Effects of regional infusion in certain painful states

Journal of Neurology, Neurosurgery, and Psychiatry, 1980, 43, 446-451

Effects of regional guanethidine infusion incertain painful statesL LOH, P W NATHAN,* G D SCHOTT, AND P G WILSON

From The National Hospital for Nervous Diseases, Queen Square, London

S U M M A R Y Guanethidine was infused by the regional intravenous technique into upper and lowerlimbs of patients with painful hyperpathic states, due to peripheral or central lesions. The relief ofpain and hyperpathia occurred within 20 minutes of infusion and lasted between three and 128hours; in a few patients, the relief has lasted for months. Great reduction in skin conductance andvasodilatation occurred, there being marked variation in the time of onset and duration of theseeffects. In some cases there was marked pilo-erection. Guanethidine given in this way did not com-

pletely block the sympathetic control of digital blood-vessels. There were no effects on sensibilityof normally innervated regions.

In a previous paper, Loh and Nathan' reportedthat the chronic painful states that were relievedby blocking the sympathetic outflow were thosecharacterised by the presence of hyperpathia andallodynia. t They concluded that this state ofabnormal sensitivity was being maintained by therelease of noradrenaline in the neighbourhoodof the peripheral nerves. The lesions in the caseswere damage to peripheral nerves or nerve rootsand, in a few cases, damage within the centralnervous system. In the work reported at thattime, the sympathetic outflow had been blockedin two ways: by the injection of local anaestheticsolutions into the sympathetic chain and ganglia,and by the infusion of guanethidine intravenouslydistal to a cuff occluding the circulation, asrecommended by Hannington-Kiff.2 The variouseffects of guanethidine reported at that time werenot considered. These actions of guanethidine,given by infusion into upper and lower limbs,are reported here. They are the times of onsetand duration of pain relief, vasodilatation, alter-ation in skin conductance and pilo-erection.

*Member of the External Scientific Staff, Medical Research Council.

tThe terms used are those recommended by the InternationalAssociation for the Study of Pain. Hyperpathia is "a painfussyndrome, characterised by delay, over-reaction and after-sensation to a stimulus, especially a repetitive stimulus."Allodynia is a state in which pain is caused by a stimuluswhich does not cause pain in normal skin.

Address for reprint requests: Dr PW Nathan, The National Hospitalfor Nervous Diseases, Queen Square, London WClN 3BG.

Accepted 22 January 1980

Here we are reporting only cases in which guan-ethidine removed the pain and hyperpathia.Although this paper is not concerned with thistechnique as therapy for these cases, it is to bestressed that six out of 30 of the patients havehad relief of this state for six months or more.

Patients and methods

Patients with chronic pain were selected on thebasis of having allodynia and cutaneous hyper-pathia; some also had deep hyperpathia. Thenumbers of patients and their diagnoses areshown in table 1. Of these 30 patients, one hadfour blocks given within a period of four weeks

Table 1

Diagnosis

Peripheral nerve lesionReflex sympathetic dystrophyCompression of nerve or nerve rootsSudeck's atrophySpinal cord lesionPainful legs and moving toes (3)CausalgiaUnclassified disorder of sympathetic systemPathological vascular state of periphery of limbAmputation stumpCarcinomatous invasion of brachial plexusArachnoiditisPain following removal of two herniated lumbar discsMidbrain injury due to closed head injuryMultiple sclerosis

Number ofpatients

654222II

ll

1I

TOTAL: 30

446

Protected by copyright.

on March 16, 2022 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.43.5.446 on 1 May 1980. D

ownloaded from

http://jnnp.bmj.com/

Effects of regional guanethidine infusion in certain painful states

and four had two blocks given within a periodof a week. Only patients in whom pain, hyper-pathia and allodynia were removed by the in-fusion of guanethidine are presented here.

Guanethidine given by Bier's blockThe limb was elevated for about a minute toempty the veins. A Kidde automatic tourniquetwas put on the arm or thigh, inflated to apressure above systolic blood pressure and main-tained for 20 to 25 minutes. Guanethidine mono-sulphate in isotonic saline was injected into avein of the hand or foot two minutes afterocclusion of the circulation. The dose was 10 to15 mg in 20 ml of saline for the upper limb and15 to 20 mg in 40 ml of saline for the lowerlimb. On one occasion 1-5 mg of guanethidinewas given in 30 ml of saline to an upper limb.There were no general effects from the guan-

ethidine on release of the cuff, except in onepatient. On one of the two occasions when thecuff was removed, she had abdominal pain feltin the region where she had had pain withcholecystitis.

Control procedures for the relief of painand hyperpathia by guanethidineIt was necessary to know that, under the con-ditions of infusing guanethidine, the relief ofpain, hyperpathia and allodynia were not due toa local anaesthetic action of guanethidine nor tothe occlusion of the circulation.

(a) As guanethidine has local anaestheticactivity, local anaethetic solutions weregiven on a few occasions by the sametechnique. Lignocaine 20 ml of 0 5 percent solution were given. Within threeminutes of deflation of the cuff, the anal-gesia had passed off and the pain and hyper-pathia had returned.

(b) In two of the patients of this series and inmany others with hyperpathia and allodynia,occlusion of the circulation was carried outfor 25 minutes without administering guan-ethidine. This sometimes relieved the con-dition for one hour but never for longerperiods.

Sensory eflectsPatients were tested by applying those stimulithat were most painful and produced the par-ticular quality of pain of which the patient com-plained. Pain was also examined by pinprick tothe skin and by squeezing the deep tissues. Touchwas examined by stroking and rubbing.

Digit temperatureThe temperature of the digits and interdigitalwebs was measured with thermocouples or mer-cury skin thermometers. Skin temperature wastaken to assess dilatation of digital blood vessels.

Digit plethysmographyThe neural control of the blood vessels of thefingers or toes was examined by mercury straingauge loops round the digits. The signal wasamplified and recorded on an ink-writing recorder.In order to induce a sympathetic response, thefollowing manoeuvres were employed: Valsalvamanoeuvre, hand-grip on a dynamometer, anevaporating cooling solution sprayed on face orneck; sudden auditory shock such as banging adoor; mental arithmetic; repeated pinprick orother painful stimuli in the locally painful area.The tests for vasodilatation, for control of skin

vessels of the digits, and for conductance werenot performed after all blocks. The reasons forthis were that it was only after we had doneseveral blocks for therapeutic purposes that wenoticed the large variation in the different effectsof guanethidine; and so at first the various testswere done only infrequently. The other reasonwas that with heated hospital wards, there wasusually full vasodilatation of the control, un-treated limbs, and so any dilatation of the treatedlimb could not necessarily be attributed to guan-ethidine.

Electrical conductance through the skinSkin resistance increases, and thus conductancedecreases, with decreasing sympathetic activity.The theoretical basis and the factors determiningconductance have been investigated by Woolley-Hart.4 5 Skin resistance was measured by anammeter (Relaxation Meter, Medici). Changes incurrent were measured by applying a constantvoltage. Two types of electrode were used, onefor the palm or sole and one for the digits. Theelectrode diameter was 1-5 cm. The distance be-tween palm or sole electrodes was 3-5 cm. Thedigit electrodes were mounted independently onVelcro strapping. The voltage was adjusted toproduce a current of either 25 ,LA or 50/A inthe control limb or in the affected limb beforeinjecting the guanethidine. Any change in cur-rent indicated a change in conductance.

Results

Relief of pain, hyperpathia and allodyniaPain, hyperpathia and allodynia were relievedwithin 18-23 minutes of the injection. Deep

447



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


448

tenderness was less affected. The relief of hyper-pathia often came on while the guanethidine wascausing severe burning pain. There was no effecton sensibility except for the removal of hyper-pathia and allodynia; and there was no effect onnormally innervated areas. For instance, in a casewith a lesion of the median nerve, the injectionof guanethidine removed the abnormal sensitivityfrom the median territory and had no effects inthe ulnar territory.The relief of the pain occurred while the cuff

was still on in 22 blocks and on removal of thecuff in the remaining 15 blocks. It is thought thatthere is no difference between these two popu-lations, for those who did not notice relief ofpain and hyperpathia until the cuff had been re-moved were probably too preoccupied with thevarious sensations and discomfort from having atight cuff on the limb to notice any change inthe underlying pain. The duration of relief ofpain and hyperpathia and allodynia is shown intable 2. In 24 out of 37 of the blocks, the relieflasted for less than 24 hours; in six patients ithas lasted for more than six months. From atherapeutic point of view, these results appearto be rather unsatisfactory. They are more sowhen one remembers that we are discussing hereonly the patients in whom guanethidine relievedpain and hyperpathia. We have seen at least afurther ten patients in whom it failed to do so.The injection in one patient of a tenth of the

usual dose (15 mg) had some slight effects inrelieving the sensitivity to stimulation but noother effects.

Vasodilatation and neural control ofdigital blood vesselsReactive hyperaemia following removal of thecuff was slight, usually raising the digital tem-perature by only 2°C.The time of onset of vasodilatation due to the

infusion of guanethidine is shown in table 3.Vasodilatation started to occur on removal of

Table 2 Duration of pain relief

3-12 h 13-24 h 25-60 h 51-128 h Months

Numberof blocks 14 10 5 2 6

Table 3 Onset of vasodilatation after removal of cuff

L Loh, P W Nathan, G D Schott, and P G Wilson

the cuff in four blocks; in table 3, this is called"immediate". The range of time of onset ofvasodilatation was wide, between 20 minutes and24 hours. This is, however, a somewhat arbitraryfigure as it took around one hour for full vaso-

dilatation to occur.The duration of vasodilatation is shown in

table 4. In most blocks this was for at least 48hours. This is an arbitrary figure for the timebetween full vasodilatation and return to thetemperature before guanethidine was usuallyabout 24 hours. By the times given in the table,the temperature of the two hands or two feetwas the same. To be sure that there was fullreturn of vascular tone, it would have beennecessary to put the patients in a cold environ-ment to see if vasoconstruction was equal in thetwo limbs. This was not done.

Tests of sympathetic control of the peripheralblood vessels showed complete block of sym-pathetic nerves in only one case. In all others,a partial block was shown by the fact that some

manoeuvres still caused some vasoconstriction.Partial block was shown within two hours of theinfusion of guanethidine in four, between twoand three hours in one, by four hours in one,and around 5.5 hours in one. Further evidencethat vasodilatation was incomplete was obtainedon two occasions by blocking the ulnar andmedian nerves with local anaesthetic solutionduring the time of the vasodilatation induced byguanethidine. In both cases, the local anaestheticblock produced a slight, further rise in digitaltemperature. Thus guanethidine did not com-

pletely block sympathetic control of the bloodvessels of the digits. This is unlikely to have beendue to an insufficient dose, as the same blockoccurred with 15mg as with 30mg.

Guanethidine often produced effects that couldhave been attributed to histamine release, suchas engorgement of the tissues, slight oedema anditching.

Skin conductanceThe effects of occluding the circulation on con-

ductance need to be considered. Edelberg6 foundthat during the time of occlusion conductanceusually stopped, and when the circulation was

restored conductance returned to its previous,or to a higher, level. In our cases, conductance

Table 4 Duration of vasodilatation

I d 2d 3d 4d

Number of blocks 6 8 3 1

Immediate Within 1I h 11-3 h 4-6 h 7-24 Iz

Numberof blocks 4 3 3 6 7



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


Effects of regional guanethidine infusion in certain painjul states

went down to zero with the current used whenthe cuff was on, and continued to remain so onremoval of the cuff in eight blocks. After fourblocks, conductance returned to its previous levelon removal of the cuff, and remained at thatlevel or two hours. After one block, it increasedwhen the cuff was removed. It was unrecordablewithin two hours in one block, within three hoursin one block, and within four hours in one block.We have insufficient evidence on the duration ofreduction of conductance.We did not undertake a study of sweating; for

sweating, being cholinergic, was not expected tobe affected by guanethidine. However, we ob-served that in limbs into which guanethidine hadbeen injected, sweating was sometimes presentand sometimes absent.

Pilo-erectionAn unexpected finding was marked erection ofthe arrectores pili muscles in the part treated byguanethidine. An example is illustrated in fig 1.It occurred in about a third of the trials. It cameon while the cuff was still on in two blocks inone patient, and immediately the cuff was re-moved in six patients. At these times, vaso-dilatation had not yet occurred. In four of thesepatients it lasted up to 24 hours and in one, itlasted eight hours.

Times of onset and duration of various effectsIt is clear that the onset and duration of variouseffects of guanethidine infusion vary consider-ably. The pain and hyperpathia are always re-moved either while the cuff is still on orimmediately it is released. At this time there isno block of sympathetic activity as judged by

skin conductance, and in most cases vaso-dilatation had not started to occur. The lack ofrelationship between relief of pain and onset andduration of vasodilatation was strikingly demon-strated in two patients in whom pain and hyper-pathia returned during the period of vaso-dilatation; a second infusion of guanethidinegiven during this period had the same thera-peutic effect as the first, the relief of pain andhyperpathia lasting between 12 and 36 hours.There was no relation between reduction in

conductance and vasodilatation. Reduction ofconductance sometimes occurred before vaso-dilatation; it often occurred as vasodilatationdeveloped. It returned to its normal level whilethe vasodilatation continued.

Discussion

Preventing sympathetic activity by injecting localanaesthetic solutions into the chain and gangliastops the pain, hyperpathia and allodynia, reducesconductance to a minimum, and causes vasodila-tation: these effects occur at the same time. Thelocal infusion of guanethidine has the sameeffects but they are spread out in time.The action of guanethidine is to release

noradrenalin and then to prevent re-uptake andthus to deplete it at noradrenergic varicosities,according to Chang et al.7

PainIn considering how guanethidine relieves certainhyperpathic states, one must first exclude anylocal anaesthetic action of the substance. Thefollowing points would seem to remove thispossibility: (a) in the control cases in which

Figure Pilo-erectionin left upper limbfollowing infusion ofguanethidine. Notethe sharp border atthe lower level of thecuff.

449



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


450

lignocaine or bupivacaine was injected, theremoval of hyperpathia and pain lasted onlythree minutes after removal of the cuff; (b)guanethidine has no effects on the sensibility ofnormally innervated areas below the cuff; (c)anaesthesia is not found over the cornea when5% guanethidine eyedrops are used; (d) if a localanaesthetic effect were occurring in our material,the small sympathetic fibres would have beenblocked immediately. But in the majority of ourtrials this was not so.Why in some lesions of the nervous system,

blocking the sympathetic nerve fibres stops theabnormal sensitivity and pain is unknown. Itis proposed that in the cases with hyperpathiaand allodynia, peripheral nerve fibres have be-come abnormal in that they have become sensi-tive to adrenaline and noradrenaline. Thissuggestion is based on experiments carried outby Wall and Gutnick8 9 on nerve fibres in aneuro-ma induced on the cut sciatic nerve of therat. After a certain time, some of the afferentmyelinated fibres started firing spontaneously.Noradrenaline and adrenaline increased theirfiring rate and activated these fibres when theywere quiet. This activity was stopped by alphaand beta blockers.

It seems to be that if there is a lesion any-where along an afferent pathway-in thethalamus, in the midbrain, within the spinalcord, in the posterior nerve roots or in theperipheral nerves-the peripheral nerve fibres maybecome abnormal. They are abnormal in thatthey become sensitive to mechanical distortion,liable to discharge impulses frequently andspontaneously, and abnormally sensitive tonoradrena,line and adrenaline. We do not knowwhen noradrenaline stores would have beendepleted following guanethidine given by Bier'sblock. Data from investigations of othersituations in species other than man provides uswith no evidence. It is likely that at the timewhen the spontaneous pain and hyperpathia hadbeen removed, the presumed firing of somaticafferent nerve fibres as a result of the effect ofnoradrenaline had ceased. Yet at that time, thesympathetic termina-ls would probably not havebeen depleted of transmitters.

Skin conductanceSkin conductance was examined as it is anothermeasure of sympathetic activity. The times ofonset of reduced conductance do not fit withwhat is known about times of depletion ofnoradrenaline. Conductance was so low that itcould not be measured after eight blocks on

L Loh, P W Nathan, G D Schott, and P G Wilson

removail of the cuff, but after nine blocks itreturned to its previous level, and started todiminish after two hours. Depletion of nor-adrenaline is said to occur in two hours andbecome fairly complete within 18 hours.10

VasodilatationWhen the sympathetic chain is blocked bylocal anaesthetic solution, the immediate effectis vasodilatation; it lasts between two and fourhours, and the skin vessels do not respond tomanoeuvres inducing sympathetic excitation.With guanethidine, however, the effect is pro-gressive. In a minority of trials, vasodilatationstarted to occur on removal of the cuff; in t;herest it came on at any time between 15 andsix hours. From the time when the digitaltemperature exceeded that of the control limbto full vasodilatation took around one hour. Itlasted from one to four days, and during thisperiod the vessels could be slightly constrictedby induced sympathetic activity. Sympatheticactivity, as judged by skin conductance, hadreturned to normal within 20 hours; butvasodilatation usually continued up to aboutfour days. Sweating was sometimes obvious at atime when there was full vasodilatation. It istherefore seen from these observations thatvasodilatation continued long after the expectedperiod of noradrenaline depletion. Thus it appears that the vasodilatation following guanethi-dine cannot be due only to sympathetic blockade.There is evidence that guanethidine can have adirect action on the blood-vessels. After depletionof noradrenaline stores by reserpine, guane,thidinestill causes a fall in blood pressure.'0 Abboudand Eckstein"' gave guanethidine intra-arteriallyto dogs that had previously been treated withreserpine or in which the limbs into which theguanethidine was injected had been denervated.They concluded that guanethidine "had a directvasodilator action on the blood vessels of thedog's denervated leg" directly affecting thevascular smooth muscle. This effect did notappear to be mediated through the release ofcatecholamines, histamine or acetylcholine. Anguset al'2 also noted the duration of vasodilatationafter guanethidine and tried to determine whetherit could be due to autonomic blockade. From anexamination of hindlimb blood flow and vascularresistance in the rabbit, they too concluded that"most of the vasodilatation following guanethi-dine is not mediated by the autonomic nervoussystem." However, they thought it was due tohistamine release, although they admitted that, inspite of their positive evidence, the vasodilatation



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


Effects of regional guanethidine infusion in certain states

was not influenced by H1 and H2 antagonists.Cooper et a13 injected guanethidine intra-arteriallyinto the upper limbs of man. They injected 5 mgin five minutes, a smaller dose than we used.They used one of the tests we used for causingvasoconstriction, ice applied to the neck. Therewas immediate vasoconstriction, which lasted for30 minutes. Maximal vasodilatation was presentat five hours and lasted for 20 hours. Judging bythe response to the ice, they concluded that ittook 30 minutes for the sympathetic control ofthe distal vessels to become complete. They con-cluded that "most of the vasodilatation followingguanethidine is not mediated by the autonomicnervous system," and they thought it was relatedto the release of histamine.

Pilo-erectionIn some of our cases there was a prominentexcitation of the arrectores pili of the limb.Guanethidine has no direct action on thesemuscles; this was shown by Hellman14 in anexcellent study of the pilomotor muscles of rats,mice, guinea pigs and cats. The innervation of themuscles was shown to be "wholly adrenergic."After giving bretylium or guanethidine, electricstimulation failed to excite the muscles; theyremained sensitive to adrenaline and noradrena-line; indeed, bretylium increased the responseto adrenaline though not to noradrenaline.Whether the early pilo-erection is due to therelease of noradrenaline by guanethidine is notknown.

Conclusions

In view of the different effects of guanethidiineinfusion to a limb occurring at such differenttimes, any general conclusions must be onlytentative. The hyperpathia and pain were alwaysstopped within minutes. Full blocking ofsympathetic control of digital blood vessels didnot occur. The prolonged vasodilatation may notbe due to sympathetic nerve blockade but toother mechanisms.

We are grateful to Ciba-Geigy Limited for sup-

plies of guanethidine and for information; andto Dr Marion C Smith for helpful criticism ofthis paper.

References

1 Loh L, Nathan PW. Painful peripheral states andsympathetic blocks. J Neurol Neurosurg Psy-chiatry 1978; 41:664-71.

2 Hannington-Kiff J. Intravenous regional sympa-thetic block with guanethidine. Lancet 1974; I:1019-20.

3 Spillane JD, Nathan PW, Kelly RE, MarsdenCD. Painful legs and moving toes. Brain 1971;94:541-56.

4 Woolley-Hart A. The role of the circulation inmeasurements of skin conductivity. Brit J Der-matol 1972; 87:213-26.

5 Woolley-Hart A. A simple technique for measur-ing skin conductivity. Med Biol Engin 1972; 10:561-2.

6 Edelberg R. Independence of galvanic skinresponse amplitude and sweat production. JInvest Derm 1964; 42:443-8.

7 Chang CC, Costa E, Brodie BB. Interaction ofguanethidine with adrenergic neurons. PharmacolExp Ther 1965; 147:303-12.

8 Wall PD, Gutnick M. Properties of afferent nerveimpulses originating from a neuroma. Nature1974; 248:740-3.

9 Wall PD, Gutnick M. Ongoing activity in per-ipheral nerves; the physiology and pharmacologyof impulses originating from a neuroma. ExpNeurol 1974; 43:580-93.

10 Carrier 0. Pharmacology of the PeripheralAutonomic Nervous System. Chicago: Year BookMedical Publishers, 1972.

11 Abboud FM, Eckstein JW. Vasodilator actionof guanethidine. Circulation Res 1962; 11:788-96.

12 Angus JA, Bobik A, Korner PI, Stoneham MT.Guanethidine-induced vasodilatation in the rabbit,mediated by endogenous histamine. Br J Phar-macol 1978; 62:7-17.

13 Cooper CJ, Fewing JD, Hodge RL, Whelan RF.Effects of bretylium and guanethidine on humanhand and forearm vessels and on their sensitivityto noradrenaline. Br J Pharmacol 1962; 21:165-73.

14 Hellman K. The isolated pilomotor muscles as anin vitro preparation. J Physiol (Lond) 1963; 169:603-20.

451



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


Documents

Effects of regional infusion in certain painful states