Sarcopenia among patients receiving hemodialysis: Weighing the evidence

Background: There is no consensus on how best to define low muscle mass in patients with end-stage renal disease. Use of muscle mass normalized to height-squared has been suggested by geriatric societies but may underestimate sarcopenia, particularly in the setting of excess adiposity. We compared four definitions of low muscle mass in a prevalent hemodialysis cohort. Methods: ACTIVE/ADIPOSE enrolled prevalent patients receiving hemodialysis from the San Francisco and Atlanta areas from June 2009 to August 2011. Whole-body muscle mass was estimated using bioelectrical impedance spectroscopy, performed before a midweek dialysis session (n=645; age 56.7±14.5years, 41% women). We defined low muscle mass as muscle mass of 2SD or more below sex-specific bioelectrical impedance spectroscopy-derived means for young adults (18-49years) from National Health and Nutrition Examination Survey and indexed to height², body weight (percentage), body surface area (BSA) by the DuBois formula, or Quételet's body mass index (BMI). We compared prevalence of low muscle mass among the four methods and assessed their correlation with strength and physical performance. Results: The prevalence of low muscle mass ranged from 8 to 32%. Muscle mass indexed to height² classified the smallest percentage of patients as having low muscle mass, particularly among women, whereas indexing by BSA classified the largest percentage. Low muscle mass/height² was present almost exclusively among normal or underweight patients, whereas indexing to body weight and BMI classified more overweight and obese patients as having low muscle mass. Handgrip strength was lower among those with low muscle mass by all methods except height². Handgrip strength was directly and modestly correlated with muscle mass normalized by percentage of body weight, BSA, and BMI (ρ=0.43, 0.56, and, 0.64, respectively) and less so with muscle/height² (ρ=0.31, P<0.001). The difference in grip strength among patients with low vs. normal muscle mass was largest according to muscle/BMI (-6.84kg, 95% CI -8.66 to -5.02, P<0.001). There were significant direct correlations of gait speed with muscle mass indexed to percentage of body weight, BSA, and BMI but not with muscle mass indexed to height². Conclusions: Skeletal muscle mass normalized to height² may underestimate the prevalence of low muscle mass, particularly among overweight and obese patients on hemodialysis. Valid detection of sarcopenia among obese patients receiving hemodialysis requires adjustment for body size.