Background-Thoracentesis and antibiotics remain the cornerstones of treatment in stage I empyema. The management of disease progression or late presentation is controversial. Open thoracotomy and decortication is perceived to be synonymous with protracted recovery and prolonged hospitalisation. Advocates of thoracoscopic adhesiolysis cite earlier chest drain removal and hospital discharge. This paper challenges traditional prejudice towards open surgery. Methods-A five year audit of empyema cases referred to a regional cardiothoracic surgical unit analysing previous clinical course, surgical management, and outcome. Results-Between February 1992 and February 1997, the number of referrals to this centre increased dramatically. Twenty two children were referred for surgery (15 boys, seven girls; age range, 0.5-16 years). Before referral, patients had been unwell for 6-50 days (median, 15), had been treated with several antibiotics, and had undergone chest ultrasound (15 patients), computed tomography (five patients), pleural aspiration attempts (13 patients), and intercostal drainage (seven patients). The organism responsible was identified in only two cases (Streptococcus pneumoniae). Three patients had intraparenchymal abscess formation. Eighteen patients underwent open thoracotomy and decortication. Drain removal was performed on the first or second day. Fever resolved within 48 hours. Median hospital stay was four days. All patients had complete clinical and radiological resolution. Conclusions-Treatment must be tailored to the disease stage. In stage II and III diseases, open decortication followed by early drain removal results in rapid symptomatic recovery, early hospital discharge, and complete resolution. In the early fibrinopurulent phase, alternative strategies should be considered. However, even in ideal cases, neither fibrinolysis nor thoracoscopic adhesiolysis can achieve more rapid resolution at lower risk.

Reoperative cardiac surgery is associated with substantial morbidity and mortality due to technical problems at sternal reentry, which can result in laceration of the right ventricle, innominate vein injury, or embolization from patent grafts. To minimize the risk associated with reentry, we adopted the method of assisted venous drainage in the cardiopulmonary bypass circuit with peripheral cannulation for cardiac reoperations. From March 1999 to May 2003, a series of 52 patients (38 males; mean age 48.7 years, range 4 months to 78 years) underwent cardiac reoperations performed with centrifugal pump venous-assisted cardiopulmonary bypass. EuroSCORE was 7.34 +/- 3.9 (range, 4-19). The reoperations were coronary artery bypass graft (25 patients), valve replacement/repair (18 patients), and complex pediatric procedures (11 patients). The studied adverse events were structural damage at reentry, mortality, blood loss, stroke, and hemolysis. Complications at sternotomy were damage to the innominate vein (1 patient) and aorta (1 patient) with blood loss of 625 and 225 mL, respectively. Four patients required intraaortic balloon pump or extracorporeal membrane oxygenation (n = 1) for hemodynamic support on weaning off cardiopulmonary bypass. Three patients died in the postoperative period. Our experience with centrifugal pump-assisted venous drainage in cardiac reoperations has shown excellent results, with reduced risk of damage to vital structures on sternal reentry. In cases in which structural damage did occur, blood loss was minimal.