OBJECTIVE: Practice varies as to whether or not suction is applied to under-water seal drains following lung surgery. We tested the null hypothesis that there is no difference with respect to air leak duration. METHODS: Patients undergoing thoracotomy or video assisted thoracoscopic surgery for lobectomy or wedge resection had either low-pressure suction or no suction applied to their underwater seal bottles postoperatively. Patients were allocated using minimization, a method of unbiased allocation ensuring balance between the arms of a trial with respect to known or suspected confounding factors. The trial was powered for duration of air leak. If an air leak persisted on the 7th post-operative day, the surgeon determined further management. Kaplan-Meier survival analysis of air leak duration and a log rank test were performed on an intention-to-treat basis, with observations censored at 144h (6 complete days). RESULTS: Of the 254 patients that entered the trial, data were available for analysis for 239 (123 no-suction and 116 suction). There was no significant difference in the cumulative persistence of air leaks between the two groups (P=0.62) and inspection of the Kaplan-Meier curves suggests that any difference is negligible. CONCLUSIONS: Applying suction to the underwater seal drains following lung surgery makes no difference in terms of air leak duration. In the light of this finding we have adopted a uniform policy of no suction being applied to the underwater seal, from the time of surgery, unless a specific clinical judgment is made to use it. The anticipated gains are that this will reduce work and cost and aid mobilization.