BMI and All-Cause Mortality in Normoglycemia, Impaired Fasting Glucose, Newly Diagnosed Diabetes, and Prevalent Diabetes: A Cohort Study

OBJECTIVE This study examined associations between BMI and mortality in individuals with normoglycemia, impaired fasting glucose (IFG), newly diagnosed diabetes, and prevalent diabetes and identified BMI ranges associated with the lowest mortality in each group. RESEARCH DESIGN AND METHODS A total of 12,815,006 adults were prospectively monitored until 2013. Diabetes status was defined as follows: normoglycemia (fasting glucose <100 mg/dL), IFG (100-125 mg/dL), newly diagnosed diabetes (>= 126 mg/dL), and prevalent diabetes (self-reported). BMI (kg/m(2)) was measured. Cox proportional hazards model hazard ratios were calculated after adjusting for confounders. RESULTS During a mean follow-up period of 10.5 years, 454,546 men and 239,877 women died. U-shaped associations were observed regardless of diabetes status, sex, age, and smoking history. Optimal BMI (kg/m2) for the lowest mortality by group was 23.5-27.9 (normoglycemia), 25-27.9 (IFG), 25-29.4 (newly diagnosed diabetes), and 26.5-29.4 (prevalent diabetes). Higher optimal BMI by worsening diabetes status was more prominent in younger ages, especially in women. The relationship between worsening diabetes status and higher mortality was stronger with lower BMI, especially at younger ages. Given the same BMI, people with prevalent diabetes had higher mortality compared with those with newly diagnosed diabetes, and this was more striking in women than men. CONCLUSIONS U-curve relationships existed regardless of diabetes status. Optimal BMI for lowest mortality became gradually higher with worsening diabetes for each sex and each age-group.