The Australian cerebral palsy research study - epidemiological and genetic associations with cerebral palsy.

Abstract

Introduction Twenty two mostly small studies have reported associations between cerebral palsy (CP) and specific single nucleotide polymorphisms (SNPs). These data require prospective confirmation in a large cohort. Only one study has examined maternal genetic risk factors for CP. The current large study of mothers and children examines the contributions of genetic and epidemiological factors to CP and their interactions. Methods Caucasian children aged between five and 18 years who were born in Australia were recruited with their mothers. Results from 587 case pairs and 1,154 control pairs were analysed. Each mother and child provided DNA using buccal swabs. Multiplex PCR was used to genotype individuals and 35 specific SNPs were included in the analysis. These candidate SNPs have been putatively associated with thrombophilia, inflammation and preterm birth. Mothers completed a health, pregnancy and delivery questionnaire. State perinatal data for each participant provided further epidemiological data, while CP registers provided cerebral palsy diagnosis data. Univariable analysis examined each epidemiological and genetic risk factor individually with Bonferroni correction for multiple testing. Subsequent multivariable analyses were performed combining risks and examining interactions. Odds ratio (OR) and 95% confidence intervals are reported. Results Univariable analysis of SNP associations with CP did not confirm the majority of associations reported in the literature after correction for multiple testing. Prothrombin gene mutation in the child remained associated with hemiplegia in term delivered infants where a maternal infection during pregnancy was reported (OR 4.52, 1.70-12.03, p=0.059 after Bonferroni correction). Epidemiological associations with CP included maternal infection during pregnancy (OR 1.55, 1.26-1.91), small for gestational age (<10th centile, OR 4.35, 2.92-6.48), gestational age <32 weeks (OR 59.20, 28.87-121.38), multiple birth (OR 6.62, 4.00-10.95), a relative with CP (OR 1.61, 1.12-2.32) and male gender (OR 1.68, 1.38-2.06). Iatrogenic heat in labour did not increase the risk of CP. Multivariable analyses of genetic and epidemiological risk factors identified significant associations of CP with male gender (OR 1.5, 1.1-2.1), reported maternal infection (OR 1.9, 1.2-3.0), a relative with CP (OR 1.8, 1.1-2.9) and maternal carriage of TGF-β1-509 (OR 1.3, 1.0-1.6). Subtype analyses showed an increased risk of quadriplegic CP with a family history of CP (OR 3.27, 1.13-9.45). The interaction of maternal TNF-α 308 with infection was inversely related to CP (OR 0.7, 0.5-0.9). Discussion Most SNP associations in the literature were not confirmed by this study, probably because those studies did not correct for multiple testing. The study confirmed the reported epidemiological associations with CP in the literature. Combined multivariable analyses suggest an association of some maternal and fetal genotypes with CP outcome, particularly when an infection was reported. These results require further study and the mechanism of association is yet to be determined. Conclusions The individual SNPs studied are unlikely to play a large role in CP causation, although maternal SNPs and interactions with infection may be significant. Specific subtypes of CP (particularly term born quadriplegia) are more likely to have a genetic origin than other types. This study has stimulated new genetic and environmental studies of CP.