Cerebral Palsy Among Very Preterm Children in Relation to Gestational Age and Neonatal Ultrasound Abnormalities: The EPIPAGE Cohort Study
OBJECTIVE. To estimate the prevalence of cerebral palsy at 2 years of age among children born very preterm, according to gestational age, infant gender, plurality, and neonatal cranial ultrasound abnormalities.
METHODS. All infants born between 22 and 32 weeks of gestation in 9 regions of France in 1997 were included in this prospective, population-based, cohort study. The main outcome measure was cerebral palsy prevalence at 2 years. Of the 2364 survivors eligible for follow-up evaluation, 1954 (83%) were assessed at 2 years of age.
RESULTS. Among the 1954 children assessed at 2 years, 8.2% had cerebral palsy. Bilateral spastic cerebral palsy, hemiplegia, and monoplegia accounted for 72%, 9%, and 10% of cases, respectively. Fifty percent of the children with cerebral palsy walked independently at the age of 2, 31% were unable to walk but could sit independently, and 19% could not sit (unable to maintain head and trunk control). The prevalence of cerebral palsy was 20% at 24 to 26 weeks of gestation, compared with 4% at 32 weeks. On the basis of ultrasound findings in the neonatal period, we found that 17% of children with isolated grade III intraventricular hemorrhage and 25% of children with white matter damage (ie, ventricular dilation, persistent echodensities, or cystic periventricular leukomalacia) had cerebral palsy, compared with 4% of children with normal ultrasound scans.
CONCLUSIONS. Despite recent improvements in survival rates, cerebral palsy remains highly prevalent among very preterm children. Severe cranial ultrasound abnormalities predict motor disability strongly, but one third of infants with cerebral palsy had no ultrasound abnormalities.
- population-based study
- very preterm infants
- white matter damage
- intraventricular hemorrhage
- prevalence of cerebral palsy
Cerebral palsy is one of the most common and severe sequelae of very preterm birth. Because of improvements in survival rates, very preterm children now account for >25% of all cases of cerebral palsy.1 However, there is no evidence of a decrease in the prevalence of cerebral palsy among very preterm children.1 Since the early 1990s, preterm infants have benefited greatly from innovations in perinatal care, including prenatal steroids, exogenous surfactants, and new ventilation techniques; the long-term effects of these innovations are still a matter of debate. Although registries of childhood impairments and large population-based studies conducted in Europe,2 Western Australia3 and North America4,5 monitor trends in rates of cerebral palsy, few give gestational age-specific rates. Given the limited available information and recent changes in perinatal care, cerebral palsy is still an important issue for very preterm children.
Since the early 1990s, the survival rates of very preterm infants have increased as a result of innovations, ie, prenatal steroids, exogenous surfactants, and in utero transfer. However, the effects of these innovations on later disability are more controversial. We report the prevalence of cerebral palsy among 2-year-old, very preterm children born in 1997, in a large, geographically defined population in France. We studied the prevalence of cerebral palsy as a function of gestational age at birth, infant gender, plurality, and ultrasound abnormalities identified in the neonatal period.
All births that occurred between 22 and 32 completed weeks of gestation in all maternity units in 9 French regions in 1997 were included in the EPIPAGE study.6 Of the 2901 children born alive, 127 died immediately after birth in the delivery room, 2774 were transferred to a neonatal care unit, and 2459 were discharged alive. All survivors born before 32 weeks were enrolled in a longitudinal follow-up study. Each region was given the option of including 1 of every 2 infants born at 32 weeks in the follow-up study, to reduce their workload (given the large number of children in this group). In 2 regions (Paris area and Haute-Normandie), one half of the infants born at 32 weeks were excluded randomly from the follow-up study (n = 77). Therefore, the initial population enrolled in the follow-up study included 2382 infants. Eighteen died after discharge from the hospital but before the age of 2 years. The parents of 106 (4%) of the 2364 survivors eligible for the follow-up study refused to participate. At 2 years of age, the treating physician (pediatrician, neonatologist, or general practitioner) who examined the infant was asked to complete a standardized questionnaire. This questionnaire was completed for 1960 children (83% of the eligible children). The neurologic status of 4 children who still required artificial respiration at 2 years was not known; therefore, these children were excluded. Two cases of cerebral palsy caused by physical violence were excluded. Therefore, a total of 1954 children were included in this analysis.
Measures and Definition of Cerebral Palsy at 2 Years
Each child was subjected to a detailed physical and neurologic examination assessing tone, reflexes, posture, and movements. A precoded standardized questionnaire, completed by each treating physician, was designed to minimize the risk of ambiguous answers, and trained pediatricians reviewed questionnaires for infants with abnormal neurologic examination results. We used the definition of cerebral palsy proposed by the European Cerebral Palsy Network.2 Children were classified as having spastic cerebral palsy if they had ≥2 of the following criteria: abnormal posture or movement, increased tone, or hyperreflexia. Children with involuntary movements were classified as having dyskinetic cerebral palsy, and those with loss of coordination were classified as having ataxic cerebral palsy.2 Subtypes of cerebral palsy were classified into 4 categories, as follows. Group 1 represented hemiplegia (ie, a unilateral neurologic abnormality), and group 2 represented bilateral spastic cerebral palsy, including children with diplegia (only the lower limbs affected) and quadriplegia (all 4 limbs affected). Children with monoplegia and children with ataxic or dyskinetic cerebral palsy were placed in groups 3 and 4, respectively. There were 4 children diagnosed as having cerebral palsy for whom the subtype was not described. Three groups were defined according to the severity of handicap, as follows: group 1, children who could walk independently; group 2, children who could not walk but who could sit independently; group 3, children who could not sit independently (unable to maintain head and trunk control).
Gestational age refers to completed weeks of amenorrhea and was the best obstetric estimate based on the date of the last menstrual period and an early prenatal ultrasound scan, which is routine practice in France. Neonatal outcomes were determined by consulting medical records. In France, cranial ultrasonography is performed routinely for very preterm children admitted to NICUs. Each infant usually undergoes 1 to 3 cranial ultrasound examinations in the first 2 weeks of life. Thereafter, a weekly examination is advised for infants with cerebral abnormalities and less-frequent examinations (every 2 weeks) are recommended for infants without such abnormalities.7 Qualified neonatologists or radiologists performed cranial ultrasonography at all participating centers. Intraventricular hemorrhage (IVH), periventricular leukomalacia (PVL), and ventricular dilation were diagnosed on the basis of cranial ultrasound images. Infants were classified as having grade I IVH if any ultrasound examination showed subependymal hemorrhage, grade II IVH if an ultrasound examination showed IVH without ventricular dilation, and grade III IVH if an ultrasound examination showed IVH with primary ventricular dilation. Intraparenchymal hemorrhage (IPH) referred to a large, unilateral, parenchymal hyperdensity or a large, unilateral, porencephalic cyst, possibly attributable to venous hemorrhagic infarction.8 PVL was defined as the presence of periventricular white matter echolucencies (cystic PVL) or echodensities persisting for >14 days without cyst formation. Ventricular dilation referred to isolated dilation of ventricles with no associated IVH. In a derived variable describing all cranial ultrasound results, infants with normal ultrasound results were distinguished from those with isolated IVH without intraparenchymal involvement (ie, grade I–II [class 1] and grade III [class 2] IVH) and from those with white matter abnormalities. The latter group included infants with persisting echodensities and ventricular dilation (class 3), unilateral cystic PVL (class 4), bilateral cystic PVL (class 5), and IPH (class 6), which are considered different forms of white matter disease.9,10
Results are presented as proportions of the number of survivors assessed at 2 years of age. The statistical analyses were weighted to take into account the differences in the proportions of children who underwent follow-up monitoring according to gestational age and region. Only weighted percentages are shown in the tables. The Pearson χ2 statistic was corrected for the survey design by using the correction described by Rao and Scott11 and was converted into an F statistic. Statistical analyses were performed with Stata software (Stata Corp, College Station, TX). The study was approved by the Commission Nationale de l'Informatique et des Libertés.
Infants who were assessed at 2 years of age (responders) had a younger gestational age at birth than did the other children (nonresponders) (Table 1). There was no difference between the 2 groups regarding cranial ultrasound abnormalities. In contrast, we observed significant socioeconomic differences between the 2 groups; young mothers, single mothers, and mothers with low educational levels were less likely to participate in the follow-up study than were other mothers (Table 1).
Of the 1954 survivors assessed, 165 had cerebral palsy (8.2%; 95% confidence interval: 7.0–9.4%). Spastic cerebral palsy and ataxia and/or dyskinetic cerebral palsy accounted for 91% and 7%, respectively, of all cases. Bilateral spastic cerebral palsy was present in 72% (diplegia: 44%; quadriplegia: 28%), hemiplegia in 9%, and monoplegia in 10%. The subtype was not recorded in 4 cases. Fifty percent of the children with cerebral palsy walked independently at the age of 2 years, 31% were unable to walk but could sit independently, and 19% could not sit (unable to maintain head and trunk control). Among the 1789 children without cerebral palsy, 1.4% could not walk at 2 years of age.
The prevalence of cerebral palsy decreased with increasing gestational age (Table 2), ie, ∼20% at <27 weeks, 12% at 27 to 28 weeks, 8% at 29 to 30 weeks, 7% at 31 weeks, and 4% at 32 weeks. Cerebral palsy was slightly but not significantly more frequent among boys (9%) than girls (7%). There was no difference in cerebral palsy rates according to plurality.
Neonatal cranial ultrasound abnormalities predicted cerebral palsy to a certain extent (Table 3); 28% and 60% of children with grade III IVH and IPH, respectively, developed cerebral palsy at 2 years, compared with 7% of children without IVH. Cerebral palsy rates were 75% among children with bilateral cystic PVL, 35% among children with unilateral cystic PVL, and 17% among children with persistent echodensities, compared with 5% in the absence of PVL. Two thirds of children with PVL in the parietal or occipital lobes were later diagnosed as having cerebral palsy. No abnormalities were identified in cranial ultrasound scans for 1238 children; however, in this group, 4.4% of children developed cerebral palsy (Table 3).
When cerebral palsy rates were compared for the 4 gestational age groups, there was a statistically significant difference in the rates of cerebral palsy among children with no ultrasound abnormalities; the more immature the infant, the greater was the likelihood of cerebral palsy (Table 4). Among children with isolated IVH, persistent echodensities, or ventricular dilation, only the most immature had an increased risk of cerebral palsy. There were no statistically significant differences according to gestational age among infants with cystic PVL or IPH.
The 56 children who had cerebral palsy and normal ultrasound findings accounted for 35% of the children with cerebral palsy (Table 5). Children with isolated IVH and those with white matter disease accounted for 14% and 52%, respectively, of cerebral palsy cases. The distribution of ultrasound results varied with gestational age. Only 17% of children born at <27 weeks who developed cerebral palsy had no ultrasound abnormalities, compared with 31% at 27 to 28 weeks, 33% at 29 to 30 weeks, and 48% at 31 to 32 weeks (Table 5). Conversely, the percentages of isolated IVH decreased with gestational age, but there was little variation in the percentages of white matter abnormalities. However, subtypes of white matter abnormalities seemed to vary with gestational age; persistent echodensities and ventricular dilation were more common at 24 to 26 weeks than after that age.
There were no statistically significant differences in the distributions of cerebral palsy subtypes among children with normal ultrasound findings, isolated IVH, or white matter abnormalities (Table 6). Although bilateral spastic diplegia was more common among children with bilateral cystic PVL (88%) than among children with unilateral cystic PVL or IPH, numbers were very small.
Among infants with cerebral palsy, 35% of infants with white matter abnormalities could walk and 31% could not sit, compared with 13% and 63%, respectively, of infants with bilateral cystic PVL (Table 7). In comparison, 55% of infants with isolated IVH and 71% of infants without cranial ultrasound abnormalities could walk independently.
In this study, we estimated the risk of cerebral palsy for a large sample of very preterm children born between 22 and 32 weeks in 1997. The population of the EPIPAGE study was defined geographically, which eliminated the referral bias inherent in studies recruiting only from selected perinatal centers. Obstetricians estimated gestational ages according to the best available information (ultrasound scan results and date of last menstrual period). Because an early scan is standard practice for almost all pregnant women in France,12 we can assume that this estimation of gestational age is reliable and of good quality. This is an important point because recruitment based on birth weight leads to overrepresentation of more-mature children with restricted growth.13 Moreover, gestational age is a better predictor than birth weight of mortality and morbidity rates.14
Loss to follow-up monitoring is an important issue in longitudinal studies, because it can bias the results. We had information for 83% of the eligible survivors. Tin et al15 showed that children monitored with difficulty are more likely to have severe disability than those monitored without difficulty. In our study, children lost to follow-up monitoring had an older gestational age at birth but did not differ from the group not lost to follow-up monitoring with respect to neonatal cerebral lesions. This suggests that they were not at higher risk of neurologic disabilities than monitored children.
It is difficult to compare the rates of cerebral palsy between studies because the rates vary according to a number of factors, including definition, age and method of ascertainment, and year of birth. To facilitate comparisons, we used the criteria recommended by the Surveillance of Cerebral Palsy in Europe network2 to diagnose cerebral palsy. However, because our study took place in 9 regions of France, many physicians were involved and the ability to diagnose cerebral palsy might have varied. Furthermore, the children were assessed at 2 years, whereas assessment at 5 years would probably result in more accurate measures of final neurologic outcomes.2 Although this might have led to an underestimation of mild nondisabling cases of cerebral palsy, the Victorian Infant Collaborative Study Group16 showed that assessments of young children might be too pessimistic. However, changes in neurologic classification with age occur infrequently.17
The EPIPAGE study provides the opportunity to investigate the neuromotor outcomes of infants born very or extremely preterm who have benefited from prenatal corticosteroids, exogenous surfactant, and in utero transfer.18,19 Since the Bavarian Longitudinal Study20 and the Project On Preterm and Small for Gestational Age Infants in the Netherlands21 were performed in the 1980s, no equivalent, large, population-based studies have examined the neurodevelopmental outcomes of this population.
In our study, 20% of infants born before 27 weeks developed cerebral palsy, accounted for 18% of very preterm children with cerebral palsy. The EPICure population-based study showed that 18% of the 283 survivors born before 26 weeks had cerebral palsy at 30 months.22 Estimates of the prevalence among children born before 27 weeks vary from 11% to 35%.23,24 Differences in sample size (range: 23–283 children), age at assessment (2–5 years), and distribution of gestational ages make it difficult to compare rates between studies or to estimate precisely the cerebral palsy rate at the lowest gestational ages. An important finding of our study is that cerebral palsy rates remain high among children born after 27 weeks. On the basis of data from European registries, the gestational age-specific rate of cerebral palsy was 5% to 6% for children born at 28 to 31 weeks.1,3,25 Because these children account for more than two thirds of cerebral palsy cases among very preterm children, additional studies are needed regarding this group.
It is likely that different subtypes of cerebral palsy have different causes.26 On the basis of MRI findings, Colver and Sethumadhavan27 argued that cerebral injuries responsible for quadriplegic, triplegic, and diplegic syndromes are very similar and should be grouped together, defining bilateral spastic cerebral palsy. According to this classification, 72% of the children with cerebral palsy in our population had a bilateral spastic form and 9% had hemiplegia. These values are very similar to the percentages (75–83% and 9–13%, respectively) estimated for other preterm populations.1,22,28
In our study, we evaluated the severity of the handicap according to the ability to walk and sit independently; 50% of infants with cerebral palsy were unable to walk at the age of 2 years and 40% of these children were severely disabled (unable to maintain head and trunk control). According to the literature, the rates of moderate/severe motor disability vary from 10% to 50%,22,28,29 and there is no evidence that the functional severity of cerebral palsy is declining among very preterm or very low birth weight children.28,29
Our study showed how the risk of cerebral palsy varies according to the presence and type of cranial ultrasound abnormalities diagnosed in the neonatal period. In a review of 15 studies, Holling and Leviton30 showed that 59% of children with echolucencies developed cerebral palsy subsequently. Our results are particularly interesting because they are population-based. Of the 76 children with cystic PVL, 44 (58%) developed cerebral palsy, which is very similar to the estimate by Holling and Leviton.30 As expected, the risk of cerebral palsy was higher when cystic PVL was bilateral and localized in the parietal and occipital lobes.30
We found that the increased risk of cerebral palsy with decreasing gestational age was partly attributable to an excess of cerebral abnormalities among the most immature infants (Table 4). In one recent study, neurodevelopmental outcomes were independent of gestational age when each lesion was considered separately.31 In our study, cerebral palsy rates increased with decreasing gestational age, even among children with no ultrasound abnormalities and among those with isolated IVH, echodensities, or ventricular dilation. Although it could be speculated that gestational age in itself has an independent effect on neurodevelopmental outcomes, deficiencies of ultrasonography in detecting subtle white matter pathologic conditions should be considered.
Almost 4% of children with normal ultrasound findings developed cerebral palsy, accounting for one third of all cases of cerebral palsy (nearly 50% of cerebral palsy cases at 31–32 weeks). However, those children were less disabled than infants with ultrasound abnormalities, which suggests that they suffer from subtle and less extensive cerebral lesions. One explanation for the low sensitivity of ultrasonography at the oldest gestational ages may be that such children undergo fewer scans than children born at younger gestational ages.6 Difficulties in the diagnosis of cerebral lesions are assumed to be attributable to the limitations of the technique itself. However, the quality of scanning and scan interpretation are possible components of this problem. In a recent study, only 59% of doctors interpreted correctly high-resolution scanned images of 6 major neonatal ultrasound abnormalities.32 As suggested by De Vries et al,33 very preterm infants need to undergo scanning until discharge and the equivalent of term age. In their study, 30% of children who developed cerebral palsy after major ultrasound abnormalities would have not been diagnosed if ultrasound scans had been restricted to the first 4 weeks after birth.
Our large population-based study of very preterm children born in 1997 showed that 1 of every 12 very preterm children had cerebral palsy; this rate is at least 80 times higher than that among term children. As expected, the risk of cerebral palsy decreased with increasing gestational age at birth, although 7% of children born at 28 to 32 weeks were affected. Paradoxically, very few studies have investigated this population, although these children represent 85% of surviving very preterm infants6 and two thirds of very preterm children with cerebral palsy. Lastly, it would be of great interest to investigate the impact of neonatal injuries on sensorineural, cognitive, and fine motor functions. Data collected at 5 years of age should make it possible to answer these questions.
Funding for this study was obtained from INSERM (French National Institute of Health and Medical Research), Merck-Sharp, Dohme-Chibret, la Fondation de la Recherche Médicale (Medical Research Foundation), and la Direction Générale de la Santé du Ministère des Affaires Sociales (Directorate General for Health of the French Ministry for Social Affairs).
The EPIPAGE Study Group was as follows: INSERM U149: B. Larroque (national coordinator), P. Y. Ancel, B. Blondel, G. Bréart, M. Dehan, M. Garel, M. Kaminski, F. Maillard, C. du Mazaubrun, P. Missy, F. Sehili, K. Supernant; Alsace: M. Durant, J. Matis, J. Messer, A. Treisser (Hôpital de Hautepierre, Strasbourg); Franche-Comté: A. Burguet, L. Abraham-Lerat, A. Menget, P. Roth, J-P. Schaal, G. Thiriez (CHU St Jacques, Besançon); Haute-Normandie: C. Lévêque, S. Marret, L. Marpeau (Hôpital Charles Nicolle, Rouen); Languedoc-Roussillon: P. Boulot, J-C. Picaud (Hôpital Arnaud de Villeneuve, Montpellier), A-M. Donadio, B. Ledésert (ORS Montpellier); Lorraine: M. André, J-L. Boutroy, J. Fresson, J. M. Hascoët (Maternité Régionale, Nancy); Midi-Pyrénées: C. Arnaud, S. Bourdet-Loubère, H. Grandjean (INSERM U558, Toulouse), M. Rolland (Hôpital des Enfants, Toulouse); Nord-Pas-de-Calais: C. Leignel, P. Lequien, V. Pierrat, F. Puech, D. Subtil, P. Truffert (Hôpital Jeanne de Flandre, Lille); Pays de la Loire: G. Boog, V. Rouger-Bureau, J-C. Rozé (Hôpital Mère-Enfants, Nantes); Paris-Petite-Couronne: P-Y. Ancel, G. Bréart, M. Kaminski, C. du Mazaubrun (INSERM U149, Paris), M. Dehan, V. Zupan (Hôpital Antoine Béclère, Clamart), M. Vodovar, M. Voyer (Institut de Puériculture, Paris).
- Accepted July 19, 2005.
- Address correspondence to Pierre-Yves Ancel, MD, PhD, Epidemiological Research Unit on Perinatal and Women's Health, INSERM U149, 16 Ave Paul Vaillant-Couturier, 94807 Villejuif Cedex, France. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵Stanley F, Blair E, Alberman E. How common are the cerebral palsies? In: Bax MCO, Hart HM, eds. Cerebral Palsies: Epidemiology and Causal Pathways London, United Kingdom: Cambridge University Press; 2000:22– 39
- ↵O'Shea T, Preisser JS, Klinepeter KL, Dillard RG. Trends in mortality and cerebral palsy in a geographically based cohort of very low birth weight neonates born between 1982 and 1994. Pediatrics.1998;101 :642– 647
- ↵Winter S, Autry A, Boyle C, Yeargin-Allsopp M. Trend in the prevalence of cerebral palsy in a population-based study. Pediatrics.2002;110 :1220– 1225
- ↵Larroque B, Bréart G, Kaminski M, et al. Survival of very preterm infants: EPIPAGE, a population based study. Arch Dis Child Fetal Neonatal Ed.2004;89 :F139– F144
- ↵Voyer M. Bilan des premiers jours et surveillance du préterme. In: Encyclopédia Médico-Chirurgicale, ed. Prématurité Paris, France: Elsevier; 1998:101– 141
- ↵Arnold CC, Kramer MS, Hobbs CA, McLean FH, Usher RH. Very low birth weight: a problematic cohort for epidemiologic studies of very small or immature neonates. Am J Epidemiol.1991;134 :604– 613
- ↵Tin W, Wariyar U, Hey E, Northern Neonatal Network. Changing prognosis for babies of less than 28 weeks' gestation in the north of England between 1983 and 1994. BMJ.1997;314 :107– 111
- ↵Tin W, Fritz S, Wariyar U, Hey E. Outcome of very preterm birth: children reviewed with ease at 2 years differ from those followed up with difficulty. Arch Dis Child Fetal Neonatal Ed.1998;79 :F83– F87
- ↵Victorian Infant Collaborative Study Group. Neurosensory outcome at 5 years and extremely low birthweight. Arch Dis Child Fetal Neonatal Ed.1995;73 :F143– F146
- ↵Fresson J, Blondel B, Truffert P; EPIPAGE Group. Régionalisation des soins pour les enfants de moins de 33 semaines d'aménorrhée. In: Collet M, Treisser A, eds. Journées Nationales de Médecine Périnatale Paris, France: Arnette; 2001:207– 218
- ↵Rijken M, Stoelhorst MSJ, Martens SE, et al. Mortality and neurologic, mental, and psychomotor development at two years in infants born less than 27 weeks' gestation: the Leiden follow-up project on prematurity. Pediatrics.2003;112 :351– 358
- ↵Doyle LW. Outcomes at 5 years of age of children 23 to 27 weeks' gestation: refining the prognosis. Pediatrics.2001;108 :134– 141
- ↵Colver AF, Sethumadhavan T. The term diplegia should be abandoned. Arch Dis Child.2003;88 :286– 290
- ↵Pharoah PO, Platt MJ, Cooke T. The changing epidemiology of cerebral palsy. Arch Dis Child Fetal Neonatal Ed.1996;75 :F169– F173
- ↵Vollmer B, Roth S, Baudin J, Stewart AL, Neville BGR, Wyatt JS. Predictors of long-term outcome in very preterm infants: gestational age versus neonatal cranial ultrasound. Pediatrics.2003;112 :1108– 1114
- ↵Reynolds PR, Dale RC, Cowan FM. Neonatal cranial ultrasound interpretation: a clinical audit. Arch Dis Child Fetal Neonatal Ed.2001;84 :F92– F95
- Copyright © 2006 by the American Academy of Pediatrics