Impact of central-line-associated bloodstream infections and catheter-related bloodstream infections: a systematic review and meta-analysis

Background: Accurate effect estimates are needed to inform input parameters of health economic models. Central-line-associated bloodstream infections (CLABSIs) and catheterrelated bloodstream infections (CRBSIs) are different definitions used for central-line bloodstream infections and may represent dissimilar patients, but previous metaanalyses did not differentiate between CLABSIs/CRBSIs. Aim: To determine outcome effect estimates in CLABSI and CRBSI patients, compared to uninfected patients. Methods: PubMed, Embase, and CINAHL were searched from January 2000 to March 2024 for full-text studies reporting all-cause mortality and/or hospital length of stay (LOS) in adult inpatients with and without CLABSI/CRBSI. Two investigators independently reviewed all potentially relevant studies and performed data extraction. Odds ratio for mortality and mean difference in LOS were pooled using random-effects models. Risk of study bias was assessed using ROBINS-E. Findings: Thirty-six studies were included. Sixteen CLABSI and 12 CRBSI studies reported mortality. The mortality odds ratios of CLABSIs and CRBSIs, compared to uninfected patients, were 3.19 (95% CI: 2.44, 4.16; I2 = 49%) and 2.47(95% CI: 1.51, 4.02; I2 = 82%), respectively. Twelve CLABSI and eight CRBSI studies reported hospital LOS; only three CLABSI studies and two CRBSI studies accounted for the time-dependent nature of CLABSIs/CRBSIs. The mean differences in LOS for CLABSIs and CRBSIs compared to uninfected patients were 16.14 days (95% CI: 9.27, 23.01; I2 = 91%) and 16.26 days (95% CI: 10.19, 22.33; I2 = 66%), respectively. Conclusion: CLABSIs and CRBSIs increase mortality risk and hospital LOS. Few published studies accounted for the time-dependent nature of CLABSIs/CRBSIs, which can result in overestimation of excess hospital LOS. (c) 2024 The Healthcare Infection Society. Published by Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.