Antibiotic resistance in Europe: burden, dissemination and potential control strategies

© 2015 Dr. Andrew James StewardsonAntimicrobial resistance (AMR) complicates the management of infections and is associated with additional morbidity, mortality, and costs. The aim of this research, involving multiple different studies across different research areas, was to contribute to knowledge regarding the burden, modifiable risk factors, and control of AMR. Burden: A single-centre pilot study (Chapter 2) informed the development of a multicentre retrospective cohort study in European hospitals to estimate the impact of AMR on the hospital mortality, excess length-of-stay, and cost attributable to bloodstream infections (BSI) caused by Staphylococcus aureus classified as methicillin-susceptible (MSSA) or resistant (MRSA), and Enterobacteriaceae classified as third-generation-cephalosporin-susceptible (3GCSE) or resistant (3GCRE) (Chapter 3–5). BSI due to 3GCRE, MSSA and MRSA increased the daily risk of death and prolonged length-of-stay compared with non-infected patients. 3GCSE BSI increased length-of-stay but not hazard of death. Third-generation-cephalosporin-resistance significantly increased the hazard of death and excess length-of-stay compared to susceptible strains, whereas methicillin resistance did not. The cost of AMR – from the hospital perspective – was extrapolated from excess length-of-stay and bed-day value. While bloodstream infections with S. aureus has a greater impact on mortality, excess length-of-stay and cost than Enterobacteriaceae per infection, the impact of AMR was greater among patients with Enterobacteriaceae BSI. Dissemination: We performed a multicentre prospective cohort study to determine the impact of antimicrobial treatment regimen on the intestinal carriage of resistant Enterobacteriaceae among outpatients with urinary tract infections, and their household contacts (Chapter 6). Fluoroquinolones – but not nitrofurans – transiently suppressed intestinal Enterobacteriaceae, followed one month later by a significant increase in prevalence and relative abundance of ciprofloxacin-resistant Enterobacteriaceae. Treatment duration didn’t modulate this effect. Fluoroquinolones were not associated with selection of extended-spectrum β-lactamase-producing Enterobacteriaceae (ESBL-PE). A microbiomic sub-study confirmed that ciprofloxacin has a transient global impact on intestinal microbiota (Chapter 7). A cross-sectional study determined the prevalence of food contamination and food-handler colonisation with ESBL-PE in a hospital kitchen and compared these strains with patient isolates (Chapter 8). While most raw chicken meat samples (92%) were contaminated with ESBL-PE, no ESBL-PE were found in prepared meals, food-handler colonisation was not frequent, and there was only modest overlap in the ESBL genes and E. coli strains of chicken and human origin. Current risk management strategies appear to minimise risk to staff and patients. Potential control strategies: Healthcare worker hand hygiene is considered important for reducing healthcare-associated AMR dissemination. The World Health Organization (WHO) recommends a multimodal hand hygiene promotion. We developed the WHO Hand Hygiene Self-Assessment Framework (HHSAF) as a self-administered survey to provide a situation analysis of hand hygiene resources and practices within healthcare organisations (Chapter 9). We performed usability and reliability testing of the HHSAF, with modification after each phase as indicated. Given hand hygiene compliance remains suboptimal despite standard multimodal promotion, we conducted a single-centre cluster-randomised controlled trial to investigate the impact of enhanced performance feedback and patient participation (Chapter 10). Wards were assigned to control; performance feedback; or performance feedback plus patient participation. The primary outcome was hand hygiene compliance. Secondary outcomes included patient colonisation and infection events. Hand hygiene performance increased in all arms, but the magnitude was greater in performance feedback plus patient participation wards than control wards. Patient-level outcomes did not follow a consistent pattern