We read with interest the article by Goulenok et al in a recent issue CHEST (April 2004),1showing that obesity, defined by a body mass index (BMI) > 27, was associated with a higher mortality rate among ICU patients than predicted by the simplified acute physiology score (SAPS) II. The authors were surprised by the fact that overweight was not taken into consideration by the SAPS II or APACHE (acute physiology and chronic health evaluation) score. Four other studies2–5 have analyzed the potential impact of obesity on ICU outcome, with conflicting results. Like Goulenok et al,1two of these studies2,5 showed increased mortality in obese ICU patients. However, in one of these studies,2the obese patients had a particularly high mean BMI of 51, and were compared to a group of nonobese patients (BMI < 30). The authors did not give their results for patients with BMI values between 30 and 40. Tremblay and Bandi3found no difference between overweight and severely obese patients in a multi-institutional database of 41,011 ICU patients. Furthermore, Garrouste-Orgeas et al4 found a lower mortality rate in obese ICU patients (BMI > 30) than in nonobese patients. We conducted a prospective, multicenter, case-control study to evaluate the prognostic significance of BMI > 35 in ICU patients receiving mechanical ventilation for > 48 h. Our preliminary results for 69 patients with a mean BMI of 42 ± 6 and a mean (± SD) SAPS II score of 45 ± 15 show that mortality was lower than predicted by SAPS II (16% vs 37%, respectively), with a standard mortality ratio of 0.46. These results disagree with those of Goulenok et al,1 even though they were obtained in patients with higher BMI values and poorer clinical status (SAPS II) on ICU admission. Thus, the impact of obesity in ICU mortality remains controversial, and it seems premature to add BMI to ICU severity scores.