The sequelae of increasing childhood obesity are of major concern. We assessed the association of BMI in late adolescence with diabetes mortality in midlife.
RESEARCH DESIGN AND METHODS
The BMI values of 2,294,139 Israeli adolescents (age 17.4 ± 0.3 years), measured between 1967 and 2010, were grouped by U.S. Centers for Disease Control and Prevention age/sex percentiles and by ordinary BMI values. The outcome, obtained by linkage with official national records, was death attributed to diabetes mellitus (DM) as the underlying cause. Cox proportional hazards models were applied.
During 42,297,007 person-years of follow-up (median, 18.4 years; range <1–44 years) there were 481 deaths from DM (mean age at death, 50.6 ± 6.6 years). There was a graded increase in DM mortality evident from the 25th to the 49th BMI percentile group onward and from a BMI of 20.0–22.4 kg/m2 onward. Overweight (85th to 94th percentiles) and obesity (the 95th percentile or higher), compared with the 5th to 24th percentiles, were associated with hazard ratios (HRs) of 8.0 (95% CI 5.7–11.3) and 17.2 (11.9–24.8) for DM mortality, respectively, after adjusting for sex, age, birth year, height, and sociodemographic variables. The HR for the 50th through 74th percentiles was 1.6 (95% CI 1.1–2.3). Findings persisted in a series of sensitivity analyses. The estimated population-attributable fraction for DM mortality, 31.2% (95% CI 26.6–36.1%) for the 1967–1977 prevalence of overweight and obesity at age 17, rose to a projected 52.1% (95% CI 46.4–57.4%) for the 2012–2014 prevalence.
Adolescent BMI, including values within the currently accepted “normal” range, strongly predicts DM mortality up to the seventh decade. The increasing prevalence of childhood and adolescent overweight and obesity points to a substantially increased future adult DM burden.