We read with interest the article by Wildman et al.1 They have carried out an interesting study on patients with chronic obstructive pulmonary disease (COPD) or asthma admitted to the Intensive Care Unit (ICU). The authors in their discussion state that their overall results may be least applicable to patients on long-term oxygen (LTO) and/or with low functional scores. Intubated patients with COPD on LTO constitute 10.4% of 394 intubated patients in this study. Only mortality results relating to this population are shown in this article. It would have been interesting to show results such as quality of life at 180 days. If deciding to intubate is difficult in patients with exacerbated COPD, it is more difficult in those patients on LTO. Clinicians tend to be pessimistic about survival and quality of life and they are very selective when admitting these patients. In 1999 we reported on a study2 which focused on patients with exacerbated COPD on LTO who were mechanically ventilated. Mortality was higher than that reported by Wildman et al in this subgroup of patients: 35% (in ICU), 50% (in hospital), 75% (at 1 year) and 85% (at 5 years). Patients who died in hospital and in the first year after discharge had a lower forced expiratory volume in 1 s (FEV₁) than survivors. Although the study was carried out in a University tertiary hospital, the sample size was small (20 patients recruited in 2 years). Our paper is, to date, the only one which focuses on patients with COPD on LTO. The small sample size of patients with COPD on LTO, both in the series of Wildman et al and in our series, supports the statement about the selective criteria considered when admitting these patients. Therefore, how many patients are not admitted who could survive with an acceptable quality of life? In our experience, of the five survivors at 1 year, two had a good functional level previously and both had a functional status similar to that prior to the ICU admission. The mean age of the three patients alive at 5 years was 63.3 years, and one of them was independent. The cost per quality-adjusted life year was US$44.602. Obviously, we have no predictors to help us in the daily decision making regarding these patients although we agree that functional status in the period of stability and FEV₁ before ICU admission can be helpful. However, our decisions can only be supported by the experiences of small studies or data from studies with different main objectives. Therefore, results regarding these patients with severe COPD would be very welcome.