CKS did not identify any authoritative, UK-based guidelines on the management of rosacea. Therefore, recommendations are based on evidence from randomized controlled trials (RCTs) where available [van Zuuren et al, 2005], and otherwise on the best available expert opinion described in narrative reviews [Berth-Jones, 2004; Powell, 2005; Diamantis and Waldorf, 2006].

• A thorough assessment of rosacea is a pragmatic recommendation that reflects the need to rule out other diagnoses (especially acne vulgaris) and accurately determine the predominant subtype of rosacea in order to carry out optimal management. Vasodilatory drugs, such as calcium-channel blockers, should be avoided where possible, as they can exacerbate flushing [Powell, 2005].
• Only papulopustular rosacea is receptive to treatments that are commonly used in primary care, with erythematotelangiectatic rosacea (i.e. persistent redness and flushing) being particularly difficult to treat [Baldwin, 2006]. In one RCT, people with poor outcomes were described as having 'redness rather than pustulation' as the predominant symptom of their rosacea [Pelle et al, 2004].

Topical treatments

• Treatment of papulopustular rosacea is mainly symptomatic in nature, as there is a lack of evidence that treatment improves the underlying pathology or the long-term prognosis of the disorder [Berth-Jones, 2004]. However, treatment can lead to remission in the shorter term, and most drugs can be used intermittently or continuously to control symptoms (see Follow up).
• Metronidazole is usually suitable as first-line treatment for mild-to-moderate papulopustular rosacea and for maintenance of more severe disease [Powell, 2005].
• The mechanism of action of topical metronidazole in rosacea is not fully understood, but it may act as an anti-inflammatory and antioxidant, rather than as an antibiotic [Nally and Berson, 2006].
• There is relatively good evidence from RCTs showing that topical metronidazole is more effective than placebo and equally as effective as oral tetracyclines; it causes fewer adverse effects than azelaic acid [van Zuuren et al, 2005].
• Azelaic acid is an alternative that may be more effective than metronidazole, but is more likely to cause adverse effects [Pelle et al, 2004].
• The mechanism of azelaic acid is likely to be similar to that of metronidazole. Azelaic acid has been shown to be superior to placebo. There is limited evidence from two head-to-head RCTs that azelaic acid is more effective than metronidazole, although further studies are required to confirm if this is clinically relevant [van Zuuren et al, 2005].
• Azelaic acid may cause burning, stinging, itching, scaling, and dry skin in 26–38% of people, although these effects are transient and do not usually affect compliance [Gupta and Gover, 2007].

Oral treatments

• Systemic treatment is recommended when there is moderate-to-severe papulopustular rosacea covering extensive areas that would be difficult to treat topically [Powell, 2005].
• Tetracyclines are effective at a dose that is subtherapeutic to their antibiotic effect; this is thought to be an anti-inflammatory effect [Baldwin, 2006].
• There is limited evidence from RCTs that tetracyclines are more effective than placebo, and as effective as both topical and oral metronidazole [van Zuuren et al, 2005]. The drugs used in these studies were tetracycline and oxytetracycline, which is reflected in their product licenses [BNF 55, 2008].
• Low-dose doxycycline has recently been reported, in two RCTs, to be effective in the treatment of rosacea [Del Rosso et al, 2007], and lymecycline is also likely to be effective. Lymecycline is not licensed for this use in the UK, however, doxycycline 40 mg modified-release capsules have recently been licensed in the UK for rosacea. Both drugs are more convenient to take than the older tetracyclines, an important consideration for a drug that is be used in the long term.
• Erythromycin is an alternative if tetracyclines are contraindicated (e.g. pregnancy, reactions to sunlight), and is recommended on the basis of clinical experience rather than controlled studies [Baldwin, 2006]. However, its effectiveness can be extrapolated from another macrolide, clarithromycin, which has been reported to be more effective, and better tolerated, than doxycycline in one RCT (n = 40) [Pelle et al, 2004].

Ocular rosacea

• Ocular rosacea is the usual cause of posterior blepharitis (Meibomian gland dysfunction), which is discussed fully in the CKS topic on Blepharitis.
• The eye is affected in about half of people with rosacea [Berth-Jones, 2004], and requires treatment to relieve symptoms and prevent deterioration.
• Evidence for the effectiveness of treatment for blepharitis is, in general, lacking. However, CKS identified one RCT (n = 35) that found that oxytetracycline was associated with more remissions than placebo [Bartholomew et al, 1982].

There is a lack of evidence from controlled trials to support lifestyle changes, as they are generally not suitable interventions for this type of study. Recommendations are therefore based on physiological and pharmacological principles, pragmatism, and clinical experience, as described in narrative reviews [Berth-Jones, 2004; Powell, 2005; Diamantis and Waldorf, 2006].

• Explaining the natural history of the disease will usually reassure the person. Although the categorization of rosacea implies a progressive disorder, this is not necessarily the case. For instance, many people with facial flushing do not develop other symptoms; this has led some experts to believe the term 'pre-rosacea' (that is sometimes used) is inappropriate [Powell, 2005]. Where severe disease does develop (e.g. rhinophyma), the person can usually be treated by surgery or laser treatment in secondary care [Berth-Jones, 2004].
• Solar radiation has been implicated in the pathogenesis of rosacea, and areas of exposed fair skin are most affected [Diamantis and Waldorf, 2006]. The use of sunscreens should limit further skin damage from sunlight, although this has not been verified by controlled studies.
• The possible trigger factors (temperature, exercise, stress, alcohol, foodstuffs) all have the potential to physically irritate skin or dilate cutaneous blood vessels and elicit blushing [Powell, 2005]. It is therefore reasonable to advise the person to avoid them, at least on a trial basis.
• Dry skin is a common complaint of rosacea, therefore CKS recommends an emollient should be used as appropriate. However:
• People with rosacea tend to have sensitive skin [Jappe et al, 2008], so products with potential irritants (e.g. perfumes) should be avoided.
• Topical corticosteroids (e.g. hydrocortisone) should not be used, except for severe inflammatory rosacea (under expert supervision) [Nally and Berson, 2006], as they can trigger, worsen, or mimic, the condition [Powell, 2005].

CKS did not identify any authoritative, UK-based, guidelines on the management of rosacea. Recommendations on follow up and subsequent maintenance of rosacea are based on expert opinion described in narrative reviews [Berth-Jones, 2004; Powell, 2005; Diamantis and Waldorf, 2006].

• CKS recommends people should be followed up after 12 weeks, as this reflects a reasonable individual trial period to assess whether treatment has been effective.
• Most randomized controlled trials (RCTs) of interventions for rosacea have assessed response at 4–15 weeks, but mainly at the higher end of this range [van Zuuren et al, 2005].
• The effect of treatment for rosacea often has a gradual onset, and some experts believe there is little point in discontinuing treatment before 3 months [Berth-Jones, 2004]. However, prescribing longer courses of treatment without monitoring is not recommended, especially for systemic drugs, as they may cause adverse effects without significant benefit [Powell, 2005].
• Rosacea is a chronic, relapsing condition, which in general does not spontaneously resolve. However, RCTs studying the disorder have generally been performed over the short term, so recommendations on maintenance are based mainly on clinical experience and pragmatism [Powell, 2005].
• Most people will benefit from maintenance or intermittent treatment, as rosacea often relapses after a successful course of treatment. For example, two studies found that the rate of relapse after stopping a course of tetracycline was 25% after 1 month, 50–60% after 6 months, and 70% after 1–4 years [Baldwin, 2006].
• One longer-term RCT randomized people who had been successfully treated with oral tetracycline and topical metronidazole (n = 88) to receive maintenance treatment with metronidazole gel or placebo (gel vehicle). After 6 months, metronidazole gel significantly reduced the likelihood of relapse and the number of residual lesions compared with placebo [Dahl et al, 1998].
• Combining oral and topical treatment has not been studied in controlled trials, but is a reasonable practice for people with severe disease.

CKS did not identify any national referral criteria or guidelines for rosacea. In the absence of established policy, these recommendations are based on pragmatism and what is accepted in the UK as good clinical practice.

• A dermatologist is in a position to offer several pharmacological treatments that are not suitable for the treatment of rosacea in primary care. In general, there is a lack of evidence from controlled trials to support these treatments, rather than evidence of no effect. Therefore, an expert may trial them on an individual basis. These include [Pelle et al, 2004; Powell, 2005; Baldwin, 2006]:
• Other topical treatments, such as benzoyl peroxide, topical antibiotics (other than metronidazole), tacrolimus, or retinoids (e.g. tretinoin).
• Other oral antibiotics, such as clarithromycin, azithromycin (useful if erythromycin is poorly tolerated), or minocycline.
• The combined oral contraceptive pill (if a hormonal cause is suspected in a woman).
• Oral isotretinoin or clonidine (for flushing).
• Cardiovascular drugs to prevent flushing (e.g. beta-blockers or spironolactone).
• Some forms of rosacea are resistant to pharmacological treatment. Referral to secondary care allows for other options to be considered:
• Persistent erythema is often mistakenly attributed to heavy drinking, and this can cause considerable stigma [Powell, 2005; Baldwin, 2006]. Erythema can be masked with camouflage makeup; the dermatology department should be able to provide the person with the relevant local contact information (see www.timewarp.demon.co.uk/redcross.html for details of the Red Cross Beauty Care and Cosmetic Camouflage Service).
• Treatment options for telangiectasia and phymatous disease in secondary care include laser treatment and corrective electrosurgery [Pelle et al, 2004; Powell, 2005]. However, these are not generally available on the NHS.