Background: To investigate cardiovascular disease mortality associated with longitudinal changes in body weight, and recreational and nonrecreational physical activity during the obesogenic transition in the United States since the 1970’s.
Methods: Data were analyzed from 4,921 individuals aged 25-74 years who participated in the National Health and Nutrition Examination Survey between 1971 and 1979 and follow-up studies to 1992. Mortality was confirmed by searching the National Death Index or proxy interview; clinical data were collected in person. Changes in self-reported recreational- and non-recreational physical activity categories over time were coded as stable, increase, or decrease. Competing-risks regression was used to determine Hazard Ratios (HRs) adjusted for covariates. A logit model in a generalization method was used to explore mediation effects of change in body weight.
Results: Compared with the ‘moderate-vigorous stable’ group (reference), individuals who remained inactive (‘inactive stable’) or reduced their participation in physical activity (‘active to inactive’) experienced the highest mortality, with a 50-176% and 22-222% relative increased HR for nonrecreational and recreational physical activity, respectively, across all models adjusted for covariates. This corresponded to significant loss of life (up to 3 years; all p<0.05). Individuals who became active (‘inactive to active’) were not at increased risk. We found weak (but non-statistically significant) evidence of mediation effects of body weight change on mortality.
Conclusions: Longitudinal changes in moderate-vigorous recreational and nonrecreational physical activity were important predictors of cardiovascular disease mortality during the obesogenic transition period in the US and were mostly unexplained by changes in body weight.