Background: Ventilator-associated pneumonia, a leading cause of sepsis in patients with acute respiratory failure, is difficult to distinguish clinically from other processes affecting patients receiving mechanical ventilation. We conducted a prospective study of patients with suspected ventilator-associated pneumonia to identify the causes of fever and densities on chest radiographs and to evaluate the diagnostic yield and efficiency of tests used alone and in combination., Methods: The 50 patients entered into the study underwent a systematic diagnostic protocol designed to identify all potential causes of fever and pulmonary densities. Diagnoses responsible for fever were established by strict diagnostic criteria for 45 of the 50 patients. The prevalence of specific conditions and diagnostic yield of individual tests were used to formulate a simplified diagnostic protocol., Results: The diagnostic protocol identified 78 causes of fever (median 2 per patient). Infections were the leading causes of fever and pulmonary densities. Of the 45 patients with fever, 37 had one or more infections identified (67 sources). Most infections (84 percent) were one of four types:pneumonia, sinusitis, catheter-related infection, or urinary tract infection. Ventilator-associated pneumonia occurred in only 42 percent. All but nine infections (87 percent) were directly or indirectly related to insertion of a catheter or a tube. Concomitant infections were frequent (62 percent), particularly in patients with sinusitis (100 percent), catheter-related infections (93 percent), and pneumonia (74 percent). Of concomitant infections, 60 percent were caused by a different pathogen. Noninfectious causes of fever were more common in the 22 patients with adult respiratory distress syndrome. Histologically proved pulmonary fibroproliferation was the only cause of fever in 25 percent of patients with adult respiratory distress syndrome. Radiographic densities were caused by an infection in only 20 patients (19 pneumonia, 1 empyema). In more than 50 percent of the 25 patients without adult respiratory distress syndrome, congestive heart failure, and atelectasis were the sole causes of pulmonary densities, and fever always originated from an extrapulmonary site of infection. Used in combination, bronchoscopy with protected sampling, computed tomographic scan of the sinuses, and cultures of maxillary sinus aspirate, central intravenous or arterial lines, urine, and blood identified 58 of the 78 sources of fever (74 percent)., Conclusions: The observations in this study document the complex nature of acute respiratory failure and fever and underscore the need for accuracy in diagnosis. The frequent occurrence of multiple infectious and noninfectious processes justifies a systematic search for source of fever, using a comprehensive diagnostic protocol. A simplified diagnostic protocol was devised based on the diagnostic value of individual tests.