Study objective: To identify the prognostic factors for pneumothorax in patients in the ICU.
Design: Retrospective cohort study.
Setting: ICU at a university-based teaching hospital.
Patients and methods: Sixty patients developed pneumothoraces in the ICU during a period of 36 months. Medical records relating to patients' age, sex, underlying diseases, associated medical conditions, reasons for admission, acute physiology and chronic health evaluation (APACHE) II scores, procedures performed before the development of pneumothorax, occurrences of tension pneumothorax, duration of chest tube placement, chest tube removal, duration of ICU stay, and patient outcomes all were analyzed. A multivariate logistic regression model was applied with variables that were significantly associated with survival in the univariate analysis. The probabilities of chest tube removal were calculated using the Kaplan-Meier method.
Results: Thirty-five patients (58%) had procedure-related pneumothoraces. The procedure that most commonly caused pneumothoraces was thoracentesis (n = 19; 54%), followed by central vein/pulmonary artery catheterization (n = 14; 40%) and bronchoscopy/transbronchial lung biopsy (n = 8; 23%). A multivariate logistic regression analysis also showed that pneumothorax due to barotrauma (p = 0.001), tension pneumothorax (p = 0.0023), and concurrent septic shock (p = 0.0476) were significantly and independently associated with death. The log-rank test revealed that the success rate of chest tube removal was higher in patients with procedure-related pneumothoraces (p = 0.0055).
Conclusions: Patients with procedure-related pneumothoraces had better outcomes. Patients with pneumothoraces occurring in the ICU due to barotrauma, or a complicating tension pneumothoraces, carry a higher risk of mortality.