What’s going on in rosacea?

‘A novel treatment for acne vulgaris and rosacea’ by Elewski1 records an impressive response to azithromycin therapy in 10 patients with acne rosacea. Although this was an uncontrolled, unmatched study, improvement was seen in all 10 patients treated, and nine of the 10 were clear after a 3-month intermittent course of therapy (5 days treatment, twice per month). The author points out that the success of azithromycin in inflammatory acne vulgaris is not unexpected in view of its high level of activity against Proprionibacterium acnes with a low minimum inhibitory concentration level and high penetration levels of the skin. The effectiveness of the pulsed (intermittent) therapy in acne vulgaris was due to the long elimination half-life of the drug he suggests, but does not postulate a mechanism of action in acne rosacea. Why is this or other antibiotic treatment effective in rosacea? The currently most widely accepted theory of the pathogenesis of rosacea is that inflammatory facial lesions result from the effects of material leaked from chronically dilated (telangiectatic) facial vessels that initiate an inflammatory cascade resulting in the formation of erythematous papules and pustules.2 The telangiectatic vessels are thought to come about as a consequence of the combined effect of frequent facial flushing and collagen disruption secondary to chronic actinic damage.3 So how do antibiotics work so effectively in clearing the skin of inflammatory lesions? Clearly, they do not significantly affect the telangiectasias or flushing (these persist after successful treatment of the papules and pustules) nor the damaged collagen, which remains unchanged after therapy. Most authors suggest that these drugs are active via anti-inflammatory mechanisms on the papules and pustules. After a variable period of remission, the papules and pustules gradually reappear in the skin in the majority of affected individuals after cessation of therapy. There are some difficulties with this view of the pathogenesis of rosacea and the effects of antibiotic therapy. Flushing and telangiectasias are not invariable precursors of inflammatory rosacea. In fact in some series flushing occurs only in the minority of patients.4 The distribution of the flush area on the face (lateral cheeks, sides of neck, full forehead) is different from the distribution of the papules and pustules that develop (mid-forehead, nose, medial cheeks, mid-chin), and many patients who flush frequently never develop inflammatory lesions at all. The development of uniform, small, dome-shaped erythematous papules and pustules consequent to ‘leaky vessels’ as postulated in rosacea is an exceptional model of cutaneous inflammation, which to my knowledge does not occur in other similar clinical circumstances. Histological examination of active pustules shows follicular orientated inflammatory change without prominent evidence of vascular damage, which again is hard to reconcile with a facial vascular-based aetiological mechanism. Oedema, which might be expected to be an early sign of leaky cutaneous vasculature, is not a consistent finding in rosacea, and when it does occur tends to be seen in advanced disease, is of a firm non-pitting fixed inflammatory type and does not usually parallel the distribution of rosacea, often being most prominent on the forehead and below the eyes. Finally, why are the dramatic ‘anti inflammatory’ effects of antibiotics on the papules and pustules of rosacea restricted to certain antibiotics, and not obtainable with other pharmaceutical agents with more potent anti-inflammatory properties? Is it possible that antibiotics work so effectively in rosacea by their antibacterial effects? This question has not been posed in a serious fashion because thinking on the pathogenesis of this disease has become so focused on vascular abnormalities as being the basis of the disease. The pustules of rosacea are thought to be sterile, but the microbiological studies on which this conclusion is based are more than 20 years old and recent studies using advanced retrieval and culture techniques are singularly lacking, as is the application of polymerase chain reaction techniques to the detection of bacterial DNA in the active lesions of this disease. The presence of bacteria with a significant pathogenic role could explain many of the features of rosacea. The clinical appearance of erythema papules and pustules could fit, as would the response to antibiotic therapy. Demodex mites, which are known to be increased in rosacea could be attracted to and multiply in follicles rich in bacteria, or they could be contaminated by these organisms and carry them in their migration across the face. The photo-distribution of rosacea (central forehead, nose, cheeks, chin, and bald scalp of male patients) could be explained by bacterial porphyrin production, which is known of certain Proprionibacterium species,5 and could initiate an inflammatory reaction within hair follicles and lead to the histological features of folliculitis that is a feature of inflammatory rosacea. Not so long ago the pathogenesis of gastroduodenal ulceration was directly linked to excessive acid production. It was not until innovative investigations probed the role of Helicobacter pylori, previously thought not to have pathogenic potential, that the true nature of the disease was uncovered. Our present understanding of the aetiology and pathogenesis of rosacea is regretfully lacking, and we need to reconsider all aspects of this disease, including the possible role of bacteria in the production of inflammatory lesions if we can begin to understand why antibiotic treatments as recorded in this issue of the JEADV are so effective.