Head-to-Head Comparison of the Prognostic Value of Actual Versus Percent Predicted 6-Minute Walk Distance in Patients With Heart Failure.

Abstract

Purpose: The 6-minute walk test is a powerful prognostic tool to predict mortality in heart failure (HF). We sought to assess whether the use of percent predicted 6-minute walk test distance (%6MWD) allows for enhancing the prognostic performance of the test versus actual 6MWD.

Methods: We studied 1425 patients with HF. The primary outcome was all-cause mortality within 1 and 3 years. Predicted 6MWD were determined using the Enright, Duncan, and Morbach equations.

Results: The area under the curve of actual and %6MWD in predicting mortality ranged from 0.713 to 0.737 at 1 year and from 0.646 to 0.692 at 3 years. In multivariable analyses, actual 6MWD and each %6MWD were independently associated with mortality risk both at 1 year and 3 years; the odds ratios ranged from 1.29 to 1.36 at 1 year and from 1.19 to 1.24 at 3 years. When added to the baseline risk model for 1-year mortality prediction, actual 6MWD and each %6MWD yielded a statistically significant, similar improvement in discriminative ability, the proportion of survivors reclassified as lower risk, and, except for %6MWDEnright, the proportion of patients with events reclassified as higher risk. Similar results were observed at 3 years.

Conclusions: Our data indicate that %6MWD, regardless of the equation used to estimate predicted values, is independently associated with mortality risk in HF and provides incremental prognostic information over well-established predictors of death. Its prognostic performance, however, is not superior to that of actual 6MWD.