Previous research found that in noncritically ill patients, thoracocentesis has an unpredictable effect on oxygenation, possibly due to re-expansion pulmonary edema and systemic hypotension. The authors performed a retrospective analysis to study the effect of tube thoracostomy on oxygenation in ICU patients, and the complications associated with it. The authors reviewed the charts of 58 ICU patients in whom 74 procedures were performed. Demographic data, APACHE II score, and indication for thoracocentesis were retrieved from the patient's file. The P(a)O(2)/FiO(2) ratio was calculated before, 12, 24, and 48 hours after tube thoracostomy. P(a)O(2)/FiO(2) ratios at the mentioned time intervals were compared using 1-way analysis of variances (ANOVA) with repeated measures. Logistic regression analysis was used to identify factors associated with a good response to treatment. Age of the patients was 53 +/- 19.0 years (range, 17-88), APACHE II score was 21 +/- 8.3 (range, 6-38), and median length of stay was 13.5 days (interquartile range, 7-25). The volume drained during the first 24 hours was 1077 +/- 667 ml. P(a)O(2)/FiO(2) ratio was 185 +/- 79.3 before chest drainage, 197 +/- 79.1 at 12 hours, 217 +/- 88.9 at 24 hours, and 233 +/- 99.8 at 48 hours. In only 54% of the procedures, a response to therapy was present. Multivariate analysis identified a P(a)O(2)/FiO(2) below 180 to be independently associated with improvement in oxygenation. At 24 and 48 hours, the P(a)O(2)/FiO(2) ratio was significantly higher than before drainage (P <.001). There were 13 complications in 11 procedures (14.9%). The authors' results suggest that tube thoracostomy can be used as an adjunct in the treatment of selected patients with hypoxemic respiratory failure in the ICU. A low P(a)O(2)/FiO(2) seems to be a good predictor of response to therapy. However, the complication rate is considerable, especially in patients with a prolonged ICU stay.