Controlling antibiotic resistance in the ICU

Abstract

Patients admitted to intensive care units (ICUs) are frequently colonized with (antibiotic-resistant) bacteria, which may lead to healthcare associated infections. Antimicrobial-resistant bacteria (AMRB), such as methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant Enterococci (VRE) and highly-resistant Enterobacteriaceae (HRE) have been increasing worldwide and are associated with delayed appropriate therapy, failure of therapy, and increased morbidity and mortality. The global increase in AMRB necessitates more effective control measures, especially in intensive care units. The main part of this thesis was an international study in 13 ICUs in which we aimed to control antibiotic resistance: the MOSAR-ICU trial. After a 6 months baseline phase, we implemented chlorhexidine body-washing (CBW), combined with hand hygiene improvement (P2) in all ICUs. In phase three (P3), ICUs were additionally randomized to universal screening (PCR-based tests for MRSA and VRE and chromogenic screening for HRE); or to screening for MRSA and VRE with chromagar without HRE screening. In all ICUs, carriers detected were treated with contact precautions. Implementation of a hand hygiene improvement program in these ICUs yielded a rapid and sustained improvement of compliance from 52% (baseline), to 69% (P2) and eventually 77% in P3. Adding feedback of results appeared to increase effectiveness. In P2, there was a 3.6% weekly reduction of MRSA-acquisition. There was a trend towards a similar effect for VRE, but no trend was observed for HRE-acquisition. This suggests that patient-to-patient transmission may not be the dominant acquisition route for HRE in ICUs. The interventions in P3 failed to further reduce acquisition rates. Thus, we recommend that all ICUs implement the hand hygiene improvement program and universal CBW; and do not recommend universal screening to control antibiotic resistance in ICUs. There was some published evidence that CBW is effective in preventing carriage, and possibly bloodstream infections, with MRSA and VRE in ICU patients. There was no evidence (or lack of evidence) that it reduces carriage or infections with HRE. Our study on molecular epidemiology of MRSA revealed that MRSA epidemiology was homogenous within ICUs of the same country, but heterogeneous between countries. LA-MRSA and CA-MRSA were nearly absent in these European ICUs. ForHRE, there waslarge heterogeneity in antibiotic resistance mechanisms and sequence types of E. coli and K. pneumoniae; and a difference in clonality between these two pathogens. If confirmed in other studies, this would demand different control policies for E. coli and K. pneumoniae. Of all patients at risk for HRE acquisition in one French participant, 13% acquired HRE carriage during their ICU stay. No carbapenemase-producing Enterobacteriaceae were detected. Prior exposure to third generation cephalosporins or to a beta-lactam inhibitor (within 3 months of ICU admission) was strongly associated with acquisition. Thus, we would not recommend empirical treatment with carbapenems in this setting. In a sub-study on duration until decolonization, 127 included patients were colonized with AMRB during their first admission and had at least one readmission. The median times until decolonization were 1-2 months for all bacterial types, with no significant differences between the survival curves of the different bacteria. In view of these findings, interventions targeted at recently readmitted patients may be most effective.