Introduction: Weight gain during early adulthood may contribute to diabetes risk above and beyond a single measure in midlife. Nonetheless, most epidemiologic studies have focused on the risks associated with a one body mass index (BMI) value in midlife.
Hypothesis: Accounting for change in BMI from early to mid-adulthood would add prognostic value for estimating risk of diabetes compared to using a single value.
Methods: We included 9,857 participants from the Atherosclerosis Risk in Communities (ARIC) Study without diabetes and with BMI ≥18.5 kg/m2 in 1996-1998 (visit 4). Participants were followed through 2015 for incident diagnosed diabetes. Current BMI was defined as calculated BMI using measured height and weight at visit 4. Early adulthood BMI was calculated from current height and self-reported weight at age 25. We classified current and early adulthood BMI into three categories (18.5-<25, 25-<30, and ≥30 kg/m2) and we examined the associations of changes in BMI category with diabetes risk using Cox proportional hazards models.
Results: Overall, 64% (6,205 of 9,857) of the participants moved from a lower to higher BMI category between age 25 and visit 4 (mean age, 63), 35% (3,446 of 9,857) remained stable, and 2% (206 of 9,857) moved from a higher to lower BMI category. Adding BMI at age 25 to a model with current BMI significantly improved model fit. In analyses using BMI change categories between age 25 and visit 4, relative to persons with normal BMI at age 25 and visit 4, persons who moved from a lower to higher BMI category and those who maintained a high BMI had a higher risk for diabetes (Table).
Conclusions: Current weight is associated with diabetes risk. Past weight gain from early adulthood did not appear to add information above and beyond current weight. Our findings highlight the importance of weight management at all ages for optimal diabetes prevention.