ORIGINAL ARTICLE, p 760
The late Archibald Leman Cochrane was an internationally renowned epidemiologist who recognized the necessity for the medical profession to organize a critical summary, specialty by specialty, of all relevant randomized controlled trials (RCTs) and to update this material periodically.1 The purpose of these critically appraised systematic reviews was to provide validated evidence to inform both healthcare providers (in policy formation as well as in practice) and healthcare users. Cochrane reviews are increasingly seen as fundamental research that should precede the design of new studies by highlighting areas where more quality RCTs are needed.2 Forty-nine of the 4500 reviews currently available in the Cochrane library are by the Skin Group. These address the common dermatology diagnoses and a substantial number report sufficient evidence to inform clinical decision-making.3
In this issue of BJD, Van Zuuren et al.4 examine the evidence-based management strategies for rosacea. This updates the existing Cochrane systematic review with additional trials and applies intention-to-treat to the meta-analyses. Fifty-eight trials comprising 6633 participants were examined. The review concludes that topical metronidazole, azelaic acid and doxycycline (40 mg) appear to be effective for moderate to severe rosacea, and that ciclosporin 0·5% ophthalmic emulsion is effective for ocular rosacea (based on one study). Quality-of-life outcomes were evaluated in a minority and participants’ assessments of improvement in rosacea severity were recorded in only half of these studies. The authors recommend that outcomes in future trials be based on a standardized scale of assessment of the treatment efficacy by the participants and physicians and that changes in quality of life as a result of the interventions be addressed. In addition they suggest that a greater emphasis be placed on the management and treatment of rosacea based on the staging pattern of the disease.
Reviews such as these are important summaries of evidence relating to the efficacy of particular treatments. They also reinforce the need for high-quality RCTs for future research. However, dermatologists should be aware that this review has limitations for clinical practice for two main reasons.
Firstly, clinical trials in rosacea are carried out almost exclusively on patients with papulopustular (subtype 2) disease, with lesion counts (papules and pustules) and grades of erythema being primary elements of assessment. The relationship of subtype 1 (erythematotelangiectatic rosacea) and subtype 3 (phymatous rosacea) to papulopustular rosacea is unclear,5 but it is unlikely that they represent ‘stages’ in progression of disease. Clinicians experienced in treating ‘rosacea’ have long recognized that the topical and systemic therapies deemed to be effective in papulopustular rosacea are ineffective or may even exacerbate other subtypes. The main conclusions of the study by Van Zuuren et al. thus apply only to subtype 2 rosacea.
Secondly, RCTs with large numbers of participants are costly and usually funded by pharmaceutical companies. As pointed out in a recent commentary on acne treatment in this journal6 this introduces a form of bias as companies are unlikely to fund evaluation of (possibly effective) cheap traditional treatments without commercial potential. Clinical experience would suggest that in the management of individual patients with rosacea several such treatments can be useful.
Finally, there is a danger that the demand by journals for statistically significant RCT studies as standard will inhibit the publication of innovative therapeutic interventions by (nonfunded) clinicians in small numbers of patients. Perhaps more than other specialists, dermatologists have had a special affection for these ‘anecdotal’ short reports. This probably reflects the vast number of conditions that affect the skin, their visible nature in evolution (to both the patient and the physician) and the variability of patient response. In our desire to conform to the ‘evidence-based’ requirements of journals let us remember the words of Oscar Wilde7 and take care that we do not ‘kill the thing we love’ by allowing the publication of such clinical observations in the traditional short report format to disappear from our premier journals.