Antibiotics in acne: time for a rethink

Over 10 years have passed since Anne Eady and her colleagues in Leeds demonstrated that therapeutic failure in acne may be related to bacterial resistance to antibiotics. The reluctance of clinicians to act on this information and to recognise the practices most likely to further this problem has led to a global increase in antibiotic resistance in P. acnes. In this issue, the Leeds group clearly show that P. acnes strains resistant to erythromycin, tetracycline, clindamycin and a variety of related antibiotics are to be found in Europe, USA, Australia and Japan. Through polymerase chain reaction, they demonstrate that the point mutations previously observed in the ribosomal RNA of resistant strains in the UK now occur across four continents and that the phenotypic pattern of resistance determined by MIC levels corresponds to the genomic pattern of mutation. For the first time the group have identified P. acnes with raised MIC levels to minocycline, although only in bacteria isolated from acne patients treated in the U.S.A. The emergence of resistant strains of bacteria due to exposure to antibiotics is regrettable and the inappropriate use of antibiotics has been condemned by the scientific, medical and lay press. A report from the House of Lords Select Committee on Science and Technology in 1998 called for national guidelines on the use of antibiotics and incentives from the pharmaceutical industry to find ways of overcoming antibacterial resistance. The work of the Leeds group raises the issue of the ecological importance of antibiotic resistance in P. acnes and challenges us to examine the place of antibiotics in the management of acne. The increase in antibiotic resistance is worrying but are we heading for cutaneous Armageddon? The mutations identified by Ross et al. in this issue are all in essential chromosomal genes and therefore not transferable between bacteria. The ermCD gene is a mobile resistance gene on a transposon that could potentially move between strains of propionibacteria and even to other coryneforms but this has proved difficult to achieve in vitro. Thus, the clinical and ecological problem is not the transfer of resistance between bacteria, but the transfer of resistant strains of bacteria between patients. The Leeds group believes that an individual may harbour different strains of P. acnes with varying degrees of antibiotic resistance. These findings should prompt us to look again at the relationship between acne and the microbe that bears its name. Propionibacterium spp. are present in both normal and abnormal pilosebaceous units but the role of these ubiquitous organisms in the pathogenesis of acne remains unclear. There is a significant increase in the number of P. acnes recoverable from the surface of the skin during adolescence but there is no, or a very weak, association between the severity of acne and the absolute number of bacteria on the surface or within superficial pilosebaceous ducts. The best evidence for a pathogenic role for these organisms in acne was provided by Eady's group in 1989, with the demonstration that therapeutic control of acne was lost when P. acnes developed resistance to erythromycin and that control could be regained when an antibiotic to which the organisms were sensitive was taken. This work has led Eady's group to develop a concept model that speculates that acne could be an infection of individual blocked pilosebaceous units. In this model there would be no reason to suppose that neighbouring pilosebaceous units would support identical or even similar populations of bacteria, indeed we know that it may be impossible to isolate any P. acnes from some pilosebaceous units of acne sufferers. Work with bacteria isolated from individual pilosebaceous units from truncal (back) skin of acne patients lends support to this theory but Koch's famous postulates for the definition of infection have not been met. At present, it is fair to assume that P. acnes spp. have a role in the pathogenesis of acne but the precise mechanism is