Twenty patients undergoing a posterolateral thoracotomy for lung resection or a nonpulmonary procedure were divided into four groups. Group 1 was the control group. Patients in Group 2 had an intercostal nerve block at the time of closure. Those in Group 3 underwent a continuous intercostal nerve block for five days. Electronic pain control was used in Group 4. An additional group of patients underwent operation through an anterolateral thoracotomy (Group 5) and was compared with the control group. Breathing performance was evaluated daily for five days with bedside spirometry, and intergroup comparison was done utilizing the unpaired t test and analysis of variance. Forced expiratory volume in one second, expressed as percent of preoperative values, was significantly better in Group 3 (continuous intercostal nerve block) at 52.4 +/- 9.2% (standard deviation; p less than 0.05) and in Group 5 (anterolateral thoracotomy) at 52.0 +/- 7.5% (p less than 0.05) than in the control group (38.4 +/- 8.8%) five days postoperatively. It is concluded that bedside spirometry is a simple and reliable technique to assess postoperative changes in ventilatory mechanics due to pain. The pain that follows posterolateral thoracotomy can be substantially decreased with a continuous intercostal nerve block. Anterolateral thoracotomy is notably less painful than posterolateral thoracotomy and should be considered the approach of choice for patients with decreased pulmonary reserve who undergo uncomplicated pulmonary resection.