Magnetic resonance imaging in neonatal hypoxic-ischemic brain injury
Liauw, L.
Citation
Liauw, L. (2009, March 19). Magnetic resonance imaging in neonatal hypoxic-ischemic brain injury. Retrieved from
https://hdl.handle.net/1887/13690
Version: Corrected Publisher’s Version
License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden
Downloaded from: https://hdl.handle.net/1887/13690
Chapter 7
Prediction of short-term neurological outcome in full-term neonates with hypoxic-ischaemic encephalopathy based on combined use of electroencephalogram and neuro-imaging
Lara Leijser Alla Vein Lishya Liauw Tzipi Strauss Sylvia Veen
Gerda van Wezel-Meijler
Neuropediatrics 2007; 38 (5): 219-227
Chapter 7
Background
In infants with hypoxic-ischaemic encephalopathy (HIE), prediction of the prognosis is based on clinical, neuro-imaging and neurophysiological parameters.
Methods
EEG, cranial ultrasound, MRI and follow-up findings of 23 infants (GA 35-42 weeks) with HIE were studied retrospectively to assess:
1. the contribution of ultrasound, MRI and EEG in predicting outcome, 2. the accuracy of ultrasound as compared to MRI, and
3. whether patterns of brain damage and EEG findings are associated.
Results
An abnormal EEG background pattern was highly predictive of adverse outcome (positive predictive value (PPV) 0.88). If combined with diffuse white and deep and/or cortical grey matter changes on ultrasound or MRI, the PPV increased to 1.00. Abnormal neuro-imaging findings were also highly predictive of adverse outcome. Abnormal signal intensity in the posterior limb of the internal capsule, and diffuse cortical grey matter damage were associated with adverse outcome. MRI showed deep grey matter changes more frequently than ultrasound. Severely abnormal neuro-imaging findings were always associated with abnormal EEG background pattern.
Conclusions
Both early EEG and neuro-imaging findings are predictive of outcome in infants with HIE.
The predictive value of EEG is strengthened by neuro-imaging.
Acknowledgements
We are grateful to Prof. Dr. FJ Walther for critically reading this manuscript and to Dr. S le Cessie for her statistical support.
Introduction
Hypoxic-ischaemic encephalopathy (HIE) is a major cause of brain damage and neurodevelopmental abnormalities in full-term newborn infants (1). In infants with HIE, different types of brain lesions are encountered (2-5). The clinical diagnosis of HIE is based on a combination of prenatal and postnatal criteria (6). The severity of HIE varies from stage 1, which is associated with normal outcome, to stage 3, which is associated with high mortality and severe neurodevelopmental abnormalities.
Following HIE stage 2 outcome varies widely, from normal neurodevelopment to severe cerebral palsy, mental retardation and/or epilepsy (6-8).
An accurate prediction of the prognosis of individual infants is important for clinicians and parents. In some cases the prediction of an adverse outcome may lead to withdrawal of intensive care. In other cases prediction of an adverse outcome will lead to early supportive care, such as physiotherapy and speech therapy.
Neurophysiological and neuro-imaging findings are helpful in establishing the prognosis in infants with HIE (9-11). The relation between conventional clinical parameters (i.e., Apgar score, Sarnat score and cord blood pH) and outcome has been assessed by various authors (6,12-14). The Sarnat score is significantly related to outcome (6,12) but Apgar score and cord blood pH correlate poorly with outcome (12-14).
Electroencephalography (EEG) may show characteristic patterns that can predict outcome after asphyxia when recordings are made within the first one to three days after birth (15-17). However, its predictive value becomes less accurate with time (11). Epileptiform discharge on EEG is much less accurate in predicting neurodevelopmental outcome than background activity (9).
Cranial ultrasonography (CUS) may not detect some frequently encountered patterns of hypoxic brain damage in full-term infants, including cortical, subcortical and/or basal ganglia lesions (18-20). MRI studies have contributed to a better understanding of brain lesions and outcome in infants with HIE (2-4,9). Studies comparing CUS and MRI in infants with HIE suggest that CUS and MRI are complementary tools in predicting outcome (18).
Neuro-imaging findings and EEG background patterns have predictive value in infants with HIE (9,21) but few data are available on the association between patterns of brain damage and EEG findings (9). In addition, the combined predictive value of CUS, MRI and EEG in infants with HIE has, to our knowledge, so far not been evaluated.
This study evaluates the combined value of neuro-imaging (both CUS and MRI) and EEG for predicting short-term neurological prognosis in (near) term infants with HIE.
Chapter 7
The aims of the study among newborn infants with HIE were:
1. to assess the contribution of CUS, MRI and EEG in predicting short-term outcome,
2. to assess the accuracy of CUS as compared to MRI, and
3. to evaluate whether patterns of brain damage, detected with neuro- imaging techniques, are associated with EEG patterns.
Patients and Methods
Patients
Infants born between January 2002 and July 2004 after a gestational age (GA) of 35 weeks or more were studied retrospectively if they had a clinical diagnosis of perinatal asphyxia, and were born at or transferred to the neonatal intensive- care unit of the Leiden University Medical Center (LUMC). Infants were excluded if septicaemia, meningitis, a metabolic disorder, congenital cardiac abnormality or a specific syndrome was suspected or proven. All infants underwent sequential CUS examinations and at least one EEG and MRI examination in the first two weeks after birth. The diagnosis of perinatal asphyxia was based on the following criteria:
• signs of fetal distress before delivery (abnormal cardiotocography recording such as decreased variability, late deceleration, baseline bradycardia),
• Apgar score < 7 at five minutes,
• arterial umbilical cord pH < 7.1, and
• clinical signs of HIE (such as irritability, convulsions, lethargy) (6).
Medical records, neuro-imaging and EEG findings, and follow-up data were reviewed. HIE was staged according to Sarnat and Sarnat (6).
Neuro-imaging Ultrasonography
CUS examinations were done using a standard scanning protocol, i.e., at least within 24 hours after birth, on the third day after birth and biweekly until discharge or transferral to another hospital. Scanning was performed with an Aloka 5500 scanner with a multifrequency transducer (Biomedic Nederland B.V., Almere, the Netherlands) in at least six coronal and five sagittal planes, using a transducer frequency of 7.5 MHz. In addition, in order to detect cortical and/or subcortical abnormalities, scanning was also done at a higher frequency (10 MHz) and for optimal visualization of deeper structures (posterior fossa) at a lower frequency (5 MHz). All images were saved on magneto-optical disks and later reviewed by two investigators (GvWM and LL) who were blinded to the EEG and MRI findings and outcome of the infants.
One or more of the following four final CUS descriptions was/were made:
1. No echodensities or only periventricular echogenic changes of less than seven days duration.
2. Echogenic changes in periventricular or deep white matter (other than mentioned in 1).
3. Echogenic changes in deep grey matter (thalamus and/or basal ganglia).
4. Echogenic changes in cortical grey matter.
Special note was made whether abnormalities were localized or diffuse and where they were located. The presence of changes consistent with haemorrhage was recorded.
MRI
MRI examinations were performed within one week after birth on a 1.5 Tesla MRI system (Philips Medical Systems, Best, the Netherlands), using a head coil especially designed for imaging the neonatal head. Conventional T1- and T2- weighted images were obtained in the transverse plane according to a standardized MRI protocol for newborn infants. T1-weighted images were also obtained in the sagittal plane. In addition, diffusion-weighted and FLAIR images were performed in the transverse plane. In most cases, contrast images (using gadolinium) were performed. MR images were evaluated immediately after the scanning procedure by the attending neuroradiologist. Investigators (LL and GvWM) who were blinded to the CUS, EEG and follow-up findings later reviewed the images. Special attention was paid to the signal intensity (SI) of white matter, deep grey matter and cortical grey matter. On T1-and T2-weighted images the distinction between cortex and subcortical white matter was recorded. In addition, the presence of haemorrhage was noted and in infants with a postconceptional age over 38 weeks, the SI of the posterior limb of the internal capsule (PLIC) was recorded as normal, equivocal or abnormal (22).
In analogy to the final CUS descriptions, one or more of the following final MRI descriptions, divided into four groups, was/were noted:
1. No ischaemic changes.
2. Ischaemic changes in periventricular or deep white matter.
3. Ischaemic changes in deep grey matter (thalamus and/or basal ganglia).
4. Ischaemic changes in cortical grey matter.
Special note was made whether abnormalities were localized or diffuse and where they were located.
Chapter 7
Grading of CUS and MRI findings
CUS and MRI findings were classified into eight grades according to the predominant pattern and severity of changes, as adapted from the classification by Mercuri et al. (23).
1. Normal: Normal basal ganglia and thalami, and normal white matter with or without localized cortical involvement (CUS or MRI description 1 or 4 localized).
2. Mild basal ganglia and thalami: Localized basal ganglia and/or thalamic lesions with or without localized cortical involvement but normal white matter and on MRI normal PLIC (description 3 localized with or without 4 localized).
3. Moderate white matter: Localized white matter lesions with or without cortical involvement but normal basal ganglia and thalami and on MRI normal PLIC (description 2 localized with or without 4 localized).
4. Moderate basal ganglia and thalami: Localized basal ganglia and/or thalamic lesions and on MRI abnormal PLIC with or without localized cortical involvement (description 3 localized and abnormal PLIC with or without 4 localized).
5. Moderate white matter and basal ganglia and thalami: Localized white matter lesions and localized basal ganglia and thalamic lesions and on MRI abnormal PLIC, with or without cortical involvement (descriptions 2 localized and 3 localized and abnormal PLIC with or without 4).
6. Severe white matter: Diffuse or localized white matter lesions with diffuse cortical involvement but normal basal ganglia and thalami and on MRI normal PLIC (description 2 localized/diffuse with 4 diffuse).
7. Severe basal ganglia and thalami with subcortical white matter: Diffuse basal ganglia and thalami and white matter lesions with abnormal PLIC on MRI and localized cortical involvement (descriptions 3 diffuse and 2 diffuse with 4 localized and abnormal PLIC).
8. Severe basal ganglia and thalami with diffuse white matter: Diffuse basal ganglia and thalami and white matter lesions with abnormal PLIC on MRI and diffuse cortical involvement (descriptions 3 diffuse and 2 diffuse with 4 diffuse and abnormal PLIC).
In order to obtain the most accurate comparison between the CUS and MRI examinations, the CUS scan performed closest to the MRI examination was used.
CUS and MRI findings were considered identical if CUS and MRI description(s) were exactly the same, overlapping if one or more but not all CUS and MRI description(s) were the same, and different if the description(s) were completely different.
Electroencephalography
EEG was performed within the first 3 days after birth, using a Nihon Kohden 2100 system (Biomedic Nederland B.V., Almere, the Netherlands). The EEGs contained 16 EEG bipolar leads of the international 10-20 system, electro-oculogram, electrocardiogram, pneumogram, and submental electromyogram. All EEGs were recorded with a simultaneous video registration for at last 45 minutes and evaluated by the attending neurophysiologist. An investigator was blinded to the neuro-imaging findings and outcome of the infants later reviewed the recordings.
The data obtained by EEG were classified according to continuity and symmetry of frequencies, amplitude of background activity and presence of electrographic epileptic seizure activity. Epileptiform activity was defined as abnormal repetitive discharge with a clear onset and typical evolution in space and time, lasting more than ten seconds. Status epilepticus was considered present it the total duration of the epileptiform activity exceeded 30 minutes or the sum of the epileptic discharges exceeded 50% of the record (24).
The EEG patterns were divided into the following four groups, according to severity:
1. Normal background and no epileptiform discharge.
2. Normal background and presence of epileptiform discharge.
3. Abnormal background (i.e. discontinuous and/or asymmetrical and/or low voltage (all activities amplitude ≤ 30 μV)) and no epileptiform discharge.
4. Abnormal background and presence of epileptiform discharge.
Special note was made in case of status epilepticus.
In each recording, it was taken into consideration that anti-epileptic treatment may have affect on the background pattern of the neonatal EEG (25). However, as we divided EEG background pattern into two groups (i.e., normal and abnormal), treatment did not interfere with the assessment of the EEG.
Follow-up
Surviving infants visited the outpatient clinic of our own hospital or the referring hospital at regular intervals until two years of age. At two years of age, all infants underwent a standardized neurological examination according to Hempel (26).
The diagnosis of cerebral palsy was made according to the criteria of Hagberg et al. (27).
The infants were assigned to one of the following four groups, based on outcome at two years of age:
1. Normal, i.e., normal neurological examination.
2. Moderately abnormal, i.e., mildly abnormal neurological findings (mild hypertonia, hypotonia and/or asymmetry).
3. Severely abnormal, i.e., cerebral palsy.
4. Death.
Special note was made of visual and/or hearing problems and/or epilepsy.
For this study, the outcome groups 1 and 2 were considered as favourable outcome, whereas outcome groups 3 and 4 were considered as adverse outcome.
Chapter 7
Statistical analysis
Statistical analysis was performed using SPSS. To assess the strength of the association between CUS findings (graded from 1 to 8), MRI findings (graded from 1 to 8) or EEG findings (graded from 1 to 4) and outcome at follow-up (graded from 1 to 4), a nonparametric Spearman correlation coefficient was calculated. To assess the correlation between CUS and MRI findings, a simple linear regression coefficient was calculated for the CUS and MRI grades. Predictive values of CUS findings, MRI findings, EEG background pattern and presence of epileptiform discharge on EEG for outcome at two years of age were calculated. For these calculations CUS and MRI grades 1 and 2 were considered normal to mildly abnormal, grades 3, 4 and 5 moderately abnormal and grades 6,7,8 severely abnormal. Level of significance was 5%.
Results
Patients
After exclusions, data of 23 infants (14 male) were studied (Table 1). The mean GA of the infants was 39.0 (range 35.0-42.3) weeks and the mean birth weight was 3150 (range 2200-4125) grams. Mean Apgar scores were 3 (range 0-6), 4 (range 1-6) and 6 (range 2-10) at respectively 1, 5 and 10 minutes after birth and mean umbilical cord pH was 7.0 (range 6.6-7.1). Two infants were classified HIE stage 1, 12 infants stage 2, and nine infants stage 3.
Clinical signs of convulsions were observed in 18 infants, all of whom received anti-epileptic drugs. Eight of the 23 infants died while hospitalized at a mean age of 7.4 (range 2-19) days after birth. Death was related to severe neurological problems in all of them. In one of these infants parental permission for autopsy of the brain was obtained.
Table 1 Perinatal and postnatal clinical data of the 23 studied infants with HIE.
Case No
Sex GA at birth (weeks)
BWt (grams)
Mode of delivery
Complication during delivery
Apgar score (1/5/10 min)
Umbilical cord pH
Postnatal support
Ventilation HIE stage
Convulsions
1 M 40+1 3770 CS Massive placental
haemorrhage
2/2/10 6.83 CC + intub Yes 3 Yes
2 F 38+3 3040 SVD Meconium staining 5/6/9 7.02 None No 3 No
3 M 35+0 3300 EmCS 1/4/6 6.85 CC + intub Yes 2 Yes
4 F 42+1 3500 SVD Meconium staining 0/n.d./n.d. 6.84 CC + intub Yes 2 Yes
5 M 36+4 2750 CS Cord prolapse 0/4/4 6.60 CC + intub Yes 2 Yes
6 M 36+3 4125 CS Meconium staining 6/4/6 7.03 None No 2 Yes
7 M 42+0 3325 SVD Meconium staining 4/4/7 7.00 None Yes 1 No
8 M 41+1 3027 Vacuum Meconium staining 3/6/8 7.06 Intub Yes 2 Yes
9 M 37+1 2685 Vacuum 0/1/2 6.90 Intub Yes 3 Yes
10 M 41+5 3860 SVD Meconium staining 5/6/8 7.05 Intub Yes 2 Yes
11 F 40+0 2200 EmCs Meconium staining 3/3/n.d. 6.90 CC + intub Yes 3 No
12 F 37+1 2630 EmCS Placental abruption 2/5/5 n.d. CC + intub Yes 2 Yes
13 M 41+1 3380 Vacuum Cord entanglement 4/5/6 6.97 Intub Yes 1 No
14 M 37+0 2900 EmCs Massive placental haemorrhage
0/2/6 n.d. CC + intub Yes 2 Yes
15 F 37+3 2460 SVD 2/6/7 n.d. None No 3 Yes
16 M 36+6 3160 SVD 4/5/6 7.04 Intub Yes 3 No
17 M 37+3 3080 Vacuum 6/6/6 7.05 Intub Yes 3 Yes
18 F 42+2 3345 EmCS Placenta previa + meconium staining
3/5/8 6.99 Intub Yes 2 Yes
19 M 38+1 3185 EmCS 0/4/6 7.09 CC + intub Yes 2 Yes
20 F 41+4 3420 SVD 1/4/4 7.05 Intub Yes 3 Yes
21 F 37+2 3100 EmCS Cord prolapse 6/6/7 7.04 Intub Yes 3 Yes
22 F 39+0 2605 SVD Massive placental haemorrhage + cord entanglement
2/6/6 7.07 Intub Yes 2 Yes
23 M 41+6 3600 CS 2/2/n.d. n.d. Intub Yes 2 Yes
No= number; GA= gestational age; BWt=birthweight; M=male; F=female; CS=Caesarian section;
SVD= spontaneous vaginal delivery; EmCS=emergency Caesarian section; n.d.= not documented;
CC=cardiac compresssion; intub= intubation
Neuro-imaging, EEG and follow-up findings
CUS, MRI and EEG findings are presented in Table 2. Motor outcome is shown in Table 3. Mean postnatal age was 4.0 (range 1-7) days at MRI examination and 2.0 (range 1-3) days at EEG.
Examples of CUS and MRI examinations of studied infants are depicted in Figures 1-6.
Chapter 7
Table 2 CUS, MRI, EEG and outcome findings, grouped according to the extent of changes on CUS compared to on MRI.
Outcome group
Case No
Pn day MRI
Pn day EEG
Sarnat score
CUS descriptions
CUS grade
MRI descriptions
PLIC MRI
grade EEG group
Extent of changes CUS vs MRI
1 13 5 3 1 2,4 localized 3 2,4 localized Normal 3 1 =
1 6 2 3 2 2,4 localized 3 2,4 localized Normal 3 1 =
1 8 4 2 2 2,4 localized 3 2,4 localized Normal 3 2 =
1 22 2 1 2 2,4 localized 3 2,4 localized Normal 3 2 =
2 7 4 3 1 4 localized 1 4 localized Normal 1 1 =
2 3 7 3 2 2 localized 3 2 localized Normal 3 3 =
3 4 2 2 2 1 1 1 Normal 1 3 =
3 11 7 1 3 2,3,4 diffuse 8 2,3,4 diffuse Abnormal 8 3 =
4 1 4 3 3 2,4 diffuse 6 2,4 diffuse Normal 6 3 =
4 16 1 3 3 2,3,4 diffuse 8 2,3,4 diffuse Abnormal 8 3 =
4 17 5 2 3 2,3,4 diffuse 8 2,3,4 diffuse Abnormal 8 4 =
4 20 5 2 3 2,3,4 localized 8 2,3,4 diffuse Abnormal 8 3 =
3 19 1 1 2 2,3 localized 5 2,3,4 diffuse,
localized
Abnormal 7 4 <
3 10 5 3 2 2,4 localized 4 2,3,4 diffuse Abnormal 8 4 <
3 23 1 1 2 2,4 localized 3 2,3,4 localized Normal 5 2 <
3 21 7 3 3 3,4 localized 4 2,3,4 diffuse,
localized
Abnormal 7 4 <
3 15 6 2 3 2,4 localized 3 4 localized Abnormal 4 4 <
4 5 1 1 2 2,4 localized 3 2,3,4 diffuse,
localized
Abnormal 7 3 <
4 2 4 3 3 2,3 diffuse 7 2,3,4 diffuse Abnormal 8 3 <
4 9 6 1 3 2,3 localized 5 2,3 diffuse,
localized
Abnormal 7 3 <
4 18 3 2 2 2,3,4 diffuse 8 2,4 diffuse Abnormal 6 4 >
1 14 7 1 2 3 localized 2 1 Normal 1 1 n.a.
2 12 4 1 2 1 1 3 localized Normal 2 3 n.a.
=, similar extent on CUS and MRI; <, less extensive changes on CUS compared to MRI;
>, more extensive changes on CUS compared to MRI
No=number; pn=postnatal; PLIC= posterior limb of internal capsule; n.a.=not applicable
Table 3 Motor outcome at 2 years of age (n= number of infants).
Motor outcome Number of infants
Outcome at 2 years
Normal 5 normal
Moderately abnormal
3 mild hypertonia of both legs, axial hypotonia (n=1);
(axial) hypotonia, dystonia (n=1);
mild axial hypotonia (n=1)
Figure 1 Parasagittal CUS scan (case 17;
CUS grade 8, MRI grade 8) showing subtle increase in echogenicity in the thalami and basal ganglia (arrows).
Figure 2 Parasagittal CUS scan (case 1; CUS grade 6, MRI grade 6) showing a widening of the hypoechogenic cortical rim (short arrows) and increased echogenicity in the periventricular white matter (long arrow).
a b
Figure 3
a) Coronal CUS scan (case 2; CUS grade 7, MRI grade 8) showing increased echogenicity in the thalami (arrows), more obvious on the right side than on the left.
b) Coronal CUS scan in the same infant showing increased echogenicity in the parietal periventricular white matter (arrows).
Chapter 7
a b
Figure 4
(same infant as in Figure 3)
a)Transverse T1-weighed MR image showing high signal intensity in the basal ganglia (long arrow) and loss of normal high signal in the posterior limb of the internal capsule (short arrow).
b) Transverse T2-weighted MR image in the same infant again showing abnormal signal intensity in the basal ganglia and posterior limb of the internal capsule, also showing increased signal intensity in the white matter (arrows) , being most obvious in the frontal region.
a b
Figure 5 a and b) Coronal and parasagittal CUS scan (case 16; CUS grade 8, MRI grade 8) showing loss of normal architecture and diffusely increased echogenicity in the white matter (arrows).
Figure 6 Transverse T2-weighted MR image in the same infant as in Figure 5 showing diffuse brain swelling, loss of normal architecture and abnormal signal intensity throughout the brain.
Relation between CUS fi ndings and MRI fi ndings
Table 2 shows the CUS and MRI fi ndings. In 12 infants CUS and MRI fi ndings were identical. The linear regression correlation coeffi cient between CUS and MRI grades was 0.83, which was highly signifi cant (p < 0.0001).
Brain autopsy fi ndings and relation with CUS and MRI fi ndings In one infant brain autopsy was performed (case 2). The results of the autopsy were in accordance with the neuro-imaging fi ndings.
Association between neuro-imaging fi ndings and EEG fi ndings Diffuse white matter and deep and/or cortical grey matter changes on CUS and/
or MRI, i.e., stages 6 to 8 (severely abnormal), were always associated with an abnormal EEG background pattern. Abnormal SI in the PLIC on MRI was also always associated with an abnormal EEG background pattern. No obvious relation was found between other patterns of brain injury and EEG fi ndings.
CUS, MRI and EEG related to outcome
Tables 2, 4 show the relation between CUS, MRI and EEG fi ndings and outcome at follow-up. The predictive values of neuro-imaging and/or EEG fi ndings for outcome are presented in Table 5.
Chapter 7
Table 4 Relation between MRI, CUS, EEG and outcome.
MRI stage CUS stage
EEG group
1 2 3 4
1 (n=3) 1 (n=2) □ ●
2 (n=1) Ο
2 (n=1) 1 (n=1) □
3 (n=5) 3 (n=5) ΟΟ ΟΟ □
4 (n=1) 3 (n=1) ●*
5 (n=1) 3 (n=1) ●
6 (n=2) 6 (n=1) ▪
8 (n=1) ▪* #
7 (n=4) 3 (n=1) ▪*
4 (n=1) ●*
5 (n=2) ▪* ●*
8 (n=6) 4 (n=1) ●*
7 (n=1) ▪*
8 (n=4) ●*▪*▪* ▪*#
Ο normal outcome; □ moderately abnormal outcome; • severely abnormal outcome;
N N=number of infants
Table 5 Predictive values of neuro-imaging findings and/or EEG findings for motor outcome at 2 years of age. The predictive values (i.e. sensitivity, specificity, PPV and NPV) of normal to mildly abnormal and severely abnormal neuro-imaging findings, alone and in combination with EEG background findings, for, respectively, favourable and adverse outcome at 2 years of age. In addition, the predictive values of normal and abnormal EEG background pattern, and absence and presence of epileptiform discharge on EEG for, respectively, favourable and adverse outcome at 2 years of age (PPV, positive predictive value; NPV, negative predictive value).
Findings Predictive values for outcome Sensitivity Specificity PPV NPV
CUS Normal-mildly abnormal 0.38 0.93 0.75 0.74
Severely abnormal 0.47 1.00 1.00 0.50
MRI Normal-mildly abnormal 0.38 0.93 0.75 0.74
Severely abnormal 0.74 1.00 1.00 0.67
EEG - background Normal 0.75 0.93 0.86 0.88
Abnormal 0.93 0.75 0.88 0.86
EEG - epileptiform discharge
Without 0.47 0.75 0.78 0.43
With 0.75 0.47 0.43 0.78
CUS + EEG Normal to mildly abnormal + normal background
0.67 1.00 1.00 0.50
Severely abnormal CUS findings were highly predictive of an adverse outcome, while normal or mildly abnormal CUS often predicted a favourable outcome. The Spearman correlation coefficient between CUS findings and outcome at follow-up was 0.67, which was highly significant (p=0.001).
Severely abnormal MRI findings were also highly predictive of an adverse outcome, while normal or mildly abnormal MRI often predicted a favourable outcome. The Spearman correlation coefficient between MRI findings and outcome at follow-up was 0.77, which was highly significant (p<0.0001).
All infants with abnormal SI in the PLIC had an adverse outcome. Three (case 1, 4, 23) of the 15 infants with adverse outcome had normal SI in the PLIC, of whom two (cases 1 and 23) had white matter, deep grey matter and/or cortical grey matter damage. The other infant (case 4) developed arterial infarction after MRI was performed on day 2. In addition, diffuse cortical grey matter damage (cases 10, 11, 18, 1, 2, 16, 17 and 20) was always associated with an adverse outcome.
Abnormal EEG background activity was highly predictive of an adverse outcome, while a normal EEG background pattern was highly predictive of a favourable outcome. The Spearman correlation coefficient between EEG findings and outcome at follow-up was 0.71, which was highly significant (p< 0.0001).
The presence of epileptiform discharge on EEG had a poor predictive value for an adverse outcome, while the absence of epileptiform discharge predicted a favourable outcome. The two infants with electroencephalographic status epilepticus had an adverse outcome (both infants died).
Combining the CUS findings with EEG background pattern did not change the predictive value of CUS findings for adverse outcome, but did improve the predictive value of normal or mildly abnormal CUS for favourable outcome. It also improved the predictive value of abnormal EEG background activity for adverse outcome and of normal EEG background pattern for favourable outcome.
Likewise, combining the MRI findings with EEG background pattern did not change the predictive value of MRI findings for adverse outcome, but improved the predictive value of normal or mildly abnormal MRI for favourable outcome. It also improved the predictive value of abnormal EEG background activity for adverse outcome and of normal EEG background pattern for favourable outcome.
The combination of abnormal EEG background pattern (EEG groups 3-4) with diffuse white matter and diffuse deep and/or cortical grey matter changes on CUS or MRI (grade 6, 7 or 8) was strongly associated with an adverse outcome (outcome groups 3-4). In all (n=7) but one infant (case 23) with normal EEG background pattern (EEG groups 1-2) combined with normal CUS/MRI or localized white matter or grey matter changes (grade 1, 2 or 3) outcome was favourable (outcome groups 1-2). In all (n=16) but one infant (case 12), abnormal EEG background pattern was associated with an adverse outcome. One infant with normal CUS and MRI (grade 1) but abnormal EEG background, was severely abnormal at follow-up (case 4). This infant initially had normal early neuro- imaging findings and was discharged early. Arterial infarction was diagnosed after discharge.
Chapter 7
Discussion
This study shows that early EEG background pattern accurately predicts motor outcome in (near) term newborn infants with HIE and that the predictive value of abnormal EEG background increases when early EEG findings are combined with neuro-imaging findings. The good predictive value of EEG background pattern found in this study is consistent with other studies, showing interictal EEG background activity during the first days to be predictive of outcome (9,17). The presence of epileptiform discharge on EEG was not predictive of an adverse outcome; of the 15 infants with an adverse outcome, eight had no epileptiform discharge, and three out of the nine infants with epileptiform discharge had favourable outcome (2 normal, 1 moderately abnormal). This is consistent with a previous study, showing that infants with epileptiform discharge on EEG may have a normal outcome (9). Consistent with previous studies (9,17), electroencephalographic status epilepticus, although only detected in two infants, was associated with an adverse outcome.
This study also shows that severely abnormal neuro-imaging findings reliably predict adverse motor outcome at two years of age, but that normal or mildy abnormal neuro-imaging findings do not guarantee a favourable outcome. The predictive value of normal or mildly abnormal neuro-imaging findings for favourable outcome increased by adding early EEG findings. Some authors suggest that in infants with HIE, CUS is not as predictive of outcome as MRI and that subtle abnormalities, particularly in the deep and cortical grey matter and subcortical white matter, are difficult to detect or are easily overlooked on CUS (18-20). In five infants in our study, deep grey matter (four infants: cases 12, 10, 23, 5) or white matter (one infant: case 21) changes remained undetected with CUS, while they were seen on MRI. In one infant (case 18), CUS showed more changes (i.e., diffuse deep grey matter changes) than MRI. CUS may have overestimated brain damage in this infant. However, it is more likely that CUS accurately detected these changes, since this infant had an adverse outcome, and abnormal CUS findings were also highly predictive of outcome. It is possible that MRI did not detect mild and/or small changes, since during the study period 4 mm slices with gaps were applied, while nowadays 2 mm slices without gap are used.
There were only slight differences in positive predictive value for adverse outcome between EEG findings, CUS findings and MRI findings, while the differences in negative predictive value were considerable; normal to mildly abnormal neuro- imaging (in particular CUS) was not a strong predictor for a favourable outcome.
These differences may have been more obvious if large numbers of infants had been studied.
Neuro-imaging findings supported the predictive value of EEG background pattern.
The combination of abnormal EEG background pattern with diffuse white matter and
In the infant with normal EEG background pattern combined with only localized white and grey matter changes but adverse outcome (case 23), EEG and MRI were performed on day 1. As abnormalities on conventional MRI may not be visible within the first few days of the hypoxic-ischemic insult (4,28), more severe abnormalities, that could explain the adverse outcome of this infant, may have been missed. Unfortunately, no diffusion-weighted imaging was performed in this infant.
One infant with normal CUS and MRI but abnormal EEG background pattern was severely abnormal (i.e., had a hemiparesis) at follow-up (case 4). This infant had the MRI examination performed on day 2 and was discharged and transferred to another hospital soon thereafter, after which arterial infarction was diagnosed.
In one infant (case 12) with abnormal EEG background and normal CUS and localized deep grey matter changes on MRI, outcome was only moderately abnormal. So, in this infant the neuro-imaging findings were more predictive than the EEG findings. The reason for the inconsistency between neuro-imaging and EEG findings in this infant is not clear.
The similarity between CUS and MRI findings for most infants in our study is inconsistent with some other studies (3), showing that CUS may fail to detect deep grey matter, cortical grey matter and/or subcortical white matter changes.
CUS techniques and protocols may have been different in those studies. MRI was complementary to CUS in most cases, which is consistent with a previous study (18). MRI provided more precise information about localization and extent of lesions. However, since CUS can be performed at the bedside and repeated easily, even in very sick, unstable infants, we feel that MRI cannot replace CUS and that both techniques are complementary. We also feel that CUS, if applied under favourable circumstances (i.e., good equipment, experienced ultrasonographer, optimal timing of serial examinations), can detect both white matter and deep and cortical grey matter changes (29). In addition, since abnormal CUS was highly predictive of adverse outcome, we feel that in full-term neonates with HIE, abnormalities that are significant for motor outcome are reliably detected with CUS.
We found that abnormal SI in the PLIC was associated with an adverse outcome.
This is consistent with the results of a study by Rutherford et al. who found abnormal outcome in all 36 infants with SI changes in the PLIC (22). Three infants (cases 4, 1, 23) with normal SI in the PLIC had an adverse outcome. In case 1, this is probably explained by the presence of white matter and cortical grey matter damage. The infant without abnormalities on CUS/MRI but adverse outcome was the infant who developed arterial infarction after early neuro-imaging was performed (case 4). Also in case 23 neuro-imaging was probably performed too early.
Diffuse cortical involvement was also associated with an adverse outcome.
Rutherford et al. have reported spastic diplegia, microcephaly and intellectual deficit in infants with widespread cortical damage (28).
Severely abnormal CUS and MRI findings were always associated with an abnormal EEG background pattern, but normal to mildly abnormal and moderately abnormal
Chapter 7
consistent with a study by Biagioni et al. (9). They also showed that severe deep grey and/or white matter lesions were associated with abnormal EEG background pattern and moderate lesions with variable background patterns. However, in their study normal to mildly abnormal MRI findings were always associated with normal EEG background pattern. This inconsistency in findings may have been due to the small number of infants we studied.
We conclude from this retrospective study that in (near) term newborns with HIE, early EEG is a reliable tool for predicting outcome of individual infants. Early EEG predicts adverse outcome as accurately as neuro-imaging, but favourable outcome is predicted better by early EEG. However, neuro-imaging is indispensable in these infants, since EEG background pattern loses its predictive value after the first few days after birth (11), and since both severely abnormal CUS and MRI were predictive of adverse motor outcome and supported the predictive value of early EEG. In addition, in several infants (cases 23, 12) neuro-imaging was more predictive than early EEG. Abnormal SI in the PLIC on MRI and diffuse cortical grey matter damage were strongly associated with an adverse outcome.
CUS and MRI findings were mostly comparable and MRI, showing more exact location and extent of abnormalities, was complementary to CUS in most cases.
Finally, severely abnormal neuro-imaging findings were always associated with an abnormal EEG background pattern.
References
1. Volpe JJ. Hypoxic-ischemic encephalopathy: neuropathology and pathogenesis. In: Volpe JJ, ed.
Neurology of the newborn. 4th ed. Philadelphia: WB Saunders Company; 2001:296-330.
2. Barkovich AJ, Truwit CL. Brain damage from perinatal asphyxia: correlation of MR findings with gestational age. AJNR 1990;11:1087-1096.
3. Keeney SE, Adcock EW, McArdle CB. Prospective observations of 100 high-risk neonates by high- field (1.5 Tesla) magnetic resonance imaging of the central nervous system. II. Lesions associated with hypoxic-ischemic encephalopathy. Pediatrics 1991;87:431-438.
4. Rutherford MA, Pennock JM, Schwieso JE, Cowan FM, Dubowitz LM. Hypoxic ischaemic encephalopathy: early magnetic resonance imaging findings and their evolution. Neuropediatrics 1995;26:183-191.
5. Triulzi F, Parazzini C, Righini A. Patterns of damage in the mature neonatal brain. Pediatr Radiol 2006;36:608-620.
6. Sarnat HB, Sarnat MS. Neonatal encephalopathy following fetal distress. A clinical and electroencephalographic study. Arch Neurol 1976;33:696-705.
7. Finer NN, Robertson CM, Richards RT, Pinnell LE, Peters KL. Hypoxic-ischemic encephalopathy in term neonates: perinatal factors and outcome. J Pediatr 1981;98:112-117.
8. Levene MI, Sands C, Grindulis H, Moore JR. Comparison of two methods of predicting outcome in perinatal asphyxia. Lancet 1986;1:67-69.
9. Biagioni E, Mercuri E, Rutherford M, et al. Combined use of electroencephalogram and magnetic resonance imaging in full-term neonates with acute encephalopathy. Pediatrics 2001;107:461- 468.
10. Eken P, Toet MC, Groenendaal F, de Vries LS. Predictive value of early neuroimaging, pulsed Doppler and neurophysiology in full term infants with hypoxic-ischaemic encephalopathy. Arch Dis Child Fetal Neonatal Ed 1995;73:F75-F80.
11. Pressler RM, Boylan GB, Morton M, Binnie CD, Rennie JM. Early serial EEG in hypoxic ischaemic encephalopathy. Clin Neurophysiol 2001;112:31-37.
12. Finer NN, Robertson CM, Peters KL, Coward JH. Factors affecting outcome in hypoxic-ischemic encephalopathy in term infants. Am J Dis Child 1983;137:21-25.
13. Nelson KB, Grether JK. Potentially asphyxiating conditions and spastic cerebral palsy in infants of normal birth weight. Am J Obstet Gynecol 1998;179:507-513.
14. Yudkin PL, Johnson A, Clover LM, Murphy KW. Clustering of perinatal markers of birth asphyxia and outcome at age five years. Br J Obstet Gynaecol 1994;101:774-781.
15. Holmes G, Rowe J, Hafford J, Schmidt R, Testa M, Zimmerman A. Prognostic value of the electroencephalogram in neonatal asphyxia. Electroencephalogr Clin Neurophysiol 1982;53:60- 72.
16. Holmes GL, Lombroso CT. Prognostic value of background patterns in the neonatal EEG. J Clin Neurophysiol 1993;10:323-352.
17. Wertheim D, Mercuri E, Faundez JC, Rutherford M, Acolet D, Dubowitz L. Prognostic value of continuous electroencephalographic recording in full term infants with hypoxic ischaemic encephalopathy. Arch Dis Child 1994;71:F97-102.
18. Blankenberg FG, Loh NN, Bracci P, et al. Sonography, CT, and MR imaging: a prospective comparison of neonates with suspected intracranial ischemia and hemorrhage. AJNR 2000;21:213-218.
19. Cowan F. Outcome after intrapartum asphyxia in term infants. Semin Neonatol 2000;5:127-140.
20. Ment LR, Bada HS, Barnes P, et al. Practice parameter: neuroimaging of the neonate: report of the Quality Standards Subcommittee of the American Academy of Neurology and the Practice Committee of the Child Neurology Society. Neurology 2002;58:1726-1738.
21. Ortibus EL, Sum JM, Hahn JS. Predictive value of EEG for outcome and epilepsy following neonatal seizures. Electroencephalogr Clin Neurophysiol 1996;98:175-185.
22. Rutherford MA, Pennock JM, Counsell SJ, et al. Abnormal magnetic resonance signal in the internal capsule predicts poor neurodevelopmental outcome in infants with hypoxic-ischemic encephalopathy. Pediatrics 1998;102:323-328.
23. Mercuri E, Ricci D, Cowan FM, et al. Head growth in infants with hypoxic-ischemic encephalopathy:
correlation with neonatal magnetic resonance imaging. Pediatrics 2000;106:235-243.
Chapter 7
24. Clancy RR, Legido A. The exact ictal and interictal duration of electroencephalographic neonatal seizures. Epilepsia 1987;28:537-541.
25. Bell AH, Greisen G, Pryds O. Comparison of the effects of phenobarbitone and morphine administration on EEG activity in preterm babies. Acta Paediatr 1993;82:35-39.
26. Hempel MS. Neurological development during toddling age in normal children and children at risk of developmental disorders. Early Hum Dev 1993;34:47-57.
27. Hagberg G, Hagberg G, Olow I. The changing panorama of cerebral palsy in Sweden 1954-1970.
II. Analysis of the various syndromes. Acta Paediatr Scand 1975;64:193-200.
28. Rutherford M, Pennock J, Schwieso J, Cowan F, Dubowitz L. Hypoxic-ischaemic encephalopathy:
early and late magnetic resonance imaging findings in relation to outcome. Arch Dis Child Fetal Neonatal Ed 1996;75:F145-F151.
29. Eken P, Jansen GH, Groenendaal F, Rademaker KJ, de Vries LS. Intracranial lesions in the full term infant with hypoxic ischaemic encephalopathy: ultrasound and autopsy correlation. Neuropediatrics 1994;25:301-307.
