Reappraisal of attributable mortality in critically ill patients with nosocomial bacteraemia involving Pseudomonas aeruginosa

In a retrospective study, population characteristics and outcome were investigated in intensive care unit (ICU) patients with hospital-acquired Pseudomonas aeruginosa bacteraemia admitted over a seven-year period (January 1992 through December 1998). A matched cohort study was performed in which all ICU patients with P. aeruginosa bacteraemia were defined as cases (N = 53). Matching (1:2 ratio) of the controls (N = 106) was based on the APACHE II classification: an equal APACHE II score ( 1 point) and an equal diagnostic category. Patients with P. aeruginosa bacteraemia had a higher incidence of acute respiratory failure, haemodynamic instability, a longer ICU stay and length of ventilator dependence (P<0.05). In-hospital mortalities for cases and controls were 62.3 vs. 47.2% respectively (P = 0.073). Thus, the attributable mortality was 15.1% (95% confidence intervals: -1.0-31.2). In a multivariate survival analysis the APACHE II score was the only variable independently associated with mortality. In conclusion, P. aeruginosa bacteraemia is associated with a clinically relevant attributable mortality (15%). However, we could not find statistical evidence of P. aeruginosa being an independent predictor of mortality. (C) 2003 The Hospital Infection Society.