Video-assisted thoracic surgery has been widely used in the treatment of spontaneous pneumothorax despite a paucity of data regarding the relative safety and long-term efficacy for this procedure. We reviewed 113 consecutive patients (68 male and 45 female patients, aged 15 to 92 years, mean 35.1) who underwent 121 video-assisted thoracic surgical procedures during 119 hospitalizations from 1991 through 1993. Recurrent ipsilateral pneumothorax was the most frequent indication for surgery and occurred in 77 patients (65%). The most common method of management was stapling of an identified bleb in the lung, which was undertaken in 105 (87%) patients. No operative deaths occurred. Complications included an air leak lasting longer than 5 days in 10 (8%) patients, two of whom required second procedures for definitive management. No episodes of postoperative bleeding or empyema occurred. The postoperative stay ranged from 1 day to 39 days (median 3 days, average 4.3 days) and 99 patients (84%) were discharged within 5 days. Mean follow-up was 13.1 months and ranged from 1 to 34 months. Eleven patients (10%) were lost to follow-up. Ipsilateral pneumothorax recurred after five of 121 procedures (4.1%). Twelve perioperative parameters (age, gender, race, smoking history, site of pneumothorax, severity of pneumothorax, operative indications, number of blebs, site of blebs, bleb ablation, method of pleurodesis, and prolonged postoperative air leak) were entered into univariate and multivariate analysis to identify significant independent predictors of recurrence. The only independent predictor of recurrence was the failure to identify and ablate a bleb at operation, which resulted in a 23% recurrence rate versus a 1.8% rate in those with ablated blebs (p < 0.001). These data suggest that video-assisted thoracic surgery is a viable alternative to thoracotomy for the treatment of recurrent spontaneous pneumothorax. It results in a short hospital stay, low morbidity, high patient acceptance, and a low rate of recurrence.