Changes in body mass index and waist circumference and concurrent mortality among Swedish women.

OBJECTIVE

Most studies on obesity and mortality use a single anthropometric measure. Less is known about the effects of weight change on mortality. This study examined changes in body mass index (ΔBMI) and waist circumference (ΔWC) and subsequent all-cause and cause-specific mortality.

METHODS

The study was conducted in the Women's Lifestyle and Health cohort, using self-reported anthropometric measures from 1991 to 1992 and 2003. Hazard ratios of mortality and 95% confidence intervals were calculated using Cox proportional hazards models. ΔBMI and ΔWC were examined in quartiles of absolute and relative change, with the second quartile (moderate gain) as the reference.

RESULTS

There was a higher risk of death in the first quartile of relative ΔBMI: HR 1.28 (1.04-1.56). Absolute ΔBMI suggested the same pattern, but the result was nonsignificant. ΔWC was not associated with mortality. In cause-specific analyses, the association remained significant for cancer mortality only. In sensitivity analyses excluding the first 5 years of follow-up, the association was, however, attenuated.

CONCLUSIONS

This study found a higher risk of death among women in the first quartile of relative ΔBMI compared with the second. It was driven by cancer mortality but may be ascribed to reverse causality. ΔWC was not associated with mortality.