In the first part of this study, 61 patients admitted for the first episode or the first recurrence of a spontaneous pneumothorax (SP) were randomly treated with thoracic drainage (TD; 28 patients) or with simple needle aspiration (NA; 33 patients). Success rate of therapy was significantly higher with TD than with NA (93%, CI 84 to 100 vs 67%, CI 51 to 83; p = 0.01). Hospital stay was similar between the two groups (7 +/- 4.6 vs 7 +/- 5.6 days), mainly because NA was delayed by 72 h in 26 patients. Recurrence rates at 3 months were 29% (CI 11 to 47%) after TD, and 14% (CI 0 to 29%) after NA (p > 0.20, NS). In the second part of the study, an additional population of 35 patients was treated by immediate NA, with a success rate of 68.5% (CI 53.5 to 83.5%), and a recurrence rate at 3 months of 30% (CI 10 to 50%). Taken together, our results indicate that NA may be proposed as a first-line treatment of SP, with a successful result in two thirds of patients and recurrence in one fifth of patients. In patients who do not heal with NA, a combined risk of TD failure and short-term recurrence of 50% may be an incentive for undelayed surgical procedures.