Inequality in risk of cerebral palsy by parental educational status–A population-based cohort study with sibling analysis

Abstract

Introduction Cerebral palsy (CP) is the most common cause of physical disability in children, affecting about 2 per 1000 live-born. In a registry-based study in Norway and Denmark using a cohort design, we found that the risk of having a child with CP decreases with higher parental education. This inverse association may reflect a causal effect of higher education on risk of CP in the child through lifestyle factors and health behaviors. However, the observed association could also be due to confounding from unobserved family-level factors including genetic factors or childhood socioeconomic status of the parents. By using a sibling case-control design comparing the association within pairs of full siblings discordant on CP status it is possible to hold constant some of the unobserved factors that are shared between siblings. By comparing results from a cohort design with results from a sibling case-control design, we aim to investigate whether such familial confounding influences the association between parental education and the risk of CP in the child. Methods The full study cohort includes 1275,819 children who survived the first year of life, registered in the Danish Medical Birth Registry (MBR) during 1981–2007. The sibling case-control cohort includes 2891 full-sibling pairs discordant on CP status. Information on parents’ highest attained education in the year before delivery was retrieved from Statistics Denmark, while CP status originated from the Danish Cerebral Palsy Registry. The association between parental educational attainment and risk of CP in the child will be estimated using logistic regression with cluster robust standard error in the full cohort and a conditional logistic regression model in the sibling case-control cohort, adjusting for year of delivery and parental age. Results Preliminary analyses show that the risk of CP in the child decreases with increasing maternal education both in the cohort analyses and in the sibling case-control analyses. Compared to mothers with only primary or lower secondary education, mothers with higher education had lower odds of having a child with CP (adjusted odds ratio (aOR) 0.73, 95% confidence interval (CI) 0.66–0.82 in the cohort analysis and aOR 0.54, 95% CI 0.29–0.99 in the sibling case-control analysis). A similar association was found for fathers (aOR 0.71, 95% CI 0.63–0.80 in cohort analyses and aOR 0.75, 95% CI 0.33–1.74 in sibling-case control analyses). Additional adjustment for parity had practical no impact on the results. Conclusions Our results may indicate that the observed inverse association between parental education and risk of CP in the child is not entirely attributable to factors unobserved family-level factors shared between siblings. There is a wide range of possible underlying causal mechanisms, at least some of which seem potentially preventable and should be further explored.