REVIEW ARTICLE Adjunctive Use of Rifampin for the Treatment of Staphylococcus aureus Infections A Systematic Review of the Literature Joshua Perlroth, MD; Melissa Kuo, MD; Jennifer Tan, MHS; Arnold S. Bayer, MD; Loren G. Miller, MD, MPH Background: Staphylococcus aureus causes severe life- efit of adjunctive rifampin use, particularly in osteomy- threatening infections and has become increasingly com- elitis and infected foreign body infection models; however, mon, particularly methicillin-resistant strains. Rif- many studies failed to show a benefit of adjunctive therapy. ampin is often used as adjunctive therapy to treat S aureus Few human studies have addressed the role of adjunc- infections, but there have been no systematic investiga- tive rifampin therapy. Adjunctive therapy seems most tions examining the usefulness of such an approach. promising for the treatment of osteomyelitis and pros- thetic device–related infections, although studies were Methods: A systematic review of the literature to iden- typically underpowered and benefits were not always seen. tify in vitro, animal, and human investigations that com- pared single antibiotics alone and in combination with Conclusions: In vitro results of interactions between rif- rifampin therapy against S aureus. ampin and other antibiotics are method dependent and often do not correlate with in vivo findings. Adjunctive Results: The methods of in vitro studies varied substan- rifampin use seems promising in the treatment of clini- tially among investigations. The effect of rifampin therapy cal hardware infections or osteomyelitis, but more de- was often inconsistent, it did not necessarily correlate with finitive data are lacking. Given the increasing incidence in vivo investigations, and findings seemed heavily de- of S aureus infections, further adequately powered in- pendent on the method used. In addition, the quality of vestigations are needed. data reporting in these investigations was often subop- timal. Animal studies tended to show a microbiologic ben- Arch Intern Med. 2008;168(8):805-819 TAPHYLOCOCCUS AUREUS INFEC- more limited than those for methicillin- tions are common, severe, and susceptible S aureus strains. Further- associated with significant more, treatment failure with standard morbidity and mortality. therapies for MRSA is common.13-15 There Staphylococcus aureus is the is a need to better understand the effi- Smost common cause of skin and soft- cacy of antimicrobial therapies for MRSA tissue infections and is a frequent cause of and difficult-to-treat S aureus infections. serious infections such as health care– Rifampin, a broad-spectrum antimicro- associated bloodstream infections,1,2 device- bial agent that is bactericidal against S au- Author Affiliations: Division of associated infections,3,4 and osteomyeli- reus, achieves high intracellular levels and Infectious Diseases tis.5,6 Worldwide, S aureus is the most is one of the few antimicrobial agents that (Drs Perlroth, Bayer, and common cause of infective endocarditis.7,8 can penetrate biofilms and kill organisms Miller) and Los Angeles 16-18 Biomedical Research Institute Of concern, the number of infections in the sessile phase of growth. Its use (Ms Tan and Dr Miller), caused by methicillin-resistant S aureus as monotherapy has been abandoned be- Department of Medicine, (MRSA) continues to rise. In intensive care cause of the rapid development of resis- Harbor-UCLA Medical Center, units, the proportion of S aureus infec- tance, which is prevented by combina- Torrance, California, tions that are MRSA in the United States tion with another active antibiotic. St Mary Medical Center, Long has been increasing by 3% per year, and Combination therapy with rifampin has Beach (Dr Kuo), and Division MRSA now constitutes more than 60% of been used to treat S aureus infections.19,20 of Infectious Diseases, David S aureus strains that cause infections.9 Nevertheless, it has been commonly used Geffen School of Medicine at Methicillin-resistant S aureus has been re- adjunctively to treat S aureus infec- the University of California, 19,20 Los Angeles (Drs Bayer and cently implicated as the causative agent of tions. However, data to support this Miller). Dr Kuo is now with life-threatening community-acquired in- practice are limited and are typically based Division of Infectious Diseases, fections such as sepsislike syndromes, nec- on small clinical studies or basic in vitro Department of Medicine, rotizing pneumonia, and necrotizing fasci- investigations. To our knowledge, there University of California, Irvine. itis.10-12 Therapeutic options for MRSA are has been no attempt to synthesize the lit- (REPRINTED) ARCH INTERN MED/ VOL 168 (NO. 8), APR 28, 2008 WWW.ARCHINTERNMED.COM 805 ©2008 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/28/2021 erature that has examined the effi- DATA EXTRACTION RESULTS cacy of rifampin therapy against S aureus. To this aim, we con- Each investigator was blinded to the ducted a systematic review of the use other investigator’s data extraction. The The results of the literature search of rifampin as adjunctive therapy to 2 reviewers independently rejected or ac- and the reasons for exclusion from treat S aureus infections. cepted each abstract based on the inclu- the systematic review are summa- sion and exclusion criteria. Article texts rized in the Figure. To facilitate of selected abstracts were reviewed, as review of our findings, results are METHODS were article texts of abstracts that could summarized herein as in vitro in- not be excluded based on abstract re- vestigations, animal investigations, LITERATURE SEARCH view alone. All disagreements between the abstractors as to whether the article and human investigations. should be included were settled by a To identify in vitro, animal, and hu- third independent reviewer (L.G.M.). IN VITRO INVESTIGATIONS man subject data regarding the efficacy Data from each trial were entered onto of rifampin as adjunctive therapy for the a standardized form, verified for accu- treatment of S aureus, 2 independent re- We identified 72 publications com- racy, and input into a computerized da- viewers (J.P. and M.K.) searched paring antibiotic efficacy with and tabase. Information extracted included PubMed, Cochrane Library, and study design, antibiotics tested, num- without rifampin using in vitro mod- EMBASE for publications containing the ber of subjects, year of publication, du- els (Table 1). Of 164 individual text phrases Staph* AND rifamp*. These ration of follow-up, clinical setting (in antibiotic trials in these 72 publica- terms were used to avoid ignoring ar- vitro, animal, or human), and interven- tions, 41 trials tested both methicillin- ticles with permutations of the words tion (dosage, frequency, and duration of susceptible S aureus and MRSA Staphylococcus and rifampin, (eg, Staph therapy or exposure). Abstracted data in- aureus or rifampicin). All abstracts were strains. We found that methods used cluded the outcome (eg, mortality, clini- printed for review. The search was lim- to determine the nature of the anti- cal failure, and colony count after treat- ited to English-language articles pub- biotic interactions were heteroge- ment) and the time of evaluation of lished between January 1, 1966, and neous and included E test, time-kill treatment outcome. Discrepancies in January 31, 2006. We reviewed only curves, checkerboard assays, serum data between abstractors were identi- manuscripts relating to S aureus and not fied and resolved via discussion among bactericidal activity, and ex vivo in- coagulase-negative staphylococci. We the investigators. For human studies, 2 tracellular bactericidal activity. Even also contacted several experts in the field independent reviewers (M.K. and J.T.) within-method differences (eg, among of adjunctive rifampin therapy to deter- reviewed the 8 human subject trials and time-kill studies) were substantial in mine if there were any pertinent re- assigned them Jadad scores; disagree- terms of the media used, the inocu- cently published meeting abstracts or ments between the reviewers were settled 4 9 published articles that we missed by our lum (from 10 to 10 bacteria), the by a third independent reviewer systematic review. growth phase (ie, stationary vs log (L.G.M.). phase), the temperature of inocula- tion (from 36°C to 38°C), the time the INCLUSION AND EXCLUSION CFU outcomes were quantified (from CRITERIA STATISTICAL ANALYSIS 6 to 48 hours after inoculation), and An investigation was included in our sys- For investigations that did not perform the concentration of the antibiotics tematic analysis if it met each of the fol- statistical analyses but reported re- used (0.25 to many times the mean lowing criteria: (1) the organism under sults, we attempted to perform statisti- inhibitory concentration). study was S aureus; (2) the study de- cal analyses. For investigations compar- Methodological variability was re- sign compared the efficacy of 1 or more ing dichotomous outcomes, we flected in the inconsistent interactions antibiotics alone and in combination performed ␹2 or Fisher exact test. For reported. Many studies using time-kill with rifampin; (3) the study outcome as- studies comparing means, if standard de- assays failed to report results in terms sessed quantitative bacterial measure- viations and group sample sizes were of synergy or antagonism, presenting ments, cure rates (or eradication of colo- available, we performed Wilcoxon