Methods
A systematic literature review was conducted using the PubMed/MedLine, CINAHL Plus, Global Health and Embase databases. We sought to retrieve all studies of reasonable quality which assessed the impact of antimicrobial cycling strategies on bacterial resistance within clinical settings. We also recorded antimicrobial consumption and morbidity and/or mortality rates as secondary outcomes for a more thorough assessment of the observed results.
As this study was part of a wider project we designed a broad search algorithm on the basis of definitions provided by major organizations: Infectious Diseases Society of America (IDSA), Center for Disease Prevention and Control (CDC)[11][12]. The search string covered three concepts, antimicrobial stewardship and its constituent strategies, antimicrobial resistance, and the hospital setting of the interventions:
  1. (antimicrobial stewardship) OR (antibiotic stewardship) OR (audit “and” feedback) OR (restriction) OR (pre?authorization) OR (antibiotic combination*) OR (antimicrobial combination*) OR (antibiotic cycling) OR (antimicrobial cycling) OR (antibiotic rotation) OR (antimicrobial rotation) OR (antibiotic time?out*) OR (antimicrobial time?out*) OR (dose adjustment) OR (dose optimi#ation) OR (antibiotic mixing) OR (antimicrobial mixing) OR (antibiotic de?escalation) OR (antimicrobial de?escalation) OR (parenteral oral conversion) OR (intravenous oral conversion) OR (procalcitonin) OR (electronic alert*) OR (electronic system*) OR (computeri#ed alert*) OR (computeri#ed system*) OR (automat* stop order*)
  2. Exp Drug Utilization
  3. 1 OR 2
  4. (antibiotic resistan*) OR (antimicrobial resistan*) OR (multi?drug resistan*) OR (bacterial resistan*) OR (bacterial susceptib*) OR (susceptib* phenotype*) OR (antibiotic susceptib*) OR (antimicrobial susceptib*)
  5. 3 AND 4
  6. (nosocomial OR hospital* OR in?patient OR intensive care OR ICU*)
  7. 5 AND 6
8,922 papers covering the period to 1st April 2020 were screened for relevance. Randomised studies were scarce and for this reason we decided to broaden inclusion criteria by considering quasi-experimental designs. However, we excluded simple before-and-after studies which examined cohorts lasting less than one year, to minimise confounding due to seasonality and to facilitate comparability of results. We also excluded studies which combined changes in infection control practices or applied multidisciplinary interventions due to confounding and constraints on comparability. Studies which lacked historical or parallel cohorts for comparison were not included as interpretation is impossible without some kind of internal control or comparator. Data provided by grey literature such as congress papers and reports from governmental and non-governmental organizations were outside our scope due to lack of peer review. Finally, studies which did not apply suitable statistical methods to evaluate the significance of the reported results were also excluded.
A main distinction from prior meta-research on the topic is the fact that we considered changes in infection control as well as the application of additional antimicrobial stewardship interventions as important confounding factors which should not be overlooked; this led to the exclusion of several papers which other reviews have included.