Little is known about the natural history of resolution of nosocomial pneumonia, and thus it is likely that we are not always using the optimal duration of therapy in all patients. For some patients, with few risk factors for a poor outcome, and infection with easily treated pathogens, we can probably treat with a more abbreviated course of therapy than is commonly used. For other patients with risk factors for a poor outcome, and infection with ;;high risk'' pathogens such as Pseudomonas aeruginosa, we may need longer durations of therapy. We review the clinical and microbiological definitions of resolution, including improvement, delayed resolution, relapse, or recurrent infection. There are also microbiological end points for resoution including eradication, persistence, and superinfection. The clinical parameters that affect resolution are patient related, microbiological, and treatment related, and these factors are summarized here. Currently, the time course of resolution is being defined using clinical end points such as the clinical pulmonary infection score (CPIS) and microbiological end points such as quantitative cultures of respiratory secretions. The hope for the future is to be able to identify whether the clinical response is adequate, at the earliest posible time point. This may allow for interventions to help the nonresponding patient, or shorten the duration of therapy in the responding patient.