Lack of evidence for attributing chlorhexidine as the main active ingredient in skin antiseptics preventing surgical site infections.

To the Editor—We read with great interest the articles by Noorani et al in the British Journal of Surgery and by Lee et al in Infection Control and Hospital Epidemiology. The authors have published almost simultaneously two systematic reviews of what is described as comparisons of chlorhexidine versus povidone-iodine for preoperative skin antisepsis. On the basis of their analyses, the authors conclude that chlorhexidine is the more effective antiseptic in preventing surgical site infections (SSIs). However, we believe that this conclusion is not sufficiently supported by the studies included. In both articles, the effect of preventing SSIs is solely attributed to chlorhexidine. However, the majority of studies were based on chlorhexidine-alcohol mixtures. When attributing effects to factors, it is necessary to look for other factors that could also be influencing these effects. The common agents for preoperative skin antisepsis are (1) alcohols, (2) chlorhexidine, and (3) povidone-iodine. Both chlorhexidine and povidone-iodine are available in aqueous formulations and in alcoholic formulations. Aqueous formulations have 1 active ingredient, whereas alcoholic formulations include 2 active ingredients. The differential antimicrobial activity of these compounds has been a topic of intense research and evaluation since the 1970s and is well described in infection control textbooks and guidelines.' If aqueous chlorhexidine or povidone-iodine is compared with various alcohols, then the immediate antimicrobial activity of the alcohols is significantly greater than that of the aqueous agents, by a factor of about 1 log (ie, a 10-fold difference). Alcohol is clearly a powerful skin antiseptic on its own, and in alcoholic chlorhexidine or povidone-iodine formulations it is the agent that contributes most to the overall activity. We have highlighted this in a previous letter to the editor. Formulations of alcohols with chlorhexidine or povidone-iodine may have some added activity or persistency, which pure alcohols do not have. In our opinion, the articles by Noorani et al and Lee et al have not shown clear and unambiguous evidence that the observed effects are solely or even mainly due to chlorhexidine, despite this being the main claim put forward by the authors. At the same time, they ignore the effects that the alcohol in the antiseptics is likely to have had. Beyond the mere presence of chlorhexidine and povidone-iodine, the studies analyzed in both articles are very heterogeneous. Some compared alcoholic chlorhexidine with aqueous povidoneiodine, one compared aqueous preparations in both study arms, and some compared alcoholic preparations in both arms. Some assessed SSI rates as the outcome, others (analyzed in Lee et al) assessed skin microbial cultures after antisepsis. One recent study with significantly different SSI rates as the outcome, contributing a large number of observations to the analyses, compared the use of alcoholic chlorhexidine to that of aqueous povidone-iodine, where the alcohol in the chlorhexidine formulation clearly confers an a priori advantage. One study involved vaginal surgery exclusively. The situation for vaginal surgery is fundamentally different from surgery through superficial skin, since the physiology of mucous membranes and the quality and quantity of the microbial flora are vastly different, and only aqueous preparations can be used. One earlier study from 1982 used alcoholic formulations in both arms but did not specify the alcohol concentrations used and whether they were in the antimicrobially active range, which means that no conclusions can be drawn. None of the studies with microbial culture results as the end points used any US or European standardized methods for antiseptic or disinfectant testing. Also, it is important to note that none of these studies specified whether they used any neutralizer substances in the experiments. Neutralizers are essential for some antiseptic testing experiments because some antiseptics continue to kill microorganisms after sampling, so that in the absence of neutralizers falsely higher kill rates will be seen. Chlorhexidine in particular appears to be prone to this effect, with falsely low colony counts being a consequence.