Weight change and risk of developing type 2 diabetes.

To examine the association between weight change and risk of type 2 diabetes and whether initial weight modifies the association. This is a prospective cohort study of 20,187 alumni from Harvard University and the University of Pennsylvania. At baseline in 1962 or 1966, men (mean age, 45.9 years) reported their weight, height, and other risk factors. They also had had their weight and height measured at university entry (mean age, 18.5 years). Participants were followed from baseline to 1998 for type 2 diabetes. During follow-up, 1223 men developed type 2 diabetes. Weight gain significantly increased the risk of this disease. The multivariate relative risks associated with BMI change from university entry to baseline of <-0.5, +/-0.5, >0.5 to 1.0, >1.0 to 1.5, >1.5 to 2.0, >2.0 to 3.0, and >3.0 kg/m(2) per decade were 0.88, 1.00 (referent), 1.29, 2.09, 2.69, 4.67, and 7.02, respectively (p for trend < 0.0001). Even among men with a low initial BMI < 21 kg/m(2), weight gain significantly increased risk; the corresponding relative risks were (no cases), 1.00 (referent), 1.00, 1.93, 2.47, 4.82, and 7.68, respectively (p for trend < 0.0001). A low initial BMI does not ameliorate the increase in risk of type 2 diabetes with weight gain. Avoidance of weight gain, even among lean individuals, is important to reduce the risk of this disease.