Skeletal muscle performance and muscle mass are commonly reduced in patients with advanced chronic obstructive pulmonary disease (COPD). It is currently unclear, however, whether negative changes in muscle structure and function are proportionately related to each other in these patients. In a cross-sectional study, 39 patients (post-bronchodilator FEV1=49.7+/-15.5% pred) and 17 controls were submitted to knee isokinetic dynamometry [peak torque (PT), isometric strength (IS), and total work (TW)] and dual energy X-ray absorptiometry for the evaluation of leg muscle mass (LMM). Muscle function (F) was normalised for LMM by using ratio standards (F.LMM-1), power function ratios (F.LMM-b, where b is usually not equal 1), and analysis of covariance (ANCOVA). Patients with COPD presented with reduced PT, IS, TW, and LMM as compared to controls: there were significant linear correlations among these variables in both groups (P<0.05). Ratio standards of PT.LMM-1 and TW.LMM-1 were, on average, 14% lower in patients than controls (P<0.01). The coefficients for allometric correction of IS and TW were significantly higher in patients as compared to controls (0.975 vs. 0.603 and 1.471 vs. 0.824, respectively, P<0.05), i.e. more LMM was needed to generate a given functional output in patients than normal subjects. In addition, adjusted means of muscle function variables by ANCOVA were 11-18% lower for patients than controls with LMM as the covariate (P<0.05). We conclude that factors other than simple atrophy (i.e. mass-independent mechanisms) might play a role in explaining the COPD-related skeletal muscle dysfunction.