Predisposing factors for nosocomial pneumonia in patients receiving mechanical ventilation and requiring tracheotomy

STUDY OBJECTIVES: To assess the incidence of nosocomial pneumonia (NP) after tracheotomy in an ICU population and to determine NP risk factors during the ICU stay, particularly on the day of tracheotomy. DESIGN: A retrospective study using prospectively collected data. SETTING: A 16-bed multidisciplinary ICU. PATIENTS: One hundred thirty-five patients requiring tracheotomy for mechanical ventilation (MV) weaning. RESULTS: The mean (+/- SD) duration of MV before tracheotomy was 17.8 +/-13.4 days. Thirty-seven cases of NP occurred in 35 patients (25.9%), 8.7+/-7.3 days after the tracheotomy procedure. NP cases were classified as early NP (n = 19) if they occurred within 5 days after the procedure (mean, 2.7+/-1.1 days), and as late NP (n = 18) if they occurred beyond the fifth day (mean, 14.4+/-6.1 days). Multivariate analysis identified the following three independent factors associated with early NP: the presence of positive endotracheal aspirates (EAs) with pathogen levels of > or =10(5) cfu/mL (p = 0.0001); hyperthermia (temperature, > or =38.3 degrees C; p = 0.002) on the day of tracheotomy; and the continuation of sedation beyond 24 h after the tracheotomy (p = 0. 0001). Accountable pathogens of early NP were present in EA on the day of tracheotomy (p = 0.001). Cases of late NP were significantly associated with the duration of sedation before the procedure (p = 0. 002) and with hyperthermia (temperature, > or =38.3 degrees C) on the day of tracheotomy (p = 0.0005). The ICU admitting diagnosis, previous NP, duration of administration of antimicrobial agents and MV before tracheotomy, indication for tracheotomy, PO(2)/fraction of inspired oxygen ratio, and use of steroids on the day of the procedure were not associated with the occurrence of NP. The mortality rate of our population was 33.3%, and NP increased this percentage to 54.3%. CONCLUSIONS: Our results could suggest that tracheotomy should be delayed in mechanically ventilated patients with bronchial colonization and hyperthermia, when sedation cannot be discontinued after the procedure, to prevent occurrence of early NP.