Impact of catheter-related bloodstream infections on the mortality of critically ill patients: A meta-analysis

Objective. There is controversy on whether catheter-related bloodstream infections (CR-BSI) affect the mortality of critically ill patients. Design., Meta-analysis of comparative studies that reported on mortality of intensive care unit (ICU) adult patients with and without CR-BSI. Methods. PubMed, Current Contents, and reference lists of retrieved publications were searched with no language or time restrictions. Heterogeneity was assessed by means of l(2)-statistic and chi-square test. Publication bias was detected by the funnel plot method using Egger's test. Pooled odds ratios (OR) and 95% confidence intervals (CI) were calculated by implementing both the Mantel-Haenszel fixed effect and the DerSimonian-Laird random effects model. Results. Eight studies, involving 2,540 ICU patients, were included. Heterogeneity was detected (l(2) = 0.67, 95% Cl 0.32-0.85, p = 0.003). Publication bias was not found (Egger's test, p = 0.28). All-cause in-hospital mortality was higher in ICU patients with CR-BSI than in those without CR-BSI (fixed effect model: OR = 1.81, 95% Cl 1.44-2.28; random effects model: OR = 1.96, 95% Cl 1.25-3.09). This was also the case for the subgroup analysis of the studies that were matched for severity of illness (fixed effect model: OR = 1.65, 95% Cl 1.28-2.13; random effects model: OR = 1.70, 95% Cl 1.00-2.90). Conclusion: Presence, as opposed to absence, of CR-BSI is associated with higher mortality in critically ill adult patients. This finding seems to justify and may enhance efforts to prevent CR-BSI in such patients. (Crit Care Med 2009; 37:2283-2289)