BACKGROUND & AIMS: Whether and how socioeconomic status (SES) influences the associations between type 2 diabetes mellitus (T2DM) and lactation remains unknown. We aimed to evaluate the associations between lactation duration and T2DM from a perspective of SES disparity. METHODS: A total of 263,859 parous women without diabetes at baseline (2004-2008) in the China Kadoorie Biobank were included. Lactation duration was counted for the first-child, per-child and lifetime. The latent class analysis of education level, household income, occupation, and residence was conducted to derive SES as low, intermediate, and high. T2DM cases were identified through linkage with records in hospitals, disease registry system and health insurance data during follow-up (2008-2015). Multivariable Cox proportional hazards regression models were applied to estimate hazard ratios (HRs) and 95 % confidence intervals for T2DM. Population attributable fraction evaluated the cases tied to insufficient lactation by SES. RESULTS: In a median 9.2-year follow-up, 8,204 cases were identified. Women who breastfed their first child for 12-24 and >
24 months respectively, had a reduced risk of diabetes (fully adjusted HR: 0.84 [0.75-0.94] and 0.81 [0.70-0.95]). Similar results were found for per-child (0.84 [0.72-0.98] and 0.71 [0.59, 0.85]), and lifetime lactation for >
36 months (0.66 [0.56, 0.77]). For dose-response associations, every 5-month increase in lactation duration (first-child, per-child, lifetime) was linked to a 7 %, 10 %, and 4 % lower T2DM risk. These associations were significant among low SES women but not for intermediate or high SES women. For low SES women, 36.42 % of diabetes cases were attributable to per-child lactation duration of <
24 months, and 5.76 %, 25.37 %, 47.29 %, 19.04 % of cases would be prevented if women lactating for 0, 0-6, 6-12, and 12-24 months extended their per-child lactation duration to at least 2 years. CONCLUSION: Longer lactation duration is associated with a decreased risk of T2DM among women, especially those with low SES. The promotion of extended breastfeeding could potentially prevent a significant proportion of diabetes events.