Nosocomial pneumonia in the ICU: a prospective cohort study.

Ventilator-associated pneumonia (VAP) is the most common intensive care unit (ICU)-acquired infection among patients requiring mechanical ventilation. A prospective surveillance programme of all patients has been implemented at the ICU, Karolinska University Hospital, Sweden since 2001. Within this programme, incidence and risk factors for ICU-acquired pneumonia and associated death over a 2-y period have been studied. Of 329 patients enrolled in the study, 221 required mechanical ventilation. 33 of 221 patients (15%) developed VAP, corresponding to a rate of 29 VAP/1000 ventilator d. Risk factors for VAP were aspiration (hazard ratio 3.79; 95% CI 1.48-9.68), recent surgery (HR 3.58; 95% CI 1.15-11.10) and trauma (HR 3.00; 95% CI 1.03-8.71). 11 patients of 33 (33%) with VAP died within 28 d compared to 46 of 288 (16%) without ICU-acquired pneumonia (odds ratio 2.73; 95% CI 0.97-7.63). We conclude that: 1) incidence of VAP was 15% and the most important risk factor was aspiration; 2) APACHE II score > or = 20 is a stronger predictor for poor outcome than VAP; 3) a minority of patients with APACHE II score > or = 20 develop VAP; and 4) continuous surveillance programmes are feasible and provide valuable data for improvement of quality of care.