Empyema continues to be an uncommon, frustrating, and potentially lethal complication of pneumonectomy. Between 1990 and 1994 we treated 16 cases of recalcitrant postpneumonectomy (partial or total) empyema with combinations of pulse lavage, sharp debridement, muscle flaps, myodermal flaps, and thoracoplasty. We performed 11 pectoralis muscle flaps, 6 serratus anterior muscle flaps, 9 latissimus dorsi muscle flaps, 6 rectus abdominis muscle flaps, and 1 trapezius muscle flap for an average of 2.1 muscle flaps per patient. There was 1 omental flap. Of these flaps, 2 were free and the rest pedicled. Ten of the muscle flaps carried deepithelialized cutaneous paddles, and 6 were larger than 150 cm3. Thoracoplasty was done in 11 patients to decrease the volume of the postpneumonectomy empyema cavity. Of 16 patients, 4 failed initially because of persistent bronchopleural fistula or infection but resolved after one additional procedure. There was 1 perioperative death, 3 reoperations for bleeding, 1 patient with upper extremity deep vein thromboses, 1 seroma, and 1 patient with significant postoperative pain syndrome. In order to determine the efficacy of different operative approaches, patients were retrospectively divided into two groups according to the number of operations using flaps needed to resolve their postpneumonectomy empyema. Group A required only one operation using flaps to eliminate the postpneumonectomy empyema. Group B required two operations using flaps to remedy the postpneumonectomy empyema. Group B operations were further classified into B1, for the first operation, and B2, for the second operation. No patient needed more than two operations using flaps. Three significant variables were identified, the number of muscle flaps, the number of ribs in any thoracoplasty, and the preoperative serum albumin level. The A and B2 groups had significantly more muscle flaps transposed (p = 0.006) and ribs resected (p = 0.0002) than the B1 group. These findings suggest that filling the postpneumonectomy empyema space with muscle and collapsing any remaining space by thoracoplasty were the most successful strategy. The B2 group's average albumin level was significantly higher (p = 0.03) than that in either the A or the B1 group, suggesting that improved nutrition may have played a role in the lack of recurrence. Our goals of single-stage closure and decontamination of empyema cavities were best achieved by following these principles: removal of infected and necrotic tissue using sharp debridement and pulsed lavage, repair of bronchopleural fistulas with muscle flaps, and minimization of the dead space with combinations of muscle flaps and thoracoplasty.