To critically review and summarize the clinical evidence relating a short duration of breast-feeding or early cow's milk exposure to insulin-dependent (type I) diabetes.
All relevant citations retrieved through comprehensive searching of the medical literature were critically reviewed and analyzed. Those case-control studies that minimized the possibility of bias were meta-analyzed to determine overall odds ratios (ORs) and 95% confidence intervals (CIs).
Ecological and time-series studies consistently showed a relationship between type I diabetes and either cow's milk exposure or diminished breast-feeding. In the case-control studies, patients with type I diabetes were more likely to have been breast-fed for <3 months (overall OR 1.43; 95% CI 1.15–1.77) and to have been exposed to cow's milk before 4 months (overall OR 1.63; 95% CI 1.22–2.17). Slightly lower ORs were obtained when all of the case-control studies were metaanalyzed in a sensitivity analysis.
Early cow's milk exposure may be an important determinant of subsequent type I diabetes and may increase the risk ∼1.5 times.