Considering resistance in systematic reviews of antibiotic treatment

Abstract

Context: Microorganisms resistant to antibiotic drugs are a threat to the health and chances of survival of patients. Systematic reviews on antibiotic drugs that ignore the topic of resistance present readers with a skewed view, emphasizing short-term efficacy or effectiveness while ignoring long-term consequences. Objectives: To examine whether systematic reviews of antibiotic treatment consider resistance; if not, to find out whether data on resistance were reported in the original trials; and based on that, to offer a framework for taking resistance into account in systematic reviews. Data sources: The Cochrane Database of Systematic Reviews (the Cochrane Library, 2001, issue 2); and MEDLINE, 1996-2000. Study selection: (i) Systematic reviews or meta-analyses of antimicrobial therapy, published during 1996-2000. (ii) Randomized, controlled trials abstracted in systematic reviews that addressed a topic highly relevant to antibiotic resistance. Data extraction: We examined each systematic review, and each article, to see whether the implications of resistance were discussed; and whether data on resistance were collected. Results: Out of 111 systematic reviews, only 44 (40%) discussed resistance. Ten reviews (9%) planned or performed collection of data on the response of patients with susceptible or resistant isolates. In 22 systematic reviews (20%), collection of data on induction of resistance was planned or performed. The topic of 41 reviews was judged highly relevant to resistance, and these reviews extracted data from 337 articles, out of which we retrieved 279 articles (83%). In 201 (72%) articles, resistance was discussed or data pertaining to it were collected. Ninety-seven articles (35%) gave actual data on resistance of pathogens to the study drugs, 71 articles (25%) data on efficacy of antibiotic drugs in patients with susceptible and resistant pathogens, and 55 articles (20%) provided data on infection or colonization with resistant strains during treatment. Conclusions: Most systematic reviews on antibiotic treatment ignored the issue of resistance, although many of the original articles referred to it and some reported relevant data. Reviewers should collect data on resistance and discuss the implications in their discussion and sections concerned with policy implications.