Bloodstream infections in the ICU: incidence and outcome

Objective

To evaluate the risk factors for the incidence and the outcome of patients with bloodstream infections (BSIs) in a multidisciplinary ICU.

Methods

During a 1-year period (August 2004-July 2005), all ICU patients with more than 48 hours ICU stay were studied. All episodes of BSIs were recorded with the exception of those due to coagulase-negative staphylococci because of difficulties in the clinical interpretation. The prognostic value of clinical and laboratory variables were determined.

Results

Of the 693 patients admitted to the ICU during the study period, 572 patients had a stay longer than 48 hours. Among them a total of 125 patients developed one or more episodes of BSI. There was a significant difference between patients with and without BSIs in the length of ICU stay (6 vs 26 days, median, respectively), and in days of mechanical ventilation (7 vs 26 days, median, respectively). The median time between ICU admission and the first BSI was 10 days. The most common isolated pathogen was Acinetobacter baumannii (32.5%), followed by Klebsiella pneumoniae (21.3%), Pseudomonas aeruginosa (13.8%), Enterobacter aerogenes (6.9%), Staphylococcus aureus(5.8%) and Enterococcus faecalis (5.3%). The APACHE II score on admission was significantly higher in patients with BSI (19 ± 6.3 vs 16 ± 7, P < 0.001). The best independent prognostic factors of the development of BSI were APACHE II score on admission (OR, 1.05; 95% CI 1.02–1.086, P < 0.001), the presence of ARDS (OR, 2.63; 95% CI 1.42–4.87, P = 0.002) and a history of diabetes (OR, 2.26; 95% CI 1.3–3.93, P = 0.004). The ICU mortality rate was 46.4% and 22% for patients with and without BSIs, respectively (P < 0.001). Independent predictors of mortality were admission APACHE II score (OR, 1.13; 95% CI 1.1–1.2, P < 0.001), and the presence of BSI (OR, 2.5; 95% CI 1.6–3.9, P < 0.001).

Conclusion

In our ICU patients the illness severity on admission, the history of diabetes and the presence of ARDS are risk factors for the incidence of BSI. The severity of illness and the presence of BSI are factors independently associated with the outcome.