Objectives: To determine whether adjustment of muscle mass for height(2) or for body mass represents a more-relevant predictor of physical performance.

Design: Cross-sectional study, using baseline data from a trial comparing upper- and lower-body training.

Setting: Women recruited from the community and gynecological practices in Connecticut.

Participants: One hundred eighty-nine healthy older (aged 67.5 +/- 4.8), active women receiving estrogen for osteoporosis over 2 years.

Measurements: Total and appendicular skeletal muscle (ASM) and fat mass (AFM) were determined using dual x-ray absorptiometry. Physical performance, muscle strength, and fitness measures were obtained at baseline.

Results: Adjusting ASM for height(2) identifies lean women who are sarcopenic according to published standards yet fails to identify overweight and obese women whose ASM adjusted for body mass is low. ASM divided by body mass (ASM/body mass) is a stronger physical performance predictor, explaining 32.5%, 13.5%, 11.6%, 6.3%, and 6.8% of the variance in maximum time on treadmill, 6-minute walk, gait speed, 8-foot walk, and single leg stance, respectively, whereas ASM divided by height in m(2) (ASM/height(2)) explained only 2.9%, 0.2%, 2.0%, 0.04%, and 0.1%. Multivariate modeling demonstrated considerable overlap in aspects of ASM/body mass and AFM/body mass associated with performance, with ASM/body mass dominant. In contrast, ASM/height(2) is a much stronger predictor of leg press 1 repetition maximum and maximum power.

Conclusion: The results suggest that relative sarcopenia with ASM adjusted for body mass is a better mobility predictor, with absolute sarcopenia a better indicator of isolated muscle group function in healthy postmenopausal women receiving estrogen replacement.