Nosocomial pneumonia in the critically III

Abstract

Despite the large body of published work emphasising the importance of nosocomial pneumonia its true significance with respect to mortality is difficult to assess. This partially stems from the limitations imposed by clinical diagnostic criteria and from the interpretation of multivariate analyses which in themselves can only confirm correlations and associations. Two key questions need to be resolved. First, what is the true incidence of nosocomial pneumonia? Microbiological data would suggest the incidence to be much lower than previously suspected and this view is supported by the relative rarity of severe lung infections manifested as lung cavitation, empyema or bacteraemia. Second, what is the true mortality attributable to nosocomial pneumonia? Clinical studies need to examine the mode of death and not just confirm the presence of pneumonia at the time of death. Only in cases where overwhelming lung infection leads to an inability to maintain respiratory or haemodynamic homeostasis can death be genuinely attributed to nosocomial pneumonia. Clinical experience suggests that the attributable mortality is low, an impression that is reinforced by the relatively small impact of selective decontamination of the digestive tract (SDD) upon survival in patients with nosocomial pneumonia. Until these questions are answered satisfactorily it is appropriate to take reasonable measures to prevent nosocomial pneumonia but to embark on antibiotic therapy only when there is supportive microbiological evidence or in the presence of clinically indisputable lung infection.