Abstract
The relative importance of sarcopenia and its individual components as independent predictors of mortality in the dialysis population has not been determined. We estimated whole-body muscle mass using pre-dialysis bioimpedance spectroscopy measurements in 645 ACTIVE/ADIPOSE-enrolled prevalent hemodialysis patients from San Francisco and Atlanta. Low muscle mass was defined as two standard deviations below sex-specific means for young adults from NHANES and indexed to height(2), body weight, body surface area, or body mass index. We evaluated the association of sarcopenia (low muscle mass) by four indexing methods, weak hand grip strength, and slow gait speed with mortality. Seventy-eight deaths were observed during a mean follow-up of 1.9 years. Sarcopenia was not significantly associated with mortality after adjusting for covariates. No muscle mass criteria were associated with death, regardless of indexing metrics. In contrast, having weak grip strength or slow walking speed was associated with mortality in the adjusted model. Only gait slowness significantly improved the predictive accuracy for death with an increase in C-statistic from 0.63 to 0.68. However, both gait slowness and hand grip weakness significantly improved the net reclassification index compared to models without performance measures (50.5% for slowness and 33.7% for weakness), whereas models with muscle size did not. Neither sarcopenia nor low muscle mass by itself was a better predictor of mortality than functional limitation alone in patients receiving hemodialysis. Thus, physical performance measures, including slow gait speed and weak hand grip strength, were associated with mortality even after adjustment for muscle size and other confounders.