Joint Effects of Physical Activity and Body Mass Index on Prevalent Diabetes in a Nationally Representative Sample of 1.9 Million US Adults.

Aim: To investigate the joint effects of physical activity (PA) and body mass index (BMI) on prevalent diabetes mellitus in a nationally representative sample of US adults. Materials and Methods: Data were pooled from five US Behavioral Risk Factor Surveillance System (BRFSS) surveys from 2011 to 2019. Cross-sectional associations between independent and combined PA and BMI status and diabetes were analysed using Poisson's log-linear regression with a robust-error variance, reported by adjusted prevalence ratios (APRs). These models were adjusted for relevant sociodemographic, behavioral, and health-related factors. Results: Data was available for 1,913,732 individuals (≥ 18 years). Considering individuals highly active and with normal weight as the reference group, there was an association between decreasing levels of PA and increasing BMI and diabetes prevalence. APRs ranged from APR = 1.09 (nonactive, normal weight group; 95% CI = 1.09-1.09), 1.67 (nonactive, overweight group; 95% CI = 1.67-1.67), 2.23 (nonactive, Class I obesity group; 95% CI = 2.23-2.23), 2.71 (nonactive, Class II obesity group; 95% CI = 2.71-2.71), and 3.17 (nonactive, Class III obesity group; 95% CI = 3.16-3.17). Conclusions: BMI appears to be a substantially larger predictor of diabetes compared to PA in a large population-level sample of US adults. PA provided modest reductions in the prevalence of diabetes but did not attenuate the detrimental impact of overweight and increasing levels of obesity on diabetes prevalence.