The study of Budweiser et al4confirms the important role of nutritional status in COPD. Previous work5in patients with COPD, largely treated with long-term home oxygen therapy, have demonstrated that male gender, older age, lower BMI, FEV1 percentage of predicted, Pao2, and Paco2 were identified as independent negative prognostic factors. In another study6 based on the ANTADIR observatory, it was demonstrated that the effect of BMI on survival was independent of age, FEV1, Pao2 and sex. Lower BMI was also a powerful predictor of duration and rate of hospitalization.,6 BMI in the study of Chailleux et al6 was significantly lower than in the study by Budweiser et al4: mean BMI, 23 kg/m2 vs 27.2 kg/m2. In the study by Chailleux et al,,668.1% of the COPD patients had a BMI < 24 kg/m2. Studies8 have reported that body composition assessment provides information in addition to BMI in different categories of COPD and that fat-free mass corrected for body surface is an independent predictor of mortality irrespective of fat mass in COPD. It would be interesting to evaluate in future studies body composition besides BMI in patients with chronic respiratory failure because depletion in fat-free mass can be found even in patients with normal or high BMI.9 A BMI of 25 kg/m2, as used by Budweiser et al,,4 was reported by Schols et al10 as a threshold value below which the mortality risk was clearly increased.