Background: It has been previously shown that physical activity lowers the risk of incident diabetes in American Indians. However, in this population at high risk for cardiovascular (CV) events, it remains to be fully explored whether physical activity reduces CV death in subjects with diabetes and whether this effect is equal between genders.
Objectives: To determine the inter-relationship between activity, diabetes, gender and CV death, we tested the hypotheses that, (i) activity lowers CV death in diabetes, and (ii) the CV benefit of activity is similar across genders.
Methods: We examined the association between activity (leisure-time plus occupational) and CV death among 1,896 diabetic adults (mean age 56.9 ± 8) in phase I of Strong Heart Study (1989-1991. Of these, 1,254 were female (mean age 57.4 ± 8) and 642 were male (mean age 55.9 ± 8). Activity was measured as METS-hours per week on a validated survey. Regression models were used to assess for the prespecified associations and adjust for traditional CV risk factors. Kaplan-Meir analysis was used to plot survival based on activity.
Results: After 14 years of follow-up (1989-1999), there were 222 CV deaths. Compared to diabetic subjects with minimal activity (quartile 1), those with higher activity levels had a lower risk of CV death (p=0.001, Fig 1A). Odds ratios were 0.60 (95% CI: 0.40-0.90; p=0.014), and 0.63 (0.41-0.95; p=0.027) for increasing activity (3rd and 4th quartiles), after adjustment for age and traditional CV risk factors. Moreover, in a subgroup analysis of female subjects with diabetes, increasing physical activity levels was significantly associated with less CV death compared to male subjects with diabetes (Fig. 1B), independent of conventional CV risk factors (p=0.003).
Conclusion: In a large population study of American Indians, physical activity lowered CV death in subjects with diabetes, a finding most pronounced in females. Additional studies are needed to further investigate the pathophysiologic basis of the disparate gender response to activity.