Relationship of catheter-associated urinary tract infection to mortality and length of stay in critically ill patients: a systematic review and meta-analysis of observational studies.

OBJECTIVE: To determine whether catheter-associated urinary tract infections are associated with increased morbidity and mortality in critically ill patients. DATA SOURCES: MEDLINE, HealthStar, EMBASE, and CINAHL databases from inception to June 2010 and bibliographies of included studies without language restriction. STUDY SELECTION: Studies reporting mortality or morbidity in adult intensive care unit patients with and without catheter-associated urinary tract infections. DATA EXTRACTION: Two authors independently selected studies and extracted data on study methodology, quality, and patient outcomes using a standardized form. Meta-analyses were performed using random-effects models. DATA SYNTHESIS: Of 720 citations, 11 studies enrolling 2,745 patients with and 60,719 patients without catheter-associated urinary tract infections met inclusion criteria. Catheter-associated urinary tract infection was associated with a significant increase in mortality (odds ratio [OR], 1.99; 95% confidence interval [CI], 1.72-2.31; p < .00001; I2 = 54%; eight studies; 62,063 patients) and length of stay in the intensive care unit (weighted mean difference of + 12 days; 95% CI, 9-15; p < .00001; I2 = 96%; seven studies; 13,011 patients) and hospital (mean difference + 21 days; 95% CI, 11-32; p < .0001; I2 = 98%; five studies; 10,183 patients). Restricting the analysis only to the two studies that adjusted for other outcome predictors, catheter-associated urinary tract infections were not associated with an increase in mortality (OR, 0.97; 95% CI, 0.82-1.16; p = .77; I2 = 0%; two studies; 5,626 patients). Although both studies individually demonstrated significantly increased intensive care unit length of stay after adjustment, pooled data showed that catheter-associated urinary tract infections were associated with a significant increase in intensive care unit length of stay using only a fixed effects model (mean difference + 2.6 days; 95% CI, 2.3-3.0; p < .00001) and not a random effects model (mean difference + 8 days; 95% CI, -13 to +28 days; p = .46) due to the high degree of heterogeneity for this outcome between the two studies (I2 = 99.6%) which results in a larger CI.
CONCLUSIONS: Catheter-associated urinary tract infection is associated with significantly increased mortality and length of stay in unmatched studies. Increased mortality and possibly increased length of stay appear to be consequences of confounding by unmeasured variables. These findings highlight the importance of evaluating risks and benefits of commonly used treatments such as antibiotics to manage catheter-associated urinary tract infection.