Graphical abstract Display Omitted

Highlights • Injuries, including vehicle accidents and falls, counted for 71% of unnatural deaths • A risk of dying from injury was linearly and reversely associated with the BMI level • An overweight adult had a risk 43% lower than that of a normal weight adult • An obese adult had a risk 63% lower than that of a normal weight adult

Abstract Objective The purpose is to exam whether excess body weight is associated with an increased risk of death from unnatural causes, particularly, injury. Method We analyzed nationally representative data of 14,453 adults (19 and older) who participated in the third National Health and Nutrition Examination Survey, 1988–1994, and were followed up with vital statuses through December 31, 2006. We used Cox proportional hazard regression to estimate the hazard ratio (HR) of death from all unnatural causes combined and specific ones. Gray’s test was performed to assess the equality of cumulative incidence functions between body mass index (BMI) levels. Results A total of 128 unnatural deaths were recorded during an 18-year follow-up with 193,019 person-years accumulated. Compared with healthy weight participants, a person with excess body weight had a low hazard of death from unnatural causes [HR=1.00 (reference), 0.58 (0.39–0.87), and 0.50 (0.30–0.82) for healthy weight, overweight and obese participants, respectively]. Injuries, including motor vehicle accidents and falls, were the major types of unnatural deaths (n =91, 71% of all unnatural deaths), and the risk of death from injuries was linearly and reversely associated with BMI. The HRs of injury were 1.00 (reference), 0.57 (0.36–0.91), and 0.36 (0.19–0.69) for healthy weight, overweight and obese participants, respectively. All these estimates were obtained after adjustment of socio-demographic variables. Discussion Excess body weight appears to be associated with a low risk of death from unnatural causes, particularly, injuries. Additional investigations on the mechanism underlying the relationship between BMI and unnatural deaths are warranted.