Background Our purpose was to evaluate the reliability, validity, and responsiveness of the 6-minute walk test (6MWT) in patients with heart failure (HF) enrolled in the Randomized Evaluation of Strategies for Left Ventricular Dysfunction (RESOLVD) pilot study. Methods A total of 768 patients was enrolled in a multicenter randomized clinical trial evaluating the effect of candesartan, enalapril, and metoprolol on left ventricular ejection fraction (LVEF), 6MWT distance, neurohormones, and quality of life. The 6MWT was performed once at screening and twice at baseline, 18 weeks, and 43 weeks by a standardized method. Results Test-retest reliability at baseline (intraclass correlation coefficient [ICC] = 0.90), 18 weeks (ICC = 0.88), and 43 weeks (ICC = 0.91) was very good. Baseline 6MWT distance was weakly inversely correlated to the quality-of-life cumulative score (r = –0.26, P =.0001) and moderately inversely correlated to the New York Heart Association functional classification (NYHA-FC) (r = –0.43, P =.001). In the RESOLVD study, the 6MWT was not responsive to change when effect sizes and standardized response means were used. Disease-specific quality of life was responsive to change in patients treated with candesartan and enalapril and NYHA-FC was responsive to change in the candesartan and enalapril combination and for enalapril alone with small effect sizes. The 6MWT, NYHA-FC, and quality of life were not responsive to change during the metoprolol or placebo phase. Conclusions The 6MWT is highly reproducible in patients with symptoms of HF. It is somewhat correlated to NYHA-FC and quality of life. Overall, quality of life was most responsive to change, whereas 6MWT and NYHA-FC were comparable but less responsive to change in the RESOLVD study. (Am Heart J 2001;142:698-703.)