O'Callaghan, Michael E. BSc (Hons); MacLennan, Alastair H. MD; Gibson, Catherine S. PhD; McMichael, Gai L.; Haan, Eric A. MBBS; Broadbent, Jessica L. BSc (Hons); Goldwater, Paul N. MBBS; Dekker, Gustaaf A. PhD; for the Australian Collaborative Cerebral Palsy Research Group
Cerebral palsy has been investigated extensively for epidemiologic risk factors with research dating back to the 1970s.1,2 A number of key factors have been well established as associated with cerebral palsy including preterm birth, low birth weight, infection in pregnancy, and twin births.3,4 These have been established from population-wide databases with near complete ascertainment and a low likelihood of bias and have assessed up to 7,242 cerebral palsy cases.5
The Australian Cerebral Palsy Research Study is an ongoing cerebral palsy gene-association study that is examining possible genetic interaction with environmental triggers.6 Part of the a priori aims of the study was to examine known epidemiologic risk factors for cerebral palsy as well as other potential risk factors that have not been investigated to date. Data that are not routinely available from state perinatal databases were collected by maternal questionnaire. For example, knowledge of maternal weight at the beginning of pregnancy and height have allowed assessment of maternal body mass index (BMI, calculated as weight (kg)/[height (m)]2) as a risk factor for cerebral palsy and, together with parity, sex, and race, have allowed the calculation of customized birth weight for gestational age percentiles. Data were also gathered about antenatal infection type and timing. Data about the use of hot packs, hot showers, or hot baths during labor allowed assessment of heat during labor as a risk factor for cerebral palsy. Data on plurality at first-trimester ultrasonography and birth were collected to assess the disappearing twin hypothesis as a risk factor for cerebral palsy. The objective of this study was to estimate epidemiologic risk factors for cerebral palsy.
MATERIALS AND METHODS
Recruitment of the cohort is described in detail elsewhere6 and the cohort demographics are summarized in Table 1. In summary, mother-child case–control pairs were recruited from around Australia between July 2008 and March 2010 with the following inclusion criteria: children aged between 5 and 18 years, born in Australia, and of Caucasian background. Cases and controls were recruited during the same time period and met the same inclusion criteria. Controls were randomly selected. Five hundred eighty-seven case families and 1,154 control families were included in the analysis. Ethics committee approval (approval number REC1946/4/10) was obtained from Adelaide Children, Youth and Women's Health Service.
Clinical data were obtained by linkage to state perinatal data sets throughout Australia and by a maternal questionnaire. The source of each data point is described in Table 2. Cerebral palsy diagnosis data were retrieved by linkage to cerebral palsy registers in each state and by contacting specialist clinicians where a link could not be made. Details can be found in the published protocol.
Univariable comparison between cases and controls was made using a χ2 test and PASW 17.0.2 ; P<.05 was considered significant. Individuals with missing variables were excluded from tests examining that variable. Where cell counts were less than five, Fisher exact test was used. Tests were conducted as per the study protocol with the addition of Apgar score at 1 and 5 minutes, sex, delivery type, maternal age, and previous miscarriage. Calculations detailed in the study protocol indicate that this cohort has sufficient power to detect an odds ratio (OR) of 1.7 with 80% power and α 0.01 for variables present in 12% of controls.
Combined reports of maternal infection during pregnancy reported in the questionnaire or the perinatal data were significantly associated with cerebral palsy outcome (Table 3). The maternal questionnaire enabled retrospective subjective assessment of infection type and timing during pregnancy based on maternal recollection, and provided information about some antenatal infections not routinely recorded by midwives after birth in state perinatal data sets. Although some infections presenting early during pregnancy were significantly associated with cerebral palsy outcome, the trend for increased risk was toward infections occurring late in pregnancy. Reassuringly, upper respiratory tract and gastrointestinal infections were not associated with a cerebral palsy outcome.
Small for gestational age calculated using customized birth weight for gestational age ranges was significantly associated with cerebral palsy (Table 4). Neonates with birth weights less than 3rd, 4th to 5th, 6th to 10th, 10th to 19th, or 40th centile were more likely to have cerebral palsy than those with birth weight in the 40th to 60th centiles. Increasing risk was apparent from below the 20th centiles. There was no significant association between cerebral palsy and birth weight between the 20th and 29th centile or above the 60th or 90th centiles when compared with those with birth weight between the 40th and 60th centiles.
Early gestational age was significantly associated with cerebral palsy outcome (Table 5). Birth before 32 weeks of gestation was a major risk factor for cerebral palsy (OR 59.2) when compared with all other gestational ages and also when compared with term-born neonates. Children born at 32 to 36 weeks were also at increased risk of a cerebral palsy (OR 5.0) compared with term neonates.
Multiple birth was significantly associated with cerebral palsy (OR 6.62 for twins) (Table 5). Only one triplet pregnancy was reported in this study and although this included a cerebral palsy case, there are insufficient data to draw any conclusions.
Children with a history of any relative with cerebral palsy were more likely to have cerebral palsy themselves (Table 5). However, having a sibling with cerebral palsy was not a significant risk. There were insufficient data to conclude that a parent with cerebral palsy is a risk factor for the condition in his or her child; however, all examples in this cohort were cases. More distant family members with cerebral palsy were a risk factor for cerebral palsy.
Breech presentation at the time of delivery was significantly associated with cerebral palsy when compared with cephalic presentation (Table 5). For neonates presenting breech in labor, vaginal delivery and emergency cesarean delivery were both associated with cerebral palsy, whereas elective cesarean delivery was not when compared with spontaneous cephalic vaginal delivery.
“Disappearing twin” was considered to be present when a multiple pregnancy was reported at the first trimester ultrasound examination and only a singleton delivery was reported subsequently. Only 10 such cases were reported and although association was not observed, it cannot be excluded with these numbers (Table 6).
Data about use of heat in labor (including hot packs, hot showers, or hot baths given in labor) were collected by maternal questionnaire. There was an inverse relationship between heat in labor and cerebral palsy outcome (Table 6).
Bleeding at any time in pregnancy, nuchal cord or cord entanglement, illicit drug use, tobacco smoking, low Apgar score at 1 and 5 minutes, male sex, emergency and elective cesarean delivery, and mothers having three or more previous miscarriages were significantly associated with cerebral palsy (Table 6).
Diabetes (including gestational diabetes and pre-existing maternal diabetes), maternal BMI (when comparing underweight [BMI less than 18.5], overweight [BMI 25–30] or obese [BMI more than 30] categories with mothers in the normal range [BMI 18.5–25]), maternal hypertension (including both pre-existing hypertension and gestational hypertension), maternal alcohol consumption during pregnancy (one to two drinks per week, three to four drinks per week, or more than four drinks per week compared with less than 1 drink per week), maternal anemia, maternal hypothyroidism, forceps delivery, vacuum delivery, and maternal age were not significantly associated with cerebral palsy outcome.
This study investigates upper respiratory tract, gastrointestinal, and herpes infection, and fever from early pregnancy through to 1 week after delivery as a risk factor for cerebral palsy. Our data did not show any association of upper respiratory tract infections or gastrointestinal infection with cerebral palsy. This may be reassuring to pregnant women who worry about the effect of these common infections on their newborns.
In this study, the most significant associations were found in the 21–40 weeks of gestation age range and included fever, labor and delivery complicated by infection, and other types of infections, for example, cytomegalovirus and chicken pox. Our data support the hypothesis that infection late in pregnancy is more likely to be associated with cerebral palsy than infection early in pregnancy.
This study used customized birth weight centiles as a proxy marker of intrauterine growth restriction. Significant associations were seen between birth weight in the low centiles and cerebral palsy. There was an increasing risk of cerebral palsy for birth weight below the 20th centile and the risk was nearly 12-fold under the third centile. This finding is supported in the literature.8,9 Inappropriate growth in utero may cause abnormal brain development, or alternatively, inappropriate growth in utero may occur a consequence of brain pathology; but retrospective case-control studies cannot determine which of these possibilities is more likely.
Our results support the association of gestational age and cerebral palsy with an OR of 70.6 (95% CI 34.38–145.04) reported for births at less than 32 weeks of gestation when compared with term-born controls.
Our study supports the association of twins with cerebral palsy. The mechanism by which multiple birth increases cerebral palsy risk is difficult to define as multiple birth is strongly associated with preterm delivery, intrauterine growth restriction, and other obstetric complications, all of which contribute to cerebral palsy outcome.
Familial aggregations of cerebral palsy have been reported previously10; however, we were unable to confirm this significant association, probably because of the low incidence of siblings with cerebral palsy (1%). Parents with cerebral palsy were noted only among the case cohort and not in control families, supporting a trend toward association. The familial aggregations of cerebral palsy and the predominance of male children with cerebral palsy suggest a genetic component in the etiology of cerebral palsy, although an environmental risk common to the family cannot be excluded.
Our data confirm a significant association between breech position and cerebral palsy and suggest that this risk is potentiated by vaginal delivery or emergency cesarean delivery, which corroborates the literature.11 Our data do not support an association of breech presentation with elective cesarean delivery and cerebral palsy, although the number of cases was small and replication is required before clinical practice recommendations could be made.
Disappearing twin as a risk factor for cerebral palsy has been reported previously; however, the numbers reported in this and past studies are small.12,13 Our study examined fetal demise in the first trimester (12 weeks) and shows a similar trend. Our sample size was comparable with the previously reported studies with 10 of 587 cases and 10 of 1,154 controls reporting fetal demise in the first trimester. Although our cohort did not use population data, it is unlikely to be biased in regard to this factor. All studies to date have been underpowered, but because they have all been performed in independent cohorts and report a common trend, it is possible that a disappearing twin does contribute to cerebral palsy.
Use of heat in labor and cerebral palsy outcome in neonates has not been investigated to date. Because intrapartum pyrexia14 is a documented risk factor for cerebral palsy and head cooling appears to have possible neuroprotective properties in selected cases,15 this study set out to investigate whether iatrogenic heating of the mother, through application of heat by allowing hot baths, hot showers, or hot packs in labor, contributed to risk of cerebral palsy. Our results show an inverse relation to cerebral palsy outcome. This association may be the result of confounding, as women who have uncomplicated pregnancies are more likely to be offered access to hot packs, baths, or showers during labor than those with obstetric risk factors for cerebral palsy such as preterm labor, who are more likely to receive fixed electronic fetal monitoring and have less access to heating interventions. More evidence is required before recommendations about the safety of this common clinical practice can be made.
Our study was able to confirm associations of cerebral palsy with smoking and illicit drug use but not with alcohol use, although maternal retrospective reporting bias may have influenced the data analyzed for these measures. These risk factors, while moderate, are modifiable.
Apgar score at 1 minute has been associated with cerebral palsy16 and is confirmed by our study. Apgar scores are a composite of clinical signs and a reflection of neonatal cardiorespiratory function at birth. They may reflect both acute and chronic fetal compromise and are not by themselves a cause of cerebral palsy. Male sex has also been associated with cerebral palsy and this is confirmed in our cohort,17 possibly suggesting an X-linked contribution to cerebral palsy. This is the first study to report an association between history of miscarriage and cerebral palsy outcome. Previous miscarriage may indicate genetic anomalies18,19 or clotting disorders,20 both of which have been associated with cerebral palsy.21,22
The data in this article are based on our maternal questionnaire, perinatal data held by state health departments relating to each pregnancy, and state cerebral palsy registers. Comparisons suggest that the cohorts in this study are representative of the Australian population but recruitment biases cannot be excluded. Biases can be avoided by conducting population-wide studies, but this would preclude the use of a detailed and customized questionnaire and reduce the data available, particularly in relation to antenatal infections. Maternal recall bias may also influence the data collected and could vary between cases and controls. Care has been taken to preferentially use the state-held data set rather than the maternal questionnaire to reduce this occurrence, where possible. Assessment of the data in this article by multivariate analysis will be described in a subsequent article when both epidemiologic and genetic risk factors along with their interactions will be included.
This article presents data supporting an association between cerebral palsy and maternal infection late in gestation, intrauterine growth restriction, early gestational age, multiple birth, family members with cerebral palsy, breech position, smoking and drug use, low Apgar scores, male sex, cesarean delivery, and previous miscarriages. Upper respiratory tract infections and gastroenteritis during pregnancy, which are common and often a concern to pregnant women, were not associated with a significantly increased risk of cerebral palsy.
1. Stanley FJ, Watson L. Methodology of a cerebral palsy register. The Western Australian experience. Neuroepidemiology 1985;4:146–60.
2. Nelson KB, Ellenberg JH. Epidemiology of cerebral palsy. Adv Neurol 1978;19:421–35.
3. Stanley F, Blair E, Alberman E. Cerebral palsies: epidemiology and causal pathways. Vol 151. London (UK): Mac Keith Press; 2000.
4. Nelson KB, Ellenberg JH. Antecedents of cerebral palsy. I. Univariate analysis of risks. Am J Dis Child 1985;139:1031–8.
5. Gilbert WM, Jacoby BN, Xing G, Danielsen B, Smith LH. Adverse obstetric events are associated with significant risk of cerebral palsy. Am J Obstet Gynecol 2010;203:328.e1–5.
6. O'Callaghan ME, MacLennan AH, Gibson CS, McMichael GL, Haan EA, Broadbent J, et al.. The Australian cerebral palsy research study–protocol for a national collaborative study investigating genomic and clinical associations with cerebral palsy. J Paediatr Child Health 2011;47:99–110.
7. Gardosi J, Francis A. Customised Weight Centile Calculator–GROW-Centile vol. 6.4. Gestation Network; 2009. Available at: www.gestation.net
. Retrieved December 2010.
8. Blair E, Stanley F. Intrauterine growth and spastic cerebral palsy. I. Association with birth weight for gestational age. Am J Obstet Gynecol 1990;162:229–37.
9. Jacobsson B, Ahlin K, Francis A, Hagberg G, Hagberg H, Gardosi J. Cerebral palsy and restricted growth status at birth: population-based case-control study. BJOG 2008;115:1250–5.
10. Hemminki K, Li X, Sundquist K, Sundquist J. High familial risks for cerebral palsy implicate partial heritable aetiology. Paediatr Perinat Epidemiol 2007;21:235–41.
11. Andersen GL, Irgens LM, Skranes J, Salvesen KA, Meberg A, Vik T. Is breech presentation a risk factor for cerebral palsy? A Norwegian birth cohort study. Dev Med Child Neurol 2009;51:860–5.
12. Newton R, Casabonne D, Johnson A, Pharoah P. A case-control study of vanishing twin as a risk factor for cerebral palsy. Twin Res 2003;6:83–4.
13. Taylor CL, de Groot J, Blair EM, Stanley FJ. The risk of cerebral palsy in survivors of multiple pregnancies with cofetal loss or death. Am J Obstet Gynecol 2009;201:41.e1–6.
14. Impey L, Greenwood C, MacQuillan K, Reynolds M, Sheil O. Fever in labour and neonatal encephalopathy: a prospective cohort study. BJOG 2001;108:594–7.
15. Schulzke SM, Rao S, Patole SK. A systematic review of cooling for neuroprotection in neonates with hypoxic ischemic encephalopathy - are we there yet? BMC Pediatr 2007;7:30.
16. Nelson KB, Ellenberg JH. Apgar scores as predictors of chronic neurologic disability. Pediatrics 1981;68:36–44.
17. Tioseco JA, Aly H, Essers J, Patel K, El-Mohandes AA. Male sex and intraventricular hemorrhage. Pediatr Crit Care Med 2006;7:40–4.
18. Hassold T, Chen N, Funkhouser J, Jooss T, Manuel B, Matsuura J, et al.. A cytogenetic study of 1000 spontaneous abortions. Ann Hum Genet 1980;44(pt 2):151–78.
19. Hassold TJ. A cytogenetic study of repeated spontaneous abortions. Am J Hum Genet 1980;32:723–30.
20. Preston FE, Rosendaal FR, Walker ID, Briet E, Berntorp E, Conard J, et al.. Increased fetal loss in women with heritable thrombophilia. Lancet 1996;348:913–6.
21. Report of the Australian Cerebral Palsy Register, Birth Years 1993–2003. ACPR Group; 2009.
22. Curry CJ, Bhullar S, Holmes J, Delozier CD, Roeder ER, Hutchison HT. Risk factors for perinatal arterial stroke: a study of 60 mother-child pairs. Pediatr Neurol 2007;37:99–107.
© 2011 by The American College of Obstetricians and Gynecologists.