Malnutrition is associated with poor prognosis in patients with chronic obstructive pulmonary disease (COPD). Body weight is not a reliable evaluation criterium. Body composition which is more useful can be determined routinely using 2 techniques: skinfold thickness anthropometry (Ant) and bioelectrical impedance analysis (BIA). The validity of this last technique has not been demonstrated in patients with COPD. Fat-free mass (FFM) in 58 patients (51 men, 7 women) with stable COPD (FEV1 < 50% of predicted value) was assessed using the 4-skinfold-thickness method (Ant) and BIA (Imp). Statistical analysis included correlation analysis, intraclass correlation coefficient, and the Bland and Altman analysis. Imp-FFM and Ant-FFM correlated well (r = 0.920; p < 0.0001). Intraclass correlation coefficient was high (rI = 0.9065). However, the values were scattered and there was a systematic bias (significant linear regression between the difference in estimates obtained by the 2 methods and the means). As anthropometric measurements are not reliable in the elderly patients, our results suggest that BIA could be a useful tool to determine FFM in patients with COPD. Its validity still has to be tested against a reference method.