Drs. Schultz and Wolthuis raise questions about the technique for weight and BMI determination in our study. At study enrollment, study personnel recorded patient weights according to medical records. We presume that weight was most commonly recorded as a part of daily clinical care and that it was determined by using bed scales in the participating centers. Study personnel measured patient height to calculate predicted body weight and, subsequently, tidal volume. Although we acknowledge that the weight assessment was not subject to a rigorous protocol, we contend that any systematic measurement errors are unlikely. There is no reason to believe that such inaccuracies were associated with any particular BMI category. In addition, use of BMI as a continuous variable (and lack of an effect in this analysis) reduces the likelihood of unappreciated misclassification of BMI. We cannot account for the patients' weights and BMIs before whatever insult led to acute lung injury. Therefore, we cannot make statements about "healthy" obese persons and their prognoses should they succumb to such injury. Instead, we can only comment on the BMI at the time of study enrollment, which was within 48 hours of the onset of acute lung injury.

Drs. Schultz and Wolthuis also ask if study tidal volumes differed among the BMI categories. We appreciate the opportunity to clarify this point. We examined the mean tidal volume over study days 1 to 3, stratified for treatment assignment. Among those assigned to the lower tidal volume strategy, no significant difference in tidal volume was seen among the BMI categories (6.16 ± 0.71 mL/kg of predicted body weight in the normal BMI group, 6.19 ± 0.78 mL/kg in the overweight BMI group, and 6.18 ± 0.85 mL/kg in the obese BMI group; P > 0.2). Similarly, there was no significant difference among the BMI groups assigned to the conventional strategy (11.85 ± 0.72 mL/kg in the normal BMI group, 11.84 ± 0.73 mL/kg in the overweight BMI group, and 11.86 ± 0.66 mL/kg in the obese BMI group; P > 0.2). However, as mentioned in our article, tidal volume per predicted body weight was higher in obese patients than in patients with normal body weight before study enrollment. This suggests that clinicians were providing larger tidal volumes on the basis of actual weight rather than predicted weight. Such process disparities could explain recent findings of an increased mortality risk among mechanically ventilated obese patients (1).