Eur J Pediatr (2000) 159: 663±670
Ó Springer-Verlag 2000
ORIGINAL PAPER
Bernhard Resch á Esther Vollaard á Ute Maurer á Josef Haas Helfried Rosegger á Wilhelm MuÈller
Risk factors and determinants of neurodevelopmental outcome in cystic periventricular leucomalacia
Received: 30 September 1999 and in revised form: 28 February 2000 / Accepted: 17 March 2000
Abstract The aim of the study was to determine risk factors for the development of cystic periventricular leucomalacia (PVL) and to correlate ultrasound ®ndings with neurodevelopmental outcome. By means of a retrospective case-control study (matched for gestational age, birth weight, sex, and year of birth) and a cohort analysis of all preterm infants with cystic PVL documented by ultrasound scans hospitalised at a local tertiary care centre between 1988 and 1998, 98 preterm infants with a gestational age ranging from 26 to 35 weeks were diagnosed as having cystic PVL. The mean day of diagnosis of periventricular echodensities was 3 2 days (range 1±11 days), and of cystic PVL 21 8 days (range 2±47 days). Of 79 infants (1988±1997) eligible for neurodevelopmental follow-up (91%), hemi-, di-, or tetraplegia was diagnosed in 61 (77%), normal mental outcome in 22 (28%), associated visual disorders in 41 (52%) and seizure disorders in 12 (15%) infants. Signi®cant risk factors associated with the development of cystic PVL were premature rupture of membranes, chorioamnionitis, and hyperbilirubinaemia (odds ratios 4.665, 6.026, and 2.460 respectively). Subgroup analysis according to gestational age (26±28, 29±32, 33±35 weeks) revealed similar results despite spontaneous labour (26±28 weeks; odds ratio 4.808) and pre-eclampsia (33±35 weeks; odds ratio 3.517). Multiple pregnancy was associated with a twofold increased risk (odds ratio 2.075). The white matter damage probably accounted for the signi®cantly higher prevalence of apnoeas (P < 0.001) and neonatal seizures (P < 0.001). Cysts located bilateral or parieto-occipital were associated with a higher risk of cerebral palsy (odds ratios 6.933 and 4.327 respectively). Solely anterior located cysts were associated with normal neurological outcome. Increasing size of the cysts was associated with increasing risk of cerebral palsy with a cut-o value of 10 mm (odds ratio 3.300 and above) and all infants with cysts of more than 20 mm diameter had cerebral palsy. Conclusion The high prevalence of premature rupture of the membranes and chorioamnionitis further supports the role of intra-uterine infection in the pathogenesis of periventricular leucomalacia. The overall prognosis of cystic periventricular leucomalacia is poor. Key words Cystic periventricular leucomalacia á Prematurity á Risk factors á Ultrasonography á Neurodevelopmental outcome
B. Resch (&) á U. Maurer á H. Rosegger á W. MuÈller Division of Neonatology, Department of Paediatrics, University Hospital Graz, Auenbruggerplatz 30, 8036 Graz, Austria e-mail:
[email protected] Tel.: +43-316-3852605; Fax: +43-316-3852678
E. Vollaard University Hospital Groningen, Groningen, The Netherlands J. Haas Department of Gynaecology and Obstetrics, University Hospital, Graz, Austria
664
Abbreviations GA gestational age á PROM premature rupture of the membranes á PVE periventricular echodensities á PVL periventricular leucomalacia á US ultrasound Introduction
Periventricular leucomalacia (PVL) refers to necrosis of the white matter in a characteristic distribution, i.e. in the white matter dorsal and lateral to the external angles of the lateral ventricles, involving particularly the centrum semiovale (frontal horn and body) and the optic (trigone and occipital horn) and acoustic (temporal horn) branches. PVL is the most severe and frequent cause of cerebral palsy in children surviving preterm birth [38]. PVL has been known to pathologists for more than a century and was ®rst described by Virchow in 1867 [36]. Banker and Larroche [3] introduced the name ``periventricular leucomalacia'' in 1962 to describe their observation of the periventricular ``white spots'' seen macroscopically and the softening (malacia) of the white matter (leukos). In 1982, Hill et al. [22] ®rst diagnosed cystic PVL by ultrasonography. The pathogenesis of PVL is not completely understood. The majority of the theories consider the necrotic foci to be hypoxic-ischaemic lesions resulting from impaired perfusion at the vascular border zones between the ventriculopedal and ventriculofugal arteries as the latter are poorly developed in preterm infants. These periventricular zones would be considered to be most susceptible to a decrease in perfusion pressure and cerebral blood ¯ow. A contrasting view focuses on the role of intra-uterine infection. Stimulation of circulating fetal mononuclear cells by microbial products would lead to cytokine production (interleukines 1b, -6 and tumour necrosis factor-a), which increase the permeability of the blood-brain barrier and allow passage into the central nervous system. These cytokines may stimulate fetal microglia to further cytokine production, may have a direct toxic eect on oligodendrocytes, and may provide a mitogenic stimulus for astrocytes and be responsible for astrocyte hypertrophy, which is considered to be an early pre-necrotic marker of white matter damage. Microbial invasion of the amniotic cavity between 28 and 32 weeks of gestation would lead to injury at the most vulnerable period of myelogenesis [9, 37]. The aim of the study was (1) to analyse risk factors by means of a retrospective case-control study and (2) to correlate ultrasound (US) scan ®ndings with neurodevelopmental outcome. Patients and methods This study is a retrospective analysis of all infants with cystic PVL documented by US admitted to the Division of Peripartal Paediatrics and Neonatology of the Department of Paediatrics of the University Hospital of Graz, Austria, between 1988 and 1998. The
medical charts, the US scans and the data from our neurodevelopmental follow-up outpatient clinic were reviewed. Risk factors For the analysis of risk factors, prenatal characteristics including maternal age, number of pregnancies, multiple pregnancy, history of abortion, maternal haemorrhage, pre-eclampsia/eclampsia, premature rupture of the membranes (PROM), chorioamnionitis, abnormal prepartal cardiotocography, breech presentation and caesarean section, and post-natal characteristics including asphyxia, bacterial infections, arterial hypotension, hyperbilirubinaemia, hypoxic-ischaemic encephalopathy, intra-/periventricular haemorrhage, post-haemorrhagic hydrocephalus, persistent ductus arteriosus and respiratory distress syndrome were compared with controls matched for gestational age (GA), birth weight, sex, and year of birth. Additionally, subgroup analyses according to GA ± 26 to 28, 29 to 32, and 33 to 35 weeks ± were performed. Because of the heterogeneity of clinical features of hypoxic-ischaemic encephalopathy, fetal distress, neurological abnormalities within 24 h of age, and Apgar score 5 or less after 5 min con®rmed the diagnosis. Respiratory distress syndrome was de®ned as dyspnoea presenting within 4 to 6 h of delivery, additional oxygen requirement to prevent cyanosis, and reticulogranular chest X-ray appearance. Arterial hypotension was de®ned as a mean arterial blood pressure, measured with Dinamap, below 95% limits [26] and requiring treatment. Neonatal seizures were de®ned as paroxysmal alterations in neurological function, either subtle, tonic, clonic, or myoclonic, with autonomic nervous system changes and dierent from the symmetrical tremor of jitteriness. US scan Cranial US scans were routinely obtained in all preterm infants on days 1, 3, 5, and thereafter once a week in case of pathological ®ndings. Real-time US scans were performed with a commercially available unit (Advanced Technology Laboratories Inc., Bothell, WA, USA) using a 7.5 MHz transducer and multiple images were obtained in the coronal and sagittal planes through the anterior fontanelle. The US scans were reviewed for the day of diagnosis of periventricular echodensities (PVE) and cystic PVL, the site of the cysts and the maximum diameter of the largest cysts. PVE was de®ned as con¯uent areas of increased echogenicity comparable with the echogenicity of the choroid plexus. Con®rmation of the ®ndings in both the coronal and sagittal planes was required before a de®nitive diagnosis of PVE could be made. The site of the cysts was described in terms of its being anterior, anterior to the frontal horn of the lateral ventricle, parietal, lateral to the body of the lateral ventricle, or occipital, adjacent and lateral to the occipital horn of the lateral ventricle [16]. The maximum diameters of the largest cysts were measured in both the coronal and sagittal planes and the maximum value was noted. Neurodevelopmental outcome For neurodevelopmental outcome, infants were examined at the corrected for prematurity age of 4, 8, 12, 18 and 24 months, thereafter once a year. Assessment of outcome was made using the developmental tests as described by Grith [20] in the ®rst 2 years, by Kaufman and Kaufman [25] after these years, and neurological examinations as described by Amiel-Tison and Stewart [2] and Touwen [32]. Classi®cation of mental outcome included normal, developmental delay, and mild to severe mental retardation. Classi®cation of neurological outcome included normal, minor
665 neurological abnormalities (including dystonia, hypotonia and asymmetry), hemiplegia (one side of body, arm more than leg), diplegia (minimal upper extremity involvement, legs more impaired) and tetraplegia (severe involvement of all extremities, legs more than arms). Classi®cation of visual disorders included normal, strabismus, myopia, severe visual impairment and blindness. Classi®cation of hearing disorders included hearing de®ciency and deafness. Other ®ndings included microcephaly and dystrophy. Neurodevelopmental outcome of the infants with cystic PVL was correlated with the site of the cysts (anterior, parietal, occipital), size of the cysts, and the ®ndings of cysts located unilaterally or bilaterally as diagnosed by US. Cerebral palsy included hemi-, di-, and/or tetraplegia. The outcome of the infants also was correlated with the presence of neonatal seizures. Mental outcome, visual disorders, seizure disorders, microcephalus and dystrophy were again correlated with the US ®ndings. Statistics Statistical analyses were done with the t-test and Wilcoxon test for numerical data after checking the normality assumption with the Kolmogorov-Smirnov-test. Categorial data were tested with Chisquare using Yates correction and Fisher's exact test as appropriate. Multivariate analysis was performed with a logistic regression model. Statistical analysis was done with a SAS package (SAS Institute Inc. Cary, N.C.), SPSS (SPSS Inc. Chicago, Ill.) and StatXact4 and LogXact (Cytel, Cambridge, Mass.).
Results
Population characteristics Out of 3187 infants with a GA £ 35 weeks who had been hospitalised at both neonatal divisions between 1988 and 1998, 98 were diagnosed as having cystic PVL. Their perinatal data are shown in Table 1 together with the matched case-controls. Dierences were found between use of maternal steroids and antibiotics, Apgar score at 5 min, umbilical cord artery pH, apnoeas and neonatal seizures. Pneumothorax was observed in one (1%) infant of the PVL group and in three (3%) infants of the control group. One patient of the control group had a fatal course of mechanical ventilation over 153 days because of severe interstitial lung disease. Excluding this patient, results of duration of mechanical ventilation Table 1 Perinatal characteristics of 98 infants with cystic PVL and 98 case-controls matched for year of birth, GA, birth weight, and gender. Data are presented as n (%) or mean SD (range)
were not signi®cant between both groups. The incidence of cystic PVL was 3.1%, varying between 1.0% and 5.8% per year. In all infants with a GA £ 32 weeks, the incidence of cystic PVL was 4.1%, varying between 1.2% and 6.5% per year. Risk factor analysis The prenatal risk factors are shown in Table 2, the postnatal risk factors in Table 3. There were signi®cant dierences regarding PROM, chorioamnionitis, and hyperbilirubinaemia, with associated higher risks of cystic PVL (odds ratios 4.665, 6.026 and 2.460 respectively), and a twofold increased risk due to multiple pregnancy. Subgroup analyses according to GA revealed similar results: odds ratios, listed as above, for PROM were 3.250, 4.242, and 7.714 respectively, for chorioamnionitis 6.314, 5.782, and 13.304 respectively, and for hyperbilirubinaemia 2.139, 3.690 and 1.691. Spontaneous labour was associated with a higher risk of cystic PVL in infants of 26 to 28 weeks GA (odds ratio 4.808), as was multiple pregnancy (odds ratio 7.000). Preeclampsia/eclampsia was associated with a higher risk of cystic PVL in infants of 33 to 35 weeks GA (odds ratio 3.517). All the other characteristics did not dier signi®cantly between the subgroups. Cranial US scans Because of poor quality of documented US scans in four patients, only the scans of 94 (96%) infants could be evaluated in detail. The mean day of diagnosis of PVE was 3 2 days (range 1±11 days), and of cystic PVL 21 8 days (range 2±47 days). A total of 77 (82%) infants had cystic PVL involving both the right and left cerebral hemispheres. Unilateral cystic PVL was present in 17 (18%) infants, involving in 53% the right and in 47% the left hemisphere. In 63 (67%) infants the cysts, independent of uni- or bilateral cystic PVL, were located anterior-parietal-occipital or parietal-occipital. All
Characteristic
PVL group
GA (weeks) Birth weight (g) Male gender Maternal steroids Maternal antibiotics Small for dates Apgar score 1 min Apgar score 5 min Apgar score 10 min Umbilical cord artery pH Capillary pHa Apnoeas Neonatal seizures Days of mechanical ventilation
31 2.5 (26±35) 31 2.5 (26±35) 1511 456 (670±2515) 1499 446 (750±2570) 65 (66%) 63 (64%) 34 (35%) 20 (20%) 23 (23%) 11 (11%) 11 (11%) 11 (11%) 6.8 2.0 (1±9) 6.2 2.5 (0±9) 8.9 1.1 (6±10) 8.4 1.6 (1±10) 9.2 1.0 (5±10) 9.0 1.1 (4±10) 7.29 0.09 (7.01±7.45) 7.26 0.1 (7.01±7.46) 7.20 0.1 (7.03±7.43) 7.19 0.1 (7.01±7.46) 68 (69%) 30 (31%) 31 (32%) 3 (3%) 9 11 (0±48) 12 18 (0±153)
a
Within 30 min after birth
Case-controls
P 0.948 0.997 0.765 0.013 0.012 1.000 0.105 0.009 0.272 0.037 0.942
