, Matsukawa T, Ishii M.Age-related changes in muscle sympathetic nerve activity inessential hypertension. Hypertension 1989;13:870–877.


INVITED REVIEW ABSTRACT: The Miller Fisher syndrome (MFS), characterized by ataxia,areflexia, and ophthalmoplegia, was first recognized as a distinct clinicalentity in 1956. MFS is mostly an acute, self-limiting condition, but there isanecdotal evidence of benefit with immunotherapy. Pathological data remainscarce. MFS can be associated with infectious, autoimmune, and neoplasticdisorders. Radiological findings have suggested both central and peripheralinvolvement. The anti-GQ1b IgG antibody titer is most commonly elevated inMFS, but may also be increased in Guillain–Barre syndrome (GBS) andBickerstaff’s brainstem encephalitis (BBE). Molecular mimicry, particularly inrelation to antecedent Campylobacter jejuni and Hemophilus influenzaeinfections, is likely the predominant pathogenic mechanism, but the roles ofother biological factors remain to be established. Recent studies havedemonstrated the presence of neuromuscular transmission defects in asso-ciation with anti-GQ1b IgG antibody, both in vitro and in vivo. Collectivefindings from clinical, radiological, immunological, and electrophysiologicaltechniques have helped to define MFS, GBS, and BBE as major disorderswithin the proposed spectrum of anti-GQ1b IgG antibody syndrome.

Muscle Nerve 36: 615–627, 2007

CLINICAL AND IMMUNOLOGICAL SPECTRUMOF THE MILLER FISHER SYNDROME

Y. L. LO, MD

Department of Neurology, National Neuroscience Institute, Singapore General Hospital,Outram Road, 169608 Singapore

Accepted 4 May 2007

The Miller Fisher syndrome (MFS) was first de-scribed clinically in 1956.40 The original article hadpostulated it as an unusual variant of acute idio-pathic polyneuritis, implying a strong relation withthe Guillain–Barre syndrome (GBS). The classictriad consisted of ophthalmoplegia, ataxia, andareflexia in an acute setting, and was first recognizedby Collier in 1932.34

In this review, published information on all as-pects of MFS will be summarized and discussed,mostly in a chronological fashion, to provide a com-prehensive account of developments pertaining tothis condition. As such, origins of the proposed anti-GQ1b IgG antibody syndrome, comprising MFS,GBS, and Bickerstaff’s brainstem encephalitis(BBE), will also be apparent.

SEARCH STRATEGY AND CRITERIA

References for this review were identified bysearches of PubMed from 1971 to April 2006 withthe terms “Miller Fisher syndrome,” “Fisher syn-drome,” “Fisher’s syndrome,” “Guillain–Barre syn-drome,” “Bickerstaff’s brainstem encephalitis,” and“anti-GQ1b antibody.” Due to the extensiveness ofthis topic, prioritization was given to publicationsrelevant to MFS. Selective data on GBS were in-cluded. Case reports were selected based on origi-nality. Other materials were from the author’s ownfiles. Only articles in the English language were in-cluded.

DIAGNOSIS

For the clinician, a diagnostic approach in confront-ing the varied features comprising diplopia, dysar-thria, facial asymmetry, motor incoordination, sen-sory disturbances, and weakness due to long-tractinvolvement is most relevant. Unfortunately, increas-ing knowledge of MFS and related disorders fromantibody-specificity studies, in relation to clinical syn-dromes, have made discrete characterization diffi-cult. Various formes frustes and overlap syndromesare increasingly reported, which expand the under-

Available for Category 1 CME credit through the AANEM at www.aanem.org.

Abbreviations: BBE, Bickerstaff’s brainstem encephalitis; GBS, Guillain–Barre syndrome; MFS, Miller Fisher syndrome; MRI, magnetic resonanceimagingKey words: anti-GQ1b IgG antibody; Bickerstaff’s brainstem encephalitis;Guillain–Barre syndrome; Miller Fisher syndrome; reviewCorrespondence to: Y. L. Lo; e-mail: [email protected]


Miller Fisher Syndrome MUSCLE & NERVE November 2007 615

standing of this clinical spectrum. A logical clinicalapproach would be to consider MFS, GBS, and BBEas separate initial categories and rationalize existingoverlapping features. Before the advent of anti-GQ1b IgG testing (Table 1), the diagnosis of theseconditions was largely clinical, assisted by imagingand electrophysiological studies.

In considering MFS as a disease entity, research-ers or clinicians may choose to characterize it fromseveral starting points: clinical features, antibodyprofile, infective triggers, or a combination of these.It is widely accepted that anti-GQ1b IgG antibody ispresent in well over 90% of MFS patients, and isabsent in normal subjects or disease controls.

CLINICAL FEATURES

The classic clinical triad originally described a com-bination of central and peripheral involvement. Thisled to constructive debates on the existence of asingle lesion site.

In the largest reported series,105 comprising 50consecutive MFS patients, strict entry criteria ofacute ophthalmoplegia, ataxia, and areflexia, with-out major limb weakness or other signs suggestingcentral nervous system involvement, were applied.Antecedent respiratory symptoms occurred in 76%and gastrointestinal symptoms in 4%. The medianinterval between infection onset and development ofneurological symptoms was 8 days. Overall, 89% hadelevated anti-GQ1b IgG antibodies. A median inter-val of 12–15 days was found between neurologicalonset and the beginning of recovery. By 6 months,all patients were free from ataxia and ophthalmople-gia. No deaths were reported. Table 2 summarizesthe neurological signs and symptoms in these pa-tients. Comparison made with another large seriesdid not reveal differences, suggesting fairly uniformclinical features.

Clinically, MFS is mostly a self-limiting condi-tion.105 However, cases progressing to respiratory

failure and requiring mechanical ventilation havebeen described,14 particularly in children.6,10,133

Other serious complications reported include co-ma,97 ballism,118 cardiomyopathy from dysautono-mia,124 lactic acidosis,150 and pain.79,100 Recurrenceof MFS,16,69,95,125,143,145,152 sometimes showing differ-ent phenotypes at different times, has been welldocumented.51

Ophthalmological and Cranial Nerve Features. Theclinical features of MFS are of great clinical interest,particularly the ophthalmological aspects. Apartfrom visual impairment from optic neuritis,22,23,33,160

other ophthalmological abnormalities include diver-gence paralysis,130 lid retraction, upper lid jerks, in-ternuclear ophthalmoplegia, convergence spasm,Parinaud’s syndrome, defective vestibulo-ocular re-flex,2 chronic ophthalmoplegia,134 areflexic mydria-sis, convergence failure,21,24 and acute angle clo-sure.159 Isolated abducens nerve palsy was suggestedto be a mild form of MFS in a recent study.159 Thefacial nerve was involved in 45.7% of patients in oneseries.12

Ataxia. The origin of ataxia in MFS is of profoundclinical interest. The original study by Fisher40 pro-

Table 2. Comparison of two large published clinical serieson MFS.

Mori et al.(n � 50)105

Lyu et al.(n � 32)93

Mean age NM 45 yearsMedian age 40 years NMGender Male

preponderanceNo genderdifference

Season Spring SpringUpper respiratory infection 76% 56%Acute gastroenteritis 4% 0%Median time to nadir 6 days 5.5 days

Clinical features (%)Pupillary abnormalities 42 NMPtosis 58 59Facial palsy 32 50Bulbar palsy 26 59Limb weakness 20 25Limb dysesthesia 24 34Superficial sensory loss 20 50Trigeminal dysfunction NM 16Abnormal vibratory and

deep sensation18 NM

Micturition disturbance 8 3

All patients had ataxia, areflexia, and ophthalmoplegia.No significant differences in clinical features were found between the twoseries (unpaired t-test, P � 0.3).NM, not mentioned.

Table 1. Clinical spectrum of the anti-GQ1b antibody syndrome.

Disorder Clinical featuresAnti-GQ1b antibody

frequency

MFS Ataxia, areflexia,ophthalmoplegia40

Up to 95%76

GBS Weakness, sensory loss,areflexia, cranialneuropathy164

Up to 26%22

BBE Ophthalmoplegia, ataxia,hypereflexia or disturbedconsciousness120

Up to 66%120

616 Miller Fisher Syndrome MUSCLE & NERVE November 2007

posed selective involvement of Ia-afferent neurons.Early case studies correlated abnormalities of Ia-afferent fibers with the severity of ataxia, suggestingthis process is indeed involved in its development.165

Ropper and Shahani138 suggested that disparity be-tween proprioceptive information from muscle spin-dles and kinesthetic information from joints may bea mechanism of ataxia, based on abnormalities ofperipheral nerves. Employing postural body swayanalysis, Kuwabara and colleagues82 concluded thatMFS patients have dysfunction of the proprioceptiveafferent system, and that sensory ataxia may becaused by selective involvement of muscle-spindleafferents.

Anti-GQ1b monoclonal antibody immunostain-ing of human dorsal root ganglion has been demon-strated, but its physiological significance remains un-certain.81 Direct staining of Ia afferents with anti-GQ1b IgG antibody has not been demonstrated todate.

In ataxic GBS, distal paresthesias, areflexia, andsensory ataxia have been documented. Some cases,however, showed features suggesting cerebellar-typeataxia. Ataxic GBS may progress to typical GBS,where limb weakness predominates. In a large seriesof 340 cases with anti-GQ1b IgG antibody positivity,6 cases were consistent with the ataxic form of GBS.Anti-GQ1b IgG antibody from these patients cross-reacted with GT1a, suggesting that autoantibodieswith the same fine specificity were seen in MFS andataxic GBS.175

In a large series of 445 patients with GBS, anti-GD1b IgG without cross-reactivity with other glyco-lipids was present in 9 cases, all with sensory distur-bance and 1 with cerebellar ataxia.102 Other GBScases reported subsequently had sensory ataxia andsimilarly elevated anti-GD1b IgG antibody.168,175 AsGD1b was found to localize to primary sensory neu-rons in humans, these reports seem to favor bothcerebellar and sensory involvement as causes for theataxia.

Immunological evidence of cerebellar involve-ment would be of interest for explaining the ataxiain MFS. A 1-year study using immunocytochemicalstaining of human cerebellum described selectivestaining of the molecular layer with sera from 3 MFSor GBS patients who had elevated IgG anti-GQ1bantibody levels.80 Western blot analysis also revealedincreased levels of anti-cerebellar antibodies in MFSpatients when compared with GBS patients orhealthy controls.60 These findings suggest the pres-ence of immunologically mediated cerebellar dys-function in MFS, but more studies are needed tofurther define the role of anti-GQ1b IgG antibody, as

well as the underlying immunological mechanism ofcerebellar dysfunction in MFS.

In summary, the issue of ataxia in MFS has notbeen fully explained. It is possible that the relativepredominance of anti-GQ1b or anti-GD1b IgG anti-bodies may contribute to the ataxia, but furtherstudies in large series will help to resolve these issues.

Areflexia. Areflexia is a clinical sign suggesting pe-ripheral nervous system involvement. It forms part ofthe clinical triad of MFS, and is an overlapping clin-ical feature with GBS. In the largest series of MFSreported to date, all patients demonstrated loss ofreflexes and, over a 4–6-month period of follow-up,two thirds still showed abnormally depressed tendonreflexes.105 The presence of areflexia corroborateselectrophysiological studies demonstrating periph-eral nerve dysfunction,66,67 including axonal neurop-athy41 and abnormal conduction in peripheral sen-sory fibers.47 These findings were very similar tothose seen in GBS.

CLINICAL ASSOCIATIONS

MFS with an immunological basis has been de-scribed in association with autoimmune and neoplas-tic conditions. It has been reported in conjunctionwith systemic lupus erythematosus,13,54 Hashimoto’sthyroiditis,132 Still’s disease,37 Hodgkin’s disease,139

leptomeningeal signet cell carcinomatosis,109 andlung carcinoma.35 Tacrolimus therapy has also beenassociated with MFS developing in a patient receiv-ing a cadaveric liver transplant.71 However, theseanecdotal reports do not in themselves qualify MFSas a paraneoplastic manifestation or rheumatologi-cal disorder.

EPIDEMIOLOGY

The annual incidence of MFS has been estimated at0.09 per 100,000 population and onset is most com-mon in spring. However, large epidemiological stud-ies on MFS remain scarce, and the majority of pub-lished data have been culled from studies on GBS.An incidence of MFS making up 25% of GBS pa-tients was recorded in a Japanese series published in2001.105 A retrospective 11-year hospital study in Tai-wan noted an unusually high percentage (18%) ofMFS among GBS patients, most commonly in win-ter.172 A separate 14-year retrospective review re-corded 7 MFS cases among 96 (7%) cases classifiedas GBS.25 This was in contrast to a prospective Italianstudy, which recorded 4 MFS cases from among 138GBS patients (3%).15 There is thus anecdotal evi-


dence to suggest a higher proportion of MFS amongGBS patients in the Far East. The exact reasonsremain speculative, but may be related to host fac-tors and genetic factors. In addition, uniform defi-nition of diagnostic criteria for MFS, as well as thosefor GBS and BBE, will enable future epidemiologicalstudies to estimate incidence more accurately.

Of the few reports addressing MFS in children,one study58 quoted MFS as being diagnosed in 2 of23 (9%) GBS cases. Both patients recovered with noresidual deficits.

PATHOLOGY

Few reports of autopsies on MFS patients have beenpublished, reflecting the mostly self-limiting clinicalcourse. Two autopsy reports failed to find a centralnervous system lesion129,137; one described demyeli-nating peripheral neuropathy, suggesting that MFSfalls within the spectrum of acute inflammatory poly-neuropathy,133 similar to GBS. Histological examina-tion has demonstrated segmental demyelination andaxonal swelling in peripheral and oculomotornerves, in addition to mild chromatolytic changesand pyknosis of the midbrain in a 64-year-old womanwith recurrent MFS.9 A 1992 review documented 6autopsied patients: 3 showed inflammatory brain-stem lesions and 2 had demyelination of the cranialnerves.12 Hence, the evidence of central pathology islimited and has been derived mostly from radiolog-ical or electrophysiological studies.

RADIOLOGICAL FINDINGS

The advent of magnetic resonance imaging (MRI)has contributed greatly to our understanding ofMFS. To provide a balanced comparison of centraland peripheral involvement, I have arbitrarily in-cluded cranial nerve findings as peripheral lesions.

Imaging is unremarkable in most cases of MFS.The earliest reports of MRI findings associated withMFS showed lesions in the brainstem42,158 and IIInerve nucleus on T2-weighted sequences,53 provid-ing evidence supporting central involvement. Otherreports have described spinocerebellar tract lesionsin the lower medulla,162 and lesions in the midbrain,pons, and middle cerebellar peduncle.32 This maysuggest an immune-mediated breach of the blood–brain barrier, as demonstrated in previous cerebro-spinal fluid studies.149

Peripheral lesions reported in MFS included le-sions in the lumbosacral roots,128 cauda equina, pos-terior column of the spinal cord,61 III or IV cranialnerves,57,107,157 and dorsal root ganglia.84

In comparison to typical GBS, there are abun-dant reports of central lesions on MRI. With im-proved imaging methods, it is likely that more le-sions will be reported, lending support to themultifocal nature of this condition.

ELECTROPHYSIOLOGICAL ASPECTS

Early studies employing nerve conduction studiesdemonstrated peripheral sensory conduction abnor-malities.47,67 Subsequent studies, including use ofthe blink reflex,142 highlighted axonal neuropathicchanges and cranial motor dysfunction,41,144 suggest-ing some differences from the electrophysiologicalfeatures of GBS. The presence of evoked potentialabnormalities (visual, auditory, somatosensory) hasprovided electrophysiological evidence of combinedcentral and peripheral lesions in this condi-tion.46,160,169 Four MFS patients progressing to severetetraparesis showed reduced amplitudes of com-pound muscle action potentials, additionally sug-gesting electrophysiological differences from classicdemyelinating GBS.70 Quantitative cardiovascularautonomic function tests may be subclinically abnor-mal in both the parasympathetic and sympatheticcomponents.94

Electroencephalographic findings were normalin all three patients in one adult series,66 whereastwo case reports in the pediatric age group demon-strated slowing of background rhythm. Abnormali-ties appeared to be mild and non-specific inMFS.10,97

The increasing awareness of combined centraland peripheral involvement in anti-GQ1b IgG anti-body–positive MFS has prompted further investiga-tion into this issue.121 Using serial transcranial mag-netic stimulation, the prolonged subclinical centralmotor conduction time, a reflection of corticospinaldysfunction, was shown to reduce and normalize intandem with clinical recovery in anti-GQ1b IgG an-tibody–positive MFS.91 A reversible corticobulbarmotor conduction time abnormality was also dem-onstrated in one MFS patient exhibiting dysarthria.92

These findings highlight the presence of clinical andsubclinical functional lesions in anti-GQ1b IgG anti-body–positive classic MFS patients or patients withMFS features, in addition to other clinical signs. Thefindings also strengthen the relationship betweenMFS and the related disorder of BBE, where uppermotor neuron signs may coexist with ophthalmople-gia, ataxia, or an alteration of consciousness.

The role of anti-GQ1b IgG antibody in centralnervous system lesions, as demonstrated by electro-physiological and radiological studies, is uncertain.


Indeed, in BBE, where upper motor neuron–typelesions and sensorium changes are most frequentlyencountered, up to 66% of patients have this anti-body. Much work remains to be done to uncover itsactions at different levels of the nervous system. Inaddition, it remains to be determined whether therole of anti-GQ1b IgG antibody in the central ner-vous system is similar to that in the peripheral ner-vous system. How significant is the involvement ofother antibodies? These are intriguing questions,and their answers may provide further insight intothis group of disorders.

There is evidence that anti-GQ1b IgG antibodieshave pathogenic effects at the neuromuscular junc-tion in vitro. Using single-fiber electromyography,patients with acute ophthalmoparesis and elevatedanti-GQ1b IgG antibody were shown to have abnor-mal jitters, which improved with clinical recovery,86

providing the first reported evidence of neuromus-cular transmission defect in patients with MFS. Sim-ilar observations in an anti-GQ1b IgG antibody–neg-ative patient were made by separate investigators,141

suggesting that other antibodies may be involved.Employing high-frequency repetitive nerve stimula-tion,87–89 a presynaptic neuromuscular transmissiondefect was demonstrated in anti-GQ1b IgG anti-body–positive MFS patients90 up to 3 months after

clinical presentation. This corroborates in vitro find-ings of presynaptic structural derangements occur-ring in the nerve terminal,50 as opposed to a tran-sient nerve-blocking phenomena. Such an effect wasnot seen in antibody-negative cases, thus providingfurther evidence of anti-GQ1b IgG antibody–relatedpresynaptic dysfunction in MFS patients. Most re-cently, using abnormal single-fiber electromyogra-phy of the extensor digitorium communis muscle ofan MFS patient, findings further validated the previ-ously mentioned electrophysiological studies.83

Although the electrophysiological studies may benon-specific and correlational, the findings provideimportant functional information linking immuno-logical and clinical data in MFS (Fig. 1).

TREATMENT

It is relevant to note that the half-lives of IgM andIgA are 5 and 6 days, respectively. The half-life ofIgG, by comparison, is 21 days. Plasmapheresis iseffective at shortening the clinical course of GBS,which reaches a nadir by 8–9 days. This suggests thatthe presence of IgG is the predominant factor in thedevelopment of GBS. The median time to nadir inMFS is 5–6 days,105 indicating a somewhat shortertime to maximal clinical deficit.

FIGURE 1. Summary of possible protean effects of anti-GQ1b IgG antibody as evaluated by various electrophysiological techniques inMFS patients.


There are no randomized, double-blind, placebo-controlled trials pertaining to the treatment of MFS.This again is likely related to its self-limiting clinicalcourse and low incidence. Nonetheless, immuno-modulatory treatments similar to those used in GBShave been reported anecdotally. In a large series of 50consecutive cases, no beneficial effect was shown inpatients who received plasmapheresis compared withthose who received no immunotherapy.104

The use of plasmapheresis85,96,171 with success hasmostly been described in case reports, as was the useof intravenous immunoglobulin.181 However, the lat-ter should be used judiciously; a case of cerebralinfarction was reported during treatment of an MFSpatient,161 and posterior reversible encephalopathyhas also occurred.108

Employment of immunoadsorption in a trypto-phan-immobilized column was effective in reducinganti-GQ1b IgG antibody titers from patients’ se-ra.123,180 Willison et al. provided in vitro rationale forimmunoadsorption plasma exchange by showingthat synthetic disialygalactose immunoadsorbentsdeplete anti-GQ1b IgG antibodies from MFS-associ-ated human sera.166 This provides the impetus formore research into developing immunotherapieswith greater antibody specificity and binding effi-cacy.

Future therapeutic directions could be focusedon use of novel agents acting at various levels of theimmunological cascade or at motor nerve terminals.

Although MFS and GBS are related disorders,they have different clinical courses and antibodyprofiles. The present immunomodulatory treat-ments of MFS are largely translated from experi-ences with GBS. To establish convincingly the effi-cacy of these treatments would require large,prospective clinical trials.

At present, clinical discretion should be employedas to when treatment can be instituted in the absenceof established guidelines. MFS is self-limiting, but it isreasonable to consider treatment in cases with rapidprogression of limb, bulbar, or respiratory weakness.

CAMPYLOBACTER AND OTHER INFECTIONS

Like GBS, MFS has been reported to follow infec-tions. Molecular mimicry4,62 has been shown to bethe likely mechanism by which infective agents trig-ger an immunological reaction.

MFS has been associated with Q fever,38 and withinfection with Mycoplasma pneumoniae,99 human im-munodeficiency virus,7 Campylobacter jejuni,3,31,43,110

Epstein–Barr virus,146 Hemophilus influenzae,75 Pastu-ella multocida,11 Helicobacter pylori,30 aspergillus,39

Streptococcus aureus, cytomegalovirus, varicella-zostervirus, and mumps virus.156 However, a large prospec-tive case–control serological study has shown thatassociated infective agents remain unknown in themajority of cases.72

Of the multiple antecedent infections reportedin MFS, C. jejuni infection, as in GBS, has been themost well studied.78 In a large, prospective case–control study, comprising 96 GBS and 7 MFS pa-tients,135 evidence of C. jejuni infection was presentin 2 of the 7 MFS patients, and was associated withaxonal degeneration, slow recovery, and residual dis-ability. C. jejuni GBS is commonly associated withformation of IgG antibodies against GM1, GM1b,GD1a, or GalNAc-GD1a.177

Jacobs et al.62 isolated C. jejuni from 3 MFS pa-tients, all with high anti-GQ1b IgG antibody titers,and demonstrated cross-reactivity of these antibodieswith surface epitopes of C. jejuni strains by enzyme-linked immunosorbent assay inhibition techniques.The findings support the hypothesis of molecularmimicry between bacteria and nerve tissue. In an-other series, C. jejuni was isolated from 3 MFS pa-tients with anti-GQ1b IgG, which cross-reacted withsialidase-sensitive epitopes in the lipopolysaccharidefractions, supporting the hypothesis that anti-GQ1bIgG antibodies are induced during preceding C. je-juni infection.64 There was further evidence to sug-gest that GBS and MFS induced long-lasting elevatedtiters of IgG1 and IgG3 antibodies, as compared withIgG2 in normal controls, suggesting that specificityof antibody isotype may be determined by precedingC. jejuni infection.63 Schwerer et al.147 showed that,in MFS following respiratory infection, IgG3 was themajor antibody detected, in contrast to IgA,76 IgM,or IgG2, following gastrointestinal infections. A Jap-anese study determined that Penner’s serotype 2 ofC. jejuni causing enteritis was particularly associatedwith MFS,122 in contrast to Penner’s serotype 19,which mainly triggered GBS.155,173

Recent research on C. jejuni has focused on itsgenetic aspects. A Campylobacter gene (cstII) wasfound to be associated with immune-mediated neu-ropathy by the demonstration that cstII sialic acidtransferase was a crucial determinant in the synthesisof GQ1b epitope.163 Oligoclonal expansions of Tcells bearing particular types of T-cell receptor Vbeta and V delta genes frequently occur in GBS/MFS, suggesting a role of T-cell mediation.77 Em-ploying Campylobacter knockout mutants and associ-ation studies of lipo-oligosaccharide biosynthesisgene locus with expression of ganglioside-mimickingstructures, it was shown that specific bacterial genesare crucial for induction of anti-ganglioside antibod-


ies.44 Koga et al.74 demonstrated that patients in-fected with C. jejuni Asn51 polymorphism were moreoften positive for anti-GQ1b IgG antibody andshowed ataxia. In contrast, patients with C. jejuniThr51 were more frequently positive for anti-GM1and anti-GD1a IgG antibody, and these patientsmanifested weakness predominantly. This findingsuggested that genetic polymorphism of C. jejunidetermines autoantibody reactivity and clinical pre-sentation in GBS/MFS.

MOLECULAR MIMICRY

Molecular mimicry represents a major pathogenicmechanism, and understanding this process mayhave important implications for treatment.

Studies of molecular mimicry in MFS, in relationto anti-GQ1b IgG antibody, have been carried outwith strategies similar to those in GBS. Yuki et al.175

isolated two strains of C. jejuni from patients withanti-GQ1b IgG antibody. The monoclonal antibod-ies cloned reacted with both lipopolysaccharide frac-tions, implying that the lipopolysaccharide of C. je-juni bears GQ1b epitopes, and mimicry betweenGQ1b and C. jejuni lipopolysaccharide has occurred.Salloway et al.140 isolated lipopolysaccharide from C.jejuni (serotype O:10) from an MFS patient and dem-onstrated that its terminal trisaccharide epitope con-sisted of two molecules of sialic acid linked to galac-tose, reflecting the terminal region of human GD3ganglioside, which is also present in the lipopolysac-charide of neuropathic O:19 strains of C. jejuni. Thissuggests a possible role in molecular mimicry of thistrisaccharide. In a separate study,8 the lipopolysac-charide of the OH 4384 strain of C. jejuni containedGM1 and GD1a-like epitopes. Immunization of micewith lipopolysaccharides of this strain inducedmonoclonal antibodies with GQ1b reactivity. It wasfurther demonstrated4 that one of the four rabbitsinjected with lipopolysaccharides from two MFS-related C. jejuni strains produced anti-GQ1b IgGantibody. These findings further attest to the pres-ence of molecular mimicry in the autoimmune de-velopment of MFS preceded by C. jejuni infection.However, animal experiments have demonstrateddifferences in the specificity of anti-ganglioside anti-body responses between rabbits immunized with li-popolysaccharides from the same Campylobacterstrain, suggesting that lipopolysaccharides onlypartly determine anti-ganglioside antibody specific-ity. Other strain-specific and host-related factors maycontribute.4

Koga et al.75 found serological evidence of H.influenzae in 7% of 70 consecutive MFS patients, all

with antecedent respiratory tract infection. Anti-GT1a IgG antibody cross-reactivity with GQ1b waspresent in all the patients. Their anti-GQ1b mono-clonal antibodies bound to the lipopolysaccharidesextracted from H. influenzae samples, suggesting thatthis lipopolysaccharide contained GT1a epitope.Molecular mimicry may thus be the likely mecha-nism for development of MFS after H. influenzaeinfection. Recent studies have suggested, however,that H. influenzae isolation may not always be indic-ative of its causal role in MFS/GBS, and testing forother infections should be undertaken during clini-cal management.73 There is also evidence to suggestthat antibodies to the vacuolating cytotoxin of Heli-cobacter pylori show homology with membrane ion-transport proteins, suggesting a role in MFS patho-genesis.30 Research into GBS/MFS has extended tocellular lipidomics, where differential effects onphospholipids activity with anti-GM1 vs. anti-GQ1bantibody was demonstrated in patients’ sera.55

In summary, collective evidence from bacterio-logical and immunological studies strongly supportsmolecular mimicry as the major pathogenic mecha-nism, but the presence of other genetic and hostdeterminants is becoming apparent. With advancesin biochemical and cellular technology, there will besignificant new developments in this field of diseasepathogenesis.

ANTI-GQ1b IgG ANTIBODY

Gangliosides are glycosphingolipids that contain asialic acid residue of N-acetylneuraminic acid at-tached to the terminal galactose of an oligosaccha-ride core. They play a role in plasma membrane andcell functions. The hydrophilic carbohydrate struc-ture is exposed extracellularly, and is capable ofacting as an autoantibody target. The gangliosideGQ1b (Fig. 2) is abundant in cranial nerves supply-ing extraocular muscles and the presynaptic neuro-muscular junction, which is devoid of a blood–nervebarrier. This may render it vulnerable to autoim-mune attack. The corresponding antibody, anti-GQ1b, is IgG in class and its titer rapidly decreaseswith clinical resolution of MFS.

Anti-GQ1b antibody is well known to be associ-ated with MFS.126 Anti-GQ1b IgG antibody titerspeak at clinical presentation and decline rapidly dur-ing the course of recovery. However, some studieshave measured its titer over the first 2 weeks afteronset and found that anti-GQ1b IgG antibody activ-ity reflected clinical severity scores, especially oph-thalmoplegia.45,103 Anti-GQ1b IgG antibody fre-quently cross-reacts with ganglioside GT1a, as well as


GD3 and GD1b.179 Such demonstrated specificitycan be useful to the managing clinician, who canconsider testing for the presence of anti-GQ1b IgGantibody activity or antibodies to these relatedepitopes, as markers for identification of MFS or itsformes frustes. Longitudinal testing has shown thatanti–C. jejuni IgG and IgA antibody titers parallelthat of anti-ganglioside antibodies, indicating that C.jejuni infection triggers antibody production.116 Anti-GQ1b IgG antibody positivity, occurring in variousfrequencies, has been reported in MFS, GBS, andBBE.5,117 From the clinical standpoint, anti-GQ1bIgG antibody positivity is most highly correlated withophthalmoplegia,154 but an association with ataxiafrom binding to dorsal root ganglia81 has been doc-umented. Associations with pure ataxia36,106 or iso-lated ophthalmoplegia174 have also been described.Anti-GQ1b IgG antibody testing was found to bestatistically superior to cerebrospinal fluid examina-tion for MFS diagnosis in the first 3 weeks of presen-tation.111 These clinical studies point to the presenceof a range of manifestations, from isolated ophthal-moplegia to the classic triad of MFS, in anti-GQ1bIgG antibody–positive patients.

It should be noted that the notion of an anti-GQ1b IgG antibody syndrome was not intended as anew diagnosis, but rather as recognition of a closeetiological relation between MFS, GBS, and BBE.117

IMMUNOLOGICAL ASPECTS

Although MFS, like GBS, is an immune-mediatedcondition, it remains uncertain as to how antibodymediation can be translated to motor manifesta-tions. Increased activity of anti-GQ1b IgG antibodywas shown to be present in most cases of MFS,29,167

and anti-GQ1b mouse monoclonal antibody immu-nostained the paranodal region of extramedullaryIII, IV, and VI cranial nerves. The paranodal regionsare important for nerve impulse conduction. It wasalso demonstrated with anti-GQ1b monoclonal anti-body that staining of GQ1b occurred specifically andwas densely localized in the paranodal myelin sheathof cranial nerves III, IV, and VI, but not in othercentral nervous system structures.27 Cranial nerve IIIwas shown to contain a larger amount of GQ1b thanspinal nerves roots, suggesting a role of anti-GQ1bIgG antibody in ocular manifestations of MFS.28

The motor effects of anti-GQ1b IgG antibodyremain poorly understood. Roberts et al.136 first in-vestigated the effects of anti-GQ1b IgG antibody–positive sera on mouse phrenic nerve–diaphragmpreparations, and showed that miniature endplatepotential frequencies increase, decline rapidly, andthen cease after nerve stimulation, suggesting failureof neuromuscular transmission from nerve terminalsin MFS. Patch-clamp techniques have shown thatreversible presynaptic neurotransmitter releaseblockade may contribute to muscle weakness inMFS,18 possibly due to the interference of calciuminflow or proteins in the exocytic apparatus.19

Further studies using similar methods followed,showing that anti-GQ1b IgG antibody binds to theneuromuscular junction, inducing a massive quantalrelease of acetylcholine, resembling effects of theneurotoxin alpha-latrotoxin. Experiments with com-plement-deficient sera suggested that anti-GQ1b IgGantibody acts in conjunction with activated comple-ment in the alternate pathways.20,131 It was subse-quently shown that circulating IgG antibodies inMFS could induce neuromuscular blockade at thepresynaptic and postsynaptic levels.17 Using immu-nofluorescence, electron microscopy, and micro-electrode recording, complement-mediated ultra-structural axon terminal and perisynaptic Schwann-cell destruction were demonstrated.49,115

These findings have led to studies on interven-tion. Intravenous immunoglobulin is commonlyused in the treatment of many immune-mediateddiseases, but the underlying mechanisms of actionare not well understood. It inhibits binding of anti-GQ1b IgG antibody to GQ1b and prevents comple-ment activation in ex vivo mouse models.65 Calpain

FIGURE 2. Molecular structure of ganglioside GQ1b, depictingthe ceramide portion within the plasma membrane, and the car-bohydrate structures exposed to extracellular fluid. Shaded: N-acetylneuraminic acid; grid: galactose; dots: N-acetylgalac-tosamine; lines: N-acetylglucosamine.


inhibitors may limit calcium influx and have beendemonstrated to reduce ultrastructural nerve termi-nal damage in mouse hemidiaphragm treated withanti-GQ1b IgG antibody, complement, or alpha-lat-rotoxin.114 Most recently, use of AP7070 (Microco-cept), a complement activation inhibitor, was shownto prevent the membrane attack complex cascadeformation both in vitro and in vivo,48 suggesting arole for nerve terminal neuroprotection in GBS orMFS. It remains to be seen whether these novelfindings can be translated to human trials in thetreatment of MFS and its related disorders.

Apart from anti-GQ1b, other antibodies associ-ated with development of MFS include anti-GT1a,48,59,113,151,175 anti-LM1,52,170 and anti-GD3,178

but anti-GM1 antibody titer elevation seems uncom-mon.98 Anti-GQ1b IgG antibody in MFS shows ex-ceptionally high cross-reactivity with gangliosideGT1a,148 but anti-GT1a antibody without GQ1b re-activity is less commonly observed. In particular, theataxic form of GBS has anti-GT1a as a commonantibody.176 Ataxic GBS has also been reported to beassociated with anti-GM1b and anti–GalNacAc-GD1aIgG antibodies.119 Cross-reaction of anti-GQ1b IgGwith GD1b was particularly evident in MFS patientswith impaired proprioception.153 Detection of anti-ganglioside antibody by agglutination assay may beas useful a rapid method as enzyme-linked immu-nosorbent assay.1 Although testing for anti-GQ1bIgG antibody is significant, recent studies have high-lighted the clinical importance of ganglioside com-plexes by showing that anti–GQ1b/GM1-positiveMFS patients were less likely to develop sensory dis-turbances.68

A study of anti-GQ1b IgG antibody–positive seraof MFS and GBS patients did not reveal any associ-ation of human leukocyte antigen types with immu-noresponse of anti-GQ1b IgG antibody.26 Elevatedlevels of interferon-gamma and T-helper 1 werefound,56 in keeping with other studies suggestingT-cell mediation in MFS pathogenesis.77 Reducedcerebrospinal fluid levels of hypocretin-1 were foundin 5 of 12 (42%) MFS patients. As hypocretin-1 isknown to be associated with sleep–wake cycles, fur-ther studies will help elucidate whether such patientshave disturbances in this physiological process.112

In summary, the pathogenesis of MFS is moststrongly associated with anti-GQ1b IgG antibody,and research into its treatment has been directed atthe effects of this antibody. However, it has movedbeyond this area, and downstream processes in theimmunological cascade may emerge as promisingtargets for future therapeutic intervention strategies.It can be speculated that research findings in this

direction may also have relevance for other post-infectious immunologically mediated medical condi-tions.

REFERENCES

1. Alaedini A, Briani C, Wirguin I, Siciliano G, D’Avino C, LatovN. Detection of anti-ganglioside antibodies in Guillain–Barresyndrome and its variants by the agglutination assay. J NeurolSci 2002;196:41–44.

2. Al-Din SN, Anderson M, Eeg-Olofsson O, Trontelj JV. Neuro-ophthalmic manifestations of the syndrome of ophthalmo-plegia, ataxia and areflexia: a review. Acta Neurol Scand1994;89:157–163.

3. Ang CW, De Klerk MA, Endtz HP, Jacobs BC, Laman JD, vander Meche FG, et al. Guillain–Barre syndrome and MillerFisher syndrome–associated Campylobacter jejuni lipopolysac-charides induce anti-GM1 and anti-GQ1b antibodies in rab-bits. Infect Immun 2001;69:2462–2469.

4. Ang CW, Laman JD, Willison HJ, Wagner ER, Endtz HP, DeKlerk MA, et al. Structure of Campylobacter jejuni lipopolysac-charides determines antiganglioside specificity and clinicalfeatures of Guillain–Barre and Miller Fisher patients. InfectImmun 2002;70:1202–1208.

5. Arai M, Odaka M, Yuki N, Hirata K. A patient with overlap-ping Bickerstaff’s brainstem encephalitis, Miller Fisher syn-drome and Guillain–Barre syndrome during the clinicalcourse. Eur J Neurol 2002;9:115–116.

6. Arakawa Y, Yoshimura M, Kobayashi S, Ichihashi K, Miyao M,Momoi MY, et al. The use of intravenous immunoglobulin inMiller Fisher syndrome. Brain Dev 1993;15:231–233.

7. Arranz Caso JA, Martinez R, Cabrera F, Tejeiro J. MillerFisher syndrome in a patient with HIV infection. AIDS 1997;15:550–551.

8. Aspinall GO, Fujimoto S, McDonald AG, Pang H, KurjanczykLA, Penner JL. Lipopolysaccharides from Campylobacter jejuniassociated with Guillain–Barre syndrome patients mimic hu-man gangliosides in structure. Infect Immun 1994;62:2122–2125.

9. Barontini F, Di Lollo S, Maurri S, Lambruschini P. Localiza-tion of the pathological process in Miller Fisher syndrome.Ital J Neurol Sci 1992;13:221–225.

10. Benito-Leon J, Bravo J, Mateos F, Simon R. Miller Fishersyndrome in infancy. Child Nerv Syst 1996;12:559–561.

11. Bennetto LP, Lyons P. Miller Fisher syndrome associatedwith Pastuerella multocida infection. J Neurol Neurosurg Psy-chiatry 2004;75:1786–1787.

12. Berlit P, Rakicky J. The Miller Fisher syndrome. Review ofthe literature. J Clin Neuroopthalmol 1992;12:57–63.

13. Bingisser R, Speich R, Fontana A, Gmur J, Vogel B, Landis T.Lupus erythematosus and Miller Fisher syndrome. ArchNeurol 1994;51:828–830.

14. Blau I, Casson I, Lieberman A, Weiss E. The not-so-benignMiller Fisher syndrome: a variant of Guillain–Barre syn-drome. Arch Neurol 1980;37:384–385.

15. Bogliun G, Beghi E, Italian GBS Registry Study Group. Inci-dence and clinical features of acute inflammatory polyra-diculopathy in Lombardi, Italy, 1996. Acta Neurol Scand2004;110:100–106.

16. Brittain CJ, Lake D. Acute angle closure in Miller Fishersyndrome. Eye 2006;20:739–740.

17. Buchwald B, Bufler J, Carpa M, Heidenreich F, Pitz R, DudelJ, et al. Combined pre- and post-synaptic action of IgGantibodies in Miller Fisher syndrome. Neurology 2001;56:67–74.

18. Buchwald B, Weishaupt A, Toyka KV, Dudel J. Immunoglob-ulin G from a patient with Miller Fisher syndrome rapidlyand reversibly depresses evoked quantal release at the neu-


romuscular junction of mice. Neurosci Lett 1995;201:163–166.

19. Buchwald B, Weishaupt A, Toyka KV, Dudel J. Pre- andpostsynaptic blockade of neuromuscular transmission byMiller Fisher syndrome IgG at mouse motor nerve terminals.Eur J Neurosci 1998;10:281–290.

20. Bullens RW, O’Hanlon GM, Goodyear CS, Molenaar PC,Conner J, Willison HJ, et al. Anti-GQ1b antibodies andevoked acetylcholine release at mouse motor endplates.Muscle Nerve 2000;23:1035–1043.

21. Caccavale A, Mignemi L. Acute onset of a bilateral areflexi-cal mydriasis in Miller Fisher syndrome: a rare neuro-oph-thalmologic disease. J Neuroophthalmol 2000;20:61–62.

22. Carpo M, Pedotti R, Allaria S, Lolli F, Mata S, Cabaletti G, etal. Clinical presentation and outcome of Guillain–Barre andrelated syndromes in relation to anti-ganglioside antibodies.J Neurol Sci 1999;168:78–84.

23. Chan JW. Optic neuritis in anti-GQ1b positive recurrentMiller Fisher syndrome. Br J Ophthalmol 2003;87:1185–1186.

24. Chan YC, Wilder-Smith E, Chee MW. Acute ophthalmople-gia with pupillary areflexia associated with anti-GQ1b anti-body. J Clin Neurosci 2004;11:658–660.

25. Cheng BC, Chang WN, Chang CS, Chee CY, Huang CR,Chen JB, et al. Guillain–Barre syndrome in southern Taiwan:clinical features, prognostic factors and therapeutic out-comes. Eur J Neurol 2003;10:655–662.

26. Chiba A, Kusunoki S, Kuwata S, Juji T, Shibata Y, KanazawaI. HLA and anti-GQ1b antibody in Miller Fisher syndromeand Guillain–Barre syndrome. J Neuroimmunol 1995;61:85–88.

27. Chiba A, Kusunoki S, Obata H, Machinami R, Kanazawa I.Ganglioside composition of the human cranial nerves, withspecial references to pathophysiology of Miller Fisher syn-drome. Brain Res 1997;754:32–36.

28. Chiba A, Kusunoki S, Obata H, Machinami R, Kanazawa I.Serum anti-GQ1b IgG antibody is associated with ophthal-moplegia in Miller Fisher syndrome and Guillain–Barre syn-drome: clinical and immunohistochemical studies. Neurol-ogy 1993;43:1911–1917.

29. Chiba A, Kusunoki S, Shimizu T, Kanazawa I. Serum IgGantibody to ganglioside GQ1b is a possible marker of MillerFisher syndrome. Ann Neurol 1992;31:677–679.

30. Chiba A, Sugiyama T, Yonekura K, Tanaka S, Matsumoto H,Fujii N, et al. An antibody to VacA of Helicobacter pylori in theCSF of patients with Miller Fisher syndrome. Neurology2004;63:2184–2186.

31. Chiba S, Sugiyama T, Yonekura K, Tanaka S, Matsumoto H,Fujii N, et al. An antibody to VacA of Helicobacter pylori incerebrospinal fluid of patients with Guillain–Barre syn-drome. J Neurol Neurosurg Psychiatry 2002;73:76–78.

32. Chikakiya H, Kunishige M, Yoshino H, Asano A, SumitomoY, Endo I, et al. Delayed motor and sensory neuropathy in apatient with brainstem encephalitis. J Neurol Sci 2005;234:105–108.

33. Colding-Jorgensen E, Vissing J. Visual impairment in anti-GQ1b positive Miller Fisher syndrome. Acta Neurol Scand2001;103:259–260.

34. Collier J. Peripheral neuritis. Edinburgh Med J 1932;39:601–618.

35. Csepany T, Boczan J, Magyar MT, Molnar S, Csiba L, DeesyJ, et al. Miller Fisher syndrome—a presenting clinical man-ifestation of lung cancer in a previously apparently healthyindividual. J Neurol 2004;251:898–900.

36. Derakhshan I. Ataxic Guillain–Barre syndrome with anti-GQ1b antibody: relation to Miller Fisher syndrome. Neurol-ogy 2000;55:1419–1420.

37. Desai SS, Allen E, Deodhar A. Miller Fisher syndrome inadult onset Still’s disease: case report and review of theliterature of other neurological manifestations. Rheumatol-ogy (Oxford) 2002;41:216–222.

38. Diaz Ortuna A, Maeztu C, Munoz JA, Reigadas R, Reigadas R,Rodriguez T, et al. Miller Fisher syndrome associated with Qfever. J Neurol Neurosurg Psychiatry 1990;53:61561–61516.

39. Finsterer J, Niedermayr A, Weigl PG, Voigtlander T. Anti-GQ1b-negative Miller Fisher syndrome with lower cranialnerve involvement from parasinusoidal Aspergilloma. J In-fect 2006; 2: 81–85.

40. Fisher M. An unusual variant of acute idiopathic polyneuritis(syndrome of ophthalmoplegia, ataxia and areflexia).N Engl J Med 1956;255:57–65.

41. Fross RD, Daube JR. Neuropathy in the Miller Fisher syn-drome: clinical and electrophysiologic findings. Neurology1987;37:1493–1498.

42. Giroud M, Mousson C, Chalopin JM, Rifle G, Dumas R.Miller Fisher syndrome and pontine abnormalities on MRI:a case report. J Neurol 1990;237:489–490.

43. Godschalk PC, Bergman MP, Gorkink RF, Simons G, vandern Braak N, Lastovica AJ, et al. Identification of DNAsequence variation in Campylobacter strains associated withthe Guillain–Barre syndrome by high-throughput AFLPanalysis. BMC Microbiol 2006;6:32.

44. Godschalk PC, Heikema AP, Gilbert M, Komagamine T, AngCW, Glerum J, et al. The crucial role of Campylobacter jejunianti-ganglioside antibody induction in Guillain–Barre syn-drome. J Clin Invest 2004;114:1659–1665.

45. Goffete S, Jeanjean A, Pierret F, Peeters A, Sindic CJ. Clinicalrelevance of the determination of anti-GQ1b antibodies inMiller Fisher and Guillain–Barre syndrome. Acta NeurolBelg 1998;98:322–326.

46. Goldberg-Stern H, Melamed E, Gadoth N. Abnormal evokedpotentials in Miller Fisher syndrome: further evidence ofcombined peripheral and central demyelination. J NeurolNeurosurg Psychiatry 1994;57:506.

47. Guiloff RJ. Peripheral nerve conduction in Miller Fishersyndrome. J Neurol Neurosurg Psychiatry 1977;40:801–807.

48. Halstead SK, Humphreys PD, Goodfellow JA, Wagner ER,Smith RA, Willison HJ. Complement inhibition abrogatesnerve terminal injury in Miller Fisher syndrome. Ann Neurol2005;58:203–210.

49. Halstead SK, Morrison I, O’Hanlon GM, Humphreys PD,Goodfellow JA, Plomp JJ, et al. Anti-disialosyl antibodiesmediate selective neuronal or Schwann cell injury at mouseneuromuscular junctions. Glia 2005;53:177–189.

50. Halstead SK, O’Hanlon GM, Humphreys PD, Morrison DB,Morgan DP, Todd AJ, et al. Anti-disialoside antibodies killpresynaptic Schwann cells and damage motor nerve termi-nals via membrane attack complex in a murine model ofneuropathy. Brain 2004;127:2109–2123.

51. Hamaguchi T, Yamaguchi K, Komai K, Yamada M, Yokoji H,Satake R, et al. Recurrent anti-GQ1b IgG antibody syndromeshowing different phenotypes in different periods. J NeurolNeurosurg Psychiatry 2003;74:1350.

52. Harukawa H, Utsumi H, Asano A, Yoshino H. Anti-LM1antibodies in the sera of patients with Guillain–Barre syn-drome, Miller Fisher syndrome, and motor neuron disease.J Peripher Nerv Syst 2002;7:54–58.

53. Hatanaka T, Higashino H, Yasuhara A, Kobayashi Y. MillerFisher syndrome: etiological significance of serial blink re-flexes and MRI study. Electromyogr Clin Neurophysiol 1992;32:317–319.

54. Hess DC, Awad E, Posas H, Sethi KD, Adams RJ. Miller Fishersyndrome in systemic lupus erythematosus. J Rheumatol1990;17:1520–1522.

55. Hirakawa M, Morita D, Tsuji S, Kusunoki S. Effects of phos-pholipids on antigangioside antibody reactivity. J Neuroim-munol 2005;159:129–132.

56. Hohnoki K, Inoue A, Koh CS. Elevated serum levels ofIFN-gamma, IL-4 and TNF-alpha/unelevated serum levels ofIL-10 in patients with demyelinating diseases during theacute stage. J Neuroimmunol 1998;87:27–32.

57. Hosoya T, Adachi M, Yamaguchi K, Kato T, Sugai Y. Abdu-cens nerve enhancement demonstrated by multiplanar re-


construction of contrast-enhanced three-dimensional MRI.Neuroradiology 201;43:295–301.

58. Hung PL, Chang WN, Huang LT, Huang SC, Chang YC,Chang CJ, et al. A clinical and electrophysiologic survey ofchildhood Guillain–Barre syndrome. Pediatr Neurol 2004;30:86–91.

59. Ilyas AA, Cook SD, Mitchen FA, Taki T, Kasama T, Handa S,et al. Antibodies to GT1a ganglioside in patients with Guil-lain–Barre syndrome. J Neuroimmunol 1998;82:160–167.

60. Inoue A, Koh C, Iwahashi T. Detection of serum anticere-bellar antibodies in patients with Miller Fisher syndrome.Eur Neurol 1999;42:230–234.

61. Inoue N, Ichimura H, Goto S, Hashimoto Y, Ushio Y. MRimaging findings of spinal posterior column involvement ina case of Miller Fisher syndrome. Am J Neuroradiol 2004;25:645–648.

62. Jacobs BC, Endtz H, van der Meche FG, Hazenberg MP,Achtereekte HA, van Doorn HA. Serum anti-GQ1b IgG an-tibodies recognize surface epitopes on Campylobacter jejunifrom patients with Miller Fisher syndrome. Ann Neurol1995;37:260–364.

63. Jacobs BC, Endtz HP, van der Meche FG, Hazenberg MP, deKlerk MA, van Doorn PA. Humoral immune responseagainst Campylobacter jejuni lipopolysaccharides in Guillain–Barre and Miller Fisher syndrome. J Neuroimmunol 1997;79:62–68.

64. Jacobs BC, O’Hanlon GM, Breedland EG, Veitch J, vanDoorn PA, Willison HJ. Human IgM paraproteins demon-strate shared reactivity between Campylobacter jejuni lipopoly-saccharides and human peripheral nerve disialylated gan-gliosides. J Neuroimmunol 1997;80:23–30.

65. Jacobs BC, O’Hanlon GM, Bullens RW, Veitch J, Plomp J,Willison HJ. Immunoglobulins inhibit pathophysiological ef-fects of anti-GQ1b-positive sera at motor nerve terminalsthrough inhibition of antibody binding. Brain 2003;126:2220–2234.

66. Jamal GA, Ballantyne JP. The localization of the lesion inpatients with acute ophthalmoplegia, ataxia and areflexia(Miller Fisher syndrome). A serial multimodal neurophysio-logical study. Brain 1988;111:95–114.

67. Jamal GA, MacLeod WN. Electrophysiologic studies in MillerFisher syndrome. Neurology 1984;34:685–688.

68. Kaida KL, Kanzaki M, Morita D, Kamakura K, Motoyoshi K,Hirakawa M, et al. Anti-ganglioside complex antibodies inMiller Fisher syndrome. J Neurol Neurosurg Psychiatry 2006;77:1043–1046.

69. Kaplan JG, Schaumburg HH, Sumner A. Relapsing ophthal-moparesis–sensory neuropathy syndrome. Neurology 1985;35:595–596.

70. Katsuno M, Ando T, Hakusui S, Yanagi T, Sobue G. Motorconduction studies in Miller Fisher syndrome with severetetraparesis. Muscle Nerve 2002;25:378–382.

71. Kaushik P, Cohen AJ, Zuckerman SJ, Vatsavai SR, Pepper SJ,Banda VR, et al. Miller Fisher variant of Guillain–Barresyndrome requiring a cardiac pacemaker in a patient ontacrolimus after liver transplantation. Ann Pharmacother2005;39:1124–1127.

72. Koga M, Gilbert M, Li J, Koike S, Takahashi M, Furakawa K, etal. Antecedent infections in Fisher syndrome: a commonpathogenesis of molecular mimicry. Neurology 2005;64:1605–1611.

73. Koga M, Koike S, Hirata K, Yuki N. Ambiguous value ofHemophilus influenzae isolation in Guillain–Barre and Fishersyndromes. J Neurol Neurosurg Psychiatry 2005;76:1736–1738.

74. Koga M, Takahashi M, Masuda M, Hirata K, Yuki N. Campy-lobacter gene polymorphism as a determinant of clinical fea-tures of Guillain–Barre syndrome. Neurology 2005;65:1376–1381.

75. Koga M, Yuki N, Tai T, Hirata K. Miller Fisher syndrome andHemophilus influenzae infection. Neurology 2001;57:686–691.

76. Koga M, Yuki N, Takahashi M, Saito K, Hirata K. Closeassociation of IgA anti-ganglioside antibodies with anteced-ent Campylobacter jejuni infection in Guillain–Barre and Fish-er’s syndrome. J Neuroimmunol 1998;81:138–143.

77. Koga M, Yuki N, Tsukada Y, Hirata K, Matsumoto Y. CDR3spectratyping analysis of the T cell receptor repetoire inGuillain–Barre and Fisher syndromes. J Neuroimmunol2003;141:112–117.

78. Koga M, Yuki N, Hirata K. Subclass distribution and thesecretory component of serum IgA anti-ganglioside antibod-ies in Guillain–Barre syndrome after Campylobacter jejuni en-teritis. J Neuroimmunol 1999;96:245–250.

79. Koga M. Yuki N, Hirata K. Pain in Miller Fisher syndrome.J Neurol 2000;247;720–721.

80. Kornberg AJ, Pestronk A, Blume GM, Lopate G, Yue J, HahnA. Selective staining of the cerebellar molecular layer byserum IgG of Miller Fisher and related syndromes. Neurol-ogy 1996;47:1317–1320.

81. Kusunoki S, Chiba A, Kanazawa I. Anti-GQ1b IgG antibody isassociated with ataxia as well as ophthalmoplegia. MuscleNerve 1999;22:1071–1074.

82. Kuwabara S, Asahina M, Nakajima M, Mori M, Fukutake T,Hattori T, et al. Special sensory ataxia in Miller Fisher syn-drome detected by postural body sway analysis. Ann Neurol1999;45:533–536.

83. Lange DJ, Deangelis T, Sivak MA. Single-fiber electromyo-graphy shows terminal axon dysfunction in Miller Fishersyndrome: a case report. Muscle Nerve 2006;34:232–234.

84. Li HF, Xie YC. Posterior column involvement in MillerFisher syndrome or dorsal root ganglion neuronopathy?Am J Neuroradiol 2005;26:1881.

85. Littelwood R, Bajada S. Successful plasmapharesis in theMiller Fisher syndrome. Br Med J (Clin Res Ed) 1981;282:778.

86. Lo YL, Chan LL, Pan A, Ratnagopal P. Acute ophthalmopa-resis with anti-GQ1b antibody: electrophysiological evidenceof neuromuscular transmission defect in the orbicularis oc-uli. J Neurol Neurosurg Psychiatry 2003;75:436–440.

87. Lo YL, Dan YF, Leoh TH, Tan YE, Nurjannah S, RatnagopalP. The effect of exercise on repetitive nerve stimulationstudies: new appraisal of an old technique. J Clin Neuro-physiol 2004;21:110–113.

88. Lo YL, Dan YF, Leoh TH, Tan YE, Ratnagopal P. Decrementin area of muscle responses to repetitive nerve stimulation.Muscle Nerve 2003;27:494–496.

89. Lo YL, Leoh TH, Dan YF, Tan YE, Nurjannah S, RatnagopalP. Repetitive stimulation of the long thoracic nerve. J NeurolNeurosurg Psychiatry 2003;74:379–381.

90. Lo YL, Leoh TH, Dan YF, Lim LL, Seah A, Fook-Chong S, etal. Presynaptic neuromuscular transmission defect in theMiller Fisher syndrome. Neurology 2006;66:148–149.

91. Lo YL, P Ratnagopal. Transcranial magnetic stimulationstudies in the Miller Fisher syndrome: evidence of cortico-spinal tract abnormality. J Neurol Neurosurg Psychiatry2001;71:210–214.

92. Lo YL, Ratnagopal P. Corticobulbar dysfunction in theMiller Fisher syndrome. Clin Neurol Neurosurg 2003;105:156–158.

93. Lyu RK, Tang LM, Cheng SY, Hsu WC, Chen ST. Guillain–Barre syndrome in Taiwan: a clinical study of 167 patients.J Neurol Neurosurg Psychiatry 1997;63:494–500.

94. Lyu RK, Tang LM, Hsu WC, Chen ST. Quantitative cardio-vascular autonomic function study in Fisher syndrome.J Neurol Neurosurg Psychiatry 2002;73:333–335.

95. Madhavan S, Geetha N, Bhargavan PV. Recurrent MillerFisher syndrome. J Assoc Physicians India 2004;52:582–584.

96. Maisey DN, Olczak SA. Successful plasmapharesis in theMiller Fisher syndrome. Br Med J (Clin Res Ed) 1981;282:1159.

97. Matsumoto H, Kobayashi O, Tamura K, Ohkawa T, Sekine I.Miller Fisher syndrome with transient coma: comparison


with Bickerstaff’s brainstem encephalitis. Brain Dev 2002;24:98–101.

98. McCombe PA, Wilson R, Prentice RL. Results of testing foranti-GM1 antibodies. J Clin Neurosci 2000;7:209–212.

99. Merkx H, Keyser J, Ebinger G. Miller Fisher syndrome asso-ciated with Mycoplasma pneumoniae infection: report of a case.Clin Neurol Neurosurg 1994;96:96–99.

100. Minguez M, Reus S, Pascual E. Dorsal-lumbar pain as thefirst manifestation of Miller Fisher syndrome. Clin ExpRheumatol 1998;16:509.

101. Minoda R, Uno K, Toriya T, Eura M, Noguchi S, MasuyamaK. Neurologic and otologic findings in Fisher’s syndrome.Auris Nasus Larynx 1999;26:153–158.

102. Miyazaki T, Kusunoki S, Kaida K, Shiina M, Kanazawa I.Guillain–Barre syndrome associated with IgG monospecificto ganglioside GD1b. Neurology 2001;56:1227–1229.

103. Mizoguchi K. Anti-GQ1b IgG antibody activities related tothe severity of Miller Fisher syndrome. Neurol Res 1998;20:617–624.

104. Mori M, Kuwabara S, Fukutake T, Hattori T. Plasmapharesisand Miller Fisher syndrome: analysis of 50 consecutive cases.J Neurol Neurosurg Psychiatry 2002;72:680.

105. Mori M, Kuwabara S, Fukutake T, Yuki N, Hattori T. Clinicalfeatures and prognosis of Miller Fisher syndrome. Neurology2001;56:1104–1106.

106. Mori M, Kuwabara S, Koga M, Asahina M, Ogawara K, Hat-tori T, et al. IgG anti-GQ1b antibody positive acute ataxiawithout ophthalmoplegia. J Neurol Neurosurg Psychiatry1999;69:668–670.

107. Nagaoku U, Kato T, Kurita K, Arawaka S, Hosoya T, Yuki N,et al. Cranial nerve enhancement on three-dimensional MRIin Miller Fisher syndrome. Neurology 1996;47:1601–1602.

108. Nakajima M. Posterior reversible encephalopathy complicat-ing intravenous immunoglobulins in a patient with MillerFisher syndrome. J Neurol 2005;54:58–60.

109. Nakatsuji Y, Sadahiro S, Watanabe S, Fujimura H, Abe K,Koguchi K, et al. Leptomeningeal signet-ring cell carcino-matosis presenting with ophthalmoplegia, areflexia andataxia. Clin Neuropathol 2001;20:272–275.

110. Neisser A, Bernheimer H, Berger T, Moran AP, Schwerer B.Serum antibodies against gangliosides and Campylobacter je-juni lipopolysaccharides in Miller Fisher syndrome. InfectImmun 1997;65:4038–4042.

111. Nishimoto Y, Odaka M, Hirata K, Yuki N. Usefulness ofanti-GQ1b IgG antibody testing in Fisher syndrome com-pared with cerebrospinal fluid examination. J Neuroimmu-nol 2004;148:200–205.

112. Nishino S, Kanbayashi T, Fujiki N, Uchino M, Ripley B,Watanabe M, et al. CSF hypocretin levels in Guillain–Barresyndrome and other inflammatory neuropathies. Neurology2003;61:823–825.

113. O’Leary CP, Vetich J, Durward WF, Thomas AM, Rees JH,Willison HJ. Acute oropharyngeal palsy is associated withantibodies to GQ1b and GT1a gangliosides. J Neurol Neu-rosurg Psychiatry 1996;61:649–651.

114. O’Hanlon GM, Humphreys PD, Goldman RS, Halstead SK,Bullens RW, et al. Calpain inhibitors protect against axonaldegeneration in a model of anti-ganglioside antibody–medi-ated motor nerve terminal injury. Brain 2003;126:2497–2509.

115. O’Hanlon GM, Plomp JJ, Chakrabarti M, Morrison I, Wag-ner ER, Goodyear CS, et al. Anti-GQ1b ganglioside antibod-ies mediate complement-dependent destruction of the mo-tor nerve terminal. Brain 2001;124:893–906.

116. Odaka M, Koga M, Yuki N, Susuki K, Hirata K. Longitudinalchanges of anti-ganglioside antibodies before and after Guil-lain–Barre syndrome onset subsequent to Campylobacter jejunienteritis. J Neurol Sci 2003;21:99–103.

117. Odaka M, Yuki N, Hirata K. Anti-GQ1b antibody syndrome:clinical and immunological range. J Neurol Neurosurg Psy-chiatry 2001;70:50–55.

118. Odaka M, Yuki N, Hirata K. Bilateral ballism in a patient withoverlapping Fisher’s and Guillain–Barre syndromes. J Neu-rol Neurosurg Psychiatry 1999;67:206–208.

119. Odaka M, Yuki N, Tatsumoto M, Tateno M, Hirata K. AtaxicGuillain–Barre syndrome associated with anti-GM1b andanti-GalNac-GD1a antibodies. J Neurol 2004;251:24–29.

120. Odaka M, Yuki N, Yamada M, Koga M, Takemi T, Hirata K,et al. Bickerstaff’s brainstem encephalitis: clinical features of62 cases and a subgroup associated with Guillain–Barre syn-drome. Brain 2003;126:2279–2290.

121. Ogawara K, Kuwabara S, Yuki N. Fisher syndrome or Bick-erstaff brainstem encephalitis? Anti-GQ1b IgG antibody syn-drome involving both the peripheral and central nervoussystems. Muscle Nerve 2002;26:845–849.

122. Ohtsuka K, Nakamura Y, Hashimoto M, Tagawa Y, TakahashiM, Saito K, et al. Fisher syndrome associated with IgG anti-GQ1b antibody following infection by a specific serotype ofCampylobacter jejuni. Ophthalmology 1998;105;1281–1285.

123. Ohtsuka K, Nakamura Y, Tagawa Y, Yuki N. Immunoadsorp-tion therapy for Fisher syndrome associated with IgG anti-GQ1b antibody. Am J Ophthalmol 1998;125:403–406.

124. Oomura M, Yamawaki T, Oe H, Moriwaki H, Miyashita K,Naritomi H, et al. Association of cardiomyopathy caused byautonomic nervous system impairment with the MillerFisher syndrome. J Neurol Neurosurg Psychiatry 2003;74:689–690.

125. Orr CF, Storey CE. Recurrent Miller Fisher syndrome. J ClinNeurosci 2004;11:307–309.

126. Panda S, Tripathi M. Anti-GQ1b antibody syndrome: clinicaland immunological range. J Neurol Neurosurg Psychiatry2002;72:418–419.

127. Paparounas K. Anti-GQ1b ganglioside antibody in periph-eral nervous system disorders: pathophysiologic role andclinical relevance. Arch Neurol 2004;61:1013–1016.

128. Pedotti R, Carpo M, Lucchi S, Righini A, Scarlato G, Nobile-Orazio E. Lumbosacral root and facial nerve enhancementin Miller Fisher syndrome. J Neurol 1998;245:753–754.

129. Phillips MS, Stewart S, Anderson JR. Neuropathological find-ings in Miller Fisher syndrome. J Neurol Neurosurg Psychi-atry 1984;47:492–495.

130. Pinchoff BS, Slavin MI, Rosenstein D, Hyman R. Divergenceparalysis as the initial sign in the Miller Fisher syndrome.Am J Ophthalmol 1986;101:741–742.

131. Plomp JJ, Molenaar PC, O’Hanlon GM, Jacobs BC, Veitch J,Daha MR, et al. Miller Fisher anti-GQ1b antibodies: alpha-latrotoxin-like effects on motor end plates. Ann Neurol1999;45:189–199.

132. Polizzi A, Ruggieri M, Vecchio I, Genovese S, Rampello L,Raffacle R. Autoimmune thyroiditis and acquired demyeli-nating polyradiculoneuropathy. Clin Neurol Neurosurg2001;103:151–154.

133. Pulitano S, Viola L, Genovese O, Chiaretti A, Piastra M,Mariotti P, et al. Miller Fisher syndrome mimicking intracra-nial hypertension following head trauma. Child’s Nerv Syst2005;21:473–476.

134. Redel SW, Barnett MH, Yan WX, Halmagyi GM, Pollard JD.Chronic ophthalmoplegia with anti-GQ1b antibody. Neurol-ogy 2000;54:1000–1002.

135. Rees JH, Soudain SE, Gregson NA, Hughes RA. Campylobacterjejuni infection and Guillain–Barre syndrome. N Engl J Med1995;333:1374–1379.

136. Roberts M, Willison H, Vincent A, Newsom-Davis J. Serumfactor in Miller Fisher variant of Guillain–Barre syndromeand neurotransmitter release. Lancet 1994;343:454–455.

137. Ropper A. Absence of CNS lesion in autopsied patients withMiller Fisher syndrome. Arch Neurol 1985;42:15.

138. Ropper AH, Shahani B. Proposed mechanism of ataxia inFisher’s syndrome. Arch Neurol 1983;40:537–538.

139. Rubio-Nazabal E, Marey-Lopez J, Torres-Carrete JP, Alvarez-Perez P, Rey Del Corral P. Miller Fisher syndrome andHodgkin’s disease. J Neurol Neurosurg Psychiatry 2002;73:344.


140. Salloway S, Mermel LA, Seamans M, Aspinall GO, Nam ShinJE, Kurjanczyk LA, et al. Miller–Fisher syndrome associatedwith Campylobacter jejuni bearing lipopolysaccharide mole-cules that mimic human GD3. Infect Immun 1996;64:2945–2949.

141. Sartucci F, Cafforio G, Borghetti D, Domenici L, Orlandi G,Murri L. Electrophysiological evidence by single fiber elec-tromyography of neuromuscular transmission impairmentin a case of Miller Fisher syndrome. Neurol Sci 2005;26:125–128.

142. Sauron B, Bouche P, Cathala HP, Chain F, Castaigne P.Miller Fisher syndrome: clinical and electrophysiologic evi-dence of peripheral origin in 10 cases. Neurology 1984;34:953–956.

143. Sawyer CN, Naik RB. A case of recurrent Miller Fisher syn-drome. Br J Hosp Med 1987;38:566.

144. Scelsa SN, Herskovitz S. Miller Fisher syndrome: axonal,demyelinating or both? Electromyogr Clin Neurophysiol2000;40:497–502.

145. Schapira AH, Thomas PK. A case of recurrent idiopathicophthalmoplegic neuropathy (Miller Fisher syndrome).J Neurol Neurosurg Psychiatry 1986;49:463–464.

146. Schnorf H, Rathgeb JP, Kohler A. Anti-GQ1b positive MillerFisher syndrome in a patient with acute Epstein–Barr virusinfection and negative Campylobacter serology. Eur Neurol1998;40:177.

147. Schwerer B, Neisser A, Bernheimer H. Distinct immunoglob-ulin class and immunoglobulin G subclass patterns againstganglioside GQ1b in Miller Fisher syndrome following dif-ferent types of infection. Infect Immun 1999;67:2414–2420.

148. Schwerer B. Antibodies against gangliosides: a link betweenpreceding infection and immunopathogenesis of Guillain–Barre syndrome. Microbes Infect 2002;4:373–384.

149. Segurado OG, Kruger H, Mertens HG. Clinical significanceof serum and CSF findings in the Guillain–Barre syndromeand related disorders. J Neurol 1986;233:202–208.

150. Shah SS, Rodriguez T, McGowan JP. Miller Fisher variant ofGuillain–Barre syndrome associated with lactic acidosis andstavudine therapy. Clin Infect Dis 2003;36:131–133

151. Shimamura H, Miura H, Iwaki Y, Kubodera T, Matsuoka T,Yuki N, et al. Clinical, electrophysiological and serologicaloverlap between Miller Fisher syndrome and acute sensoryneuropathy. Acta Neurol Scand 2002;105:411–413.

152. Sitajayalakshmi S, Borgohain R, Mani J, Mohandas S. Recur-rent Miller Fisher syndrome: a case report. Neurol India2002;50:365–367.

153. Susuki K, Yuki N, Hirata K. Fine specificity of anti-GQ1b IgGand clinical features. J Neurol Sci 2001;185:5–9.

154. Suzuki T, Chiba A, Kusunoki S, Chikuda M, Fujita T, Misu K.Anti-GQ1b ganglioside antibody and ophthalmoplegia ofundetermined cause. Br J Ophthalmol 1998; 82: 916–918.

155. Takahashi M, Koga M, Yokoyama K, Yuki N, Epidemiology ofCampylobacter jejuni isolated from patients with Guillain–Barre and Fisher syndromes in Japan. J Clin Microbiol 2005;43:335–339.

156. Takano H, Yuki N. Fisher’s syndrome associated with chick-enpox and anti-GQ1b antibody. J Neurol 1995;242:255–256.

157. Tanaka H, Yuki N, Hirata K. Trochlear nerve enhancementon three-dimensional magnetic resonance imaging in Fishersyndrome. Am J Ophthalmol 1998;126:322–324.

158. Taphoorn MJ, Uitdehaag BM, Lanting P. A brainstem lesionin the Miller Fisher syndrome demonstrated by CT and MRI.J Neurol 1991;238:243.

159. Tatsumoto M, Odaka M, Hirata K, Yuki N. Isolated abducensnerve palsy as a regional variant of Guillain–Barre syndrome.J Neurol Sci 2006;243:35–38.

160. Toshniwal P. Demyelinating optic neuropathy with MillerFisher syndrome. The case for overlap syndromes with cen-tral and peripheral demyelination. J Neurol 1987;234:353–358.

161. Turner B, Wills AJ. Cerebral infarction complicating intra-venous immunoglobulin therapy in a patient with MillerFisher syndrome. J Neurol Neurosurg Psychiatry 2000;68:790–791.

162. Urushitani M, Udaka F, Kameyama M. Miller Fisher–Guil-lain–Barre overlap syndrome with enhancing lesions in thespinocerebellar tracts. J Neurol Neurosurg Psychiatry 1995;58:241–243.

163. Van Belkum A, van den Braak N, Godschalk P, Ang W,Jacobs B, Gilbert M, et al. A Campylobacter jejuni gene associ-ated with immune-mediated neuropathy. Nat Med 2001;7:752–753.

164. Van der Meche FG, Van Doorn PA, Meuistee M, JennekensFG, GBS-consensus group of the Dutch Neuromuscular Re-search Support Center. Diagnostic classification criteria forthe Guillain–Barre syndrome. Eur Neurol 2001;45:133–139.

165. Weiss JA, White JC. Correlation of 1a afferent conductionwith ataxia of Fisher syndrome. Muscle Nerve 1986;9:327–332.

166. Willison HJ, Townson K, Veitch J, Boffey J, Isaacs N,Andersen S, et al. Synthetic disialygalactose immunoadsor-bents deplete anti-GQ1b antibodies from autoimmune neu-ropathy sera. Brain 2004;127:680–691.

167. Willison HJ, Veitch J, Paterson G, Kennedy PG. Miller Fishersyndrome is associated with serum antibodies to GQ1b gan-glioside. J Neurol Neurosurg Psychiatry 1993;56:204–206.

168. Willison HJ, Veitch J. Immunoglobulin subclass distributionand binding characteristics of anti-GQ1b antibodies inMiller Fisher syndrome. J Neuroimmunol 1994;50:159–165.

169. Wong V. A neurophysiological study in children with MillerFisher syndrome and Guillain–Barre syndrome. Brain Dev1997;19:197–204.

170. Yako K, Kusunoki S, Kanazawa I. Serum antibody against aperipheral nerve myelin ganglioside, LM1, in Guillain–Barresyndrome. J Neurol Sci 1999;168:85–89.

171. Yeh JH, Chen WH, Chen JR, Chiu HC. Miller Fisher syn-drome with central involvement: successful treatment withplasmapharesis. Ther Apheresis 1999;3:69–71.

172. Yuan CL, Wang YJ, Tsai CP. Miller Fisher syndrome: a hos-pital based retrospective study. Eur Neurol 2000;44:79–85.

173. Yuki N, Miyatake T. Guillain–Barre syndrome and Fisher’ssyndrome following Campylobacter jejuni infection. Ann NYAcad Sci 1998;845:330–340.

174. Yuki N, Odaka N, Hitata K. Acute ophthalmoparesis (with-out ataxia) associated with anti-GQ1b IgG antibody: clinicalfeatures. Ophthalmology 2001;108:196–200.

175. Yuki N, Susuki K, Hirata K. Ataxic Guillain–Barre syndromewith anti-GQ1b antibody: relation to Miller Fisher syndrome.Neurology 2000;54:1851–1853.

176. Yuki N, Taki T, Takahashi M, Saito K, Yoshino H, Tai T, et al.Molecular mimicry between GQ1b ganglioside and lipopoly-saccharides of Campylobacter jejuni isolated from patients withFisher’s syndrome. Ann Neurol 1994;36:791–793.

177. Yuki N, Yoshino H, Sato S, Miyatake T. Acute axonal poly-neuropathy associated with anti-GM1 antibodies followingCampylobacter jejuni enteritis. Neurology 1990;40:1900–1902.

178. Yuki N. Infectious origins of, and molecular mimicry inGuillain–Barre and Fisher syndromes. Lancet Infect Dis2001;1:29–37.

179. Yuki N. Molecular mimicry between gangliosides and lipo-polysaccharides of Campylobacter jejuni isolated from patientswith Guillain–Barre syndrome and Miller Fisher syndrome.J Infect Dis 1997;176(suppl):S150–S153.

180. Yuki N. Tryptophan-immobilized column absorbs immuno-globulin G anti-GQ1b antibody from Fisher’s syndrome: anew approach to treatment. Neurology 1996;46:1644–1651.

181. Zifko U, Drlicek M, Senautka G, Griswold W. High doseimmunoglobulin therapy is effective in the Miller Fishersyndrome. J Neurol 1994;241:178–179.


ABSTRACT: Excitability measurements on human motor and sensorynerves have provided new insights into axonal membrane changes in pe-ripheral nerve disorders. The aim of this study was to establish an in vivo ratpreparation suitable for threshold tracking of sensory nerve action potentials(SNAPs) to model clinical sensory nerve excitability studies. In Sprague–Dawley rats anesthetized with ketamine and xylazine, current stimuli wereapplied to the base of the tail and SNAPs recorded from distal needleelectrodes. Multiple excitability data were obtained as previously describedfor human nerves and compared to recordings from the motor tail axons andto sensory recordings from human median and ulnar nerves. The pattern ofexcitability changes in rats was broadly similar to that in humans, althoughsome parameters differed significantly. Individual recordings were stable forat least 3 h. These data show that the rat tail enables excitability propertiesof sensory as well as motor axons to be studied experimentally, e.g., inmodels of nerve disease and during pharmacological interventions.

Muscle Nerve 36: 628–636, 2007

MULTIPLE MEASURES OF AXONAL EXCITABILITY INPERIPHERAL SENSORY NERVES: AN IN VIVO RAT MODEL

ANNETTE GEORGE, MD, and HUGH BOSTOCK, PhD

Sobell Department of Motor Neuroscience and Movement Disorders, Institute of Neurology,University College London, Queen Square, London WC1N 3BG, United Kingdom

Accepted 17 May 2007

Assessment of nerve excitability in human periph-eral nerves provides insight into peripheral nervefunction and underlying axonal membrane proper-ties. Using the noninvasive technique of thresholdtracking in human peripheral nerves, multiple nerveexcitability measures can be obtained by monitoringthe changes in threshold current required for 40%of a maximal compound action potential duringvarious stimulation protocols: (1) during changes ofstimulus duration (strength–duration relationship);(2) during long-lasting subthreshold polarizing cur-rent pulses (threshold electrotonus, current–thresh-old relationship); or (3) following a single supra-maximal stimulus (recovery cycle). The availabilityof a standardized battery of these threshold trackingtechniques in an automated excitability testing pro-tocol (trond recording protocol4,5) has improvedclinical applicability of nerve excitability tests andhas proved extremely helpful in identifying and

characterizing axonal membrane properties undernormal conditions and in peripheral nerve dis-ease.8,10

The molecular mechanisms that underlie theseaxonal membrane properties in humans are imper-fectly understood, and animal models can contrib-ute greatly to improving understanding. For exam-ple, an in vivo rat model for recording thresholdelectrotonus in motor axons was described by Yanget al.20 using the rat tail. This model has recentlybeen used to help identify the molecular nature ofthe slow potassium current in motor axons, since thecontributions of this current to accommodation to100-ms subthreshold depolarizing currents and tolate subexcitability were blocked by the specificKCNQ channel blocker XE991.19 However, nomethod has yet been described for applying stan-dardized multiple nerve excitability recordings to ratsensory nerves in vivo.

Establishing a model for sensory trond record-ing in vivo is warranted for two reasons. First, nerveexcitability properties in human sensory nerves aredistinct in some respects from human motor nerves(such as lower rheobase, different stimulus responseslope, less superexcitability, and refractoriness dur-ing recovery cycle) and their excitability changesduring activity-dependent hyperpolarization or isch-emia are regulated differentially.1,6,7 Thus, the un-

Abbreviations: CMAP, compound muscle action potential; ISI, interstimulusinterval; RRP, relative refractory period; SNAP, sensory nerve action potential;TEd (TEh), depolarizing (hyperpolarizing) threshold electrotonusKey words: membrane potential; recovery cycle; sensory nerve action po-tential; superexcitability; threshold electrotonusCorrespondence to: H. Bostock; e-mail: [email protected]


628 Nerve Excitability in Rat MUSCLE & NERVE November 2007

derlying molecular mechanisms and ion-channelregulation might be distinct and this can only beinvestigated in separate animal models of motor andsensory nerve excitability. Second, some neuropa-thies preferentially impair sensory axons, and maytherefore require studies on sensory axons in ananimal model to gain insights into the disease-induced changes in axonal membrane propertiesinvolved.

Therefore, in the present study we applied thestandardized protocol for multiple measures ofnerve excitability that is used clinically5 to peripheralsensory nerves in a rat model. The aim of the studywas to establish a tool that would be suitable forexperimental sensory nerve excitability recordingsand for subsequent studies with pharmacological in-tervention in peripheral nerve disease models or ingenetically modified animals. We also compared theexcitability properties in rat sensory axons with thosein rat motor axons and in human sensory axons.

MATERIALS AND METHODS

Animals. Experiments were performed in adult fe-male Sprague–Dawley rats weighing 250–300 g ob-tained from B&K Universal (Hull, UK). All experi-mental procedures were carried out under licensefrom the U.K. Home Office, following approval bythe Ethical Review Panel of the Institute of Neurol-ogy. After the recording session the animal was killedby an overdose of anesthetic.

Subjects. Eighteen healthy human volunteers(nine women, nine men; median age, 28 years;range 21–45 years) were recruited to study excitabil-ity parameters in the median and ulnar sensorynerve. The study was approved by the Joint ResearchEthics Committee of the National Hospital for Neu-rology and Neurosurgery and the Institute of Neu-rology, London. Written informed consent was ob-tained from all subjects.

Compound Action Potential Recordings in Rat Periph-

eral Nerves. Rats were anesthetized with an initialdose of 90 mg/kg IP ketamine and 10 mg/kg IPxylazine to produce deep anesthesia. To maintainthe level of anesthesia, supplementary doses, equalto one-third of the initial dose, were administered asrequired. The lower back aspect was shaved and therat was placed prone on a heated under-blanket,which was controlled by a rectal thermistor probe tokeep the body temperature close to 37°C. Themethod of peripheral nerve excitability testing issimilar to that described recently for rat motor tail

nerves.19,20 Nonpolarizable Ag/AgCl disk electrodes(Red Dot, 3M Health Care, Borken, Germany) wereused for stimulation. The cathode was attached tothe lateral aspect of the tail base. The anode wasplaced on the skin of the hip. Stainless steel needleelectrodes were used for recording and grounding.To record compound muscle action potentials(CMAPs), the electrode was inserted into the ipsilat-eral aspect of the tail muscle 50–60 mm distal to thestimulating electrode. To record SNAPs the elec-trode was inserted into the ipsilateral aspect of thetail skin, �100 mm distal to the stimulating elec-trode. The reference electrode was inserted nearby.A ground electrode was inserted between the stimu-lating and recording electrodes. The temperature ofthe skin was measured with a thermocouple placedadjacent to the stimulating electrode. Signals wereamplified and filtered (2 Hz–2 kHz) (Disa EMGamplifier type 14C13; Skovlunde, Denmark) andline interference was removed with an online noiseeliminator (HumBug; Quest Scientific, North Van-couver, Canada). Signals were digitized by aDAQ2000 A/D board (Iotech; Cleveland, Ohio) at asampling rate of 10 kHz. Stimulation was controlledby a personal computer running qtracs software(written by H. Bostock, © Institute of Neurology,London) connected to the data acquisition unit andthe stimulator (DS5 prototype; Digitimer, WelwynGarden City, UK).

Separation of Motor and Sensory Action Potentials.

The lower part of Figure 1 shows four recordingsmade from the same rat tail. CMAP recordings weremade from site 1 with 1 ms stimuli, set supramaximalto record maximum peak-to-peak CMAP amplitude(trace 1A), and during threshold tracking to evoke atarget response close to 40% of the maximum (trace1B). The recordings in traces 2A and 2B were mademore distally, at site 2, with stimulus currents oflower amplitude and 0.5-ms duration, and at highergain. By careful positioning of the recording elec-trodes it was usually possible to record a maximalSNAP peak-to-peak response, as in trace 2A, withoutdistortion due to overlap with the low-level CMAP(asterisked). The SNAPs appeared earlier than theCMAPs because of the synaptic delay at the neuro-muscular junction and slower conduction in the mo-tor nerve terminals. When the stimulus was furtherreduced to track the target response level (trace 2B),contamination by CMAP was usually not detectable.However, when the test stimuli were superimposedon strong hyperpolarizing currents during recordingof the current–threshold relationship, and at shortinterstimulus intervals during the recovery cycle,

Nerve Excitability in Rat MUSCLE & NERVE November 2007 629

contamination by CMAPs could interfere withthreshold tracking.

Compound Action Potential Recordings in Human Me-

dian Nerve. Recordings were performed as de-scribed by Kiernan et al.5 Briefly, the stimuluscurrents were applied via nonpolarizable elec-trodes (Red Dot, 3M Health Care) with the cath-ode over the median or ulnar nerve at the wrist,and the anode about 10 cm more proximal. Sur-face electrodes (disposable ring electrodes; ViasysHealthcare, Old Woking, UK) were attached to theindex or little finger as recording electrodes. Thetemperature of the skin was measured with a ther-mocouple placed adjacent to the stimulating elec-trode. The stimulation protocol was as describedbelow.

Protocol of Multiple Nerve Excitability Measures.

Multiple nerve excitability measurements were re-corded by using the threshold tracking programqtracs (Windows qtrac v1.0.0; © Institute of Neu-rology, London). Two qtracs protocols were used:the trondcsw protocol was used for sensory nerverecordings with test stimulus pulse width of 0.5 ms,and the trondcmw protocol was used for motornerve recordings with a pulse width of 1 ms. Sweepduration was 230 ms and the test stimulus was deliv-ered at regular intervals of 0.8 s. Both protocolsenable the threshold changes to be monitored dur-ing various stimulation procedures on multiplechannels in an automated fashion. They differ fromthe earlier trondx protocols described by Kiernanet al.4,5 only in the method of estimating strength–duration time constant and rheobase. Instead of

FIGURE 1. Schematic diagram of method of recording from rat tail and representative action potential waveforms. (A) Recordingarrangement, showing positions of nonpolarizable stimulation electrodes and needle electrodes for motor (1) and sensory (2) actionpotentials. (B) CMAPs (1) and SNAPs (2) recorded from the same rat tail at different levels of stimulus current (indicated). Plots 1A and2A indicate maximal action potentials, with lines indicating peak-to-peak amplitude measurement. Arrows indicate onset of stimulusartifact. Plots 1B and 2B are examples of action potentials recorded while threshold tracking, in which the computer adjusted the stimuluslevel on a trial-by-trial basis to match the amplitude to the target response level (indicated by horizontal gray lines). The asterisk in Plot2A indicates a low-level CMAP that started to appear when the SNAP was close to maximal.


using stimulus–response curves for two differentstimulus durations, a single stimulus–response curvewas recorded (0.5 ms stimulus duration for sensory,1 ms for motor). A target response amplitude was setto the steepest part of the stimulus–response curvebetween 30% and 50% of the maximum response,and then a 5-point strength–duration relationship(0.5 to 0.1 ms for sensory fibers, or 1.0 to 0.2 ms formotor) was recorded for the target response bythreshold tracking. Threshold electrotonus, cur-rent–threshold relationship, and recovery cycle wererecorded as previously described,4,5 with the thresh-old to the test stimulus alone monitored throughoutthe recording and compared with the threshold al-tered by conditioning currents.

Data Analysis. The trond recordings were ana-lyzed and plotted with qtracp. This program en-ables automated analysis of multiple excitability pa-rameters. First the raw stimulus and response valuesfor each stimulus condition were used to estimatethe threshold stimulus required to elicit the targetresponse. These data were then used for furtherstatistical analysis and for plotting, as previously de-scribed for human excitability data.4,5 One MEM(multiple excitability measures) file was generatedfrom one trond recording and contained, in addi-tion to the threshold estimates plotted as in Figure 2,a list of 32 derived excitability parameters that wereused for statistical analysis. Several MEM files wereindexed in an MEF file, and qtracp also providesfor statistical comparisons between groups of record-ings specified by the MEF files. Parameters wereassumed to be normally distributed unless they failedthe Lilliefors test for normality (P � 0.05%). Meanparameter values were compared between groupsusing parametric tests for normally distributed andnonparametric tests for nonnormally distributeddata. Parameter variability is indicated by the stan-dard deviation (SD) or interquartile range (IQR).Statistical significance was assumed with P � 0.05.

RESULTS

Sensory Nerve Excitability Parameters in Rat Tail Sen-

sory Nerves. Nerve excitability parameters derivedfrom SNAP recordings of the tail nerves in Sprague–Dawley rats are summarized in Figure 2, column A,and compared with equivalent data from rat motoraxons (column B) and human median sensory axons(column C). The mean rat peak SNAP amplitudewas 30 � 1.8 �V (mean � SD; n � 8). Strength–duration time constants for SNAP recordings (Fig.2.A1) ranged from 0.28 to 0.7 ms (mean 0.39, me-

dian 0.32, IQR 0.30–0.43 ms). Figure 2.A2 illustratesthe threshold changes during the threshold electro-tonus recording. In response to the 40% depolariz-ing current pulses (plotted upwards), after the initialfast threshold change, threshold reduction reacheda peak (56 � 4.6%) at around 10 ms after the onsetof the current pulse and returned to a steady stateafter around 50 ms (48 � 3.9%). With the hyperpo-larizing current pulses, after the initial fast thresholdchange, threshold reduction fell continuously to aminimum at 100 ms (�125 � 11% for 40%). Ontermination of the polarizing current pulses, thresh-old overshoot was observed in most but not all re-cordings and was generally small (3.6 � 2.5% thresh-old reduction after hyperpolarization). Figure 2.A3illustrates the current–threshold relationship. Theincrease in slope with depolarizing currents (upperright) indicates outward rectification, and thesmaller increase in slope with hyperpolarizing cur-rents (lower left) indicates inward rectification. Thisrecording is incomplete, since in most cases the sen-sory threshold during strong hyperpolarizing currentswas high enough that the SNAP became contaminatedby a low-level CMAP (cf. Fig. 1.2A) and complete cur-rent–threshold relationships were only recorded in twoanimals. Figure 2.A4 shows the mean changes inthreshold during the recovery cycle. Again, in someSNAP recordings there was contamination by CMAPsat short interstimulus intervals, so that complete recov-ery cycles were only recorded in four animals. Latesubexcitability was small, with a maximum (1.9 �0.9%) at an interstimulus interval (ISI) of around 50ms. At shorter intervals it was preceded by the period ofsuperexcitability, which reached a peak thresholdchange of �15.7 � 3.2% at �3 ms. Unlike in humanrecordings, the nerves were already slightly superexcit-able at the shortest interval of 2 ms in the standardprotocol.

Comparison of Rat Sensory Nerve Excitability Parame-

ters with Rat Motor Data. Column B of Figure 2summarizes nerve excitability parameters derivedfrom eight CMAP recordings of rat tail motor nerves,stimulated at the same site as the sensory nerves. Themean peak amplitude of the CMAPs was 4.5 � 1.5mV (mean � SD; n � 8), about 100 times greaterthan the peak SNAP amplitude. The strength–dura-tion time constants for motor axons (Fig. 2.B1) werenot significantly different from the sensory axons(mean 0.40 ms, median 0.28, IQR 0.25–0.29; P �0.38, Mann–Whitney test). The threshold electroto-nus (B2) and recovery cycle data (B4) correspondclosely to that previously found in Wistar rats.19 Asfar as comparison between sensory and motor nerve


FIGURE 2. Multiple excitability measurements from three preparations. Column A, rat tail sensory axons. Column B, rat tail motor axons.Column C, human median sensory axons. Row 1, charge–duration relationship. Row 2, threshold electrotonus. Row 3, current–thresholdrelationship. Row 4, recovery cycle.


excitability measures in the rat model is concerned,absolute values from threshold electrotonus cannotbe directly compared between the sensory and themotor recordings since stimulus durations were dif-ferent. However, Figure 2 shows that the time courseand pattern of these excitability changes observed insensory axons were similar to those observed in mo-tor axons. Nerve excitability parameters that wereclearly distinct in sensory from motor recordingswere observed during the recovery cycle recording(Fig. 2A4,B4). There was less subexcitability in sen-sory axons (1.9 � 0.9% in sensory vs. 4.7 � 1.8% inmotor axons; P � 0.01; unpaired t-test with Welchcorrection) and sensory axons were on average su-perexcitable rather than refractory at an ISI of 2 ms(�7.4 � 8% change in threshold in sensory nerves,compared with 15.3 � 10% in motor axons; P �0.05; Mann–Whitney test).

Comparison of Rat Sensory Nerve Excitability Parame-

ters with Human Sensory Data. To test whether thenerve excitability parameters obtained for the rat invivo model resemble those in humans, we comparedrat sensory nerve excitability data with sensory me-dian nerve and sensory ulnar nerve recordings in 18healthy control subjects, using the same trond re-cording protocol as used in the animal studies.These new sensory recordings were very similar to

those described previously for the median nerve.5

There was little difference between the median andulnar recordings, so only the median data is illus-trated in Figure 2C. Comparing columns A and C inFigure 2, the patterns of nerve excitability plots ob-tained for rat sensory tail axons are qualitativelysimilar to those in human studies, and the variabilityof the measurements was, if anything, less. However,absolute values derived from the recordings in therat model differed significantly from the humandata, as shown more clearly in Figure 3.

The threshold electrotonus waveforms in Figure3A were on average similar in rat (filled circles) andhuman (open circles) studies but threshold reduc-tions during depolarizing currents were less pro-nounced in the rat. Thus TEd (10–20 ms) (i.e., themean threshold reduction 10–20 ms after the start ofthe 40% depolarizing current) was 56% in rat and67% in human median sensory nerve recordings(P � 0.001; unpaired t-test Welch corrected). TEd(90–100 ms) was 46% in the rat and 52% in thehuman median sensory nerve recordings (P � 0.01).During a 100-ms hyperpolarizing current, TEh(10–20 ms) was �71% in rat and �83% in humanrecordings (P � 0.0001), whereas TEh (90–100 ms)was not significantly different between rat and hu-man median sensory nerve recordings (P � 0.68;

FIGURE 3. Threshold electrotonus and recovery cycle in rat and human sensory nerves. (A) Threshold electrotonus. Note that thresholdchanges in the rat recordings are less pronounced (small arrow) and the overshoot after a 100-ms hyperpolarizing current is much less(long arrow) in the rat compared to the human recordings. (B) Recovery cycle. Note that the degree of superexcitability in the ratrecordings is similar to the human sensory recordings, whereas the relative refractory period is shorter (short arrow), and there is less latesubexcitability (long arrow). Filled circles, mean values of recordings from rat sensory tail nerves (n � 8). Open circles, human mediansensory nerves (n � 18). Open triangles, human ulnar sensory nerves (n � 18).


Mann–Whitney test). After a 100-ms hyperpolarizingcurrent, recovery to baseline values was slower in ratrecordings and overshoot was less pronounced.Mean TEh (overshoot) was 3.6 � 2.5% in rat and19 � 3.9% in human median sensory nerve record-ings (P � 0.0001). Figure 3B shows that the amountof superexcitability in the rat sensory recording wascomparable to the degree of superexcitability mea-sured in human median or ulnar sensory nerve re-cordings. Features that were significantly distinct inthe rat sensory from the human median sensorynerve recordings were relative refractoriness (P �0.001) and amount of late subexcitability (P �0.001). These findings are in line with the data formotor axons: less pronounced threshold changesduring TE recordings and less pronounced late sub-excitability during the recovery cycle are exhibitedby rat tail motor axons19 in comparison with humanmotor axons.4

Stability of Individual Recordings. To test the stabil-ity of the recordings from rat sensory tail nerves, werepeated them at 30-min intervals for up to 3 h. Weaddressed this question since pharmacological inter-vention studies might require repeated recordingswithin the same animal. Figure 4 shows one exampleof repeated threshold electrotonus (A) and recoverycycle (B) recordings, and the stability of importantexcitability parameters over time. Similar recordingswere obtained from two other animals. Mean valuesfor peak SNAP, strength–duration time constant,TEd (10–20 ms), TEd (peak), TEh (90–100 ms),TEd (90–100 ms), superexcitability (%), and latesubexcitability (%) were similar when baseline valueswere compared to values after 180 min (n � 3;Wilcoxon matched pairs test). For example, peaksuperexcitability was �17.3 � 1% at baseline and�16.8 � 1.2% after 180 min (n � 3; mean � SD; nosignificant difference between groups, Wilcoxonmatched pairs test). In summary, these data indicatestability of the recordings for at least 3 h (last time-point investigated).

DISCUSSION

This study describes in vivo measurements of multi-ple excitability parameters in peripheral sensory ax-ons in the rat. The patterns of excitability changeresemble in many respects those previously de-scribed for human sensory nerves, suggesting thatthe rat tail may provide a useful model for clinicalstudies of excitability properties of human sensoryaxons. However, there are substantial differences insome excitability parameters, and others could not

be measured consistently in this preparation becauseof contamination by motor action potentials.

The rat tail was used because the long conduc-tion distance available allows sensory and motor re-sponses to be separated on the basis of latency, andbecause the tail permits noninvasive stimulation ofmotor and sensory axons with nonpolarizable elec-trodes, which are essential for accurate recording ofresponses to polarizing currents. For this model tobe useful, it is not necessary that the individual ex-citability parameters be indistinguishable from thosein human peripheral nerve, but only that they showcomparable features, ascribable to the function ofcorresponding active and passive membrane proper-ties. For example, when motor axons are subject to100-ms subthreshold depolarizing currents, the re-duction in threshold reaches a maximum after10–20 ms, and is then followed by a slow increase inthreshold, i.e., accommodation (Fig. 2, row 2). Al-though this accommodation is less in rat than hu-man motor fibers (as it is less in rat than humansensory fibers in Fig. 3A), the family resemblancemakes it highly likely that corresponding ion chan-nels (shown to be KCNQ2 potassium channels inlarge rat motor axons19) are responsible. In the sameway, the family resemblance between the plots inFigure 2 makes it likely that the rat tail can providea useful model for human sensory nerves.

The differences in the recovery cycles shown inFigure 3B are more substantial. The relative refractoryperiod (RRP) is much shorter in the rat, and in futurestudies it will be necessary to extend the recovery cyclemeasurements to intervals as short as 1 ms to recordrefractoriness adequately. The reason for this discrep-ancy between rat and human cutaneous sensory axonsis unclear. Refractoriness is sensitive to temperature,but the skin temperatures in the rat were similar tothose in the humans (32.7 � 0.75°C vs. 33.5 � 0.89°C,respectively). The refractory period is closely related toaction potential duration, since the membrane poten-tial has to become repolarized before inactivation ofsodium channels is removed and they can generate afresh action potential. In mammalian myelinated ax-ons, action potential duration depends primarily onsodium channel inactivation, as nodal fast potassiumconductance is small.17,18 Although an increased nodalfast potassium conductance could help shorten actionpotential duration, it is unlikely that this would reducethe RRP, since RRP also depends on superexcitability,which would be reduced. Our data therefore suggestthat sodium channel inactivation may be faster in ratthan human sensory axons. No such species differencewas found by Reid16 when voltage clamp measure-ments on human nodes of Ranvier (that included so-


dium channel inactivation kinetics) were comparedwith corresponding data from rat nodes.14 However, inthose studies motor and sensory fibers were not distin-guished, and the RRP is longer in rat motor than ratsensory axons (Fig. 2A4,B4). Further studies areneeded to determine the cause of this surprising spe-cies difference in recovery cycle recordings, but wenote that similarly short RRPs have been recordedfrom rat cutaneous A-fibers in an in vitro skin–nervepreparation (unpublished observations; Maurer, Bos-tock, and Koltzenburg).

Strength–duration time constants in human sen-sory axons have regularly been found to be longerthan those in motor axons,12,15 a difference attrib-uted to a greater level of persistent sodium current atthe resting potential in sensory axons.2 In our ratrecordings, however, motor and sensory strength–duration time constants were not significantly differ-ent. This may reflect in part a high variability of therat time constant estimates, but sensory strength–

duration time constants were significantly shorter inrat than human recordings (P � 0.001).

These discrepancies between rat and humannerve excitability measurements indicate that the ratmodel has to be used with caution as a tool forunderstanding clinical observations, especially thoseconcerning refractoriness and strength–durationtime constant. However, the recordings are repro-ducible and stable over at least 3 h, and faithfullyreproduce many of the characteristics of human sen-sory nerve excitability.

Comparison with Other Methods. Experimentalnerve excitability testing in rodents has previouslybeen described for motor axons in vivo13,19,20 and forunmyelinated axons in vitro.3,9,11 In the presentstudy, trond recordings were performed in vivo inanesthetized animals. This has the advantage over invitro recordings in that normal blood supply andperfusion of the nerve is maintained. Thus, the ax-

FIGURE 4. Stability of individual TROND recordings in rat sensory tail nerves. Top, seven superimposed recordings of thresholdelectrotonus (A) and recovery cycle (B) made from the same tail at half-hour intervals. Bottom, plots of four key SNAP parameters,showing absence of any trend over time.


onal microenvironment remains relatively undis-turbed, which is important for preserving restingmembrane potential and other functions thatstrongly depend on membrane potential, which in-clude all nerve excitability properties. Another ad-vantage of in vivo nerve recordings is that the nerveis preserved in a good condition to be harvestedafterwards for histological and immunohistochemi-cal procedures. This allows nerve excitabilitychanges to be correlated with nerve histology (e.g.,the degree of fiber loss in neuropathy models) orwith the detection, quantification, and cellular local-ization of an axonal membrane protein of interest(e.g., an ion channel protein). In addition, since thein vivo recordings are only minimally invasive, re-peated recordings can be made from the samenerve, allowing longitudinal studies to monitor thedevelopment of pathology in nerve disease modelsand their changes during treatment.

As described in clinical studies,5 one limitation ofthe trond sensory recording is that threshold track-ing may become impossible in cases of small SNAPamplitudes, e.g., in severe experimental neuropathymodels. For accurate threshold tracking, the signal-to-noise ratio must be sufficiently large that the tar-get response remains above the noise level. Thesoftware allows the signal-to-noise ratio to be im-proved by averaging, but this extends the recordingperiod. Also, interference of the tracking of theSNAP by a concomitant motor response can disturbthe in vivo trond recording, which is not the casewith in vitro recordings. The in vivo preparationallows the study of the immediate impact of drugs orother pharmacological interventions administeredby the IP or IV route at therapeutic doses.19 How-ever, the dosage is limited by toxicity, and the con-centration of the drug reaching the axons cannot becontrolled. This is in contrast to the in vitro bathrecordings with desheathed nerves or spinal roots,where the nerve environment can be controlledmuch more directly, in particular by the relativelyrapid administration and removal of drugs.3,9,11 Thein vitro preparations also avoid the possibility thatthe anesthetics used for the in vivo nerve recordingsmay have an effect on nerve excitability.

In conclusion, our study shows that minimallyinvasive threshold tracking of SNAPs as well asCMAPs is possible using the rat tail, allowing multi-ple excitability parameters of sensory and motor ax-ons to be assessed in vivo. This model may be usefulfor pharmacological studies and studies in nervedisease models.

Supported by The Wellcome Trust.

REFERENCES

1. Bostock H, Burke D, Hales JP. Differences in behaviour ofsensory and motor axons following release of ischaemia.Brain 1994;117:225–234.

2. Bostock H, Rothwell JC. Latent addition in motor and sensoryfibres of human peripheral nerve. J Physiol (Lond) 1997;498:277–294.

3. Grafe P, Quasthoff S, Grosskreutz J, Alzheimer C. Function ofthe hyperpolarization-activated inward rectification in non-myelinated peripheral rat and human axons. J Neurophysiol1997;77:421–426.

4. Kiernan MC, Burke D, Andersen KV, Bostock H. Multiplemeasures of axonal excitability: a new approach in clinicaltesting. Muscle Nerve 2000;23:399–409.

5. Kiernan MC, Lin CS, Andersen KV, Bostock H. Clinical eval-uation of excitability measures in sensory nerve. Muscle Nerve2001;24:883–893.

6. Kiernan MC, Lin CS, Burke D. Differences in activity-depen-dent hyperpolarization in human sensory and motor axons.J Physiol (Lond) 2004;30:667–672.

7. Kiernan MC, Mogyros I, Burke D. Differences in the recoveryof excitability in sensory and motor axons of human mediannerve. Brain 1996;119:1099–1105.

8. Kiernan M, Burke D, Bostock H. Nerve excitability measures:biophysical basis and use in the investigation of peripheralnerve disease. In: Dyck PJ, Thomas PK, editors. Peripheralneuropathy, 4th ed. Philadelphia: Elsevier Saunders; 2005. p113–129.

9. Lang PM, Moalem-Taylor G, Tracey D, Bostock H, Grafe P.Activity-dependent modulation of axonal excitability in unmy-elinated peripheral rat nerve fibres by the 5-HT(3) serotoninreceptor. J Neurophysiol 2006;96:2963–2971.

10. Lin CS-Y, Kiernan M, Burke D, Bostock H. Assessment ofnerve excitability in peripheral nerve disease. In: Kimura J,editor. Clinical neurophysiology of peripheral nerve disease.Handbook of clinical neurophysiology. Edinburgh: Elsevier;2006. p 381–403.

11. Moalem G, Grafe P, Tracey DJ. Chemical mediators enhancethe excitability of unmyelinated sensory axons in normal andinjured peripheral nerve of the rat. Neuroscience 2005;134:1399–1411.

12. Mogyoros I, Kiernan MC, Burke D. Strength-duration prop-erties of human peripheral nerve. Brain 1996;119:439–447.

13. Moldovan M, Krarup C. Evaluation of Na�/K� pump func-tion following repetitive activity in mouse peripheral nerve.J Neurosci Methods 2006;155:161–171.

14. Neumcke B, Schwarz JR, Stampfli R. A comparison of sodiumcurrents in rat and frog myelinated nerve: normal and mod-ified sodium inactivation. J Physiol (Lond) 1987;382:175–191.

15. Panizza M, Nilsson J, Roth BJ, Rothwell J, Hallett M. The timeconstants of motor and sensory peripheral nerve fibers mea-sured with the method of latent addition. ElectroencephalogrClin Neurophysiol 1994;93:147–154.

16. Reid G. Ion channels in human axons. PhD thesis, Universityof London, 1996.

17. Schwarz JR, Eikhof G. Na currents and action potentials in ratmyelinated nerve fibres at 20 and 37 degrees C. Pflugers Arch1987;409:569–577.

18. Schwarz JR, Reid G, Bostock H. Action potentials and mem-brane currents in the human node of Ranvier. Pflugers Arch1995;430:283–292.

19. Schwarz JR, Glassmeier G, Cooper E, Kao T, Nodera H,Tabuena D, et al. KCNQ channels mediate Iks, a slow K�

current regulating excitability in the node of Ranvier.J Physiol (Lond) 2006;573:17–34.

20. Yang Q, Kaji R, Hirota N, Kojima Y, Takagi T, Kohara N, et al.Effect of maturation on nerve excitability in an experimentalmodel of threshold electrotonus. Muscle Nerve 2000;23:498–506.


ABSTRACT: The baroreflex maintains a stable blood pressure (BP) bydynamically adjusting heart rate (vagal component) and total peripheralresistance (adrenergic component). Vagal baroreflex sensitivity (BRS-v) iswidely used but no methodology existed to quantitate adrenergic baroreflexsensitivity (BRS-a) until we developed the indices of BP recovery time (PRT)and BRS-a. The aims of this study were to generate a normative databaseand to evaluate whether there is an age effect on the cardiovagal andadrenergic sensitivities. We evaluated recordings of heart rate (HR) and BPin 255 normal subjects during the Valsalva maneuver (VM) and determinedboth BRS-v and BRS-a sensitivities. PRT increased with age whereas allother parameters declined with age. The adrenergic parameters correlatedwell with each other but not significantly with BRS-v. The results indicate thatboth BRS-a and BRS-v become blunted with increasing age and that theseindices behave independently of each other.

Muscle Nerve 36: 637–642, 2007

EFFECT OF AGE ON ADRENERGIC AND VAGALBAROREFLEX SENSITIVITY IN NORMAL SUBJECTS

CHIH-CHENG HUANG, MD,1,2 PAOLA SANDRONI, MD, PhD,1 DAVID M. SLETTEN,1

STEPHEN D. WEIGAND, MS,3 and PHILLIP A. LOW, MD1

1 Mayo Clinic, Department of Neurology, 200 First Street SW, Rochester, Minnesota 55905, USA2 Chang Gung Memorial Hospital–Kaohsiung Medical Center, Department of Neurology, Taiwan3 Mayo Clinic, Department of Biostatistics, Rochester, Minnesota, USA


The baroreflex is responsible for maintaining a stableblood pressure (BP) in spite of changes in body posi-tion.2,7 Changes in position, such as standing from alying position, result in a reduction in venous return tothe heart. Baroreceptors are unloaded and the barore-flex, by adjusting heart rate (vagal component) andtotal peripheral resistance (sympathetic adrenergiccomponent), prevents a change in BP.15 Neurologicdisease such as multiple system atrophy, pure auto-nomic failure, or the autonomic neuropathies result inlesions of the baroreflex and there is ensuing ortho-static hypotension, supine hypertension, and loss ofdiurnal variation in BP.9 Baroreflex sensitivity (BRS) iswidely used to quantitate the vagal component of thereflex.13 However, in disorders such as the autonomicneuropathies, the adrenergic component may be dif-ferentially or selectively involved.9 Although the mea-

surement of orthostatic BP or plasma norepinephrineincrement in response to standing provides indices ofadrenergic function,3,6 their sensitivity and specificityare relatively poor. Currently, quantitative evaluationof adrenergic function requires microneurogra-phy,4,5,15 a method that is too invasive and time con-suming for routine clinical use.

We have recently described an index of the adren-ergic component of the baroreflex, the BP recoverytime (PRT).18 We then described a more completeevaluation that relates PRT to the BP change thatdrives PRT,17 hence generating an index of adrenergicbaroreflex sensitivity, and validated the method againstdirectly recorded muscle sympathetic nerve activity(MSNA) from microneurographic recordings duringValsalva maneuver (VM).17 The aims of this study were:(1) to provide normative data on PRT and other pa-rameters of adrenergic baroreflex sensitivity; (2) toevaluate the effect of age and gender on the adrener-gic component of the baroreflex; and (3) to studywhether there is a differential effect of age on thecardiovagal and adrenergic components.


We evaluated 255 normal subjects, primarily ofnorthern European descent, of either gender who

Abbreviations: BP, blood pressure; BRS, baroreflex sensitivity; BRS-a, ad-renergic baroreflex sensitivity; BRS-v, vagal baroreflex sensitivity; HR, heartrate; HRDB, heart rate response to deep breathing; II-E, early phase II; II-L, latephase II; MSNA, muscle sympathetic nerve activity; PRT, blood pressurerecovery time; VM, Valsalva maneuver; VR, Valsalva ratioKey words: adrenergic; age; baroreflex sensitivity; cardiovagal; Valsalva ma-neuverCorrespondence to: P. A. Low; e-mail: [email protected]


Adrenergic Baroreflex and Age MUSCLE & NERVE November 2007 637

were free of disorders or medication known toaffect autonomic function. All subjects underwentroutine autonomic testing to evaluate cardiovagaland adrenergic function, and no coffee, food, ornicotine was permitted for 4 h before the test-ing.9,10 Heart rate (HR) was derived from contin-uously recorded standard three-lead ECG (Ivy Bio-medical, model 101; Branford, Connecticut).Arterial BP was continuously measured at the fin-ger using beat-to-beat photoplethysmographic re-cordings (Finapres blood pressure monitor 2300;Ohmeda, Englewood, Ohio).

The heart rate response to deep breathing(HRDB) and the VM were performed with thesubject supine. For HRDB subjects were instructedto pace their breathing at a rate of 10 s perbreath.8 Compliance to the breathing rate wasestablished using an oscillating light (WR MedicalElectronics, Stillwater, Minnesota). For the VM,subjects were instructed to maintain a column ofmercury at 40 mmHg for 15 s via a tube with an airleak (to ensure an open glottis). The maneuverwas repeated until two reproducible responseswere obtained. We utilized strict criteria for anacceptable maneuver. We excluded: (1) expiratorypressure less than 30 mmHg or 10 s; (2) nonpro-ducible qualitative configuration of phases; and(3) a “flat-top” response or the response with sys-tolic pressure of early phase II (II-E) or phase IIIabove baseline.

The baseline values of BP (systolic, diastolic, andmean) were determined for all subjects over a 30-s

interval directly preceding the VM. PRT was definedas time in seconds from the valley of phase III untilBP returned to baseline.18 Adrenergic baroreflexsensitivity was defined in two ways, the standard(BRS-a) and the alternative (BRS-a1), as illustrated inFigure 1.17 BRS-a was defined as the systolic BP dec-rement associated with phase III divided by PRT.BRS-a1 was defined as systolic BP decrement dividedby PRT, of which BP decrement was the BP fall inphase II-E plus three-fourths of the amplitude ofphase III.17 The amplitude of phase III was measuredfrom the end of late phase II (II-L) to the valley ofphase III.1 In addition, the parameters of vagalbaroreflex sensitivity (BRS-v), expressed as the slopeof regression of heart period over systolic BP duringphase II-E, Valsalva ratio (VR, i.e., the ratio betweenthe highest HR reached in phase II and the lowestHR of phase IV reflex bradycardia), and HRDB werealso measured for each subject. Finally, the productsof BRS-v with BRS-a and with BRS-a1 were calculatedas global measures for baroreflex function and de-noted BRS-g and BRS-g1, respectively.

Data Analysis and Statistics. Percentiles were calcu-lated following the methods described by O’Brienand Dyck.12 We began by taking a base 10 loga-rithm transformation of the dependent variablesbecause they were found to be skewed. To evaluatewhether a dependent variable was related to age,we fit a model with both age and age squared.Since for none of the endpoints was the squaredterm significant, we removed this term from themodel. With the linear age term in the model, wethen tested whether gender was significant. Thefinal models were found to involve age only. To

FIGURE 1. A blood pressure profile during Valsalva maneuverillustrates the calculating of BRS-a and BRS-a1. BRS-a was definedas the systolic BP decrement associated with phase III (bottom ofphase III to baseline; highlighted) divided by PRT. BRS-a1 wasdefined as systolic BP decrement divided by PRT, where BP dec-rement was the BP fall in phase II-E (A) plus three-fourths of theamplitude of phase III (B).

Table 1. Demographic, heart rate response to deep breathing,and Valsalva ratio data.

N Mean � SD

Age (years) 175 43.6 � 16.1Male 79 44.1 � 15.7Female 96 43.2 � 16.5

Height (cm) 166 170.3 � 9.7Male 73 177.8 � 7.6Female 93 164.4 � 6.6

Weight (kg) 169 77.0 � 16.7Male 75 87.4 � 16.1Female 94 68.8 � 11.9

HRDB 167 19.9 � 9.0Male 77 19.0 � 8.5Female 90 20.8 � 9.4

Valsalva ratio 175 1.94 � 0.41Male 79 1.90 � 0.38Female 96 1.98 � 0.43

638 Adrenergic Baroreflex and Age MUSCLE & NERVE November 2007

evaluate whether variability about the mean de-pended on age, we calculated the Spearman non-parametric rank correlation between the absolutevalue of the residuals and age. To evaluatewhether variability varied by gender, we per-formed a Wilcoxon rank-sum test comparing theabsolute value of residuals for men and women.We did not find that variability depended on age

or gender, and therefore proceeded to calculatez-scores and percentiles following steps 5–7 out-lined by O’Brien and Dyck.12 Finally, the base 10logarithm of the dependent variables was undonefor the reporting of percentiles. The associationsbetween measurements were evaluated usingSpearman’s nonparametric rank-order correla-tion, denoted by rho, due to skewness. For all tests,

Table 2. Regression of barosensitivity indices with age and gender.

Dependentvariable

Interceptcoefficient

Age coefficient � SE(P-value)

Coefficient ofdetermination (r2)

Age2

P-value*GenderP-value†

Log10 PRT �0.195 0.005 � 0.001 (�0.001) 0.094 0.286 0.332Log10 BRS-a 1.462 �0.002 � 0.001 (0.023) 0.029 0.599 0.868Log10 BRS-a1 1.722 �0.003 � 0.001 (0.014) 0.034 0.964 0.340Log10 BRS-v 1.115 �0.007 � 0.001 (�0.001) 0.238 0.622 0.480Log10 BRS-g 2.557 �0.009 � 0.001 (�0.001) 0.266 0.995 0.472Log10 BRS-g1 2.837 �0.010 � 0.001 (�0.001) 0.225 0.771 0.192

*Based on a model with age and age2.†Based on a model with age and gender.The unit is seconds for PRT, mmHg/s for BRS-a and BRS-a1, ms/mmHg for BRS-v, and ms/s for BRS-g and BRS-g1.

Table 3. Clinically relevant percentiles of the parameters of baroreflex sensitivity.

Age (by decade)

Percentile 20 30 40 50 60 70

PRT(s)2.5 0.21 0.24 0.27 0.30 0.33 0.385 0.27 0.30 0.33 0.37 0.42 0.4795 2.39 2.68 3.01 3.38 3.79 4.2597.5 2.89 3.24 3.63 4.08 4.57 5.13

BRS-a (mmHg/s)2.5 11.6 11.0 10.5 10.1 9.6 9.25 12.9 12.3 11.8 11.2 10.7 10.295 56.8 54.3 51.8 49.5 47.3 45.197.5 61.1 58.4 55.7 53.2 50.8 48.6

BRS-a1 (mmHg/s)2.5 15.1 14.1 13.1 12.2 11.4 10.75 17.2 16.1 15.0 14.0 13.1 12.295 129.2 120.6 112.5 105.0 98.0 91.597.5 182.6 170.4 159.1 148.5 138.5 129.3

BRS-v (ms/mmHg)2.5 3.5 3.0 2.6 2.2 1.9 1.65 4.4 3.8 3.2 2.7 2.3 2.095 20.6 17.5 14.9 12.7 10.8 9.297.5 22.3 19.0 16.2 13.8 11.7 10.0

BRS-g (ms/s)2.5 91 74 60 49 40 325 95 78 63 51 42 3495 623 506 412 335 272 22197.5 810 658 535 435 353 287

BRS-g1 (ms/s)2.5 116 92 73 58 46 375 136 108 86 68 54 4395 1680 1334 1060 842 669 53197.5 2123 1687 1340 1064 845 671


P-values less than 0.05 were considered significant.Continuous variables are expressed as mean �standard deviation.

RESULTS

Of our recordings from 255 subjects, 175 recordingswere accepted for analysis—79 from men and 96 fromwomen. The excluded data were mainly comprised ofVM with a “flat-top” response, i.e., where BP changesfailed to fall below baseline. Demographic, HRDB, andVR data are presented in Table 1.

Gender and Age Effect on Parameters of Baroreflex

Sensitivity. As mentioned above, the final modelswere found to involve age only. Male and femaledata were therefore combined. Each of the log-trans-formed baroreflex sensitivity parameters were foundto be linearly associated with age. The age coefficientfor PRT was positive, indicating that PRT increasedwith age, and negative for all other parameters

(BRS-a, BRS-a1, BRS-v, BRS-g, and BRS-g1), suggest-ing that baroreflex sensitivity of both adrenergic andcardiovagal components decreased with age. Sum-maries of the regression models used to calculatepercentiles are shown in Table 2. The 2.5th, 5th,95th, and 97.5th percentiles for each parameter bydecades are listed in Table 3. Figure 2 illustrates theparameters of PRT, BRS-a, BRS-a1, and BRS-v as afunction of increasing age.

Correlations between Each Parameter of Baroreflex

Sensitivity. There were no significant correlationsbetween BRS-v and the parameters of the adrenergiccomponent of baroreflex sensitivity (PRT, BRS-a,and BRS-a1). As expected, correlations among pa-rameters of adrenergic baroreflex sensitivity weresignificant (Table 4).

Correlations between Parameters of Baroreflex Sensi-

tivity and Hemodynamic Indices of VM. We correlatedthe parameters of baroreflex sensitivity to some in-dices of interest of VM including phase II-L (i.e., theincrement of BP during phase II-L) and phase IV(i.e., the highest BP above baseline in phase IV)because these two indices have been widely used asindicators of adrenergic function. The correlationsbetween phase II-L and the parameters of PRT,BRS-a, BRS-a1, and BRS-g1 were significant. Phase IValso had significant correlations with these parame-ters in addition to BRS-v. The correlation coeffi-cients are listed in Table 5.

Correlations between Vagal Component of Baroreflex

Sensitivity and HRDB. HRDB, known to be a pure car-diovagal index, had strong correlations with BRS-v(rho � 0.53, P � 0.001), BRS-g (rho � 0.51, P �0.001), and BRS-g1 (rho � 0.42, P � 0.001), but nosignificant correlation with PRT, BRS-a, or BRS-a1.

FIGURE 2. Indices about baroreflex sensitivity as a function ofincreasing age. Lines show the 2.5th, 5th, 95th, and 97.5thpercentiles

Table 4. Correlations between the parameters of baroreflex sensitivity.

PRT BRS-a BRS-a1 BRS-v BRS-g BRS-g1

PRT — �0.49* �0.85* �0.10(P � 0.18)

�0.43* �0.73*

BRS-a — 0.71* �0.09(P � 0.23)

0.59* 0.47*

BRS-a1 — �0.07(P � 0.38)

0.42* 0.71*

BRS-v — 0.70* 0.60*BRS-g — 0.82*BRS-g1 —

*P � 0.001.


DISCUSSION

In a previous study, we documented that PRT pro-vides a good quantitative index of adrenergic barore-flex sensitivity and validated it against patients withdifferent severities of adrenergic failure.18 Consider-ing the antecedent BP decrement (that drives PRT)is different with each VM, an improved method thatrelates PRT to the antecedent BP change was intro-duced. The intuitive index is the ratio of BP deficit atthe beginning of recovery and PRT (phase III) yield-ing BRS-a. In the study of Schrezenmaier et al.,17 itwas found that PRT correlates better with MSNAfrom phase II-E and phase III combined than fromeach phase alone. MSNA in phase II-E comprises alarger response than MSNA in phase III by 25%.Thus, the formula was modified with the numeratoras the BP fall in early phase II-E plus 75% of anadditional fall in BP due to phase III (BRS-a1).

The effect of age on cardiovagal function hasbeen reported previously.11 BRS-v reflects the cardio-vagal component of the reflex and had a negativeage coefficient in regression analysis. PRT was pro-longed with age and both BRS-a and BRS-a1 de-creased with age, suggesting that the adrenergiccomponent also becomes blunted with age. How-ever, the coefficients of determination for PRT,BRS-a, and BRS-a1 in our age model were relativelylow compared with BRS-v, although they were allsignificant. This may imply that the adrenergic partis more robust during the aging process. However,there are some confounding factors to be consid-ered. The main stimulus for vasoconstriction (i.e.,responsible for BP recovery) is the preceding BPdecline in phase II-E,17 which was reported to in-crease with age in previous studies1,14 and also in thisstudy. Thus, although adrenergic baroreflex sensitiv-ity is diminished during aging, aging is associatedwith a larger stimulus to activate adrenergic vasocon-striction. The effect is more obvious in BRS-a andBRS-a1 than in PRT because the former have BPdecline in the numerator. Additionally, the BP re-covery curve is nonlinear and the recovery of BP islikely a combined result of baroreflex-mediated va-soconstriction and elastic recoil.

The indices of phase II-L and phase IV have beenwell accepted as indices to track adrenergic activa-tion. Phase II-L is mainly a result of �-adrenergicactivation.16 The mild to moderate correlations be-tween phase II-L and the parameters that reflectadrenergic barosensitivity (PRT, BRS-a, and BRS-a1)rather than vagal component (BRS-v) support thisnotion. In contrast, the mechanism of phase IV ismore complex, as it is more dependent on cardiacthan peripheral adrenergic tone.16 There was signif-icant correlation between phase IV and BRS-v as wellas the parameters pertaining to the adrenergic com-ponents (PRT, BRS-a, and BRS-a1), suggesting theinterplay of �-adrenergic, �-adrenergic, and vagalactivities in phase IV.

The indices of adrenergic barosensitivity (PRT,BRS-a, and BRS-a1) correlated well with each other,but not with BRS-v. This observation is consistentwith a previous report that there is no correlationbetween adrenergic sympathetic activity and vagalbaroreflex gains.15 These findings are also consistentwith clinical and laboratory observations that eachcomponent may be selectively or differentially in-volved in certain autonomic disorders. What are thepractical implications of these findings? First, thebetter correlation of BRS-a1 than BRS-a with phaseII-L, as well as the correlations with microneurogra-phy, suggest that BRS-a1 is the better index. Second,PRT remains an excellent index and, because of itsease of application, is suitable for routine use,whereas BRS-a1 is more appropriate for researchquantitation.

Supported in part by National Institutes of Health (grants NS32352, NS 44233, NS 43364), Mayo CTSA (grant U54RR 24150),and Mayo Funds.

REFERENCES

1. Denq JC, O’Brien PC, Low PA. Normative data on phases ofthe Valsalva maneuver. J Clin Neurophysiol 1998;15:535–540.

2. Eckberg DL, Harkins SW, Fritsch JM, Musgrave GE, GardnerDF. Baroreflex control of plasma norepinephrine and heartperiod in healthy subjects and diabetic patients. J Clin Invest1986;78:366–374.

3. Eckberg DL, Rea RF, Andersson OK, Hedner T, Pernow J,Lundberg JM, et al. Baroreflex modulation of sympathetic

Table 5. Correlations between baroreflex sensitivity and hemodynamic indices of Valsalva maneuver.

PRT BRS-a BRS-a1 BRS-v BRS-g BRS-g1

Phase II-L �0.55† 0.17* 0.66† �0.007 (P�0.93) 0.099 (P�0.19) 0.50†

Phase IV �0.44† 0.43† 0.52† �0.17* 0.15 (P �0.05) 0.28†

*Correlations are significant at the 0.05 level.†Correlations are significant at the 0.01 level.


activity and sympathetic neurotransmitters in humans. ActaPhysiol Scand 1988;133:221–231.

4. Fagius J, Wallin BG. Sympathetic reflex latencies and conduc-tion velocities in normal man. J Neurol Sci 1980;47:433–448.

5. Fagius J, Wallin BG. Sympathetic reflex latencies and conduc-tion velocities in patients with polyneuropathy. J Neurol Sci1980;47:449–461.

6. Freeman R. Assessment of cardiovascular autonomic func-tion. Suppl Clin Neurophysiol 2004;57:369–375.

7. Kirchheim HR. Systemic arterial baroreceptor reflexes.Physiol Rev 1976;56:100–177.

8. Low PA. Autonomic nervous system function. J Clin Neuro-physiol 1993;10:14–27.

9. Low PA. Testing the autonomic nervous system. Semin Neu-rol 2003;23:407–421.

10. Low PA, Benrud-Larson LM, Sletten D, Opfer-Gehrking TL,Weigand SD, O’Brien PC, et al. Autonomic symptoms anddiabetic neuropathy: a population-based study. Diabetes Care2004;27:2942–2947.

11. Low PA, Denq JC, Opfer-Gehrking TL, Dyck PJ, O’Brien PC,Slezak JM. Effect of age and gender on sudomotor and car-

diovagal function and blood pressure response to tilt in nor-mal subjects. Muscle Nerve 1997;20:1561–1568.

12. O’Brien PC, Dyck PJ. Procedures for setting normal values.Neurology 1995;45:17–23.

13. Parati G, Di Rienzo M, Mancia G. How to measure baroreflexsensitivity: from the cardiovascular laboratory to daily life.J Hypertens 2000;18:7–19.

14. Piha SJ. Autonomic responses to the Valsalva manoeuvre inhealthy subjects. Clin Physiol 1995;15:339–347.

15. Rudas L, Crossman AA, Morillo CA, Halliwill JR, TahvanainenKU, Kuusela TA, et al. Human sympathetic and vagal barore-flex responses to sequential nitroprusside and phenylephrine.Am J Physiol 1999;276:H1691–1698.

16. Sandroni P, Benarroch EE, Low PA. Pharmacological dissec-tion of components of the Valsalva maneuver in adrenergicfailure. J Appl Physiol 1991;71:1563–1567.

17. Schrezenmaier C, Singer W, Muenter-Swift N, Sletten D, LowPA. Adrenergic and vagal baroreflex sensitivity in autonomicfailure. Arch Neurol 2007;64:381–386.

18. Vogel ER, Sandroni P, Low PA. Blood pressure recovery fromValsalva maneuver in patients with autonomic failure. Neu-rology 2005;65:1533–1537.


ABSTRACT: The decremental response of the compound muscle actionpotential (CMAP) to provocative tests is not characterized in genetically verifiedmyotonic disorders. We therefore studied the relationship between decrementalresponses and mutation type in 10 patients with recessive myotonia congenita(rMC), two with paramyotonia congenita (PMC), nine with myotonic dystrophytype 1 (DM1), four with DM2, and 14 healthy people. CMAPs were measured atrest, just after a short exercise test (SET), and during short, 5- and 10-HZ,repetitive nerve stimulation (RNS) trains at 32°C and at 20°C. The degree ofdecrement was not related to the severity of clinical myotonia. Controls andPMC patients had similar responses when warm, but with cooling PMC patientshad a persistent decrement of CMAPs. In the rMC patients the decrementalresponses were related to the type of mutation of the CLCN1 gene, as adecrement was encountered in the T268M, R894X, IVS17�1 G�T, K248X,and 2149 del G, but not with the IVS1�3 A�T, F167L, or dominant A313Tmutations. In DM1 patients there was no relationship between decrement andCTG repeats. The degree of partial inexcitability in myotonic muscle membranetherefore depends on the mutation type rather than degree of clinical myotonia.RNS at 10 HZ is more sensitive than SET for demonstrating abnormalities inrMC patients when warm; differences are less marked when cold, which isuseful to diagnose PMC. Provocative tests are therefore useful in myotonias todemonstrate muscle inexcitability, which depends on the chloride or sodiumchannelopathy.

Muscle Nerve 36: 643–650, 2007

COMPARATIVE EFFICACY OF REPETITIVE NERVESTIMULATION, EXERCISE, AND COLD INDIFFERENTIATING MYOTONIC DISORDERS

PATRIK MICHEL, MD,1 DAMIEN STERNBERG, PhD,2 PIERRE-YVES JEANNET, MD,3 MURIELLE DUNAND, MD,1

FRANCINE THONNEY, PhD,4 WOLFRAM KRESS, MD, PhD,5 BERTRAND FONTAINE, MD, PhD,6

EMMANUEL FOURNIER, MD, PhD,7 and THIERRY KUNTZER, MD1

1 Nerve-Muscle Unit, Neurology Service, CHU Vaudois and University of Lausanne, Room BH7/469, 1011 Lausanne, Switzerland2 Biochemical Unit of Molecular and Cellular Cardiogenetics and Myogenetics, Universite Pierre et Marie Curie, Paris, France3 Neuropaediatrics Unit, Department of Paediatrics, CHU Vaudois and University of Lausanne, Lausanne, Switzerland4 Division of Medical Genetic, CHU Vaudois and University of Lausanne, Lausanne, Switzerland5 Institute of Human Genetics University of Wuerzburg Biozentrum, Wuerzburg, Germany6 Institut National de la Sante et de la Recherche, UMR546, RESOCANAUX and Universite Pierre et Marie Curie, Paris, France7 Department of Physiology, Groupe Hospitalier Pitie-Salpetriere and Universite Pierre et Marie Curie, Paris, France


Diagnosis of myotonic disorders is often made byclinicians in electrodiagnostic laboratories, but dis-tinction of the various disorders associated with myo-

tonic discharges has largely been based on a combi-nation of clinical features and genetic patterns. DNAtesting has certain shortcomings: it can take severalweeks to obtain the results, the mutation cannotalways be identified, certain myotonic disorders(e.g., those found in the context of a paralytic at-tack) may be secondary to acquired diseases, and, insome regions, the cost of analysis may be prohibitive.Certain clinical signs may be distinctive but, in addi-tion, some myotonic disorders may have character-istic changes in their muscle membrane excitability,with a decrement in electrically elicited compoundmuscle action potential (CMAP) amplitude follow-ing repetitive nerve stimulation (RNS),1,3,15 shortexercise,21,23 or long exercise.12,16 Most of the early

This article includes Supplementary Material available via the inter-net at http://www.mrw.interscience.wiley.com/suppmat/0148-639X/suppmat/

Abbreviations:: CMAP, compound muscle action potential; DM1, DM2,myotonic dystrophy type 1 or 2; MC, myotonia congenita; PMC, paramyoto-nia congenita; RNS, repetitive nerve stimulation; SET, short exercise testKey words: cold; muscle channelopathies; mutations; myotonia; repetitivenerve stimulation; temperatureCorrespondence to: T. Kuntzer; e-mail: [email protected]


Differentiating Myotonic Disorders MUSCLE & NERVE November 2007 643

studies used RNS trains and found decremental re-sponses (or progressive decrease in amplitude) inthe CMAPs of patients with myotonia congenita(MC) or myotonic dystrophy. Attempts to define thedistinguishing features of the CMAP decrement insmall series of patients showed that (1) some decre-ment of the CMAP occurred in all forms of myotoniawhen the duration of nerve stimulation was longenough,1 (2) a large decrement occurred especiallyin recessive MC,1,3,15,20,21,23 and (3) the presence of adecrement varied according to the specific mutationtype in MC.4,5,8

Besides RNS studies, exercise tests have alsoproved useful in the differential diagnosis of myo-tonic syndromes. Recent studies of patients withnondystrophic myotonic syndromes showed thatchloride and sodium channel mutations can be dis-tinguished by combining repeated short with longexercise tests8 and with cold.9 However, little isknown about the relative sensitivity of RNS and ex-ercise in distinguishing myotonic syndromes.

We recently defined the responses in healthysubjects to a short exercise test (SET) and duringshort RNS trains when warm and cold.13 We havenow examined the electrophysiological responses inpatients with several types of dystrophic and nondys-trophic myotonias with the same standardized pro-tocol in order to determine whether muscle mem-brane dysfunction correlates with genetic features ofthe underlying myotonia.


Patients. Twenty-five patients with a known myo-tonic disorder were recruited by personal contact.The patients were assigned to one of four groups(Table 1) on the basis of clinical manifestations,family history, and DNA testing or ancillary testsaccording to established criteria published by theEuropean NeuroMuscular Center (http://www.en-mc.org/nmd/diagnostic.cfm). Ten had recessiveMC (rMC), two had paramyotonia congenita (PMC),nine had myotonic dystrophy type 1 (DM1), andfour had a proximal myotonic myopathy (PROMMor dystrophia myotonica type 2, DM2). In the 10rMC patients (from eight different families) themode of inheritance was compatible with an autoso-mal-recessive trait and initial symptoms were presentin the first decade; in nine the diagnosis was con-firmed by determination of the mutations in one orboth alleles of the muscle chloride channel gene(CLCN1). The two PMC patients from two differentfamilies were the only affected members among thefamily, and both harbored a mutation in the muscle

sodium channel gene (SCN4A). The nine DM1 pa-tients (from eight different families) all had theDM-associated chromosome 19 CTG repeat. Thefour DM2 patients (from two families, one previouslypublished11) all had a CCTG expansion in intron 1of the ZNF9 gene on chromosome 3. All patientswere medication-free for �1 month before the ex-periments. Our Medical Ethics Committee approvedthe study and all subjects gave written informedconsent.

Electrophysiology. Clinical examination and elec-tromyographic studies with nerve conduction studieswere performed on a separate day prior to the pro-tocol evaluation. None of the patients had per-formed strenuous exercise before the examination.All studies were performed using a Viking IV ma-chine (Nicolet Viasys Healthcare, Madison, Wiscon-sin) according to a standardized protocol describedpreviously.13 In short, CMAPs were recorded fromthe abductor digiti minimi (ADM) muscle of theright hand with surface steel disc electrodes. Theulnar nerve was stimulated supramaximally at thewrist and two to five CMAPs were recorded at rest toensure a stable baseline response. The ADM musclewas then exercised for 10 s, corresponding to theshort exercise test (SET), and stimulated immedi-ately after its cessation and then once every 10 s for1 min. After 5-min rest, a 10-s RNS train was deliv-ered at 5 Hz. After another 5-min rest, a 5-s RNStrain was delivered at 10 Hz. Temperature was main-tained above 32°C (range 32.2–34.6°C) for the firstpart of the protocol. The muscle was then cooled to20–25°C and the three previous steps were repeated.Cooling below 25°C (range 19.7–24.6°C) was ob-tained with immersion of the hand for 5 min in coldtap water with electrodes being retaped to their orig-inal site, which had been defined by pen marks. Atthe end of the study the muscle was rewarmed toabove 32° by immersion of the hand in warm water,and two single CMAPs were obtained at rest to checkfor reversibility of the cold-induced changes inCMAPs. A surface temperature probe located 1 cmfrom the recording electrode recorded cutaneoustemperature constantly. All healthy and myotonicsubjects underwent the entire protocol except that theyoungest PMC patient participated in only part of theprotocol (warm and cold CMAPs at rest, warm SET).

Statistics. Relative changes in amplitude, duration,and area at different temperatures and before andafter provocative tests were calculated in the sameway as in controls.13 CMAP amplitude measurementswere made from baseline to negative peak, but for

644 Differentiating Myotonic Disorders MUSCLE & NERVE November 2007

Table 1. Demographic, phenotype, and mutations found in the patients.

Patient no. Sex AgeMutations or no. of CTG

repeats Warming-upTransient

weakness

Muscle bulk (H,hypertrophic; A,

amyotrophic)Hand

weaknessHand

myotonia

Recessive myotonia congenita1 M 19 IVS1�3 A�T �/�

(intron 1, splicing);F167L �/� (exon 4,missense)

� � 0 0 �

2 M 19 T268M �/� (exon 7,missense); R894X�/� (exon 23,nonsense)

�� H �� 0 ��

3 M 39 T268M �/� (exon 7,missense); possiblemissing secondmutation

�� H �� 0 ��

4 F 33* IVS17�1 G�T �/�(intron 17, splicing)

�� H � 0 �

5 M 37* IVS17�1 G�T �/�(intron 17, splicing)

�� H �� 0 �

6 M 29† K248X �/� (exon 6,nonsense); 2149 delG (exon 17,frameshift)

�� H ��

7 F 26† K248X �/� (exon 6,nonsense); 2149 delG (exon 17,frameshift)

�� H ��

8 M 32 F167L �/� (exon 4,missense); possiblemissing secondmutation

0 0 0 0 0

9 M 38 no mutation found inCLCN1

0 0 0 0 0

10 F 35 A313T �/� (exon 8,missense, dominant)

0 0 H � 0 �

Paramyotonia congenita (PMC)11 M 57 R1448H 0 0 0 0 012 M 12 T1700_E1703del 0 0 0 0 �

Dystrophia myotonica type 1 (DM1)13 M 39 500 0 0 A� � �14 M 22 670 0 0 0 � �15 F 51 870 0 0 A� � �16 F 49 300 0 0 A� � �17 F 56 670 � 0 A�� 18 F 36 670 � 0 0 � �19 M 46 670 � 0 A� � �20 M 28 200 � 0 0 � �21 F 58 670 � 0 A� � �

Dystrophia myotonica type 2 (DM2)22 F 48 CCTG expansion 0 0 0 0 023 M 61 CCTG expansion 0 0 A� � �24 M 54 CCTG expansion 0 0 A� � 025 F 58 CCTG expansion 0 0 0 0 �

Mutations are in the gene coding for the muscle chloride channel (CLCN1) in patients 1 to 10 and for the muscle sodium channel gene (SCN4A) in patients 11and 12. Hand myotonia was evaluated both by exercise or percussion. NF, all exons screened and no mutations found; 0, no (absent); �, mild; ��, obviousfindings.*Family 1.†Family 2.


the RNS studies, change in amplitude was measuredfrom peak to peak. Area and duration of the nega-tive phase of the CMAP were measured from onset tobaseline crossing. Median values were calculated be-cause of the small numbers of subjects and becausea parametric distribution could not be assumed. TheWilcoxon signed-rank test was used to determinesignificance in relative changes from baseline and indifferences between the warm and cold conditions.To compare the absolute values at warm and coldconditions, the Mann–Whitney rank-sum test wasused, as electrodes were removed for cooling andrewarming in water. Differences were consideredsignificant with a probability of P � 0.05.

RESULTS

Patients. The basic characteristics, main clinicaldata, and genetic status of the patients are shown inTable 1. The warming-up phenomenon was seen inmost rMC patients, and paradoxical myotonia exclu-sively in PMC patients. The severity of hand weaknessand myotonia varied widely among patients, exceptin the DM1 group, where weakness of hand musclesand myotonia were always encountered.

Eight different mutations (three missense, two non-sense, one frameshift, two splice site mutations) of theCLCN1 gene were found in the nine patients classifiedas rMC and in the woman (patient 10) with the mildestform of myotonia, a fluctuating myotonia occurringonly during her three pregnancies. Among the ninerMC patients, six (patients 1, 2, and 4–7) had a “full”recessive genotype (one homozygous mutation or twodifferent heterozygous mutations) highly in favor ofrecessive MC (or Becker generalized myotonia); twopatients had only one heterozygous mutation, with thatmutation being reported as recessive MC mutation,thus suggesting that a second heterozygous mutation

may have escaped molecular analysis; and in one, noCLCN1 mutation was found, thus weakening the diag-nosis of MC.

The six patients with the “full” recessive MC ge-notypes, and one patient with only one recessivemutation (patient 3) presented with the warming-upphenomenon and had stereotypic transient weak-ness at rest. This correlation did not seem to beinfluenced by the type (missense, nonsense, splicing,frameshift) or the pattern of combination of CLCN1mutations. In one (patient 8) with only one recessivemissense mutation in CLCN1, there was no warmingup or transient weakness at rest. One dominant mis-sense mutation (A313T) was found in patient 10 withthe pregnancy-induced myotonia, an already recog-nized cause of fluctuating myotonia.14 This molecu-lar finding corresponds to the diagnosis of dominantMC (Thomsen type).

CMAPs Evoked at Rest. The CMAP parameters ob-tained at rest are shown in Table 2. When subjectswere warm their CMAPs did not differ from controls.After cooling there was a tendency for the medianamplitude, area, and duration to increase in the rMCand DM groups, but contrary to the findings incontrols,13 the magnitude of this increase was notsignificant (Table 2). In the two PMC patients therewas a 30% decrease in both amplitude and area anda 60% increase in duration of the CMAPs. Afterrewarming, the median parameters were not signifi-cantly different from the first values obtained atbaseline, except for the two PMC patients, for whomneither amplitude, area, or duration recoveredwithin the 2 min that followed.

Short Exercise Test (SET). When warm and just afterthe SET, there was a significant increase in the me-

Table 2. CMAP parameters in the 14 controls and 25 myotonic patients.

Control rMC PMC DM1 DM2

Warm Cold Warm Cold Warm Cold Warm Cold Warm Cold

Amplitude (mV) 10 13.7† 10.5 12 9.6 7.1* 8 12.2 7.5 6.2Range 7.5;13.9 9.4;20.2 8;15.5 9.1;20.6 8.8;10.4 6.3;8 4;13.4 5.7;17.4 6;8.6 5.8;7.8P to control — — n.s. n.s. n.s. �0.001 n.s. n.s. �0.05 �0.05

Area (mVms) 34.1 52.7† 27.8 43.1* 25.9 17.8* 28.8 50.1* 21 33Range 20.2;44.9 35.9;86 18.6;45 24.6;96.8 22.3;29.5 16.5;19.2 14.6;45.8 18.9;78.7 20.2;24 20.2;58.4P to control — — n.s. �0.05 n.s. �0.05 n.s. n.s. �0.05 �0.05

Duration (ms) 5.7 8.3† 5.3 6.8* 5.15 8† 7.2 8.8 6.3 8.5*Range 4.6;7.2 5.8;12.4 3.9;6.6 5.2;10.3 4.8;5.5 8;8.1 6.4;10.6 7.3;12.1 5.9;7.4 7.8;14.7P to control — — �0.05 �0.01 �0.05 �0.05 �0.001 n.s. n.s. n.s.

P to controls indicates statistical differences between patients and controls. Statistical differences between values obtained when warm and then at coldtemperatures: *P � 0.01; †P � 0.01; ns, not significant.


dian value of CMAP amplitude in controls13 and inPMC patients, no change in the DM2 group, and asignificant decrease in the rMC and DM1 groups(Supplementary Fig. 1). No control subject showed adecrement in amplitude. A few myotonic patientsshowed a normal potentiation (two rMC, two DM1,one DM2), so that a decrement was not completelysensitive for myotonic disorders. Of the rMC pa-tients, eight had a decrement; a decrement exceed-ing 40% was only seen in patients 3, 6, and 7, and didnot resolve within 30 s as in the other patients andcontrols.

CMAP area also tended to decrease more in myo-tonias than in controls and PMC patients (greaterthan �20% in 0 controls; six rMC; all DM1, and twoDM2 patients). The shortest CMAP durations afterSET were obtained in the rMC and DM1 groups(�25% reduction as compared to controls in tworMC and eight DM1). When the SET was repeatedafter cooling (Fig. 1) there was a highly significantdecrease in the median value of CMAP amplitudeand area in the rMC and PMC groups, but not incontrols or DM groups. Six patients had a decreaseof greater than 40%, all in the rMC or PMC groups.This decrement resolved within 30 s in the rMCgroup, but no resolution was seen during the 60-sobservation period after the SET in the PMC pa-tients.

Repetitive Nerve Stimulation. When warm, there wasa significant increase in amplitude, a decrease induration of the 20th and 50th CMAPs (Supplemen-tary Fig. 2), and no change in area in controls and inthe PMC patients. A significant decrease in ampli-tude and areas was recorded in the rMC group at5-Hz and 10-Hz stimulation, the decrease beinggreater than 40% in six (patients 2–7) at 10 Hz. Inthe DM groups, no significant changes occurred inamplitude and area.

When the RNS was repeated with cooling, nostatistical changes in the median value of CMAPamplitude were observed in controls or DM groups,but the amplitude was significantly decreased in therMC and PMC groups. The decrease was �40% in 5of the 10 rMC and in the PMC patients, with changesobserved at 10 Hz but not at 5 Hz. In the DM1 groupa significant decrease in CMAP duration was ob-served.

Correlations between SET, RNS, and Genotype.

Changes in CMAP parameters were similar after SETand RNS. However, the sensitivity to the provocativetests was not the same, and varied according to thetype of myotonic disorder. When a decrement inamplitude or area was clearly abnormal (i.e., notseen in controls) following SET, the decrement wasalso abnormal after cooling and RNS trains at 10 Hzafter warming and cooling, but within the rMCgroup these results varied. A decrement greater than40% was indeed observed in three patients followingSET while warm, two other patients had an abnormaldecrement following SET in the cold. One furtherpatient had an abnormal decrement during RNStrains at 10 Hz in warm and cold conditions. In thisgroup we therefore found rMC patients with andwithout a decremental response, perhaps dependingon the type of mutations or mutation combinations:a decrement was encountered in patients 2–7, withmutations T268M, R894X, IVS17�1 G�T, K248X,and 2149 del G, but not in patients 1, 8, and 10 withdifferent mutations such as IVS1�3 A�T, F167L,and A313T (Fig. 2). Interestingly, there was no cor-relation in disease severity and provocative test re-sponse (Table 1). It is worth noting that the decre-ment was reproducible in the four rMC patientstested on 2 different days. For the two sibling pairs inthe rMC group, no influence of sex was observed infamily 1 (patient 4, woman; patient 5, man) or family2 (patient 6, man; patient 7, woman).

In the two PMC patients, an abnormal and per-sistent (see above) decrement was only seen aftercooling, to either SET or RNS; when warm, theresponses were similar to control subjects. In DMpatients the decrement varied widely during thetests, and the only consistent observation was a non-significant negative correlation in the magnitude ofthe amplitude of the decrement when warm andcold with the number of CTG repeats.

DISCUSSION

We have demonstrated that provocative tests mayinduce transient or persistent abnormal changes in

FIGURE 1. Percent change in CMAP amplitude obtained afterSET in controls and 25 myotonic patients. Note there are DM1patients who have a postexercise decrement of up to 30%, andtherefore a decrement greater than 40% is needed to be specif-ically encountered in rMC patients (patients 3, 6, and 7).


CMAP parameters in myotonic patients. Thesechanges are variable in magnitude and type, andcorrespond to a partial inexcitability of the myotonicmuscle membrane. This study confirms previousdata published before genetic investigations wereperformed, provides information on the usefulnessof several provocative tests, and shows that they candiscriminate between different groups of myotonicdisorders depending on the genotype.

Responses to Cooling the Recorded Muscle at Rest.

In healthy controls, we have previously demon-strated a potentiation of almost 30% in the threemeasured CMAP parameters following supramaxi-mal stimulations at temperatures between 20–25°C,13

and hypothesized a reduced functioning of theNa�-K� pump to explain this phenomenon. No dif-ference was observed between the rMC and DM1patients and controls; in the four DM2 patients,cooling induced a mild reduction in CMAP ampli-tude with no change in area.

In the two PMC patients, cooling induced a sig-nificant and persistent reduction of CMAP ampli-tude and area, in agreement with earlier electrophys-iological results.19,24 These patients had the R1448Hand T1700_E1703del mutations, respectively, in themuscle sodium channel gene. This muscle inexcit-ability points to a possible dysfunction of the voltage-gated channel gating process recently demonstratedin PMC, in which cold induces malfunction of Na�

channel gating kinetics,2 with a shift in the steady-state activation curve to hyperpolarized potentials atlower temperatures. The resulting increase in thewindow current, slower inactivation and faster recov-ery after inactivation may explain the pronouncedclinical symptoms when cold.17 The overall result isan enhanced and prolonged Na� influx into musclecells, causing sustained depolarization of the cellmembrane. Although the number of PMC patients issmall, our results suggest that provocative tests aftercooling could be a simple way to discriminate asubset of PMC due to sodium channel mutationsfrom other myotonic syndromes.

Effects Induced by RNS and SET. The decrement inCMAP amplitude and area demonstrated whenwarm in the rMC and DM patients is in agreementwith the findings of Streib et al.21,23 The decrementwas previously linked to the concept of transientweakness, defined by a sudden lapse in power duringsustained activity after rest, which was described as avery early disabling feature of rMC.1,6,18 However, wefound that the presence of a decrement did notcorrelate with disease severity, type of mytonia, orassociated signs, in agreement with what has alsobeen described by others.1,3,5

Two new findings were observed in the presentstudy. First, a large decrement could be demon-strated when warm in two thirds of our rMC groupfollowing a 10-Hz RNS train but was present in onlyone third of the cases following SET. Although RNStrains are less comfortable for patients, our resultsclearly suggest that RNS is more sensitive than SETin defining rMC patients with mutant chloride chan-nels that induce partial inexcitability. Changes werenot statistically significant at 5 Hz and therefore RNSshould be performed at 10 Hz.13 Second, the decre-ment was larger in rMC than DM patients, with adecrement larger than 40% clearly indicating rMCpatients.

In the PMC patients, no decrement was recordedby SET or RNS when muscle was warm, arguingagainst the idea that some degree of inexcitabilitycan be encountered in PMC patients.16,22 In ourstudy, persistent inexcitability was only observed ei-

FIGURE 2. Correlation between the presence of a decrementand mutation type in the rMC group. Patients 1–4 were examinedtwice on alternate days, with the same results. Genotypes includ-ing mutations T268M, R894X (patients 2, 3), IVS17�1 G�T(patients 4, 5), K248X and 2149 del G (patients 6, 7) induce anobvious decrement after RNS trains at 10 HZ, but genotypesF167L �/� and IVS1�3 A�T � (patients 1, 8), and the dominantA313T mutation, are not associated with a decremental re-sponse. The decrement was not uniform when present. No cor-relation was found between decrement and clinical signs (Table1). The 20th and 50th CMAPs were compared with the firstCMAP percent changes in area.


ther by SET or RNS when muscles were cold. Theseresults are in agreement with recent studies in alarge series of PMC patients demonstrating (1) vari-able changes in CMAP amplitude after a first SET atroom temperature, and (2) drastic declines in CMAPamplitude and area when SET is repeated8 or com-bined with cooling.9 The combined effect of coldand SET could enhance the biophysical alterationsinduced by certain muscle sodium channel muta-tions, leading to a persistent membrane depolariza-tion and muscle inexcitability.

Overall, our results confirm the usefulness ofSET and of RNS for discriminating between rMCand DM and for recognizing PMC patients in coldconditions, thereby enhancing the diagnostic sensi-tivity of the tests.

Correlations between rMC, DM, and Genotype. InrMC, it has been suggested that CMAP decrementmay be related to variation of CLCN1 alleles,5,8,9

which include many types of mutated alleles (non-sense, splice site mutations, frameshift mutations,missense mutations) and combinations. Our resultsalso suggest such a relationship, as a decrement wasencountered in individuals (1) with homozygosity orcompound heterozygosity for nonsense, splice site,and/or frameshift mutations, resulting in genotypesthat probably lead to low or null expression of theCLCN1 channel through decay of nonsense tran-scripts (mutations IVS17�1 G�T, K248X, and 2149del G), or (2) with the T268M or R894X mutations,which probably lead to the expression of specificmissense or truncated channel subunits. In contrast,no decrement was observed in individuals with com-pound heterozygosity for IVS1�3 A�T and F167L,which probably lead to the expression of F167Ldimers only, or simple heterozygosity for the domi-nant A313T mutation, which leads to mixed expres-sion of mutated A313T subunits and wildtype sub-units (Fig. 2). Other missense mutations may belocated in different areas of the channel subunitsand result in variant electrophysiological propertiesfor CLCN1 homodimers or heterodimers. This wasclearly shown for missense mutations affecting thegate selectivity of the channel.10 The influence ofmutation type on CMAP decrement needs to beconfirmed and refined by further studies. It is possi-ble that the same mutation may be associated withvarious RNS responses, in the same way that carriersof the R894X mutation exhibit different phenotypes,with a recessive or dominant expression.7

RNS responses may also vary within a family dueto a differential allelic expression in dominant ped-igrees.4 In the DM1 patients, we were surprised to

encounter an inverse correlation between the decre-ment and the number of CTG repeats, although thistrend was not significant. As severity of DM1 diseaseis correlated with the number of repeats, the ex-pected response would be an increase in decrementwith increasing repeats. However, decrement is di-rectly related to the transient postexercise weaknessand not to severity of the disease. One explanationcould be the extent of the dystrophic process thatmasks the functional changes observed by decre-ment, but this needs to be studied with a largergroup of patients.

We thank Drs. Eric Berrut, Michel R. Magistris, and FrancoisOchsner for allowing us to examine their patients.

REFERENCES

1. Aminoff MJ, Layzer RB, Satya-Murti S, Faden AI. The declin-ing electrical response of muscle to repetitive nerve stimula-tion in myotonia. Neurology 1977;27:812–816.

2. Bouhours M, Sternberg D, Davoine CS, Ferrer X, Willer JC,Fontaine B, et al. Functional characterization and cold sensi-tivity of T1313A, a new mutation of the skeletal muscle so-dium channel causing paramyotonia congenita in humans.J Physiol (Lond) 2004;554:635–647.

3. Brown JC. Muscle weakness after rest in myotonic disorders;an electrophysiological study. J Neurol Neurosurg Psychiatry1974;37:1336–1342.

4. Colding-Jorgensen E. Phenotypic variability in myotonia con-genita. Muscle Nerve 2005;32:19–34.

5. Colding-Jorgensen E, DunO M, Schwartz M, Vissing J. Decre-ment of compound muscle action potential is related to mu-tation type in myotonia congenita. Muscle Nerve 2003;27:449–455.

6. Deymeer F, Cakirkaya S, Serdaroglu P, Schleithoff L, Leh-mann-Horn F, Rudel R, et al. Transient weakness and com-pound muscle action potential decrement in myotonia con-genita. Muscle Nerve 1998;21:1334–1337.

7. Duno M, Colding-Jorgensen E, Grunnet M, Jespersen T, Viss-ing J, Schwartz M. Difference in allelic expression of theCLCN1 gene and the possible influence on the myotoniacongenita phenotype. Eur J Hum Genet 2004;12:738–743.

8. Fournier E, Arzel M, Sternberg D, Vicart S, Laforet P, EymardB, et al. Electromyography guides toward subgroups of muta-tions in muscle channelopathies. Ann Neurol 2004;56:650–661.

9. Fournier E, Viala K, Gervais H, Sternberg D, Arzel M, LaforetP, et al. Cold extends electromyography distinction betweenion channel mutations causing myotonia. Ann Neurol 2006;60:356–365.

10. Grunnet M, Jespersen T, Colding-Jorgensen E, Schwartz M,Klaerke DA, Vissing J, et al. Characterization of two newdominant ClC-1 channel mutations associated with myotonia.Muscle Nerve 2003;28:722–732.

11. Kohler A, Burkhard P, Hefft S, Bottani A, Pizzolato GP,Magistris MR. Proximal myotonic myopathy: clinical, electro-physiological and pathological findings in a family. Eur Neu-rol 2000;43:50–53.

12. Kuntzer T, Flocard F, Vial C, Kohler A, Magistris M, Labarre-Vila A, et al. Exercise test in muscle channelopathies andother muscle disorders. Muscle Nerve 2000;23:1089–1094.

13. Kuntzer T, Michel P. Muscle membrane polarisation afterprovocative tests, and after cooling: the normal CMAPchanges to be expected. Clin Neurophysiol 2004;115:1457–1463.


14. Lacomis D, Gonzales JT, Giuliani MJ. Fluctuating clinicalmyotonia and weakness from Thomsen’s disease occurringonly during pregnancies. Clin Neurol Neurosurg 1999;101:133–136.

15. Lambert EH, Millikan CH, Eaton LM. Stage of neuromuscu-lar paralysis in myotonia [Abstract]. Am J Physiol 1952;171:741.

16. McManis PG, Lambert EH, Daube JR. The exercise test inperiodic paralysis. Muscle Nerve 1986;9:704–710.

17. Mohammadi B, Mitrovic N, Lehmann-Horn F, Dengler R,Bufler J. Mechanisms of cold sensitivity of paramyotonia con-genita mutation R1448H and overlap syndrome mutationM1360V. J Physiol (Lond) 2003;547:691–698.

18. Ricker K, Meinck HM, Stumpf H. Neurophysiologische Un-tersuchungen uber das Stadium passagerer Lahmung by Myo-tonia congenita und Dystrophia myotonica. Z Neurol 1973;204:135–148.

19. Ricker K, Samland O, Peter A. Elektrische und mechanischeMuskelreaktion bei adynamia episodica und paramyotoniacongenita nach Kalteeinwirkung und Kaliumgabe. J Neurol1974;208:95–108.

20. Rossi B, Rossi A, Sartucci F. Repetitive nerve stimulation inthe differential diagnosis of congenital myotonia. Ital J Neu-rol Sci 1984;5:385–390.

21. Streib EW. Differential diagnosis of myotonic syndromes.Muscle Nerve 1987;10:603–615.

22. Streib EW. Successful treatment with tocainide of recessivegeneralized congenital myotonia. Ann Neurol 1986;19:501–504.

23. Streib EW, Sun SF, Yarkowski T. Transient paresis in myotonicsyndromes: a simplified electrophysiological approach. Mus-cle Nerve 1982;5:719–723.

24. Subramony SH, Malhotra CP, Mishra SK. Distinguishingparamyotonia congenita and myotonia congenita by electro-myography. Muscle Nerve 1983;6:374–379.


ABSTRACT: We determined the prevalence of muscle acetylcholine re-ceptor (AChR) antibodies in patients with adult-acquired generalized myas-thenia gravis (MG), the seroconversion rate at 12 months, and the preva-lence of muscle-specific tyrosine kinase (MuSK) antibody amongpersistently seronegative patients. We identified 562 consecutive MayoClinic patients with MG based on clinical and electrophysiological criteria. Atpresentation, 508 patients (90.4%) tested positive for AChR binding orAChR modulating antibodies. After 12 months, 15.2% of initially seronega-tive patients had become seropositive, yielding a seronegativity rate of 8.2%(95% confidence interval: 6.2–9.6%). Among seronegative patients not re-ceiving immunosuppressants, 38% were MuSK antibody-positive and 43%were seropositive for nonmuscle autoantibodies. Classification as seroneg-ative MG should be reserved for nonimmunosuppressed patients with gen-eralized MG who lack muscle AChR binding, AChR modulating, or MuSKantibodies at presentation and at follow-up of at least 12 months.

Muscle Nerve 36: 651–658, 2007

FREQUENCY OF SERONEGATIVITY INADULT-ACQUIRED GENERALIZEDMYASTHENIA GRAVIS

KOON HO CHAN, MD,1 DANIEL H. LACHANCE, MD,1,2

C. MICHEL HARPER, MD,2 and VANDA A. LENNON, MD, PhD1–3

1 Department of Laboratory Medicine & Pathology, Mayo Clinic, College of Medicine,200 First Street S.W., Rochester, Minnesota 55905, USA

2 Department of Neurology, Mayo Clinic, College of Medicine,Rochester, Minnesota, USA

3 Department of Immunology, Mayo Clinic, College of Medicine,Rochester, Minnesota, USA


Nicotinic acetylcholine receptors (AChR) of inner-vated skeletal muscle are the principal interactionsites for pathogenic autoantibodies in patients withacquired myasthenia gravis (MG). Pathogenic mech-anisms documented for this autoantibody includecomplement-mediated destruction of the postsynap-tic AChR-bearing membrane in muscle, accelerateddegradation of AChR initiated by cross-linking

through binding of bivalent IgG,7,25 and, least com-monly, blockade of the neurotransmitter bindingsite on AChR.1,18 The assay employed most com-monly to detect AChR antibody in serum is a radio-immunoprecipitation assay, initially reported to de-tect AChR antibody in �87% of MG patients.30

The seronegativity frequency reported subsequentlyhas ranged from 7% to 34% for all acquired MGpatients, and from 6%–25% for generalizedMG.6,23,26,30,35,37,39,40 It is agreed universally thatAChR antibody is detected least frequently in pa-tients whose MG is restricted clinically to extraocularmuscles or who are in clinical remission.26,29,30,40,42

The mean onset age of generalized MG in AChRantibody–negative patients is younger than in AChRantibody–positive patients,2 but a caveat in assessingseronegativity rates is that inclusion of juvenile casesincreases the risk of inadvertently including congen-ital myasthenic syndromes because those patientspresent most frequently in childhood. Their neuro-logical presentations and electromyographic abnor-malities often resemble those of acquired MG and


Abbreviations: AChE, acetylcholinesterase; AChR, nicotinic acetylcholinereceptor; CI, confidence interval; CT, computed tomography; DM-1, diabetesmellitus, type 1; EMG, electromyography; GAD65, glutamic acid decarboxyl-ase-65; IgG, immunoglobulin G; MG, myasthenia gravis; MGFA, MyastheniaGravis Foundation of America; MUP, motor unit potential; MuSK, muscle-specific tyrosine kinase; NIS, Neuropathy Impairment Score; SFEMG, single-fiber EMGKey words: AChR antibody; adult-acquired myasthenia gravis; autoantibodyevaluation; MuSK antibody; seronegative myasthenia gravisCorrespondence to: D. H. Lachance; e-mail: [email protected].


Seronegative Myasthenia Gravis MUSCLE & NERVE November 2007 651

the family history may be uninformative becausemost congenital syndromes are autosomal-recessivetraits.10

Observations that plasma exchange and immu-nosuppressant therapy benefit some seronegativeMG patients,15,33 and that serum IgG from such pa-tients may impair neuromuscular transmission wheninjected into mice,4,34 suggested that some seroneg-ative cases are autoantibody mediated. This sugges-tion was confirmed when antibodies to MuSK wereidentified as a novel marker for a variant of autoim-mune MG.11,17,31,36 MuSK is the muscle postsynapticmembrane receptor for the neurotrophic factoragrin and promotes stable integration of AChR atthat site.12 MuSK antibody has been reported in�40% of seronegative generalized MG patients of allages. This form of autoimmune MG is benefited byplasmapheresis and other immunomodulating ther-apies. Most patients lack thymic pathology and mostreports suggest thymectomy is not benefi-cial.11,21,22,36,38 For three decades, “seronegative MG”has been defined as the occurrence of clinical andelectrophysiological evidence of acquired general-ized MG in the absence of AChR antibody detectedby a single assay employing detergent-solubilized hu-man skeletal muscle as the source of AChR (and125I-�-bungarotoxin as a tracer for AChR immuno-precipitation). Sensitivity is optimal with antigen de-rived from a combination of innervated and dener-vated primate muscle14 and is lower when theantigen source is a human muscle cell line express-ing only fetal-type AChR.3 Results for assays of mus-cle AChR-modulating antibody25 have not beentaken into consideration in determining the fre-quency of seronegativity in patients with acquiredMG. Furthermore, most studies have included pa-tients undergoing treatment with immunosuppres-sants, which can cause apparent seronegativity.24 Re-sults of serological evaluation repeated at 12 monthsor more after onset of MG symptoms have rarelybeen reported for patients who are initially seroneg-ative to determine whether AChR binding or AChRmodulating antibodies are detectable later.

In this study we assessed the prevalence of bothof these autoantibodies to determine the rate ofseronegativity in a large cohort of patients present-ing to the Neuromuscular Clinic at Mayo Clinic(Rochester, Minnesota) with symptoms and signsconsistent with generalized autoimmune MG begin-ning in adulthood. We report clinical, electrophysi-ological, and serological characteristics in MG pa-tients whom we classified as seronegative.


Patients. This retrospective study was approved byour Institutional Review Board. Study subjects wereascertained through detailed review of records for allpatients whose final diagnosis coded in the MayoClinic Electromyography Laboratory (in the periodJanuary 2, 1987, to August 21, 2000) was MG or amyasthenic syndrome. Clinical observations, neuro-logical deficits, electrophysiological findings, and se-rological and radiological results were abstracted.Patients classified as having acquired MG had thefollowing attributes compatible with that diagnosis:(1) clinical history; (2) objective weakness on physi-cal examination (including, but not limited to, ocu-lar muscle); and (3) electromyographic abnormali-ties, either 10% decrement in the compound muscleaction potential amplitude or area during 2-Hz re-petitive stimulation of at least two nerves, or wide-spread motor unit potential (MUP) variation or ab-normal single-fiber EMG (SFEMG) in at least onecranial or limb muscle. When SFEMG was employed,either the frontalis, extensor digitorum communis,or any other clinically involved muscle was examinedas long as established normal reference values forthat muscle were available for our laboratory. A pos-itive edrophonium (Tensilon) test or improvementof muscle strength by oral acetylcholinesterase inhib-itor were considered supportive, but not essential,diagnostic criteria.

Patients were excluded from the study if: (1) thefinal neurological diagnosis was other than MG (e.g.,neuropathy, radiculopathy, motor neuron disease,myopathy, or Lambert–Eaton syndrome); (2) theclinical history or electrophysiological studies (in-cluding microelectrode studies in some cases) wereconsistent with a congenital myasthenic syndrome;(3) the neurological symptoms began before age 17years; (4) the final diagnosis was uncertain on thebasis of clinical and electrophysiological findings; or(5) serological testing was not done.

We recorded age, sex, and all serological data.For seronegative patients, we recorded details ofneurological history and examination, electrophysi-ological data, and history or clinical evidence ofautoimmunity. We tested available sera additionallyfor MuSK antibody and other organ-specific andnonorgan-specific autoantibodies. We classified pa-tients clinically according to recommendations ofthe Myasthenia Gravis Foundation of America(MGFA)19 and we graded neurological deficits by amodified Neuropathy Impairment Score (NIS), us-ing the following clinical scale for assessment ofmuscle strength compared to normal for age and

652 Seronegative Myasthenia Gravis MUSCLE & NERVE November 2007

sex: 0, normal strength; 1, mild weakness (25%weaker than normal); 2, moderate weakness (50%weaker than normal); 3, severe weakness (75%weaker than normal); and 4, unable to generate anyforce.8 We summated the scores for each of 22 mus-cle groups bilaterally to obtain a total score between0 and 176. This scale has been validated in patientswith peripheral neuropathy9 and was used to assessthe efficacy of 3,4-diaminopyridine in patients withLambert–Eaton syndrome.32 We present the greatestseverity score recorded for each patient.

Serological Tests. Tests were performed in theMayo Clinic’s Neuroimmunology Laboratory. AChRbinding antibody was detected by immunoprecipita-tion assay using 125I-�-bungarotoxin to label AChRssolubilized in Triton X-100 from membranes of am-putated ischemic human skeletal muscle.14 AChR-modulating antibody was determined by a bioassayin which the patient’s serum is incubated with amonolayer of cultured human myogenic cells for16 h at 37°C before adding 125I-�-bungarotoxin todetermine the percent loss of surface AChR.18 Stria-tional antibodies were detected by enzyme immuno-assay using human skeletal muscle sarcomeric pro-teins as antigen.5,14 Neuronal voltage-gated calciumchannel antibodies of P/Q-type and N-type weredetected by immunoprecipitation using 125I-�-conopeptide MVIIC or 125I-�-conopeptide GVIA asligands to label the respective channel proteins sol-ubilized in digitonin from membranes of autopsiedhuman cerebral cortex.27 MuSK antibody was de-tected by immunoprecipitation using 125I-labeled re-combinant extracellular domain of human MuSK.17

RESULTS

A total of 562 patients fulfilled criteria for classifica-tion as adult-acquired generalized MG. The meanonset age was 51.9 years (range, 17–87 years) and302 patients (54%) were male. At symptom onset, 39patients (6.9%) had MG that clinically was ocularrestricted (MGFA Class I), but electrophysiologicalfindings were compatible with generalized MG; 302patients (53.7%) had mild generalized MG (MGFAClass II); 211 patients (37.5%) had moderate gener-alized MG (MGFA Class III), and 10 patients (1.8%)had severe generalized MG (MGFA Class IV). Figure1 illustrates the serological findings for this cohort atdiagnosis and at 12 or more months after diagnosis.At initial testing, 508 patients (90.4%) had positiveserology for AChR antibodies: 500 had AChR bind-ing antibody (98.4%; values ranged from 0.05–3,295nmol/L; Fig. 2), and 450 had AChR modulating

antibody (88.4%; values ranged from 37%–100%AChR loss). Eight of these 508 patients were consis-tently seropositive for AChR modulating antibodyonly. Thus, the frequency of seronegativity for AChRantibodies at initial testing was 9.6% (54 of 562). Ofthis seronegative group, 19 patients (35%) were re-ceiving immunosuppressant therapy. Among thewhole group of 562 patients, 200 (35.6%) had stria-tional antibodies upon initial testing (titers rangedfrom 120–122,880). One patient of the 54 who wereseronegative for AChR antibodies was seropositivefor striational antibody.

A follow-up serum specimen was available for 33of the 54 initially seronegative patients; 21 were un-available for follow-up evaluation (including the sin-gle patient who was seropositive for striational anti-body without AChR antibodies). For those withfollow-up serum, the mean onset age was 46.7 years(range 31–82 years), 21 patients (61%) were female;at onset 3 (9%) belonged to MGFA Class I, 17 (52%)to Class II, and 13 (39%) to Class III. For the 21patients without follow-up serum, the mean onsetage was 50.3 years (range 17–81 years), 14 (67%)were female, and at onset 2 (10%) belonged toMGFA Class I, 11 (52%) to MGFA Class II, and 8(38%) to MGFA Class III. All patients in both groupshad normal chest computed tomography (CT) with-out evidence of thymoma or thymic hyperplasia.Thus, the two groups did not differ in terms of meanonset age, sex, duration of symptoms, severity(MGFA classification), proportion with disease clin-ically restricted to extraocular muscles, proportiontaking immunosuppressant medication when serumwas drawn, or chest CT evidence of thymic hyperpla-sia. Five of the 33 patients whose serological testing

FIGURE 1. Serological evaluation of 562 adult patients at pre-sentation of acquired generalized myasthenia gravis and at 12months or later.


was repeated 12 months or more after initial testinghad positive results: three had AChR binding anti-body only and two had AChR modulating antibodyonly. None of the five seroconverted patients hadreceived immunosuppressant therapy. If this ob-served 15.2% rate of seroconversion at 12 months ispresumed for the 21 initially seronegative patientswho lacked a follow-up serum specimen at 12months, we estimate three more seropositive pa-tients (95% confidence interval [CI]: 3–27) and atotal of 8 (95% CI: 5–24) seropositive among 54patients who were seronegative at initial testing. Thisyields an overall seronegativity rate of 8.2% for mus-cle AChR antibodies (95% CI: 5.9%–10.5%). This islikely a high estimate because 35% of the initiallyseronegative patients were receiving immunosup-pressant medication at the time of testing.

Among the 28 patients whose follow-up serologyremained negative, seven were receiving immuno-suppressant medication. Removal of those seven pa-tients from consideration identified 21 individualsamong the initial study group who were unequivo-cally seronegative at 12 months after clinical diagno-sis. Table 1 summarizes their clinical characteristics.None were seropositive for P/Q-type or N-type cal-cium channel antibodies. Eight patients (38%) wereseropositive for MuSK antibody; seven of those werefemale (88%). The mean age at MG onset for MuSKantibody-positive patients was 38 years (range 31–48). All had generalized weakness affecting ocular,

bulbar, and trunk muscles with or without limb in-volvement. The mean maximal modified NIS scorewas 38.4 (range 18–52). A myasthenic crisis was re-corded for two patients. Six patients had significantdecrement during repetitive nerve stimulation (onlytwo in noncranial muscles). Electrophysiological ab-normality in two patients was detected only by single-fiber EMG. All had normal chest CT scans.

The 13 seronegative patients who lacked MuSKantibody were older (mean age of MG onset 51.3years; range 34–82 years) and six were female(46%). The mean maximal modified NIS score dur-ing follow-up was 21.5 (range 5–44). No patient hada myasthenic crisis. Eleven had weak extraocular andlimb muscles, with or without bulbar and truncalmuscle involvement. Weakness in the remaining twopatients was restricted to extraocular muscles clini-cally, but EMG abnormalities in limb muscles wereconsistent with generalized MG (varying motor unitpotentials on needle EMG or abnormal SFEMG, seeMaterials and Methods). Eight patients had signifi-cant decrement on repetitive nerve stimulation (fivein noncranial muscles). The other five patients hadabnormalities on SFEMG, and three of them hadvarying motor unit potentials on needle EMG. Thethymus in all 13 patients was normal for age by chestCT scanning.

We have summarized the treatments, outcomes,and immunological characteristics for the 21 patientsidentified as unequivocally seronegative in a supple-mentary table at http://www.mrw.interscience.wiley.com/suppmat/0148-639X/suppmat/. Six of the eightMuSK antibody-positive patients received immunosup-pressant therapies and all responded favorably.Thymectomy was performed in two patients; histologywas normal for age in both patients and no clinicalimprovement ensued for the single patient treated withthymectomy without immunosuppressant medication.Five MuSK antibody-positive patients (63%) had evi-dence of coexisting autoimmunity: three had an unre-lated autoimmune disease, one had a family history ofautoimmune disease, and two had coexisting organ-specific or nonorgan-specific autoantibodies: thyro-globulin antibody (one); thyroid peroxidase antibody(one); and antinuclear antibody (one). Six of the 13patients lacking MuSK antibody received immunosup-pressant medications and all responded favorably.Thymectomy was performed in four patients (three ofthe four also received immunosuppressant therapies)and histology was normal for age in all of them. Thesingle patient treated with thymectomy without immu-nosuppressant medication had satisfactory clinical im-provement. Ten of the 13 patients lacking MuSK anti-body (77%) had evidence of coexisting autoimmunity:

FIGURE 2. Serum values for muscle AChR binding antibodymeasured by radioimmunoprecipitation of 125I-�-bungarotoxin-complexed AChR solubilized from ischemic limb muscle fromseveral patients (mostly diabetic). All sera yielding values greaterthan 0.02 nmol/L were retested and tested additionally with 125I-�-bungarotoxin alone. The latter value was subtracted from thefinal results to exclude false-positive results.14 The median valuefor the 500 patients who were seropositive in this assay was 6.62nmol/L (range 0.03–3295 nmol/L). None of 178 adult healthycontrol subjects (male and female, age 18–83) were positive.


Table 1. Clinical features, treatments, and outcomes in patients with adult-acquired generalized MG and seronegativity for AChRantibodies.

Patientnumber

Sex/ageat onset

MGFAclass

WorstmNIS Crises Treatment

Treatmentresponse

Evidence for autoimmunity

Coexistingdisease

Autoantibodiesdetected

Familyhistory

1 F/48 IIIa 48 No AChE inhibitor,azathioprine

Improved None MuSK Yes

2 F/42 IIb 18 No AChE inhibitor,prednisone

Improved Graves’disease, DM-1

MuSK, Tg, TPO No

3 F/35 IIa 32 Yes Thymectomy Unchanged Raynaud’sphenomenon,lichen planus

MuSK No

4 F/31 IIb 36 No AChE inhibitor,prednisone

Improved Hashimoto’sthyroiditis

MuSK No

5 F/34 IIIb 52 Yes AChE inhibitor,prednisone

Improved None MuSK No

6 F/35 IIIb 50 Yes AChE inhibitor,prednisone

Improved None MuSK No

7 M/41 IIb 38 No Thymectomy,AChEinhibitor,prednisone,azathioprine

Minimalmanifestation

None MuSK, ANA No

8 F/39 IIb 33 No AChE inhibitor Improved None MuSK No9 M/68 I 8 No AChE inhibitor Improved Raynaud’s

phenomenonNone No

10 F/44 IIIa 44 No Thymectomy,AChEinhibitor,azathioprine

Improved None None No

11 M/34 IIa 22 No AChE inhibitor Unchanged None Tg, GAD65 No12 F/48 I 5 No None Unknown Pernicious

anemia,Hashimoto’sthyroiditis

GPC, GAD65,Tg, TPO

No

13 F/65 IIa 18 No AChE inhibitor,prednisone

Improved Graves’ disease Tg, TPO Yes

14 M/82 IIIa 25 No AChE inhibitor,azathioprine

Improved None AMA, SMA No

15 F/35 IIb 21 No Thymectomy,AChEinhibitor,azathioprine

Improved Rheumatoidarthritis

RF Yes

16 M/41 IIa 19 No AChE inhibitor Improved None None No17 M/67 IIb 22 No AChE inhibitor,

prednisonePartial

remissionHashimoto’s

thyroiditisNone Yes

18 M/50 IIb 20 No Thymectomy,AChEinhibitor

Improved None None None

19 F/36 IIIb 43 No Thymectomy,AChEinhibitor,prednisone,azathioprine

Improved Hashimoto’sthyroiditis

None None

20 F/50 IIa 15 No AChE inhibitor Minimalmanifestation

Graves’ disease Tg, TPO None

21 M/47 IIa 18 No AChE inhibitor Minimalmanifestation

None SMA None

AchE, acetylcholinesterase; AMA, antimitochondrial; GAD65, glutamic acid decarboxylase-65; GPC, gastric parietal cell; MGFA, Myasthenia Gravis Foundationof America; mNIS, modified Neuropathy Impairment Score; MuSK, muscle-specific kinase; RF, rheumatoid factor; SMA, smooth muscle antibody; Tg,thyroglobulin; TPO, thyroperoxidase.


seven had an unrelated autoimmune disease and sevenhad coexisting organ-specific or nonorgan-specific au-toantibodies: thyroglobulin antibody (four); thyroidperoxidase antibody (three); GAD65 antibody (two);smooth muscle antibody (two); gastric parietal cell an-tibody (one); antimitochondrial antibody (one); orrheumatoid factor (one).

We performed additional testing on stored se-rum from 23 of the 28 initially seronegative patientswho failed to fulfill our strict criteria for seronega-tivity either because a 12-month follow-up specimenwas lacking (21 patients) or because they were re-ceiving immunosuppressant medication at the timeof testing (seven patients). None had P/Q-type orN-type calcium channel antibodies. Serum was avail-able from eight patients for MuSK antibody testing(seven had both initial serum and follow-up serumtested but were receiving immunosuppressant med-ication); none were positive. One or more non-muscle autoantibodies were detected in 14 of the 23patients (61%): gastric parietal cell antibody (six);thyroid peroxidase antibody (six); thyroglobulin an-tibody (three); GAD65 antibody (three); antinuclearantibody (three); smooth muscle antibody (two) andantimitochondrial antibody (one).

DISCUSSION

Our study group of 562 patients with adult-acquiredgeneralized MG was ascertained exclusively on thebasis of clinical and electrophysiological abnormali-ties typical of acquired MG. Their mean onset age ofdisease (�52 years) and the slight male predomi-nance were likely due to exclusion of patients withonset age before 17 years (early-onset MG is charac-terized by female predominance and late-onset MGby mild male predominance).20 We estimated theseronegativity frequency to be less than 8.2%, pro-vided that tests for AChR binding and AChR modu-lating antibodies were performed before commence-ment of immunosuppressant therapy, and wererepeated at 12 months after symptom onset if ini-tially negative. The observed seroconversion rate of15.2% at 12 months is essentially identical to the15.8% seroconversion rate reported by Sanders etal.35 among 95 patients initially seronegative forAChR binding antibody when retested 6 or moremonths after disease onset. It is remarkable that inthat study one patient seroconverted 9 years afterclinical onset of MG.

Immunosuppressant therapy is underappreci-ated as a source of false seronegativity in patientswith MG. It is our experience that initiation of im-munosuppressant therapy weeks or months before

serological evaluation can cause apparent seronega-tivity.24 The Sanders et al.35 study reported that 9%of 143 initially seropositive MG patients became se-ronegative when retested in clinical remission follow-ing treatment. Seronegativity was documented asearly as 1 month after initiating corticosteroid ther-apy and 4 months after thymectomy. The interpre-tation of a patient’s serological and clinical statuscan be complicated further when characteristic find-ings of MG are obscured by a superimposed steroid-induced myopathy. This emphasizes the importanceof performing comprehensive serological evaluationbefore initiating immunosuppressant therapy.

The diagnosis of MG in a seronegative patientmust be made by meticulous clinical and electromyo-graphic criteria because other neuromuscular disor-ders can mimic acquired MG, both clinically and incertain electrophysiological characteristics, andsometimes in response to immunomodulatory ther-apies. Inadequate serological testing may miss thediagnosis of truly seropositive MG, and hence delayoptimal therapy, including consideration of earlythymectomy. A false diagnosis of “seronegative” au-toimmune MG may lead to unnecessary and poten-tially harmful immunosuppressant therapy. The 38%frequency of MuSK antibody in our patients whowere unambiguously seronegative for AChR antibod-ies, as well as their clinical characteristics, accordwith the frequency and phenotype of MuSK anti-body-positive MG reported by other investiga-tors.11,36,38

We emphasize that the diagnosis of acquired MGin all patients of this study was made on the basis ofcompatible clinical features and electrophysiologicalabnormalities. Detailed study by an experiencedEMG specialist is essential in the early stages of MGwhen electromyographic abnormalities can mimicearly Lambert–Eaton syndrome or a congenital my-asthenic syndrome.16 Serological testing facilitatesthe diagnosis of acquired MG. Although muscleAChR binding, AChR modulating, or striational an-tibodies are positive in 13% of patients with classicLambert–Eaton syndrome (with or without can-cer),28 patients with MG never have calcium channelantibodies (except in rare nonthymomatous para-neoplastic cases).13,25,41 Based on our laboratory’s 25years of clinical–serological correlative experience,we consider reproducible seropositivity for any skel-etal muscle antibody (AChR binding, AChR modu-lating, or striational) in the absence of a P/Q-type orN-type calcium channel antibody to be an appropri-ate serological criterion for confirming the diagnosisof acquired MG in a nonimmunosuppressed patientwho has unambiguous clinical and electrophysiolog-


ical findings supporting that diagnosis. However, ifstandard electrophysiological evaluation does notsupport a diagnosis of MG, other clinical associationsof muscle AChR binding, AChR modulating, andstriational antibodies must be considered. For exam-ple, seropositivity is encountered in �35% of pa-tients with autoimmune liver diseases, in �5% ofpatients with a variety of autoimmune neurologicaldisorders with or without cancer,25 and in an un-known percentage of patients with graft-versus-hostdisease.5

Our study has revealed that the actual frequencyof seronegativity (i.e., undetectable muscle AChR orMuSK antibodies) in adult-acquired generalized MGis 5%. Among this small group of “seronegative” MGpatients, we documented a high prevalence of mis-cellaneous autoantibodies, suggesting that those pa-tients also may prove to have an autoimmune basisfor their disease. It is conceivable that some seroneg-ative patients may have AChR or MuSK antibody ofhigh affinity but at a level too low to detect due toadsorption to autoantigen in vivo. Alternatively, ad-ditional rare motor endplate autoantibodies withpathogenic potential may remain to be discovered.Serological testing for other organ-specific and non-organ-specific autoantibodies is a valuable ancillaryinvestigation in evaluating seronegative acquiredgeneralized MG of adult onset.24,25 Positive resultsmay justify a trial of immunosuppressant therapy.

Supported by the M. Lee Pearce Foundation and the AdmadjajaThymoma Research Program. We thank Denice Bredlow, TomKryzer, Sue Nodtvedt, and the technical staff of the Mayo Clinic’sNeuroimmunology Laboratory for valuable assistance and Dr. K.Meng Tan for editorial suggestions.

REFERENCES

1. Almon RR, Appel SH. Interaction of myasthenic serum glob-ulin with the acetylcholine receptor. Biochim Biophys Acta1975;393:66–77.

2. Andrews PI, Massey JM, Howard JF Jr, Sanders DB. Race, sex,and puberty influence onset, severity, and outcome in juve-nile myasthenia gravis. Neurology 1994;44:1208–1214.

3. Beeson D, Amar M, Bermudez I, Vincent A, Newsom-Davis J.Stable functional expression of the adult subtype of humanmuscle acetylcholine receptor following transfection of thehuman rhabdomyosarcoma cell line TE671 with cDNA encod-ing the epsilon subunit. Neurosci Lett 1996;207:57–60.

4. Burges J, Vincent A, Molenaar PC, Newsom-Davis J, Peers C,Wray D. Passive transfer of seronegative myasthenia gravis tomice. Muscle Nerve 1994;17:1393–1400.

5. Cikes N, Momoi MY, Williams CL, Howard FM Jr, HoaglandHC, Whittingham S, et al. Striational autoantibodies: quanti-tative detection by enzyme immunoassay in myasthenia gravis,thymoma, and recipients of D-penicillamine or allogeneicbone marrow. Mayo Clin Proc 1988;63:478–481.

6. Compston DA, Vincent A, Newsom-Davis J, Batchelor JR.Clinical, pathological, HLA antigen and immunological evi-dence for disease heterogeneity in myasthenia gravis. Brain1980;103:579–601.

7. Drachman DB, Angus CW, Adams RN, Michelson JD, Hoff-man GJ. Myasthenic antibodies cross-link acetylcholine recep-tors to accelerate degradation. N Engl J Med 1978;298:1116–1122.

8. Dyck PJ, Kratz KM, Lehman KA, Karnes JL, Melton LJ 3rd,O’Brien PC, et al. The Rochester Diabetic Neuropathy Study:design, criteria for types of neuropathy, selection bias, andreproducibility of neuropathic tests. Neurology 1991;41:799–807.

9. Dyck PJ, Litchy WJ, Lehman KA, Hokanson JL, Low PA,O’Brien PC. Variables influencing neuropathic endpoints:the Rochester Diabetic Neuropathy Study of Healthy Subjects.Neurology 1995;45:1115–1121.

10. Engel AG, Ohno K, Sine SM. Congenital myasthenic syn-dromes. In: Engel A, editor. Myasthenia gravis and myas-thenic disorders. New York: Oxford University Press; 1999. p251–297.

11. Evoli A, Tonali PA, Padua L, Monaco ML, Scuderi F, BatocchiAP, et al. Clinical correlates with anti-MuSK antibodies ingeneralized seronegative myasthenia gravis. Brain 2003;126:2304–2311.

12. Glass DJ, Bowen DC, Stitt TN, Radziejewski C, Bruno J, RyanTE, et al. Agrin acts via a MuSK receptor complex. Cell1996;85:513–523.

13. Griesmann GE, Harper CM, Lennon VA. Paraneoplastic my-asthenia gravis and lung carcinoma: distinction from Lam-bert–Eaton syndrome and hypothesis of aberrant muscle ace-tylcholine receptor (AChR) expression. Muscle Nerve 1998;7:S122.

14. Griesmann GE, Kryzer TJ, Lennon VA. Autoantibody profilesof myasthenia gravis and Lambert–Eaton myasthenic syn-drome. In: Rose NR, Hamilton RG, Detrick B, editors. Manualof clinical and laboratory immunology, 6th ed. Washington,DC: ASM Press; 2002. p 1005–1012.

15. Grob D, Simpson D, Mitsumoto H, Hoch B, Mokhtarian F,Bender A, et al. Treatment of myasthenia gravis by immuno-adsorption of plasma. Neurology 1995;45:338–344.

16. Harper CM, Lennon VA. Lambert–Eaton Syndrome. In: Ka-minski HJ, editor. Current clinical neurology: myastheniagravis and related disorders. Totowa, NJ: Humana Press; 2007.

17. Hoch W, McConville J, Helms S, Newsom-Davis J, Melms A,Vincent A. Auto-antibodies to the receptor tyrosine kinaseMuSK in patients with myasthenia gravis without acetylcho-line receptor antibodies. Nat Med 2001;7:365–368.

18. Howard FM Jr, Lennon VA, Finley J, Matsumoto J, ElvebackLR. Clinical correlations of antibodies that bind, block, ormodulate human acetylcholine receptors in myasthenia gra-vis. Ann N Y Acad Sci 1987;505:526–538.

19. Jaretzki A 3rd, Barohn RJ, Ernstoff RM, Kaminski HJ, KeeseyJC, Penn AS, et al. Myasthenia gravis: recommendations forclinical research standards. Task Force of the Medical Scien-tific Advisory Board of the Myasthenia Gravis Foundation ofAmerica. Neurology 2000;55:16–23.

20. Keesey JC. Clinical evaluation and management of myasthe-nia gravis. Muscle Nerve 2004;29:484–505.

21. Lauriola L, Ranelletti F, Maggiano N, Guerriero M, Punzi C,Marsili F, et al. Thymus changes in anti-MuSK-positive and-negative myasthenia gravis. Neurology 2005;64:536–538.

22. Lavrnic D, Losen M, Vujic A, De Baets M, Hajdukovic LJ,Stojanovic V, et al. The features of myasthenia gravis withautoantibodies to MuSK. J Neurol Neurosurg Psychiatry 2005;76:1099–1102.

23. Lefvert AK, Bergstrom K, Matell G, Osterman PO, PirskanenR. Determination of acetylcholine receptor antibody in myas-thenia gravis: clinical usefulness and pathogenetic implica-tions. J Neurol Neurosurg Psychiatry 1978;41:394–403.

24. Lennon VA. Serological diagnosis of myasthenia gravis andthe Lambert–Eaton myasthenic syndrome. In: Lisak R, editor.Handbook of myasthenia gravis and myasthenic syndromes.New York: Marcel Dekker; 1994. p 149–164.


25. Lennon VA. Serological profile of myasthenia gravis and dis-tinction from the Lambert–Eaton myasthenic syndrome. Neu-rology 1997;48:S23–S27.

26. Lennon VA, Howard FM Jr. Serologic diagnosis of myastheniagravis. In: Nakamura R, editor. Clinical laboratory molecularanalyses. Orlando, FL: Grune & Straton; 1985. p 29–44.

27. Lennon VA, Kryzer TJ, Griesmann GE, O’Sulleabhain PE,Windebank AJ, Woppmann A, et al. Calcium-channel anti-bodies in Lambert–Eaton myasthenic syndrome and otherparaneoplastic syndromes. N Engl J Med 1995;332:1467–1474.

28. Lennon VA, Lambert EH, Whittingham S, Fairbanks V. Au-toimmunity in the Lambert–Eaton myasthenic syndrome.Muscle Nerve 1982;5:S21–25.

29. Limburg PC, The TH, Hummel-Tappel E, Oosterhuis HJ.Anti-acetylcholine receptor antibodies in myasthenia gravis.Part 1. Relation to clinical parameters in 250 patients. J Neu-rol Sci 1983;58:357–370.

30. Lindstrom JM, Seybold ME, Lennon VA, Whittingham S,Duane DD. Antibody to acetylcholine receptor in myastheniagravis: prevalence, clinical correlates, and diagnostic value.1975. Neurology 1998;51:933–939.

31. McConville J, Farrugia ME, Beeson D, Kishore U, Metcalfe R,Newsom-Davis J, et al. Detection and characterization ofMuSK antibodies in seronegative myasthenia gravis. Ann Neu-rol 2004;55:580–584.

32. McEvoy KM, Windebank AJ, Daube JR, Low PA. 3,4-Diamin-opyridine in the treatment of Lambert–Eaton myasthenicsyndrome. N Engl J Med 1989;321:1567–1571.

33. Miller RG, Milner-Brown HS, Dau PC. Antibody-negative ac-quired myasthenia gravis: successful therapy with plasma ex-change. Muscle Nerve 1981;4:255.

34. Mossman S, Vincent A, Newsom-Davis J. Myasthenia graviswithout acetylcholine-receptor antibody: a distinct disease en-tity. Lancet 1986;1:116–119.

35. Sanders DB, Andrews PI, Howard JF, et al. Seronegative my-asthenia gravis. Neurology 1997;48:S40–S45.

36. Sanders DB, El-Salem K, Massey JM, McConville J, Vincent A.Clinical aspects of MuSK antibody positive seronegative MG.Neurology 2003;60:1978–1980.

37. Sanders DB, Howard JF, Massey JM, Mihovilovic M, OlanowCW, Roses AD, et al. Seronegative myasthenia gravis. AnnNeurol 1987;22(Suppl):P27.

38. Scuderi F, Marino M, Colonna L, Mannella F, Evoli A, Prov-enzano C, et al. Anti-p110 autoantibodies identify a subtype of“seronegative” myasthenia gravis with prominent oculobulbarinvolvement. Lab Invest 2002;82:1139–1146.

39. Soliven BC, Lange DJ, Penn AS, Younger D, Jaretzki A 3rd,Lovelace RE, et al. Seronegative myasthenia gravis. Neurology1988;38:514–517.

40. Tindall RS. Humoral immunity in myasthenia gravis: bio-chemical characterization of acquired antireceptor antibodiesand clinical correlations. Ann Neurol 1981;10:437–447.

41. Vernino S, Lennon VA. Autoantibody profiles and neurolog-ical correlations of thymoma. Clin Can Res 2004;10:7270–7275.

42. Vincent A, McConville J, Farrugia ME, Newsom-Davis J. Sero-negative myasthenia gravis. Semin Neurol 2004;24:125–133.


ABSTRACT: Surgical treatment of lateral femoral cutaneous neuropathy(LFCN) is performed only after failure of conservative management. Wereexamined 167 cases (7 bilateral) of LFCN of various etiologies (idiopathic,abdominal surgery, iliac crest bone grafting, trauma, and total hip arthro-plasty) operated on between 1987 and 2003. Average follow-up was 98months (20–212). The intervention was performed under local anesthesia in139 cases (83%). Surgical release of the nerve was performed in 153 cases(92%) and transection in 14 cases (8%). Surgical treatment of LFCN led toimprovement and patient satisfaction in 130 cases (78%). The results de-pended on several factors, especially the underlying etiology, duration ofsymptoms before intervention, and integrity of the nerve. Nerve releaseremains the first-line surgical technique, improving painful symptoms inmany cases while preserving sensation of the thigh. It can be performedunder local anesthesia by an experienced surgeon.

Muscle Nerve 36: 659–663, 2007

LATERAL FEMORAL CUTANEOUS NEUROPATHYAND ITS SURGICAL TREATMENT:A REPORT OF 167 CASES

IGOR BENEZIS, MD, BENOIT BOUTAUD, MD, JEROME LECLERC, MD,

THIERRY FABRE, MD, and ALAIN DURANDEAU, MD

CHU Pellegrin, Universite Victor Segalen, Bordeaux, France


Meralgia paresthetica is a lateral femoral cutaneousneuropathy (LFCN). This neuropathy may be idio-pathic, traumatic, compressive (lumbar disk hernia,pelvic tumor), metabolic, or immunological in etiol-ogy. The lateral femoral cutaneous nerve is a sensorynerve that arises from the second and third lumbarroots, follows an intrapelvic course on the anteroin-ternal face of the iliac muscle, and then becomesvertical by circumventing the anterosuperior iliacspine to pass under the inguinal ligament. It crossesthe fascia lata and divides into two or three terminalbranches. The nerve may be injured at any level, butthe most frequent site of compression is where itemerges from the pelvis, because at this level it iswedged into an inextensible narrow space under theinguinal ligament and inside the anterosuperior iliacspine. Bernhardt7 provided the first description of

this syndrome in 1878. Conservative management ofLFCN leads to satisfactory improvement in morethan 90% of cases,13,36 and involves analgesics, non-steroidal anti-inflammatory drugs, gabapentin, andlocal injection of anesthetics and anti-inflammatorysteroids that reduce or eliminate the aggravatingfactors. Surgical treatment is performed only afterfailure of conservative approaches. This study pre-sents our experience with surgical treatment ofLFCN in 167 cases operated on between 1987 and2003.


This retrospective study included 187 patients (80women and 107 men) operated on between 1987and 2003 by the same surgeon. Of these patients, 160(69 women and 91 men) were reexamined during2005 by an independent specialist. Seven patientswere operated on bilaterally (making a total of 167cases). The mean follow-up was 98 months (20–212months). Mean age of the patients was 52 years(17–80 years). Mean body mass index (BMI) was 27(17–46; normal, lower than 25) and 50 patients(31%) were considered obese (BMI �30). All pa-tients were operated on after a failure of conservativetreatment, except 16 who were treated for recur-rence after previous surgery.

Abbreviations: BMI, body mass index; EMG, electromyogram; LFCN, lateralfemoral cutaneous neuropathy; SNAP, sensitive nerve action potential;SNCV, sensitive nerve conduction velocityKey words: lateral femoral cutaneous nerve; meralgia paresthetica; nerveentrapment; nerve release; nerve transectionCorrespondence to: I. Benezis, CHU Pellegrin, Service du Professeur Du-randeau, Place Amelie Raba Leon 33 000, Bordeaux, France; e-mail:[email protected]


Lateral Femoral Cutaneous Neuropathy MUSCLE & NERVE November 2007 659

Twelve cases of LFCN (7%) were secondary totrauma and 22 (13%) were secondary to abdominalsurgery (3 parietal rupture repair, 11 inguinal her-niorraphy, and 8 gynecological or urological surgeryby Pfannenstiel incision). Twenty cases (12%) weresecondary to an iliac bone graft harvest and 8 casesto a total hip arthroplasty on the homolateral side.Finally, 105 cases (62%) were considered to be idio-pathic. All had electrodiagnostic studies that con-firmed the diagnosis in 156 cases (93%), showingattenuated sensory nerve action potential (SNAP)amplitudes or reduced sensory nerve conductionvelocity (SNCV).24,31 In the 11 other cases, it did notcontribute to the diagnosis although the patientshad characteristic clinical signs. In 45 cases (27%),spinal imaging studies had been performed becausethe initial diagnosis of the referring physician was oflumbar disease, but these were unhelpful.

The mean interval between the onset of clinicaldisturbances and intervention was 46 months (1–311months). The intervention was performed underlocal anesthesia in 139 cases (83%). General anes-thesia was necessary in 28 cases, including somecases presumed difficult because of previous abdom-inal surgery or harvesting of the iliac bone graft, andwas also required in some anxious or particularlyobese patients.

The procedure was performed on an outpatientbasis. It consisted in injection of about 15 ml of 1%lidocaine two fingerwidths medial to and under theanterosuperior iliac spine, corresponding to the lo-cation of Tinel’s sign. After a 4-cm horizontal cuta-neous section, the fat tissue was dissected to reach

the fascia lata, which at this level is the continuationof the iliac fascia. Additional local anesthesia underthe fascia was given. Then the nerve was carefullydetached from the sartorius. Once located, it wasdecompressed downwards by opening the fascia lataand upwards by sectioning the inguinal ligament.The nerve was then released in its pelvic course. Thesubcutaneous tissue was sutured by resorbablethreads and the skin was closed by interrupted su-tures without drainage. A 2-kg sand bag was placedon the scar for 6 hours.

In 29 cases (17%), the nerve was in an abnormalanatomical position (Fig. 1). In 7 cases (4%), thenerve crossed the iliac crest behind the anterosupe-rior iliac spine. In 9 cases (5%), the nerve passedthrough a split in the inguinal ligament on its inser-tion into the anterosuperior iliac spine. In 5 cases(3%), the nerve was located medially under the in-guinal ligament near the genitofemoral nerve. In 8cases (5%), the nerve had already divided into two orthree branches during its passage under the inguinalligament. During surgery, we found 69 cases (41%)with macroscopic changes of the nerve, including 48cases of narrowing, 11 of apparent enlargement, and10 neuromas.

The nerve was released 153 times (92%) andtransection was performed in 14 cases (3 cases afterfailure of release, 10 because the nerve presented aneuroma, and 1 because it presented considerablenarrowing). We operated on 16 cases of failed pre-vious nerve decompression (9 idiopathic, 4 iliacbone graft harvest, 2 post-traumatic, and 1 post-inguinal herniorraphy). Nerve release was per-

a b c d

FIGURE 1. Anatomical variations in the inguinal portion of the lateral cutaneous nerve of the thigh in 167 cases. (a) The nerve crossesthe iliac crest behind the anterosuperior iliac spine (4%). (b) The nerve passes through a split in the inguinal ligament (5%). (c) The nerveis very medial, near the genitofemoral nerve (3%). (d) The nerve divides into two or three branches during its passage under the inguinalligament (5%).

660 Lateral Femoral Cutaneous Neuropathy MUSCLE & NERVE November 2007

formed again in 11 cases and the nerve was sectionedin 5 cases.

For follow-up purposes, we drew up a simple andreproducible questionnaire for all the patients. Weasked the patients whether they felt better or not andwhether they were satisfied with the outcome of theirsurgery. Results were analyzed according to variousfactors, especially the etiology of LFCN, the durationof symptoms, the incidence of excess weight, and theelectrodiagnostic data. Statistical assessment wasdone using the Mantel–Haenszel chi-square test.

RESULTS

Of our patients, 102 (61%) recovered fully and werecompletely satisfied, and 28 (17%) experienced par-tial improvement and were also satisfied. Thirty-seven (22%) had no improvement and were some-what (12%) or totally dissatisfied (10%). No patientexperienced worsening of symptoms. Results werevariable within etiologic groups (Table 1) and ac-cording to the duration of symptoms (Table 2).

The average period until pain regressed was 53days (range, 1–365 days). Improvement in cutaneoussensitivity was obtained on average after 113 days(1–365 days). Twenty-two patients (13%) improvedbut their dysesthesias persisted, and 3 were dissatis-fied with the outcome.

Nerve transection was performed in 14 cases(8%). Five of these were improved or the patient wassatisfied and experienced no subsequent pain. Ninepatients were dissatisfied and experienced no or only

partial relief of symptoms, with persisting dysesthe-sias and pain. Among the 16 cases of re-operation(11 nerve releases and 5 transections), only 8 wereimproved (4 total recovery and 4 partial). Dissatis-faction was total in 5 patients in whom transectionwas performed after failure of a previous decompres-sion.

There were 12 postoperative complications(7%): 8 hematomas requiring drainage and 4 woundruptures requiring re-intervention. Return to workwas possible within an average of 24 days followingsurgery (1–180 days), but 14 patients (8.4%) werenot able to work again because of persisting anddisabling pain.

DISCUSSION

Nerve release gave good results, with 78% of ourpatients indicating satisfaction. The outcome de-pended on several factors, especially the etiology andduration of symptoms.

Idiopathic meralgia paresthetica is comparableto an entrapment syndrome of the lateral femoralcutaneous nerve.3,14,26 The nerve is compressed in anarrow fibrous space limited by the anterosuperioriliac spine externally, the inguinal ligament above,and the tendon of the sartorius muscle below. Open-ing this channel relieves the pain due to nerve com-pression.15,21,26 Surgical repair of an inguinal herniamay increase the tension and even modify the direc-tion of the inguinal ligament, leading to compres-sion or stretching of the nerve. In five instances thenerve was found to be wedged under an extraperi-toneal prosthetic mesh, and once under a fixingstaple. Nerve release decreased parietal abdominaltension and provided total recovery in 77% of pa-tients having undergone abdominal surgery. Regard-ing the Pfannenstiel incision, we believe that theperioperative injury was due to the installation ofside valves that retracted the rectus abdominis mus-cles and likely compressed the nerve against the iliaccrest.25,29,32

Table 1. Results of surgical treatment in 167 cases of LFCN.

Cause Number of cases Total recovery

Improvement

Partial None

Idiopathic 105 (63%) 67 (64%) 21 (20%) 17 (16%)Abdominal surgery 22 (13%) 17 (77%) 2 (9%) 3 (14%)Iliac bone graft 20 (12%) 8 (40%) 3 (15%) 9 (45%)Trauma 12 (7%) 5 (42%) 2 (16%) 5 (42%)Total hip arthroplasty 8 (5%) 5 (62.5%) 0 3 (37.5%)Total 167 (100%) 102 (61%) 28 (17%) 37 (22%)

Table 2. Results of surgical treatment in 167 cases of LFCN.

Duration ofsymptoms

Number ofcases

Totalrecovery

Improvement

Partial None

�6 months 26 22 (85%) 1 (4%) 3 (12%)6–12 months 20 13 (65%) 2 (10%) 5 (25%)�12 months 121 67 (55%) 25 (21%) 29 (24%)Total 167 102 (61%) 28 (17%) 37 (22%)


Transient LFCN after iliac crest bone graft har-vest occurs in approximately 30% of patients,27 butthe pain persists after 3 months in only 1%.4,30 Thenerve lesion is a direct injury due to excessive sam-pling or harvesting too near the superoanterior iliacspine. Concerning nerve injury after total hip arthro-plasty, the main mechanism in our 8 patients wasprobably incorrectly positioned anterior paddedsupports on the anterosuperior iliac spine, as re-ported by others.22,27 Another reported mechanismis limb length discrepancy,19 but we did not identifythis as a cause in our patients.

Fifty cases (30%) were of iatrogenic origin. Pre-vention of iatrogenic lesions requires special atten-tion when positioning the patient for surgery andgood knowledge of the nerve course and its anatom-ical variations.1,9,28 Trauma by compression againstthe iliac crest is a well-known cause of LFCN.6,27,34

The anterosuperior iliac spine constitutes a solidblock on which the nerve may be compressed. Thenature of the trauma is variable but generally resultsfrom a high-energy impact. However, it can also bedue to repeated microtrauma (e.g., from a belt orbrace worn too tightly).8,17,23

The results of surgical treatment of LFCN wereworse (P � 0.04) when the nerve had suffered adirect injury. Forty cases were secondary to iliac boneharvesting, trauma, and total hip arthroplasty. Only18 of these patients (45%) fully recovered. Laguenyet al.24 indicated that trauma and repeated trauma ofthe lateral femoral cutaneous nerve may be respon-sible for involvement of the unmyelinated fibers orfor stimulation of their activity, but also suspectedinterference with the process of regeneration.

Sectioning the nerve is a significant cause of poorresults. Symptoms improved in 78% of cases withsurgical release of the nerve and only in 35% withtransection (P � 0.004), confirming our preferencefor surgical release. Nevertheless, there is no consen-sus on this issue. Benini5 obtained better results withrelease in 36 cases, whereas Antoniadis et al.,2 in astudy of 29 cases, reported improvement in 82% ofcases with transection and 72% of cases with nerverelease. van Eerten et al.35 found similar results in 21patients. These series were quite small, however, anddo not demonstrate with certainty the superiority oftransection. Whatever the technique used, failuresremain frequent and often unexplained. The diag-nosis must therefore be made again by repeat clini-cal examination and complementary tests in order toseek another cause of symptoms.10,16,20,33

Obesity is frequently a causative factor ofLFCN.12,13,18,31 The weight of the abdominal belttends to lower the inguinal ligament, thereby nar-

rowing the nerve passage. In our series, obesity prob-ably had an etiological role because its frequency(30%) was higher than in the general population(11% in France,11 P � 0.0009). However, obesitydoes not seem to influence the results of surgicaltreatment. Postoperative complications or poor re-sults were not more numerous among obese pa-tients; the surgical procedure was simply longer andmore difficult to perform. Revision surgery (16cases) was difficult to perform and often gave poorresults, although 8 cases were improved after a newnerve release. In the light of this series, our ap-proach after failure of a correctly performed nervedecompression is now surgical abstention.

Analysis of the electrophysiological data did notallow any factors predictive of prognosis or surgicalresponse to be established. SNAP amplitude indi-cates neither the severity of symptoms nor the qualityof recovery.24 However, electromyography (EMG) isuseful for diagnosing LFCN, and its specificity isgreater than 98% in the hands of experienced neu-rophysiologists.31 In our series, EMG was performedby 27 different neurophysiologists, which likely ac-counts for our low specificity (93%) and the diffi-culty of analyzing precisely our electrophysiologicaldata.

The duration of symptoms influences the prog-nosis. Ecker and Woltman13 observed that the poten-tial for cure was decreased by 50% if symptoms hadlasted more than 2 years. We also observed a signif-icant decrease in total recovery in relation to theduration of symptoms before surgical intervention.Total recovery was 55% when patients were operated1 year after the onset of pain, 65% when operatedwithin 6–12 months of onset, and 85% within 6months (P � 0.0073).

Surgical treatment should be temporarily with-held depending on the etiology of the LFCN. Inidiopathic cases, a delay of 6 months allows for ap-propriate conservative management. For post-trau-matic or post-surgical etiologies, this interval can beshortened as determined on an individual basis. Inour opinion, surgical release of the nerve remainsthe operative technique of choice because it im-proved painful symptoms in 78% of cases while pre-serving sensation of the thigh.

REFERENCES

1. Andrew DR, Gregory RP, Richardson DR. Meralgia paraes-thetica following laparoscopic inguinal herniorrhaphy. Br JSurg 1994;81:715.

2. Antoniadis G, Braun V, Rath S, Moese G, Richter HP. Meral-gia paraesthetica and its surgical treatment. Nervenarzt 1995;66:614–617.

662 Lateral Femoral Cutaneous Neuropathy MUSCLE & NERVE November 2007

3. Aszmann OC, Dellon ES, Dellon AL. Anatomical course of thelateral femoral cutaneous nerve and its susceptibility to com-pression and injury. Plast Reconstr Surg 1997;100:600–604.

4. Banwart JC, Asher MA, Hassanein RS. Iliac crest bone graftharvest donor site morbidity. A statistical evaluation. Spine1995;20:1055–1060.

5. Benini A. Meralgia paresthetica. Pathogenesis, clinical aspectsand therapy of compression of the lateral cutaneous nerve ofthe thigh. Schweiz Rundsch Med Prax 1992;81:215–221.

6. Beresford HR. Meralgia paresthetica after seat-belt trauma.J Trauma 1971;11:629–630.

7. Bernhardt M. Ueber isolirt im Gebiete des N. cutaneus fem-oris externus vorkommende Parasthesien. NeurologischesCentralblatt 1895;14:242–244.

8. Boyce JR. Meralgia paresthetica and tight trousers. JAMA1984;251:1553.

9. Broin EO, Horner C, Mealy K, Kerin MJ, Gillen P, O’Brien M,et al. Meralgia paraesthetica following laparoscopic inguinalhernia repair. An anatomical analysis. Surg Endosc 1995;9:76–78.

10. Cedoz ME, Larbre JP, Lequin C, Fischer G, Llorca G. Upperlumbar disk herniations. Rev Rhum Engl Ed 1996;63:421–426.

11. Charles MA, Basdevant A, Eschwege E. Prevalence of obesityin adults in France: the situation in 2000 established from theOBEPI Study. Ann Endocrinol (Paris) 2002;63:154–158.

12. Deal CL, Canoso JJ. Meralgia paresthetica and large abdo-mens. Ann Intern Med 1982;96:787–788.

13. Ecker AD, Woltman HW. Meralgia paresthetica: a report ofone hundred and fifty cases. JAMA 1938;110:1650–1652.

14. Edelson JG, Nathan H. Meralgia paresthetica. An anatomicalinterpretation. Clin Orthop Relat Res 1977;122:255–262.

15. Edelson R, Stevens P. Meralgia paresthetica in children.J Bone Joint Surg Am 1994;76:993–999.

16. Flowers RS. Meralgia paresthetica. A clue to retroperitonealmalignant tumor. Am J Surg 1968;116:89–92.

17. Garcia-Albea E, Palomo F, Tejeiro J, Cabrera F. Occupationalmeralgia paraesthetica. J Neurol Neurosurg Psychiatry 1990;53:708.

18. Ghent WR. Further studies on meralgia paresthetica. CanMed Assoc J 1961;85:871–875.

19. Goel A. Meralgia paresthetica secondary to limb length dis-crepancy: case report. Arch Phys Med Rehabil 1999;80:348–349.

20. Jiang GX, Xu WD, Wang AH. Spinal stenosis with meralgiaparaesthetica. J Bone Joint Surg Br 1988;70:272–273.

21. Keegan JJ, Holyoke EA. Meralgia paresthetica. An anatomicaland surgical study. J Neurosurg 1962;19:341–345.

22. Kitson J, Ashworth MJ. Meralgia paraesthetica. A complica-tion of a patient-positioning device in total hip replacement.J Bone Joint Surg Br 2002;84:589–590.

23. Kotler D. Meralgia paresthetica. Case report in policewoman.JAAPA 2000;13:39–42, 47.

24. Lagueny A, Deliac MM, Deliac P, Durandeau A. Diagnosticand prognostic value of electrophysiologic tests in meralgiaparesthetica. Muscle Nerve 1991;14:51–56.

25. Luijendijk RW, Jeekel J, Storm RK, Schutte PJ, Hop WC,Drogendijk AC, et al. The low transverse Pfannenstiel incisionand the prevalence of incisional hernia and nerve entrap-ment. Ann Surg 1997;225:365–369.

26. Macnicol MF, Thompson WJ. Idiopathic meralgia pares-thetica. Clin Orthop Relat Res 1990:270–274.

27. Mirovsky Y, Neuwirth M. Injuries to the lateral femoral cuta-neous nerve during spine surgery. Spine 2000;25:1266–1269.

28. Murata Y, Takahashi K, Yamagata M, Shimada Y, Moriya H.The anatomy of the lateral femoral cutaneous nerve, withspecial reference to the harvesting of iliac bone graft. J BoneJoint Surg Am 2000;82:746–747.

29. Raiga J, Barakat P, Diemunsch P, Maillot C, Treisser A, BrettesJP. Femoral neuropathy after transversal suspubic laparoto-mies. Etiopathological explanation on the basis of an anatom-ical study. J Gynecol Obstet Biol Reprod (Paris) 2002;31:183–186.

30. Schnee CL, Freese A, Weil RJ, Marcotte PJ. Analysis of harvestmorbidity and radiographic outcome using autograft for an-terior cervical fusion. Spine 1997;22:2222–2227.

31. Seror P, Seror R. Meralgia paresthetica: clinical and electro-physiological diagnosis in 120 cases. Muscle Nerve 2006;33:650–654.

32. Sippo WC, Burghardt A, Gomez AC. Nerve entrapment afterPfannenstiel incision. Am J Obstet Gynecol 1987;157:420–421.

33. Suber DA, Massey EW. Pelvic mass presenting as meralgiaparesthetica. Obstet Gynecol 1979;53:257–258.

34. Thanikachalam M, Petros JG, O’Donnell S. Avulsion fractureof the anterior superior iliac spine presenting as acute-onsetmeralgia paresthetica. Ann Emerg Med 1995;26:515–517.

35. van Eerten PV, Polder TW, Broere CA. Operative treatment ofmeralgia paresthetica: transection versus neurolysis. Neuro-surgery 1995;37:63–65.

36. Williams PH, Trzil KP. Management of meralgia paresthetica.J Neurosurg 1991;74:76–80.


ABSTRACT: This study was undertaken to evaluate collateral sproutingcapability in an end-to-side repair model with long regenerative distance.Forty-five rats were used and divided into four groups, according to thereparative procedure following peroneal nerve division: (A) “double” end-to-side neurorrhaphy with a regenerative distance of 0.6 cm; (B) “double”end-to-side neurorrhaphy with a regenerative distance of 1.2 cm; (C) end-to-end neurorrhaphy; and (D) nerve stumps buried into neighboring mus-cles. In all animals the contralateral healthy side served as a control.Functional assessment of nerve regeneration was performed at intervals upto 5 months using the Peroneal Function Index (PFI). Evaluation 150 daysafter surgery included peroneal and tibial nerve histologic and morphometricexamination and wet weights of the tibialis anterior muscle. Functionalevaluation and axonal counting data demonstrated that there was no sta-tistically significant difference between groups A and B, or between groupsA and C. There was no functional or histologic evidence of donor nervedeterioration. In conclusion, the present study confirms that “double” end-to-side neurorrhaphy may be useful for the repair of divided human nerveswith long gaps.

Muscle Nerve 36: 664–671, 2007

CAN END-TO-SIDE NEURORRHAPHY BRIDGE LARGEDEFECTS? AN EXPERIMENTAL STUDY IN RATS

MARIOS G. LYKISSAS, MD,1 ANASTASIOS V. KOROMPILIAS, MD,1 ANNA K. BATISTATOU, MD,2

GREGORY I. MITSIONIS, MD,1 and ALEXANDROS E. BERIS, MD1

1 Department of Orthopaedic Surgery, University of Ioannina, School of Medicine,Ioannina, P.C. 45110, Greece

2 Department of Pathology, University of Ioannina, School of Medicine, Ioannina, Greece


Development in microsurgical techniques and in-strumentation in combination with better under-standing of the physiology of nerve injury and regen-eration has led to the evolution of repair methodsfor peripheral nerve injuries. However, direct recon-struction of large nerve defects is not always possible.Autologous nerve grafts remain the gold standardfor treating large peripheral nerve defects. Use ofautologous nerve grafts is bounded by the limitedamount of available tissue and the increased donorsite morbidity.16,30 The above-mentioned limitationsand the poor clinical results necessitate a further

search for alternative methods of nerve reconstruc-tion.

Among others, a large number of conduits, withor without addition of peptides, Schwann cells,growth factors, or cytokines, have been used to in-crease the maximum nerve gap distance successfullybridged.10,13,15,20 Unfortunately, nerve regenerationis not possible in acellular conduits with lengthsexceeding 40 mm,9,30,31 due to the limited migrationcapacity of Schwann cells. Consequently, the lengthof the natural or synthetic materials that are used tobridge peripheral nerve defects is also limited.

In the last 15 years, end-to-side neurorrhaphy hasbeen added to the surgical options, although thistechnique was described at the beginning of the lastcentury. End-to-side neurorrhaphy was first de-scribed as an alternative approach for the treatmentof facial palsy,5 but the approach was neglected untilViterbo et al.,23–26 beginning in 1992, reintroducedend-to-side technique in studies in animals and af-terwards in patients with facial nerve palsy. Sincethen, several other investigators have carried outexperimental and clinical studies on end-to-side neu-rorrhaphy.1,3,12,18,19,21,27,28,32,33 This study was con-

Abbreviations: EIT, experimental intermediary toe spread; EPL, experimen-tal print length; ETS, experimental toe spread; IT, intermediary toe spread;NIT, normal intermediary toe spread; NPL, normal print length; NTS, normaltoe spread; PFI, peroneal function index; PL, print length; rHuEpo, recombi-nant human erythropoietin; SFI, sciatic function index; TFI, tibial function in-dex; TS, toe spreadKey words: collateral sprouting; end-to-side neurorrhaphy; nerve injury;nerve repair; regenerative capacityCorrespondence to: M. G. Lykissas; e-mail: [email protected]


664 End-to-Side Neurorrhaphy MUSCLE & NERVE November 2007

ducted to determine collateral sprouting capability inan end-to-side repair model with long regenerativedistance between two end-to-side neurorrhaphies.


The procedures for this investigation were per-formed according to protocols approved nationally.The experiment was carried out on 45 male Wistarrats weighing 190–330 g. Before and after the oper-ation the animals were kept in individual cages andmaintained on standard rat chow and water ad libi-tum, with a 12-h light-dark cycle. The animals weredivided into four groups according to the operativeprocedure: end-to-side repair with 0.6-cm long tibialnerve bridge (group A; n � 12); end-to-side repairwith 1.2-cm long tibial nerve bridge (group B; n �12); end-to-end nerve repair (group C; n � 12); andsegmentary resection (group D; n � 9).

Surgical Procedure. Just before the operation allanimals were given conditioning trials on an 8 � 62cm walking track for comparison with normativedata.7 The animals were then weighed and operatedon under anesthesia with a single 5 mg/kg intraperi-toneal ketamine injection (Narketan 10, 100 mg/ml;Vetoquinolag, Belp-Bern, Switzerland). In all ani-mals the right hindlimb was used as the experimen-tal limb and the left hindlimb as the control limb.

The lateral aspect of the right thigh and hip wereshaved and washed with antiseptic solution. Theright sciatic, tibial, and peroneal nerves were ex-posed and dissected through a semitendinosus–biceps femoris muscle-splitting incision under mag-nification with a Zeiss operating microscope (modelOP16-S; Zeiss, Munich, Germany). The commonperoneal nerve was transversely divided 2 mm dis-tally to its emergence from the sciatic nerve trunk.The distal and proximal stumps of the commonperoneal nerve were then immediately repaired ac-cording to grouping.

In group A the proximal segment of the pero-neal nerve was sutured end-to-side to the trunk of

the intact tibial nerve and the distal segment wasrepaired by the same method 0.6 cm distal to thefirst end-to-side neurorrhaphy (Fig. 1A). In groupB the proximal segment of the peroneal nerve wassutured end-to-side to the trunk of the intact tibialnerve, as in group A, but the distal segment wassutured end-to-side to the tibial nerve 1.2 cm distalto the first end-to-side neurorrhaphy (Fig. 1B). Inboth groups A and B an epineural window wasmade on the tibial trunk before each end-to-sidenerve coaptation took place. In group C the per-oneal nerve was repaired in a classic end-to-endfashion. In groups A, B, and C the neurorrhaphieswere performed under microscope magnificationusing three epineural 10/0 nylon interrupted su-tures with sharp cutting head, placed at 120° in-tervals. In group D (control) both proximal anddistal nerve stumps were buried in the neighbor-ing muscles and fixed with a single 8/0 nylonsuture. This was done to prevent regeneration ofthe proximal stump from interference with thefinal results. In all groups, wound closure wasachieved by 4/0 vicryl interrupted sutures for themuscles and 3/0 nylon interrupted sutures for theskin, and the animals were kept under the super-vision of a veterinarian.

Walking Track Analysis. Walking track analysis wasfirst proposed by de Medinaceli et al.8 as an index forthe functional condition of rat sciatic nerve. Bain etal.4 described three indexes for the evaluation ofcomplete sciatic (SFI; Sciatic Function Index), per-oneal (PFI; Peroneal Function Index), and tibialnerve (TFI; Tibial Function Index) lesions in the rat.These indexes have a statistical basis and have be-come reliable parameters for the functional evalua-tion of peripheral nerve regeneration.

At 15, 30, 60, 90, and 150 days after the operationall animals were given conditioning trials on an 8 �62 cm walking track darkened at one end.8 Thepaper strips containing the footprints were copiedwith a high-resolution scanner. Afterwards, the digi-

FIGURE 1. Illustration of surgical procedures.

End-to-Side Neurorrhaphy MUSCLE & NERVE November 2007 665

tized footprints were analyzed in a computer using agraphic software program, which automatically cal-culates the PFI and TFI. The parameters measuredin the footprints were print length (PL), toe spread(TS), and intermediary toe spread (IT) both in thenormal (NPL, NTS, and NIT, respectively) and ex-perimental paw prints (EPL, ETS, and EIT, respec-tively). At least three prints of each paw were mea-sured for each animal. All measurements were madeby two observers blinded to the identity of the digi-tized footprints at the end of the study according tothe formula of Bain et al.4:

PFI � 174.9�EPL � NPL�

NPL

� 80.3�ETS � NTS�

NTS� 13.4

TFI � � 37.2�EPL � NPL�

NPL� 104.4

�ETS � NTS�

NTS

� 45.6�EIT � NIT�

NIT� 8.8

Morphometric Studies. On the 150th postoperativeday, all animals were anesthetized and prepared.The experimental as well as unoperated controlsides were exposed through a posterolateral ap-proach to the thigh and the nerves were dissectedand removed. The tibialis anterior muscle was har-vested from the experimental and the control sideand the animals were killed with a lethal dose ofintraperitoneal sodium pentobarbital (100 mg/kg).

In groups A and B, sections 0.5 cm in length wereremoved from: (1) the tibial nerve midway betweenthe proximal and the distal end-to-side nerve coap-tation, for evaluation of axons in the outerepineurium of the donor nerve; (2) the distal pero-neal nerve segment with the site of distal end-to-siderepair; and (3) the tibial nerve distal to end-to-sideneurorrhaphies, to evaluate any potential damage tothe donor nerve resulting from the procedure. Inanimals in group C, a section (0.5 cm long) wasremoved distal to the site of end-to-end nerve repair.In D, a segment (0.5-cm long) of the distal stump ofthe peroneal nerve was harvested. In all animalscorresponding sections from the contralateral con-trol sides were harvested.

All sections were immersed in a solution of 2.5%glutaraldehyde buffered with cacodylate, washed insodium cacodylated buffer (0.2 mol/L, pH 7.4),postfixed in 1% osmium tetroxide, dehydrated in anethyl alcohol solution of increasing concentrations

(50%, 70%, 95%, and 100%), and embedded inepoxy resin for axonal counting. Cross-sections 1-�mthick were cut using a Reichert-Jung ultracut E witha diamond knife (Reichert-Jung, Weiss, Austria)from the nerve segments, stained with toluidineblue, and examined with a light microscope (Ax-ioscop; Zeiss) connected to a computer using thepublic domain software NIH Image (http://rsb.info.nih.gov/nih-image/).

Initially, the entire fascicle image was capturedand the total fascicle area was measured. Quantita-tive evaluation followed, with myelinated nerve fi-bers counted in all nerve segments. The final resultswere expressed as the ratio of the density of thefibers in the experimental side to the contralateralunoperated control side. Quantitative evaluationalso included the calculation of percent neural tissue(100 � neural area/intrafascicular area), expressedas the ratio of the experimental to control sides. Allmeasurements were done by a single blinded inves-tigator.

Muscle Weights. On the 150th postoperative day,tibialis anterior was harvested with its tendon fromthe experimental and control sides. Wet muscleweights were measured and reported as the ratio ofthe experimental to the contralateral control side.Afterwards, specimens from each muscle were im-mersed in 10% formalin and embedded in paraffinfor histologic analysis.

Statistical Analysis. Peroneal function index, mus-cle mass ratios, and nerve fiber density values wereanalyzed using the nonparametric Kruskal–Wallis,and Mann–Whitney U-tests. All tests were calculatedwith use of the SPSS statistic package for personalcomputers, v. 13.0 (SPSS, Chicago, Illinois). In allinstances P � 0.05 was regarded as statistically signif-icant.

RESULTS

Walking Track Analysis. The preoperative PFI wasnever 0, but oscillated between –22.15 and –7.57.Before the surgical procedure no significant differ-ence was noted between the four groups. On the15th day after surgery, PFI values were not signifi-cantly different between group A and group B, but adifference was detected between groups A and B andgroup C (P � 0.001), with group C revealing betterresults (Table 1). During the second postoperativeevaluation on the 30th day, no significant differencewas recorded between groups A and B, or betweengroups A and C. Group B and group C were different


(P � 0.001) and remained so on day 60. At day 90 nosignificant difference was detected between groupsA, B, and C. At the end of the study, on day 150,differences between group A and group B, as well asbetween group A and group C, remained insignifi-cant. However, a significant difference was detectedbetween group B and group C (P � 0.037). More-over, differences between experimental groups A, B,and C and control group D were significant (P �0.001).

At the end of the study the mean TFI was �14.88and �11.51 in groups A and B, respectively. Thisdifference was not statistically significant (P � 0.77).TFI for groups C and D were not calculated, sincethe tibial nerve remained intact in these groups.

Morphometric Studies. At the end of the study theratios of peroneal nerve fiber density (axons/mm2)for groups A, B, and C were estimated (Table 2).Severe axonal degeneration was present in group D.Morphometric evaluation was not performed in thisgroup. The results demonstrated that there was nostatistically significant difference between groups A,B, and C.

The percentage neural tissue data are shown inTable 3. The difference between the end-to-side neu-rorrhaphy group with regenerative distance of 0.6cm and the end-to-side neurorrhaphy group withdouble regenerative distance (1.2 cm) was not sig-nificant. Moreover, the ratio of percentage neuraltissue did not differ between the end-to-side neuror-

rhaphy group with a regenerative distance of 1.2 cmand the end-to-end nerve repair group.

Microscopic examination of all nerve segmentswas in accordance with quantitative histomorphom-etry of nerves (Fig. 2). In groups A and B, examina-tion of the bridging part of the tibial nerve revealeda large number of small myelinated fibers with rela-tively small calibers traveling in the outerepineurium (Fig. 3).

In groups A and B there was no microscopicevidence of donor nerve (tibial nerve) deterioration.Furthermore, no significant difference was obtainedbetween groups A and B in either the percentage ofneural tissue data and the ratio of percentage neuraltissue in the tibial nerve distal to end-to-side neuror-rhaphies.

Muscle Weights. Macroscopically, in groups A, B,and C the muscle atrophy on the experimental sidewas negligible (Fig. 4). However, in group D tibialisanterior muscle atrophy was obvious on the experi-mental side compared with the normal unoperatedside.

The tibialis anterior muscle weight data areshown in Table 4. The differences between groups Aand B, as well as between groups A and C, were notsignificant. However, there was a significant differ-ence between groups B and C (P � 0.05). Further-more, the difference between the three experimen-tal groups and the control group, in which the

Table 1. Peroneal function index scores.

Group A Group B Group C Group D

Pre-op �16.62 � 3.56 �14.05 � 3.35 �14.33 � 2.69 �14.35 � 4.64D 15 �44.02 � 5.29 �41.66 � 4.24 �35.41 � 2.12 �46.78 � 8.19D 30 �33.19 � 5.59 �36.63 � 3.50 �28.41 � 4.16 �46.89 � 5.89D 60 �24.12 � 4.73 �28.09 � 4.39 �22.92 � 3.62 �49.69 � 5.69D 90 �18.89 � 4.50 �19.93 � 4.69 �18.01 � 2.49 �62.81 � 2.55D 150 �16.09 � 2.79 �16.84 � 3.42 �13.94 � 2.68 �71.32 � 3.21

Average PFI scores in groups A, B, C, and D obtained at different postoperative intervals. All values are expressed as mean � standard deviation.

Table 2. Peroneal nerve fiber density.

Group Axon density (n/mm2) Ratio of axon density

A 13900 � 2900 (9400–18900) 1.73 � 0.35 (1.24–2.52)B 13100 � 4200 (5700–19000) 1.57 � 0.62 (0.67–2.68)C 13500 � 2600 (8800–17500) 1.95 � 0.46 (0.97–2.47)

No significant differences were noted between groups.Average peroneal nerve fiber density and density ratio of the experimentalto contralateral side on postoperative day 150. All values are expressed asmean � standard deviation.

Table 3. Neural tissue data and ratio of the experimental tocontralateral side of the peroneal nerve.

GroupPercentage neural tissue

(%) Ratio of neural tissue

A 38.44 � 4.69 (28.7–46.2)† 0.76 � 0.07 (0.60–0.85)B 35.98 � 10.81 (19.7–52.3)* 0.72 � 0.17 (0.42–0.94)C 48.86 � 5.48 (37.9–54.7)*† 0.83 � 0.05 (0.76–0.89)

*Significant differences between groups (P � 0.005).†Significant differences between groups (P � 0.001).All values are expressed as mean � standard deviation.


peroneal nerve was transected and buried into theneighboring muscles to prevent reinnervation, washighly significant (P � 0.001). According toTrumble,22 the relative loss of tibialis anterior mus-cle weight closely reflects the degree of denervation.

Muscle Histology. The tibialis anterior muscles ofthe experimental and the contralateral control sidewere stained with hematoxylin-eosin and evaluatedhistologically (Fig. 5). There was no significant dif-ference between groups A, B, and C in the histologicappearance of the muscles from the affected limbs.Furthermore, muscles of the experimental side weresimilar to muscles of the unoperated control side. Inthese groups transverse sections of muscle fasciclesshowed polygonal myofibers, with mostly subsar-colemmal nuclei and little intervening endomysialconnective tissue (Fig. 5A2,B2,C2). However, histo-logic examination of tibialis anterior muscles in con-trol group D revealed small bundles of atrophicfibers with small diameter and centrally placed nu-clei mingled with connective and adipose tissue (Fig.5D2).

DISCUSSION

Nerve regeneration after end-to-side neurorrhaphyinvolves: (1) the induction of collateral sprouting;(2) the ability of the growing axons to penetrate thedifferent layers; and (3) the readjustment of mo-toneurons that have adopted new motor units.29

There is a limit to the distance that axons regeneratespontaneously, which is less than 10 mm for ratsciatic nerves.11 When the nerve gap is longer than10 mm nerve regeneration will not occur, probablydue to the decrease of neurotropism from the distal

FIGURE 2. Transverse section of the distal part of the peroneal nerve 150 days postoperatively (toluidine blue, original magnification�400). (A) In group A histologic sections revealed a high number of myelinated axons. (B) Histologic appearance of the operated sidein group B was similar to group A. (C) In end-to-end group (group C), a normal pattern of nerve regeneration was obtained. (D) In groupD Wallerian degeneration was present.

FIGURE 3. Transverse section of the bridging part of the tibialnerve in group B (toluidine blue, original magnification �400).Small myelinated fibers with relatively small calibers are travelingin the outer epineurium.


nerve segment as the gap increases.6,17 In order todetermine the capacity for spontaneous nerve regen-eration in the rat model using histologic parameters,Mackinnon et al.14 demonstrated a mean regenera-tive distance of 20 mm.

The present study was conducted to determineperipheral nerve regenerative capacity after end-to-side neurorrhaphy. Functional and histologic evi-dence showed that the intact tibial nerve functionedas a bridge for regenerating axons not only forgroup A, but also for group B. Notably, when thedistal segment of the peroneal nerve was suturedend-to-side to the tibial nerve 1.2 cm distal to itsemergence from the sciatic nerve trunk, the growingaxons traveled in the outer epineurium of the tibial

nerve with approximately double the speed of that ingroup A (half distance), based on the time thatfunctional recovery occurred among experimentalgroups. In group B functional recovery of anteriortibialis muscle was present 3 months postoperatively,when PFI values were not significantly different frompreoperative values. The same duration was neces-sary for functional recovery of anterior tibialis mus-cle in groups A and C. Consequently, nerve gapscould be bridged over a long distance using end-to-side nerve coaptation. However, confirmation of thepresent findings is required.

Histologic studies showed no evidence that theend-to-side technique might have impaired tibial nervefunction in group A or B. In these two experimentalgroups normal tibial nerve function permitted partialweight-bearing on the experimental side as early as thefirst postoperative functional evaluation. Over the nextweeks functional improvement was gradual, with theimprint reaching a nearly normal appearance by the90th day. At the end of the study there was no signifi-cant difference in wet muscle weights between groupsA and B. This fact may indicate that adequate musclereinnervation is possible with a long regenerative dis-tance using end-to-side neurorrhaphy.

FIGURE 4. Macroscopic appearance of the tibialis anterior muscles of control (left in each picture) and experimental (right in each picture)limbs. There was minimal atrophy in muscles of the experimental limbs in groups A, B, and C. In contrast, in group D the atrophy wasevident.

Table 4. Average wet muscle weights.

Group Mean � SD Minimum Maximum

A 0.83 � 0.07 0.71 0.92B 0.81 � 0.09 0.68 0.97C 0.88 � 0.04 0.81 0.94D 0.19 � 0.04 0.13 0.28

Average wet muscle weights reported as the ratio of the experimental to thecontralateral normal side on postoperative day 150.


Viterbo et al.24 described the use of “double”end-to-side neurorrhaphy. This technique stimulatesaxonal growth by a supercharged effect comparedwith end-to-end repair. We used “double” neuror-rhaphy in end-to-side repair groups A and B. Usingthis technique the distal peroneal nerve is believedto be regenerated by axons that arise from the prox-imal peroneal stump without any contaminationfrom the tibial nerve. An extra group where theproximal stump of the peroneal nerve was not su-tured in an end-to-side fashion to the tibial nerve

would determine the contribution from the tibialnerve to distal repair. Retrograde labeling would benecessary to know the source of regenerating axonsinto the distal peroneal nerve. However, this wasbeyond the purpose of the present study.

A significant variable affecting axonal regenera-tion after end-to-side neurorrhaphy is the use ofepineurial rather than perineurial sutures.2 Peri-neurial sutures are more likely to induce collateralsprouting than epineurial sutures. However, it is pos-sible that the better axonal regeneration seen in

FIGURE 5. Histologic appearance of the experimental tibialis anterior muscles in groups A (A2), B (B2), and C (C2) was similar(hematoxylin-eosin, �100). In group D (D2) experimental tibialis anterior muscles were atrophic with small, angulated fibers. Tibialisanterior muscles of unoperated limbs in groups A (A1), B (B1), C (C1), and D (D1) were used as controls.


repairs with perineurial sutures might be due to agreater degree of donor nerve damage. Lundborg etal.12 noticed that end-to-side anastomosis through anepineurial window results in a greater rate of axonalregeneration. In our study, epineurotomy was madeby meticulous technique without damaging donornerve fascicles.

In conclusion, “double” end-to-side neuror-rhaphy can provide satisfactory functional recoveryfor the peroneal nerve, without any deterioration ofthe tibial nerve function. Functional evaluation andaxonal counting demonstrated that nerve regenera-tion can be supported by the use of an intact nervebridge technique with the same results for a distanceof 0.6 cm or 1.2 cm in a rat model. Consequently, a“double” end-to-side technique may be a valuabletool when end-to-end coaptation is not possible.

REFERENCES

1. Al-Qattan MM. Prevention and treatment of painful neuro-mas of the superficial radial nerve by the end-to-side nerverepair concept: an experimental study and preliminary clini-cal experience. Microsurgery 2000;20:99–104.

2. Al-Qattan MM, Al-Thunyan A. Variables affecting axonal re-generation following end-to-side neurorraphy. Br J Plast Surg1998;51:238–242.

3. Aszmann OC, Korak KJ, Rab M, Grunbeck M, Lassmann H,Frey M. Neuroma prevention by end-to-side neurorrhaphy: anexperimental study in rats. J Hand Surg (Am) 2003;28:1022–1028.

4. Bain JR, Mackinnon SE, Hunter DA. Functional evaluation ofcomplete sciatic, peroneal, and posterior tibial nerve lesionsin the rat. Plast Reconstr Surg 1989;83:129–138.

5. Balance CA, Balance HA, Stewart P. Remarks on the operativetreatment of chronic facial palsy of peripheral origin. BrMed J 1903;2:1009–1013.

6. Brunelli GA, Vigasio A, Brunelli GR. Different conduits inperipheral nerve surgery. Microsurgery 1994;15:173–178.

7. Dellon ES, Dellon AL. Functional assessment on neurologicalimpairment: track analysis in diabetic and compression neu-ropathies. Plast Reconstr Surg 1991;88:686–694.

8. De Medinaceli L, Freed WJ, Wyatt RJ. An index of the func-tional condition of rat sciatic nerve based on measurementsmade from walking tracks. Exp Neurol 1982;77:634.

9. Hadlock T, Elisseeff J, Langer R, Vacanti J, Cheney M. Atissue-engineered conduit for peripheral nerve repair. ArchOtolaryngol Head Neck Surg 1998;124:1081–1086.

10. Kim DH, Connolly SE, Kline DG, Voorhies RM, Smith A,Powell M, et al. Labeled Schwann cell transplants versus suralnerve grafts in nerve repair. J Neurosurg 1994;80:254–260.

11. Lundborg G, Dahlin LB, Danielsen N, Hansson HA, Johan-nesson A, Longo FM, et al. Nerve regeneration across anextended gap: a neurobiological view of nerve repair and thepossible involvement of neuronotrophic factors. J Hand Surg(Am) 1982;7:580–587.

12. Lundborg G, Zhao Q, Kanje M, Danielsen N, Kerns JM. Cansensory and motor collateral sprouting be induced from in-tact peripheral nerve by end-to-side anastomosis? J Hand Surg(Br) 1994;19:227–282.

13. Lykissas MG, Sakellariou E, Vekris MD, Kontogeorgakos VA,Batistatou AK, Mitsionis GI, et al. Axonal regeneration stim-ulated by erythropoietin: an experimental study in rats. J Neu-rosci Methods 2007;164:107–115.

14. Mackinnon SE, Hudson AR, Hunter DA. Histologic assess-ment of nerve regeneration in the rat. Plast Reconstr Surg1985;75:384–388.

15. Madison RD, Da Silva CF, Dikkes P. Entubulation repair withprotein additives increases the maximum nerve gap distancesuccessfully bridged with tubular prostheses. Brain Res 1988;447:325–334.

16. Martini R, Schachner M, Brushart TM. L2/HNK-1 carbohy-drate is preferentially expressed by previously motor axon-associated Schwann cells in reinnervated peripheral nerves.J Neurosci 1994;14:7180–7191.

17. McCallister WV, Cober SR, Norman A, Trumble TE. Usingintact nerve to bridge peripheral nerve defects: an alternativeto the use of nerve grafts. J Hand Surg (Am) 2001;26:315–325.

18. Noah EM, Williams A, Jorgenson C, Skoulis TG, Terzis JK.End-to-side neurorraphy: a histologic and morphometricstudy of axonal sprouting into an end-to-side nerve graft. JReconstr Microsurg 1997;12:99–106.

19. Pelissier P, Riahi R, Casoli V, Martin D, Baudet J. Les anasto-moses nerveuses terminolaterales. Rapport clinique prelimi-naire a propos de dix cas. Ann Chir Plast Esthet 2001;46:129–133.

20. Strauch B. Use of nerve conduits in peripheral nerve repair.Hand Clin 2000;16:123–130.

21. Tham SK, Morisson WA. Motor collateral sprouting throughan end-to-side nerve repair. J Hand Surg (Am) 1998;23:844–851.

22. Trumble TE. Peripheral nerve transplantation: the effects ofpredegenerated grafts and immunosuppression. J NeuralTransplant Plast 1992;3:39–49.

23. Viterbo F. A new method for treatment of facial palsy: thecross-face nerve transplantation with end-to-side neuror-rhaphy. Rev Soc Bras Cir Plast Estet Reconstr 1993;8:85.

24. Viterbo F, Trindade JC, Hoshino K, Mazzoni Neto A. Latero-terminal neurorrhaphy without removal of the epineuralsheath: experimental study in rats. Rev Paul Med 1992;110:267–275.

25. Viterbo F, Trindade JC, Hoshino K, Mazzoni Neto A. Twoend-to-side neurorrhaphies and nerve graft with removal ofthe epineural sheath: experimental study in rats. Br J PlastSurg 1994;47:75–80.

26. Viterbo F, Trindade JC, Hoshino K, Mazzoni Neto A. End-to-side neurorrhaphy with removal of the epineural sheath: anexperimental study in rats. Plast Reconstr Surg 1994;94:1038–1047.

27. Voche P, Ouattara D. End-to-side neurorrhaphy for defects ofpalmar sensory digital nerves. Br J Plast Surg 2005;58:239–244.

28. Yuksel F, Peker F, Celikoz B. Two applications of end-to-sidenerve neurorrhaphy in severe upper-extremity nerve injuries.Microsurgery 2004;24:363–368.

29. Zhang F, Fischer KA. End-to-side neurorrhaphy. Microsurgery2002;22:122–127.

30. Zhang Y, Campbell G, Anderson PN, Martini R, Schachner M,Lieberman AR. Molecular basis of interactions between re-generating adult rat thalamic axons and Schwann cells inperipheral nerve grafts. I. Neural cell adhesion molecules.J Comp Neurol 1995;361:193–209.

31. Zhang Y, Campbell G, Anderson PN, Martini R, Schachner M,Lieberman AR. Molecular basis of interactions between re-generating adult rat thalamic axons and Schwann cells inperipheral nerve grafts. II. Tenascin-c. J Comp Neurol 1995;361:210–224.

32. Zhang Z, Soucacos PN, Bo J, Beris AE. Evaluation of collateralsprouting after end-to-side nerve coaptation using a fluores-cent double-labeling technique. Microsurgery 1999;19:281–286.

33. Zhao J, Chen Z, Chen T. Nerve regeneration after termino-lateral neurorraphy: experimental study in rats. J ReconstrMicrosurg 1997;13:31–37.


ABSTRACT: Laminin �2 deficiency causes �50% of human congenitalmuscular dystrophies. Muscle in the corresponding dy/dy mouse model hasreduced force but increased fatigue resistance during isometric contractions.To determine whether a similar pattern of alterations is present duringisotonic contractions, dy/dy diaphragm was studied in vitro. During 20%load, dystrophic diaphragm had significantly reduced shortening, shorteningvelocity, work and power deficits, which persisted during the fatigue-inducingstimulation. In contrast, during 40% load, isotonic contractile performance ofdiseased muscle was impaired only mildly and only for some contractileparameters. At both loads, rate of isotonic fatigue when expressed relativeto initial contractile values was similar for dystrophic and normal muscle, orin some instances slightly higher for dystrophic muscle. Therefore, fatigueresistance is considerably impaired during isotonic contractions relative tothat reported previously for isometric contractions. This has important impli-cations for increased susceptibility to respiratory failure in laminin �2–deficient muscular dystrophy.

Muscle Nerve 36: 672–678, 2007

ISOTONIC FATIGUE IN LAMININ �2–DEFICIENTdy/dy DYSTROPHIC MOUSE DIAPHRAGM

JENNIFER POLLARINE, BA, MICHELLE MOYER, MS, and ERIK VAN LUNTEREN, MD

Department of Medicine, Cleveland Department of Veterans Affairs, Medical Center, andCase Western Reserve University, 10701 East Boulevard, Cleveland, Ohio 44106, USA


The muscular dystrophies are a heterogeneousgroup of inherited disorders that are characterizedby abnormal muscle wasting and weakness.7,37 Vari-ation in muscle-fiber size, necrosis, abnormal regen-eration of muscle tissue, invasion of muscle by mac-rophages, and replacement of muscle by connectivetissue and fat occur in most forms of muscular dys-trophy.8 These physiological abnormalities result inmuscle weakness that can lead to severe respiratoryproblems and death.3,7 Patients with congenitalforms of muscular dystrophy exhibit muscle weak-ness at birth or shortly thereafter. Infants are de-scribed as “floppy” and are unable to lift their limbsand head.16 All motor development is delayed inthese children and many fail to walk. Among thesechildren, 50% are lacking the important muscle pro-tein laminin �2 (also known as merosin).4

Laminin �2 is a component of the basal laminain muscle and peripheral nerves and is important inmaintaining muscle integrity because of its connec-tions to the sarcolemmal cytoskeleton and basal lam-ina. This muscle protein has roles in muscle celladhesion, differentiation, growth, shape, and migra-tion, and its absence causes large-scale degenerationof myofibrils.14 Absence of this protein results in theclassic form of congenital muscular dystrophy.19

Several rodent models are available for the studyof congenital muscular dystrophy, including thelaminin �2–deficient dy/dy mouse.4 These mice de-velop muscle weakness several weeks after birth,which worsens progressively, leading to a markedlyshortened lifespan.5 Variation in muscle-fiber size,connective tissue proliferation, and muscle-fiber ne-crosis also occur.5 Previous in vitro studies have sug-gested that laminin �2–deficient dy/dy diaphragmexhibits high resistance to fatigue during isometriccontractions, despite the muscle weakness associatedwith this disease.13,32 Connolly et al.5 found similarresults (in forearm) using the intact animal and agrip–pull protocol. Most of these studies attributedincreased fatigue resistance to changes in fiber typecomposition of the laminin �2–deficient musclebased on an increase in the proportion of slow-twitchfibers and a decrease in fast-twitch fibers.13,32 These


Abbreviations: ANOVA, analysis of variance; EDL, extensor digitorum lon-gus; Lo, optimal lengthKey words: contraction, laminin, muscular dystrophy, shorteningCorrespondence to: J. Pollarine; e-mail: [email protected]


672 Isotonic Fatigue in dy/dy Diaphragm MUSCLE & NERVE November 2007

results are supported by histology of dy/dy muscle, aswell as myosin composition and gene expressionanalysis.9,13,34 Gene expression array studies on dy/dydiaphragm revealed other biological processes thatwere affected by laminin �2 deficiency, such as cellmotility, development, defense/immune response,and cell adhesion.34

Initial studies on muscles with laminin �2 defi-ciency dealt with isometric contractions. The dia-phragm functions by shortening and elongating dur-ing breathing, and the magnitude of these lengthchanges is modulated by factors such as hypercapnia,airway occlusion, and bronchoconstriction.6,22,28,29 Itis therefore important to study the isotonic contrac-tile properties of the diaphragm to obtain a morecomprehensive understanding of the extent towhich laminin �2 deficiency affects the respiratorymuscle system. There are important differences be-tween isometric and isotonic contractions, in thatmore energy is consumed and heat generated dur-ing isotonic contractions than isometric contrac-tion.12,15,18,25,27 Laminin deficiency could cause moreimpairment of shortening than isometric contrac-tions because of the larger energy and mechanicaldemands during the former. This could be caused bydisturbances in cell-to-cell adhesion as well as alteredmetabolic and contractile properties resulting fromthe fiber subtype shifts.10,11,25,26 Therefore, testingdy/dy muscle using isotonic contractions is importantfor better elucidating the contractile deficits of lami-nin �2–deficient dy/dy skeletal muscle. We hypothe-size that the isotonic contractile properties of dy/dydystrophic diaphragm muscle differ considerablyfrom that of wildtype muscle.

MATERIAL AND METHODS

We used male homozygous dy/dy dystrophic mice(129P1/ReJ-Lama2dy; n � 11) and their phenotypi-cally normal (�/?) controls (n � 13). Mice wereobtained from Jackson Laboratories (Bar Harbor,Maine) and given free access to food and water.Studies were conducted when mice were 7–9 weeksold, at which time dystrophic mice were smaller thannormal (13.6 � 0.4 g vs. 24.1 � 0.4 g, P � 0.001).The mice were anesthetized with an intraperitonealinjection of rodent anesthetic cocktail (ketamine,xylazine, and acepromazine). The diaphragm wassurgically removed and special care was taken tokeep intact the connective tissue at the margin of therib cage and the central tendon of the diaphragm.Once removed, the diaphragm was kept in an aer-ated bath of physiological Krebs solution. The dia-phragm was cut into �3-mm-wide strips and surgical

thread was tied at both the rib and central tendonends. The muscle strips were mounted vertically in adouble-jacketed bath containing aerated (95% O2,5% CO2) physiological solution kept at a constant37°C. The physiological solution was comprised asfollows (in mM): 135 NaCl, 5 KCl, 2.5 CaCl2, 1MgSO4, 1 NaH2PO4, 15 NaHCO3, and 11 glucosewith an adjusted pH of 7.35–7.45.

A dual-mode servo-controlled force transducerwas used to measure the performance of the muscle(model 300B; Aurora Scientific, Ontario, Canada).This force transducer measured both length andforce separately and was able to hold the force con-stant while the length was measured. It had a re-sponse time of 1.3 ms, and the sinusoidal frequencyresponse specification was 530 Hz at �3dB. Themuscle strip underwent supramaximal electricalstimulation (0.2-ms duration) via parallel platinumelectrodes placed �4 mm apart with the musclesituated in the middle. All muscle strips were stimu-lated at optimal length (Lo), and isometric force (50and 75 Hz) was determined both in absolute termsand normalized for calculated cross-sectional areas[muscle mass (g)/ Lo (cm) � muscle density]. Theabsolute force values for each muscle sample formedthe basis for the maximal loads during subsequentfatigue testing.

Muscle fatigue was assessed by repeatedly stimu-lating the muscle strip over a 5-min period usingfour protocols chosen based on the basis of previousstudies1,21,24,35: 50 Hz at 20% maximal load, 75 Hz at20% maximal load, 50 Hz at 40% maximal load, and75 Hz at 40% maximal load. The muscle was stimu-lated every 3 s with a train duration of 333 ms (therelatively long time between trains was necessary dueto limitations of the software that controlled andcollected data from the servo-controlled trans-ducer), which is similar to other studies.13,21 Datawere relayed to a computer using the data acquisi-tion and analysis program Dynamic Muscle Control(Aurora Scientific). Muscle fatigue was evaluated bymeasuring the maximum amount of shortening,work, shortening velocity, and power. Shorteningvelocity was measured beginning 10 ms after the firstdetectable length change in the muscle strip and fora duration of 20 ms.36 Work was calculated as theproduct of the isotonic afterload (load against whichthe muscle contracts under isotonic conditions) andthe amount of maximum shortening. Power was cal-culated as the product of the isotonic afterload andshortening velocity.35

Statistical analysis between two groups was doneusing unpaired t-tests. Fatigue data were analyzedusing a two-way repeated-measures analysis of vari-

Isotonic Fatigue in dy/dy Diaphragm MUSCLE & NERVE November 2007 673

ance (ANOVA). This was followed in cases of signif-icance with the Tukey and Neuman–Keuls tests(which gave identical results). Probability values ofP � 0.05 were considered statistically significant.Data appear as mean � SE.

RESULTS

Isometric physical and contractile characteristics ofthe muscle strips used in this study were significantlydifferent in dystrophic compared with normal mice.The optimal length of dystrophic muscle was de-creased by 11% when compared to normal (8.34 �0.30 vs. 9.33 � 0.24 mm; P � 0.014). Compared withnormal mice, diaphragm of dystrophic mice gener-ated 17% less twitch force (6.25 � 0.44 vs. 7.57 �0.36 N/cm2; P � 0.028), 36% less force during 50-Hzstimulation (9.74 � 1.40 vs. 15.18 � 1.21 N/cm2; P �0.003), and 23% less force during 75-Hz stimulation(14.13 � 0.91 vs. 18.40 � 1.11 N/cm2; P � 0.006).

At 20% load baseline values for maximum short-ening were decreased for dy/dy muscle by �35% atboth stimulation frequencies, and work was de-creased by 64% and 55% for 50 Hz and 75 Hz,respectively (Table 1). Shortening velocity was de-creased by 43% for 50 Hz and power was decreasedby 68% and 44% for 50 Hz and 75 Hz, respectively.In contrast, at 40% load dystrophic muscle baselinevalues differed only for power at 75 Hz when com-pared with normal muscle (Supplementary Table 1).

Examples of isotonic contractions in normal anddystrophic muscle strips during repetitive 50-Hzstimulation at 20% load are shown in Figure 1. Dys-trophic muscle did not perform as well as normalmuscle during the 5-min period, and there weresignificant differences between normal and dystro-phic muscle strips for all four indices of muscleperformance at both 50- and 75-Hz stimulation fre-quencies (Fig. 2). Furthermore, there were signifi-cant interactions between the presence or absence ofdisease and time during repetitive stimulation. Dif-ferences between the two groups were most highlyand consistently significant at early points during the

fatigue test. Differences between dystrophic and nor-mal muscle persisted even when normalized relativeto smaller size of the muscles from dystrophic ani-mals (Supplementary Fig. 1).

Data for muscle isotonic contractile performanceduring repetitive contractions at 40% load are shownin Figure 3. Both dystrophic and normal diaphragmfatigued appreciably over the 5-min time period.Repeated-measures ANOVA indicated there were nosignificant differences in overall contractile perfor-mance over time between normal and dystrophicmuscle strips at 40% load. However, in many in-stances there were significant interactions betweenthe presence or absence of disease and time duringrepetitive stimulation, so that early on during thefatigue tests the contractile parameters differed be-tween normal and dystrophic muscle. Generally sim-ilar results were found for 20% and 40% load whendata were normalized to muscle size (SupplementaryFig. 2).

To factor out the effects of differences in base-line performance between dystrophic and normalmuscle, the rate of fatigue was determined by exam-ining the normalized (% initial) values for the 20%and 40% load protocols. In most cases, normal anddystrophic muscle fatigued at the same rate duringthe 20% load (Fig. 4). Significant differences werepresent for maximum shortening and work for the75-Hz 20% load. However, dystrophic muscle fa-

Table 1. Normalized baseline values for maximum shortening, work, shortening velocity, and power during 50- and 75-Hz stimulation at20% load.

Measurements

50-Hz 20% load 75-Hz 20% load

Normal Dystrophic P-value Normal Dystrophic P-value

Maximum shortening (Lo) 0.30 � 0.02 0.19 � 0.02 �0.001 0.31 � 0.02 0.21 � 0.21 0.003Work (joules/m2) 71.2 � 7.76 25.6 � 6.08 �0.001 97.0 � 14.2 43.49 � 8.43 0.006Shortening velocity (Lo/s) 4.09 � 0.50 2.33 � 0.23 0.007 4.66 � 0.36 3.65 � 0.49 0.104Power (watts/m2) 1007 � 161 322 � 64 0.001 1428 � 209 796 � 156 0.030

Values are means � SE.

FIGURE 1. Examples of length tracings from normal and laminin�2-deficient dystrophic muscle strips at three timepoints (initial,2.5 min, 5 min) during isotonic fatigue (50-HZ 40% load).


tigued faster than normal muscle during these pro-tocols. In contrast, there was no difference in therate of fatigue, and no interaction between the pres-ence/absence of disease and time during the 40%load protocols (Supplementary Fig. 3).

DISCUSSION

We found that laminin �2–deficient dy/dy musclehad lower maximum shortening, work, shorteningvelocity, and power baseline values than normal mus-cle when stimulated at a 20% load. These decreaseswere maintained to varying degrees during the 5-minisotonic fatigue. In contrast, dy/dy tissue did not havesignificantly lowered contractile performance at ahigher load (40%). Furthermore, there was no dif-ference in fatigue during the 40% load, althoughthere were significant interactions between the pres-ence or absence of disease and time during repeti-tive stimulation. Nonetheless, increased isotonic fa-tigue resistance was not a feature of dystrophic

muscle, irrespective of load or stimulation fre-quency.

There have been several isometric experiments,all of which have found increased fatigue resistancein dy/dy muscle.13,31–33 In contrast, we did not findincreased fatigue resistance in laminin �2–deficientdy/dy muscle during isotonic contractions. In fact, wefound two parameters (maximum shortening andwork during 75-Hz 20% load) for which dystrophicmuscle fatigued at a significantly higher rate. It isunknown why there was no increased fatigue resis-tance found in dystrophic muscle during isotoniccontractions. It is possible that the larger energydemands placed on the shortening muscle exceedthe benefits that the increased slow-fiber composi-tion could provide in muscle fibers that are alreadyseverely impaired by the laminin �2 deficiency.

Several studies have examined isotonic contrac-tile performance in another model of muscular dys-trophy, dystrophin-deficient mdx muscle. Attal et al.2

FIGURE 2. Maximum shortening, work, shortening velocity, and power, in absolute terms, during 50- and 75-HZ stimulation at 20% load.Values are means � SE. P-values indicate results of ANOVA testing differences between normal and dystrophic muscle; Pint valuesindicate results of ANOVA testing interactions between disease and time. Asterisks indicate differences (P � 0.05) between normal anddystrophic muscle strips at each timepoint.


found decreases in shortening velocity (16%) andpower (66%) values in the sternohyoid muscle ofdystrophin-deficient mdx mice. Lynch et al.20 studiedthe dystrophin-deficient mdx diaphragm and found a54% decrease in power when compared to normal.The present study found similar decreases in laminin�2-deficient dy/dy shortening velocity as that foundin dystrophin-deficient mdx mice.

We are not aware of any isotonic fatigue studiesthat have been done in any dystrophic model. How-ever, Watchko et al.35 studied diaphragm isotonicfatigue properties in a genetically modified mousemodel (CK�/�) that was deficient in the muscleenzyme creatine kinase. The diaphragm from thesemice had significantly decreased maximum shorten-ing, work, shortening velocity, and power values dur-ing fatigue contractions. Furthermore, the isotoniccontractile values of the CK�/�-deficient diaphragmstrips declined to zero in �60% of the time that ittook the normal muscle. Watchko et al.35 performed

their experiments at the load that had the largestmaximal power (40% load) and a stimulation fre-quency of 75 Hz. It would be interesting to seewhether the isotonic fatigue of CK�/�-deficient dia-phragm is affected by changes in load to the samedegree as noted for dy/dy dystrophic muscle in thepresent study.

Regarding mechanisms accounting for the al-tered isotonic contractile properties in our study,several previous investigations have examined theeffects of laminin-deficient muscular dystrophy onmyosin isoform or fiber subtype distributions. Hayesand Williams13 found that dystrophic soleus musclecompared with normal muscle had an increased pro-portion of type I slow fibers and a decrease in IIa fastfibers. Fitzsimons and Hoh9 examined the myosincomposition of dystrophic soleus, extensor digito-rum longus (EDL), and gastrocnemius (the latterhas similar fiber type composition as the dia-phragm). They found that dystrophic soleus had an

FIGURE 3. Maximum shortening, work, shortening velocity, and power, in absolute terms, during 50- and 75-HZ stimulation at 40% load.Values are means � SE. P values indicate results of ANOVA testing differences between normal and dystrophic muscle; Pint valuesindicate results of ANOVA testing interactions between disease and time. Asterisks indicate differences (P � 0.05) between normal anddystrophic muscle strips at each timepoint.


abnormally high percentage of slow myosin and adecrease in intermediate myosin, dystrophic EDLhad increases in intermediate (which is normally notfound in EDL) and slow myosin, and dystrophicgastrocnemius had increased proportions of slowand intermediate myosin and decreased proportionsof fast myosins. The soleus and gastrocnemius haddecreased levels of type IIb fibers, whereas the type Ifibers were relatively spared and actually appeared ingreater numbers. Those authors also suggest thatthere is selective atrophy in dystrophic muscle wheretype II fibers are damaged and type I fibers remainundamaged and in greater numbers. Diaphragmstrips from human patients with Fukuyama musculardystrophy, another form of congenital muscular dys-trophy with deficient laminin levels, were analyzedhistologically. A type 1 fiber predominance wasfound as well as a type IIb deficiency.23 Kihira andNonaka17 found similar results in biceps brachii,quadriceps femoris, and gastrocnemius.

In addition to shifts among fiber types, geneticdeficiency of laminin �2 and other structural pro-teins leads to a large number of other downstreamchanges.30 For example, a gene expression arraystudy34 found 69 gene expression changes in dy/dymouse diaphragm, which were grouped into fourbasic themes: cell motility, development, defense/immune response, and cell adhesion. The cell mo-tility (muscle contraction) theme had overexpres-sion for contractile proteins and molecules that bindto contractile proteins. The development (muscle)theme included changes in morphogenesis, organo-genesis, cell differentiation, and regulation of devel-opment. Among these, Myog plays a role in regulat-ing enzyme activity, favoring oxidative over glycolyticmetabolism, which likely impacts the component offatigue resistance that is related to energy supplies.The fact that the disease-induced changes in myosinisoform distributions and fiber subtype distributionsdo not explain all of the contractile changes in the

FIGURE 4. Maximum shortening, work, shortening velocity, and power expressed as percentage initial values during 50- and 75-HZ

stimulation at 20% load. Values are means � SE. Asterisks indicate differences (P � 0.05) between normal and dystrophic muscle strips.P values indicate results of ANOVA testing differences between normal and dystrophic muscle; Pint values indicate results of ANOVAtesting interactions between disease and time. Asterisks indicate differences (P � 0.05) between normal and dystrophic muscle strips ateach timepoint.


present study suggests that these or other as yetunidentified factors may also be playing a role inmodulating contractile properties.

Supported in part by National Institutes of Health (NIH) grantHL-70697, as well as by funding from the research service of theDepartment of Veterans Affairs. We thank Dr. Jon Watchko foradvice in calculating work and power values.

REFERENCES

1. Attal P, Coirault C, Chemla D, Blanc FX, Rocher P, Pourny JC,et al. Isotonic mechanisms of a pharyngeal dilator muscle anddiaphragm in the rat before and after fatigue. Eur Respir J2000;15:308–313.

2. Attal P, Lambert F, Marchand-Adam S, Bobin S, Pourney J,Chemla D, et al. Severe mechanical dysfunction in pharyngealmuscle from adult mdx mice. Am J Respir Crit Care Med2000;162:278–281.

3. Blake DJ, Weir A, Newey SE, Davies KE. Function and geneticsof dystrophin and dystrophin-related proteins in muscle.Physiol Rev 2002;82:291–329.

4. Cohn RD, Campbell KP. Molecular basis of muscular dystro-phies. Muscle Nerve 2000;23:1456–1471.

5. Connolly AM, Keeling RM, Mehta S, Pestronk A, Sanes JR.Three mouse models of muscular dystrophy: the natural his-tory of strength and fatigue in dystrophin, dystrophin/utro-phin-, and laminin �2-deficient mice. Neuromuscul Disord2001;11:703–712.

6. Decramer M, De Troyer A, Kelly S, Macklem PT. Mechanicalarrangement of costal and crural diaphragms in dogs. J ApplPhysiol 1984;56:1484–1490.

7. Emery AH. The muscular dystrophies. BMJ 1998;317:991–995.

8. Emery AH. The muscular dystrophies. Lancet 2002;359:687–695.

9. Fitzsimons RB, Hoh JFY. Myosin isoenzymes in fast-twitch andslow twitch muscles of normal and dystrophic mice. J Physiol(Lond) 1983;343:539–550.

10. Geiger PC, Cody MJ, Sieck GC. Force-calcium relationshipdepends on myosin heavy chain and troponin isoforms in ratdiaphragm muscle fibers. J Appl Physiol 1999;87:1894–1900.

11. Geiger PC, Cody MJ, Macken RL, Sieck GC. Maximum spe-cific force depends on myosin heavy chain content in ratdiaphragm muscle fibers. J Appl Physiol 2000;89:695–703.

12. Han Y, Geiger PC, Cody MJ, Macken RL, Sieck GC. ATPconsumption rate per cross bridge depends on myosin heavychain isoform. J Appl Physiol 2003;94:2188–2196.

13. Hayes A, Williams DA. Examining potential drug therapies formuscular dystrophy utilizing the dy/dy mouse: I. Clenbuterol.J Neurol Sci 1998;157:122–128.

14. Helbling-Leclerc A, Zhang X, Topaloglu H, Cruaud C, TessonF, Weissenbach J, et al. Mutations in the laminin �2 chaingene (LAMA2) cause merosin-deficient congenital musculardystrophy. Nat Genet 1995;11:216–218.

15. Hill AV. The heat of shortening and dynamic constants ofmuscle. Proc R Soc Lond B Biol Sci 1938;126:136–195.

16. Hoffman EP. Muscular dystrophies. In: Dulbecco R, editor.Encyclopedia of the human body, Vol 5. San Diego: AcademicPress; 1997. p 901–906.

17. Kihira S, Nonaka I. A histochemical study with morphometricanalysis on biopsied muscles. J Neurol Sci 1985;70:139–149.

18. Kushmerick MJ. Energetics of muscle contraction. In: PeachyLD, Adrian RH, Geiger SR, editors. Handbook of physiology,Sect. 10. Bethesda, MD: American Physiology Society; 1983. p189–236.

19. Leyten QH, Gabreels FJM, Renier WO, ter Laak HJ. Congen-ital muscular dystrophy: a review of the literature. Clin NeurolNeurosurg 1996;98:267–280.

20. Lynch GS, Rafael JA, Hinkle RT, Cole NM, Chamberlain JS,JA Faulkner. Contractile properties of diaphragm muscle seg-ments from old mdx and old transgenic mdx mice. Am JPhysiol Cell Physiol 1997;41:C2063–C2068.

21. Machiels HA, Van Der Heijden HFM, Heunks LMA, Dekhui-jzen PNR. The effect of hypoxia on shortening contractions inrat diaphragm muscle. Acta Physiol Scand 2001;173:313–321.

22. Newman S, Road J, Bellemare F, Clozel JP, Lavigne CM,Grassino A. Respiratory muscle length measured by sonomi-crometry. J Appl Physiol 1984;56:753–764.

23. Nonaka I, Sugita H, Takada K, Kumagai K. Muscle histochem-istry in congenital muscular dystrophy with central nervoussystem involvement. Muscle Nerve 1982;5:102–106.

24. Seow CY, Stephens NL. Fatigue of mouse diaphragm musclein isometric and isotonic contractions. J Appl Physiol 1988;64:2388–2393.

25. Sieck GC, Han Y, Prakash YS, Jones KA. Cross-bridge cyclingkinetics, actomyosin ATPase activity and myosin heavy chainisoforms in skeletal and smooth respiratory muscles. CompBiochem Physiol B Biochem Mol Biol 1998;119:435–450.

26. Sieck GC, Prakash YS. Cross-bridge kinetics in respiratorymuscles. Eur Respir J 1997;10:2147–2158.

27. Sieck GC, Prakash YS, Han Y, Fang Y, Geiger PC, Zhan W.Changes in actomyosin ATP consumption rate in rat dia-phragm muscle fibers during postnatal development. J ApplPhysiol 2003;94:1896–1902.

28. van Lunteren E, Cherniack NS. Electrical and mechanicalactivity of respiratory muscles during hypercapnia. J ApplPhysiol 1986;61:719–727.

29. van Lunteren E, Haxhiu MA, Deal EC Jr, Arnold JS, Cherni-ack NS. Respiratory changes in thoracic muscle length duringbronchoconstriction. J Appl Physiol 1987;63:221–228.

30. van Lunteren E, Leahy P. Gene expression microarrays andrespiratory muscles. Respir Physiol Neurobiol 2007;156:103–115.

31. van Lunteren E, Moyer M. Improvement of dy/dy dystrophicdiaphragm by 3,4-diaminopyridine. Muscle Nerve 2002;26:71–78.

32. van Lunteren E, Moyer M. Reduced fatigue in diaphragmmuscle of merosin-deficient dy/dy dystrophic mice. Respira-tion 2003;70:636–642.

33. van Lunteren E, Moyer M. Sternohyoid muscle fatigue prop-erties of dy/dy dystrophic mice, an animal model of merosin-deficient congenital muscular dystrophy. Pediatr Res 2003;54:547–553.

34. van Lunteren E, Moyer M, Leahy P. Gene expression profilingof diaphragm muscle in �2-laminin (merosin) deficient dy/dydystrophic mice. Physiol Genomics 2005;25:85–95.

35. Watchko, JF, Daood MJ, Sieck GC, LaBella JJ, Ameredes BT,Koretsky AP, et al. Combined myofibrillar and mitochondrialcreatine kinase deficiency impairs mouse diaphragm isotonicfunction. J Appl Physiol 1997;82:1416–1423.

36. Watchko, JF, Daood MJ, Wieringa B, Koretsky AP. Myofibrillaror mitochondrial creatine kinase deficiency alone does notimpair mouse diaphragm isotonic function. J Appl Physiol2000;88:973–980.

37. Watchko JF, O’Day TL, Hoffman EP. Functional characteris-tics of dystrophic skeletal muscle: insights from animal mod-els. J Appl Physiol 2002;93:407–417.


ABSTRACT: The X-linked recessive disease phosphoglycerate kinase(PGK) deficiency is caused by altered expression of the PGK1 enzyme,which causes muscle stiffness, hemolytic anemia, and mental retardation. Inthis study we characterized the PGK1 gene in a family of two brothers, twosisters, and their parents. A single mutation in exon 6, which was associatedwith the pattern of inheritance of PGK1 deficiency, was observed. This silentG213G; c.639C�T mutation was localized to the conserved exon–intronboundary. We have developed a method for quantification of PGK1 mRNAand demonstrated a marked reduction in PGK1 mRNA in both brothers withthe disease. A smaller decrease in PGK1 expression was observed in onesister with symptoms of PGK deficiency and in her mother. Only the normalPGK1 allele was expressed in the two heterozygous women. Whereas mostknown PGK1 mutations cause amino acid alterations, our study indicatesthat inhibition of the transcription mechanism is the cause of PGK deficiency.

Muscle Nerve 36: 679–684, 2007

ALTERED EXPRESSION OF PGK1 IN A FAMILYWITH PHOSPHOGLYCERATE KINASE DEFICIENCY

EVA K. SVAASAND, MSc,1 JAN AASLY, MD, PhD,2,3 VESLEMØY MALM LANDSEM, MSc,4 and

HELGE KLUNGLAND, PhD4

1 Department of Pathology and Medical Genetics, St. Olav Hospital, Trondheim, Norway2 Department of Neurology, St. Olav Hospital, Trondheim, Norway3 Norwegian University of Science and Technology, Faculty of Medicine, Department of

Neuromedicine, Trondheim, Norway4 Norwegian University of Science and Technology, Faculty of Medicine, Department of

Laboratory Medicine, Children and Women’s Health, Trondheim, Norway


Phosphoglycerate kinase 1 (PGK1) is a glycolyticenzyme that plays an important role in glycolysis,catalyzing the reversible conversion of 1,3-disphos-phoglycerate to generate one molecule of adenosinetriphosphate. PGK1 deficiency is a rare genetic de-fect that has variable clinical symptoms. According toSugie et al.,16 the clinical manifestations are hetero-geneous and can be divided into two main types: oneis characterized by hemolytic anemia or mental re-tardation, whereas myopathy dominates in the other.Neurological symptoms consist mainly of variablemental retardation,8 emotional instability, or epilep-sy,7 and other symptoms include exercise-inducedmuscle stiffness and cramping, often accompanied

by rhabdomyolysis and severe myoglobinuria andsometimes followed by acute renal failure.7,11

Among known cases of PGK1 deficiency, hemo-lytic anemia combined with central nervous systemdefects is common. In some patients, myopathy isobserved exclusively, whereas all three major clinicalsymptoms have been observed in only two cases.9,10

Based on the monomeric nature of the PGK1 en-zyme, the wide variety of clinical symptoms is surpris-ing. The enzyme is expressed in all somatic tissues,and it seems likely that the biochemical properties ofindividual PGK1 mutations may explain the distinctfeatures of individual patients.7

The diagnosis of PGK1 deficiency cannot bemade clinically. Enzymatic muscle tests are requiredto distinguish this disorder from the many otherdiseases in the glycolysis/glycogenolysis and fattyacid metabolism pathways that present similar symp-toms.1,5

A previous report by Aasly et al.1 described twoyoung brothers who experienced muscle pain,cramps, and stiffness following heavy exercise. Bothhad very low PGK1 levels in muscle (2%–3% ofcontrols), as well as in blood cells and fibroblasts.

This article includes Supplementary Material available via the internet atwww.mrw.interscience.wiley.com/suppmat/0148-639X/suppmat/

Abbreviations: IVS, intervening sequences; mRNA, messenger RNA; PCR,polymerase chain reaction; PGK, phosphoglycerate kinase; RNA, ribonucleicacid; RT-PCR, reverse transcriptase–polymerase chain reactionKey words: mRNA expression, myoglobinuria, myopathy, PGK1, phospho-glycerate kinase deficiencyCorrespondence to: H. Klungland; e-mail: [email protected]


Altered Expression of PGK1 MUSCLE & NERVE November 2007 679

Since PGK1 deficiency is an X-linked trait, moststudies of PGK1 defects are performed on men. Inthe family observed in the present study, a heterozy-gous sister also showed clinical symptoms.

The purpose of our study was to find a molecularexplanation for the reduced PGK1 enzyme activityobserved in two brothers affected by PGK1 defi-ciency. We initiated our study by searching for anucleotide change in the coding sequence and pro-moter of the PGK1 gene and extended our investi-gations to examine the level of PGK1 messengerRNA (mRNA) expression.


Family Material and Isolation of Nucleic Acids. Bloodsamples were taken from the two brothers, their twosisters, and their mother and father (Fig. 1). Bothbrothers had suffered several episodes of severe mus-cle stiffness and cramping induced by physical exer-cise, sometimes accompanied by myoglobinuria.1One of the sisters, D2, developed mild muscle stiff-ness during hard physical exercise, but had no epi-sodes of myoglobinuria. The other sister and parentswere asymptomatic.

Two unrelated individuals, one of each gender,served as controls in all experiments, and 50 addi-tional random controls from the same geographicalregion were used to verify the findings. DNA wasisolated from blood with ethylenediaminetetraaceticacid anticoagulant according to standard protocols(QIAmp Blood Kit; Qiagen, Valencia, California).Additional samples for mRNA analysis were takenfrom the whole family (except from the father, whodid not have the silent mutation), and from twocontrols, again one of each gender. Two PAXgene

Blood RNA Tubes (PreAnalytiX, Hombrechtikon,Switzerland) were obtained from each individualand RNA was isolated according to the protocol. Theconcentration of RNA was normalized using theNanoDrop spectrophotometer (NanoDrop Technol-ogies, Wilmington, Delaware) and dilution in dieth-ylpyrocarbonate-treated water.

Amplification and Sequencing of the Complete PGK1

Coding Sequences. Primers (1–24), described inSupplementary Table 1, were used to amplify all 11exons and the promoter region of the PGK1 gene.Polymerase chain reactions (PCRs) were run in a50-�l reaction containing 5 �l GeneAmp 10 � PCRbuffer (100 mM Tris-HCl, pH 8.3, 500 mM KCl), 2mM MgCl2 (exon 6 with 3 mM and exon 9 with 8mM MgCl2), 0.25 �M dNTP, 0.3 �M of each primer,2.0 U AmpliTaq GoldDNA Polymerase (Applied Bio-systems, Foster City, California) and 75 ng genomicDNA. After a 10-min incubation at 94°C, 35 cycles ofPCR were run at 94°C for 30 s, 55°C for 30 s (exon5 at 57°C and exon 1 at 59°C), and 72°C for 30 s,followed by a 10-min elongation step at 72°C. Suc-cessful PCR reactions were verified on agarose gel,purified using QIAamp PCR Purification Kit (Qia-gen), and sequenced using the Big Dye TerminatorCycle Sequencing Reaction Kit v. 3.0 from AppliedBiosystems. Sequencing reactions were analyzed onABI 310 and ABI 3100 sequencers (Applied Biosys-tems).

Amplification, Quantification, and Sequencing of PGK1

mRNA. Primers (25–31), described in Supplemen-tary Table 1 online, were used to amplify and se-quence the PGK1 mRNA. One-step reverse transcrip-tase (RT)-PCR was performed with primers 27 and28 in a 50-�l reaction mix containing 10 �l 5�Qiagen One Step RT-PCR buffer (2.5 mM MgCl2final concentration), 400 �M of each dNTP, 0.6 �Mof each gene-specific primer (27 and 28), 2.0 �l

FIGURE 1. Family tree illustrating the family unit evaluated in thisstudy. Among the four children, both sons (S1 and S2) areaffected, the father and one of the daughters (D1) are unaffected,whereas the other daughter (D2) and the mother are carriers.Females are indicated with a circle, males with a square. Thehealthy individuals have clear symbols, the affected individualshave black symbols, and the carriers have bull’s-eye symbols.

Table 1. Distribution of polymorphisms found in the PGK1 intron5 and exon 6.

Intron 5 polymorphismIVS5-24C�T

Exon 6 polymorphismG213G; c.639C�T

Mother G/A C/TFather G CDaughter 1 G/A C/CDaughter 2 G/G C/TSon 1 G TSon 2 G TControl, female G/A C/CControl, male G C

680 Altered Expression of PGK1 MUSCLE & NERVE November 2007

Qiagen One Step RT-PCR enzyme mix, and 1 pg to2 �g total RNA. PCR was performed using the fol-lowing conditions: 50°C for 30 min, 95°C for 15 min,35 cycles of PCR run at 94°C for 1 min, 63°C for 1min, and 72°C for 1 min 45 s, followed by a 10-minelongation step at 72°C.

Two-step RT-PCR was performed with primers 25and 26. Reverse transcriptase was incubated in a20-�l RT-reaction mix containing 4 �l 5� RT-PCRbuffer, 2.5 mM MgCl2, 250 �M of each dNTP, 10mM DTT, 1.25 �M random hexamer, 10 U RNaseInhibitor, 15 U MultiScribe reverse transcriptase(Applied Biosystems) and �0.1 �g of DNase-treatedtotal RNA. The reaction mix was incubated at 25°Cfor 10 min, followed by incubation at 42°C for 12min. PCR was performed in a 50-�l reaction mixcontaining 9.4 �l 5� RT-PCR buffer, 1.75 mMMgCl2, 200 �M of each dNTP, 0.15 �l of each primer(25 and 26), 2.5 U AmpliTaq Gold DNA polymerase(Applied Biosystems), and 3 �l of the RT reaction.After a 10-min incubation at 94°C, 43 cycles of PCRwere run at 94°C for 1 min, 55°C for 1 min, and 72°Cfor 1 min, followed by a 10-min elongation step at72°C.

Successful PCR reactions were verified on aga-rose gel, purified using the QIAamp PCR Purifica-tion Kit (Qiagen), and sequenced using the Big DyeTerminator Cycle Sequencing reaction kit v. 3.0from Applied Biosystems. Sequencing reactions wereanalyzed on ABI 310 and ABI 3100 sequencers. Rel-ative quantification of the mRNA was performedusing the Versadoc equipment (BioRad, Hercules,California) with the VersaDoc Quantity One v. 4.4.1.software, applying the global method.

RESULTS

PGK1 enzyme activity had previously been measuredin the two brothers, as described by Aasly et al.1 PGKactivity in muscle was deficient in both brothers,being only 2%–3% of levels observed in mean con-trols.1 Myopathy and myoglobinuria are the mostdominant clinical manifestations of PGK1 defi-ciency, so enzyme measurements were performed inmuscle biopsies. The invasive nature of a musclebiopsy and the young age of the daughters dictatedthat it was not possible to collect muscle enzymemeasurements from these family members.

Sequencing of the complete coding regions andthe promoter of the PGK1 gene unveiled three poly-morphisms/mutations in the family studied. One ofthese alterations, a single base substitution in intron6 (IVS6�28C�T), was found in all individualstested, including the two controls. This mutation

could represent an error in the published PGK1sequence (Access. No. M11963) or a variation in ourNorwegian population. Two other mutations local-ized to intron 5 and exon 6 were distributed asshown in Table 1. For intron 5, the mutation located24 bases prior to the beginning of exon 6 (IVS5-24C�T) was not associated with the disease, and alsooccurred in a significant number of the controls.The third mutation, localized to the end of exon 6,followed the expected inheritance of the disease(Fig. 2). Fifty additional samples from unrelated in-dividuals were tested and all were homozygous/hemizygous for the normal allele. However, theG213G; c.639C�T alteration does not cause anyamino acid alteration that could directly explain thephenotype.

Based on these findings, new blood samples weretaken from which to isolate and quantify PGK1mRNA. One-step RT-PCR using exon 1 F and exon11 R primers (Supplementary Table 1) resulted inthe amplification of two fragments that were visual-ized on agarose gels. Whereas the smaller fragmenthad the expected length of PGK1 complementaryDNA (verified by sequencing), sequencing of thelonger fragment verified a sequence identical to thePGK1P2 pseudogene (Fig. 3A; Access. No.AC010422). This competitive PCR was used forquantification of PGK1 mRNA as shown in Table 2.For the two affected sons, the expression levels wereestimated to be 11% and 14% of the expressionestimated in individuals with a normal PGK1 se-quence. The two heterozygous women (mother anddaughter 2) also showed significantly reduced PGK1mRNA (32.5% and 38.2%, respectively). When com-plementary DNA from these two women was ana-lyzed the mutated sequence was not detected amongthe PGK1 mRNA (Fig. 2A).

A second experiment was set up specifically toamplify PGK1 mRNA with primers positioned in ar-eas with low similarity to known pseudogenes (Prim-ers 27 and 28 in Supplementary Table 1). A singlespecific band was amplified (Fig. 3B). A very similarexpression profile was observed in this nonquantita-tive experiment (Table 2, lower line), probably indi-cating that the PCR was still in the exponentialphase.

DISCUSSION

Molecular analysis of a family with PGK1 deficiency1

revealed a silent mutation following the X-linkedmode of inheritance. We focused our search formutations in the PGK1 gene and examined the com-plete coding sequence, splice donor and acceptor


sites, as well as the promoter. Further investigationsinto the levels of PGK1 mRNA expression revealedmarked changes in gene expression that correlatedwith the decreased enzyme activity described in theaffected brothers and two heterozygous women (Ta-ble 2). This observation is supported by previousmeasurements of muscle enzyme activity conductedon the two brothers,1 but enzyme studies could notbe performed on the heterozygous women.

Mutations that underlie a variety of genetic dis-eases, including PGK1 deficiency, are often found tomodify amino acid sequences. These allelic variantsprovide straightforward explanations for the func-tional changes observed. Less obvious is the effect ofmutations within splice donor or splice acceptor sitesthat do not directly alter the protein sequence. Se-quencing of the amplified PGK1 complementaryDNA failed to verify the expression of the mutatedallele in the two heterozygous women. Since thisallele is expressed at a low level in hemizygous men,low expression or a lack of expression could alsoexplain this negative observation in the women. Al-ternatively, expression of PGK1 mRNA with altered

FIGURE 2. Exon-6 mutation localized between two triple-G se-quences. (A) A normal C (GGGCGGG) was observed in thefather as well as in daughter 1 and all controls. (B) HeterozygousN � C/T (GGGC/TGGG) sequence was observed in two carriers(mother and daughter 2). (C) A hemizygous T (GGGTGGG)variant was observed in both affected sons.

FIGURE 3. Quantitative PCR. Competitive PCR involving thepseudogene PGK1P2 (A) and primer amplification of the PGK1mRNA exclusively (B). (A) Based on sequencing, the largestband of 1586 bp was shown to be the pseudogene, whereas theshorter 1304-bp fragment represents the PGK1 transcript. (B)PGK1 mRNA amplified with primers localized to exclude theamplification of pseudogenes (primers 25–26; SupplementaryTable 1 online) resulting in an 1167-bp product. Lane 1: Marker(0.15–2.1 kbp ladder consisting of pBR 328 DNA digested withBgl I and pBR 328 DNA digested with Hinf III from BoehringerMannheim, Germany); lane 2: mother; lane 3: daughter 1; lane 4:daughter 2; lane 5: son 1; lane 6: son 2; lane 7: female control;lane 8: male control.


splicing may explain the clinical observation in thedaughter. However, we were not able to detect anyvariant mRNA in our study. In a recent study byShirakawa et al.,13 an intronic mutation in the PGK1gene resulted in two distinct mRNAs: one normaland one longer fragment containing the 5� region ofintron 7. The adult patient displayed mental retar-dation as well as myoglobinuria without hemolyticanemia.

The mutation in our family with PGK1 defi-ciency, G213G; c.639C�T in exon 6, is localized tothe exon–intron transition. Similar mutations arefound to affect splicing of mRNA, which often resultsin the presence of alternative mRNAs producingproteins with altered or no function.3 The lack ofalternative transcripts in our experiments might beexplained by a high sensitivity to degradation (non-sense-mediated mRNA decay) or failure of the PCRprimers to amplify mRNA of alternative length.3 In acomprehensive study evaluating conserved se-quences within the exon–intron boundary, Cartegniet al.4 published a significant number of sequencesthat are of importance for obtaining normal exonsplicing. The normal GGGCGGG mRNA sequencematches the transcriptional exonic splice enhancerGGGCGGA with a value above threshold, and thenormal CGGAGCT matches the exonic splice en-hancer CGGAGCT in a similar way. However, theadditional mismatch caused by the C-T transitionexcludes binding of the specific serine/arginine-richproteins (SR2/ASF) in both cases (software availableat http://exon.cshl.edu/ESE/). In either case, wemight have a similar situation as PGK Antwerp with apoint mutation located at the end of exon 7.12

The PGK1 gene is localized on the X-chromo-some and, consequently, a single allele is active ineach cell. X-inactivation occurs early in mammalianfemale development, to transcriptionally turn offone of the X chromosomes. Skewed inactivation,however, may result in heterozygous females thatmanifest X-linked diseases usually seen only inmales.2 When the same X chromosome is inactivatedin all somatic stem cells, the somatic cells of a femalewill show the same phenotype as those in a female

homozygote or a male hemizygote. Clinical variationwithin X-linked diseases is common in heterozygouscarriers, and can be extreme, ranging from a normalexpression to complete expression of the defect.This is clearly demonstrated in PGK Munchen,6 andis also a putative explanation for the clinical symp-toms observed in the heterozygous daughter in ourstudy, since her mother lacked obvious clinical symp-toms. A selection for X-inactivation can vary betweentissues; in our study PGK1 mRNA levels were exam-ined exclusively in blood lymphocytes.

Based on our findings, we propose that the exon6 mutation alters PGK1 expression at the level oftranscription. Decreased PGK1 expression is clearlyseen in the two hemizygous brothers, and to a lesserextent in the heterozygous carriers. The altered tran-scription explains the distorted expression of PGK1in leukocytes and the reduced activity of PGK1 inmuscle, leukocytes, and fibroblasts.1 We suggest thatone or two exonic splice enhancer sites are de-stroyed, leading to a less efficient splicing of thePGK1 mRNA transcript.

REFERENCES

1. Aasly J, van Diggelen OP, Boer AM, Brønstad G. Phosphoglyc-erate kinase deficiency in two brothers with McArdle-likeclinical symptoms. Eur J Neurol 2000;7:111–113.

2. Brown CJ. Skewed X-chromosome inactivation: cause or con-sequence? J Natl Cancer Inst 1999;91:304–305.

3. Cartegni L, Chew SL, Krainer AR. Listening to silence andunderstanding nonsense: exonic mutations that affect splic-ing. Nat Rev Genet 2002;3:285–298.

4. Cartegni L, Wang J, Zhu Z, Zhang MQ, Krainer AR. ES-Efinder: a web resource to identify exonic splicing enhancers.Nucleic Acid Res 2003;31:3568–3571.

5. DiMauro S, Dalakas M, Miranda AF. Phosphoglycerate kinasedeficiency: another cause of recurrent myoglobinuria. AnnNeurol 1983;12:11–19.

6. Fujii H, Krietch WKG, Yoshida A. A single amino acid substi-tution (Asp3Asn) in a phosphoglycerate kinase variant (PGKMunchen) associated with enzyme deficiency. J Biol Chem1980;255:6421–6423.

7. Hamano T, Mutoh T, Sugie H, Koga H, Kuriyama M. Phos-phoglycerate kinase deficiency: an adult myopathic form witha novel mutation. Neurology 2000;54:1188–1190.

8. Maeda M, Yoshida A. Molecular defect of a phosphoglyceratekinase variant (PGK-Matsue) associated with hemolytic ane-mia: leu3pro substitution caused by T/A3C/G transition inexon 3. Blood 1991;77:1348–1352.

Table 2. Quantification of RNA.

MotherGly213C/T

Daughter 1Gly213C/C

Daughter 2Gly213C/T

Son 1Gly213T

Son 2Gly213T

Control, femaleGly213C/C

Control, maleGly213C

RNA/DNA ratio 27.9% 81.7% 43.5% 11.1% 14.9% 97.4% 121.2%RNA quantity 32.5% 103.0% 38.2% 7.3% 6.6% 89.2% 107.6%

RNA/DNA ratio in the upper line refers to the estimations carried out with PGK1 mRNA amplification relative to the PGK2 pseudogene. The average value forthe two controls and daughter 1 were used as the standard. Ideally these values should be 100% each, but due to the sensitivity of the method smalldifferences in optical density of the electrophoresis bands will be detected, leading to the individual values deviating somewhat from the theoretical value.


9. Morimoto A, Ueda I, Hirashima Y, Sawai Y, Usuku T, Kano G,et al. A novel missense mutation (1060G3C) in the phospho-glycerate kinase gene in a Japanese boy with chronic haemo-lytic anemia, developmental delay and rhabdomyolysis. Br JHaematol 2003;122:1009–1013.

10. Noel N, Flanagan J, Kalko SG, Bajo MJR, del Mar Manu M,Fuster JLG, et al. Two new phosphoglycerate kinase mutationsassociated with chronic haemolytic anaemia and neurologicaldysfunction in two patients from Spain. Br J Haematol 2005;132:523–529.

11. Ohno K, Tanaka M, Sahasi K, Ibi T, Sato W, Yamamoto T, etal. Mitochondrial DNA deletions in inherited recurrent myo-globinuria. Ann Neurol 1991;29:364–369.

12. Ookawara T, Dave V, Willems P, Martin JJ, de Barsy T, MatthysE, et al. Retarded and aberrant splicing caused by single exon

mutation in a phosphoglycerate kinase variant. Arch BiochemBiophys 1996;327:35–40.

13. Shirakawa K, Takahashi Y, Miyajima H. Intronic mutation inthe PGK1 gene may cause recurent myoglobinuria by abber-ant splicing. Neurology 2006;66:925–927.

14. Singer-Sam J, Keith DH, Tani K, Simmer RL, Shively L, LindsayS, et al. Sequence of the promoter region of the gene for humanX-linked 3-phosphoglycerate kinase. Gene 1984;32:409–417.

15. Spanu C, Oltean S. Familial phosphoglycerate kinase defi-ciency associated with rhabdomyolysis and acute renal failure:abnormality in mRNA splicing? Nephrol Dial Transplant2003;18:445–449.

16. Sugie H, Sukie Y, Ito M, Fukuda T. A novel missense mutation(837T3C) in the phosphoglycerate kinase gene of a patientwith a myopathic form of phosphoglycerate kinase deficiency.J Child Neurol 1998;13:95–97.


ABSTRACT: In order to gain insight into intracellular mechanisms involvedin longitudinal growth of skeletal muscle, we determined gene expression ofubiquitin-ligases (MAFbx/atrogin-1, E3 alpha, and MuRF-1) and deubiquiti-nating enzymes (UBP45, UBP69, and USP28) at different time-points (24,48, and 96 h) of continuous stretch of the soleus and tibialis anterior (TA)muscles. In the soleus, real-time polymerase chain reaction (PCR) showedthat MAFbx/atrogin-1, E3 alpha, and MuRF-1 gene expression was down-regulated, peaking at 24–48 h. Gene expression of all deubiquitinatingenzymes increased with continuous stretch of soleus. In the TA, geneexpression of the ubiquitin-ligases MAFbx/atrogin-1 and MuRF-1 was ele-vated, whereas expression of UBP45 and UBP69 was downregulated.Western blot analysis showed that the overall ubiquitination level decreasedin the soleus and increased in the TA during stretch. These results suggestthat ubiquitin-ligases and deubiquitinating enzymes are involved in longitu-dinal growth induced by continuous muscle stretch.

Muscle Nerve 36: 685–693, 2007

UBIQUITIN-LIGASE AND DEUBIQUITINATING GENEEXPRESSION IN STRETCHED RAT SKELETAL MUSCLE

ANTONIO GARCIA SOARES,1 MARCELO SALDANHA AOKI, PhD,2 ELEN HARUKA MIYABARA, PhD,3

CAMILA VALENTIM DELUCA,1 HELCIO YOGI ONO, DDS,1 MARCELO DAMARIO GOMES, PhD,4 and

ANSELMO SIGARI MORISCOT, PhD1

1 Department of Cell and Developmental Biology, Biomedical Sciences Institute, University of Sao Paulo,Avenida Lineu Prestes 1524, 05508-900 Sao Paulo, SP, Brazil

2 School of Arts, Sciences and Humanities, University of Sao Paulo, Sao Paulo, SP, Brazil3 Faculty of Physical Therapy, University of Sao Paulo City, Sao Paulo, SP, Brazil4 Department of Biochemistry and Immunology, Ribeirao Preto School of Medicine,

Sao Paulo University, Ribeirao Preto, SP, Brazil


Skeletal muscles have a remarkable capacity to ad-just their mass to different stimuli. Mechanical over-load results in radial growth of skeletal muscle (par-allel sarcomere addition),7,30 whereas stretch maylead to serial sarcomere addition, resulting in longi-tudinal growth, especially in more stretchable mus-cles, such as the soleus, due to their anatomicalposition.11 Radial growth depends upon a balancebetween protein synthesis and degradation, and pro-teolysis plays a major role in controlling musclemass.19 Mechanisms controlling protein levels dur-

ing longitudinal growth remain elusive. Informationon the control of longitudinal growth is thereforeimportant to better understand differential mecha-nisms underlying the control of muscle mass.

Calpains13 and components of the proteasomalpathway25 may play important roles in protein deg-radation in skeletal muscle. Accordingly, the protea-somal pathway, an adenosine triphosphate (ATP)–dependent proteolytic complex, is a major regulatorof skeletal muscle wasting in several metabolic con-ditions such as injury, uremia, glucocorticoid treat-ment, sepsis, cancer, trauma, and aging.16 Decreasedmechanical load also activates the proteasomal path-way, as observed in denervation models, joint immo-bilization, and bed rest.3 In order to direct the pro-tein to be degraded in the proteasomal complex, keysteps involve: (1) ubiquitin activation by the E1 fam-ily of enzymes; (2) ubiquitin conjugation by theubiquitin conjugating (E2) family of enzymes; and(3) ubiquitination, performed by the E3 family ofubiquitin-ligases. E1 gene expression appears to beunaffected by stimuli that increase activity of theproteasomal pathway. It is speculated that upregula-

Abbreviations: ANOVA, analysis of variance; ATP, adenosine triphosphate;Ct, cycle threshold; E1, enzyme ubiquitin activating; E3, enzyme ubiquitinligating; E3 alpha, N-end rule ligase; FOXO, forkhead transcription factor;IGF-1, insulin-like growth factor-1; MAFbx, muscle atrophy F-box; MuRF-1,muscle RING-finger 1; MyoD, myogenic differentiation factor; PCR, polymer-ase chain reaction; PI3K-AKT, phosphatidylinositol-3-kinase–protein kinaseB; S6K1, ribosomal S6 kinase 1; TA, tibialis anterior; UBP, ubiquitin-bindingprotease; USP, ubiquitin-specific proteasesKey words: deubiquitinating enzyme; muscle stretch; sarcomere; skeletalmuscle; ubiquitin-ligaseCorrespondence to: A. S. Moriscot; e-mail: [email protected]


Ubiquitin–Proteasome in Stretched Skeletal Muscle MUSCLE & NERVE November 2007 685

tion may be unnecessary because E1 protein is con-stitutively expressed at high levels and is not thoughtto regulate pathway specificity. Multiple ubiquitinmolecules are conjugated with target proteins andtagged by E3 enzymes in a substrate- and tissue-specific manner.25 In contrast to E1 enzymes, geneexpression of E3 enzymes is highly modulated inskeletal muscle under several conditions, especiallyin disuse, sepsis, and starvation.17

Recent findings have highlighted an importantrole for enzymes able to remove ubiquitin from ed-ited proteins, the so-called UBPs (ubiquitin-bindingproteases).20 UBPs have been implicated in a num-ber of biological processes such as cancer,4 neuronaldiseases, development, and genome integrity, al-though the physiological functions of most of theseenzymes remain largely unknown.35 Deubiquitinat-ing enzymes may be involved in ubiquitin removalfrom substrates, sparing them from degradation.6Deubiquitinating enzymes, such as UBP6,18 ubiq-uitin C-terminal hydrolase,15 and Rpn11,32 are asso-ciated with the 26S proteasome, corroborating with acoordinated role in protein ubiquitin editing.

To our knowledge, only one study has been con-ducted to address the role of UBPs in skeletal musclemass maintenance using catabolic models (fasting,streptozotocin-induced diabetes, dexamethasonetreatment, and cancer)5 and no studies have ad-dressed UBP gene expression in disuse models andskeletal muscle stretch. Although the E3 family playsa key role in skeletal muscle mass loss,3 the balancebetween ubiquitination and deubiquitination is stilluncertain in skeletal muscle under mechanical stim-uli, including overload and stretch. Therefore, thepresent study was undertaken to determine geneexpression of ubiquitin-ligases and UBPs during skel-etal muscle longitudinal growth induced by stretch-ing. We focused on gene expression of the E3 en-zymes MAFbx (muscle atrophy F-box)/atrogin-1, E3alpha, and MuRF-1 (muscle RING-finger 1), whichare highly expressed in skeletal muscle and seem toplay important functional roles in skeletal musclegrowth regulation. MAFbx/atrogin-1 is robustlyoverexpressed in atrophy models induced by meta-bolic conditions such as food deprivation and ure-mia.10 E3 alpha is clearly involved in skeletal musclecatabolism in models of cancer–cachexia.14 MuRF-1is related to degradation of master proteins involvedin skeletal muscle cellular specification.29 Consider-ing that the regulation of deubiquitinating enzymegene expression is largely unknown in skeletal mus-cle, we chose to study deubiquitinating genes thatare relatively well expressed in this tissue (UBP45,UBP69, and USP28).22,31


All protocols described in this study conformed tothe Ethical Principles in Animal Research followedby the Brazilian College of Animal Experimentationand were approved by our institutional review board.

Male Wistar rats (6–8 weeks old; 220 � 5 g bodyweight) obtained from the animal breeding facilityat our institution were used for this study. The ani-mals were divided into four groups at random andhoused in cages under a 12-h light–dark cycle withfood and water ad libitum.

Rats were anesthetized with ketamine and xyla-zine (30 mg/kg and 10 mg/kg, respectively) duringthe immobilization procedure. In order to study so-leus and TA, different immobilization procedureswere performed. The first procedure involved fixingthe left hindlimb in total dorsiflexion and wasachieved by casting the limb. The soleus muscle waskept immobilized for three different times (24, 48,and 96 h). Intact animals were used as control andrepresent the group at 0 h. The second procedureinvolved casting the left hindlimb in total plantarextension. Care was taken to ensure that the cast didnot cause ischemia.

Determination of Sarcomere Number and Length. At24, 48, and 96 h after the beginning of the experi-ments all animals were anesthetized and the soleusand TA muscles were dissected free from surround-ing tissue. The tendons of these muscles wereclamped with the ankle joint in a 90° position, whichwas considered to be Lo (initial length). Then eachsoleus and TA muscle was divided longitudinally intotwo similar parts with a surgical knife: the medialpart was removed and used for gene expression anal-ysis, whereas the lateral portion was fixed with glu-taraldehyde in situ in order to determine sarcomeremeasurements.

The method used to determine the number andlength of sarcomeres along a single muscle fiber wasdeveloped by Williams and Goldspink.33 The musclewas fixed for 3 h in 2.5% glutaraldehyde and thenremoved, placed in 30% HNO3 for 2 days, and sub-sequently stored in 50% glycerol.

Five single muscle fibers of each whole soleus andTA muscle were teased out from tendon to tendonand mounted in glycerin. Sarcomere number andlength along a 300-�m distance were quantified atdifferent points of the middle region of each singlefiber using a projection microscope. The total num-ber of sarcomeres in each single muscle fiber wasdetermined by the correlation between the numberof sarcomeres identified along a distance of 300 �m

686 Ubiquitin–Proteasome in Stretched Skeletal Muscle MUSCLE & NERVE November 2007

and the total fiber length.34 There is a certain varia-tion in the number and length of sarcomeres alongthe muscle fiber, but in the present study it wasassumed that sarcomere length along the entirelength of the muscle fiber was homogeneous.21

Gene Expression Analysis. Muscles were quickly re-moved, frozen in liquid nitrogen, and stored at�70°C. Total RNA was isolated from soleus and TAsamples using Trizol (Invitrogen, Carlsbad, Califor-nia) following the manufacturer’s recommenda-tions. Concentrations of RNA were determined bymeasuring absorbance at 260 nm. The purity of theRNA was determined by calculating the absorbanceratio at 260 nm and 280 nm, and by ethidium bro-mide staining. Isolated RNA was stored at �70°Cuntil analysis by real-time polymerase chain reaction(PCR).

For reverse transcription, 1 �g of total RNA wastypically used in a reaction containing oligo-dT (500�g/ml), 10 mM of each dNTP, 5� first-strandbuffer, 0.1 M dithiothreitol, and 200 U of reversetranscriptase (SuperScript II, Invitrogen). Reversetranscription was performed at 70°C for 10 min,followed by incubation at 42°C for 60 min and at95°C for 10 min.

Primer sets for rat MAFbx/atrogin-1, E3 alpha,UBP45, UBP69, and USP28 were designed usingPrimer Express software v2.0 (Applied Biosystems,Foster City, California). Primer sequences forMuRF-1 were obtained from Wray et al.38 and tran-scription factor IID (TFIID) from Stockholm et al.27

All primers were synthesized by IDT (Coralville,Iowa) (Table 1).

For each sample, PCR was performed in dupli-cate in a 25-�l reaction volume of 5–20 ng of cDNA,12.5 �l Syber Green Master Mix (Applied Biosys-tems), and 50–200 nM of each primer. PCR analyseswere carried out using the following cycle parame-ters: 50°C for 2 min, 95°C for 10 min, followed by 40cycles of 95°C for 15 s, and 60°C for 1 min. Fluores-

cence was quantified and analyses of amplificationplots were performed with the ABI Prism 5700 Se-quence Detector System (Applied Biosystems). Re-sults were expressed using the comparative cyclethreshold (Ct) method as described by the manufac-turer. The �Ct values were calculated in every sam-ple for each gene of interest as Ctgene of interest minusCtreporter gene; TFIID was the control gene. The cal-culation of the relative changes in the expressionlevel of one specific gene (��Ct) was performed bysubtraction of the �Ct from the control group (0 h,used as a calibrator) to the corresponding �Ct (atdifferent times) from the stretch groups (24, 48, and96 h). The values and ranges given in Figures 1 and2 were determined as follows: 2��Ct with ��Ct, andthe control levels were arbitrarily set to 1.

Western Blot Analysis. Soleus and TA muscles werehomogenized in RIPA buffer (0.625% Nonidet P-40,0.625% sodium deoxycholate, 6.25 mM sodiumphosphate, and 1 mM ethylene-diamine tetraaceticacid at pH 7.4) containing 10 �g/ml of proteaseinhibitor cocktail (Sigma-Aldrich, St. Louis, Mis-souri). Homogenates were centrifuged at 12,000 gfor 10 min at 4°C, the supernatant was saved, andprotein concentration was determined by Bradfordassay (Bio-Rad, Hercules, California) with bovine se-rum albumin as a reference.

Equal amounts of protein (100 �g) were runon 7.5% sodium dodecylsulfate–polyacrylamidegel electrophoresis (SDS-PAGE) and transferredto a polyvinylidene fluoride membrane (Immo-bilon-P, Millipore, Bedford, Massachusetts). Themembranes were stained with Ponceau S to con-firm the protein amount and then rinsed withTween Tris-buffered saline solution (0.5 M NaCl,50 mM Tris-HCl at pH 7.4, and 0.1% Tween 20).All membranes were incubated with Super BlockBuffer (Pierce, Rockford, Illinois) and 5% nonfatdried milk for 2 h at 4°C. Membranes were thenincubated overnight with rabbit polyclonal anti-

Table 1. Primers used in real-time PCR.

Primers Forward Reverse

MAFbx/atrogin-1 TACTAAGGAGCGCCATGGATACT GTTGAATCTTCTGGAATCCAGGATMuRF-1 TGACCAAGGAAAACAGCCACCAG TCACTCCTTCTTCTCGTCCAGGATGGE3 alpha AGCACAGTGCAGCATTTCAGTT TTGCACAATCCCTACAGGAATATGTUBP45 CAGCATGCGTACCTCCTACACC ACTCTTTGAATTCTTGGCTTTGTTGAUBP69 CCGGACACAGCCCATGAG GTAGCGGGACGATTCTGTATAGCUSP28 AAAGGCCAGTAATGGTGACATCA GTCGTGACTGGGCTCCTTAACTTFIID ACGGACAACTGCGTTGATTTT ACTTAGCTGGGAAGCCCAAC

Accession codes for MAFbx/atrogin1, MuRF-1, E3alpha, UBP45, UBP69, USP28, and TFIID genes are AF441120, AY059627, AF061555, AF106658,AF106659, XM_236240, and D01034, respectively.


body against ubiquitin (BostonBiochem, Cam-bridge, Massachusetts) at 4°C. The antibody wasdiluted 1:1500 in 5% nonfat dried milk and 0.1%Tween 20 in 0.1 M phosphate-buffered saline(PBS, pH 7.4). After a 30-min wash in TweenTris-buffered saline solution, membranes were in-cubated with horseradish peroxidase– conjugatedantibody against rabbit immunoglobulin G for 2 hat room temperature and diluted 1:2000 in 5%nonfat dried milk with Tween Tris-buffered saline.Detection of the labeled proteins was done usingthe enhanced chemiluminescence system (ECL,Amersham, UK).

Statistical Analysis. One-way analysis of variance(ANOVA) following Tukey’s multiple comparison orStudent–Newman–Keuls tests was used to comparemore than two groups. For all comparisons, P � 0.05was considered statistically significant.

RESULTS

Sarcomere Number and Length in Stretched Muscles.

Soleus muscles immobilized in a lengthened posi-tion for 48 and 96 h, but not 24 h, had their sarco-mere number significantly increased compared withcontrol (�19% and �29%, respectively; P � 0.05;

FIGURE 1. Gene expression of MAFbx/atrogin-1 (A), E3� (B), and MuRF-1 (C) in the soleus and TA muscles, respectively, of control (0h) and groups stretched for different periods (24, 48, and 96 h). Data are expressed as mean � SE; n � 5 per group; *P � 0.05 vs. control(one-way ANOVA followed by Tukey’s procedure for multiple comparisons).


Table 2). Sarcomere length was significantly reducedin both the 48- and 96-h groups compared withcontrol (�15% and �21%, respectively; P � 0.05;Table 2), whereas it was unchanged in the 24-hgroup.

Sarcomere number and length in TA musclewere unaltered at all times evaluated as comparedwith control (Table 2).

Ubiquitin-Ligase Gene Expression. Our data showthat stretch substantially represses gene expressionof all three investigated components of the ubiquit-in–proteasome system (MAFbx/atrogin-1, E3 alpha,and MuRF-1) in the soleus muscle. MAFbx/atrogin-1and E3 alpha mRNA levels decreased drastically after24-h stretch (Fig. 1A and B). The expression of thesetwo genes in the soleus muscle was partially recov-ered after 48- and 96-h stretch (Fig. 1A and B).MuRF-1 mRNA levels also decreased after 24-hstretch (Fig. 1C). Repression of MuRF-1 mRNA lev-els remained around 60% up to 48 h and returned tobasal levels at 96-h stretch (Fig. 1C).

Gene expression of MAFbx/atrogin-1, E3 alpha,and MuRF-1 was also determined in the TA muscle.After continuous stretch for 48 h, MAFbx/atrogin-1gene expression increased 3.2-fold compared withcontrol, returning to basal levels at 96 h (Fig. 1A). E3alpha gene expression decreased after 24-h stretch(Fig. 1B) and remained at 96 h (Fig. 1B). MuRF-1gene expression in the TA muscle decreased to 46%at 24 h and increased 2.7-fold at 48 h compared withcontrol (Fig. 1C). At the end of the stretch protocol,MuRF-1 gene expression returned to basal levels(Fig. 1C).

Deubiquitin Gene Expression. In addition to ubiq-uitin-ligase gene expression we also addressed hall-

mark genes coding for proteins of the deubiquitinfamily (UBP45, UBP69, and USP28).

UBP45 mRNA levels in the soleus muscle re-mained unchanged after 24 h and, after 48 and 96 hof stretch, these values increased 2.9- and 3.8-fold,respectively (P � 0.05), compared with control (Fig.2A). Interestingly, UBP69 mRNA levels in the soleusfollowed the same tendency compared with UBP45mRNA during stretch, although the increase inmRNA levels was less pronounced (Fig. 2B). Forty-eight hours of stretch resulted in a 1.8-fold increase(P � 0.05) in UBP69 mRNA levels compared withcontrol (Fig. 2B) and remained at these levels at 96 hof stretch (P � 0.05; Fig. 2B). USP28 mRNA levels inthe soleus muscle increased significantly duringstretch at all time-points addressed in this study,although only modestly (Fig. 2C).

UBP45 gene expression in the TA muscle wasclearly reduced at all times of continuous stretch(Fig. 2A). Similarly, UBP69 gene expression in theTA decreased at all time-points evaluated (Fig. 2B).Finally, USP28 gene expression was not alteredthroughout continuous stretch in the TA (Fig. 2C).

Ubiquitin–Protein Conjugate Expression. We also de-termined the expression of ubiquitinated proteins inthe soleus and TA muscles of control and stretchgroups by Western blot analysis (Fig. 3). The blotsshowing ubiquitin–protein conjugates in the soleus(Fig. 3B) and TA muscles (Fig. 3D) are displayednext to the respective membranes stained with Pon-ceau S (Fig. 3A and C). In this preparation, ubiqui-tinated proteins are particularly visible at higher mo-lecular weights as a smear (Fig. 3B and D). Ourresults show that at 48- and 96-h stretch the ubiquit-in–protein conjugate expression in the soleus mus-cles decreased when compared with their controls(Fig. 3B). Interestingly, the expression of ubiquitin–protein conjugates in the TA muscles significantlyincreased at 24 h and up to 48 h compared withcontrol (Fig. 3D).

DISCUSSION

The mechanical stimulus induced by stretchingdrives rapid and significant muscle mass gain, result-ing in sarcomere addition on preexisting myofibrils,so-called longitudinal growth, particularly in morestretchable muscles due to their anatomical position,such as the soleus. Longitudinal growth does notresult in increased force but leads to augmentedjoint range-of-motion. This is important in prevent-ing osteoarticular injury, particularly in sports involv-ing bouncing and jumping activities with a high

Table 2. Sarcomere number and length (�m) at different times ofcontinuous stretch of the soleus and TA muscles.

Muscles GroupsSarcomere

numberSarcomerelength (�m)

Soleus 0 h 6338 � 120.5 3.27 � 0.224 h 6858 � 590.8 3.23 � 0.348 h 7534 � 900.4* 2.79 � 0.4*96 h 8174 � 740.7* 2.60 � 0.3*

TA 0 h 9013 � 260.8 2.17 � 0.0724 h 9057 � 220.3 2.21 � 0.1148 h 9146 � 150.6 2.19 � 0.0796 h 9097 � 200.6 2.20 � 0.10

Values expressed as mean � SD; n � 6 or 7. TA, tibialis anterior muscle.*P � 0.05 vs. matched control (one-way ANOVA followed by Tukey’sprocedure for multiple comparisons).


intensity of stretch–shortening cycles, such as soc-cer.37 Longitudinal growth is also important in thereestablishment of functional status after certain in-juries. Stretch strongly drives protein accumulationand also increases expression of growth factors suchas insulin-like growth factor-1 (IGF-1).2,9 Inhibitionof mTOR (mammalian target of rapamycin) by rapa-mycin significantly reduces longitudinal growth in-duced by stretching.2 Nitric oxide plays a central rolein longitudinal growth induced by stretching.28

Other cellular and molecular mechanisms probablyalso play important roles in longitudinal growth. Theproteasomal system is a good candidate, consideringthat longitudinal growth is associated with protein

accumulation, and the proteasomal system is a majorregulator of protein degradation in skeletal muscle.Understanding the cellular processes that controlprotein accumulation in such circumstances is notonly important for understanding longitudinalgrowth itself but also for clarifying the differentialmechanisms governing parallel and serial sarcomereincorporation.

Our first goal in the present study was to testwhether the stretch procedure adopted was success-ful in increasing sarcomere number in the soleus. Asexpected, continuous stretch of the soleus muscleresulted in an increase in sarcomere number as earlyas 48 h after onset of the stretch procedure. Sarco-

FIGURE 2. Gene expression of UBP45 (A), UBP69 (B), and USP28 (C) in the soleus and TA muscles, respectively, of control (0 h) andgroups stretched for different periods (24, 48, and 96 h). Data are expressed as mean � SE; n � 5 per group; *P � 0.05 vs. control(one-way ANOVA followed by Tukey’s procedure for multiple comparisons).


mere number continued to increase for up to 96 hafter the onset of stretch compared with unstretchedcontrols (Table 2). Increase in sarcomere numbercoincided with a decrease in sarcomere length be-cause we observed a significant reduction in sarco-mere length at 48 and 96 h after onset of the stretchprotocol (Table 2). Therefore, the stretch protocolused in the present study is suitable to address theeffects of stretch on gene expression in the soleus.We also studied the TA muscle, which is not maxi-mally stretched due to its anatomical position. Theresults show that neither sarcomere number norsarcomere length changed significantly with thisstretch protocol.

In the soleus, at least three ubiquitin-ligases havetheir gene expression robustly and rapidly down-regulated by stretch (MAFbx/atrogin-1, E3 alpha,and MuRF-1). In addition, two deubiquitinating en-zymes are significantly upregulated (UBP45 andUBP69) in a similar time frame (gene expression ofubiquitin-ligases is modulated slightly earlier thandeubiquitinating enzymes), raising the possibilitythat these proteasome-associated proteins work incoordination to accumulate protein in stretchedskeletal muscle. The decrease in ubiquitin-ligase

gene expression was transient (nadir occurs within24 h after stretch) and preceded the increase insarcomere number (48–96 h after stretch), suggest-ing an adaptive mechanism resulting in a pulse ofprotein accumulation. However, the presence ofmultiple proteasome-associated enzymes expressedin the same tissue suggests that specific substratesmay be targeted selectively under various pathophys-iological conditions. In fact, several muscle-specificsubstrates have been recently identified for MAFbx/atrogin-129 and MuRF-1.36

It is still unclear whether the E3 substrates iden-tified in vitro occur in vivo and whether these sub-strates can be ubiquitinated selectively. At present,few substrates for E3s have been well characterizedfunctionally. For example, MyoD, a myogenic tran-scription factor, can be spared from selective ubiq-uitination when interacting with promoters locatedin target genes. Therefore, in situations whereMyoD/DNA interactions decrease (i.e., disuse),MyoD would be more vulnerable to degradation bythe proteasomal system.1 Similarly, selective ubiquiti-nation regulates the activity of the transcriptionfactor NF-�B (nuclear factor–kappaB) in skeletalmuscle during disuse.12 Specific targets for deubiq-

FIGURE 3. Representative Western blots of ubiquitin–protein conjugates in control (0 h) and stretched soleus (B) and TA (D) musclesat 24, 48, and 96 h are shown on the right, with respective membranes stained with Ponceau S on the left (A, C) for soleus and TA,respectively. Muscle samples from control (0 h) and stretched groups (24, 48, and 96 h) are identified at the top of each membrane/blot.Mw, molecular weight.


uitinating enzymes and their biological significancein skeletal muscle are even more elusive. To ourknowledge, the only example is an antagonistic roleof splice variants of UBP69 and UBP45 in myogen-esis. Both myoblast fusion and myosin heavy-chainaccumulation are stimulated by the stable expressionof UBP69 and antagonized by forced expression ofUBP45, highlighting putative differential roles ofproteasome-related enzymes.22

Considering the limited knowledge of targets forproteasome-related enzymes in vivo in skeletal mus-cle, interpretations of the biological significance ofthe observed changes in gene expression are limited.The identification and characterization of sets ofproteins targeted by the proteasomal system in vivounder different pathophysiological conditions wouldhelp to increase knowledge of the molecular biologyof skeletal muscle.

Although it is currently uncertain how skeletalmuscle stretch may downregulate E3s and simulta-neously upregulate most UBPs, IGF-1 is a good can-didate as a regulatory factor. Continuous stretchleads to a rapid and robust increase in local IGF-18

and, more recently, the effects of IGF-1 on musclegrowth have been attributed to activation ofthe phosphatidylinositol-3-kinase–protein kinase B(PI3K-AKT) pathway.23 This pathway is involved inthe phosphorylation of translation factors such asS6K1 (ribosomal S6 kinase 1), driving increased pro-tein synthesis and FOXO (forkhead transcriptionfactor) phosphorylation, inactivating its ability totransactivate target genes such as MAFbx/atrogin-1and MuRF-1.24,25 Thus, an important component ofprotein accumulation in stretched skeletal musclemay involve the transcriptional suppression of atro-genes via IGF-1, decreasing the degradation of myo-fibrillar proteins. It remains to be determinedwhether the PI3K-AKT pathway also regulates geneexpression of UBPs.

In order to investigate whether stretch leads tochanges in the overall level of protein ubiquitinationin the stretched muscles addressed in this work, weused an antibody specific for ubiquitinated proteinsin a Western blot approach. The results show a de-crease in the level of ubiquitinated proteins in thesoleus, more easily identified at high molecularweights (Fig. 3). Therefore, the coordinated modu-lation of ubiquitin ligases and deubiquitinases in thesoleus leads to a net decrease in the level of ubiqui-tinated proteins, which is likely involved in proteinaccumulation during stretch.

When we analyzed ubiquitin and deubiquitingene expression in the stretched TA, ubiquitin li-gases were mostly upregulated and deubiquitinases

were mostly downregulated, in contrast to the re-sponses obtained in the soleus. These results, com-bined with the observation that overall levels of ubi-quitinated proteins in TA increased starting at 48 hof stretch and that sarcomere number was unalteredby the stretching protocol, suggest that the immobi-lization effect overcomes the stretch effect in the TA,likely due to differences in the nature of anatomicalcharacteristics of the musculoskeletal units crossingthe ankle joint.26 In fact, the results also suggest thatmechanical stress is the main factor for sustaininglongitudinal growth, although decreased neural in-put into the immobilized muscles may also play arole in limiting longitudinal growth.

This work was supported by FAPESP (Fundacao de Amparo aPesquisa do Estado de Sao Paulo, Brazil).

REFERENCES

1. Abu Hatoum O, Gross-Mesilaty S, Breitschopf K, Hoffman A,Gonen H, Ciechanover A, et al. Degradation of myogenictranscription factor MyoD by the ubiquitin pathway in vivoand in vitro: regulation by specific DNA binding. Mol CellBiol 1998;18:5670–5677.

2. Aoki MS, Miyabara EH, Soares AG, Saito ET, Moriscot AS.mTOR pathway inhibition attenuates skeletal muscle growthinduced by stretching. Cell Tissue Res 2006;324:149–156.

3. Bodine SC, Latres E, Baumhueter S, Lai VK, Nunez L, ClarkeBA, et al. Identification of ubiquitin ligases required for skel-etal muscle atrophy. Science 2001;294:1704–1708.

4. Chauhan D, Hideshima T, Anderson KC. Proteasome inhibi-tion in multiple myeloma: therapeutic implication. Annu RevPharmacol Toxicol 2005;45:465–476.

5. Combaret L, Adegoke OA, Bedard N, Baracos V, Attaix D,Wing SS. USP19 is a ubiquitin-specific protease regulated inrat skeletal muscle during catabolic states. Am J Physiol En-docrinol Metab 2005;288:E693–E700.

6. Curcio-Morelli C, Zavacki AM, Christofollete M, Gereben B,de Freitas BC, Harney JW, et al. Deubiquitination of type 2iodothyronine deiodinase by von Hippel–Lindau protein-in-teracting deubiquitinating enzymes regulates thyroid hor-mone activation. J Clin Invest 2003;112:189–196.

7. Glass DJ. Skeletal muscle hypertrophy and atrophy signalingpathways. Int J Biochem Cell Biol 2005;37:1974–1984.

8. Goldspink DF, Cox VM, Smith SK, Eaves LA, Osbaldeston NJ,Lee DM, et al. Muscle growth in response to mechanicalstimuli. Am J Physiol 1995;268:E288–E297.

9. Goldspink G, Yang SY. Effects of activity on growth factorexpression. Int J Sport Nutr Exerc Metab 2001;11(suppl):S21–S27.

10. Gomes MD, Lecker SH, Jagoe RT, Navon A, Goldberg AL.Atrogin-1, a muscle-specific F-box protein highly expressedduring muscle atrophy. Proc Natl Acad Sci USA 2001;98:14440–14445.

11. Griffin GE, Williams PE, Goldspink G. Region of longitudinalgrowth in striated muscle fibres. Nat New Biol 1971;232:28–29.

12. Hunter RB, Stevenson E, Koncarevic A, Mitchell-Felton H,Essig DA, Kandarian SC. Activation of an alternative NF-kappaB pathway in skeletal muscle during disuse atrophy.FASEB J 2002;16:529–538.

13. Kramerova I, Kudryashova E, Venkatraman G, Spencer MJ.Calpain 3 participates in sarcomere remodeling by actingupstream of the ubiquitin–proteasome pathway. Hum MolGenet 2005;14:2125–2134.


14. Kwak KS, Zhou X, Solomon V, Baracos VE, Davis J, BannonAW, et al. Regulation of protein catabolism by muscle-specificand cytokine-inducible ubiquitin ligase E3alpha-II duringcancer cachexia. Cancer Res 2004;64:8193–8198.

15. Lam YA, Xu W, DeMartino GN, Cohen RE. Editing of ubiq-uitin conjugates by an isopeptidase in the 26S proteasome.Nature 1997;385:737–740.

16. Lecker SH, Solomon V, Mitch WE, Goldberg AL. Muscleprotein breakdown and the critical role of the ubiquitin–proteasome pathway in normal and disease states. J Nutr1999;129(suppl):227S–237S.

17. Lecker SH, Jagoe RT, Gilbert A, Gomes M, Baracos V, BaileyJ, et al. Multiple types of skeletal muscle atrophy involve acommon program of changes in gene expression. FASEB J2004;18:39–51.

18. Leggett DS, Hanna J, Borodovsky A, Crosas B, Schmidt M,Baker RT, et al. Multiple associated proteins regulate protea-some structure and function. Mol Cell 2002;10:495–507.

19. Li JB, Goldberg AL. Effects of food deprivation on proteinsynthesis and degradation in rat skeletal muscles. Am J Physiol1976;231:441–448.

20. Liu S, An H, Li N, Yu Y, Lin N, Wan T, et al. Cloning andidentification of a novel human ubiquitin-like protein, DC-UbP, from dendritic cells. Biochem Biophys Res Commun2003;300:800–805.

21. Matano T, Tamai K, Kurokawa T. Adaptation of skeletal mus-cle in limb lengthening: a light diffraction study on the sar-comere length in situ. J Orthop Res 1994;12:193–196.

22. Park KC, Kim JH, Choi EJ, Min SW, Rhee S, Baek SH, et al.Antagonistic regulation of myogenesis by two deubiquitinat-ing enzymes, UBP45 and UBP69. Proc Natl Acad Sci USA2002;99:9733–9738.

23. Rommel C, Bodine SC, Clarke BA, Rossman R, Nunez L, StittTN, et al. Mediation of IGF-1-induced skeletal myotube hy-pertrophy by PI(3)K/Akt/mTOR and PI(3)K/Akt/GSK3pathways. Nat Cell Biol 2001;3:1009–1013.

24. Sacheck JM, Ohtsuka A, McLary SC, Goldberg AL. IGF-Istimulates muscle growth by suppressing protein breakdownand expression of atrophy-related ubiquitin ligases, atrogin-1and MuRF1. Am J Physiol Endocrinol Metab 2004;287:E591–E601.

25. Sandri M, Sandri C, Gilbert A, Skurk C, Calabria E, Picard A,et al. Foxo transcription factors induce the atrophy-relatedubiquitin ligase atrogin-1 and cause skeletal muscle atrophy.Cell 2004;117:399–412.

26. Spector SA, Simard CP, Fournier M, Sternlicht E, EdgertonVR. Architectural alterations of rat hind-limb skeletal musclesimmobilized at different lengths. Exp Neurol 1982;76:94–110.

27. Stockholm D, Herasse M, Marchand S, Praud C, Roudaut C,Richard I, et al. Calpain 3 mRNA expression in mice afterdenervation and during muscle regeneration. Am J PhysiolCell Physiol 2001;280:C1561–C1569.

28. Tatsumi R, Liu X, Pulido A, Morales M, Sakata T, Dial S, et al.Satellite cell activation in stretched skeletal muscle and therole of nitric oxide and hepatocyte growth factor. Am JPhysiol Cell Physiol 2006;290:C1487–C1494.

29. Tintignac LA, Lagirand J, Batonnet S, Sirri V, Leibovitch MP,Leibovitch SA. Degradation of MyoD mediated by the SCF(MAFbx) ubiquitin ligase. J Biol Chem 2005;280:2847–2856.

30. Tsika RW, Herrick RE, Baldwin KM. Time course adaptationsin rat skeletal muscle isomyosins during compensatory growthand regression. J Appl Physiol 1987;63:2111–2121.

31. Valero R, Bayes M, Francisca Sanchez-Font M, Gonzalez-An-gulo O, Gonzalez-Duarte R, Marfany G. Characterization ofalternatively spliced products and tissue-specific isoforms ofUSP28 and USP25. Genome Biol 2201;2:43.1–43.10.

32. Verma R, Aravind L, Oania R, McDonald WH, Yates JR III,Koonin EV, et al. Role of Rpn11 metalloprotease in deubiqui-tination and degradation by the 26S proteasome. Science2002;298:611–615.

33. Williams PE, Goldspink G. Longitudinal growth of striatedmuscle fibres. J Cell Sci 1971;9:751–767.

34. Williams PE, Catanese T, Lucey EG, Goldspink G. The impor-tance of stretch and contractile activity in the prevention ofconnective tissue accumulation in muscle. J Anat 1988;158:109–114.

35. Wing SS. Deubiquitinating enzymes—the importance of driv-ing in reverse along the ubiquitin–proteasome pathway. IntJ Biochem Cell Biol 2003;35:590–605.

36. Witt SH, Granzier H, Witt CC, Labeit S. MURF-1 and MURF-2target a specific subset of myofibrillar proteins redundantly:towards understanding MURF-dependent muscle ubiquitina-tion. J Mol Biol 2005;350:713–722.

37. Witvrouw E, Mahieu N, Danneels L, McNair P. Stretching andinjury prevention: an obscure relationship. Sports Med 2004;34:443–449.

38. Wray CJ, Mammen JM, Hershko DD, Hasselgren PO. Sepsisupregulates the gene expression of multiple ubiquitin ligasesin skeletal muscle. Int J Biochem Cell Biol 2003;35:698–705.


ABSTRACT: The reflex torque responses of the elbow and shoulder toconstant velocity angular extensions of the full comfortable range of thespastic elbow were measured in 16 people with unilateral stroke and 6neurologically intact controls in order to identify the interjoint reflex couplingthat occurs after stroke. The resulting responses showed a substantial reflextorque at the elbow and shoulder in subjects with stroke, with 12 of the 16subjects producing adduction of the shoulder in response to passive exten-sion of the elbow. The presence of simultaneous shoulder flexion torque withelbow flexion torque and with an identical waveform indicated an active roleof biarticular elbow/shoulder flexors, such as the biceps. As the bicepsmuscle produces a shoulder abduction moment, shoulder adduction pro-duced during elbow extension was thought to be associated with neuralrather than biomechanical coupling. These results suggest that spasticity inpeople with stroke is more complex than its traditional perception as ahyperexcitable stretch reflex, and includes potent heteronymous reflex path-ways. The reflex coupling observed between the shoulder and elbow shouldbe considered in the diagnosis and clinical management of spasticity. Thepotential impact of this reflex on the coordination of volitional arm move-ments will be examined in future studies.

Muscle Nerve 36: 694–703, 2007

MULTIJOINT REFLEXES OF THE STROKE ARM: NEURALCOUPLING OF THE ELBOW AND SHOULDER

SAMIR G. SANGANI, BS,1 ANDREW J. STARSKY, MPT,2

JOHN R. MCGUIRE, MD,3 and BRIAN D. SCHMIT, PhD1

1 Department of Biomedical Engineering, Marquette University, P.O. Box 1881,Milwaukee, Wisconsin 53201-1881, USA

2 Department of Physical Therapy, Marquette University, Milwaukee, Wisconsin, USA3 Department of Physical Medicine and Rehabilitation, Medical College of Wisconsin, Milwaukee, Wisconsin, USA


The purpose of this study was to identify reflextorque coupling of the shoulder to imposed move-ments of the elbow in individuals who had had astroke. Clinical observations suggest that spasticityis not necessarily a single joint phenomenon, withimposed movements of a joint producing reflexactivity throughout the arm. As a result, we postu-lated that reflex coupling of the elbow and shoul-der, if it exists, could contribute to problems witharm posture and movement in people after stroke.Stretch reflex activation of muscles that cross theelbow could produce joint torques at the shoulderthrough biomechanical coupling of biarticular el-bow/shoulder muscles21 or by neural couplingthrough heteronymous reflex activation of shoul-

der muscles by elbow muscle afferents.20,29 Wepostulated that imposed movements of the elbowwould produce shoulder torques consistent withboth biomechanical and neural coupling mecha-nisms. These effects may have important conse-quences in the quantification of spasticity and maypredict coupling of the elbow and shoulder duringfunctional movements of the arm.

Assessment of spasticity typically focuses on hom-onymous, single-joint stretch reflexes. Specifically,spasticity has been defined as a disorder of the sen-sorimotor system characterized by a velocity-depen-dent increase in tonic stretch reflexes with exagger-ated tendon jerks.27 Consequently, clinical scales forspasticity assess the resistance to a manually imposedmovement of the same joint.3 Biomechanical mea-surements of spasticity expand on these concepts,using a motor to impose a movement to the jointwhile measuring the torque or electromyographic(EMG) responses of the same joint.32,33 Althoughhyperexcitability of the homonymous stretch re-flexes has been documented, heteronymous reflex

Abbreviations: ANOVA, analysis of variance; EMG, electromyographic;MANOVA, multivariate analysis of varianceKey words: joint torque; neural coupling; spasticity; stretch reflex; strokeCorrespondence to: B. D. Schmit; e-mail: [email protected]


694 Reflexes of Stroke Arm MUSCLE & NERVE November 2007

effects on other joints have been measured infre-quently in people after stroke. Heteronymous reflexcoupling of muscles that cross the elbow and shoul-der could have important implications for voluntarycontrol of arm movements in people with stroke.

Coupled activation of muscles crossing differentjoints is a phenomenon that is commonly associatedwith volitional movements in people with chronicstroke. Further, groupings of muscles into functionalunits, often termed “muscle synergies,” can result inabnormal muscle activation patterns during voli-tional movements.9 The synergy relations of the el-bow and shoulder have been measured in detail afterstroke6,8,11; most notably, volitional efforts to extendthe elbow result in activation of shoulder adductorsand, conversely, voluntary elbow flexion producessynergistic shoulder abduction.14 Although themechanism of these couplings is unknown, proprio-spinal pathways that integrate descending com-mands and proprioceptive feedback may be in-volved.30 These same pathways could producestretch reflex coupling of the elbow and shoulder.

In this study we hypothesized the presence ofheteronymous shoulder reflex responses to stretchof the spastic elbow in persons with a spastic hemi-paresis following stroke. This hypothesis was testedby measuring the elbow and shoulder torque re-sponses to motor-controlled, imposed movements ofthe elbow. We anticipated that the reflex responsesto imposed elbow movements would be comprisedof both elbow and shoulder responses involving bothbiomechanical coupling and heteronymous reflexresponses.


Sixteen subjects with stroke, recruited through anoutpatient stroke clinic, and six neurologically intactage-matched controls participated in this study.Stroke subjects were tested between 1 and 18 yearsafter stroke, 10 had left hemiparesis, the Fugl–Meyerscale16 for the upper extremity ranged from 9 to 47(scale range is 0–66), and the Ashworth score3

ranged from 0 to 3. The stroke subject inclusioncriteria for participation in the study included atleast 21 years of age and a history of stroke resultingin clinical spasticity of the upper extremity. Exclu-sion criteria included the presence of fixed contrac-tures or a history of tendon transfer in the affectedlimb; inability to give informed consent; a diagnosisof any other neuromuscular disease; use of amino-glycoside antibiotics, curare-like agents, or otheragents that might interfere with neuromuscularfunction; profound atrophy or excessive weakness of

the muscles of the arm; or the presence of a systemicinfection. Control subjects were required to have nohistory of stroke or any other pathology of the upperextremity. The study was initiated after informedconsent was obtained; it was conducted in accordwith the Helsinki Declaration of 1975 and approvedby our Institutional Review Board.

The experimental apparatus and subject prepa-ration used in this study have been described previ-ously.32,33,35 Briefly, displacements of elbow anglewere imposed at a constant angular velocity using aBiodex Rehabilitation/Testing System 3 (BiodexMedical Systems, Shirley, New York), hereafter re-ferred to as Biodex. The affected hand and wrist ofthe subject were clamped securely and comfortablyonto a manipulandum that extended from the Bio-dex motor using a customized fixture. A six-axis loadcell (JR3, Woodland, California) was attached to themanipulandum, which measured the correspondingforces and moments along three principal axes. Theelbow axis of rotation was centered over the motorand proper placement was verified visually by anabsence of upper arm translation during manualmovement of the manipulandum. The motor posi-tion was adjusted to achieve a shoulder abduction of80° and a shoulder flexion of 20°. The wrist wasstrapped in a neutral position with a relaxed lightgrip of the fingers. With full elbow extension definedas 180°, the minimum and maximum range of mo-tion of the elbow was between 55°–100° and 40°–176°, respectively. Subjects were initially movedthrough the full, comfortable range of motion of theelbow, thus defining the range of elbow movementthat was to be used for each subsequent trial. Carewas taken to avoid translation or rotation of theshoulder joint.

Each test consisted of 12 imposed movements ofthe elbow. Each movement began with the elbow inthe flexed posture, and then a constant velocity ex-tension movement was imposed at the elbow untilthe elbow reached the predetermined extension pos-ture. The elbow was then held in extension for 5 sand returned to the flexed posture using a constantvelocity of the same magnitude. Movements wereconducted at four different test velocities: 6, 30, 60,and 90°/s while the subject was instructed to relax.The data gathered at the 6°/s test velocity were usedexclusively to measure the passive torque, since thisvelocity did not produce a stretch reflex response, asevidenced by an absence of detectable EMG signals.A minimum of 1 min for rest was allowed betweenmovement trials. The different velocities were ap-plied in a series of three epochs, with each epoch

Reflexes of Stroke Arm MUSCLE & NERVE November 2007 695

consisting of one trial at each velocity, in randomorder.

Surface EMG recordings were made of the teresmajor, medial deltoid, pectoralis major, biceps, andlateral head of the triceps. Surface EMG electrodes(Motion Lab Systems, Baton Rouge, Louisiana) wereplaced over the muscle bellies on lightly abradedskin and the signals were amplified (�10,000) andlow-pass filtered (500 Hz) prior to sampling. Loadcell, velocity, and position signals were low-pass fil-tered with a customized hardware circuit at 250 Hzand then all signals (including the EMGs) were dig-itized at 1,000 samples/s using a personal computerwith an SCB-100 data acquisition board (NationalInstruments, Austin, Texas) and customized Lab-View software (National Instruments). EMG datawere used to confirm reflexive joint torque measure-ments and for interpretation of the results.

Reflex torques at the elbow and shoulder jointwere calculated from the load cell measurementsand used as the primary measurement of the reflexresponse to imposed elbow movements. The loadcell data consisted of forces and moments along thethree principal axes during the entire movementcycle. The manipulandum, along with the load cell,was tilted during forearm placement in order toplace subjects in a comfortable position. This offsetload cell angle (�off) was measured along with theoffset distance between the shoulder joint and thecenter of the load cell denoted as loff (offset length).The reflex torque was calculated only for the exten-sion portion of the elbow perturbation. First, all loadcell signals were low-pass-filtered at 10 Hz using azero phase delay filter (the filtfilt function of Matlab,MathWorks, Natick, Massachusetts). The activestretch reflex torque was obtained by subtracting thepassive torque measured at slow velocity (6°/s) fromeach signal obtained at the higher velocities (30, 60,and 90°/s). This process accounts for artifacts result-ing from the weight of the arm or from the nonvis-cous passive properties of the joints. We then as-sumed that the passive viscous elements werenegligible, which is considered a reasonable assump-tion for the elbow.19

In order to obtain reflex torques from the loadcell data, we aligned the z axis with the elbow jointand set the axis perpendicular to the lateral hu-merus. As a result, the torque in the z direction wasa direct measure of elbow torque. The x and y axeswere defined to form an orthogonal axes system asshown in Figure 1. The load cell data were then usedto calculate the shoulder torques for abduction, ex-tension, and external rotation using the transforma-tion matrix shown below, which was derived by math-

ematically translating the forces and momentsobtained at the load cell to the elbow and shoulderjoints.

��AB

�EXT

�EROT

��sin��loff � cos��loff lhum � cos�� sin�� 0cos(�)lhum sin��lhum 0 0 0 1�cos��lhum � sin��lhum 0 � sin�� cos�� 0

��Fx

Fy

Fz

�x

�y

�z

�� e � �off

where �AB is shoulder abduction torque; �ext is shoul-der extension torque; �erot is shoulder external rota-tion torque; �e is elbow angle and �off is offset angle;lhum is the length of the humerus; and loff � offsetlength. Fx, Fy, Fz, �x, �y, and �z are the forces andtorques obtained from the load cell along the x, y,and z axis, respectively.

The velocity dependence of the torque measure-ments was then characterized by testing the correla-tion of the peak reflex torques at the elbow andshoulder to the test velocities (30, 60, and 90°/s) foreach subject. Linear regression was performed onthe peak reflex joint torques for each test velocity(30, 60, and 90°/s) and for each subject (� � 0.05).Further, to test velocity dependence the peak multi-joint torque values were compared across all sub-jects, trials, and test velocities using a multivariateanalysis of variance (MANOVA) (� � 0.05). Post-hocANOVAs were used to test the effect of velocity oneach torque measurement (elbow flexion/exten-sion, shoulder flexion/extension, shoulder abduc-tion/adduction, and shoulder internal/external ro-tation) (� � 0.05).

In order to determine whether the size of thereflex torques at the shoulder correlated with the

FIGURE 1. Schematic for the torque calculations. Using the armlength and elbow and shoulder angle measurements, the torquesproduced at the shoulder were measured. Shoulder abduction/adduction was defined along the y axis at the shoulder, flexion/extension along the z axis, and internal/external rotation alongthe x axis.


size of the elbow stretch reflex, a linear regression ofthe elbow torque and individual shoulder torqueswas performed across the test group. Peak torquewas calculated for each of the test velocities (30, 60,and 90°/s). The peak torques calculated for the90°/s test velocity were used for subsequent analysis,as they were more consistently produced across sub-jects than the torques obtained at the other testvelocities (30 and 60°/s). The mean torques for the90°/s trials were calculated for each subject and alinear regression analysis of each peak shouldertorque (flexion/extension, abduction/adduction,and internal/external rotation) versus peak elbowtorque (� � 0.05) was conducted across the testgroup.

Additional interpretation of the responses wasinvestigated by determining the relative delay of theshoulder torques (abduction/adduction, flexion/extension, and internal/external rotation) with re-spect to the elbow torque using a cross-correlationanalysis with the data obtained at the test velocity of90°/s. The start point for cross-correlation of theelbow and each shoulder torque was taken as thepoint immediately after the inertial artifact at thebeginning of the velocity profile and the endpointwas taken as the point immediately before the iner-tial artifact during the end of the velocity profile.These points were selected using customized Matlab

code where the elbow torque response was plottedalong with the velocity profile to determine the ini-tial and final artifacts. The mean delays betweeneach shoulder torque and the elbow torque werethen compared across the test group using a pairedt-test (� � 0.05) for the shoulder flexion/extension,abduction/adduction, and internal/external rota-tion.

RESULTS

The stretch reflex responses at the elbow and shoul-der of 16 hemiparetic individuals with stroke wereevaluated by examining torque and EMG responsesto constant-velocity ramp stretches. The elbow reflextorque response for a single trial at 90°/s is shown inFigure 2, along with EMGs generated at the elbowand shoulder muscles. A reflexive elbow flexiontorque was produced during the extension of theelbow, and an extension torque was produced dur-ing the flexion movement. The biceps and tricepsEMGs, representing elbow muscle activity in re-sponse to the imposed movement of the elbow, werefound to be consistent with the torque resistance tomotion. Heteronymous activation of shoulder mus-cles, which included the pectoralis major and medialdeltoid, was also observed despite the fact that thesemuscles are not lengthened by the elbow perturba-

FIGURE 2. Torque and EMG responses to an imposed elbow movement. The elbow position, elbow flexion/extension torque, and EMGsof the biceps, triceps, pectoralis, major and medial head of the deltoid are shown during a single test trial. The velocity of movement forthis trial was 90°/s. Note that the imposed elbow extension produced a reflex torque of the elbow in flexion, resisting the imposedmovement, with activation of the biceps. Conversely, imposed flexion activated the elbow extensors. Activation of the shoulder muscleswas also observed, suggesting a multijoint reflex response.


tion. Specifically, the EMGs at the shoulder sug-gested that elbow perturbation produced heterony-mous reflex activity in the shoulder in subjects withhemiparesis following stroke. Although reflex re-sponses were produced during both elbow flexionand extension, responses to elbow flexion were onlyobserved in 4 of our 16 subjects in this study and,thus, only the reflex response to elbow extension wasused for subsequent analysis.

Elbow perturbations produced substantial reflextorques at the shoulder that generally correspondedwith stretch reflex torque produced at the elbow. Atypical joint torque response to an imposed elbowextension movement is shown in Figure 3 for a hemi-paretic and control subject. In the control subjectsthe reflex torques at the shoulder (abduction/ad-duction, extension/flexion, and internal/externalrotation) and elbow (flexion/extension) were ab-sent at all test velocities (30, 60, and 90°/s) as shownin Figure 3B for a single trial at 90°/s. In subjectswith stroke, the stretch reflex response in the elbowwas generally accompanied by shoulder flexion andadduction torques. The shoulder flexion reflextorque was proportional to the elbow torque, consis-tent with a two-joint action of the biceps (elbow andshoulder flexion).2,5 In contrast, the shoulder adduc-tion torque, when it occurred, often produced awaveform slightly different from the elbow flexiontorque as shown in Figure 3A, with a rise that waslater than the elbow torque. In some cases (alsoevidenced in Fig. 3A), shoulder adduction torquemagnitude exceeded the elbow flexion torque. A netshoulder adduction was found to be more prevalentthan shoulder abduction, occurring in 12 of 16 testsubjects. In the four cases where shoulder abductionwas observed, it occurred simultaneously with elbowflexion and had an identical waveform.

The reflex torques at both the elbow and shoulderjoints were velocity dependent. Peak torque responsesfor each trial are shown in Figure 4 for a representativesubject. The magnitudes of the elbow and shoulderflexion torques and their velocity dependence weretypically similar, consistent with a mechanical couplingof the joint torques (e.g., by activation of a biarticularmuscle). Shoulder adduction/abduction torques alsoincreased with velocity, although the magnitude usu-ally differed substantially from the elbow flexiontorque and the relative amount of shoulder adductiontorque varied by subject. The regression results foreach subject, for each torque, showed a significantvelocity dependence across all subjects, with mean r2

values of 0.80 for elbow flexion, 0.82 for shoulderadduction/abduction, 0.75 for shoulder flexion, and0.76 for shoulder external/internal rotation (P � 0.05

in each subject). Thus, the peak elbow and shouldertorques across all subjects were found to significantlydepend on the test velocities of the imposed elbowmovement.

FIGURE 3. Reflex elbow and shoulder torques. The reflextorques produced at the elbow and shoulder by an imposedelbow extension at 90°/s are shown. (A) Reflex torques observedin stroke subjects. The reflex response consisted of an elbowflexion, which resisted the movement, along with a shoulderadduction, shoulder flexion, and shoulder external rotation. Thisresponse, shown for the same subject as in Figure 2, was typicalof the responses observed across most of the test group. Notethat the reflex responses at the shoulder showed a trend similarto that of the elbow in that they generally increased in magnitudewith progressive extension. (B) Absence of reflex torques in anormal control subject. The absence of response at both theelbow and the shoulder is representative of similar absence ofreflex responses observed across all control subjects.


Across the study group, the peak multijointtorques depended on velocity of the imposed elbowmovement. The MANOVA for all subjects and testvelocities indicated that larger test velocities pro-duced significantly greater reflex torque responsesacross all subjects (P � 0.05). Post-hoc ANOVA testsindicated significant effect of velocity (P � 0.0001)on each of the torque measurements.

A strong linear relationship was found when cor-relating peak elbow and peak shoulder flexiontorques across the population, but not for the elbowflexion and shoulder abduction/adduction torques.Specifically, shoulder flexion torque was found to bestrongly correlated with elbow flexion response (r2 �0.94, P � 0.05), consistent with the shoulder re-sponse being mediated by stretch reflex activation ofthe biceps (Fig. 5A). The shoulder adduction/ab-duction torque was not significantly correlated withthe elbow flexion torque response (r2 � 0.007, P �0.755), likely because of the four abduction (nega-tive torque) responses. Despite this lack of correla-tion, it appeared that the magnitude of the shoulderadduction increased with increasing elbow flexiontorque (Fig. 5B). The regression analysis was re-

peated in the 12 subjects demonstrating shoulderadduction torque to investigate this possibility. Theresulting correlation was significant (r2 � 0.54, P �

FIGURE 4. Velocity dependence of the reflex torques. The peakreflex torque for elbow flexion, shoulder adduction, shoulder flex-ion, and shoulder external rotation are shown for each velocity,for one subject (same as in Fig. 2). Similar results were observedacross all subjects tested. Linear regression lines are superim-posed on the data points. All elbow and shoulder torque re-sponses were significantly correlated to test velocity (r2 value was0.80 for elbow flexion, 0.82 for shoulder adduction/abduction,0.75 for shoulder flexion, and 0.76 for shoulder external/internalrotation) (P � 0.05).

FIGURE 5. Results of the linear regression of shoulder and elbowtorques. A significant correlation was observed between each peakshoulder torque and the peak elbow torque across the test group.Each symbol represents the mean data from one subject and theline calculated from the linear regression is shown. (A) The peakshoulder flexion torque showed a tight linear correlation with peakelbow torque (r2 � 0.94, P � 0.05). (B) The peak shoulder adduction(filled symbols) and abduction (open symbols) were also correlatedwith peak elbow flexion torque, although the deviation from the linewas larger (r2 � 0.48, P � 0.05). (C) The internal (filled symbols)and external (open symbols) rotation torques were less consistentacross subjects, although a significant correlation was still observed(r2 � 0.34, P � 0.05).


0.01), indicating that the shoulder adduction reflexresponse did not directly depend on elbow flexionreflex response, but may be related to the overallmagnitude of the response (Fig. 5B). It was alsonoted that one of the four subjects with shoulderabduction demonstrated internal rotation, a torqueresponse inconsistent with activation of the biarticu-lar elbow flexors (i.e., the biceps muscle producesexternal rotation5). The rotation torques were notsignificantly correlated to elbow flexion torque foreither internal rotation (n � 8, r2 � 0.49, P � 0.051)or external rotation (n � 8, r2 � 0.3, P � 0.156) (Fig.5C).

The analysis of the delay between the elbowtorque and shoulder torques, summarized in Figure6, indicated that relative to the elbow torque, shoul-der adduction torque had a larger delay than thedelay between the shoulder flexion and elbow flex-ion torques. In fact, shoulder flexion torque ap-peared to occur simultaneously with the elbow flex-ion torque response, whereas the elbow torquepreceded shoulder adduction torque. The mean de-lay between shoulder and elbow flexion was 0.1 ms,mean delay between shoulder adduction and elbowflexion was 25.28 ms, and mean delay between shoul-der rotation and elbow flexion was 29.19 ms. Thus,both shoulder adduction and shoulder rotationtorques were found to be significantly delayed withrespect to elbow flexion torque based on a pairedt-test (P � 0.0001).

The EMG measurements of the deltoid and pec-toralis major muscle groups generally suggested apattern of coactivation at the shoulder. All 16 sub-jects demonstrated detectable EMG activity in the

deltoid during elbow extension. Despite this obser-vation, 12 subjects produced a net shoulder adduc-tion torque. Conversely, activity in the pectoralismajor (a shoulder flexor, adductor, and internalrotator) occurred in only six subjects during elbowextension. Of these six subjects, three produced anet shoulder adduction and three internal rotation.Reliable recordings of signals from the teres majorwere obtained in only one subject, who demon-strated increases in EMG amplitude during elbowextension. Note that EMG signals were not measuredin a number of other shoulder muscles that are likelyto have contributed to the reflexive shoulder torquesobserved in this study.

In order to confirm that shoulder adduction wasnot caused by a biarticular action of the biceps, thebiceps of two neurologically intact controls wereelectrically activated using a 300-�s biphasic 50-Hzpulse, with an intensity that elicited a strong contrac-tion of the biceps. The arm was placed in the sametest position (elbow at 175–45° flexion). The datashowed no evidence of shoulder adduction torquethroughout the passive range of motion. Conversely,electrical activation of the biceps produced shoulderabduction torques, along with the expected shoulderflexion.

There appeared to be no trends in the relationbetween the measured shoulder torques and theclinical measures of motor function (Fugl–Meyerscore) or spasticity (Ashworth score).

DISCUSSION

We concluded from this study that imposed stretchperturbations of the elbow produce joint torques atthe shoulder including both biomechanically cou-pled joint torques and heteronymous reflex cou-pling of elbow and shoulder muscles. Specifically, ashoulder flexion torque was observed that appearedto be coincident with reflex activation of the elbowflexors. This torque was consistent with activation ofthe biceps muscle, which is a biarticular elbow/shoulder flexor.5 Shoulder adduction was also seenin 12 of 16 subjects in response to imposed elbowmovements and, based on differences in elbow flex-ion and shoulder adduction waveforms, appeared tobe produced by a heteronymous reflex coupling tothe imposed elbow extension. All responses werevelocity-dependent, suggesting that velocity-sensitiveafferents (e.g., Ia muscle afferents) likely mediatethe reflex responses. The delay between shoulderadduction and elbow reflex responses suggests thatshoulder adduction reflexes might involve interneu-ronal or supraspinal pathways. Functionally, this

FIGURE 6. Delay of the shoulder torque responses with respectto the elbow stretch reflex response. The delay of the shouldertorque with respect to the elbow reflex response was calculatedusing a cross-correlation of the two signals. Note that the delayfor the shoulder flexion torque was approximately zero across allsubjects tested. The other torque responses were delayedslightly from the elbow torque response.


raises the question whether heteronymous reflexcoupling might affect the coupling of elbow exten-sion and shoulder adduction observed during voli-tional movement tasks.

Mechanical Coupling of the Elbow and Shoulder. Thesimultaneous presence of shoulder flexion and el-bow flexion torques, illustrated by similar waveformsof the two torque profiles, could be attributed to astretch reflex activation of the biceps. Both heads ofthe biceps are biarticular muscles that cross theshoulder and elbow and produce a flexion torque atthe shoulder.5 We confirmed the flexion moment ofthe biceps using electrical stimulation, which pro-duced comparable elbow and shoulder torques in asimilar ratio to those observed during elbow stretch(Fig. 5A). There was also clear evidence of EMGactivity in the biceps during imposed stretch pertur-bations that coincided with the shoulder flexiontorque production. Thus, the most likely explana-tion for the presence of the shoulder flexion torqueduring imposed elbow extension was activation ofthe biarticular elbow/shoulder flexors. However, theoccurrence of shoulder adduction cannot be ex-plained by the same mechanism.

Cadaver studies, with the shoulder placed in aposture similar to the one used here, indicate thatbiceps produces a small abduction moment at theshoulder.5,28 This was confirmed using surface elec-trical stimulation of the biceps in two neurologicallyintact subjects. Biceps activation (EMG) has alsobeen observed during voluntary shoulder abductiontasks31; however, activation appears to depend onthe nature of the task4,17 and under some circum-stances the biceps does not appear to contribute toabduction at all.17 Together, these results suggestthat stretch reflex activation of biarticular elbow flex-ors could not have produced the observed shoulderadduction torque. The shoulder adduction andshoulder rotation observed in response to imposedelbow extension are likely to be produced by musclesthat do not cross the elbow, presumably throughheteronymous reflex pathways.

Shoulder adduction torque could have been pro-duced by a number of shoulder muscles includingthe teres major, teres minor, and pectoralis ma-jor.5,21 EMG activity was observed in the deltoids andpectoralis major muscles, which were not stretchedby the imposed elbow extension (Fig. 2). This activitycould have contributed to the observed shouldertorques. The pectoralis major acts as an adductor aswell as an internal rotator,5,21 and the posterior del-toid also has the potential to produce shoulder ad-duction, although this effect may be counteracted by

anterior deltoid action.5 EMG recordings from theteres major were inconclusive; however, there wassubstantial subcutaneous adipose tissue over themuscle in almost all subjects and the skin recordingsmay not have had sufficient sensitivity. The biceps,which do cross the elbow, act as shoulder externalrotators5 and thus could not account for the internalrotation observed in 8 of the 16 subjects. Overall, weconcluded that the shoulder adduction and shoul-der internal rotation torques must be produced bymuscles that do not cross the elbow.

The shoulder internal/external rotation torqueresponses also suggest that shoulder muscles that donot cross the elbow were activated by the imposedelbow extension movement. The pectoralis majormight have accounted for some of the internal rota-tion moment in these subjects, as EMG activity wascommonly observed in this muscle, but detectablepectoralis major activity was only observed in 3 of the8 subjects that produced internal rotation. EMGmeasurements were not made from a large numberof shoulder rotator muscles and, therefore, it is dif-ficult to interpret the sources of the shoulder rota-tion torques. Given the inconsistency of the shoulderinternal/external rotation response, coactivation ofthe shoulder rotators was a likely response to theimposed elbow extension, but this was not measureddirectly.

Reflex Feedback from the Elbow onto the Shoulder

Muscles. The pattern of the shoulder reflex torquethroughout the elbow range of motion was generallysimilar to the response commonly observed at theelbow: a sigmoidal shape with a gradual rise withstretch and a plateau at larger elbow angles.33,34 Ithas been hypothesized that this leveling of the reflextorque at larger joint angles could be related to anadditional reflex loop, possibly involving force-feed-back inhibition.22 The magnitude of the adductionresponse appeared to depend on the elbow re-sponse, which in turn depended on stretch velocity;therefore, the shoulder adduction response mightoriginate in the velocity-sensitive afferents of theelbow or, alternatively, as a secondary response tothe elbow muscle activation.

One explanation for the shoulder reflex re-sponse to imposed elbow movements is an enhancedheteronymous reflex coupling of the musclesthroughout the arm after stroke. In individuals withintact nervous systems, a monosynaptic coupling ofelbow and shoulder muscles has been described.29 Inaddition, excitatory reflex connections between mus-cle afferents throughout the arm and scapular mus-cles have been observed.1 Normally, these reflex


pathways are only observed under experimental con-ditions and are not activated by slower, large-ampli-tude perturbations of the arm joints. It is plausiblethat an enhanced excitability of these heteronymousstretch reflexes might coincide with increased hyper-excitability of homonymous stretch reflexes that of-ten accompany stroke.

Although volitional intervention at the shoulderduring the imposed elbow movement is unlikely, itcannot be completely eliminated from consider-ation. Subjects were instructed to relax during thetrial and all had substantial hemiparesis affecting thearm. Further, the responses were repeatable and hada consistent monotonically increasing response, sim-ilar to those associated with reflex actions. Voluntaryinterventions during large-amplitude stretch pertur-bations can disrupt the torque recordings, but gen-erally do not produce the type of repeatable, mono-tonically increasing responses that we observed. Inaddition, the responses we observed were correlatedto the velocity of perturbation and had a net delaywith respect to the elbow flexion torque.

Reflex Coupling as Part of a Motor Control Strategy.

Functionally, the observed shoulder activity mayhave been a consequence of an exaggeration of nor-mal stabilization of the shoulder that occurs duringvolitional arm movements. During multijoint armmovements, the nervous system anticipates andplans for mechanical effects that arise from motionof the linked segments, reflecting a shoulder-centered pattern that stabilizes the joint.18 Theshoulder requires a stable base and a wide move-ment range, which is maintained by various groupsof muscles including the teres major, pectoralis ma-jor, and deltoid.15 Load perturbations to the elbowin normal subjects produce multijoint responses thathave been attributed to activation of muscles actingabout the elbow and shoulder joints.26 The biceps, inparticular, may play an important role, since cadaverstudies suggest that it can provide stability to theshoulder, particularly in the anterior direction.23,24

The shoulder thus plays an important role in provid-ing the arm with a large range of motion and simul-taneously providing a stable platform for arm move-ments.

The shoulder muscles might demonstrate a hy-persensitivity to reflex activation after stroke becausenormal motor control of the shoulder relies morestrongly on stretch reflex pathways. For example,stretch reflexes play a relatively larger role in posturestabilization of the shoulder than the elbow.13 Fur-ther, postural stabilization of the endpoint of thearm incorporates a large multijoint control compo-

nent, which could be actualized either through biar-ticular muscles or heteronymous reflexes.13 The re-sults of our study could be the consequence of anincreased reflex gain at the shoulder, particularlyinvolving heteronymous reflexes for shoulder stabi-lization. That is, the shoulder response to imposedelbow extension could be a heteronymous reflex forpostural stabilization that is disinhibited by neuraldamage associated with stroke.

Heteronymous reflex coupling of the elbow andshoulder is consistent with a neural coupling ofmonoarticular muscles at the elbow and shoulder,which could compensate for limits in the biome-chanical coupling created by biarticular muscles. Forexample, postural tasks in primates suggest a neuralcoupling of elbow flexors/shoulder extensors and ofelbow extensors/shoulder flexors that compensatefor the biomechanical coupling produced by biartic-ular elbow/shoulder flexors (biceps) and elbow/shoulder extensors (triceps long head).25 A similarpattern may also occur in the elbow flexors/exten-sors and shoulder abductors/adductors in humansfollowing stroke. The biceps acts to flex the elbowand abduct the shoulder; therefore, the elbow flexoractivity could be neurally coupled with monoarticu-lar shoulder adductors in order to compensate forshoulder abduction produced by the biceps duringelbow flexion. Interestingly, the neural coupling thatwe observed was produced by a reflex test, whereasthe coupling in primates is produced by a volitionaltask,10 raising questions about the effects of the cor-tical damage resulting from stroke, its effects on thecoupling of muscles crossing multiple joints, and theneuroplastic adaptations that may occur during re-habilitation.

The reflex coupling of the elbow and shouldermay be a corollary to muscle synergies reported dur-ing volitional movement after stroke. One of themain components of “extensor synergy” observed inindividuals with spastic hemiparesis consists of exten-sion of the elbow and adduction of shoulder, with astrong influence on the pectoralis major. This syn-ergy pattern is believed to disrupt multijoint reach-ing movements after stroke.7,12,14 The muscle activitypatterns observed in response to imposed elbow ex-tensions are consistent with a movement synergy,rather than a synergy produced by volitional com-mand. That is, stretch of the elbow produced activityof the elbow flexors and shoulder adductors. Thisresult is opposite to the activation of elbow extensorsand shoulder adductors that is associated with voli-tional synergy patterns. In contrast, shoulder adduc-tion was produced during imposed elbow extensionmovements. This observation suggests that synergis-


tic patterns of muscle activity might result, at least inpart, from reflex coupling of the shoulder adductorsto stretch of the elbow flexors.

Supported by NIH grant R01-NS052509 and AHA grants0575084N and 0550109Z.

REFERENCES

1. Alexander CM, Harrison PJ. Reflex connections from forearmand hand afferents to shoulder girdle muscles in humans.Exp Brain Res 2003;148:277–282.

2. An KN, Hui FC, Morrey BF, Linscheid RL, Chao EY. Musclesacross the elbow joint: a biomechanical analysis. J Biomech1981;14:659–669.

3. Ashworth B. Preliminary trial of carisoprodol. Practitioner1964;192:540–542.

4. Basmajian JV, Latif A. Integrated actions and functions of thechief flexors of the elbow: a detailed electromyographic anal-ysis. J Bone Joint Surg Am 1957;39–A:1106-1118.

5. Bassett RW, Browne AO, Morrey BF, An KN. Glenohumeralmuscle force and moment mechanics in a position of shoul-der instability. J Biomech 1990;23:405–415.

6. Beer RF, Dewald JP, Dawson ML, Rymer WZ. Target-dependentdifferences between free and constrained arm movements inchronic hemiparesis. Exp Brain Res 2004;156:458–470.

7. Beer RF, Dewald JP, Rymer WZ. Deficits in the coordinationof multijoint arm movements in patients with hemiparesis:evidence for disturbed control of limb dynamics. Exp BrainRes 2000;131:305–319.

8. Bourbonnais D, Vanden Noven S, Carey KM, Rymer WZ.Abnormal spatial patterns of elbow muscle activation in hemi-paretic human subjects. Brain 1989;112:85–102.

9. Brunnstrom S. Movement therapy in hemiplegia: a neuro-physiological approach. New York: Harper and Row; 1970.

10. Cabel DW, Cisek P, Scott SH. Neural activity in primary motorcortex related to mechanical loads applied to the shoulderand elbow during a postural task. J Neurophysiol 2001;86:2102–2108.

11. Cirstea MC, Levin MF. Compensatory strategies for reachingin stroke. Brain 2000;123:940–953.

12. Cirstea MC, Ptito A, Levin MF. Arm reaching improvementswith short-term practice depend on the severity of the motordeficit in stroke. Exp Brain Res 2003;152:476–488.

13. de Vlugt E, Schouten AC, Van der Helm FC. Quantification ofintrinsic and reflexive properties during multijoint arm pos-ture. J Neurosci Methods 2006;155:328–349.

14. Dewald JP, Beer RF. Abnormal joint torque patterns in theparetic upper limb of subjects with hemiparesis. Muscle Nerve2001;24:273–283.

15. Engin AE. On the biomechanics of the shoulder complex.J Biomech 1980;13:575–590.

16. Fugl-Meyer AR, Jaasko L, Leyman I, Olsson S, Steglind S. Thepost-stroke hemiplegic patient. 1. a method for evaluation ofphysical performance. Scand J Rehabil Med 1975;7:13–31.

17. Furlani J. Electromyographic study of the m. biceps brachii inmovements at the glenohumeral joint. Acta Anat (Basel)1976;96:270–284.

18. Galloway JC, Koshland GF. General coordination of shoulder,elbow and wrist dynamics during multijoint arm movements.Exp Brain Res 2002;142:163–180.

19. Given JD, Dewald JP, Rymer WZ. Joint dependent passivestiffness in paretic and contralateral limbs of spastic patientswith hemiparetic stroke. J Neurol Neurosurg Psychiatry 1995;59:271–279.

20. Gracies JM, Meunier S, Pierrot-Deseilligny E, Simonetta M.Pattern of propriospinal-like excitation to different species ofhuman upper limb motoneurones. J Physiol (Lond) 1991;434:151–167.

21. Halder AM, Itoi E, An KN. Anatomy and biomechanics of theshoulder. Orthop Clin North Am 2000;31:159–176.

22. Hidler JM, Schmit BD. Evidence for force-feedback inhibitionin chronic stroke. IEEE Trans Neural Syst Rehabil Eng 2004;12:166–176.

23. Itoi E, Kuechle DK, Newman SR, Morrey BF, An KN. Stabilis-ing function of the biceps in stable and unstable shoulders.J Bone Joint Surg Br 1993;75:546–550.

24. Kumar VP, Satku K, Balasubramaniam P. The role of the longhead of biceps brachii in the stabilization of the head of thehumerus. Clin Orthop Relat Res 1989;244:172–175.

25. Kurtzer I, Pruszynski JA, Herter TM, Scott SH. Primate upperlimb muscles exhibit activity patterns that differ from theiranatomical action during a postural task. J Neurophysiol2006;95:493–504.

26. Lacquaniti F, Soechting JF. EMG responses to load perturba-tions of the upper limb: effect of dynamic coupling betweenshoulder and elbow motion. Exp Brain Res 1986;61:482–496.

27. Lance JW. The control of muscle tone, reflexes, and move-ment: Robert Wartenberg lecture. Neurology 1980;30:1303–1313.

28. Lucas DB. Biomechanics of the shoulder joint. Arch Surg1973;107:425–432.

29. McClelland VM, Miller S, Eyre JA. Short latency heterony-mous excitatory and inhibitory reflexes between antagonistand heteronymous muscles of the human shoulder and upperlimb. Brain Res 2001;899:82–93.

30. Pierrot-Deseilligny E. Propriospinal transmission of part ofthe corticospinal excitation in humans. Muscle Nerve 2002;26:155–172.

31. Sakurai G, Ozaki J, Tomita Y, Nishimoto K, Tamai S. Electro-myographic analysis of shoulder joint function of the bicepsbrachii muscle during isometric contraction. Clin OrthopRelat Res 1998;354:123–131.

32. Schmit BD, Dewald JP, Rymer WZ. Stretch reflex adaptationin elbow flexors during repeated passive movements in uni-lateral brain-injured patients. Arch Phys Med Rehabil 2000;81:269–278.

33. Schmit BD, Dhaher Y, Dewald JP, Rymer WZ. Reflex torqueresponse to movement of the spastic elbow: theoretical anal-yses and implications for quantification of spasticity. AnnBiomed Eng 1999;27:815–829.

34. Schmit BD, Rymer WZ. Identification of static and dynamiccomponents of reflex sensitivity in spastic elbow flexors using amuscle activation model. Ann Biomed Eng 2001;29:330–339.

35. Starsky AJ, Sangani SG, McGuire JR, Logan B, Schmit BD.Reliability of biomechanical spasticity measurements at theelbow of people poststroke. Arch Phys Med Rehabil 2005;86:1648–1654.


SHORT REPORT ABSTRACT: We report a patient with autosomal-dominant amyotrophiclateral sclerosis (ALS) and a sequence variation in the SOD1 promoterregion, located in the conserved TATA box motif (TATAAA3TGTAAA).Functional promoter studies of this variant in an in vitro system showedmoderate reduction in transcriptional activity of SOD1. This variant waspresent in only two of 301 individuals with sporadic amyotrophic lateralsclerosis, was not detected in 396 matched controls, and was recentlyreported in dbSNP (rs7277748). Our data suggest that this TATA box defectis not a disease-causing mutation or susceptibility factor for ALS but rathera rare polymorphism with a potential effect on SOD1 gene expression.

Muscle Nerve 36: 704–707, 2007

ANALYSIS OF A GENETIC DEFECT IN THE TATA BOXOF THE SOD1 GENE IN A PATIENT WITH FAMILIALAMYOTROPHIC LATERAL SCLEROSIS

STEPHAN NIEMANN, MD, PhD,1,2 WENDY J. BROOM, PhD,1 and ROBERT H. BROWN JR., DPhil, MD1

1 Cecil B. Day Laboratory for Neuromuscular Research, Harvard Medical School,Mass General Institute for Neurodegenerative Disease, Massachusetts General Hospital (East),Building 114, 16th Street, Charlestown, Massachusetts 02129, USA

2 Institut fur Humangenetik, Justus-Liebig-Universitat, Giessen, Germany


Amyotrophic lateral sclerosis (ALS) is a clinicallyand genetically heterogeneous group of neurode-generative disorders primarily affecting motor neu-rons in the brain, brainstem, and spinal cord.6,7,10

ALS is familial (FALS) in �10% of cases, mostly inher-ited as an autosomal-dominant trait.6,7 Mutations in thegene encoding Cu/Zn superoxide dismutase (SOD1)account for �15%–20% of FALS cases and were alsofound in some sporadic ALS (SALS) cases.9 SOD1 hasbeen highly conserved through evolution and is con-stitutively expressed in all eukaryotic cells.

At least 117 different SOD1 mutations have beenidentified in FALS and SALS. The vast majority ofthese mutations are missense mutations occurring inall five exons of the SOD1 coding region. Two exonicdeletions and three exonic insertions, resulting inframeshift and premature termination of the pro-tein, two in-frame exonic deletions, and two non-sense mutations have been reported. In addition,two intronic mutations have been demonstrated to

result in alternative splicing of the SOD1 transcript(ALS online database, www.alsod.org).

Here we report an FALS patient carrying a ge-netic defect in the highly conserved TATA-box motif(TATAAA3TGTAAA) of the SOD1 promoter. Thisvariant results in decreased expression of SOD1 in anin vitro system, but is unlikely to be the cause for ALSin this patient. Our study yields new insight intoSOD1 gene regulation.


Patient. The index case (marked by arrow) is from athree-generation German family with autosomal-dom-inant ALS (Fig. 1A). He first noted muscle cramps inthe legs and weakness in the left foot at the age of 42.On physical examination at the age of 43, he hadweakness of the left foot and later of the right foot,hyperreflexia of the legs, and fasciculations in all ex-tremities. There were no other neurological abnormal-ities. Electromyography at 43 years showed abnormalspontaneous activity, large motor unit potentials, and areduced recruitment pattern in both lower limbs. Hethen developed rapidly progressive disease with atro-phy in all extremities, weakness of neck extension, anddyspnea. Electromyographic findings at 46 years wereof abnormal spontaneous activity and a neurogenicpattern in both upper limbs (biceps brachii) and inboth lower limbs (tibialis anterior). He was tetraplegicand died at 47 years of age.

Abbreviations: ALS, amyotrophic lateral sclerosis; FALS, familial ALS; PCR,polymerase chain reaction; SALS, sporadic ALS; SNP, single nucleotide poly-morphismKey words: amyotrophic lateral sclerosis; SOD1; superoxide dismutase;TATA box; promoter assayCorrespondence to: S. Niemann; e-mail: [email protected]


704 Short Reports MUSCLE & NERVE November 2007

Family history revealed that all other affectedrelatives had developed disease symptoms in thefifth decade and had died a few years after diseaseonset.

The SOD1 TATA box motif was assayed in (1) thepatient and another 80 FALS cases from Germany,which proved negative for mutations in the codingsequence of SOD1, and (2) 301 SALS cases from theUnites States. All individuals tested were of Euro-pean ancestry. In addition, 396 individuals withoutany neurological disorder and from the same ethnicbackground served as controls.

Genetic Analysis. DNA was extracted from periph-eral blood with informed consent from all partic-ipants. The SOD1 promoter region was amplifiedby polymerase chain reaction (PCR) with the prim-ers SOD1Ex1F-S (5�-TTCTCCACATTTCGGGGTTC-3�), located at �261 to �242, and SOD1Ex1F-AS(5�-GTGACTCAGCACTTGGCAC-3�), located at 198–217 (corresponding to the SOD1 initiation codon;GenBank Access. No. NT_011512). This was fol-lowed by direct sequencing of both strands of PCRproducts on an ABI-prism 3100 DNA sequencer (Ap-plied Biosystems, Foster City, California) or a Beck-man Coulter CEQ 8800 Genetic Analysis System se-quencer (Beckman Coulter, Fullerton, California).SOD1 exons 2–5 were analyzed by direct sequencingof PCR products as described previously.8

Functional Studies. Approximately 2,200 bp ofSOD1 promoter sequence were amplified by PCRusing the following primers: 5�-AGGCTCGAG-AGAATCACTTGAACCCAGCA-3� and 5�-CGTAAG-CTTCGCCATAACTCGCTAGGCCACGC-3�. PCRproducts were cloned into the pGL3-Enhancer Lu-ciferase Reporter Vector (Promega, Madison, Wis-consin) at XhoI and HindIII restriction enzymesites. A mutation corresponding to the TATA boxsequence variation in the index patient was cre-ated using the Quick change site-directed mu-tagenesis kit (Stratagene, La Jolla, California) andwas verified by sequence analysis.

HeLa cells were grown in Dulbecco’s modified Ea-gle’s medium (DMEM) supplemented with 10% fetalcalf serum. Cells were plated in 96-well plates at 30,000cells/well and were cotransfected with 0.3 �g of eitherthe mutant or the wildtype SOD1 promoter pGL3 fire-fly luciferase reporter construct and with 0.3 �g of thepRL-TK renilla luciferase construct using lipo-fectamine reagents according to the manufacturer’sinstructions (Invitrogen, Carlsbad, California). ThepRL-TK plasmid was used as an internal control toassess transfection efficiency. Forty-eight hours aftertransfection, cells were washed in phosphate-bufferedsaline (PBS), lysed, and luciferase activity of cell ex-tracts was determined with the Dual-Luciferase Re-porter Assay System (Promega). Reporter firefly lumi-nescence activities were normalized to control pRL-TKrenilla luciferase activities. The normalized activity ofthe wildtype construct was set at 100% and the percent-age luciferase activity of the mutant construct calcu-lated accordingly. Data are reported as means andstandard deviation of an experiment performed inreplicates of 16.

FIGURE 1. (A) Pedigree structure of the family with the SOD1promoter TATA box variant. The filled symbols denote the af-fected individuals. The arrow indicates the proband; �/� indi-cates heterozygous mutation. (B) Sequence of the promoterregion of the human SOD1 gene. The first base of the ATG startcodon is indicated as 1. The TATA box at nucleotide position–110 to –105 (in relation to the start codon) is boxed. BC041449is an mRNA adjacent to the 5� region of the SOD1 gene andtranscribed in the opposite direction. (C) Electropherograms ofSOD1 promoter region in a patient (bottom) and control (top).Sequence analysis of PCR products reveals a heterozygous A toG transition at nucleotide position –109 (in relation to the ATGstart codon) in the patient. The sequence variant is located atnucleotide position 2 of the SOD1 TATA box, which is alteredfrom TATAAA to TGTAAA.

Short Reports MUSCLE & NERVE November 2007 705

RESULTS

Sequence analysis of the SOD1 promoter in the FALSpatient identified a heterozygous A to G transition atnucleotide position –109 with respect to the adenineof the ATG start codon numbered as 1 (Fig. 1C;GenBank Access. No. NT_011512). This sequencevariation (TATAAA3TGTAAA) is located in theTATA box motif, which is highly conserved in SOD1among various species (Fig. 1B). No sequence vari-ations were found in the coding region of SOD1(exons 1–5) in this patient. The TATAAA3TGTAAAvariant was not detected in another 80 FALS casesand 396 nonneurological controls, and was observedin only two of 301 (0.003) SALS cases. The variantwas reported as a single nucleotide polymorphism(SNP) rs7277748 at dbSNP, submitted as ss10979656by the Baylor College of Medicine (Houston, Texas).SNP rs7277748/ss10979656 was initially discoveredby alignment of sequence traces from a pool of eightanonymous samples from healthy adult donors re-cruited from the Baylor Polymorphism Resource.This SNP was only detected in a single aligned seg-ment, i.e., ascertained in two chromosomes, and wastherefore deposited as a nonvalidated SNP at db-SNP. Subsequent to our study, this SNP was validatedby the submission of ss32469332 to the clusterrs7277748 including genotype data obtained in apanel of 89 individuals. The allele frequency of theTATAAA3TGTAAA variant was reported as 0.022.

We evaluated the potential functional effect ofthis variant by introducing the TATA box mutationinto an SOD1 promoter cloned upstream of a lucif-erase reporter gene. Constructs carrying either themutant or normal allele and the pRL-TK controlplasmid were transiently transfected into HeLa cells.In the presence of the TATAAA3TGTAAA muta-tion, a decrease in luciferase activity was detected,which was �67.8% compared to the activity re-corded for the wildtype construct (data not shown).This finding is in agreement with earlier studiesdemonstrating that mutations within the TATA boxusually result in a decrease in gene expression.3

DISCUSSION

The index case from a family with autosomal-dominantALS presented here carries a heterozygous sequencevariation (A3G) in the promoter element of the SOD1gene (TATAAA3TGTAAA). This variant lies in theTATA box and replaces adenine, the second residue ofthe TATA box element. Comparative sequence analy-ses of eukaryotic promoter regions revealed that ade-nine at position 2 is the most highly conserved residueof the TATA box element.4,5 Mutagenesis studies have

shown that single base changes in the TATA box ele-ment decrease transcription initiation and accuracy.Loss of function due to reduced gene expression hasbeen demonstrated to be the mechanism of TATA boxmutations in very few Mendelian disorders.2,3 Conse-quently, we investigated the functional impact of theSOD1 TATA box variant in a transfection assay anddemonstrated that it resulted in reduced expression ofa reporter gene construct. Although our functionalstudies indicate that this TATA box variant may de-crease the transcription of SOD1, we believe it is un-likely to be the cause of ALS in the reported patient butrather represents one of the cited rare polymorphisms.

First, the variant is listed as a polymorphism(rs7277748) at dbSNP, initially observed in the align-ment of only two chromosomes (ss10979656). Therecent report of this SNP with an allele frequency of0.022, and its absence in 792 control chromosomesscreened in this study, indicate that this variant is arare polymorphism.

Second, current evidence supports the concept thatmutations in SOD1 exert their deleterious effect througha novel toxic property and not by loss or reduction ofSOD1 activity.7 Reduction of SOD1 activity has beenobserved in some but not all SOD1 mutations. In addi-tion, none of the 117 different hitherto reported SOD1mutations associated with FALS are true null alleles,which would be expected if loss of activity played a rolein the pathogenesis of the disease.1 Correspondingly,hetero- or homozygous SOD1 knockout mice do notdevelop an overt disease phenotype.

In conclusion, our results suggest that the A to Gtransition in the SOD1 promoter element, althoughlocated in the highly conserved TATA box, is not offunctional relevance in the patient described here. Inaddition, the presence of the TATA box defect in onlytwo of 301 (0.003) SALS individuals does not indicatethat this variant acts as a susceptibility factor in ALS.The variant rather represents a rare polymorphism, thefunctional significance of which is uncertain. The ob-served effect of this variant on gene expression, al-though only moderate and detected in an in vitrosystem, may give rise to the speculation that it could berelevant in modifying the pathogenic process of otherdisorders. However, there are currently no data tosubstantiate this concept. Alternatively, one may hy-pothesize that disruption of the TATA box is com-pletely innocent and has little or no effect in mediatingtranscription initiation in SOD1 in vivo. In fact, lack ofa TATA box and the presence of multiple transcriptionstart sites are characteristic features of most housekeep-ing genes, a group to which SOD1 belongs. In supportof this hypothesis, clusters of SOD1 mRNAs with 5� endsof different lengths have been reported, some of which


even encompass the TATA box motif (e.g., GenBankDB490082, DB453617, BF965365).

REFERENCES

1. Andersen PM, Sims KB, Xin WW, Kiely R, O’Neill G, Ravits J,et al. Sixteen novel mutations in the Cu/Zn superoxide dis-mutase gene in amyotrophic lateral sclerosis: a decade ofdiscoveries, defects and disputes. Amyotroph Lateral SclerOther Motor Neuron Disord 2003;4:62–73.

2. Antonarakis SE, Irkin SH, Cheng TC, Scott AF, Sexton JP,Trusko SP, et al. beta-Thalassemia in American blacks: novelmutations in the “TATA” box and an acceptor splice site. ProcNatl Acad Sci U S A 1984;81:1154–1158.

3. Bosma PJ, Chowdhury JR, Bakker C, Gantla S, de Boer A,Oostra BA, et al. The genetic basis of the reduced expressionof bilirubin UDP-glucuronosyltransferase 1 in Gilbert’s syn-drome. N Engl J Med 1995;333:1171–1175.

4. Breathnach R, Chambon P. Organization and expression ofeucaryotic split genes coding for proteins. Annu Rev Biochem1981;50:349–383.

5. Bucher P. Weight matrix descriptions of four eukaryotic RNApolymerase II promoter elements derived from 502 unrelatedpromoter sequences. J Mol Biol 1990;212:563–578.

6. Gros-Louis F, Gaspar C, Rouleau GA. Genetics of familial andsporadic amyotrophic lateral sclerosis. Biochim Biophys Acta2006;1762:956–972.

7. Hand CK, Rouleau GA. Familial amyotrophic lateral sclerosis.Muscle Nerve 2002;25:135–159.

8. Niemann S, Joos H, Meyer T, Vielhaber S, Reuner U, Gleich-mann M, et al. Familial ALS in Germany: origin of the R115GSOD1 mutation by a founder effect. J Neurol NeurosurgPsychiatry 2004;75:1186–1188.

9. Rosen DR, Siddique T, Patterson D, Figlewicz DA, Sapp P,Hentati A, et al. Mutations in Cu/Zn superoxide dismutasegene are associated with familial amyotrophic lateral sclerosis.Nature 1993;362:59–62.

10. Rowland LP, Shneider NA. Amyotrophic lateral sclerosis.N Engl J Med 2001;344:1688–1700.


SHORT REPORT ABSTRACT: Many patients with Duchenne muscular dystrophy (DMD) areeventually diagnosed with sleep-disordered breathing (SDB). SDB is asso-ciated with reduced ventilation, decreased arterial oxygen tension, andincreased respiratory muscle recruitment during sleep, factors that could beespecially detrimental to respiratory muscles in DMD. To assess whetherSDB impacts dystrophin-deficient respiratory muscle function and fibrosis,diaphragm strength, and collagen content were evaluated in dystrophic mice(Dmdmdx) exposed to experimental SDB. Diurnal exposure to episodic hyp-oxia resulted in a 30% reduction in diaphragm strength without affectingcollagen content. Episodic hypoxia secondary to SDB can exacerbate re-spiratory muscle dysfunction in DMD.

Muscle Nerve 36: 708–710, 2007

EPISODIC HYPOXIA EXACERBATES RESPIRATORYMUSCLE DYSFUNCTION IN DMDMDX MICE

GASPAR A. FARKAS, PhD, KATHLEEN M. MCCORMICK, PhD, and LUC E. GOSSELIN, PhD

Department of Exercise and Nutrition Sciences, Room 405, Kimball Tower, School of Public Healthand Health Professions, University at Buffalo, Buffalo, New York 14214, USA


Respiratory muscle failure is a leading cause ofdeath in Duchenne muscular dystrophy (DMD).Since respiratory muscle weakness represents theprimary impairment of the respiratory system inthese patients,5,12,13,28 elucidating mechanisms thatcontribute to diaphragm dysfunction is clinically im-portant. Developing strategies for preserving respi-ratory muscle function is critical for maximizing lifeexpectancy in DMD patients.

Sleep represents a vulnerable period for DMDpatients.4,6 Many DMD patients undergo nocturnalarterial oxygen desaturations as a consequence oftheir disorder. Smith et al.26 noted repeated night-time oxygen desaturations from 95% saturation atbaseline to a nadir of 74% (range 58%–90%) in 9 of14 DMD patients. In six clinically stable DMD pa-tients, Barbe et al.3 reported that oxygen desatura-tions were evident during 25% of total sleep time.The severity of the sleep-related respiratory distur-bances in DMD has been correlated with daytimeoxygen saturation levels and increasing age.3,15 Be-cause of inherent membrane fragility attributed tothe absence of dystrophin, the respiratory muscles of

DMD patients may be especially at risk from SDB.Despite the pervasiveness of SDB in this patient pop-ulation, whether and to what extent nighttime oxy-gen desaturations contribute to respiratory muscledysfunction has not been investigated.

The mutant Dmdmdx mouse lacks dystrophin andits respiratory muscles exhibit alterations consistentwith DMD.2,8,9,16,19,27 To assess whether SDB impactsdystrophin-deficient respiratory muscles, we evalu-ated the effect of long-term diurnal (sleep phase)exposure to episodic hypoxia on diaphragm func-tion and fibrosis in Dmdmdx mice.


Studies were conducted on 18 6-month-old maledystrophic (Dmdmdx) mice purchased from JacksonLaboratories (Bar Harbor, Maine). Animals wererandomly divided into two equal groups: (1) a room-air (RA) control group, and (2) an experimentalSDB group, exposed to episodic hypoxia (EH) 8h/day, 5–6 days/week for 12 weeks. During EH ex-posure, chamber O2 levels (90-s cycle length) fluctu-ated between 21% and 5%.22 Food and water wereavailable ad libitum. Studies were approved by ourInstitutional Review Board.

At the end of the experimental period animalswere deeply anesthetized (ketamine/xylazine) andthe diaphragm was rapidly removed in toto andplaced in oxygenated Krebs solution at 37°C. Dia-phragm samples were prepared for in vitro strength

Abbreviations: DMD, Duchenne muscular dystrophy; EH, episodic hypoxia;RA, room air; SDB, sleep-disordered breathingKey words: Dmdmdx mouse model; Duchenne muscular dystrophy; respira-tory muscle function; sleep apnea; sleep-disordered breathingCorrespondence to: G. A. Farkas; e-mail: [email protected]



and collagen concentration measurements, assess-ments of function, and fibrosis, respectively.

Functional Characteristics Isometric contractilecharacteristics were measured in isolated diaphragmsegments.10,11 Force output was measured with aFT03C force transducer (Grass Instruments, Quincy,Massachusetts). Twitch characteristics and force–frequency profile were determined at optimallength. Maximal isometric force (Po) was deter-mined and forces were normalized for cross-sec-tional area and expressed in Newtons per squarecentimeter (N/cm2).

Assessment of Fibrosis The concentration of hy-droxyproline was measured by high-performance liq-uid chromatography14 to assess diaphragm collagenconcentration.

Statistics Data from RA and EH groups were com-pared using the t-test. A value of P � 0.05 was con-sidered statistically significant. Results are presentedas means � SD.

RESULTS

Studies were initially carried out in 10 Dmdmdx mice(five RA and five EH), and subsequently repeated ina second age-matched cohort of eight Dmdmdx mice(four RA and four EH). The results from both co-horts revealed similar findings and were thus com-bined. In total, nine Dmdmdx mice were exposed toEH and nine Dmdmdx mice served as RA age-matched controls.

Body weights of EH-exposed mice remained con-stant during the experimental period, whereas RAmice gained weight. At the end of the experimentalperiod RA mice weighed more than EH mice(35.1 � 1.6 g vs. 28.8 � 4.1 g, P � 0.01). Failure ofanimals to gain weight during EH exposure is con-sistent with previous studies of EH.22

Diaphragm forces were significantly reduced(range, 27%–31% decrease) in the EH-exposedcompared to the RA group across all stimulationfrequencies (Fig. 1). Maximal tetanic force (Po) was11.9 � 1.4 and 8.6 � 2.4 N/cm2 (P � 0.01) in RAand EH-treated Dmdmdx mice, respectively. By com-parison, Po was 18.6 � 3.4 N/cm2 in five non-DMDcontrol mice (various ages and strains) used for dryruns prior to collecting our experimental data.

Collagen content was measured in diaphragmsegments from the first test cohort of mice. Dia-phragm collagen content was unaltered by EH expo-sure (8.5 � 2.9 vs. 8.7 � 4.4 �g hydroxyproline/mg

dry weigh, RA and EH, respectively). Diaphragmdry/wet ratio was also unaltered following EH expo-sure (0.230 � 0.008 vs. 0.218 � 0.012, RA and EHvalues, respectively).

DISCUSSION

In non-DMD rodent models, long-term EH exposurefails to negatively impact diaphragm force.17,22 Incontrast, the current findings (Fig. 1) demonstratethat EH exposure exacerbates diaphragm dysfunc-tion in Dmdmdx mice. Reductions in force per unitcross-sectional area can indicate that contractility iscompromised, or alternatively that infiltration byconnective tissue artificially reduces forces when cor-rected for cross-sectional area. The absence ofchange in collagen content and dry/wet ratio inEH-exposed Dmdmdx mice suggest that the func-tional force decline likely represents dysfunction ofmyofibrillar contractility.

The mechanisms contributing to the observedforce decline are not known. Exposure to EH leadsto the enhanced recruitment of respiratory musclesabove resting levels. Due to inherent mechanicalfragility from the lack of dystrophin,19 increased re-cruitment could cause sarcolemmal rupturing lead-ing to muscle dysfunction.1,7,19 Inflammation or ab-normal Ca�� levels in response to the muscleinjury/repair cycle could further contribute to mus-cle necrosis and dysfunction.18,23,24 In addition, dueto the lack of nNOS in dystrophin-deficient muscles,increased muscle activation during EH could result

FIGURE 1. Force–frequency profile of diaphragm from Dmdmdx

mice exposed to episodic hypoxia (MDX EH) and room air (MDXRA). Note that at all stimulation frequencies, diaphragm forcesare reduced in EH-exposed mice compared to RA controls.


in heightened oxidative stress and tissue ischemia,factors that could also promote respiratory muscledysfunction in DMD patients with SDB.20,21,25

Nocturnal oxygen desaturations are often notedin DMD patients presenting with normal daytimepulmonary functions, indicating that early in thecourse of the disease SDB will go undetected andthat some patients will fail to receive appropriateinterventions at a time that the putative effects ofnighttime episodic oxygen desaturation are beingmanifest. Detecting SDB early in the course of thedisease combined with identifying key putativemechanisms will provide therapeutic strategies tominimize respiratory muscle dysfunction in DMD.

Supported by Departmental Funds, from an Individual Develop-ment Grant awarded to G.F. from the United University Profes-sions Union, and from a Research Grant awarded to L.G. from theMuscular Dystrophy Association.

REFERENCES

1. Alderton JM, Steinhardt RA. Calcium influx through calciumleak channels is responsible for the elevated levels of calcium-dependent proteolysis in dystrophic myotubes. J Biol Chem2000;275:9452–9460.

2. Attal P, Lambert F, Marchand-Adam S, Bobin S, Pourny JC,Chemla D, Lecarpentier Y, et al. Severe mechanical dysfunc-tion in pharyngeal muscle from adult mdx mice. Am J RespirCrit Care Med 2000;162:278–281.

3. Barbe F, Quera-Salva MA, McCann C, Gajdos P, Raphael JC,de Lattre J, et al. Sleep-related respiratory disturbances inpatients with Duchenne muscular dystrophy. Eur Respir J1994;7:1403–1408.

4. Barthlen GM. Nocturnal respiratory failure as an indication ofnoninvasive ventilation in the patient with neuromusculardisease. Respiration 1997;64(Suppl 1):35–38.

5. Black LF, Hyatt RE. Maximal static respiratory pressures ingeneralized neuromuscular disease. Am Rev Respir Dis 1971;103:641–650.

6. Bourke SC, Gibson GJ. Sleep and breathing in neuromusculardisease. Eur Respir J 2002;19:1194–1201.

7. Clarke MS, Khakee R, McNeil PL. Loss of cytoplasmic basicfibroblast growth factor from physiologically wounded myofi-bers of normal and dystrophic muscle. J Cell Sci 1993;106:121–133.

8. Coirault C, Lambert F, Marchand-Adam S, Attal P, Chemla D,Lecarpentier Y. Myosin molecular motor dysfunction in dys-trophic mouse diaphragm. Am J Physiol 1999;277:C1170–1176.

9. Dupont-Versteegden EE, McCarter RJ. Differential expressionof muscular dystrophy in diaphragm versus hindlimb musclesof mdx mice. Muscle Nerve 1992;15:1105–1110.

10. Farkas GA, Gosselin LE, Zhan WZ, Schlenker EH, Sieck GC.Histochemical and mechanical properties of diaphragm mus-

cle in morbidly obese Zucker rats. J Appl Physiol 1994;77:2250–2259.

11. Gosselin LE, Barkley JE, Spencer MJ, McCormick KM, FarkasGA. Ventilatory dysfunction in mdx mice: impact of tumornecrosis factor-alpha deletion. Muscle Nerve 2003;28:336–343.

12. Gozal D. Pulmonary manifestations of neuromuscular diseasewith special reference to Duchenne muscular dystrophy andspinal muscular atrophy. Pediatr Pulmonol 2000;29:141–150.

13. Hahn A, Bach JR, Delaubier A, Renardel-Irani A, Guillou C,Rideau Y. Clinical implications of maximal respiratory pres-sure determinations for individuals with Duchenne musculardystrophy. Arch Phys Med Rehabil 1997;78:1–6.

14. Hartel JV, Granchelli JA, Hudecki MS, Pollina CM, GosselinLE. Impact of prednisone on TGF-beta1 and collagen indiaphragm muscle from mdx mice. Muscle Nerve 2001;24:428–432.

15. Khan Y, Heckmatt JZ. Obstructive apnoeas in Duchenne mus-cular dystrophy. Thorax 1994;49:157–161.

16. Lang JM, Esser KA, Dupont-Versteegden EE. Altered activityof signaling pathways in diaphragm and tibialis anterior mus-cle of dystrophic mice. Exp Biol Med (Maywood) 2004;229:503–511.

17. McGuire M, MacDermott M, Bradford A. Effects of chronicintermittent asphyxia on rat diaphragm and limb musclecontractility. Chest 2003;123:875–881.

18. Nguyen HX, Tidball JG. Interactions between neutrophilsand macrophages promote macrophage killing of rat musclecells in vitro. J Physiol (Lond) 2003;547:125–132.

19. Petrof BJ, Shrager JB, Stedman HH, Kelly AM, Sweeney HL.Dystrophin protects the sarcolemma from stresses developedduring muscle contraction. Proc Natl Acad Sci U S A 1993;90:3710–3714.

20. Rando TA, Disatnik MH, Yu Y, Franco A. Muscle cells frommdx mice have an increased susceptibility to oxidative stress.Neuromuscul Disord 1998;8:14–21.

21. Rando TA. Role of nitric oxide in the pathogenesis of mus-cular dystrophies: a “two hit” hypothesis of the cause of mus-cle necrosis. Microsc Res Tech 2001;55:223–235.

22. Ray AD, Magalang UJ, Michlin CP, Ogasa T, Krasney JA,Gosselin LE, et al. Intermittent hypoxia reduces upper airwaystability in lean but not obese Zucker rats. Am J Physiol RegulIntegr Comp Physiol 2007 (to be published).

23. Reid MB, Li YP. Tumor necrosis factor-alpha and musclewasting: a cellular perspective. Respir Res 2001;2:269–272.

24. Reid MB, Lannergren J, Westerblad H. Respiratory and limbmuscle weakness induced by tumor necrosis factor-alpha: in-volvement of muscle myofilaments. Am J Respir Crit CareMed 2002;166:479–484.

25. Row BW, Liu R, Xu W, Kheirandish L, Gozal D. Intermittenthypoxia is associated with oxidative stress and spatial learningdeficits in the rat. Am J Respir Crit Care Med 2003;167:1548–1553.

26. Smith PE, Calverley PM, Edwards RH. Hypoxemia duringsleep in Duchenne muscular dystrophy. Am Rev Respir Dis1988;137:884–888.

27. Stedman HH, Sweeney HL, Shrager JB, Maguire HC, Panet-tieri RA, Petrof B, et al. The mdx mouse diaphragm repro-duces the degenerative changes of Duchenne muscular dys-trophy. Nature 1991;352:536–539.

28. Vingos P. Respiratory function and pulmonary infection inDMD. Isr J Med Sci 1977;13:207–214.


SHORT REPORT ABSTRACT: Muscle cramps are difficult to study scientifically because oftheir spontaneity and unpredictability. Various laboratory techniques to inducemuscle cramps have been explored but the best technique for inducing crampsis unclear. Electrical stimulation appears to be the most reliable, but there is aperception that it is extremely painful. Data to support this perception arelacking. We hypothesized that electrical stimulation is a tolerable method ofinducing cramps with few side effects. We measured cramp frequency (HZ),pain during electrical stimulation, and soreness before, at 5 s, and 30, 60, and90 min after cramp induction using a 100-mm visual analog scale. Group 1received tibial nerve stimulation on 5 consecutive days; Group 2 received it onalternate days for five total treatments. Pain and soreness were mild. Thehighest ratings occurred on Day 1 and decreased thereafter. Intersessionreliability was high. Our study showed that electrical stimulation causes littlepain or soreness and is a reliable method for inducing cramps.

Muscle Nerve 36: 711–714, 2007

PAIN AND SORENESS ASSOCIATED WITH APERCUTANEOUS ELECTRICAL STIMULATIONMUSCLE CRAMPING PROTOCOL

KEVIN C. MILLER, MS, and KENNETH L. KNIGHT, PhD

Human Performance Research Center, Brigham Young University, 106 Smith Fieldhouse,Provo, Utah 84602, USA


Muscle cramps are common, but despite their prev-alence their cause is unknown. Most of what isknown about exercise-associated muscle cramps isbased on observations and anecdotes rather thanscientific research. Muscle cramps are difficult tostudy scientifically because of their unpredictabilityand spontaneity; thus, attempts have been made toinduce them in a laboratory setting.

The three main models for inducing cramps in alaboratory are through exercise,8 magnetic stimula-tion,3 and electrical stimulation.13 Exercise models areonly 50% effective8 and are associated with confound-ing variables (e.g., hydration status, lactate accumula-tion, electrolyte imbalances) that create confusion as towhether the cramp was induced by exercise or theseother variables. Magnetic stimulation is reliable3 butlacks precision regarding the focality of the magneticstimulus.4,11 Electrical stimulation is also highly reli-able,13 but has been associated by some with extreme

pain, possibly due to increased sensory nerve activationand higher current density.3 This appears to be thegreatest disadvantage for using electrical stimulation toinduce muscle cramps.

Pain or soreness associated with electrically in-duced cramps has been described3,4 but not quanti-fied. We hypothesized that electrical stimulation is areliable method for inducing muscle cramps anddoes not cause great levels of pain or soreness.Therefore, the purposes of this study were to deter-mine: (1) the amount of pain and soreness experi-enced with electrically induced cramps, (2) howlong soreness persists following the conclusion of theprotocol, (3) whether the technique can be alteredto reduce pain and soreness, (4) how long it takesfor the body to accommodate to the cramp thresh-old frequency, and (5) whether a positive linearrelationship exists between pain and threshold fre-quency or soreness and threshold frequency, and (6)to replicate a prior muscle cramping reliability study.


A 2 � 5 � 6 factorial design guided data collection.The independent variables were: group (electricalstimulation every day or on alternative days), day (1,2, 3, 4, 5), and time (prestimulation, during electri-cal stimulation, 5 s after stimulation, and 30, 60, 90

Abbreviations: EMG, electromyograph(y); ICC, interclass correlation coeffi-cientsKey words: accommodation; reliability, threshold frequency; visual analogscaleCorrespondence to: K. C. Miller; e-mail: [email protected]



min after cramp induction). The dependent vari-ables were: pain, soreness, and threshold frequency(i.e., the frequency in Hz at which two consecutivecramps were elicited in the flexor hallicus brevis).Pain was measured once (during electrical stimula-tion) every day. Soreness was measured 5 times (pre-stimulation, 5 s after stimulation, 30, 60, 90 min aftercramp induction) daily. Threshold frequency wasalso recorded daily.

Subjects. Twenty-three healthy, college-aged stu-dents (10 men, 13 women; mean age, 21.4 � 2.2years; height, 174.2 � 10.6 cm; weight, 68.7 � 10.1kg) were randomly assigned to one of two groups.Exclusion criteria included: (1) pregnancy, (2) in-jury to the dominant lower limb within the last 6months, (3) prior experience with an electricallyinduced muscle cramp protocol, and (4) neurolog-ical, cardiovascular, or neuromuscular disease. Theprocedures were approved by our university’s Insti-tutional Review Board and subjects provided writteninformed consent.

Instrumentation. The compound muscle action po-tentials of the flexor hallicus brevis were sampled at2,000 Hz for 15 s using the MP150 system and Acq-Knowledge 3.7.3 software (Biopac Systems, Goleta,California). Disposable, long-term measurementelectrodes (Biopac, EL502-10) were used to collectelectromyographic (EMG) data of the subjects’ dom-inant limb. The total EMG recording consisted ofbaseline (1 s), stimulation (2 s), and poststimulusactivity (12 s). A Grass S88 stimulator with SIU5Stimulus Isolation Unit (Astro-Med, West Warwick,Rhode Island) with an 8-mm Ag–AgCl shielded stim-ulating electrode (Biopac, EL258S) was used to de-liver the train of electrical stimuli.

Pain and soreness ratings were quantified usingan ungraduated 100-mm visual analog scale withanchors marking the extremes of the scale (i.e., nopain/soreness and worst pain/soreness experi-enced).

Procedures. Subjects were randomly assigned toone of two groups. Prior to each testing session,subjects rated their soreness, which was defined asdistress, discomfort, or hurtful sensations experi-enced in the medial ankle before or after stimula-tion. Subjects lay supine with their ankle hanging offthe edge of a table. Standard EMG preparatory pro-cedures13 were performed at the medial plantar as-pect of the foot, area around the medial and lateralmalleoli, and ipsilateral tibial tuberosity. An 8-mmstimulating electrode was placed slightly inferior to

the medial malleolus. The tibial nerve was submaxi-mally stimulated 2–4 times with 1-ms electrical stim-uli at 80 V to determine the site around the medialmalleolus that caused the greatest hallux flexion. An8-cm square dispersive electrode was placed over thelateral malleolus. Electrodes were secured with med-ical tape and an elastic wrap at these locations. TwoEMG measurement electrodes were placed 2 cmapart over the mid-belly of the flexor hallucis breviswith a single ground measurement electrode overthe ipsilateral tibial tuberosity.

Stimulus intensity was initially set at 80 V andincreased by 10-V increments according to subjecttolerance. Subjects donned headphones and lookedat the ceiling to eliminate noise or distracting stimuliand received two trains of electrical stimuli (onetrain/s) beginning at a train frequency of 4 Hz(eight total stimuli on the first trial). Immediatelyfollowing stimulation subjects rated their pain,which was defined as distress, discomfort, or hurtfulsensations in their medial ankle during stimulation.Subjects were instructed to rate the pain caused bythe stimulation, not by the cramp. Approximately 5 slater, subjects rated soreness.

If a cramp did not occur at 4 Hz, subjects restedfor 1 min and train frequency was increased by 2 Hz.This was repeated until the flexor hallucis breviscramped. A muscle cramp was defined as an invol-untary, painful contraction of the flexor hallucisbrevis immediately following stimulation, and wasverified by involuntary, sustained great toe flexionand an average EMG root mean square amplitudegreater than 2 SD above the 1-s baseline EMG aver-age root mean square amplitude.13 The final stimu-lus intensity and threshold frequency were recorded.

If a cramp failed to resolve spontaneously, thegreat toe was stretched until the cramp was allevi-ated. Following rest for 1 min the same stimulusintensity and frequency was reapplied. If the subjectscramped a second consecutive time, they rated theirpain and soreness and no further electrical stimula-tion was applied that day. The location of the EMGelectrodes, stimulating electrode, and EMG siteswere then marked for future testing sessions. Sub-jects lay on the table for 90 min, during which theyrated the amount of soreness in their medial ankleevery 30 min.

The same procedures were performed on subse-quent testing sessions with the exception of findingthe threshold frequency and stimulus intensity. Thestimulus intensity and threshold frequency that elic-ited a muscle cramp in the previous session wasapplied and cramping was verified. If the subject didnot cramp, the procedures for eliciting cramp were


repeated from the stimulus intensity and frequencythat induced the cramp previously. The same trainnumber (two) and duration (2 s) and pulse duration(1 ms) were used for all subsequent testing sessions.

Statistical Analysis. Five-day mean threshold fre-quencies, pain during the electrical stimuli, andsoreness before, at 5 s, and 30, 60, and 90 minpost-cramp induction were used for statistical analy-sis (Number Cruncher Statistical Software 2001,Kaysville, Utah). A mixed-model analysis of variancewas used to assess the effects of group, day, and timeand the interaction of these factors on pain, sore-ness, and threshold frequency. Tukey–Kramer post-hoc tests were used to determine within-subjecteffects. Pearson correlation coefficients were com-puted to describe the relationships between pain,soreness, and threshold frequency. Interclass corre-lation coefficients (ICC [3, 1]) estimated interses-sion reliability. Significance was set at P � 0.05.

RESULTS

There were no significant differences in pain(F1,21 � 0.53, P � 0.48), soreness (F1,21 � 0.83, P �0.37), or threshold frequency (F1,21 � 0.0002, P �0.95) between groups, nor were there significantgroup by day interactions for pain (F4,84 � 0.17, P �0.95) or threshold frequency (F4,84 � 0.48, P �0.75). Therefore, the groups were condensed forpost-hoc analysis. There were significant differencesin threshold frequency over days (F4,88 � 11.6, P �0.001), soreness over days (F4,88 � 7.4, P � 0.001),time (F4,88 � 17.4, P � 0.001), and day by time(F16,352 � 6.7; P � 0.001).

Pain was significantly greater on Day 1 than onDays 2–5 (F4,88 � 6.7, P � 0.001, Table 1). Pain onDays 2–5 were not significantly different.

Soreness was greater on Day 1, at 5 s post-crampinduction than all other combinations of day andtime (P � 0.05, Table 1). Soreness returned to rest-ing values within 30 min each day.

Threshold frequency was significantly lower on Day 1than on Days 3–5 (P � 0.05, Fig. 1). Threshold frequencyon Day 2 was significantly lower than Days 4 and 5 (P �0.05), but Days 3, 4, and 5 were not significantly differentfrom each other (P � 0.05, Fig. 1).

Pain and threshold frequency exhibited a weak,negative correlation between days (r � �0.33, P �0.13). Soreness immediately following electricalstimulation and threshold frequency exhibited aweak, negative correlation on Days 1 and 2 (r ��0.25, P � 0.25), and a weak, positive correlation onDays 3, 4, and 5 (r � 0.23, P � 0.29). Intersessionreliability was high (ICC [3,1] � 0.99).

DISCUSSION

The extreme pain purported to occur when induc-ing muscle cramps with percutaneous electrical stim-ulation3 did not occur. Our subjects’ highest painwas less than other accepted laboratory techniquesin the health professions such as cryotherapy treat-ments,7 intramuscular injections,1 and delayed-onset

Table 1. Visual analog scale pain and soreness ratings (mm on 100-mm scale) over 5 days.

Pain duringstimulation

Soreness Soreness post–cramp induction

Prestimulation 5 s 30 min 60 min 90 min

Day 1 13.5 � 15.2 0.0 � 0.0 8.5 � 11.9† 1.5 � 2.9 0.9 � 2.4 0.2 � 0.5Day 2 6.8 � 9.7* 0.0 � 0.0 2.2 � 3.4 0.5 � 0.9 0.3 � .9 0.1 � 0.3Day 3 6.5 � 8.7* 0.04 � 0.2 2.1 � 2.8 0.3 � 0.7 0.2 � 0.4 0.2 � 0.5Day 4 6.4 � 7.5* 0.2 � 0.9 1.8 � 2.4 0.3 � 1.1 0.1 � 0.5 0.2 � 0.7Day 5 5.8 � 7.2* 0.04 � 0.2 1.9�2.6 0.2 � 0.5 0.04 � 0.02 0.04 � 0.2

*Indicates a difference compared to Day 1 (P � 0.05).†Indicates difference from all other days and times (P � 0.05).

FIGURE 1. Threshold frequency over time. *Significantly differentthan Day 1 (P � 0.05). †Significantly different than Day 2 (P �

0.05).


muscle soreness protocols6 (13.5 vs. 21–49 mm inthe literature). Moreover, in one study comparingelectrical stimulation and magnetic stimulation, atechnique considered not to be highly painful,3,4

pain during electrical stimulation was no more pain-ful than magnetic stimulation in half of the patients.4

The immediate poststimulation reduction inpain (50%, 13.5 to 6.5 mm) and soreness (74%, 8.5to 2.2 mm) between Days 1 and 2 suggests a learningeffect or adaptation. Accommodation to stimulationintensity can occur shortly after application,12 andpain has been observed to decrease with repeatedelectrical stimulation.5,9,10 Pain and soreness do notappear to decrease significantly following the secondday of testing; thus, one testing session is sufficientfor pain and soreness accommodation.

Soreness appears to be less of a concern thanpain. Soreness immediately poststimulation on Day 1was only 63% of the pain experienced, and returnedto baseline levels within 30 min on each day oftesting. This suggests any soreness due to stimulationis short-lived and independent of pain. Subjects canbe assured that any discomfort they feel during theelectrical stimulation will dissipate within 30 min.Also, the lack of a difference in pain or sorenessbetween sessions indicates that subjects may betested daily without fear of accumulating pain orsoreness.

Threshold frequency appears unrelated to theamount of pain or soreness experienced. We begantesting at 4 Hz because others have observed cramp-ing to occur at this low of a stimulation frequency.2We believe starting stimulation at lower frequenciesensures a more accurate threshold frequency anddecreases the likelihood of subject apprehensionand muscle guarding. If the goal is to study suscep-tibility to muscle cramps, lower frequencies shouldbe used to detect changes in threshold frequencies.However, if muscle cramp relief interventions arethe goal, higher starting stimulation frequencies maybe used without fear of causing great pain or sore-ness. Moreover, our high intersession reliability over5 days of testing replicates and extends the previousreport of changes in threshold frequency observedover 3 days.13 Changes in threshold frequency areunlikely due to measurement variability.

Finally, we believe it is good technique to induceconsecutive cramps to verify subjects’ true thresholdfrequency. We allowed 1 min of rest before attempt-ing to induce a second cramp, whereas others13 havegiven 30 min of rest between cramps. As a result ofsuch a short rest period, we observed that the second

cramping episode was often less intense and ofshorter duration than the first cramp. Alpha motorneuron inhibition or motor unit fatigue may explainthis. Activation of muscle receptors due to the pas-sive stretch may have caused inhibition of the alphamotor neuron, preventing a strong cramp within 1min of stretching. Also, because the flexor hallicusbrevis is small, it is not unreasonable that most of themotor units were activated during the first cramp,and that motor unit fatigue may have inhibited somemotor units from inducing a second cramp so soonafter the first was initiated. One minute of rest maynot be long enough for these effects to dissipate.Future research should explore how much time isneeded to prevent this inhibition.

We thank the Mary Lou Fulton Endowment for their generousfunding of this research project, Mitch Radigan for help with datacollection, and Dr. Marcus Stone for his advice, encouragement,and guidance.

REFERENCES

1. Barnhill B, Holbert M, Jackson N, Erickson R. Using pressureto decrease the pain of intramuscular injections. J Pain Symp-tom Manage 1996;12:52–58.

2. Benatar M, Chapman K, Rutkove S. Repetitive nerve stimula-tion for the evaluation of peripheral nerve hyperexcitability.J Neurol Sci 2004;221:47–52.

3. Caress J, Bastings E, Hammond G, Walker F. A novel methodof inducing muscle cramps using repetitive magnetic stimu-lation. Muscle Nerve 2000;23:126–128.

4. Chokroverty S. Magnetic stimulation of the human peripheralnerves. Electromyogr Clin Neurophysiol 1989;29:409–416.

5. Colloca L, Benedetti F, Pollo A. Repeatability of autonomicresponses to pain anticipation and pain stimulation. Eur JPain 2005;10:659–665.

6. Dannecker E, Koltyn K, Riley J, Robinson M. Sex differencesin delayed onset muscle soreness. J Sports Med Phys Fitness2003;43:78–84.

7. Hawkins J, Knight K, Long B. Cold modalities decrease painfollowing orthopedic injuries. Salt Lake City, UT: RockyMountain Athletic Trainers Association; 2006.

8. Jung A, Bishop P, Al-Nawwas A, Dale R. Influence of hydra-tion and electrolyte supplementation on incidence and timeto onset of exercise-associated muscle cramps. J Athl Train2005;40:71–75.

9. Lahoda R, Stacher G, Bauer P. Experimentally induced pain:measurement of pain threshold and pain tolerance using anew apparatus for electrical stimulation of the skin. Int J ClinPharmacol Biopharm 1977;15:51–56.

10. Milne R, Kay N, Irwin R. Habituation to repeated painful andnon-painful cutaneous stimuli: a quantitative psychophysicalstudy. Exp Brain Res 1991;87:438–444.

11. Olney R, So Y, Goodin D, Aminoff M. A comparison ofmagnetic and electrical stimulation of peripheral nerves. Mus-cle Nerve 1990;13:957–963.

12. Starkey C. Therapeutic modalities. Philadelphia: FA Davis;1999.

13. Stone M, Edwards J, Babington J, Ingersoll C, Palmieri R.Reliability of an electrical method to induce muscle cramp.Muscle Nerve 2003;27:122–123.


CASE OF THE MONTH ABSTRACT: A progressive radial neuropathy of unknown etiology despite1.5T magnetic resonance imaging (MRI) and surgical exploration was iden-tified as an intraneural perineurioma by a localized Tinel’s sign, an enlargedradial nerve at the spiral groove by 3.0T MRI, and a fascicular biopsy. Thedistinction between the initial diagnoses of inflammatory, demyelinatingpolyneuropathy and perineurioma was made by immunohistochemistry andelectron microscopy. A slowly progressing, localized mononeuropathyshould include perineurioma in the differential diagnosis.

Muscle Nerve 36: 715–720, 2007

INTRANEURAL PERINEURIOMA OF THE RADIALNERVE VISUALIZED BY 3.0 TESLA MRI

DORIS NGUYEN, BS,1 P. JAMES DYCK, MD,2 and JASPER R. DAUBE, MD2

1 Mayo Clinic College of Medicine, Rochester, Minnesota, USA2 Department of Neurology, Mayo Clinic College of Medicine, 200 1st Street SW,

Rochester, Minnesota 55905, USA

Accepted 5 March 2007

Perineurioma is a rare, benign, peripheral nervelesion with an indolent course.3,7,9,12 The literaturehas classified two distinctive types based on localiza-tion: intraneural and extraneural (soft tissue).2–5 In-traneural perineurioma is composed of perineurialcells that surround the nerve as a diffusion barrierbetween blood and the nerve.14,21,25,26 The tumorcells stain positively for vimentin, epithelial mem-brane antigen (EMA), and glucose transporter pro-tein 1.11,17–19 Electron microscope studies have fur-ther characterized intraneural perineuriomas.17,20

The cells stain negatively for Schwann cells (S-100).18

Imaging studies to aid in the diagnosis of perineu-riomas have included high-resolution sonographyand magnetic resonance imaging (MRI).2,3,19,22 Re-cent studies have identified a chromosomal 22 dele-tion associated with intraneural perineuriomas, sug-gesting neoplasm rather than local compression ortrauma as the etiology.7 We describe a case of intra-neural perineurioma involving the radial nerve diag-nosed and confirmed by electromyography (EMG),MRI, and fascicular biopsy.

CASE REPORT

A healthy, physically active 30-year-old cashier wasreferred for a 4-year history of progressive right armweakness. Her symptoms started with right wrist painnot attributable to injury or illness, without weaknessor sensory loss. The wrist pain continued for 6months despite analgesics and physical therapy. AnEMG performed elsewhere suggested “right ulnarblockage.” Over the next 6 months the pain subsidedand was replaced with gradual-onset weakness ofright thumb extension. Paresthesias were present onthe dorsum of her right thumb. For 2 years there wasno recurrence of pain, but she had slowly progressiveweakness of finger and wrist extension on the right,with progressive radial distribution sensory loss forpain and touch.

Twenty-one months after onset, a right-elbowMRI was performed and showed atrophy of forearmmuscles innervated by the posterior interosseousnerve. Two months later, radial nerve exploration inthe distal third of the arm and the elbow regiondemonstrated no abnormality; neurolysis was per-formed without a biopsy. Six months after the nerveexploration an EMG and nerve conduction studies(NCS) of the right arm performed at another insti-tution were reported to show partial radial nerveconduction block at the spiral groove. Her rightforearm extensor muscles showed reduced recruit-ment, fibrillation potentials, and large motor unitpotentials (original report unavailable).

She was treated with intravenous immune globu-lin 1.0 g/kg monthly for 9 months. The deficit con-

Abbreviations: CIDP, chronic inflammatory demyelinating polyneuropathy;EM, electron microscopy; EMA, epithelial membrane antigen; EMG, electro-myography; FSE, fast spin echo; FOV, field of view; H&E, hematoxylin andeosin; MRI, magnetic resonance imaging; NCS, nerve conduction studies;RF, radiofrequencyKey words: 3.0T MRI; chronic inflammatory demyelinating polyneuropathy;perineurioma; radial neuropathy; Tinel’s signCorrespondence to: J. R. Daube; e-mail: [email protected]

© 2007 Wiley Periodicals, Inc.Published online 30 April 2007 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/mus.20795

Radial Nerve by 3.0T MRI MUSCLE & NERVE November 2007 715

tinued to progress slowly despite treatment. She wasgiven Rebif injections twice per week for 3 months,with no improvement.

A second EMG was reported to suggest conduc-tion block of the right median and ulnar nerves(original report unavailable). Three months later acervical MRI revealed no abnormalities of neuralstructures. Her tentative diagnosis was of multifocalmotor neuropathy.

For 3 months prior to our evaluation, she expe-rienced generalized fatigue and exhaustion. Shecomplained of bilateral calf, forearm, and paraspinalmuscle aching that increased with activity and pro-gressive cramps in her calves. She denied twitching,tremor, or other abnormal movements. She did nothave weakness or sensory loss in the other limbs. Shehad no difficulty in swallowing or shortness ofbreath. She was referred for suspected amyotrophiclateral sclerosis or focal motor neuron disease. Areview of her past medical history and family historyidentified no systemic disorders clearly related to herprimary neurological problem. Her medications in-cluded treatments for migraine, edema, hirsutism,depression, and muscle and joint pain that appearedunrelated to her right upper-extremity problem.There were no other significant medical problems orlaboratory findings.

Neurological examination revealed severe rightforearm extensor weakness and atrophy with relativesparing of extensor carpi radialis longus and brachi-oradialis. Stretch reflexes were normal, although re-inforcement was required for the quadriceps re-flexes. Loss of superficial pain and temperaturesensation was limited to the dorsum of the right

hand. There was no other sensory disturbance. Pe-ripheral nerve enlargement and fasciculation wereabsent. Blood test studies, including heavy metaltesting, were normal.

Conduction studies of the right upper limb (Ta-ble 1) showed a low-amplitude right radial com-pound muscle action potential with a normal four-point radial motor conduction studies. Right radialsensory nerve action potential was absent. NCS werenormal for the rest of the right arm, right leg, andleft arm (Table 1). Needle EMG of right radial-innervated muscles distal to the triceps and anco-neus muscles revealed fibrillation potentials, re-duced recruitment, and large motor unit potentials.Needle EMG of the following muscles were normal:right pronater teres, right first dorsal interosseous,right deltoid, right biceps brachii, right triceps (longhead), right low cervical paraspinal, right vastus me-dialis, right anterior tibial, right medial gastrocne-mius, and left extensor digitorum communis. Thefindings were those of a long-standing, severe rightradial neuropathy distal to the branches to the tri-ceps and anconeus muscles.

MRI of the right elbow without contrast com-pared to her previous MRI showed progressive atro-phy of posterior interosseous forearm muscles. Anassociated mild increase in T2 signal suggested bothchronic and subacute changes. No focal lesions orabnormal neurovascular structures were seen. TheMR images were interpreted as showing posteriorinterosseous neuropathy. Six months later an MRI ofthe brachial plexus bilaterally showed only mild de-generative changes in the right acromioclavicularjoint. Two weeks later the patient’s symptoms be-

Table 1. Nerve conduction studies.

Nerve Stimulus site Recording site Amplitude (mV/�V) Velocity (m/s)Distal latency

(ms) F-wave latency (ms)

SensoryL. radial Elbow Dorsal hand 33 (�20) — 2.5 (�2.9) —R. radial Elbow Dorsal hand 0 — No response —R. median Wrist Index finger 61 (�15) 62 (�56) 3.2 (�3.6) —R. ulnar Wrist Fifth finger 48 (�10) 70 (�54) 2.3 (�3.1) —R. sural 14 cm proximal

to ankleAnkle 11 (�6) — 3.5 (�4.5) —

MotorL. radial Elbow EDC 11.8 123 2.1 —R. radial Elbow EDC 2.1 76 2.1 —R. median Elbow APB 10.2 (�4.0) 53 (�48) 3.6 (�4.5) 25.8R. ulnar Elbow ADM 10.9 (�6.0) 58 (�51) 2.7 (�3.6) 24.9R. peroneal Ankle EDB 5.6 (�2.0) 45 (�41) 4.4 (�6.6) 47.4R. tibial Ankle AH 12.2 (�4.0) 49 (�40) 5.6 (�6.1) 49.9

APB, abductor pollicis brevis; EDC, extensor digitorum communis; ADM, abductor digiti minimi; EDB, extensor digitorum brevis; AH, abductor hallucis.Normal values are in parentheses.Dashes indicate that no values were recorded.

716 Radial Nerve by 3.0T MRI MUSCLE & NERVE November 2007

came progressively worse. Examination now revealedpercussion tenderness (Tinel’s sign) in the mid-armover the course of the radial nerve; 3T MRI of herright humerus was performed. The MR images re-vealed an enlarged radial nerve at the spiral groove(Fig. 1).

Exploration of the radial nerve was performed ata level proximal to the previous neurolysis. A focalenlargement of the radial nerve was noted in themid-arm (Fig. 2); fascicular biopsy was performed of

a nonfunctioning portion of the nerve to determinethe nature of the abnormality. The procedure waswell tolerated without new deficit.

Hematoxylin-and-eosin (H&E) staining of thefascicular nerve biopsy showed onion-bulb–likestructures made of whorls of spindle cells sur-rounded by collagen. The spindle cells were charac-terized by extended nuclei and bipolar, elongated,eosinophilic cytoplasmic processes. The onion-bulb–like structures were cellular rather than true,Schwann cell onion-bulbs, such as are seen in hyper-trophic neuropathies like chronic inflammatory de-myelinating polyneuropathy (CIDP). Immunohisto-chemical preparations showed antibody staining forEMA but were negative for S-100 (Fig. 3). Electronmicroscopy (EM) of the nerve demonstrated elon-gated cytoplasmic processes with incomplete basallamina (Fig. 3). These findings confirmed the diag-nosis of perineurioma with pseudo–onion-bulbsrather than true onion-bulbs from Schwann cells.Unlike perineuriomas, CIDP reacts negatively withEMA and positively with S-100.15 A decision wasmade to perform standard radial nerve tendon trans-fers at a later stage. The outcome of tendon transferswas more favorable and predictable than resection ofthe lesion and nerve grafting given the relativelylong segment of abnormal nerve, the chronicity ofmuscle denervation, and the potential for recur-rence of perineurioma.

A year after the patient was referred to us and 5years after onset of symptoms, she underwent recon-struction with tendon transfers: pronater teres toextensor carpi radialis brevis, palmaris longus to ex-tensor pollicis longus, and flexor carpi radialis toextensor digitorum communis. Three months afterthe tendon transfers, the patient rated her rightupper extremity improvement at 80%. She was pain-free, off medications, and able to perform all activi-ties of daily living, except heavy lifting. She demon-strated excellent right wrist, finger, and thumbextension but continued physical therapy for 2 moremonths to regain more strength. A 10-month fol-low-up showed recovery of functional use of herright hand with only limited independent functionof finger extension. She had no pain. She now worksas a clerk and uses her right arm effectively at work.She is able to do all activities of daily living.

DISCUSSION

At least 72 intraneural perineurioma cases werefound by PubMed and Medline literature searches,with involvement of hand and upper extremitiesconstituting 60% of cases; the radial nerve was af-

FIGURE 1. (A) T1-weighted fast spin echo axial 3.0T MR image(TR � 950 ms, TE � 16) of right arm without contrast. All imagingwas performed using a custom-built receive-only extremity RFcoil. (B) T2-weighted fast spin echo axial 3.0T MR image of rightarm without contrast with fat suppression. Note the enlargedradial nerve (arrow) at the level of the spiral groove. There is mildartifact in both images related to respiratory motion of the adja-cent chest wall.


fected in 10% of those cases.5,15 The incidence ofperineurioma is further complicated by its manysynonyms: storiform perineurial fibroma, localizedhypertrophic neuropathy, or hypertrophic neuri-tis.4,10,12

Focal or multifocal immune-based, inflammatoryneuropathies are the most common chronic monon-europathies that have no apparent cause.1,16 Ourpatient demonstrates that a perineurioma should bepart of the differential diagnosis for any patient pre-senting with an idiopathic, indolent, localizedmononeuropathy. EMG and NCS can readily localizeand define the severity of a mononeuropathy. Attimes the presence of dispersion can suggest an in-flammatory demyelinating process, but differentia-tion of pathologic entities can only be reliably made

FIGURE 3. Representative pictures from the targeted fascicular radial nerve biopsy. (A) A transverse epoxy section, stained withmethylene blue showing pseudo–onion-bulbs typical of perineurioma—they are densely compacted, circumferentially arranged profilesthat occasionally have a myelinated fiber in the center. (B) An electron micrograph of a myelinated fiber surrounded by the processes ofthe perineurioma. (C) A paraffin transverse section that shows that the centers of the myelinated fibers react but that circumferentialprocesses do not react to S-100 (a Schwann cell marker) in contrast to (D), the EMA preparation (perineurial membranes) in which thecircumferential processes are reactive. These preparations demonstrate that the onion-bulb–like formations in this case are from aperineurial origin (pseudo–onion-bulbs) and from a Schwann cell onion (real onion-bulbs.) These findings are diagnostic of perineurioma.

4™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™FIGURE 2. (A) The mildly enlarged radial nerve can be seen atthe proximal extent of the surgical exposure (i.e., near the mid-arm). Its fascicles are enlarged at that level. The nerve appearsnormal more distally. (B) A single nonfunctioning fasciculargroup, measuring �2 mm, was selected for biopsy (in vasolooplabeled by the asterisk). This individual fascicle is abnormal, witha tapered proximal portion and expanded distal portion.


with biopsy. Fascicular biopsy is an effective mode ofdiagnosis with limited risk of increased deficit.6,23,27

Fascicular biopsies require sufficient preopera-tive localization that the surgeon can reliably deter-mine where to perform the biopsy. A combination ofthe findings on physical examination (including alocalized Tinel’s sign), EMG/NCS, and MRI isneeded for optimal guidance. Imaging of nerves isfinding increased value in patients with mononeu-ropathies without a clear etiology. The power of a 3TMRI scan early in the course of disease can expeditethe localization of the disease process and facilitatethe diagnosis of perineurioma. Tesla measures themagnetic field strength of the scanner and 3.0T isequivalent to roughly 30,000 times the earth’s mag-netic field.24 The 3T MRI has twice the field strengthof conventional 1.5T MRI scanners, allowing for sig-nificantly improved resolution in a clinically feasibleexamination time due to the roughly double avail-able signal-to-noise ratio.9 Because of improvedspatial resolution, small lesions can be accuratelylocalized and better characterized prospectively, es-pecially when targeted based on clinical and EMGfindings. In this patient’s case, the 1.5T and 3Texaminations were performed using similar RF coils(both custom-made at the Mayo Clinic with nearlyidentical geometry so that coil differences were notcontributory). Identical pulse sequences were usedat both strengths (FSE T1 and FSE T2 with chemicalfat saturation). The difference lies in the in-planespatial resolution and slice thickness. The 3T exam-ination was performed using a smaller 12-cm field ofview (FOV) with a larger imaging matrix (384 � 256,pixel size �0.46 mm) and thinner sections (5 mm)than the 1.5T examinations, which used a 16-cmFOV with a matrix of 256 � 192 (pixel size �0.84mm) and a slice thickness of 6 mm, leading to avoxel size �3–4 times larger than at 3T. Identifica-tion of this small (8 mm) lesion immediately adja-cent to vessels (that confused the interpretation)resulted from the higher spatial resolution, whichclearly showed the lesion margins and detail withoutsignificant volume averaging even though patientmotion was present and slightly blurred some of theimages. In retrospect, the lesion is not identifiableon the 1.5T study, which was performed in a veryappropriate and standard fashion.

The improved resolution not only allows narrow-ing of differential diagnostic considerations but alsoimproves the accuracy of targeted fascicular biopsiesfor definitive diagnosis. Earlier studies have reliedon the identification of peripheral nerve abnormal-ities through the use of high-quality T2-weightedimaging with robust fast suppression, which was the

technique used on both the 1.5T and 3T examina-tions here.

The added value of high-field MRI has been sig-nificant in the area of brain, spine, body, and mus-culoskeletal imaging.9,24 The precise pathologic di-agnosis of a focal nerve lesion, however, requires afascicular nerve biopsy. Biopsy allows histologic, im-munohistochemical, and EM studies to confirm thediagnosis of a perineurioma. A surgeon experiencedin peripheral nerve surgery facilitates the isolationand removal of a nerve segment with limited loss ofsensory and motor function, and infrequent residualpain.

In summary, the success in identifying this lesiondepended on a cooperative effort among specialistswhich, through communication of a full knowledgeof the patient’s symptoms and suspected area ofinvolvement, led to a proper imaging study and thento a favorable outcome.

The authors thank Dr. Kimberly K. Amrami for developing 3TMRI scanning of the peripheral nerve, providing the images, andwriting parts of the article; Dr. Robert J. Spinner for performingthe fascicular nerve biopsy and editing the article; Dr. Joel P.Felmlee for designing the custom coil and the protocol for 3Tupper arm imaging; and Ms. JaNean Engelstad for the histologicalpreparations. Material from this case report was presented at theannual meeting of the American Academy of Neurology in April2005 at Miami, Florida.

REFERENCES

1. Allen DC, Smallman CA, Mills KR. Multifocal acquired demy-elinating sensory and motor neuropathy presenting as a pe-ripheral nerve tumor. Muscle Nerve 2006;34:373–379.

2. Beekman R, Schoemaker MC, van der Plas JPL, van den BergLH, Franssen H, Wokke JHJ, et al. Diagnostic value of high-resolution sonography in ulnar nerve neuropathy at the el-bow. Neurology 2004;62:767–773.

3. Beekman R, Slooff WM, Van Oosterhout FM, Lammens M,Van Den Berg LH. Bilateral intraneural perineurioma pre-senting as ulnar neuropathy at the elbow. Muscle Nerve 2004;30:239–243.

4. Bilbao JM, Khoury NJ, Hudson AR, Briggs SJ. Perineurioma(localized hypertrophic neuropathy). Arch Pathol Lab Med1984;108:557–560.

5. Cortes W, Cheng J, Matloub H. Intraneural perineurioma ofthe radial nerve in a child. J Hand Surg 2005;30A:820–825.

6. Dyck PJB, Amrami KK, Spinner RJ, Klein CJ, Engelstad J, DyckPJ. Fascicular biopsy of proximal nerves in selected cases withMRI abnormality is diagnostically informative. J PeripherNerv Syst 2003;8:14–17.

7. Emory TS, Scheithauer BW, Hirose T, Wood M, Onofrio BM,Jenkins RB. Intraneural perineurioma: a clonal neoplasmassociated with abnormalities of chromosome 22. Am J ClinPathol 1995;103:696–704.

8. Erlandson RA. The enigmatic perineurial cell and its partici-pation in tumors and in tumorlike entities. Ultrastruct Pathol1991;15:335–351.

9. Fischbach F, Lehmann TN, Ricke J, Bruhn H. Vascular com-pression in glossopharyngeal neuralgia: demonstration byhigh-resolution MRI at 3 Tesla. Neuroradiology 2003;45:810–811.


10. Hamazaki S, Fujiwara K, Okada S. Intraneural perineuriomainvolving the ulnar nerve. Pathol Int 2004;54:371–375.

11. Hirose T, Tani T, Shimada T, Ishizawa K, Shimada S, Sano T.Immunohistochemical demonstration of EMA/Glut1-positiveperineurial cells and CD34-positive fibroblastic cells in pe-ripheral nerve sheath tumors. Mod Pathol 2003;16:293–298.

12. Isaac S, Athanasou NA, Pike M, Burge PD. Radial nerve palsyowing to localized hypertrophic neuropathy (intraneuralperineurioma) in early childhood. J Child Neurol 2004;19:71–75.

13. Jazayeri MA, Robinson JH, Legolvan DP. Intraneural perineu-rioma involving the median nerve. Plast Reconstr Surg 2000;105:2089–2091.

14. Johnson PC, Kline DG. Localized hypertrophic neuropathy:possible focal perineurial barrier defect. Acta Neuropathol1989;77:514–518.

15. Kline DG, Gruen JP, Mitchell W. Resection and graft repairfor localized hypertrophic neuropathy. Neurosurgery 1998;43:78–83.

16. Koller H, Kieseier BC, Jander S, Hartung H. Chronic inflam-matory demyelinating polyneuropathy. N Engl J Med 2005;352:1343–1354

17. Lazarus SS, Trombetta LD. Ultrastructural identification of abenign perineurial cell tumor. Cancer 1978;41:1823–1829.

18. Mitsumoto H, Wilbourn AJ, Goren H. Perineurioma as thecause of localized hypertrophic neuropathy. Muscle Nerve1980;3:403–412.

19. Perentes E, Nakagawa Y, Ross GW, Stanton C, Rubinstein LJ.Expression of epithelial membrane antigen in perineurial

cells and their derivatives. An immunohistochemical studywith multiple markers. Acta Neuropathol (Berl) 1987;75:160–165.

20. Rankine AJ, Filion PR, Platten MA, Spagnolo DV. Perineu-rioma: a clinicopathological study of eight cases. Pathology2004;36:309–315.

21. Saft C, Andrich JE, Neuen-Jacob E, Schmid G, Schols L,Amoiridis G. Supracubital perineurioma misdiagnosed as car-pal tunnel syndrome: case report. BMC Neurol 2004;4:19–24.

22. Simmons Z, Mahadeen ZI, Kothari MJ, Powers S, Wise S,Towfighi J. Localized hypertrophic neuropathy: magnetic im-aging findings and long-term follow-up. Muscle Nerve 1999;22:28–36.

23. Stevens JC, Dyck PJ. Sensory loss from whole sural nervebiopsy. Ann Neurol 1983;14:493–494.

24. Takahashi S, Yonezawa H, Takahashi J, Kudo M, Inoue T,Tohgi H. Selective reduction of diffusion anisotropy in whitematter of Alzheimer disease brains measured by 3.0 Teslamagnetic resonance imaging. Neurosci Lett 2002;332:45–48.

25. Tranmer BI, Bilbao JM, Hudson AR. Perineurioma: a benignperipheral nerve tumor. Neurosurgery 1986;19:134–138.

26. Tsang WY, Chan JK, Chow LT, Tse CC. Perineurioma: anuncommon soft tissue neoplasm distinct from localized hy-pertrophic neuropathy and neurofibroma. Am J Surg Pathol1992;16:756–763.

27. Wong BZY, Amrami KK, Wenger DE, Dyck PJB, ScheithauerBW, Spinner RJ. Lipomatosis of the sciatic nerve: typical andatypical MRI features. Skeletal Radiol 2006;35:180–184.


CASE OF THE MONTH ABSTRACT: Inflammatory myopathies (IM) are a heterogeneous group ofdiseases characterized by immune-mediated damage to skeletal muscle.Sensory abnormalities are rare in patients with IM. We report two patients,one with dermatomyositis and the other with inclusion-body myositis, whopresented with unexpected sensory abnormalities due to probable immune-mediated damage to dorsal root ganglia. We emphasize the importance ofcombined neuroimaging and neurophysiological assessment for proper di-agnosis.

Muscle Nerve 35: 721–725, 2007

MYOSITIS WITH SENSORY NEURONOPATHY

MARCONDES C. FRANCA Jr., MD, ANDREIA V. FARIA, MD, PhD,

LUCIANO S. QUEIROZ, MD, PhD, and ANAMARLI NUCCI, MD, PhD

Departments of Neurology, Radiology, and Pathology, School of Medical Sciences,Campinas State University (UNICAMP), “Zeferino Vaz,” Campinas, SP 13083-970, Brazil

Accepted 5 March 2007

Inflammatory myopathies (IM) are a heterogeneousgroup of diseases characterized by immune-medi-ated damage to skeletal muscle. Three major formsof IM are now recognized: dermatomyositis (DM);polymyositis (PM); and inclusion-body myositis(IBM).1 DM is essentially a humorally mediated dis-ease of endomysial capillaries, whereas PM and IBMare related to cell-mediated damage to muscle fi-bers.1 In IBM, autoimmune inflammatory and de-generative features coexist and are found through-out the course of the disease.2 Despite thesedifferences, proximal muscle weakness is the usualinitial complaint in all patients with IM. The tempo-ral course is subacute in DM/PM and chronic inIBM. In addition, a predilection for elderly patientsand therapeutic refractoriness are two distinctive fea-tures of IBM.

Sensory assessment, both clinical and neurophys-iological, in patients with IM is frequently normal.However, there are a few reports of sensory abnor-malities in IM,3–5 particularly IBM.4,5 These aremainly supported by nerve conduction studies

(NCS) or pathological analyses of sural nerve speci-mens, and clinical manifestations are generally mild.We report two patients with IM who had unusualsensory abnormalities and emphasize the impor-tance of combined magnetic resonance imaging(MRI) and neurophysiological assessment forproper diagnosis. We also highlight some conditionsthat must be included in the differential diagnosisand discuss possible pathogenetic mechanisms.

CASE REPORTS

Case 1. A 71-year-old man was evaluated because ofslowly progressive lower-limb weakness and numb-ness over the last 5 years. He was unable to climbstairs and frequently fell while walking. His past med-ical history was unremarkable and he had no historyof diabetes mellitus, thyroid dysfunction, alcoholconsumption, or long-term medication use.

On neurological examination, cognitive perfor-mance was adequate for age and cranial nerves werenormal. He had symmetric weakness of finger flex-ors and proximal lower-limb muscles, with markedquadriceps wasting, global areflexia, and sensory im-pairment of all modalities up to the knees and el-bows.

Laboratory studies revealed a slightly increasedserum creatine kinase (CK) level (290 U/L), eryth-rocyte sedimentation rate of 42 mm in the first hour,and positive anti-nuclear antibodies at 1 in 160 (ho-mogeneous pattern). Screening for Sjogren’s syn-drome, which included anti-Ro and anti-La antibod-ies, Schirmer’s test and parotid scintigraphy, andhuman immunodeficiency virus (HIV) and human

Abbreviations: CK, creatine kinase; CMAP, compound muscle action po-tential; COX, cytochrome c oxidase; CT, computerized tomography; DM,dermatomyositis; DRG, dorsal root ganglia; EMG, electromyography; HIV,human immunodeficiency virus; HTLV, human T-cell lymphotropic virus; IBM,inclusion-body myositis; IM, inflammatory myopathy; MRI, magnetic reso-nance imaging; MUAP, motor unit action potential; NCS, nerve conductionstudies; PM, polymyositis; SNAP, sensory nerve action potential; SND, sen-sory neuron disease; SS, Sjogren’s syndromeKey words: dermatomyositis; inclusion-body myositis; magnetic resonanceimaging; sensory neuronopathy; spinal cordCorrespondence to: A. Nucci; e-mail: [email protected]

© 2007 Wiley Periodicals, Inc.Published online 27 April 2007 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/mus.20783

Myositis with Sensory Neuronopathy MUSCLE & NERVE November 2007 721

T-cell lymphotropic virus (HTLV) serologies, werenegative. Thyroid function tests and blood cellcount, electrolytes, serum electrophoresis, fastingblood glucose, serum vitamin B12, and venous lactatewere normal. Cerebrospinal fluid examination wasunrewarding, including tests for neurosyphilis (Ve-nereal Disease Research Laboratory and Treponemapallidum hemagglutination assay). The patient un-derwent chest and abdominal computed tomogra-phy, upper gastrointestinal endoscopy, and colonos-copy to search for underlying neoplasia. Tumoralmarkers (carcinoembryonic antigen, alpha-fetopro-tein, prostate-specific antigen, and cancer antigen19-9) were within normal limits.

Motor and sensory nerve conduction studies areshown in Table 1. F-wave latencies in arms and legswere within the normal range but tibial H reflexcould not be obtained on either side. At rest, needleelectromyography (EMG) showed positive sharpwaves and fibrillation potentials in proximal anddistal muscles in all limbs. There were frequentshort-duration, polyphasic, and small-amplitude mo-tor unit action potentials (MUAPs) associated with afew large-amplitude and long-duration MUAPs inproximal muscles.

Quadriceps femoris biopsy showed mild endomy-sial mononuclear inflammatory infiltrate withoutperifascicular atrophy, some fibers with central nu-clei, rimmed vacuoles (sometimes several per fiber),and necrotic myofibers interspersed with regenera-tive ones (Fig. 1A and B). There were several cyto-chrome c (COX)–negative fibers but no ragged-redfibers or type grouping. Blood vessels had normalcaliber and showed no infiltrates.

Spinal cord magnetic resonance imaging (MRI)showed dorsal column (both gracile and cuneatus

fasciculi) hyperintense lesions on T2-weighted im-ages extending from C2 to T1.

The patient was started on azathioprine (3mg/kg daily) and followed on a regular basis. After1 year, there was a slight reduction of serum CKlevels but no modification of motor strength. Al-though sensory complaints improved, residual pro-prioceptive, vibratory, and tactile dysfunction arestill present.

Case 2. A 41-year-old woman was referred for neu-romuscular evaluation because of generalized weak-ness, myalgia, and hand paresthesias evolving over 2months. She had been healthy up to admission anddenied any relevant family history of neurologicaldisease.

There was proximal (grade 3 on the MedicalResearch Council scale) and distal (grade 4) muscleweakness associated with global areflexia when shewas first examined. Significant wasting and fascicu-lations were not found. Sensory testing revealedasymmetric areas of tactile and vibratory hypesthesiain the upper and lower limbs. Cranial nerve nucleiwere not affected. Gottron’s sign and a heliotroperash were found.

Serum CK level was increased (2300 U/L), but apanel of autoantibodies (including Jo-1, anti-Ro, andanti-La), erythrocyte sedimentation rate, thyroidfunction tests, and venous lactate were normal. RoseBengal staining of the cornea was normal as wasparotid gland scintigraphy. A systemic search forunderlying neoplasia that included computerized to-mography (CT) of the chest, abdomen, and pelvis aswell as mammography was normal. Upper gastroin-testinal endoscopy, colonoscopy, and tumor markerswere also normal.

Table 1. Motor and sensory nerve conduction studies in patients 1 and 2.

Normal Patient 1 Patient 2

SNAP (�V) NCV (m/s) SNAP (�V) NCV (m/s) SNAP (�V) NCV (m/s)

Sensory nervesRadial �12 �44 NR NR NR NRUlnar �12 �44 NR NR NR NRSural �5 �38 NR NR NR NR

CMAP (mV) NCV (m/s) CMAP (mV) NCV (m/s) CMAP (mV) NCV (m/s)

Motor nervesMedian �4 �49 16.4 50.0 0.75 52.3Ulnar �6 �49 13.1 56.2 0.97 50.1Peroneal �2 �41 5.7 43.7 1.10 47.2Posterior tibial �3 �41 6.8 44.8 0.72 47.8

SNAP, sensory nerve action potential; CMAP, compound muscle action potential; NR, no response obtained.

722 Myositis with Sensory Neuronopathy MUSCLE & NERVE November 2007

Nerve conduction studies are detailed in Table 1.Late responses (F waves and H reflex) were notobtained. Repetitive stimulation at 3 and 20 Hzshowed no abnormalities. Needle EMG revealedabundant positive sharp waves and fibrillation poten-tials at rest and short, polyphasic MUAPs with earlyrecruitment during voluntary contraction. Thesefindings were present in all muscles examined (del-toid, biceps brachii, flexor carpi ulnaris, rectus fem-oris, tibialis anterior), but were most pronounced inproximal muscles, especially deltoid and rectus fem-oris.

A biceps brachii biopsy revealed perifascicularatrophy, perivascular mononuclear cell infiltrates,capillary abnormalities and depletion, and myofibernecrosis and regeneration (Fig. 1C). As sensory ab-normalities are not usually seen in dermatomyositis,further neuroimaging investigations were under-taken. Spinal cord T2-weighted images revealed hy-perintense lesions in the posterior fasciculi (gracilisand cuneatus) at the cervical and thoracic levels(Fig. 2).

The patient was started on oral steroids (pred-nisone 1 mg/kg daily) with gradual recovery ofstrength and decline of serum CK levels. Althoughshe had almost complete motor recovery, she wasstill unable to walk unassisted due to persistent pro-prioceptive hypesthesia with sensory ataxia after 1year. At that time, extensive screening for underlyingneoplasia was again negative. She was later lost tofollow-up.

DISCUSSION

The first patient had a slowly progressive myositiswith rimmed vacuoles and marked involvement offinger flexors and quadriceps femoris, which are

FIGURE 1. Quadriceps femoris biopsy (patient 1). (A)Lymphomononuclear infiltrate among the muscle fibers, variationin fiber size and internal nuclei. Hematoxylin–eosin (H&E) stain;original magnification �100. (B) A muscle fiber containing sev-eral rimmed vacuoles. Biceps brachii biopsy (patient 2). (C) Se-vere non-specific perivascular inflammatory infiltrate in perimi-sium. Variation of fiber size and perifascicular atrophy. H&E stain;original magnification �100.

FIGURE 2. Cervical spinal cord MRI of the second patient.T2-weighted images showing hyperintense lesions in the dorsalfuniculi (arrow).


typical findings of IBM.6 In the second patient, bi-opsy and clinical features including classic cutaneoussigns were indicative of DM. Despite the distinctsubtypes of IM, sensory deficits had the same ana-tomical substrate in both patients. NCS and spinalcord MRI abnormalities indicated simultaneousdamage to central and peripheral sensory pathways,suggesting the dorsal root ganglia as the major site ofthe lesion.7 Diffuse areflexia in both patients and thetruncal sensory deficits found in the second case arenot typically found in axonopathies and indeed sup-port a proximal lesion. On clinical grounds, overtsensory ataxia was possibly overshadowed by motordeficits at onset. In case 2, sensory ataxia was moreeasily noticed at follow-up because of steroid-induced motor improvement.

In the first patient, the dissociation of abnormal-ities in motor and sensory NCS reinforces the possi-bility of a sensory neuron disease (SND).8 We believethe second patient also had SND despite the simul-taneous abnormalities found in sensory and motorNCS. In this individual, sensory impairment, mani-fested by complete absence of SNAPs, diffuseareflexia, and disabling sensory ataxia, would be bet-ter explained by damage to DRGs than by an ax-onopathy. However, low-amplitude CMAPs must alsobe explained and three possibilities merit consider-ation. Lambert–Eaton myasthenic syndrome is onepossibility but repetitive stimulation at 20 Hz did notreveal significant increments in CMAP amplitudes. Itis more likely that a superimposed motor axonopa-thy accounts for the low-amplitude CMAPs, but somefindings do not support this hypothesis. First, therewas marked and rapid clinical improvement on ste-roids, which is unusual in axonal neuropathies. Sec-ond, gracile and cuneate lesions were present andcannot be explained by damage to distal peripheralaxons. Third, the motor NCS abnormalities mayhave resulted from the myositis itself. Although un-likely, low-amplitude CMAPs may eventually befound in active forms of IM, especially in DM9,10 anddistal myopathies such as myotonic dystrophy type1.11 Follow-up NCS would have been useful to sup-port this hypothesis but unfortunately could not beperformed.

SNDs are a heterogeneous group of diseasescharacterized by DRG involvement and subsequentdegeneration of both peripheral and central projec-tions.7,12 The neurophysiological hallmark of theseconditions is a diffuse and non–length-dependentaxonal compromise of SNAPs with preserved motorNCS and needle EMG.8 In some patients, there areassociated motor abnormalities. Diagnosis basedsolely on NCS and clinical evaluation is challenging.

In such situations, spinal cord MRI has been em-ployed as a useful adjunctive diagnostic tool13 toshow lesions in the dorsal fasciculi caused by degen-eration of DRG central projections, as in our cases.

Some potentially treatable diseases have beenassociated with SND. Sjogren’s syndrome is the mostfrequent immune-mediated SND.7,12 In this situa-tion, sensory deficits evolve rapidly but immunosup-pressive therapy is effective.14 None of our patientscomplained of dry mouth or eyes and an extensivescreening excluded Sjogren’s syndrome. In additionto tropical spastic paraplegia, neurological manifes-tations of HTLV infection may include SND andmyositis, namely IBM,1,15 as seen in the first patient.His serological examinations for both HIV andHTLV were negative. Paraneoplastic SND is a well-characterized entity particularly associated withbronchogenic carcinomas and non-Hodgkin lym-phomas.7,12 In fact, SND was first described in 1948by Denny-Brown in patients with lung cancer whodeveloped subacute myositis and a sensory neurop-athy.16 Postmortem analysis revealed severe loss ofDRG cells and skeletal muscle inflammatory infil-trates resembling DM. Understanding of paraneo-plastic SND has since evolved and serum markershave been identified, such as Hu and CMRP-5 anti-bodies.17,18 In our patients, neoplasia was not foundeven after extensive investigation. Two additionalfeatures of the first patient that argue against para-neoplastic syndrome are the IBM phenotype that hasa weak association with tumors19 and a long fol-low-up period (�5 years) without malignancy. Neo-plastic diseases are more frequently related to DM asseen in the second patient. Paraneoplastic DM, how-ever, usually affects individuals older than 50 years ofage.19 Screening for anti-Hu antibodies would havebeen useful for the second patient, but she was lostto follow-up.

Despite this, few studies focusing on the associa-tion of SND and DM outside of an oncological con-text have been published.3 Based on histologicalfindings of sural nerve biopsies, some investigatorshave suggested that humoral immune activation andvasculitis of the vasa nervorum may be the underly-ing mechanism of axonal damage.10 As capillariessupplying the DRG have a loose blood–nerve barrier,pathogenic auto-antibodies could easily reach theDRG, which would explain the deficits in patient 2.We are unaware of any previous reports of an asso-ciation between IBM and SND, such as occurred inour first patient. Although a chance association can-not be firmly excluded, our report further extendsthe list of neuromuscular manifestations of IBM.20

724 Myositis with Sensory Neuronopathy MUSCLE & NERVE November 2007

This study was supported by Fundacao de Amparo a Pesquisa doEstado de Sao Paulo (FAPESP).

REFERENCES

1. Mastaglia FL, Garlepp MJ, Phillips BA, Zilko PJ. Inflammatorymyopathies: clinical, diagnostic and therapeutic aspects. Mus-cle Nerve 2003;27:407–425.

2. Dalakas MC. Inflammatory, immune, and viral aspects ofinclusion-body myositis. Neurology 2006;66(suppl 1):33–38.

3. Vogelgesang SA, Gutierrez J, Klipple GL, Katona IM. Polyneu-ropathy in juvenile dermatomyositis. J Rheumatol 1995;22:1369–1372.

4. Arnardottir S, Svanborg E, Borg K. Inclusion body myositis–sensory dysfunction revealed with quantitative determinationof somatosensory thresholds. Acta Neurol Scand 2003;108:22–27.

5. Hengstman GJ, van Engelen BG, ter Laak HJ, Gabreels-FestenAA. Familial inclusion body myositis with histologically con-firmed sensorimotor axonal neuropathy. J Neurol 2000;247:882–884.

6. Mastaglia FL, Phillips BA. Idiopathic inflammatory myopa-thies: epidemiology, classification and diagnostic criteria.Rheum Dis Clin N Am 2002;28:723–741.

7. Sghirlanzoni A, Pareyson D, Lauria G. Sensory neuron dis-eases. Lancet Neurol 2005;4:349–361.

8. Lauria G, Pareyson D, Sghirlanzoni A. Neurophysiologicaldiagnosis of acquired sensory ganglionopathies. Eur Neurol2003;50:146–152.

9. Amato AA, Dumitru D. Acquired myopathies. In: Amato AA,Dumitru D, Zwarts MJ, editors. Electrodiagnostic medicine,2nd ed. Philadelphia: Hanley & Belfus; 2002. p 371–1432.

10. Matsui N, Mitsui T, Endo I, Oshima Y, Kunishige M, Matsu-moto T. Dermatomyositis with peripheral nervous system in-volvement: activation of vascular endothelial growth factor(VEGF) and VEGF receptor (VEGFR) in vasculitic lesions.Intern Med 2003;42:1233–1239.

11. Pfeilsticker BH, Bertuzzo CS, Nucci A. Electrophysiologicalevaluation in myotonic dystrophy: correlation with CTGlength expansion. Arq Neuropsiquiatr 2001;59:186–191.

12. Kuntzer T, Antoine JC, Steck AJ. Clinical features and patho-physiological basis of sensory neuronopathies (ganglionopa-thies). Muscle Nerve 2004;30:255–268.

13. Mori K, Koike H, Misu K, Hattori N, Ichimura M, Sobue G.Spinal cord magnetic resonance imaging demonstrates sen-sory neuronal involvement and clinical severity in neuronopa-thy associated with Sjogren’s syndrome. J Neurol NeurosurgPsychiatry 2001;71:488–492.

14. Takahashi Y, Takata T, Hoshino M, Sakurai M, Kanazawa I.Benefit of IVIG for long-standing ataxic sensory neuronopa-thy with Sjogren’s syndrome. Neurology 2003;60:503–505.

15. Shimazaki R, Ueyama H, Mori T, Mori M, Fujimoto S, Kum-amoto T, Tsuda T. Chronic sensory neuronopathy associatedwith human T-cell lymphotropic virus type I infection. J Neu-rol Sci 2002;194:55–58.

16. Denny-Brown D. Primary sensory neuropathy with muscularchanges associated with carcinoma. J Neurol Neurosurg Psy-chiatry 1948;11:73–87.

17. Dalmau J, Furneaux HM, Rosenblum MK, Graus F, Poster JB.Detection of the anti-Hu antibody in specific regions of thenervous system and tumor from patients with paraneoplasticencephalomyelitis/sensory neuronopathy. Neurology 1991;41:1757–1764.

18. Yu Z, Kryzer TJ, Griesmann GE, Kim K, Benarroch EE, Len-non VA. CRMP-5 neuronal autoantibody: marker of lungcancer and thymoma-related autoimmunity. Ann Neurol2001;49:146–154.

19. Hill CL, Zhang Y, Sigurgeirsson B, Pukkala E, Mellemkjaer L,Airio A, et al. Frequency of specific cancer types in dermato-myositis and polymyositis: a population-based study. Lancet2001;357:96–100.

20. Amato AA, Gronseth GS, Jackson CE, Wolfe GI, Katz JS, BryanWW, et al. Inclusion body myositis: clinical and pathologicalboundaries. Ann Neurol 1996;40:581–586.


Documents

Muscle & Nervelib.ajaums.ac.ir/booklist/904492.pdf · BASIC ANATOMY AND PHYSIOLOGY OF THE SYMPATHETIC SYSTEM ... Sympathetic Innervation of the Skin. Compared to muscle, the sympathetic