Abstract

The etiology of childhood cancer is not well understood, but may be linked to prenatal and perinatal factors, such as maternal diabetes. However, this association has not been examined in depth. We aimed to determine if maternal diabetes is associated with risk of childhood brain tumor (CBT), leukemia (all types combined and acute lymphoblastic leukemia [ALL] separately), and lymphoma. All children born in Sweden between 1973 and 2014 (n=4,239,965) were followed from birth until first cancer diagnosis, age 15 years, or December 31, 2015. Data on maternal diabetes, childhood cancer, and covariates were obtained from nationwide health registers. Incidence rate ratios (IRRs) and 95% confidence intervals (CIs) were calculated using Cox regression adjusted for potential confounders/mediators. Additionally, we performed an exploratory analysis using results from published genome-wide association studies and functional annotation. Maternal diabetes was associated with lower risk of CBT (adjusted IRR [95% CI]: 0.56 [0.35-0.91]) and higher risk of leukemia (adjusted IRR: 1.47 [1.13-1.92] for all leukemia combined and 1.64 [1.23-2.18] for ALL). These associations were similar for both maternal type 1 diabetes and gestational diabetes. Associations of five previously identified genetic loci were compatible with a causal effect of diabetes traits on neuroblastoma and common Hodgkin's lymphoma. Children whose mother had diabetes had lower risk of CBT and higher risk of leukemia, compared with children whose mother did not have diabetes. Our results are compatible with a role of prenatal and perinatal glycemic environment in childhood cancer etiology.