Prognostic value of ankle-brachial index and dobutamine stress echocardiography for cardiovascular morbidity and all-cause mortality in patients with peripheral arterial disease

Background. Peripheral arterial disease (PAD) is associated with an excessive risk for cardiovascular events and mortality. To determine measures prognostic of adverse events, ankle-brachial index (ABI) was compared with dobutamine stress echocardiography (DSE) in patients referred to our vascular center for the evaluation of PAD. Methods. The medical records of consecutive patients referred for the concurrent evaluation of PAD and coronary artery disease (CAD) between 1992 and 1995 were reviewed for subsequent cardiovascular events and death. Results: Among 395 patients (mean age, 69.7 +/- 9.6 years; 40% women), 341 had abnormal ABI and 268 had abnormal DSE (95 fixed and 173 stress-induced wall motion abnormalities). During a mean follow-up of 4.7 years, 27.3% of patients experienced a cardiovascular event, and 39.4% died. By multivariate analysis, ABI provided the strongest prediction of all-cause mortality (hazard ratio [HR], 2.34; 95% confidence interval [CI], 1.36 to 4.05; P = .002). Conversely, DSE with inducible or fixed wall motion abnormalities showed no association with cardiovascular events or increased mortality in multivariate analysis. The only DSE variable independently predictive of mortality was decreased left ventricular ejection fraction (< 50%) at peak stress (HR, 1.70; 95% CI, 1.22 to 2.36; P = .002). Statin and aspirin therapy, but not beta-blockers, were protective. There was no relation between ABI and wall motion index score at rest or after stress. Conclusions. In high-risk patients referred to our vascular center for the evaluation of PAD, the assessment of ABI provided a strong independent prediction of all-cause mortality. Therefore, proper interpretation of this simple, affordable, and reproducible measure extends beyond the assessment of PAD severity. Although a poor left ventricular response to dobutamine was also predictive, other echo variables were not.