Abstract
CONTEXT: Recent studies have demonstrated a strong association between cardiorespiratory fitness (CRF) and mortality, but bias due to differences in the distribution of baseline variables has not been adequately considered. We studied a cohort of veterans with and without Type-2 diabetes using a propensity score matching method. METHODS: Males with (n=592) and without (n= 6,167) Type-2 diabetes were studied. Propensity scores were used to balance covariate distributions between groups with and without Type-2 diabetes. All-cause mortality was the end point. RESULTS: Predictors of mortality included hypertension, smoking, Type-2 diabetes, BMI and CRF. For each 1 MET increase in CRF in the unmatched group, the adjusted HR was 0.83 in those with diabetes (95% CI 0.77-0.89; p<0.0001) compared to 0.87 in those without diabetes (95% CI 0.86-0.89; p<0.0001). Similar trends were observed for the matched dataset: the adjusted HRs were 0.83 (95% CI 0.77-0.90; p<0.0001) and 0.88 (95% CI 0.82-0.94; p<0.0001) for those with and without diabetes, respectively. CONCLUSIONS: CRF is a strong predictor of mortality in veterans with and without Type-2 diabetes. Although the trend in the association between CRF and all-cause-mortality was similar for matched and unmatched data, the mortality risks were relatively inflated when using unmatched data.