Rapid identification of infected patients and accurate selection of antimicrobial agents for initial treatment of hospital-acquired pneumonia represent important clinical goals, because it seems that better treatment of this infection might have a major impact on hospital-associated mortality and morbidity. Persistently high mortalities for pneumonia in the critical care unit argue, however, for a continued reassessment of the current modalities of therapy and definition of better protocols. More active and less toxic antibacterial agents are still needed. It should be emphasized that in the event that one or several specific etiologic agents are identified by a reliable technique, the choice of antimicrobial drugs is much easier, because the optimal treatment may be selected in light of the susceptibility pattern of the causative pathogens without resorting to broad-spectrum drugs or risking inappropriate treatment. Great efforts should be placed to obtain reliable pulmonary specimens for direct microscopic examination and cultures in each patient clinically suspected of having developed nosocomial pneumonia before new antibiotics are administered.