The role of barbiturate therapy in the development of pneumonia in head trauma patients in the intensive care unit of a university hospital was studied retrospectively. A total of 151 ventilated head trauma patients were included in the study. Intravenous thiopentone was administered to 75 patients (Group A), and 76 patients were managed without thiopentone therapy (Group B). Pneumonia was diagnosed when a new persistent pulmonary infiltrate appeared, with at least two of the following: (a) fever greater than 38°C, (b) a white blood cell count greater than 15 000/mm3, or (c) the presence of purulent bronchial secretions. On admission, there were no differences in the acute physiology and chronic health evaluation 2 revision (APACHE II) and Glasgow Coma Score between Groups A and B. Fifty-three per cent of the patients treated with thiopentone (Group A) developed pneumonia compared with 35% in Group B (odds ratio 1.85, 95% confidence interval 0.97-3.51). Gram-negative and Staphylococcus aureus organisms were the most frequently isolated in all cases. No differences in age, sex, APACHE II, Glasgow Coma Score, nutritional status or dexamethasone treatment were observed between the groups with and without pulmonary infection. In the multivariate analysis, prolonged mechanical ventilation before pneumonia and thiopentone treatment remained as the only independent risk factors for the appearance of pneumonia in head trauma patients (p=0.001 for both). Nevertheless, thiopentone did not increase the rate of mortality in patients with pneumonia. In conclusion, head trauma victims treated with thiopentone have a greater risk for the development of nosocomial pneumonia independent of mechanical ventilation.