Letter regarding "Maternal smoking during pregnancy and risk of childhood-onset type 1 diabetes in offspring: A systematic review and meta-analysis".

This correspondence discusses the recent meta-analysis by Zhang et al. examining maternal smoking during pregnancy and the risk of childhood-onset type 1 diabetes in offspring. The reported inverse association aligns with findings from previous meta-analyses and family-based epidemiological studies. However, interpreting this association remains challenging because observational perinatal research is susceptible to residual confounding, selection effects, and collider bias. Maternal smoking is closely linked with socioeconomic and behavioral factors that may independently influence immune development and autoimmune susceptibility in offspring. Mechanistic evidence also remains inconsistent, as experimental and developmental studies have generally associated prenatal tobacco exposure with adverse immune and metabolic outcomes rather than sustained autoimmune protection. The correspondence further highlights the absence of a formal GRADE certainty assessment and emphasizes that statistical precision should not be interpreted as evidence of causality. Further studies utilizing biomarker-confirmed exposure assessment, Mendelian randomization, and other causal inference approaches are warranted.