P2-259 Cancer risk in children with birth defects: a longitudinal, population-based assessment among 2.7 million births

Introduction The published literature, to date, is largely inconclusive regarding cancer risk among children with birth defects. To improve knowledge of such risk, we studied population-based (statewide) birth cohorts from Arizona, Iowa, and Utah selected from among 2.7 million births delivered from 1983 to 2006. Methods Birth defect and cancer diagnoses were identified from linked population-based surveillance systems. A population-based cohort of over 43 000 children with major birth defects (including trisomies 13, 18, and 21) was compared to a cohort of nearly 148 000 births without birth defects, randomly sampled from the same underlying birth population and frequency-matched to the birth defects cohort by birth year. Kaplan-Meier time-to-event analysis, accounting for censoring by death, was used to estimate cancer risk up to age 15 years. Results Compared to the reference cohort, children with birth defects had a statistically significant increase in cancer risk (RR, 2.73). Risk was highest among children with Down syndrome (RR, 13.2), and was driven largely by leukaemias. Cancer risk was moderately increased among children with a birth defect but without chromosomal anomalies (RR, 1.82). In this group, cancer risk was driven largely by brain tumours and embryonal tumours, and occurred mainly in children with brain defects, cleft palate, rectal defects, and some heart defects. Conclusion These population-based findings support and extend previous findings that suggest increased cancer risk in children with birth defects, including non-chromosomal defects, and suggest selected defect groups in which further research could help identify a common genetic susceptibility to cancer and birth defects.