Videothoracoscopy versus thoracotomy for the treatment of spontaneous pneumothorax: a propensity score analysis

Few randomized controlled trials have been published on outcomes after treatment of spontaneous pneumothorax. The objective of this study was to assess recurrence, pulmonary complications, prolonged air leak, and hospital duration of stay in patients undergoing videothoracoscopic surgery (VATS) or thoracotomy for spontaneous pneumothorax.

From January 2005 to December 2012, 7,396 patients underwent operations for spontaneous pneumothorax and were entered into the French national database. The propensity score, which is the conditional probability of assignment to a particular treatment given a vector of observed covariates, was used for the analysis. Three statistical analyses were performed: matching, subclassification, and the inverse probability of treatment weighting. The primary end point was recurrence, defined as a pneumothorax requiring a chest tube or new operation. The secondary end point was pulmonary complications, prolonged air leak, and hospital duration of stay.

VATS was performed in 6,419 patients and thoracotomy in 997 patients. Pleurodesis was performed by abrasion or pleurectomy in 5,873 patients (79%) and by using a chemical agent in 1,523 patients (21%). The median time to recurrence was 3 months (range, 1 to 76 months). The recurrence rate was higher in the VATS group regardless of the statistical analysis that was used: 2.1 for unmatched samples, 2.5 for matched samples, 2.3 for subclassification, and 1.7 for the inverse probability of treatment weighting. VATS significantly reduced the hospital duration of stay by 1 day but did not significantly reduce pulmonary complications or prolonged air leak.

VATS reduced the hospital duration of stay, but the risk of recurrence was higher. This information should be delivered to patients before pneumothorax operations.