AimsThe association between relative muscle mass (RMM) and incidence of type 2 diabetes (T2DM) is largely unknown. We examined whether RMM predicted development of T2DM in an apparently young healthy population.MethodsThis cohort study was comprised of 113,913 men and 89,854 women, free of T2DM at baseline, who underwent a health checkup examination and were followed-up annually or biennially for an average of 2.9 years. We used skeletal muscle mass index (SMI) as an indicator of RMM. SMI (%) [total skeletal muscle mass (kg)/body weight (kg)×100] was estimated using a bioelectrical impedance analyzer. The study outcome was incident T2DM, defined as fasting serum glucose ≥126 mg/dL, HbA1C ≥6.5%, or use of medication for T2DM.ResultsDuring 589,098.8 person-years of follow-up, 4,264 individuals developed T2DM (incidence rate, 7.2 per 1000 person-years). Median age (range) at baseline was 39.1 years (18.1–87.1). RMM was negatively associated with incidence of T2DM in a dose-response manner. The multivariate-adjusted hazard ratios (95% CIs) for incident T2DM comparing quartiles 3, 2 and 1 of RMM to the highest quartile were 1.32 (1.14–1.52), 1.63 (1.42–1.86), and 2.21 (1.94–2.51), respectively, for males and 1.18 (0.88–1.58), 1.46 (1.11–1.91), and 1.96 (01.51–2.53) for females (P for trend <0.001; 0.011). This association was stronger in younger or premenopausal subjects.ConclusionsRMM was negatively associated with development of T2DM in a large sample of young and middle-aged Korean adults. Further research is required to determine whether preservation of muscle mass through intervention affects the risk of T2DM.