To the Editor:-The recent article by Thille et al. 1 and the accompanying editorial by Rouby 2 discuss the effects of positive end-expiratory pressure (PEEP) on pulmonary recruitability in patients with adult respiratory distress syndrome (ARDS).Thille et al. found that alveolar recruitment was similar in pulmonary and extrapulmonary ARDS and concluded that PEEP levels should not be determined based on cause of ARDS. 1 However, the type of ARDS could not be classified in 37% of the patients in their retrospective study, 1 an important point that is further discussed in Rouby's editorial. 2 In the study by Thille et al., the response to PEEP was assessed by pressure-volume curves only, and the effect of PEEP on oxygenation is not reported.More than 30 yr ago, we described the variable effect of PEEP on oxygenation in patients with ARDS. 3,4 We found that both the shortterm and the long-term effects of PEEP on oxygenation response in patients with sepsis were statistically smaller compared with that in patients without sepsis, and we concluded that ARDS associated with sepsis seems to be the result of a more severe pulmonary insult. 3 In a subgroup of trauma patients, we found again that the improvement in oxygenation after the application of PEEP in patients who developed sepsis was significantly smaller than that of patients without sepsis and without lung contusion. 4 As imperfect as they are by today's standards, these two very old studies do have relevance to the ongoing discussion about the effectiveness of pulmonary recruitment in ARDS because, in contrast to Thille et al.'s findings, they support the view that the effectiveness of PEEP does depend on the etiology of ARDS.It is unfortunate that for too many years we have not made much progress in better defining the various subgroups of ARDS.The clinician at the bedside is still left with one option only, namely, apply PEEP and assess its consequences.In fact, the response to PEEP may help in making the diagnosis.