Natural history and decolonization strategies for ESBL/carbapenem-resistant Enterobacteriaceae carriage: systematic review and meta-analysis.

Abstract

ESBL-producing Enterobacteriaceae and carbapenem-resistant Enterobacteriaceae (CRE) are rapidly spreading worldwide. Their natural reservoir is intestinal. We carried out a systematic review and meta-analysis to estimate CRE and ESBL carriage duration and to evaluate the effect of decolonization therapy. We included cohort and comparative studies examining the natural history of CRE/ESBL colonization, examining rates of carriage following decolonization or comparing decolonization and no decolonization conducted in the healthcare setting or in the community. A comprehensive search was conducted until November 2015. We compiled carriage rates at 1, 3, 6 and 12 months with and without decolonization therapy and assessed the effect of decolonization. Thirty-seven studies fulfilled inclusion criteria. In healthcare settings, pooled ESBL/CRE colonization rates decreased without intervention from 76.7% (95% CI = 69.3%-82.8%) at 1 month to 35.2% (95% CI = 28.2%-42.9%) at 12 months of follow-up. Following decolonization, the rate was 37.1% (95% CI = 27.5%-47.7%) at end of therapy and 57.9% (95% CI = 43.1%-71.4%) at 1 month. In two randomized trials, carriage was significantly reduced at end of therapy (risk ratio = 0.42, 95% CI = 0.25-0.65), but the effect was not significant after 1 month (risk ratio = 0.72, 95% CI = 0.48-1.05), with no longer follow-up. Heterogeneity was explained by surveillance methodology, with no differences observed between ESBLs and CREs. Among community dwellers, ESBL colonization decreased from 52.3% (95% CI = 29.5%-74.2%) at 1 month to 19.2% (95% CI = 9.7%-34.4%) at 6 months. A significant proportion of ESBL and CRE carriers remain colonized up to 1 year in the healthcare setting. While short-term decolonization therapy reduces carriage during therapy, its longer-term effects are unclear.