Because of a lack of clinical trials, the American Thoracic Society published a consensus statement on nosocomial pneumonia that included recommendations for antibiotic therapy. Almost concurrently, a multicenter study of the need to modify antibiotic therapy in ventilator-associated pneumonia (VAP) and a report specifically studying Pseudomonas VAP independently demonstrated both the appropriateness and some of the inadequacies of the therapies recommended by the American Thoracic Society. Anaerobic involvement as a copathogen was documented in early-onset VAP but probably impacts antibiotic choices little because of the presence of aerobic pathogens. Potential improvements in aminoglycoside treatment were suggested by meta-analysis of once-daily dosing and aerosolization. The most provocative study used decision-analysis techniques to suggest that antibiotic therapy based on clinical diagnosis of VAP would result in greater mortality than withholding antibiotics. Diagnosis of VAP by bronchoscopic and nonbronchoscopic quantitative cultures resulted in improved outcome with antibiotic treatment, possibly the result of documentation of inappropriate empiric antibodies in approximately 40% of cases of late-onset VAP.