LAUSR

AIMS The aim of this study was to determine the ability to predict all-cause mortality using established per cent-predicted (%PRED) equations for peak oxygen consumption (VO2peak) estimated by a submaximal walk test in outpatients with cardiovascular disease. METHODS Male patients (N = 1491) aged 62 ± 10 years at baseline underwent a moderate and perceptually regulated (11-13 on the 6-20 Borg scale) 1-km treadmill-walking test to estimate VO2peak. %PRED was derived from the Fitness Registry and the Importance of Exercise: A National Data Base (FRIEND) and the Wasserman/Hansen equations. RESULTS There were 215 deaths during a median 9.4-year follow-up. The FRIEND prediction equation provided better prognostic information with receiver operating curve analysis showing significantly different areas under the curve (0.72 and 0.69 for the FRIEND and the Wasserman/Hansen equations respectively, p = 0.001). Overall mortality rate was higher across decreasing tertiles of %PRED using FRIEND, with 26%, 11% and 5% for the least fit, intermediate and high fit tertiles, respectively (p for trend < 0.0001). Compared with the least fit tertile, the adjusted hazard ratios for the second and third tertiles were 0.54 (95% confidence interval 0.34-0.87, p = 0.01) and 0.45 (95% confidence interval 0.25-0.81, p = 0.008), respectively. Each 1% increase in %PRED conferred a 3% improvement in survival (p = 0.0004). CONCLUSION Low %PRED VO2peak in cardiac outpatients determined by the FRIEND equation was associated with a high mortality rate independent of traditional cardiovascular risk factors and clinical history. The FRIEND equation may provide a suitable normal standard when applied to clinically stable outpatients with cardiovascular disease.