The charts of 110 children with community acquired bacterial pneumonia were reviewed. A subset of children who required surgical intervention for empyema or parapneumonic effusion was identified. Patients were divided into two treatment groups: antibiotics/tube thoracostomy alone (group 1) versus operative intervention (group 2). Overall, 33 (30%) of the children required surgical intervention for complications. Seventeen (15%, group 1) were successfully treated with antibiotics/tube thoracostomy alone, while 16 (15%) in group 2 were treatment crossovers, failing this initial therapy. Of group 2 children, 4 (25%) underwent thoracotomy and lobectomy, while 12 (75%) underwent video-assisted thoracoscopic surgery (VATS). Although group 2 children were younger than those in group 1 (4.4 +/- 3.6 versus 6.3 +/- 4.1 years, p<0.05) and had longer hospitalizations (20.1 +/- 10.1 versus 8.2 +/- 3.9 days, p<0.05), symptom duration, preoperative antibiotics, fibrinolytic use, and leukocytosis were similar (p>0.05). Group 1 children had 13 (76%) positive cultures, 92% with pan-sensitivities, in contrast to group 2, which had 12 (75%) positive cultures, but only 33% were sensitive to first-line antibiotics (p<0.01). Group 2 patients were also more likely to have complex multi-loculated empyemas, pneumatoceles, or pulmonary necrosis identified on imaging studies (100% versus 24%, p<0.01). These data suggest that the natural history of pneumonia in children is heavily influenced by bacterial virulence. Tube thoracostomy and appropriate antibiotics remain effective for pan-sensitive, simple parapneumonic effusions and empyema. Complex parapneumonic effusions and empyema, however, which occur more frequently in the setting of first-line antibiotic resistance, often fail more conservative managements and may be best treated by earlier operative debridement.