237 Similar prognostic abilities of different modes of calculation of the ankle-brachial index to predict mortality in patients with coronary artery disease. A systematic prospective study
Résumé
BackgroundThe traditional method of calculation of the ankle-brachial index (ABI) and the usual 0.90 threshold have been recently challenged by alternate methods and values proposed as more sensitive to predict cardiovascular mortality. However the overall discriminative value of the ABI using different modes of calculation has not been systematically performed.MethodsWe assessed the ABI to predict the 5-years mortality of 705 patients who underwent coronary bypass surgery. We compared the ROC curves of 6 modes of ABI calculation, combining 3 different numerators (the highest (high), the mean (mean) and the lowest (low) pressure of posterior tibial and anterior tibial arteries) with 2 denominators (the highest (high) or the mean (mean) systolic pressure of both arms).ResultsDuring follow-up, 97 (13.8%) patients died. The area under the curves (AUC) of the 6 methods to predict 5-years mortality were similar (Table). Only the AUC of method “High/High” was significantly (p<0.05) higher than method “High/Mean” and “Low/Mean”. The most accurate threshold to predict mortality differed for each method.ConclusionEach mode of ABI calculation presents its own optimal cut-off value to predict mortality. The use of the single 0.90 threshold to compare different modes of ABI calculation is inappropriate. The ABI mode of calculation and the threshold should be considered altogether. The ability of the ABI to predict mortality is almost similar among different modes of calculation.