Body Mass Index and Mortality in Individuals With Type 1 Diabetes

Context The relationship between body mass index (BMI) and mortality may differ between patients with type 1 diabetes and the general population; it is not known which clinical characteristics modify the relationship. Objective Our aim was to assess the relationship between BMI and mortality and the interaction with clinically meaningful factors. Design, Setting, and Participants This prospective study included 5836 individuals with type 1 diabetes from the FinnDiane study. Main Outcome Measure and Methods We retrieved death data for all participants on 31 December 2015. We estimated the effect of BMI on the risk of mortality using a Cox proportional hazards model with BMI as a restricted cubic spline as well as effect modification by adding interaction terms to the spline. Results During a median of 13.7 years, 876 individuals died. The relationship between baseline BMI and all-cause mortality was reverse J-shaped. When analyses were restricted to those with normal albumin excretion rate, the relationship was U-shaped. The nadir BMI (BMI with the lowest mortality) was in the normal weight region (24.3 to 24.8 kg/m2); however, among individuals with diabetic nephropathy, the nadir BMI was in the overweight region (25.9 to 26.1 kg/m2). Diabetic nephropathy, diabetes-onset age, and sex modified the relationship between BMI and mortality (Pinteraction < 0.05). Conclusions Normal weight is optimal for individuals with type 1 diabetes to delay mortality, whereas underweight might be an indication of underlying complications. Maintaining normal weight may translate into reduced risk of mortality in type 1 diabetes, particularly for individuals of male sex, later diabetes-onset age, and normal albumin excretion rate.