The objective of this study was to determine those factors which are predictive of outcome in cases of thoracic empyema that are solely community acquired. All patients admitted with a diagnosis of thoracic empyema in the Auckland region between 1993 and 1995 were reviewed retrospectively. Both clinical and radiological outcomes were determined at 3-6 month follow up. Radiological outcomes were scored on admission, discharge and follow up by blinded chest radiograph review based on lung field opacification, extent and thickness of pleura. Forty-six cases fulfilled inclusion criteria. Mean age was 56 +/- 21 years, 67% were male and there was no ethnic preponderance. Forty-six per cent of cases had specific underlying risk factors. Multi-loculated cases of empyema occurred in 63%, pleural aspirate was culture positive in 59% and Streptococcus milleri represented 42% of all positive cultures. Delay in appropriate management was 5.2 +/- 4.2 days. Fifty-six per cent of total procedures incorporated intrapleural streptokinase as an adjunct to drainage. The cure rate was 93.3% without requirement for surgical intervention. Two patients died early giving a 4.3% mortality rate. Radiological scores were 8.6 +/- 0.6 on admission, 6.3 +/- 1.5 on discharge and 2.5 +/- 2.5 at follow up. Improvement in radiological scores was statistically significant from admission to discharge and follow up (P < 0.001). The total duration of hospitalization was 22 +/- 10 days and by univariate analysis was predicted by: pleural sepsis not involving S. milleri (P = 0.001), specific predisposing factors (P = 0.02), and frank pus on pleural aspiration (P = 0.03). Total delay in management was suggestive in prolonging the duration of hospitalization (P = 0.07). Parapneumonic thoracic empyema is an important source of morbidity which is determined by both patient and management factors. Good results can be obtained by prompt medical therapy including tube drainage and intrapleural streptokinase without need for surgical intervention.