To describe the epidemiology of nosocomial pneumonia in trauma patients and its impact on outcome, we performed a retrospective case-control analysis. Quantitative bronchoscopic cultures were collected from 62 intubated patients with suspected pneumonia. Patients with proven pneumonia had higher abdominal injury scores. Those with bronchoscopy-negative pneumonitis were older. Age and injury severity were used to match two controls to each case. The incidence of pneumonia was 5.8% Streptococci and Hemophilus were common pathogens, but gram-negative rods were isolated more frequently after lengthier intubation. Polymicrobial infections were common. There were no serious complications of bronchoscopy, and culture results often led to antibiotic therapy. No excess mortality could be attributed to pneumonia. Patients with pneumonia and those with bronchoscopy-negative pneumonitis required prolonged care compared with others (p < 0.05). Patients with pneumonia did not receive excess ventilation or hospitalization but incurred hospital charges 1.5 times higher than controls (p = 0.04). Pneumonia was confirmed in less than half of those suspected of having it on the basis of clinical findings. When severity of injury was considered, pneumonia was associated with neither increased mortality nor increased hospital care, but the clinical features suggesting respiratory infection identified trauma patients requiring prolonged hospitalization and incurring higher costs.