The Big Bang Theory (season 12) | Crime Series | Diane a les paules FRENCH WEBRIP 2018

Multimodality evoked potentials as a prognostic tool in term asphyxiated newborns

Multimodality evoked potentials as a prognostic tool in term asphyxiated newborns

Electroencephalography and clinical Neurophysiology 108 (1998) 199–207 Multimodality evoked potentials as a prognostic tool in term asphyxiated newbo...

212KB Sizes 0 Downloads 9 Views

Electroencephalography and clinical Neurophysiology 108 (1998) 199–207

Multimodality evoked potentials as a prognostic tool in term asphyxiated newborns E. Scalais a ,*, A. Franc¸ois-Adant b, C. Nuttin a, A. Bachy c, J.M. Gue´rit d a

Department of Pediatrics, Pediatric Neurology, Entite´ Hospitalie`re, Centre Hospitalier Espe´rance St-Joseph, Clinique Saint-Vincent, 207 Rue Fr. Lefevbre, 4000 Rocourt, Lie`ge, Belgium b Centre Ne´onatal, Clinique St-Vincent, Department of Pediatrics, Entite´ Hospitalie`re, Centre Hospitalier Espe´rance St-Joseph, Rocourt, Lie`ge, Belgium c Centre Ne´onatal, Clinique Notre-Dame, Charleroi, Belgium d Unite´ d’Explorations Electrophysiologiques du Syste`me Nerveux, Cliniques Universitaires St-Luc, University of Louvain Medical School, Brussels, Belgium Accepted for publication: 1 September 1997

Abstract Hypoxic-ischemic (HI) events may cause permanent brain damage, and it is difficult to predict the long-term neurological outcome of survivors. Multimodality evoked potentials (MEPs), using flash visual (fVEPs), somatosensory (SEPs), and brain-stem auditory evoked potentials (BAEPs) may assess the cerebral function in term neonates. MEPs were recorded in 40 hypoxic-ischemic term or near-term neonates during the first week of life in order to predict the neurological outcome. A 3 point grading system registered either mild, moderate, or severe abnormalities. At 24 months of corrected age, the infants were assessed with a blind protocol to determine neurological development. Grade 0 fVEPs and SEPs were associated with a normal neurological status with 100% (P , 0.001) of the infants. Abnormal SEPs or total grade (VEPs + SEPs) . I were not associated with normal outcomes (P , 0.0001). Normal BAEPs did not predict a normal outcome, but severely abnormal BAEPs did predict an abnormal outcome. A significant correlation was found between EP (VEPs + SEPs) grade (r = 0.9, P , 0.0001), Sarnat stage (r = 0.6, P , 0.001), and clinical outcome. This study confirmed that both fVEPs and SEPs are more accurate as prognostic indicators for term neonates. EPs (VEPs + SEPs) also are more accurate in predicting the ultimate neurological outcome compared with the Sarnat scoring.  1998 Elsevier Science Ireland Ltd. Keywords: Multimodality evoked potentials; Asphyxia; Newborns

1. Introduction Several methods have been used for the early prediction of neurological outcome after perinatal asphyxia. These methods include estimation of hypoxic-ischemic (HI) encephalopathy based on clinical assessment (Sarnat and Sarnat, 1976; Levene et al., 1985) and imaging techniques (Adsett et al., 1985; Lipp-Zwahlen et al., 1985; Lipper et al., 1986; Fitzhardinge et al., 1981; Barkovich, 1992). Neuroimaging techniques such as computed tomography (CT), ultrasonography (US), and magnetic resonance imaging (MRI) provide information about the morphology of the nervous system without assessing its function. This can be done by * Corresponding author. Tel.: +32 424 66265.

0168-5597/98/$19.00  1998 Elsevier Science Ireland Ltd. All rights reserved PII S0168-5597 (97 )0 0076-2

electrophysiological techniques. Neonatal EEG has prognostic power in definite abnormal and normal tracings (Rose and Lombroso, 1970; Monod et al., 1972; Sarnat and Sarnat, 1976; Watanabe et al., 1980; Holmes et al., 1982; Pezzani et al., 1986; van Lieshout et al., 1995), although early prognosis may be difficult (Watanabe et al., 1980) in intermediate grade abnormalities. The prognostic value of evoked potentials (EPs) in asphyxia have also been studied. Flash visual evoked potentials (fVEPs; Gambi et al., 1980; Whyte et al., 1986; Muttitt et al., 1991) and somatosensory evoked potentials (SEPs) (Go¨rke, 1986; Majnemer et al., 1987; Willis et al., 1987; De Vries et al., 1991; Gibson et al., 1992b; De Vries, 1993) have been shown to predict the neurological outcome. The predictive power is less accurate with brain-stem auditory evoked

EEP 96663

200

E. Scalais et al. / Electroencephalography and clinical Neurophysiology 108 (1998) 199–207

potentials (BAEPs) (Hecox and Cone, 1981; Stockard et al., 1983; Guinard et al., 1988; Karmel et al., 1988; Majnemer et al., 1988) than with fVEPs and SEPs. A few studies have combined SEPs with fVEPs (Hrbek et al., 1977; Taylor et al., 1992) or BAEPs (Majnemer et al., 1990; Majnemer and Rosenblatt, 1995) and VEPs with BAEPs (Hakamada et al., 1981). The present study analyses the predictive value of multimodality evoked potentials (MEPs) and Sarnat scoring with respect to the neurological and developmental outcome, in asphyxiated neonates. It also emphasizes the interest of analyzing the degree of EP abnormality by using a scoring system including both VEPs and SEPs.

2. Methods 2.1. Subjects Forty infants were studied. All were over 36 weeks gestation with appropriate weight for gestational age and were admitted to a neonatal intensive care unit (between 1988 and 1992). The mean gestational age was 39.3 ± 1.3 weeks and the mean birth weight was 3262 ± 106 g. Asphyxia was considered if the 5 min Apgar score was ,3 or by clinical histories suggestive of asphyxia including two of the following observations: (i) abnormal neurological examination consistent with asphyxia (prolonged altered tone or consciousness, early seizures), (ii) intrapartum history suggestive of asphyxia (late decelerations in fetal monitoring, meconium staining of the liquor), and (iii) CT compatible with hypoxic-ischemic changes. HI encephalopathy was graded according to the criteria of Sarnat (Sarnat and Sarnat, 1976) and each infant was classified according to the most severe stage recorded over the first 3 days of life. Infants with congenital neurological malformations or dysmorphic features were not included in the study. Informed consent was obtained from the parents. Infants who presented seizures were treated with phenobarbital (PB). Phenytoin and/or benzodiazepine were added if the seizures were difficult to control. EEG studies and cranial ultrasound examinations could not be serially conducted on all infants. In severe asphyxia, CT was performed after the first week of life. A control group (VEPs: n = 14; SEPs and BAEPs: n = 13) was formed of normal 37–41 weeks gestation-age infants. In the studied population, the infants were followedup. The assessment was performed at 24 months corrected age by a pediatrician who was not informed of the electrophysiological results. The infants had a neurological examination with a quantitative assessment of motor disabilities (none; mild = abnormal posture; moderate = locomotion impaired; severe = locomotion impossible) and a psychodevelopmental index (PDI) (Brunet-Lezine or Griffith; nor-

mal = PDI . 85; mildly abnormal = PDI 70–84; moderately abnormal = PDI 51–69; severely abnormal = PDI , 50). 2.2. MEP testing The MEPs were recorded mostly during the first 3 days of life and if abnormal, repeated weekly until discharge. For 8 infants, the MEPs were performed between day 4 and day 7. In one infant, BAEPs were studied on day 7. If the infants had more than one examination during the first days of life, the most severe test recorded was retained. The subjects were usually asleep except for the fVEPs. A Nicolet Compact-Four and Spirit were used. The VEPs were obtained by binocular stimulation with light-emitting diode (LED) goggles. The VEPs were recorded on two channels from Oz referenced to A1 + A2 (linked ears) (Channel I) and to Cz (Channel II). If fVEPs were severely abnormal, electroretinograms were recorded from an outer canthus lead that was also referenced to Cz. Electrode impedance was less than 5 kQ. The band pass was 1–100 Hz and the sweep time was 1000 ms. A gain of 10 K was used. At least two averages of 64 trials were obtained. SEPs were evoked by median nerve stimulation at the wrist at minimal twitch level (electrical square wave pulses: 200 ms; rate: 1.1/s; trials: 400–500). For spinal recording, a rate of 4.1/s was employed. Electrodes were placed at the following levels: Channel I = (CII): at the second cervical vertebra (reference: linked ears; band pass: 30–3000 Hz; sweep time: 30 ms; gain: 40 K). Channel II = (F3’ or F4’) and III = (C3’ or C4’): at the contralateral prerolandic (F3’ or F4’: 2 cm behind F3 or F4 as defined by the 10–20 international system) and the postrolandic area (C3’ or C4’: 2 cm behind C3 or C4) and referenced to A1 + A2 (linked ears). The band pass was 1–250 Hz, the sweep time 200 ms and the gain was 10 or 20 K. The ground was placed on the upper limbs. BAEPs were obtained with monaural clicks (90 dB HL, alternating polarity, rate: 21.7/ s, trials: 2100–3000). The band pass was 150–3000 Hz, the sweep time was 15 ms and the gain was 10 K. The contralateral ear was used as ground. 2.3. Naming of components and normal values (A) The major components of fVEPs are P200, N300 and, P400 (Fig. 1A) (N300 may be bifid, and P400 not always evident). Similar responses using cephalic (Oz−-Cz+) or linked ears (Oz−-linked ears+) references were observed. Identification of components was made with the linked ears reference. (B) SEPs recorded from the spinal cord comprise early components corresponding to spinal potentials. It was the only single prominent negative peak which was measured for latency (N13). If the peak was bifid, the second peak was taken. SEPs recorded from the sensory cortex comprise early components generated in the brain-stem (only the P14 was measured, but identification is not always evident in neonates), and short (N20 or N1) and late-latency

E. Scalais et al. / Electroencephalography and clinical Neurophysiology 108 (1998) 199–207

201

I: increased latency or unusual morphology, poorly reproducible. II: missing components (P200, N300), low amplitude (,2 mV) and poorly reproducible wave, abnormal features may be more pronounced on the linked ears reference. III: absence of identifiable components. SEPs were classified into 4 grades (Fig. 2A–D). In the identification of spinal activity short (N20 or N1) and late parietal cortical (N2) components were taken into consideration. With unilateral or bilateral abnormalities, the best side was used and the asymmetry was noted. Late components may be unstable. A short (Pa) prerolandic cortical component was noted. 0: normal spinal and cortical activities. I: normal spinal activity; increased latencies (N20 or N1), poorly reproducible or asymmetric. II: normal spinal activity; poorly reproducible wave; identification of N20 or N1 is problematic. III: normal spinal activity; bilateral absence of short (N20 or N1) and late cortical (N2) waves. Fig. 1. The traces show fVEPs of Grade 0 (A), I (B), II (C), and III (D). Negativity at the active electrode is represented as an upward deflection. Identification of components was made with the linked ears reference. Grade 0 (A): Normal VEPs with P200, N300, and P400 components. (N300 may be bifid, and P400 not always evident.) Grade I (B): Increased latency, unusual morphology, or poorly reproducible wave. Grade II (C): Missing components (P200, N300), low amplitude (,2 mV) and poorly reproducible wave. Abnormal features may be more pronounced on the linked ears reference. Grade III (D): Absence of identifiable components. (1) The persisting occipital negativities (Wave I) coincide in time with the ERG’s b waves (2) obtained with left monocular stimulation and recording at the outer canthus (OC) of the left eye.

(N2) (identification is not always present) cortical components (Fig. 2A). For the cortical SEP recordings, a linked ears reference was used in order to differentiate the pre- (Pa) and postrolandic (N20 or N1) SEP components adequately. C: The BAEPs consist of several components (Waves I, II, III, IV and V). 2.4. Data scoring A 3 point grading system registered mild, moderate and severe abnormalities. Normal response (Grade 0) represents normal activity of the generators. In the presence of peripheral activity, an absence of response (Grade III) represents a loss of activity of the generators. Increased latency implies integrity of the generators but slower functioning or delayed transmission. Low amplitude and missing components could imply a decreased activity or partial loss of activity of the neurons that generate EPs. The waves were reviewed by one of the authors (G.J.M.) who was not aware of the patients’ identities. fVEPs were classified into 4 grades (Fig. 1A–D). 0: normal.

BAEPs were classified into 4 grades. 0: normal. I: increased interpeak latencies (IPL) or abnormal amplitude ratio V/I. II: Waves I and III identifiable. III: Wave I only. 2.5. Statistical analysis Fisher’s exact probability test was used to determine if a normal or an abnormal result of EPs was associated with a normal or abnormal outcome. The binomial test was used to evaluate the probability of a normal or abnormal outcome in relation to the EP grade. The relationship between the Sarnat score, the EP grade (VEPs + SEPs), and the clinical outcome was studied by means of the Spearman rank correlation coefficient test. The sensitivity, specificity, and accuracy were calculated.

3. Results 3.1. Clinical data The study population consisted of 40 infants. According to the classification of Sarnat, 5 infants were at Stage I, 26 at Stage II, and 9 at Stage III. Eleven of the 40 infants died. They were analyzed separately and were not included in the statistical analysis. These infants had severe hypoxic-encephalopathy. Three were at Stage II and 8 at Stage III. Nine infants died during the first days of life because of severe asphyxia. One infant died at 2 weeks of age because of

202

E. Scalais et al. / Electroencephalography and clinical Neurophysiology 108 (1998) 199–207

respiratory insufficiency (CT showed bilateral encephalomalacia). One with bilateral pyramidal signs and poor visual function died at 6 months. The overall mortality was 27%. Of the 29 surviving infants, 10 were normal and 19 had an abnormal outcome at follow-up. Twelve were severely handicapped, 4 moderately handicapped, and 3 mildly handicapped. 3.2. MEP data Table 1 gives the results of Fisher’s exact test. Table 2 illustrates the correlation between the fVEP and SEP grades and the number of infants with a normal outcome. Eight infants had normal fVEPs during the first week of life and all had a normal outcome except one infant with a moderate handicap (P , 0.01). The one exception had an abnormal SEP (Grade I). Grade fVEPs I were associated with a normal outcome if the SEPs were Grade 0 (P , 0.05). All the neonates with a Grade fVEP II were severely handicapped except for one infant with mild abnormalities. Eleven infants had normal SEPs during the first week of life, 10 of whom had a normal outcome (P , 0.001). The one

exception had abnormal fVEPs (Grade II). All the 18 infants with abnormal SEPs during the first week of life had neurological sequelae (P , 0.001). The combination of both fVEP and SEP gives better predictive power. Grade 0 VEPs and SEPs were associated with normal outcomes in all of the infants (P , 0.001). Abnormal SEPs or total grade (fVEP + SEP) . I was associated with abnormal outcomes (P , 0.0001). Moderate to severe sequelae were observed with Grade SEP III (P , 0.05). In contrast, neurological sequelae (mild n = 2; moderate n = 3; severe n = 9) were found in 14 infants with normal BAEPs. And Grade I BAEPs were only associated with abnormal outcome in 4 infants (mild n = 1; moderate n = 1; severe n = 2). Thus BAEPs added no further prognostic value to this study. In a few infants, EPs were studied on the first day of life and repeated on day 2 or 3. One infant had Grade VEP 0 on day one, and Grade VEP II on day 3. After the first week of life, persistent abnormal VEPs and SEPs were always associated with sequelae (P , 0.05). Severe neurological handicaps were found in infants with persistent grade VEP or SEP ≥ II. The comparison of prognostic reliability between the Sar-

Fig. 2. The traces show SEPs of Grade 0 (A), I (B), II (C), and III (D). Negativity at the active electrode is represented as an upward deflection. Linked ears were used as the reference site for both the contralateral pre- and postrolandic electrodes. Short (N20 or N1) and late (N2) components are taken into consideration. Late components may be unstable. A short (Pa) prerolandic cortical component was noted. With unilateral or bilateral abnormalities, the best side was taken into consideration and the asymmetry was noted. Grade 0 (A): Left and right upper traces: normal cortical postrolandic (N20 or N1) and prerolandic (Pa) SEPs to right and left median nerve stimulation. Identification of P14 and N2 is not always evident. Left and right lower traces: spinal SEPs to right and left median nerve stimulation. Grade I (B): Left and right upper traces: cortical (postrolandic and prerolandic) SEPs to right and left median nerve stimulation. After left median nerve stimulation, N20 or N1 latency is delayed. Left and right lower traces: spinal SEPs to right and left median nerve stimulation. Grade II (C): Left and right upper traces: SEPs to right and left median nerve stimulation. Notice the poorly reproducible wave after left median nerve stimulation; identification of N20 or N1 is problematic. Notice the absence of short (N20 or N1) and late (N2) cortical waves after left median nerve stimulation. Left and right lower traces: spinal SEPs to right and left median nerve stimulation. Grade III (D): Left and right upper traces: SEPs to right and left median nerve stimulation. Notice the absence of short (N20 or N1) and late (N2) cortical waves after right and left median nerve stimulation. Left and right lower traces: spinal SEPs to right and left median nerve stimulation.

203

E. Scalais et al. / Electroencephalography and clinical Neurophysiology 108 (1998) 199–207 Table 1

Table 3

2 × 2 tables of VEPs, SEPs, BAEPs and outcomes after more than 2 years corrected age

Comparison of predictive value between the Sarnat stage and VEP and SEP grades

Outcome

Normal VEPs

Abnormal VEPs

Total

Normal Abnormal Total

7 1 8

3 18 21

10 19 29

Fischer’s exact test (P , 0.001). Outcome

Normal SEPs

Abnormal SEPs

Total

Normal Abnormal Total

10 1 11

0 18 18

10 19 29

Fischer’s exact test (P , 0.00001). Outcome

Normal BAEPs

Abnormal BAEPs

Total

Normal Abnormal Total

8 14 22

2 4 6

10 18 28

Fischer’s exact test (NS).

nat classification and the EP grade is shown in Table 3. Stage I encephalopathy was associated with normal outcomes (n = 5), Stage III encephalopathy with severe abnormal outcomes (n = 1). In Stage II (n = 18), infants with fVEP grade ≤ I and SEP grade = 0 had normal outcomes. But infants with fVEP grade . I and SEP grade ≥ I had poor outcomes. A significant correlation was found between the EP (VEPs + SEPs) grade (r = 0.9, P , 0.0001), the Sarnat stage (r = 0.6, P , 0.001), and the clinical outcome. The sensitivity, specificity and accuracy were 94, 70, and 86% for VEPs, and 94, 100, and 96% for SEPs, and 22, 80, and 42% for BAEPs. Table 4 summarizes the MEP results in the group of infants who died. Almost all infants who died in the neonatal period had Grade II or III VEP and Grade III SEP. BAEPs also were frequently abnormal.

Normal outcome (VEP grade ≤ I and SEP grade = 0) Poor outcome (VEP grade . I or SEP grade ≥ I)

SARNAT I (n = 5) 5

SARNAT II (n = 24) 5

SARNAT III (n = 1) 0

0

18(a,a,a,b,b,b,b,c, c,c,c,c,c,c,c,c,c,c )

1 (c)

Infants who died not included. Mild handicap. Moderate handicap. c Severe handicap. Spearman rank correlation coefficient test EPs: r = 0.9, P , 0.0001; Sarnat: r = 0.6, P , 0.001. a

b

4. Discussion The evaluation of birth asphyxia is problematic (Nelson and Leviton, 1991). Asphyxia is a clinical diagnosis and the correlation with neurological outcome remains difficult. Several techniques have been used to predict the neurological status. Evoked potentials represent electrical activity induced by sensory stimulations of the nervous system. The type of stimuli used determines the nervous structures which are studied: fVEPs reflect the hemispheric structures; SEPs reflect different levels of the neuraxis including the peripheral nerve, spinal cord, central brain-stem, thalamus, and cortex; and BAEPs reflect the cochlea and the brain-stem auditory pathways. The use of one modality gives only a focal cerebral assessment because it only looks at the visual, sensory or auditory pathway, while MEPs give a more global assessment. 4.1. Normal responses and maturation All modalities of EPs, i.e, visual, somatosensory and auditory potentials, have been extensively analyzed in term neonates. The present study replicates the data of pre-

Table 2 Correlation between the EP grade (EPs in the first week of life) and the number of infants with normal outcomes

Table 4

SEPs

Grade

O I II III Total

VEPs 0

I

II

7/7*** 0/1b

3/3* 0/2a,b 0/5a,b,c,c,c 0/2b,c 3/12

0/1a 0/1c

7/8**

0/5c,c,c,c,c 0/7#

III

Total

Group of infants who died (n = 11) (Sarnat stage and MEPs)

VEPs

10/11*** 0/4 0/2c,c 0/2

0/5 0/9# 10/29

SEPs BAEPs

a

Mild handicap. Moderate handicap. Severe handicap. Favorable outcome (binomial test): *P , 0.05; **P , 0.01; ***P , 0.001. Unfavorable outcome: #P , 0.05. b c

a

Sarnat II (n = 3) 0 (n = 1) I (n = 1) II (n = 1)

Sarnat III (n = 8)

a

III (n = 3) 0 (n = 2) I (n = 1)

II III III 0 I II III

(n (n (n (n (n (n (n

= = = = = = =

3) 5) 8) 1) 1) 3) 3)

Died at 6 months with bilateral pyramidal signs and poor visual function (SEPs = III).

204

E. Scalais et al. / Electroencephalography and clinical Neurophysiology 108 (1998) 199–207

Table 5

A. Mean latencies and SD of main response components of fVEPs in normal neonates GA (weeks) fVEPs

X SD (n = 14) U. limits

40.2 0.4

N100 (ms)

P200 (ms)

N300 (ms)

Amplitude (mV)

P400 (ms)

Amplitude (mV)

168.3 26.2 233.8

219.0 20.8 271.2

295.2 20.5 346.5

20.9 18.5

359.5 24.2 420.2

15.8 9.6

B. Mean latencies and SD of main response components of spinal and cortical SEPs obtained after median nerve stimulation in normal neonatesa GA (weeks) SEPs n X SD U. limits

39.5 1.1

N13 (ms) 13 10.18 0.85 12.32

Amplitude (mV) 2.20 2.57

P14 (ms) 9 17.94 2.63 24.47

N20 (ms) 13 35.75 7.61 54.77

Amplitude (mV) 2.19 1.42

C. Mean latencies and SD of main response components of BAEPs in normal neonates GA (weeks) BAEPs

X SD (n = 13) U. limits

40.2 0.4

I

III

V

V-I

V/I

1.73 0.18 2.18

4.65 0.31

6.87 0.3

5.14 0.27 5.82

0.89 0.35

GA, gestational age; U. limits, X + (2.5 × SD); X, mean. N20 or N1 could be identified in every infant, whereas P14 were not always evident.

a

vious studies of fVEPs and BAEPs in normal neonates (Hecox and Galambos, 1974; Stockard and Stockard, 1983; Taylor et al., 1987). The values of fVEP and BAEP latencies (Table 5) agree with those in the literature (Stockard and Stockard, 1983; Taylor et al., 1987). The present study also confirms that cortical SEPs were almost constantly obtained (George and Taylor, 1991; Gibson et al., 1992a). But previous studies describe cortical SEPs with a frontal reference where cephalic frontal reference produces cervical distortions and cancels far-field potentials (Desmedt and Cheron, 1981; Desmedt, 1987). With the linked ears reference, the cortical SEPs display a negative positive wave recorded in the parietal region (N20 or N1) and a positive negative wave recorded in the fronto-central region (Pa). The phase reversal of parietal N20 or N1 and frontal Pa may also be helpful to identify the early cortical potentials. The mean peak latency of the main parietal negative wave at the cortical level was increased because of different filter (Fig. 3) settings (Laureau et al., 1988). The late N2 parietal components are not always present (Gallai et al., 1986). All types of EPs, i.e., visual, somatosensory and auditory have been observed to mature with progressive decreases in absolute latencies with increasing definition of the various waveform components (Desmedt et al., 1976; Stockard and Stockard, 1983; Willis et al., 1984; Gallai et al., 1986; Klimach and Cooke, 1988; Taylor et al., 1987; Laureau et al., 1988; Bongers-Schokking et al., 1990; George and Taylor, 1991; Gibson et al., 1992a). Our data with normal neonates replicate the observation that a more mature wave pattern is seen with term infants.

4.2. Studied population This paper replicates and extends several points previously made about the role of multi-model EPs in the prognostication of term asphyxia. Both fVEPs and SEPs can be efficient predictors. BAEPs are less sensitive. In the earliest studies, VEPs, SEPs and BAEPs were generally used separately, and not tested during the first week of life, and the criteria of inclusion were different with term and preterm infants (Hrbek et al., 1977; Gambi et al., 1980; Hakamada et al., 1981; Willis et al., 1987; Majnemer et al., 1987; Majnemer et al., 1988). The more recent studies using VEPs (Whyte et al., 1986; Muttitt et al., 1991) or SEPs (De Vries et al., 1991; Gibson et al., 1992b) showed high predictive power, and the combination of both fVEPs and SEPs yielded higher predictive reliability (Taylor et al., 1992). Others (Hecox and Cone, 1981; Stockard et al., 1983; Guinard et al., 1988; Karmel et al., 1988; Majnemer et al., 1988) have also investigated the use of BAEPs as an indicator of early brain damage. As in our study, BAEPs were not as effective in predicting neurological outcomes in asphyxiated infants because of false negatives. This study however emphasizes the interest of analyzing the degree of EP abnormality by using a scoring system including both VEPs and SEPs rather than the BAEPs. In the study of Hrbek et al. (1977) the authors developed a scoring system of VEP and SEP abnormalities and reported the infants with the highest score (worst SEPs or VEPs) to have the worst asphyxia. But there is no data on follow-up. In the study of Majnemer et al. (1990), the degree of SEP

E. Scalais et al. / Electroencephalography and clinical Neurophysiology 108 (1998) 199–207

Fig. 3. The effects of filter settings. Left (1–250 Hz) traces: cortical (postrolandic and prerolandic) SEPs to left median nerve stimulation. The linked ears were used as the reference site for both the contralateral preand postrolandic electrodes. Right upper (30–3 K) and lower (5–1.5 K) traces: cortical (retrorolandic) SEPs to left median nerve stimulation. Frontal electrode was used as the reference site.

abnormality reflects the degree of neurological impairment. Abnormalities which persisted or worsened correlated with severe neurological impairment, whereas an abnormal somatosensory evoked response, which improved or normalized in infancy, was associated with mild to moderate neurological impairment. But the combination of both SEPs and BAEPs did not yield higher accuracy. A scoring system was used in the study of Gibson et al. (1992b) including both the severity and the duration of the SEP abnormalities. Infants with the highest score (with the worst SEPs: pattern D) had severe encephalopathy and died. But there was no clear difference between the infants with cerebral palsy and the dystonic infants in terms of length of persistence of SEP abnormality or of its pattern (Gibson et al., 1992b). In the study of Taylor et al. (1992), the relation between the degree of EP abnormality and the severity of the neurological handicap could not be fully established because all the infants with abnormal outcome had severe sequelae. This paper also points out a difference concerning the fact mentioned in some studies (Gibson et al., 1992b; Taylor et al., 1992) that abnormal SEPs were not necessarily a bad prognostic sign. In our study, abnormal SEPs were consistently associated with sequelae, and Grade SEP III was always associated with abnormal VEPs. This discrepancy could probably be explained by the difference of methodology. Linked ears reference was used instead of frontal reference. Moreover, poor waveforms can be caused by noisy recordings (Taylor et al., 1992). This is why in our study, great care was taken during the recording of SEPs by eliminating electrical interference generated by surrounding monitoring equipment. We also saw that in case of difficult identification of early cortical waves due to deep sleep, the recording was repeated. In terms of duration of EP abnormality, the present data confirmed that persistently abnormal EPs (grade VEPs or SEPs ≥ II) predicted severe disability during follow-up. MEPs were also more accurate in predicting the ultimate

205

clinical outcome compared with the Sarnat staging. In the group of infants who died, severe neurophysiological abnormalities were observed with total (VEPs + SEPs) grade ≥ III. The discrepancy of accuracy between fVEPs, SEPs, and BAEPs could be expected because subcortical and cortical injury in a parasagittal distribution is the principal ischemic lesion in term neonates, the postcentral gyri (somesthetic) and the calcarine (visual) being particularly vulnerable (Volpe, 1995). In terms of prediction, it is not clear if EPs are more powerful than EEGs. In the literature, the accuracy ranges from 69% to 89% (Monod et al., 1972; Watanabe et al., 1980; Holmes et al., 1982; Pezzani et al., 1986; van Lieshout et al., 1995). A normal EEG usually carries a good prognosis (Rose and Lombroso, 1970; Monod et al., 1972; Sarnat and Sarnat, 1976; Watanabe et al., 1980; Holmes et al., 1982; Pezzani et al., 1986; van Lieshout et al., 1995). Severely abnormal EEGs like isoelectric recordings (Sarnat and Sarnat, 1976; Holmes et al., 1982) or persistent low voltage states (Holmes et al., 1982) are associated with poor prognoses. CTs have an accuracy for predicting outcome reaching 77% (Fitzhardinge et al., 1981; Adsett et al., 1985; Lipper et al., 1986). The best predictive values are obtained when CTs are performed after the first week of life (Lipp-Zwahlen et al., 1985). The prognostic value of cranial US with regard to neurodevelopmental outcome is less accurate (Hill, 1991) except in one recent study using 5, 7.5, and 10 MHz transducers (De Vries, 1993). We confirmed that both VEPs and SEPs correlate well with neurological outcomes in term hypoxic-ischemic infants. The use of a scoring system including both fVEPs and SEPs rather than BAEPs (testing sub-cortical and cortical structures) predicts more accurately the presence and the extent of the neurological handicap. The efficiency of VEPs, SEPs, and BAEPs probably varies in function of the selective vulnerability of the brain’s regions to HI events. EPs (VEPs + SEPs) also are more accurate in predicting the ultimate neurological outcome compared with the Sarnat scoring. Therefore, considering the topographic and temporal limits of the information received by the EPs, MEPs are a useful tool in the evaluation of the hypoxic-ischemic term neonates.

References Adsett, D.B., Fitz, C.R. and Hill, A. Hypoxic-ischaemic cerebral injury in the term newborn: correlation of CT findings with neurological outcome. Dev. Med. Child Neurol., 1985, 27: 155–160. Barkovich, A.J. MR and CT evaluation of profound neonatal and infantile asphyxia. AJNR, 1992, 13: 959–972. Bongers-Schokking, J.J., Colon, E.J., Hoogland, R.A., Van den Brande, J.L. and de Groot, C.J. Somatosensory evoked potentials in term and preterm infants in relation to postconceptional age and birth weight. Neuropediatrics, 1990, 21: 32–36.

206

E. Scalais et al. / Electroencephalography and clinical Neurophysiology 108 (1998) 199–207

Desmedt, J.E. Physiologie et physiopathologie des sensations somatiques e´tudie´es chez l’homme par la me´thode des potentiels e´voque´s. J. Physiol. (Paris), 1987, 55: 1–74. Desmedt, J.E. and Cheron, G. Non-cephalic reference recording of early somatosensory potentials to finger stimulation in adult or aging normal man: differentiation of widespread N18 and contralateral N20 from the prerolandic P22 and N30 components. Electroenceph. clin. Neurophysiol., 1981, 52: 553–570. Desmedt, J.E., Brunko, E. and Debecker, J. Maturation of the somatosensory evoked potentials in normal infants and children, with special reference to the early N1 component. Electroenceph. clin. Neurophysiol., 1976, 40: 43–58. De Vries, L.S. Somatosensory-evoked potentials in term neonates with postasphyxial encephalopathy. Clin. Perinatol., 1993, 20: 463–482. De Vries, L.S., Pierrat, V., Eken, P., Minami, T., Daniels, H. and Casaer, P. Prognostic value of early somatosensory evoked potentials for adverse outcome in full-term infants with birth asphyxia. Brain Dev., 1991, 13: 320–325. Fitzhardinge, P.M., Flodmark, O., Fitz, C.R. and Ashby, S. The prognostic value of computed tomography as an adjunct to assessment of the term infant with postasphyxial encephalopathy. J. Pediatr., 1981, 99: 777– 781. Gallai, V., Mazzotta, G., Cagini, L., Del Gatto, F. and Agnelotti, F. Maturation of SEPs in preterm and full-term neonates. In: V. Gallai (Ed.), Maturation of the CNS and Evoked Potentials. Elsevier, Amsterdam, 1986, pp. 95–106. Gambi, D., Rossini, P.M., Albertini, G., Sollazzo, D., Torrioli, M.G. and Polidori, G.C. Follow-up of visual evoked potential in full-term and preterm control newborns and in subjects who suffered from perinatal respiratory distress. Electroenceph. clin. Neurophysiol., 1980, 48: 509–516. George, S.R. and Taylor, M.J. Somatosensory evoked potentials in neonates and infants: developmental and normative data. Electroenceph. clin. Neurophysiol., 1991, 80: 94–102. Gibson, N.A., Brezinova, V. and Levene, M.I. Somatosensory evoked potentials in the term newborn. Electroenceph. clin. Neurophysiol. 1992a, 84: 26–31. Gibson, N.A., Graham, M. and Levene, M.I. Somatosensory evoked potentials and outcome in perinatal asphyxia. Arch. Dis. Child. 1992b, 67: 393–398. Go¨rke, W. Somatosensory evoked cortical potentials indicating impaired motor development in infancy. Dev. Med. Child Neurol., 1986, 28: 633–641. Guinard, C., Fawer, C.L., Despland, P.A. and Calame, A. Auditory brainstem responses and ultrasound changes in a high-risk infant population. Helv. Paediat. Acta, 1988, 43: 377–388. Hakamada, S., Watanabe, K., Hara, K. and Miyazaki, S. The evolution of visual and auditory evoked potentials in infants with perinatal disorder. Brain Dev., 1981, 3: 339–344. Hecox, K. and Galambos, R. Brain-stem auditory evoked responses in human infants and adults. Arch. Otolaryngol., 1974, 99: 30–33. Hecox, K.E. and Cone, B. Prognostic importance of brain-stem auditory evoked responses after asphyxia. Neurology, 1981, 31: 1429– 1433. Hill, A. Current concepts of hypoxic-ischemic cerebral injury in the term newborn. Pediatr. Neurol., 1991, 7: 317–325. Holmes, G., Rowe, J., Hafford, J., Schmidt, R., Testa, M. and Zimmerman, A. Prognostic value of the electroencephalogram in neonatal asphyxia. Electroenceph. clin. Neurophysiol., 1982, 53: 60–72. Hrbek, A., Karlberg, P., Kjellmer, I., Olsson, T. and Riha, M. Clinical application of evoked electroencephalographic responses in newborn infants. I: Perinatal asphyxia. Dev. Med. Child Neurol., 1977, 19: 34– 44. Karmel, B.Z., Gardner, J.M., Zappulla, R.A., Magnano, C.L. and Brown, E.G. Brain-stem auditory evoked responses as indicators of early brain insult. Electroenceph. clin. Neurophysiol., 1988, 71: 429–442. Klimach, V.J. and Cooke, R.W.I. Maturation of the neonatal somatosen-

sory evoked response in preterm infants. Dev. Med. Child Neurol., 1988, 30: 208–214. Laureau, E., Majnemer, A., Rosenblatt, B. and Riley, P. A longitudinal study of short latency somatosensory evoked responses in healthy newborns and infants. Electroenceph. clin. Neurophysiol., 1988, 71: 100– 108. Levene, M.I., Kornberg, J. and Williams, T.H.C. The incidence and severity of post-asphyxial encephalopathy in full-term infants. Early Hum. Dev., 1985, 11: 21–26. Lipper, E.G., Voorhies, T.M., Ross, G., Vannucci, R.C. and Auld, P.A.M. Early predictors of one-year outcome for infants asphyxiated at birth. Dev. Med. Child Neurol., 1986, 28: 303–309. Lipp-Zwahlen, A.E., Deonna, T., Chrzanowski, R., Micheli, J.L. and Calame, A. Temporal evolution of hypoxic-ischaemic brain lesions in asphyxiated full-term newborns as assessed by computerized tomography. Neuroradiology, 1985, 27: 138–144. Majnemer, A. and Rosenblatt, B. Prediction of outcome at school entry in neonatal intensive care unit survivors, with use of clinical and electrophysiologic techniques. J. Pediatr., 1995, 127: 823–830. Majnemer, A., Rosenblatt, B., Riley, P., Laureau, E. and O’Gorman, A.M. Somatosensory evoked response abnormalities in high-risk newborns. Pediatr. Neurol., 1987, 3: 350–355. Majnemer, A., Rosenblatt, B. and Riley, P. Prognostic significance of the auditory brain-stem evoked response in high-risk neonates. Dev. Med. Child Neurol., 1988, 30: 43–52. Majnemer, A., Rosenblatt, B. and Riley, P.S. Prognostic significance of multimodality evoked response testing in high-risk newborns. Pediatr. Neurol., 1990, 6: 367–374. Monod, N., Pajot, N. and Guidasci, S. The neonatal EEG: statistical studies and prognostic value in full-term and pre-term babies. Electroenceph. clin. Neurophysiol., 1972, 32: 529–544. Muttitt, S.C., Taylor, M.J., Kobayashi, J.S., MacMillan, L. and Whyte, H.E. Serial visual evoked potentials and outcome in term birth asphyxia. Pediatr. Neurol., 1991, 7: 86–90. Nelson, K.B. and Leviton, A. How much of neonatal encephalopathy is due to birth asphyxia?. Am. J. Dis. Child., 1991, 145: 1325– 1331. Pezzani, C., Radvanyi-Bouvet, M.F., Relier, J.P. and Monod, N. Neonatal electroencephalography during the first twenty-four hours of life in fullterm newborn infants. Neuropediatrics, 1986, 17: 11–18. Rose, A.L. and Lombroso, C.T. A study of clinical, pathological, and electroencephalographic features in 137 full-term babies with a longterm follow-up. Pediatrics, 1970, 45: 404–425. Sarnat, H.B. and Sarnat, M.S. Neonatal encephalopathy following fetal distress. A clinical and electroencephalographic study. Arch. Neurol., 1976, 33: 696–705. Stockard, J.E. and Stockard, J.J. Recording and analysing. In: E.J. Moore (Ed.), Bases of Auditory Brain-Stem Evoked Responses. Grune and Stratton, New York, 1983, pp. 255–286. Stockard, J.E., Stockard, J.J., Kleinberg, F. and Westmoreland, B.F. Prognostic value of brain-stem auditory evoked potentials in neonates. Arch. Neurol., 1983, 40: 360–365. Taylor, M.J., Menzies, R., MacMillan, L.J. and Whyte, H.E. VEPs in normal full-term and premature neonates: longitudinal versus cross-sectional data. Electroenceph. clin. Neurophysiol., 1987, 68: 20–27. Taylor, M.J., Murphy, W.J. and Whyte, H.E. Prognostic reliability of somatosensory and visual evoked potentials of asphyxiated term infants. Dev. Med. Child Neurol., 1992, 34: 507–515. van Lieshout, H.B.M., Jacobs, J.W.F.M., Rotteveel, J.J., Geven, W. and v’t Hof, M. The prognostic value of the EEG in asphyxiated newborns. Acta Neurol. Scand., 1995, 91: 203–207. Volpe, J.J. Hypoxic-ischemic encephalopathy: neuropathology and pathogenesis. In: J.J. Volpe (Ed.), Neurology of the Newborn. W.B. Saunders, Philadelphia, PA, 1995, pp. 279–313. Watanabe, K., Miyazaki, S., Hara, K. and Hakamada, S. Behavioral state cycles, background EEGs and prognosis of newborns with perinatal hypoxia. Electroenceph. clin. Neurophysiol., 1980, 49: 618–625.

E. Scalais et al. / Electroencephalography and clinical Neurophysiology 108 (1998) 199–207 Whyte, H.E., Taylor, M.J., Menzies, R.M., Chin, K.C. and MacMillan, L.J. Prognostic utility of visual evoked potentials in term asphyxiated neonates. Pediatr. Neurol., 1986, 2: 220–223. Willis, J., Seales, D. and Frazier, E. Short latency somatosensory evoked

207

potentials in infants. Electroenceph. clin. Neurophysiol., 1984, 59: 366– 373. Willis, J., Duncan, C. and Bell, R. Short-latency somatosensory evoked potentials in perinatal asphyxia. Pediatr. Neurol., 1987, 3: 203–207.