We would like to thank Dr. Haranath and colleagues for their helpful comments. Malnutrition, as quantified by body mass index (BMI), is accepted as an important predictor of long-term survival in advanced respiratory diseases,1–2 particularly COPD.3Consequently, BMI became part of a multidimensional scoring system that is in widespread use.4Our finding of an association between BMI and mortality in chronic hypercapnic respiratory failure5is fully in line with the literature, and malnutrition is especially prevalent in these patients.6–7 As correctly pointed out by Dr. Haranath and mentioned by us, functionally active fat-free mass might better reflect the nutritional state than BMI.8–9 We also agree that fluid retention could have influenced the predictive value of BMI, particularly on initial admission. There are, however, several arguments to be considered. First, with regard to therapy and/or signs for cor pulmonale, our population was fairly homogeneous (74% diuretics), and in the diuretics subgroup BMI was also predictive. Second, as suggested, we also have performed analyses of patients in a stable clinical state at routine clinical follow-up. This study10 demonstrated BMI as a predictor, similar to previous results.1–4 Third, patients with severe edema (and cor pulmonale) are likely to be at higher risk for death; we thus would have underestimated, not overestimated, the risk by using BMI. Therefore, the association of BMI with long-term mortality is a consistent finding, and BMI is certainly a good proxy for nutritional state in clinical routine.