Two thirds of cerebral palsy arises in the 97% of singletons born at or after 35 weeks of gestation.1 The prevalence of cerebral palsy in these relatively mature neonates, unlike that of survivors of very preterm birth, has not fallen in recent decades.1,2 Our knowledge of brain lesions in cerebral palsy has improved with advances in neuroimaging, but the etiology and prognostic value of these lesions remain imperfectly understood.3 Historically, research on the etiologies of cerebral palsy in term and late preterm births has focused on asphyxial birth events. More recently, the diversity of cerebral palsy etiology has been explored with studies examining antenatal factors including inflammation, suboptimal intrauterine growth, malformations, multiple gestations, genetic factors,4 and how risk factors may interact to form causal pathways to cerebral palsy5; we aimed to build on this work.
The objective of this study was to gain more specific information concerning pathways to perinatal death or cerebral palsy in term and late preterm singletons. We sought to identify etiologically more homogenous groups of cerebral palsy and neonatal death by stratifying according to newborn neurologic status, then quantified the contributions of four major risk factors: potentially asphyxial birth events, indicators of inflammation, fetal growth restriction, and birth defects, alone or in combination, to each of these outcomes.
MATERIALS AND METHODS
This article reports on the total population case–control study of cerebral palsy and perinatal death in Western Australian births from 1980 to 1995.6–8 Cerebral palsy was defined as a disorder of movement, posture, or both affecting activities of daily living resulting from nonprogressive lesions or abnormalities of the developing brain.1 Eligible cases of cerebral palsy for this study comprised all registrants of the Western Australian Cerebral Palsy Register (now called the Western Australian Register of Developmental Anomalies–Cerebral Palsy)1 born in Western Australia between January 1, 1980, and December 31, 1995, excluding those whose cerebral palsy was acquired postneonatally.
Perinatal and vital outcome data are available for all Western Australian births in the Maternal Child Health Research Database, which links statutory birth and death registries with statutory pregnancy and delivery information and includes more than 99.5% of registered births. From this database, we selected controls (matched for gestational age [within 1 week], year of birth [within 12 months], and plurality) and intrapartum stillbirths and neonatal deaths (deaths in the first 28 days of life) in birth years specified in Figure 1. We included all 791 children with cerebral palsy born in or after 1980 (the year that the gestational age variable was added to the statutory birth data set) who were available at the time of initiating data collection. We limited etiologic heterogeneity by selecting only singletons born at or after 35 weeks of gestation, resulting in 508 participants in the control group and 494 cerebral palsy cases. With these numbers, and considering that the direction of any association would for most exposures be self-evident, we estimated there was 80% or higher power at P<.05 to detect associations likely to be of clinical importance (odds ratios [ORs] less than 0.4 or greater than 1.8) if the exposure was observed in 5–50% of participants in the control group.
This study was approved by the Princess Margaret Hospital/King Edward Memorial Hospital Human Research and Ethics Committee, individual hospital and region human research and ethics committees, ratified by the University of Sydney Human Research and Ethics Committee and approved by the Confidentiality of Health Information Committee of the Western Australia Department of Health.
To achieve more homogenous etiologic groups, we categorized cerebral palsy and neonatal death according to presence or absence of moderate or severe neonatal encephalopathy. Moderate or severe neonatal encephalopathy was defined as any admission to special or intensive care (for neonatal death) for 2 days or more (for cerebral palsy) with seizures, abnormal consciousness (lethargic or comatose), or abnormal tone.9–11 Neonatal encephalopathy was categorized as hypoxic–ischemic only if there was also a clinical diagnosis of birth asphyxia or hypoxic–ischemic encephalopathy in the medical record and is now referred to as hypoxic–ischemic encephalopathy.
We considered seven adverse outcomes: intrapartum stillbirth; neonatal death without encephalopathy, after neonatal encephalopathy, and after hypoxic–ischemic encephalopathy; and cerebral palsy without encephalopathy, after neonatal encephalopathy, and after hypoxic–ischemic encephalopathy; and two combined outcomes “all cerebral palsy” and “all neonatal deaths” (Fig. 1). We examined the association of each of these outcomes with four repeatedly identified risk factors: 1) potentially asphyxial birth events: uterine rupture, amniotic embolism, tight nuchal cord (described by the treating clinician as “tight”; eg, requiring cutting for delivery), cord prolapse, placental abruption, severe intrapartum hemorrhage (a minimum of 100 mL fresh blood), maternal cardiac arrest, or severe shoulder dystocia12,13; 2) signs of inflammation: maternal pyrexia (greater than 37.5°C), uterine tenderness, malodorous amniotic fluid, high leukocyte count, maternal or fetal tachycardia, and inflammatory placental histology14; 3) fetal growth restriction: birth weight at least two standard deviations below optimal for gestational age and gender, maternal height and parity,15 or to a decrease the risk of false-negatives with this restrictive criterion, a diagnosis of growth restriction noted in the medical record; and 4) birth defects: a structural or functional abnormality that is present at conception or occurs before the end of pregnancy. We identified birth defects by linking with the State Registry for birth defects (Western Australian Register of Developmental Anomalies), which collects information on defects diagnosed by age 6 years.16
For those with cerebral palsy, the predominant motor impairment was categorized as spastic hemiplegia, diplegia or quadriplegia, dyskinesia (dystonia or athetosis), or other (ataxia or isolated hypotonia).
For each outcome group, we estimated frequencies and proportions of each risk factor and predominant motor impairment. Odds ratios for each outcome with each risk factor were estimated by unconditional logistic regression using SAS 9.2 and SPSS 19. Statistical significance was accepted at α <.05. The proportion of each outcome group that would be prevented were the risk factor removed in isolation was estimated from the OR obtained on univariate analysis using the equation: attributable fraction=p(OR-1)/(1+p[OR-1]), where p is the proportion of participants in the control group exposed to that factor.17 Median and interquartile ranges were calculated for 5-minute Apgar scores.
Data were available for 508 participants in the control group, 494 children with cerebral palsy, 100 neonatal deaths, and 73 intrapartum stillborn singletons born at or after 35 weeks of gestation. There were no differences among these four groups with respect to maternal age, number of previous births and number of previous pregnancy losses, maternal epilepsy, intellectual disability or other neurologic disorders, coagulation disorders, or thyroid disease. Two preconceptional factors were statistically different from the control group. Of those with nonmissing data, 0.8% of participants in the control group, 1.6% of cerebral palsy, 3.1% of neonatal deaths, and 1.4% of intrapartum stillbirths had received treatment for infertility. Private health insurance was held by 52.8% of mothers of participants in the control group, 46.8% of cerebral palsy, 36% of neonatal deaths, and 34.2% of intrapartum stillbirths and was the only marker of social status widely available.
Eight children with cerebral palsy and missing data for neonatal neurologic status were retained for analyses of “all cerebral palsy” only. Of the 486 remaining children with cerebral palsy, 66.5% did not exhibit encephalopathy, 12.4% exhibited neonatal encephalopathy, and 21.2% were diagnosed with hypoxic–ischemic encephalopathy (Table 1). Of the 100 neonates who died in the neonatal period, 22% were diagnosed with hypoxic–ischemic encephalopathy before death, a further 24% exhibited neonatal encephalopathy, and 54% were not considered encephalopathic (Table 2). Three participants in the control group met our criteria for neonatal encephalopathy but none for hypoxic–ischemic encephalopathy. All 508 participants in the control group were used as the comparison group. Apgar scores at 5 minutes were consistent with assigned neonatal neurologic status: the median and interquartile range was 9 (interquartile range 9–9) for participants in the control group and children with cerebral palsy without encephalopathy, 9 (8–9) for children with cerebral palsy after neonatal encephalopathy, 4 (3–6) for children with cerebral palsy after hypoxic–ischemic encephalopathy, 9 (6–9) for neonatal deaths without encephalopathy, 7 (5–8) for neonatal deaths with encephalopathy, and 3 (1–5) for neonatal deaths with hypoxic–ischemic encephalopathy.
The value for at least one of the four risk factors was missing for no participants in the control group, 35 (7.1%) cerebral palsy cases, two (2%) neonatal deaths, and one (1.4%) intrapartum stillbirth. Of those with complete data, most (83%) of participants in the control group, 40.5% of cerebral palsy, 23% of neonatal deaths and 40.3% of intrapartum stillbirths experienced none of the four risk factors (Tables 1 and 2).
Potentially asphyxial birth events occurred with a similar low frequency in control children and those with cerebral palsy or neonatal death who were not encephalopathic in the newborn period. Potentially asphyxial birth events occurred most frequently in intrapartum stillbirths (34.3%) and in children with cerebral palsy or neonatal death after hypoxic–ischemic encephalopathy (21.6% and 22.7%, respectively) (Tables 1 and 2). Tight nuchal cord, intrapartum hemorrhage (including placental abruption), and cord prolapse accounted for 87 of the 100 events reported and there were no instances of amniotic embolism (see Appendix, available online at http://links.lww.com/AOG/A420).
Indicators of inflammation showed a similar pattern to potentially asphyxial birth events, being rare in participants in the control group and in cerebral palsy with or without neonatal encephalopathy, but occurring in significantly higher proportions of intrapartum stillbirths, neonatal deaths, and cerebral palsy with hypoxic–ischemic encephalopathy (Tables 1 and 2). Inflammation occurred in combination with potentially asphyxial birth events in 8.2% of intrapartum stillbirths but infrequently in neonatal death or cerebral palsy (Table 3).
Our criteria for growth restriction (see Appendix, http://links.lww.com/AOG/A420) were met by 5.3% of participants in the control group and at least 13% of each poor outcome, being highest in poor outcome groups after neonatal encephalopathy. For each subgroup of neonatal neurologic status, growth restriction was associated with higher odds of neonatal death than of cerebral palsy (Tables 1 and 2).
Birth defects recognized by age 6 years were associated with significantly elevated ORs and with the highest attributable fraction of the four risk factors examined for all outcome groups except for intrapartum stillbirths (Tables 1 and 2). They were identified in almost half of the largest outcome group, cerebral palsy without encephalopathy. Birth defects occurring in combination with a potentially asphyxial birth event or inflammation were seen in only a small proportion of both cerebral palsy and neonatal death cases. Birth defects and growth restriction were the most common combination of risk factors, particularly in neonatal death and cerebral palsy (Table 3). In children with cerebral palsy, a potentially asphyxial birth event, inflammation, or both was experienced by 12.6%, whereas growth restriction, a birth defect, or both was experienced by 48.6% (P<.001), making these the dominant antecedents of term and late preterm singletons.
All cerebral palsy subtypes were represented across all outcome groups (Table 4). Despite confirming the association between quadriplegia or dyskinesia and hypoxic–ischemic encephalopathy, 65% of those with quadriplegia or dyskinesia were not diagnosed hypoxic–ischemic encephalopathy, and conversely of 103 cases of cerebral palsy after hypoxic–ischemic encephalopathy, 36 were not classified quadriplegic or dyskinetic. Of the four risk factors, only birth defects (OR 1.6, 95% confidence interval [CI] 1.1–2.3), growth restriction (OR 1.7, 95% CI 1.1–2.8), and the combination of growth restriction and a birth defect (OR 1.9, 95% CI 1.1–3.6) significantly predicted quadriplegia or dyskinesia.
By defining more etiologically homogenous outcome groups, limiting study participants to singletons at or after 35 weeks of gestation, and stratifying by neonatal neurologic status, our study identified stronger associations for these risk factors and cerebral palsy or perinatal death than previously published.18 We were also able to identify poor outcome groups that had little or no association with these specific risk factors. This more informative approach to considering the etiologic pathways to term and late preterm cerebral palsy and perinatal death allowed us to see which outcomes were most related to each of these risk factors.
The term hypoxic–ischemic encephalopathy is often understood to imply a uniform etiology. However, in this population of children with cerebral palsy who had been diagnosed with neonatal hypoxic–ischemic encephalopathy, only one child in five had a clinically recognized potentially asphyxial birth event, whereas one in eight had markers of inflammation, one in seven was growth-restricted, one in four had a birth defect recognized by age 6 years, and two in five had none of these factors. Similar etiologic heterogeneity was found in neonatal deaths preceded by a diagnosis of hypoxic–ischemic encephalopathy. The most etiologically homogenous groups were neonatal death or cerebral palsy after neonatal encephalopathy in which four in five and three in five children, respectively, had a recognized birth defect.
As anticipated,19 potentially asphyxial birth events were most important for intrapartum stillbirth with a population-attributable fraction of 31%. However, population-attributable fractions for cerebral palsy or neonatal death after hypoxic–ischemic encephalopathy were higher for birth defects than for potentially asphyxial birth events, an unanticipated finding. Potentially asphyxial birth events were not associated with cerebral palsy in the absence of neurologic abnormality in the newborn period, confirming the position of the consensus statement.12
Intrauterine inflammation was the least identified risk factor but markers of inflammation available in population studies are neither sensitive nor specific and may underestimate its role. Inflammation can produce clinical findings that closely mimic birth asphyxia, so it is possible that in some children with a diagnosis of hypoxic–ischemic encephalopathy, the initiating pathologic process was inflammatory. Optimal clinical management and strategies for prevention require distinguishing neonatal neurologic depression resulting from asphyxial injury from that associated with inflammation or other processes.20 That differential diagnosis will require incorporation of information from placental examination and may also require distinguishing inflammation resulting from infection (chorioamnionitis or funisitis) from that resulting from immunologic processes (chronic villitis).21,22
Fetal growth restriction was an important factor in all examined poor outcome groups and has been shown in other studies to be an important predictor of cerebral palsy,4 neonatal encephalopathy,9,23 and stillbirth.19 Growth restriction is itself etiologically heterogenous with maternal, fetal, and placental antecedents that vary in their strength of association with cerebral pathology.7,24 Of note, the majority of neonates with growth restriction and adverse outcomes in this study also had a birth defect recognized by early childhood.
Birth defects have been recognized as risk factors for cerebral palsy at least since 195525 but in the current study were identified in a far greater proportion than is usually reported with a population-attributable fraction of 39% for total cerebral palsy. Only 1.7% considered deformational, that is, possibly a result of the brain damage that also caused cerebral palsy. This difference may be related to our focus on term and late preterm singletons because defects are identified in greater proportions of term than preterm born children with cerebral palsy,26 the inclusion of all defects occurring before delivery (because many studies of cerebral palsy etiology exclude birth defects, at least those of the brain and spinal cord), and the inclusion of defects recognized up to the age of 6 years.27 Acceptance of defects with delayed recognition raises the possibility of outcome bias; in an effort to counter this bias, minor defects were included in these analyses only if they were identified neonatally, before the outcome of cerebral palsy was known. The combination of growth restriction and a birth defect was the strongest predictor of neonatal death and quadriplegic or dyskinetic cerebral palsy. At present, both birth defects and marked growth restriction are exclusion criteria for trials of therapeutic hypothermia. Birth defects are also exclusion criteria for trials to improve neurologic outcomes associated with growth restriction. Given the considerably elevated risk this group faces, it may be necessary to investigate antecedents and approaches to management for these two risk factors, especially when they co-occur.
The strengths of this study include its prospective design in a total geographically defined population and ascertainment of birth defects as recorded in the State Register up to 6 years of age. It includes perinatal deaths and attempts to identify etiologically more specific pathways by stratifying by neonatal neurologic status. Among its limitations, antepartum stillbirths were not included because retrospective data for these occurrences were of relatively poor quality. Population-attributable fractions estimate the clinically interpretable fraction of the outcome preventable on the isolated removal of an exposure by considering both frequency of exposure and relative risk; however, they must be interpreted cautiously. Our case–control study design necessitates estimation of ORs, which overestimate relative risk particularly when risks are high, resulting in overestimation of population-attributable fractions even if the exposure is indeed causal. The medical records from which our data were extracted were not created for research and certain potentially important observations were not regularly available, the most significant omission being placental histology. Cerebral imaging was not routinely performed in this era so perinatal stroke, a common cause of hemiplegic cerebral palsy in term neonates,28 is not reported. Although these data are based on birth years 1980–1995, the rate of cerebral palsy in term and late preterm singletons remained constant throughout the study period and has been unchanged since.1,2
Concurrent investigation of major risk factors in a total population and categorization of outcomes by neonatal neurologic status allows a better understanding of the association of specific antecedents with perinatal death and cerebral palsy than was previously available. When 33 research priorities for the etiology and prevention of cerebral palsy were recently agreed on, more than one third focused on infection or inflammation and hypoxia–ischemia, whereas none addressed birth defects or growth restriction.29 Surely, research priorities in cerebral palsy need reconsideration. This study adds weight to the evidence that in singletons born at or after 35 weeks of gestation, very significant proportions of cerebral palsy and of perinatal death are associated with antenatal maldevelopment.
1. Watson L, Blair E, Stanley F. Report of the Western Australian Cerebral Palsy Register to birth year 1999. Perth, Australia: Telethon Institute for Child Health Research; 2006.
2. Himmelmann K, Hagberg G, Uvebrant P. The changing panorama of cerebral palsy in Sweden. X. Prevalence and origin in the birth-year period 1999–2002. Acta Paediatr 2010;99:1337–43.
3. Korzeniewski SJ, Birbeck G, DeLano MC, Potchen MJ, Paneth N. A systematic review of neuroimaging for cerebral palsy. J Child Neurol 2008;23:216–27.
4. Himmelmann K, Ahlin K, Jacobsson B, Cans C, Thorsen P. Risk factors for cerebral palsy in children born at term. Acta Obstet Gynecol Scand 2011;90:1070–81.
5. Blair E. Epidemiology of the cerebral palsies. Orthop Clin North Am 2010;41:441–55.
6. McIntyre S, Badawi N, Brown C, Blair E. Population case-control study of cerebral palsy: neonatal predictors for low risk term singletons. Pediatrics 2011;127:e667–73.
7. Blair E, DeGroot J, Nelson K. Placental infarction identified by macroscopic examination and risk of cerebral palsy in infants at 35 weeks of gestational age and over. Am J Obstet Gynecol 2011;205:124.e1–7.
8. Taylor CL, De Groot J, Blair EM, Stanley FJ. The risk of cerebral palsy in survivors of multiple pregnancies with co-fetal loss or death. Am J Obstet Gynecol 2009;201:41.e1–6.
9. Badawi N, Kurinczuk JJ, Keogh JM, Alessandri LM, O’Sullivan F, Burton PR, et al.. Antepartum risk factors for newborn encephalopathy: the Western Australian case-control study. BMJ 1998;317:1549–53.
10. Sarnat HB, Sarnat MS. Neonatal encephalopathy following fetal distress. A clinical and electroencephalographic study. Arch Neurol 1976;33:696–705.
11. Jacobs S, Hunt R, Tarnow-Mordi W, Inder T, Davis P. Cooling for newborns with hypoxic ischaemic encephalopathy. The Cochrane Database of Systematic Reviews 2007, Issue 1. Art. No.: CD003311. DOI: 10.1002/14651858.CD003311.pub3.
12. MacLennan A. A template for defining a causal relation between acute intrapartum events and cerebral palsy: international consensus statement. BMJ 1999;319:1054–9.
13. Perlman JM. Intrapartum asphyxia and cerebral palsy: is there a link? Clin Perinatol 2006;33:335–53.
14. Shatrov JB, Birch SC, Tam L, Quinlivan J, McIntyre S, Mendz G. Chorioamnionitis and cerebral palsy: a meta-analysis. Obstet Gynecol 2010;116;387–92.
15. Blair EM, Liu Y, de Klerk NH, Lawrence DM. Optimal fetal growth for the Caucasian singleton and assessment of appropriateness of fetal growth: an analysis of a total population perinatal database. BMC Pediatr 2005;5:13.
16. Bower C, Rudy E, Callaghan A, Quick J, Cosgrove P, Watson L. Report of the Western Australian Register of Developmental Anomalies 1980-2010. Available at: http://www.kemh.health.wa.gov.au/services/register_developmental_anomalies/warda.htm
. Retrieved January 10, 2013.
17. Rockhill B, Newman B, Weinberg C. Use and misuse of population attributable risk. Am J Public Health 1998;88:15–9.
18. McIntyre S, Taitz D, Keogh J, Goldsmith S, Badawi N, Blair E. A systematic review of risk factors for cerebral palsy in children born at term in developed countries. Dev Med Child Neurol 2013;55:499–508.
19. Flenady V, Koopmans L, Middleton P, Frøen JF, Smith GC, Gibbons K, et al.. Major risk factors for stillbirth in high-income countries: a systematic review and meta-analysis. Lancet 2011;377:1331–40.
20. Wintermark P, Boyd T, Gregas MC, Labrecque M, Hansen A. Placental pathology in asphyxiated newborns meeting the criteria for therapeutic hypothermia. Am J Obstet Gynecol 2010;203:579.e1–9.
21. Kim MJ, Romero R, Kim CJ, Tarca AL, Chhauy S, LaJeunesse C, et al.. Villitis of unknown etiology is associated with a distinct pattern of chemokine up-regulation in the feto-maternal and placental compartments: implications for conjoint maternal allograft rejection and maternal anti-fetal graft-versus-host disease. J Immunol 2009;182:3919–27.
22. Hayes BC, Cooley S, Donnelly J, Doherty E, Grehan A, Madigan C, et al.. The placenta in infants >36 weeks gestation with neonatal encephalopathy: a case control study. Arch Dis Child Fetal Neonatal Ed 2013;98:F233–40.
23. Bukowski R, Burgett AD, Gei A, Saade GR, Hankins GD. Impairment of fetal growth potential and neonatal encephalopathy. Am J Obstet Gynecol 2003;188:1011–5.
24. Halliday HL. Neonatal management and long-term sequelae. Best Pract Res Clin Obstet Gynaecol 2009;23:871–80.
25. Eastman NJ, Deleon M. The etiology of cerebral palsy. Am J Obstet Gynecol 1955;69:950–61.
26. Rankin J, Cans C, Garne E, Colver A, Dolk H, Uldall P, et al.. Congenital anomalies in children with cerebral palsy: a population-based record linkage study. Dev Med Child Neurol 2010;52:345–51.
27. Bower C, Rudy E, Callaghan A, Quick J, Nassar N. Age at diagnosis of birth defects. Birth Defects Res A Clin Mol Teratol 2010;88:251–5.
28. Wu YW, March WM, Croen LA, Grether JK, Escobar GJ, Newman TB. Perinatal stroke in children with motor impairment: a population-based study. Pediatrics 2004;114:612–9.
29. McIntyre S, Novak I, Cusick A. Consensus research priorities for cerebral palsy: a Delphi survey of consumers, researchers and clinicians. Dev Med Child Neurol 2010;52:270–5.