Aims/hypothesis Earlier age at menarche is associated with increased BMI and obesity risk from early childhood through to adulthood. We hypothesised that earlier age at menarche would also predict subsequent diabetes risk. Methods This was a population-based prospective cohort study of 13,308 women, who were aged 40 to 75 years between 1993 and 1997 and participating in the Norfolk cohort of the European Prospective Investigation into Cancer and Nutrition (EPIC-Norfolk). We used data on age at menarche and ascertained diabetes incidence to 2005. Results There were 734 cases of diabetes (363 incident and 371 prevalent cases). Mean age at menarche was lower in women with diabetes than in non-diabetic women (12.8 vs 13.0 years, p=0.008). Compared with the earliest quintile (menarche at 8-11 years), women in the oldest quintile (menarche at 15-18 years) had lower BMI (25.5 vs 27.4 kg/m 2 , p<0.0001) and a reduced risk of diabetes (OR 0.66 [95% CI 0.51-0.86] adjusted for age, family history, physical activity, smoking, occupational social class, parity and use of hormonal preparations). The association between age at menarche and diabetes was linear (adjusted OR 0.91 [95% CI 0.87-0.96] per 1 year later menarche) and appeared to be completely mediated by adult BMI or waist circumference , further adjusted for BMI at age 40-75 years). Conclusions/interpretation Earlier age at menarche increases the risk of diabetes in women and this association appears to be mediated by increased adiposity. History of earlier menarche may help to identify women with increased subsequent risk of diabetes.