A nationwide Korean cohort study published online in Scientific Reports found no statistically significant association between maternal diabetes and the overall risk of childhood cancer, according to an Aug. 4 report. The research analyzed data from more than half a million singleton infants born during 2003–2004, using the Korean National Health Insurance database, which covers 97% of all Koreans.
Researchers compared the risk of childhood cancer among offspring of mothers with and without diabetes, adjusting for factors such as maternal age, hypertension, preterm birth, offspring sex, and various neonatal conditions. Of the mothers studied, 6% had diabetes during pregnancy and were generally older with higher rates of hypertension and preterm births. Their children also showed increased rates of respiratory distress syndrome and necrotizing enterocolitis.
During follow-up until the end of 2021, a total of 6,049 children (1.1%) developed cancer. After adjustments for confounding factors, there was no statistically significant link between maternal diabetes and overall childhood cancer risk or any specific age-at-diagnosis category.
However, when stratified by cancer type, researchers observed an 81% increase in relative risk for urinary tract cancers among children born to mothers with diabetes—a finding that remained statistically significant after correction for multiple comparisons but represented only a small absolute increase: about 37 additional cases per 100,000 children. The authors cautioned against overinterpreting this result due to its low absolute incidence and called it “an epidemiologic signal requiring replication rather than as evidence of a large absolute cancer burden.”
Separate analyses suggested nominal associations between gestational diabetes and certain non-blood cancers but did not reveal clear differences by subtype after formal testing. Limitations included reliance on insurance claims data rather than clinical assessments; lack of information on birth order or weight; missing paternal characteristics; potential selection bias due to inclusion only of live births; possible exposure misclassification; limited generalizability beyond Korea; and small numbers for some site-specific cancers.
The authors concluded that while their findings are broadly reassuring regarding overall childhood cancer risk in relation to maternal diabetes during pregnancy—particularly given existing links between adult diabetes and several other cancers—the observed urinary tract signal should be further investigated in diverse populations.
