ABNORMAL SHARP TRANSIENTS ON ELECTROENCEPHALOGRAMS IN PRETERM INFANTS WITH PERIVENTRICULAR LEUKOMALACIA AKIHISA OKUMURA, MD, FUMIO HAYAKAWA, MD, TORU KATO, MD, KOICHI MARUYAMA, MD, TETSUO KUBOTA, MD, MOTOMASA SUZUKI, MD, HIROYUKI KIDOKORO, MD, KUNIYOSHI KUNO, MD, AND KAZUYOSHI WATANABE, MD
Objective To determine the clinical significance of abnormal sharp transients other than positive rolandic sharp waves (PRS), electroencephalograms were used for the diagnosis of periventricular leukomalacia (PVL). Study design We evaluated 126 electroencephalograms from 93 preterm infants; 31 infants had PVL, and 62 were control infants. Frontal sharp waves (FS) were defined as sharp transients of positive polarity with an amplitude >100 lV. Occipital sharp waves (OS) were defined as those of negative polarity with an amplitude >150 lV. FS, OS, or PRS were considered to be present when there were >0.1 per minute. Results The number of FS per minute was significantly higher in the PVL group than in the control group during days 0 to 4 and 5 to 7. The number of OS per minute was also significantly higher in the PVL group than in the control group during days 0 to 4, 5 to 7, and 8 to 14. The sensitivity of FS or OS was relatively high but that of PRS was low. The presence of two or more types of abnormal sharp transients was correlated with a poor outcome. Conclusions
FS or OS may be useful for predicting which infant will have PVL. (J Pediatr 2003;143:26-30)
ositive rolandic sharp waves (PRS) on the electroencephalogram (EEG) of preterm infants are well-known abnormal findings predictive of adverse neurologic outcomes. PRS are associated with white matter injuries caused by periventricular leukomalacia (PVL) as well as intraventricular hemorrhage.1-5 We also previously reported that PRS were related to severe deep white matter injuries because PRS were observed in 14 of 21 infants with moderate or severe PVL but were not seen in infants with mild PVL.6 The average age of the first appearance of PRS was 7.6 days and was 2 weeks earlier than the ultrasonographic abnormalities.6 In addition, PRS appeared in combination with disorganized patterns and should be considered with other chronic-stage EEG abnormalities.6 While analyzing PRS, we noticed that some abnormal transients were present in association with PRS. We hypothesized that these abnormal transients may also be related deep white matter injuries and could be another marker of PVL. The aim of this study was to clarify the clinical significance of abnormal sharp transients other than PRS for the diagnosis of PVL.
P
METHODS Patient Population The study patients were from the population of 238 preterm infants who were born at 32 weeks’ gestation or less that were reported previously.6 The control group was 62 infants without ultrasonographic abnormalities, whose gestational ages and birth weights were matched with the 31 patients with PVL. We retrospectively evaluated the EEG findings. PVL was diagnosed by ultrasonography as periventricular echodensity and/or cystic EEG FS OS PRS PVL
26
Electroencephalogram Frontal sharp waves Occipital sharp waves Positive rolandic sharp waves Periventricular leukomalacia
From the Department of Pediatrics, Nagoya University Graduate School of Medicine, the Department of Pediatrics, Okazaki City Hospital, and the Department of Pediatrics, Anjo Kosei Hospital, Nagoya, Aichi, Japan. Submitted for publication Mar 29, 2002; revisions received Aug 13, 2002, and Feb 28, 2003; accepted Mar 13, 2003. Reprint requests: Akihisa Okumura, MD, Department of Pediatrics, Nagoya University, Graduate School of Medicine, 65 Tsurumai-cho, ShowaKu, Nagoya, Aichi, 466-8550, Japan. E-mail
[email protected]. Copyright Ó 2003, Mosby, Inc. All rights reserved. 0022-3476/2003/$30.00 + 0 10.1016/S0022-3476(03)00182-3
Fig 1. Samples of frontal and occipital sharp waves. EEG at day 6 (31 weeks’ postconceptional age) of infants with severe periventricular leukomalacia. Calibration: 100 lV, 1 second. a, Frontal sharp waves (arrows). b, Occipital sharp waves (arrows). Positive rolandic sharp wave is also observed (arrowhead ).
changes during the neonatal period. All infants had follow-up until 24 months of corrected age or older. The severity of PVL was classified into 3 groups according to that of diplegia at 2 years of corrected age for the following two reasons. First, diplegia in the infants was attributable to PVL, because MRI did not reveal other brain lesions in any infants. Second, the extent of white matter injury correlated well with the severity of diplegia in infants with PVL.7 Diplegia was judged as mild when an infant could walk with or without support, moderate when an infant could sit but could not stand, and severe when an infant could not sit. The severity of PVL was judged as mild in 10 infants, moderate in 11, and severe in 10. All 62 infants in the control group had normal development at 24 months of age or later.
EEG Recordings The EEGs were recorded polygraphically with a Nihon Kohden electroencephalograph, using a bipolar montage with 8 surface electrodes (AF3, AF4, C3, C4, O1, O2, T3, T4) according to the 10-20 international system combined with electro-oculographic, electrocardiographic, and respiratory movement.8-10 AF3 and AF4 were located halfway between Fp1 and F3, and Fp2 and F4, respectively. EEGs were performed at the infant’s bedside for >40 minutes during wakefulness and spontaneous sleep, which included all sleep states. EEGs recorded within 3 weeks of birth were investigated. The initial EEG was performed within 1 week of birth, generally within the first 3 days after birth. When abnormal findings were found in the initial EEG or any clinical deterioration was noted, additional EEG recordings were done. Abnormal Sharp Transients on Electroencephalograms in Preterm Infants With Periventricular Leukomalacia
Two kinds of abnormal sharp waves were investigated. To detect truly abnormal components, we adopted relatively high cutoff levels. Frontal sharp waves (FS) were defined as sharp transients of positive polarity in the frontal regions with an amplitude >100 lV (Fig 1). Occipital sharp waves (OS) were defined as those of negative polarity in the occipital regions with an amplitude >150 lV (Fig 1). We evaluated only those that were isolated and sharply differentiated from background activities or artifacts. We also evaluated PRS defined as sharp transients of positive polarity appearing in the rolandic regions with an amplitude >100 lV. All EEGs were evaluated by three neonatal neurologists who were unaware of ultrasonographic findings. FS, OS, and PRS were judged as present only when all of the 3 EEG analysts agreed, and their number per minute was calculated for each record. The results were divided into 4 groups according to the age when the EEG recordings were made; days 0 to 4, days 5 to 7, days 8 to 14, and days 15 to 21.
Neuroimaging Ultrasonography was performed at least twice within the first week of life and thereafter 1 to 3 times per week. Coronal and sagittal sections from anterior fontanel were examined. Persistent periventricular echodensity was judged as present when it was present for >2 weeks. Cystic changes were diagnosed when multiple cystic formations of >3 mm in diameter were observed. When an infant had spastic diplegia, MRI was performed during late infancy (between 10 and 18 months of CA). MRI showed ventriculomegaly with an irregular 27
margin accompanied by periventricular abnormal high intensity areas on T2-weighed images in all 31 infants of the PVL group. Other brain lesions were not seen in these infants. Infants with mild diplegia had localized volume loss of periventricular white matter around the trigonal zone to the occipital horns. Those with moderate diplegia showed loss of periventricular white matter extended to the bodies of lateral ventricles. Those with severe diplegia had diffuse loss of periventricular white matter extending to the frontal horns.
Statistical Analysis FS, OS and PRS events per minute were reported as median (range) and were analyzed by means of MannWhitney U test. Gestational age, birth weight, and the date of the first EEG were described as mean ± SD and analyzed by unpaired t test. P values were considered significant at < .05.
RESULTS Patient Characteristics Gestational age of the infants was 29.4 ± 2.0 weeks in PVL group and 29.6 ± 2.0 weeks in control group. The birth weight was 1358 ± 333 g in the PVL group and 1403 ± 324 g in the control group. The age of the first EEG was 3.7 ± 2.9 days in PVL group and 4.3 ± 2.6 days in the control group. Ultrasonography demonstrated cystic changes in bilateral deep white matter in 24 infants. The average date of the first appearance of cysts was 23 days (range, 10-42 days). Cystic changes were first observed within 14 days of age in 1 infant, between 15 and 21 days of age in 12, and later than 21 days of age in 11. Persistent periventricular echodensity without cystic changes was observed in the remaining 7 infants. A total of 126 EEGs were analyzed; 2 EEG recordings were obtained in 31 infants (22 of the PVL group and 9 of the control group) and 3 recordings in 1 infant with PVL. EEGs were performed in 75 infants (26 of PVL group and 49 of the control group) during days 0 to 4, 25 infants (12 of PVL group and 13 of control group) during days 5 to 7, 20 infants (14 of PVL group and 6 of control group) during days 8 to 14, and 6 infants (3 of PVL group and 3 of control group) during days 15 to 21.
Abnormal Sharp Waves The number of FS per minute in the PVL group was 0.07 (0-0.51) during days 0 to 4, 0.30 (0.03-0.77) during days 5 to 7, 0.13 (0-1.17) during days 8 to 14, and 0.35 (0-0.89) during days 15 to 21 (Fig 2). The number of FS in the control group was 0 (0-0.44) during days 0 to 4, 0 (0-0.15) during days 5 to 7, 0.05 (0-0.15) during days 8 to 14, and 0 (0-0.03) during days 15 to 21. The number of FS was significantly increased at days 0 to 4 and days 5 to 7 in the PVL group. The number of OS per minute in the PVL group was 0.05 (0-1.44) during days 0 to 4, 0.35 (0-2.10) during days 5 to 7, 0.49 (0-3.05) during days 8 to 14, and 0.12 (0-0.30) during 28 Okumura et al
Fig 2. Number of frontal, occipital, and positive rolandic sharp waves during each period. **P < .01.
days 15 to 21 (Fig 2). The number of OS in the control group was 0 (0-0.10) during days 0 to 4, 0 (0-0.24) during days 5 to 7, 0 (0-0.02) during days 8 to 14, and 0 (0-0) during days 15 to 21. The number of OS was significantly increased at days 0 to 4, days 5 to 7, and days 8 to 14 in the PVL group. The number of PRS per minute in the PVL group was 0 (0-2.59) during days 0 to 4, 0.40 (0-4.71) during days 5 to 7, 0 (0-3.26) during days 8 to 14, and 0 (0-1.32) during days 15 to 21 (Fig 2). PRS were not seen in any EEGs in control group. FS, OS or PRS were further analyzed when the number was >0.1 per minute on at least one EEG recording. In the PVL group, FS, OS, and PRS were present in 23, 25, and 14 infants, respectively (Table I). At least one kind of abnormal sharp waves was present in 27 (87%) of 31 infants in the PVL group, and no abnormal sharp waves were seen in 51 (82%) of 62 infants in the control group. FS and OS were observed even in some infants with mild PVL, whereas PRS were not seen in these infants. The Journal of Pediatrics July 2003
Table I. Relation between abnormal sharp transients and neurologic outcome FS waves
Control PVL Severe Moderate Mild Sensitivity Specificity PPV NPV
OS waves
PRS waves
FS (ÿ)
FS >0.1/min
OS (ÿ)
OS > 0.1/min
PRS (ÿ)
PRS >0.1/min
53 8 0 5 3
9 23 10 6 7
59 6 0 2 4
3 25 10 9 6
62 17 1 6 10
0 14 9 5 0
0.74 0.82 0.72 0.87
0.81 0.90 0.89 0.91
0.45 0.82 1.00 0.78
FS, Frontal sharp waves; OS, occipital sharp waves; PRS, positive rolandic sharp waves; PVL, periventricular leukomalacia; PPV, positive predictive value; NPV, negative predictive value. Sensitivity, specificity, PPV, and NPV referred to the detection of all grades of PVL.
Table II. Types of abnormal sharp transients and neurologic outcome
FS + OS + PRS FS + OS FS + PRS OS + PRS FS only OS only PRS only None
Severe PVL
Moderate PVL
Mild PVL
Control
9
3
0
0
1 0 0 0 0 0 0
3 0 1 0 2 1 1
6 0 0 1 0 0 3
1 0 0 8 2 0 51
FS, Frontal sharp waves; OS, occipital sharp waves; PRS, positive rolandic sharp waves; PVL, periventricular leukomalacia.
Consequently, sensitivity of FS or OS was relatively high, and higher than for PRS. Specificity, positive or negative predictive value was high for the three types of sharp waves. When all three types of abnormal sharp transients were recognized, all infants had moderate or severe PVL (Table II). Among 11 infants with FS plus OS all but 1 had PVL. However, the severity of PVL was mild in 6 of them. Two of 4 infants with OS only had moderate PVL, whereas PVL was not recognized in all but 1 infant with FS only. PVL was not seen in most of those without any types of abnormal sharp transients.
DISCUSSION We tried to delineate the presence of abnormal sharp transients other than PRS in this study. For this purpose, it is essential to distinguish abnormal sharp transients from physiologic background activities or artifacts. Abnormal sharp transients were judged as present only when all of the 3 EEG analysts agreed to detect definitely abnormal components and to avoid mistaking physiologic activities for abnormal ones. In Abnormal Sharp Transients on Electroencephalograms in Preterm Infants With Periventricular Leukomalacia
addition, we adopted relatively higher cutoff levels. As a result, the number per minute was small, but the differences between PVL and control groups were clear. Both FS and OS are considered to be truly abnormal components related to white matter injuries due to PVL. The recognition of FS or OS will be useful for the prediction of PVL. FS and OS were seen in some infants with mild PVL, in whom PRS were not detected. Previous studies demonstrated that PRS are useful markers of white matter injuries due to PVL or intraventricular hemorrhage.1-6,11-13 However, sensitivity of FS or OS was higher than that of PRS. Their specificity and positive or negative predictive values were almost equal to those of PRS. The presence of two or more types of abnormal sharp transients was closely correlated with poor prognosis. Moderate or severe PVL was observed in all infants who had the three types of abnormal sharp transients. FS plus PRS or FS plus OS were almost always associated with PVL, although it was less severe than those with the three types of abnormal sharp transients. In contrast, infants with FS or OS only had relatively favorable outcomes. The serial changes in the number of FS or OS were similar to those of PRS. The numbers of FS or OS were small during days 0 to 4, increased to days 8 to 14, and then decreased again. FS and OS should be incorporated into the interpretation of chronic stage EEG abnormalities, which appear in parallel with improvement of acute stage EEG abnormalities. OS were usually associated with poor outcome, even if other types of abnormal sharp transients were not present. Neurologic outcome was often good among infants with, FS and without other types of abnormal sharp transients. This suggests that OS are more pathologic findings than FS and will be more closely related to deep white matter injuries. Lesions in patients with mild PVL were usually located exclusively in the deep white matter around the trigonal zones, for example, in the parieto-occipital regions.7 Thus, OS were often observed among infants with mild PVL. In contrast, the 29
frontal regions were involved only in infants with more severe PVL. Such infants usually have white matter lesions also in the centro-parieto-occipital regions. This is consistent with the result that FS accompanied by OS and/or PRS were correlated with poor outcome but those without OS or PRS were not. FS without OS or PRS may be less pathologic among preterm infants. In conclusion, FS and OS were observed in infants with milder PVL than PRS and will be useful for the prediction of mild PVL. FS and OS appeared in parallel with improvement of acute-stage EEG abnormalities. The presence of two or more types of abnormal sharp waves was correlated with poor prognosis.
REFERENCES 1. Cukier F, Andre M, Monod N, Dreyfus-Brisac C. Apport de l’EEG au diagnostic des he´morragies intraventriculairs du pre´mature´. Revue d’Electroencephalographie et de Neurophysiologie Clinique 1972;2:318-22. 2. Marret S, Parain D, Jeannot E, Eurin D, Fessard C. Positive rolandic sharp waves in the EEG of the premature newborn: a five year prospective study. Arch Dis Child 1992;67:948-51. 3. Marret S, Parain D, Me´nard JF, Blanc T, Devaux AM, Ensel P, et al. Prognostic value of neonatal electroencephalography in premature newborns less than 33 weeks of gestational age. Electroencephalogr Clin Neurophysiol 1997;102:178-85. 4. Bejar R, Coen RW, Meritt TA, Vaucher Y, Trice J, Centeno R, et al. Focal necrosis of the white matter (periventricular leukomalacia): sonographic,
pathologic, and electroencephalographic features. Am J Neuroradiol 1986; 7:1073-80. 5. Novotny EJ, Tharp BR, Coen RW, Bejar R, Enzmann D, Vaucher YE. Positive rolandic sharp waves in the EEG of the premature infant. Neurology 1987;37:1481-6. 6. Okumura A, Hayakawa F, Kato T, Kuno K, Watanabe K. Positive rolandic sharp waves in preterm infants with periventricular leukomalacia: their relation to background electroencephalographic abnormalities. Neuropediatrics 1999;30:278-82. 7. Yokochi K, Aiba K, Horie M, Inukai K, Fujimoto S, Kodama M, et al. Magnetic resonance imaging in children with spastic diplegia: correlation with the severity of their motor and mental abnormality. Dev Med Child Neurol 1991;33:18-25. 8. Hayakawa F, Okumura A, Kato T, Kuno K, Watanabe K. Dysmature EEG pattern in EEGs of preterm infants with cognitive impairment: maturation arrest caused by prolonged mild CNS depression. Brain Dev 1997;19:122-5. 9. Watanabe K. The neonatal electroencephalogram and sleep-cycle patterns. In: Eyre JA, eidtor. The neurophysiological examination of the newborn infant. New York: Mac Keith Press; 1992. p. 11-47. 10. Watanabe K, Hayakawa F, Okumura A. Neonatal EEG: a powerful tool in the assessment of brain damage in preterm infants. Brain Dev 1999;21:361-72. 11. Aso K, Scher MS, Barmada MA. Neonatal electroencephalography and neuropathology. J Clin Neurophysiol 1989;6:103-23. 12. Baud O, d’Allest AM, Lacaze-Masmonteil T, Zupan V, Nedelcoux H, Boithias C, et al. The early diagnosis of periventricular leukomalacia in premature infants with positive rolandic sharp waves on serial electroencephalography. J Pediatr 1998;132:813-7. 13. Blume WT, Dreyfus-Brisac C. Positive rolandic sharp waves in the neonatal EEG. Electroencephalogr Clin Neurophysiol 1982;53:277-82.
FELLOWSHIPS Fellowships available in pediatric subspecialties and those for general academic pediatric training are listed once a year, in January, in The Journal of Pediatrics. Each June, forms for listing fellowships available for the academic year beginning 18 months after publication are sent to the Chairman of the Department of Pediatrics at major hospitals in the United States and Canada. In addition, a copy of the application form appears in the July, August, and September issues of The Journal (please use the current form). Should you desire to list fellowships, a separate application must be made each year for each position. All applications must be returned to Mosby, Inc., by October 15 preceding the listing year to ensure publication. Additional forms will be supplied on request from the Periodical Editing Department, Mosby, Inc, 11830 Westline Industrial Dr, St Louis, MO 63146-3318 (phone: 800-325-4177, ext 5928, or 314-579-5928).
30 Okumura et al
The Journal of Pediatrics July 2003
