F70 Archives of Disease in Childhood 1996; 74: F70-F76

Use of evoked potentials in preterm neonates

M J Taylor, E Saliba, J Laugier

INSERM Unite 316,et Centre de P6diatrieGatien de Clocheville,Tours, FranceM J TaylorE SalibaJ Laugier

Correspondence to:Dr M J Taylor, Division ofNeurology, The Hospital forSick Children, 555University Avenue, Toronto,Ontario M5G 1X8, Canada.

Premature infants are at high risk of neuro-logical dysfunction, yet clinical neurologicalevaluation is difficult in this population. Evokedpotentials provide objective measures of CNSfunction, and can be an important adjunct tothe clinical examination in preterm infants.Evoked potentials are averaged electrical

responses to sensory stimuli that can berecorded from peripheral and central aspectsof the nervous system. Evoked potentials areobtained in response to repetitive stimuli(auditory, visual, or somatosensory), and are anon-invasive means of examining the func-tional integrity of the sensory pathways withinthe nervous system. Norms are mandatory asthe evoked potentials change rapidly in theneonatal period, reflecting maturation of thenervous system. Abnormalities in latencies andamplitudes of the evoked potential waveformsreflect damage or dysfunction along thesepathways. In this paper we review briefly themethodology for preterm evoked potentials,early developmental changes, and someemerging clinical applications.

Methodology and early developmentAs numerous methodological details differ inrecording evoked potentials in pretermneonates and in children or adults, methodsfor recording the various types of evokedpotentials will be reviewed. There are consid-erable differences among norms from variouslaboratories as a result of such factors as band-pass and wave identification in the neonates.The difficulty is increased by the very rapidmaturational changes in the evoked potentialsover the preterm period. Thus it is importantto establish norms for each laboratory, or care-fully to follow the methods of the laboratoryonce they have been established. The nomen-clature of the component waves of the variousevoked potentials differ from laboratory tolaboratory. An 'N' denotes a negative wave,followed by a number (for example, N20),occurring at 20 milliseconds (ms) afterstimulus presentation. Similarly, 'P200' woulddenote a positive peak occurring at 200 ms.However, some laboratories only order thesuccessive peaks (N1, P1, N2, etc), whileothers use the mean latency from the age groupwith which they are working. The exception isthe BAEP, as virtually all laboratories useRoman numerals to denote the positivecomponents.

BRAINSTEM AUDITORY EVOKED POTENTIALS(BAEPS)The BAEP reflects conduction along theauditory brainstem pathway. Five componentwaves are routinely measured: waves I and II

arise from the distal and rostral portions of theeighth nerve, respectively; wave III is from thepons and waves IV and V are from the mid-brain. Broad band clicks are the only stimulifor which there are normative data for infants.Click intensity should be calibrated relative tonormal hearing thresholds (nHL) for adults inthe laboratories. The earphone should be heldabove the ear, because the weight of the ear-phone can collapse the ear canal in prematureinfants. A recognisable BAEP can be recordedin premature infants of 28-30 weeks in gesta-tional age,I 2 with stimuli presented at highintensities and slow rates.The BAEP amplitude is smaller than in full

term neonates; the latencies of all the compo-nents decrease with increasing conceptionalage.3 4 Wave V shows greater age relatedlatency decreases than wave I; hence, the I-Vinterpeak latency (central conduction time)also decreases with age. In neonates wave I canbe double peaked, wave II replaced by aprominent negativity after wave I, and wavesIII, IV, and V usually merged. Despite overallamplitudes being low, the wave I:V amplituderatio can be high, and the relative amplitudeschange during the preterm period.5

Latencies decrease rapidly in preterminfants6 7 and prematurity itself can affect thelatencies. Yamamoto et a18 found noticeabledecreases in the wave I latencies and in I-IIIinterpeak latencies related to extrauterinerather than conceptual age. Collet et a19 sug-gested that smaller head size could account forthe shorter BAEP interpeak latencies that theyalso found in preterm infants with longerextrauterine life, although this is unlikely to bethe only explanation. The effects of prema-turity, even of only a few weeks, can extend forsix months.'I

VISUAL EVOKED POTENTIALS (VEPS)The VEP is a cortical response measured overthe occipital lobes to flashing light or patternedvisual stimuli. The primary componentsmeasured in neonates to flash are P100, P200,and N300, and to pattern the P280. Patternstimuli are more sensitive measures of visualfunction and the only way accurately to estimatevisual acuity neurophysiologically, however, thisrole of the pattern VEPs appears reliable onlyafter term."I Thus in preterm neonates it is theflash VEPs that are widely used.

Flash VEPsFlash VEPs are usually recorded in responseto red light emitting diodes (LED) set intogoggles.'2 13 The variability of flash VEPs ishigh, and the range of normal is wide. Therecording epoch needs to be long, as the

on July 18, 2021 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.74.1.F

70 on 1 January 1996. Dow

nloaded from

Evoked potentials in preterm neonates F71

components appear at 200-500 ms; a sweep of1 second is recommended. The waveforms arelarge, however, and only a few trials (15-30)are needed per average. Pryds et al 14 reportreliable VEPs in premature infants recordingonly three trials, using very long interstimulusintervals. With faster rates of presentation,there can be decreases in amplitude andincreases in latency.'5

During the premature period the LED VEPsshow remarkable developmental changes.Starting at about 24 weeks of gestational agethere is -only a single, late negative peak at300 ms (N300). A late positive peak emerges(P450) by 26 weeks and by 37 weeks an earlierpositive peak at 200 ms (P200) is seen.16 17Taylor et al 17 found no age related changes inlatencies of the VEP over the preterm period;Pryds et al,18 studying preterms in the first dayof life, found latencies and amplitudes of theN300 to be inversely related to age. The P200emerges at 200 ms during the late pretermperiod, and decreases in latency after term.The morphological changes are thought to

index maturational changes in the occipitallobes.'7 19 There seems to be more rapidmaturation of the visual pathway in prematureneonates in the extrauterine environment, ascomponents emerge earlier in longitudinalthan cross-sectional series.17 Thus VEP normsneed to be established for infants in the prema-ture period, as a function of their gestationalage at birth. The effects ofpreterm birth on theVEPs have been shown to last for severalmonths by some researchers.10

Flash VEPs can be recorded in sleepingpreterm infants, but significant amplitudedecrease may be seen with deep sleep.20 TheVEPs are otherwise very resistant to variousperinatal factors,'8 and are the most readilyrecorded of the evoked potentials in pretermbabies.

Pattern reversal VEPsIn the late preterm period VEPs can beobtained with a small portable televisionmonitor, with large (20) checks. Due to thelarge amplitude of the VEPs only about 20trials per average are required.I1 21 The wave-forms are distinct from those ofthe flash VEPs,containing only a single late positive wave(P280). Latency decreases correlate with post-menstrual or gestational age,"I 21 but inpreterm infants do not reflect refractive error.

SOMATOSENSORY EVOKED POTENTIAIS (SEPS)SEPs assess the functional integrity ofperipheral, spinal, subcortical and corticalaspects of the somatosensory pathway. Mediannerve (MN) SEPs are generally very well toler-ated in infants; posterior tibial nerve (PTN)SEPs are considerably more difficult and timeconsuming to record.

Median nerve SEPsMedian nerve SEPs can be recorded inpreterm neonates from 25 weeks gestational

age on wards, using a low bandpass (1-100HZ) and long sweep 200-300 ms, with eitherunilateral or bilateral stimulation.22-24 Inpreterm neonates under 30 weeks gestationalage the N20 has a latency of over 70 ms,22 anddecreases very rapidly over the prematureperiod. Most studies, however, have includedprimarily infants older than 30 weeks gesta-tional age.22 25 26 Differences between termand preterm infants tested at the same concep-tual age have not been found22 27 28 apart fromtransitory latency increases in the first week oflife.29 Birthweight, however, influences theSEPs; more mature waveforms (and shorterlatencies) correlate with weight at birth.27Several other authors have also reported wave-form complexity increasing with increasingpreterm age.25 30 However, Karniski et a126 inan extensive study, found that cortical com-ponents and scalp topography changed little,except in amplitude, over the preterm period.The component that they described as N2(equivalent to the Ni or N20 in other studies)was very stable in topography; they also foundan early N1, P2, and a later P2 in all thepreterm infants (31-40 weeks gestational age)studied. The differences between this studyand those that have found morphologicalchanges with age may be due to recordingparameters. That study was one of the few thathas recorded in response to bilateral stimula-tion.26 This may well produce a more stablecortical distribution, and as often in clinicalpractice, information regarding functionalasymmetries is sought, unilateral stimulationis likely to be more frequently used.Nevertheless, this is an impressive study, andsuggests that by 31 weeks of age, the orienta-tion of the dipoles in the somotosensory cortexis mature; latency and amplitude changesreflect the ongoing myelination process.Some authors have reported that 10-12% of

premature infants may not have consistentlyrecordable SEPs.2223 This has not been ourexperience. We have recently investigated thematurational patterns of the cortical SEP in theearly preterm period (27-31 weeks gestationalage).24 At birth, all of the infants (n=20) hadnormal neurological examinations, no clinicalor laboratory evidence of asphyxia, normalcranial ultrasound scans, and on follow up (atperiods of 4-18 months corrected age), hadnormal neurodevelopment. SEPs were seen inall of the infants. Probably the very strict inclu-sion criteria were critical to our findings ofSEPsin all the infants. The waveform consisted ofpositive, negative, and positive components (at26 weeks at 80 ms, 110 ms, and 188 ms, respec-tively). All three of these components (probablythe P15, N20, and P22) decreased rapidly inlatency over the preterm period, such that by 32weeks ofgestational age they were at 42, 55, and113 ms, respectively. As with studies in theolder preterm infants, there seemed to be no dif-ferences in the maturational changes as a func-tion of gestational age at birth.

Posterior tibial nerve SEPsMany studies on lower limb SEPs in neonates

on July 18, 2021 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.74.1.F

70 on 1 January 1996. Dow

nloaded from

F72 Taylor, Saliba, Laugier

have had difficulty obtaining responsesreliably, with success rates around 55%, lowerwhen preterm babies below 30 weeks wereincluded and higher when post-term babieswere included.31-33 This may be due to tech-nical reasons. Using a lower bandpass (as isuseful for median nerve SEPs in preterminfants), White and Cooke34 reported thatsuccessful PTN SEPs were recorded in 93% ofinfants born between 26 and 41 weeks ofgestational age and tested within the first weeksof life; reliable normative data were obtainedafter 27 weeks of gestational age. Similarly,Pike and Marlow (personal communication,1993), using a 1-100 bandpass, 25-100stimuli per average (at a rate of 0-5 Hz),reliably obtained unilateral posterior tibialnerve SEPs in all (n=93) normal pretermsfrom 26 weeks of gestational age. Latencies ofthe spinal SEP decrease steadily over thepreterm period.3' 34 Thus, despite the increas-ing size of the neonates, the latencies decrease,reflecting the substantial effect of myelinationand maturation. The cortical components alsodecrease in latency over this period.3' 34 Theprimary cortical component (P1 or P35)occurs at a latency of 75-85 ms in the youngestpreterms, but has shortened to about 35 ms byterm.33 34 It is also reported that, unlike themedian nerve SEPs, prematurity can have sub-stantial, long-lasting effects on the posteriortibial nerve SEPs,32 although this latter studyrecorded only with high bandpass and asuccess rate of 57% in the neonates.

Clinical applicationsAUDIOLOGYEven mild hearing impairment early in life cansignificantly affect the normal development ofspeech and language, if undetected for morethan six months.35 It is therefore very import-ant that premature neonates, who are at risk ofhearing impairment, be evaluated during theneonatal period.636 The BAEP is the mostwidely used and accurate technique for assess-ing hearing in neonates, but the BAEPs needto be recorded after the infants have reachedterm and before 3 months corrected age.During the premature period, the BAEPs arenot reliable predictors of audiological orneurological sequelae.6 37 The recording pro-cedures for neonatal screening have been welldescribed.38-40 Some recent studies have inves-tigated the viability of other techniques as ameans of audiological assessment in pretermneonates, and although they look promising,these are not yet established measures.41-43

DIAGNOSTIC AND PROGNOSTIC VALUE INLESIONS OF PREMATURITYIn order to predict neurodevelopmental out-come in preterm infants, numerous methodshave been proposed.44 Currently, one of themost widely used is ultrasound imaging,because of its ready applicability and use.45The reliability of techniques such as ultra-sound, however, remain imperfect prognosti-cally. The sensitivity, specificity, and positive

predictive value of these findings in deter-mining an adverse outcome in preterm infantsranges from 850/o-91% for diffused increasedechogenicities on ultrasound; 560/%-86% forabnormal resistive index with Doppler; and560/6-96% with abnormal EEG findings.45Prediction of good outcome has been lessemphasised in published findings. Evokedpotentials have not been thoroughly investi-gated as prognostic measures in preterminfants, and may prove to be valuable in sup-plementing the information from other clinicalmeasures.A number of authors have demonstrated the

value of VEPs and SEPs for assessing corticaldamage and predicting outcome in theasphyxiated term newborn.30 4"8 In termasphyxia evoked potentials can predict bothpoor and good outcome. Abnormal VEPspredict poor outcome,47 while normal SEPscan predict good outcome.46 48 These studiesdo not seem to be generally applicable, how-ever, to preterm infants.7A recent study investigated the accuracy of

VEPs in predicting neurodevelopmental out-come in preterm infants (Ekert P G, abstractpresented at IEPS V, Milan, September 1994).VEPs were recorded in the first and secondweeks of life, in 45 followed up to a mean ageof 21 (4) months. Abnormal outcome (deathor cerebral palsy (CP) and/or corrected mentaldevelopmental index (CMDI) <84) occurredin 17 infants (30%). Multiple regressionanalyses showed that periventricular leuco-malacia (PVL) (P=0002) and birthweight(P=0-014) were predictive of adverse out-come. VEPs did not predict adverse outcome(with a sensitivity and accuracy of 33% and63%, respectively). Abnormal VEPs in infantswith head ultrasound abnormalities were alsonot predictive of adverse outcome, with sensi-tivity and accuracy of 18% and 65%, respec-tively. Mean CMDI in infants with andwithout VEP abnormalities were not signifi-cantly different (P=0.57). Thus, in contrast tothe term asphyxiated infant, VEPs were poorpredictors of neurodevelopmental outcome inpreterm infants. An exception is the study byPlaczek et al,16 who found that VEPs corre-lated with visual acuity measures in preterminfants, as well as neurological development,although the follow up period was very short.Pike et al 49 also reported that the VEPrecorded at term in infants born prematurelycould contribute to the prediction of visualimpairment; this was an extensive study onvisual outcome, and not on general neurologi-cal sequelae. De Vries et al50 had previouslyshown that abnormal or absent VEPs inpreterm infants with PVL predicted abnormalvisual outcome.The prognostic role of SEPs in preterm

infants has also been investigated. Due to thepathophysiology of brain damage in preterminfants it is expected that the SEPs would be abetter measure of the damage and more sensi-tive to long term sequelae than VEPs. Thesomatosensory afferent pathways pass throughthe periventricular region, which is most at riskof injury in the preterm brain - intraventricular

on July 18, 2021 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.74.1.F

70 on 1 January 1996. Dow

nloaded from

Evoked potentials in preterm neonates F73

haemorrhage (IVH) and periventricular leuco-malacia (PVL). PVL is strongly associated withadverse neurological outcome in preterminfants. PVL is often heralded by early headultrasound scan changes (periventricularechoes (PVE) which may or may not resolve,hence delaying the definitive diagnosis ofPVL. Abnormal SEPs may be associated withdeveloped PVL, or may predict the develop-ment of PVL.

Klimach and Cooke5l found a close relationbetween SEPs recorded close to term and out-come at 6-16 months: the head ultrasoundscans and clinical picture were much lessreliable prognostically. Infants with asym-metrical SEPs developed hemiplegia, whichwas not necessarily the case with infants withunilateral lesions. Bilateral SEP abnormalitiesalways predicted neurological sequelae.However, in a larger series, de Vries et a152studied primarily preterm infants with cysticPVL and found the SEPs recorded close toterm were no more reliable than ultrasono-graphy.

Consistently abnormal SEPs did not alwayspredict sequelae, nor did normal SEPs guaran-tee normal outcome, although the specificity ofSEPs was still 92%. This research alsoincluded repeat studies within the pretermperiod in a proportion of infants, and in the 19with persistently abnormal SEPs only two werenormal on outcome. Pierrat et al53 recordedSEPs in preterm infants with extensive cysticleucomalacia. They also found that the SEPsrecorded at term were not useful measures,whereas those cases in whom they hadrecorded SEPs during the earlier pretermperiod were all abnormal, consistent withthe outcomes. De Vries et al 52 suggested thatthere may be a role for SEPs in the assessmentof preterm infants at risk of neurologicalsequelae, but that the studies would need to berepeated during the preterm period. This wasthe protocol of a recent study (Ekert PG, et al,paper submitted).We recorded SEPs in the first and second

weeks of life in 88 preterm infants (<32 weeksgestation) at a mean age of 6-5 days for the firststudy and 13-4 days for the second study(Ekert P G, submitted for publication). Themean (SD) weight was 948 (27) g and themean gestational age was 27-5 weeks.Abnormal SEPs recorded early in life are influ-enced not only by maturational factors, suchas gestational age and postnatal age, but alsoby periventricular haemorrhage or PVE.Intraventricular haemorrhage was associatedwith changes in the SEP. Grades I and IIhaemorrhages are associated with eithernormal SEPs or the loss of some late com-ponents. More clinically important haemor-rhages, including those involving theperiventricular white matter (grades III andIV) may be associated with the loss of earlycomponents as well. The association betweenabnormal SEPs and subsequent cystic PVLwas significant (P<OOO14). Abnormal SEPswere also associated with PVE (P<OOO12);absent components were associated with PVE(P<O 004) but not with subsequent PVL

(P<0 123). However, abnormal SEPs in thesecond week of life were strong independentpredictors of cerebral palsy on follow up(P<0-025). The fact that the SEPs in thesecond week of life are more reliable predictorsis consistent with the findings of Pierrat et al.29These data suggest that the SEPs are sensitive

measures of perturbations in cerebral functionin preterm infants, but probably because of theresilience ofpreterm brain, very early abnormal-ities may not necessarily predict longer termoutcome; the SEPs recorded close to term arealso not highly predictive. Thus, there may beonly a relatively narrow window, definitelyduring the second week of life as seen in theabove study (Ekert P G, submitted for publica-tion), in which the SEPs should be recorded forpredictive purposes in this population.As sequelae more often affect the lower than

upper limbs, the posterior tibial nerve (PTN)SEPs may prove more valuable predictors ofoutcome in premature infants. In a very recentstudy White and Cooke54 have shown that infact PTN SEPs were accurate predictors ofoutcome at 3 years of age. The study includedprimarily preterm infants (median gestationalage of 30 weeks), with the SEPs being recordedat the end of their hospital stay (at a mean of37*5 weeks). There was a significant correla-tion between abnormal PTN SEPs and thesubsequent presence of cerebral palsy. Allinfants but one with bilaterally normal SEPshad normal outcomes (24/25), and of those(n= 14) with bilaterally abnormal PTN SEPs,12 had cerebral palsy and one global delay ofinfants with unilateral abnormalities (n= 9),five were normal on outcome. Thus, these datasuggest that, in fact, PTN SEPs will be morevaluable than any of the other evoked poten-tials in predicting neurological sequelae inpreterm infants; the drawback is the tremen-dous time investment that is needed to performthese studies in this population.

APNOEAA few studies have examined BAEPs inpreterm infants, including some with apnoeicepisodes.55 56 Henderson-Smart et al 57 studiedexclusively apnoeic infants and found abnor-mal BAEPs in preterm infants with apnoea,and that when the apnoeas ceased, the BAEPsreturned to within normal limits. They sug-gested that apnoea in preterm infants is relatedto neural function of the brainstem. However,the suggestion that the BAEPs can reflectimmaturity of the brainstem pathways needs tobe tempered with the result of other studiesthat have found the BAEPs to be too variablein the preterm period to be consideredaccurate measures of CNS maturity.7 58 Chenet a159 found that BAEP interpeak latenciesdecreased with aminophylline treatment inapnoeic infants, but that these decreases didnot predict which preterm infants with apnoearesponded to the treatment. Further studyis needed on this question, given the numberof these infants that suffer apnoeic episodes,and the lack of clear understanding as to theaetiology of preterm apnoea.

on July 18, 2021 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.74.1.F

70 on 1 January 1996. Dow

nloaded from

F74 Taylor, Saliba, Laugier

HYPERBILIRUBINAEMIANeonatal hyperbilirubinaemia can lead toserious neurological sequelae and auditoryimpairment,60 61 but the toxicity ofbilirubin onthe neonatal CNS depends on many factors,and outcome cannot be accurately predicted.Thus, investigators have examined evokedpotentials particularly BAEPs, in hopes thatthey may indicate whether the infant wouldbenefit from exchange transfusion or if theinfant was likely to suffer neurological oraudiological impairment. A concern, particu-larly in preterm infants, is to determine whatbilirubin concentrations are safe. ProlongedBAEP interpeak latencies while infants arehyperbilirubinaemic, and improvement follow-ing exchange transfusions, have been widelyreported.62-64 In these studies, however, BAEPabnormalities were not always correlated withbilirubin concentrations, and follow up datawere not obtained to determine if these tran-sient abnormalities were significant (transientBAEP abnormalities are frequently seen in thepreterm population).6 A few recent reportsfound no relation between bilirubin andBAEPs.65 66 The latter study suggested that theotoneurological toxicity of bilirubin is relatedto the bilirubin binding properties of serumalbumin. A replication of these data with alarger series would be important to confirm therole of BAEPs in hyperbilirubinaemia.

SMALL FOR GESTATIONAL AGE (SGA)Evoked potentials have been used to assess theeffects of intrauterine growth retardation onneurological development. Watanabe et al67recorded VEPs and suggested that neurologi-cal development in SGA infants progressedrelatively normally and independently of theintrauterine conditions which caused theslower physical growth. In contrast, Hrbeket al68 found that although the VEPs of SGAinfants generally corresponded to that of agematched normal neonates, there was a ten-dency towards longer latencies in the SGAinfants, as well as the presence of late slowwaves, which the authors felt were a sign ofneurological immaturity. Both of thesefeatures, however, were transient. Pryds et al 18examined VEPs in eight SGA infants andfound no abnormalities. However, Pettigrewet al 69 and Soares et al 70 found that the BAEPinterpeak latencies were shorter in SGA infantsthan age matched controls, although Soareset al 70 suggested that this was not due to pre-cocious development of auditory brainstempathway, but to immaturity of the cochlea.This was because the shorter I-V interval wasdue to a longer latency wave I, and not to ashorter than usual latency of wave V. Sardaet at7' found that the BAEPs in SGA infantswere closely associated with the presence orabsence of maternal hypertension, suggestingthat aetiology of the growth retardation affectsthe evoked potentials. As other studies havenot correlated aetiological factors with theevoked potentials, this may explain thereported variability.

Gallai et alt25 found no differences in SEPs

between SGA and normal babies, both inpreterm and term populations, although amore recent study found a proportion of SGAinfants to have abnormal SEPs (de Vries L S,1993; personal communication). These abnor-malities were not associated with neurologicalsequelae in infants with normal head ultra-sound scans, but in a smaller number withboth abnormal head ultrasound and abnormalSEPs, cerebral palsy was seen. In the SGAbabies growth retardation does not seem toaffect the CNS routinely or uniformly, accord-ing to the evoked potentials, but as someauthors have found abnormalities in theevoked potentials, not always associated withnervous system disorders, some caution needsto be exercised in interpreting studies in thispopulation.

ConclusionsThis paper has reviewed the techniques usedfor recording evoked potentials in the prema-ture infant and the early developmentalchanges. The maturational changes in theevoked potentials, including morphologicalchanges, and the very rapid latency changeswithin the first months of life, provide aninvaluable means for assessing and monitoringdevelopment within the central nervoussystem. The maturational changes are suchthat normative values are requisite, and thenorms must take into account both the infant'sgestational age at birth as well as the posmatalage. These norms can then be used to aid inthe assessment of gestational age, and whetherthere has or has not been normal maturationaldevelopment, either in utero or during thepostnatal preterm period.

Evoked potentials are of increasing valueclinically in preterm neonates, primarilybecause of the difficulty in obtaining reliableneurological evaluation of these infants.Median nerve SEPs may provide reliable infor-mation in preterm infants at risk of PVL, andwhen recorded in the second week of life,predict cerebral palsy. PTN SEPs seem to beeven more reliable indicators of outcome, butthe difficulty in obtaining them in preterminfants needs to be taken into consideration.Further study is needed in some areas, such asin apnoeic preterm babies clearly to establishthe role that evoked potentials (in this caseBAEPs) may have in understanding both theaetiology and the clinical course of this dys-function. In other conditions, such as delayedintrauterine growth, that may lead to neuro-logical sequelae, evoked potentials can provideobjective CNS assessment. Evoked potentialsmay also prove useful in the monitoring oftreatment modalities for preterm infants.72-74The evoked potentials are a valuable adjunct inthe assessment of preterm neonates and, astheir value is recognised, we expect their use toincrease.

1 Starr A, Amlie RN, Martin WH, Sanders S. Development ofauditory function in newborn infants revealed by auditorybrainstem potentials. Pediatris 1977; 60: 831-9.

2 Schulman-Galambos C, Galambos R. Brainstem auditoryevoked responses in premature infants. JX Speech Hear Res1975; 18: 456-65.

on July 18, 2021 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.74.1.F

70 on 1 January 1996. Dow

nloaded from

Evoked potentials in preterm neonates F75

3 Fawer C, Dubowitz LMS. Auditory brain stem response inneurologically normal preterm and fuilterm newbominfants. Neuropaediatrics 1982; 13: 200-6.

4 Mayer M. Potentiels evoques et electromyographie en pidiatrie.Paris: Masson, 1990.

5 Salamy A, Mendelson T, Tooley W. Developmental profilesfor the brainstem auditory evoked potential. Early HumDev 1982; 6: 331-9.

6 Picton TW, Taylor MJ, Durieux-Smith A. Electrodiagnosis inclinical neurology. 3rd edn. London: Churchill, 1992.

7 Mercuri E, von Siebenthal K, Daniels H, Guzzetta F,Casaer P. Multimodality evoked responses in the neuro-logical assessment of the newbom. Eur J Pediatr 1994;153: 622-31.

8 Yamamoto N, Watanabe K, Sugiura J, Okada J, Nagae H,Fugimoto Y. Marked latency change of auditory brain-stem response in preterm infants in the early postnatalperiod. Brain Dev 1990; 12: 766-9.

9 Collet L, Soares I, Morgon A, Salle B. Is there a differencebetween extrauterine and intrauterine maturation onBAEP? Brain Dev 1989; 11: 293-6.

10 Uysal S, Renda Y, Topcu M, Erdem G, Karacan R. Evokedpotentials in full-term and premature infants: a compara-tive study. Child Nerv Sys 1993; 9: 88-92.

11 Grose J, Harding GFA, Wilton AY, Bissenden JG. Thematuration of the pattern reversal VEP and flash ERG inpre-term infants. Clin Vis Sci 1989; 4: 239-46.

12 Mushin J, Hogg CR, Dubowitz LMS, Skouteli H, ArdenGB. Visual evoked responses to light emitting diode(LED) photostimulation in newborn infants.Electroencephalogr Clin Neurophysiol 1984; 58: 317-20.

13 Taylor MJ. Visual evoked potentials. In: Eye JA, ed. Theneurophysiological examination of the newborn infant.London: MacKeith, 1992: 93-111.

14 Pryds 0, Greisen G, Trojaborg W. Visual evoked potentialsin preterm infants during the first hours of life.Electroencephalogr Clin Neurophysiol 1988; 71: 257-65.

15 Pryds 0. Stimulus rate-induced VEP attenuation in preterminfants. Electroencephalogr Clin Neurophysiol 1992; 84:188-91.

16 Placzek M, Mushin J, Dubowitz LMS. Maturation of thevisual evoked response and its correlation with visualacuity in preterm infants. Dev Med Child Neurol 1985; 27:448-54.

17 Taylor MJ, Menzies R, MacMillan U, Whyte HE. VEPs innormal full-term and premature neonates: longitudinalversus cross-sectional data. Electroencephalogr ClinNeurophysiol 1987; 68: 20-7.

18 Pryds 0, Trojaborg W, Carlsen J, Jensen J. Determinants ofvisual evoked potentials in preterm infants. Early HumDev 1989; 19: 117-25.

19 Kurtzberg D, Vaughan HG. Maturation of the CNS andevoked potentials. Amsterdam: Elsevier, 1986.

20 Whyte HE, Pearce JM, Taylor MJ. Changes in the VEP inpreterm neonates with arousal states, as assessed by EEGmonitoring. Electroencephalogr Clin Neurophysiol 1987; 68:223-5.

21 Harding GFA, Grose J, Wilton A, Bissenden JG. Thepattern reversal VEP in short-gestation infants.Electroencephalogr Clin Neurophysiol 1989; 74: 76-80.

22 Klimach VJ, Cooke RWI. Maturation of the neonatalsomatosensory evoked response in preterm infants. DevMed Child Neurol 1988a; 30: 208-14.

23 Kamiski W. The late somatosensory evoked potential inpremature and term infants. I. Principal componenttopography. Electroencephalogr Clin Neurophysiol 1992; 84:32-43.

24 Taylor MJ, Boor R, Ekert PG, Whyte HE. Normal prema-ture maturation of the somatosensory evoked potentials(in press).

25 Gallai V, Mazzotta G, Cagini L, Del Gatto F, Agnelott F.Maturation ofSEPs in preterm and full term neonates. In:Maturation of the CNS and evoked potentials. Amsterdam:Elsevier, 1986: 95-106.

26 Karniski W, Wyble L, Lease L, Blair RC. The latesomatosensory evoked potential in premature and terminfants. II. Topography and latency development.Electroencephalogr Clin Neurophysiol 1992; 84: 44-54.

27 Bongers-Schokking JJ, Colon EJ, Hoogland RA, Van denBrande JLV, de Groot CJ. Somatosensory evoked poten-tials in term and preterm infants in relation to postcon-ceptional age and birth weight. Neuropediatrics 1990; 21:32-6.

28 Majnemer A, Rosenblatt F, Willis D, Lavallee J. The effectof gestational age at birth on somatosensory-evokedpotentials performed at term. Y Child Neurol 1990; 5:329-35.

29 Pierrat V, de Vries LS, Minami T, Casaer P. Somatosensoryevoked potentials and adaptation to extrauterine life: alongitudinal study. Brain Dev 1990; 12: 376-9.

30 Gibson NA, Graham M, Levene MI. Somatosensoryevoked potentials and outcome in perinatal asphyxia. ArchDis Child 1992; 67: F393-8.

31 Gilmore R, Brock J, Hermansen MC, Baumann R.Development of lumbar spinal cord and cortical evokedpotentials after tibial nerve stimulation in the pretermnewborns: effects of gestational age and other factors.Electroencephalogr Clin Neurophysiol 1987; 68: 28-39.

32 Tranier 5, Chevallier B, Lemaigre D, Liot F, Lagardere B,Gallet JP. Potentiel evoque somesthesique du membreinferieur chez le nouveau-ne premature. Neurophys Clin1990; 20: 463-79.

33 Laureau E, Marlot D. Somatosensory evoked potentials aftermedian and tibial nerve stimulation in healthy newborns.Electroencephalogr Clin Neurophysiol 1990; 76: 453-8.

34 White CP, Cooke RWI. Maturation of the cortical evokedresponse to posterior-nerve stimulation in the pretermneonate. Dev Med Child Neurol 1989; 31: 657-64.

35 Davis JM, Elfenbein J, Schum R, Bentler RA. Effects ofmild and moderate hearing impairments on languageeducational and psychosocial behavior of children.J Speech Hear Disorders 1986; 55: 53-62.

36 McClelland R, Watson DR, Lawless V, Houston HG,Adams D. Relability and effectiveness of screening forhearing loss in high risk neonates. BMJ 1992; 304: 806-9.

37 Stockard JE, Stockard ni, Kleinberg F, Westrnoreland BF.Prognostic value of brainstem auditory evoked potentialsin neonates. Arch Neurol 1983; 40: 360-5.

38 Durieux-Smith A, Picton TW. Predictive value of BERA inthe assessment of hearing loss in high-risk infants. AudiolPractice 1988; 2: 4-6.

39 Adelman C, Levi H, Under N, Sohmer H. Neonatal audi-tory brain-stem response threshold and latency: 1 hour to5 months. Electroencephalogr Clin Neurophysiol 1990; 77:77-80.

40 Hyde ML, Riko K, Malizia K. Audiometric accuracy of theclick ABR in infants at risk for hearing loss. J Am AcadAudiol 1990; 1: 59-66.

41 Weber BA, Roush PA. Application of maximum lengthsequence analysis to auditory brainstem response testingin premature newborns. J Am Acad Audiol 1993; 4:157-62.

42 Bonfils P, Francois M, Avan P, Londero A, Trotoux J,Narcy P. Spontaneous and evoked otoacoustic emissionsin preterm neonates. Laryngoscope 1992; 102: 182-6.

43 Smurzynski J. Longitudinal measures of distortion-productand click-evoked otoacoustic emissions of preterminfants: preliminary results. Ear Hear 1994; 15: 210-23.

44 Fawer C, Calame A. Neurodevelopmental outcome. In:Haddad J, Saliba E, eds. Perinatal asphyxia. Paris:Springer-Verlag, 1993: 255-64.

45 Saliba E, Laugier J. Perinatal asphyxia. Berlin: SpringerVerlag, 1993.

46 de Vries LS, Pierrat V, Eken P, Taketsugu M, Daniels H,Casaer P. Prognostic value of early somatosensory evokedpotentials for adverse outcome in full-term infants withbrain asphyxia. Brain Dev 1991; 13: 320-5.

47 Muttitt SC, Taylor MJ, Kobyashi J, Whyte HE. Serialevoked visual potentials and outcome in full term birthasphyxia. PediatrNeurol 1991; 7: 86-90.

48 Taylor MJ, Murphy WJ, Whyte HE. Prognostic reliability ofSEPs and VEPs in asphyxiated tern infants. Dev MedChild Neurol 1992; 34: 507-15.

49 Pike MG, Holmstrom G, de Vries LS, Pennock JM, DrewKJ, Sonksen P. Patterns of visual impairment associatedwith lesions of the preterm infant brain. Dev Med ChildNeurol 1994; 36: 849-62.

50 de Vries LS, Connell J, Dubowitz LMS, Oozeer RC,Dubowitz V, Pennock JM. Neurological, electrophysio-logical and MRI abnormalities in infants with extensivecystic leukomalacia. Neuropediatrics 1987; 18: 61-6.

51 Klimach VJ, Cooke RWI. Short-latency cortical somato-sensory evoked responses of preterm infants with ultra-sound abnormality of the brain. Dev Med Child Neurol1988b; 30: 215-21.

52 de Vries LS, Eken P, Pierrat V, Daniels H, Casaer P.Prediction of neurodevelopmental outcome in thepreterm infant: short latency cortical somatosensoryevoked potentials compared with cranial ultrasound.Arch Dis Child 1992; 67: Fl 171-81.

53 Pierrat V. Eken P, Duquennoy C, Rousseau S, de Vries LS.Prognostic value of early somatosensory evoked potentialsin neonates with cystic leucomalacia. Dev Med ChildNeurol 1993; 35: 683-90.

54 White CP, Cooke RWI. Somatosensory evoked potentialsfollowing posterior tibial nerve stimulation predict latermotor outcome. Dev Med Child Neurol 1994; 36: 34-40.

55 McCulloch K, Robertson Y, Hamilton LR. Comparison ofauditory brainstem responses and pneumograms in inten-sive care nursery infants. Am J Perinatol 1988; 5: 172-6.

56 Cox C, Hack M, Aram D, Borawski E. Neonatal auditorybrainstem response failure of very low birth weightinfants: 8-year outcome. Ped Res 199 1; 31: 68-72.

57 Henderson-Smart DJ, Pettigrew AG, Cambell DJ. Clinicalapnea and brain-stem neural function in preterm infants.N EnglJMed 1983; 308: 353-7.

58 Lina-Granade G, Collet L, Morgon A, Salle B. Maturationand effect of stimulus rate on brainstem auditory evokedpotentials. Brain Dev 1993; 15: 263-9.

59 Chen YJ, Liou CS, Tsai CH, Yeh TF. Effect of amino-phylline on brain stem auditory evoked potentials inpreterm infants. Arch Dis Child 1994; 71: F20-3.

60 Johnston WH, Angara V, Baumal R, Hawke WA,Johnson RH, Keet S, et al. Erythroblastosis fetalis andhyperbilirubinemia: A five-year follow-up with neurologi-cal, psychological, and audiological evaluation. Pediatrics1967; 39: 88-92.

61 de Vries LS, Lary S, Dubowitz LMS. Relationship of serumbilirubin levels to ototoxicity and deafness in high risklow-birth-weight infants. Pediatrics 1985; 76: 351-4.

62 Perlman M, Fainmesser P, Sohmer H, Tamari H, Wax Y,Pevsmer B. Auditory nerve-brainstem evoked responsesin hyperbilirubinemic neonates. Pediatrics 1983; 72:658-64.

63 Streletz U, Graziani U, Branca PA, Desai H, Travis SF,Mikaelian DO. Brainstem auditory evoked potentials infuilterm and preterm newborns with hyperbilirubinemiaand hypoxemia. Neuropediatrics 1986; 17: 66-71.

64 Nakamura H, Takada 5, Shimabuku R, Matsuo M, MatsuoT, Negishi H. Auditory nerve and brainstem responses in

on July 18, 2021 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.74.1.F

70 on 1 January 1996. Dow

nloaded from

F76 Taylor, Saliba, Laugier

newborn infants with hyperbilirubinemia. Pediatrics 1985;75: 703-8.

65 Soares I, Collet L, Delorme C, Salle B, Morgon A. Areclick-evoked BAEPs useful in case of neonate hyperbiliru-binemia? Int J Pediatr Otorhinolakyngol 1989; 17: 231-7.

66 Esbjorner E, Larsson P, Leissner P, Wranne L. The serumreserve albumin concentration for monoacetyldiamino-diphenyl sulphone and auditory evoked responses duringneonatal hyperbilirubinaemia. Acta Paediatr Scand 1991;80: 406-12.

67 Watanabe K, Iwase K, Hara K. Maturation ofvisual evokedresponses in low-birthweight infants. Dev Med ChildNeurol 1972; 14: 425-35.

68 Hrbek A, Iversen N, Olsson T. Clinical applications ofevokedpotentials in neurology. New York: Raven Press, 1982.

69 Pettigrew AG, Edwards DA, Henderson-Smart DJ. Theinfluence of intra-uterine growth retardation on brainstemdevelopment of preterm infants. Dev Med Child Neurol1985; 27: 467-72.

70 Soares I, Collet L, Morgon A, Salle B. Effect of brainstemauditory evoked potential stimulus intensity variations inneonates of small for gestational age. Brain Dev 1988; 10:174-7.

71 Sarda P, Dupuy RP, Boulot Prieu D. Brainstem conductiontime abnormalities in small for gestational age infants.J Perinatal Med 1992; 20: 57-63.

72 de Zegher F, de Vries LS, Pierrat V, Daniels H, Spitz B,Casaer P. Effect of prenatal betamethasone/thyrotropinreleasing hormone treatment on somatosensory evokedpotentials in preterm newborns. Ped Res 1992; 32:212-14.

73 Friss HE, Wavrek D, Martin WH, Wolfson MR. Brain-stemauditory evoked responses to hypercarbia in preterminfants. Elecrroencephalogr Clin Neurophysiol 1994; 90:331-6.

74 Pryds 0, Schneider S. Aminophylline reduces cerebralblood flow in stable, preterm infants without affecting thevisual evoked potential. Eur _J Pediatr 1991; 150: 366-9.

on July 18, 2021 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.74.1.F

70 on 1 January 1996. Dow

nloaded from