Rosacea: Current state of epidemiology Jerry Tan, MD,a and Mats Berg, MDb Windsor, Ontario, Canada, and Uppsala, Sweden Case definitions are critical in epidemiologic research. However, modern disease indicators must now consider complex data from gene-based research along with traditional clinical parameters. Rosacea is a skin disorder with multiple signs and symptoms. In individuals, these features may be multiple or one may predominate. While studies on the epidemiology of rosacea have previously been sparse, there has been a recent increase in research activity. A broader body of epidemiological information that includes a greater variety of countries beyond Northern Europe and general population-based demographics is needed. As there are operational issues in current case definitions of rosacea subtypes—rationalization and standardization—universal consistent applications in future research is also imperative. Further improvement in disease definition combining new research information along with clinical pragmatism should increase the accuracy of rosacea case ascertainment and facilitate further epidemiological research. ( J Am Acad Dermatol 2013;69:S27-35.) Key words: comorbidities; epidemiology; prevalence; risk factors; rosacea; subtype.

T

he World Health Organization defines ‘‘epidemiology’’ as the study of distribution and determinants of disease in populations and the application of this information to disease control or prevention. Methods for study of disease distribution include surveillance and descriptive investigations, whereas determinants are evaluated by analytical investigations. A pivotal issue in epidemiology is definition of the target disease for case ascertainment. Although disease was historically defined as a cluster of symptoms, it has now also come to include laboratory, pathological, and/or genetic abnormalities. Rosacea is an inflammatory facial skin disorder that comprises multiple clinical featureseany one of which may predominate in a particular patient. These include flushing or transient facial erythema, persistent central facial erythema, inflammatory papules and pustules, and telangiectasia. Additional features include burning or stinging sensations, erythematous plaques, facial dryness and scaling, edema, peripheral location, phymatous changes, and ocular features.1,2 Although the cause of rosacea is unclear, recent progress in understanding pathogenesis includes aberrations in innate immunity, dermal

From the Department of Medicine, University of Western Ontarioa and the Division of Dermatology, University of Uppsala Academic Hospital.b Publication of this article was supported by a grant from Galderma International. Editorial support provided by Galderma International. Disclosure: Dr Tan has served on an advisory board and as a consultant and speaker for Galderma and received grants and

Abbreviations used: CI: GST: NRSEC: OR: UV:

confidence interval glutathione S-transferase National Rosacea Society Expert Committee odds ratio ultraviolet

matrix degradation, vasodilation, and fibrogenesis.3,4 Nevertheless, epidemiologic data on rosacea remain sparse and somewhat disparate. In this review, our purpose is to provide an updated perspective on the current state of rosacea epidemiologyeits distribution and potential determinants. Within this context, we also highlight shortcomings in methodology and current gaps in knowledge to guide future research.

ROSACEA: CASE DEFINITION AND SUBTYPES The hallmark of rosacea is central facial persistent erythema, typically affecting the facial convexities such as the cheeks, chin, forehead, and nose while sparing the perioral and periocular regions.5-7 This feature has been considered to be the sole requisite criterion for the diagnosis of rosacea.7 In 2002, honoraria in compensation. Dr Berg has no conflicts of interest to declare. Accepted for publication April 21, 2013. Reprint requests: Jerry Tan, MD, Department of Medicine, University of Western Ontario, 2224 Walker Rd, Suite 300, Windsor, Ontario, Canada N8W 5L7. E-mail: [email protected]. 0190-9622/$36.00 Ó 2013 by the American Academy of Dermatology, Inc. http://dx.doi.org/10.1016/j.jaad.2013.04.043

S27

J AM ACAD DERMATOL

S28 Tan and Berg

DECEMBER 2013

National Rosacea Society Expert Committee (NRSEC) for inaccuracy based on recall, flushing may not be a developed a classification and staging system to useful diagnostic criterion.2 The overlapping criteria for subtype 1 and subserve as a provisional diagnostic instrument for type 2 whereby both share a common feature of clinical, epidemiologic, and interventional investigapersistent central facial erythema and are distintions.1 One goal of that initiative was to further study the epidemiologic associations and relationships of guished by flushing in subtype 1, and papules and rosacea subtypes and potential variants. This group pustules in subtype 2, is a source of potential defined 4 subtypes of rosaconfusion. When combined cea and 1 variant based on with the NRSEC statement CAPSULE SUMMARY clinical features that they diregarding the potential for vided into primary and secconcurrent subtypes, these Epidemiology for rosacea has been thinly ondary. Furthermore they serve to rationalize combinexplored, and data are therefore scarce designated that patients may ing subtype 2 into subtype 1. and uncoordinated. present with characteristics However, this aggregation of The available data are also prone to of more than a single rosacea subtypes reduces the potencertain analytical deficiencies. Consistent subtype. Subtype 1, or erythtial for detecting different case identification criteria and ematotelangiectatic rosacea, clinical and determinant asminimization of biases are 2 important is mainly characterized by sociations. Accordingly, for parts of accurate epidemiologic data. flushing and persistent cenclarity in analysis, subtypes tral facial erythema. Subtype 1 and 2 should be considered Further development of case 2, or papulopustular rosacea, distinct. identification criteria and a broader body is characterized by persistent The definition of subtype of research should benefit rosacea central facial erythema with 2epersistent central facial erdiagnosis and disease management. transient papules/pustules in ythema and papulopustulese a central facial distribution. requires further definition of Differentiation of this subtype from acne vulgaris is time frame for papulopustules. This is particularly based on the following: an older age group is relevant if such lesions are not visible at the time of typically affected, and absent are comedones and diagnosis. An open-ended question on whether the truncal lesions. Furthermore, this subtype may be patient had ever had papulopustules may elicit falseseen concurrently with or after development of positive responses in the form of a history of acne subtype 1. Subtype 3, or phymatous rosacea, comvulgaris or folliculitis. In a recent survey of patients prises thickened skin with irregular nodules and with rosacea, 40% of those with subtype 2 reported localized enlargement. Subtype 4, or ocular rosacea, not having papules and pustules throughout the year. is diagnosed based on 1 or more of the following: The median duration of remission reported was 3 watery or bloodshot appearance, foreign body senmonths.9 Accordingly, restricting the time frame for questioning regarding papulopustules to the previous sation, burning or stinging, dryness, itching, light 6 months may enhance diagnostic accuracy. sensitivity, blurring of vision, conjunctival and marFor ocular rosacea, subtype 4, a requisite for ginal telangiectasia, lid or periocular erythema, concomitant presence of cutaneous features of rosablepharitis, conjunctivitis, chalazion, hordeolum, cea may enhance diagnostic accuracy. and corneal complications including corneal infilDetermination of persistent central facial erytrates ulcers and keratitis. Granulomatous rosacea thema is particularly difficult in dark phototypes. was recognized as the sole variant and is characterA practical solution may be to modify the essential ized by firm, indurated papules/nodules. diagnostic criteria in dark phototypes to include Potential shortcomings of this classification scheme secondary features such as burning/stinging, dry include: inclusion of flushing as a diagnostic criterion appearance, edema, and sensation of flushing. This for subtype 1; overlapping criteria for subtypes 1 and issue will require further inquiry when diagnostic 2; the lack of temporal framing for papules/pustules in accuracy is improved by genetic and/or immunohissubtype 2; and the absence of defined validated and tochemical methods. predictive features for subtype 4. The inclusion of flushing as a diagnostic criterion is controversial with some authors suggesting it PREVALENCE AND DISTRIBUTION OF unnecessary,2,7,8 whereas others suggest it is imperROSACEA ative.5 Nevertheless, elicitation of an accurate history Reports of rosacea prevalence range from less of flushing may not be straightforward.5 In view of than 1% to more than 20%, but direct comparison the subjective nature of this feature, and the potential between studies is confounded by differences in d

d

d

J AM ACAD DERMATOL

Tan and Berg S29

Extension of Schaefer et al13 study; prospective Dermatologist examination; prospective Dermatologist examination; subtype 1 was not evaluated; prospective Retrospective database Dermatologist examination; prospective Database Dermatologist examination; prospective Dermatologist examination; prospective NA NA NA 18 NA NA NA NA 82 NA 1.34 2.2 2.1 10 0.09

NA 22% 2.7 NA 78% NA 2.3 22 2.7

methodology, populations, and cultural and social perceptions of disease.10,11 Furthermore, studies on prevalence of rosacea subtypes may be inherently flawed by operational shortcomings noted above. Findings from general population surveys, which are more reflective of overall prevalence, will be presented distinct from surveys obtained in dermatology clinics that represent a selected population.

NA, Not available.

United States Germany United States Sweden Denmark (Faroe Islands) Romanowicz et al,19 2008 Schaefer et al,13 2008 Bamford et al,18 2006 Berg and Liden,16 1989 Lomholt,12 1964

91,000 Workers 348 Workers 1000 Office and outdoor workers 9,151,174 48,665 Workers 1450 809 Office workers 10,984 Germany Estonia Ireland 14

Augustin et al, 2011 Abram et al,17 2010 McAleer et al,15 2010

N Country Study

Table I. General population surveys of rosacea prevalence

Prevalence of rosacea, %

Subtype 1, %

Subtype 2, %

Comment

VOLUME 69, NUMBER 6

General population surveys General population prevalence is optimally determined by surveying the entire population using specific case finding criteria with evaluations performed by trained observers. Although the latter 2 criteria have been fulfilled in some studies, the groups surveyed have been largely limited to those of working populations (Table 1). The earliest comprehensive population-based prevalence study in rosacea was performed in 1948 and involved dermatologic examination of almost 11,000 residents of the Faroe Islands. In that study, the prevalence of rosacea was 0.09%. However, criteria for diagnosis was not explicitly stated and may have selected for the most severe cases.12 Systematic whole-body dermatologic assessments for all cutaneous abnormalities by dermatologists in 150 German companies involving almost 49,000 subjects provided a rosacea point prevalence rate of 2.2%.13 Rosacea subtyping was not performed and case finding criteria were not explicitly stated. An extension of that study involving 90,880 subjects in over 300 German companies showed an almost identical prevalence of 2.3% (95% confidence interval [CI] 2.2-2.4).14 In a cohort study conducted in Ireland conducted by dermatologists, evaluating hospital and outdoor workers totaling 1000 subjects, the prevalence of subtype 2 was 2.7%. Data on subtype 1 were unavailable as the authors thought it impossible to clinically distinguish subtype 1 from heliodermatitis.15 Smaller cross-sectional surveys conducted by dermatologists have observed higher prevalence rates. A survey of 809 Swedish office employees examined by a dermatologist determined that 10% had rosacea. Gender distribution was 14% women and 5% men.16 In a cross-sectional survey conducted by a dermatologist of 348 employees from commercial institutions in Estonia, the overall prevalence of rosacea was 22%. Notable was the exclusion of subjects younger than 30 yearsethe rationale for which was not provided. Furthermore, flushing was not considered a prerequisite for subtype 1.17 Large retrospective US database studies have compiled prevalence rates of 1.3% to 2.1%.18,19 However, prevalence rates based on administrative

J AM ACAD DERMATOL

S30 Tan and Berg

DECEMBER 2013

Table II. Specific population surveys of rosacea prevalence Study

Country

N

Prevalence of rosacea, %

Subtype 1, %

Subtype 2, %

Comment

0.2

12

69

2

NA

NA

Subtype 3 was 4%; retrospective database Dermatologist examination; retrospective database Dermatologist examination; database Diagnostic ascertainment unclear Evaluated chief symptom; retrospective database Evaluated chief symptom; retrospective database

Khaled et al,23 2010

Tunisia

Gutierrez et al,25 2010

Peru

Kyriakis et al,22 2005

Greece

50,237

1.2

72

28

Lal Khatri,24 2004 Doe et al,21 2001

Yemen United Kingdom Ghana

13,840 3383

0.4 1.8

NA NA

NA NA

0

NA

NA

Doe et al,21 2001

244 3294

2254

NA, Not available.

databases may also be flawed as they are not representative of the entire population and may underrecord comorbidities or conditions considered to be of lesser morbidity or mortality risk.20 Prevalence of rosacea: Specific population surveys A number of studies have evaluated the prevalence of rosacea in dermatology clinics (Table II). However, determining such rates is not reflective of prevalence in the general population. Further shortcomings include the following potential biases: populations that may be more care-seeking, preselection for subjects with other skin disorders, selection for more severe presentations of rosacea, and variations in sociocultural norms. In a comparative evaluation of skin diseases diagnosed in dermatology clinics in the United Kingdom (N = 3383) compared with Ghana (N = 2254), rosacea was diagnosed in 1.8% compared with 0%, respectively.21 Criteria used to establish rosacea diagnosis were not explicitly stated. However, patients in the former were from secondary referrals, usually general practitioners, whereas those in the latter were self-referred. Differences in locale and in referral source may have led to selection of diagnoses of greater perceived severity. In a cross-sectional survey of rosacea performed in an outpatient dermatology clinic in Greece of more than 50,000 cases, rosacea prevalence was 1.2%. Subtype 1 accounted for 72% of cases, and subtype 2 for 28%.22 In contrast, a retrospective chart survey (N = 244) of an outpatient dermatology clinic of a Tunisian hospital provided a prevalence rate of only 0.2% and distribution of 12% for subtype 1 and 69% for subtype 2.23 Similarly in Yemen, a dermatology clinic sample of almost 14,000 consecutive new patients yielded a rosacea prevalence of only 0.4%.24 A database survey of diagnostic codes (N = 3294) over

a 3-year period in a hospital dermatology clinic setting revealed a 2% prevalence of rosacea in Lima, Peru. Subtype distribution was not detailed in that study. This is likely an underestimate as more than one dermatologic diagnosis does not appear to have been captured based on data presentation.25 These retrospective methods based on diagnostic coding or chief symptom may underestimate rosacea prevalence as other dermatologic conditions may have motivated the clinic visit. Prevalence by subtype Accurate information on rosacea subtype prevalence is hindered by variations in case finding definitions and the potential for aggregating subtype 2 into subtype 1.26 Overall, subtype 1 is most prevalent followed by subtype 2.17,22 In the general population studies evaluating subtypes, the prevalence of subtype 1 was approximately 4-fold greater than subtype 2.16,17 Although subtype 3, phymatous rosacea, is uncommon,17,22,27 males are predominantly affected. In the study of Estonian workers where 22% had rosacea, only 1 (1%) was classified as subtype 3.17 Of the 4 rosacea subtypes, subtype 4 ocular rosacea is particularly difficult to appraise because of absence of definitive diagnostic criteria. Surveys on the prevalence of rosacea typically evaluate subtype 4 ocular rosacea based on the prevalence of general ocular signs and symptoms. Accordingly, the potential for false-positive reporting is high and may explain the wide reported prevalence range of 6% to 58%.28 A recent study, establishing the diagnosis of ocular rosacea as presence of ocular features in the presence of cutaneous criteria, determined a 33% prevalence of ocular disease among 100 patients with rosacea. The most frequent symptoms were burning and tearing whereas the most frequent signs were conjunctivitis and blepharitis. Patients with

J AM ACAD DERMATOL VOLUME 69, NUMBER 6

rosacea had tear excretion rates and tear break-up times that were significantly lower than that of healthy individuals, implying a higher degree of eye dryness.27 In a study of 176 randomly selected patients with rosacea equally distributed from departments of ophthalmology and dermatology, comparison of ocular findings revealed a higher prevalence of meibomian gland dysfunction (P \.001), telangiectasia (P = .004), interpalpebral conjunctival hyperemia (P = .005), and anterior blepharitis (P = .008) in ophthalmology patients. The major and most easily observable ocular problems in patients with rosacea were lid diseaseerelated manifestations.29 A large database survey linking rosacea with ocular diagnostic codes established that ocular disorders were diagnosed in 21%, compared with 13% in control subjects. The odds ratio (OR) was 1.82 (95% CI 1.76-1.88) and prevalence of ocular symptoms was similar between genders. Blepharitis was the ocular diagnosis with the largest difference between rosacea cases and controls with OR of 3.57 (95% CI 3.17-4.02).30 These findings could support an update in the diagnostic criteria for ocular rosacea. In particular, validation of the positive predictive value of specific ocular features for rosaceaesuch as meibomian gland dysfunction, telangiectasia, interpalpebral conjunctival hyperemia, and anterior blepharitise may enhance diagnostic accuracy for subtype 4. Incidence The incidence of rosacea was recently evaluated using the United Kingdomebased general practice research database.30 This study identified 60,042 cases of rosacea between 1995 and 2009 with a calculated rosacea incidence rate of 1.65 per 1000 person years. A high incidence of rosacea was observed in patients older than 30 years (80%) as was a significant presence of ocular symptoms (21%).30 Unfortunately, this study was not able to distinguish rosacea subtypes.30

DISEASE DETERMINANTS AND RISK FACTORS Multiple risk factors for rosacea have been proposed, including age, phototype, gender, alcohol consumption, genetics, and ultraviolet (UV) exposure. Several of these hypotheses have been addressed in recent studies. Age and gender Although rosacea is generally stated to be most prevalent in adults older than 30 years,15 this may reflect reporting bias as large studies have only been conducted in working adults. Reports of rosacea in

Tan and Berg S31

childhood exist,31-33 but prevalence is unknown as there have been no formal pediatric surveys. A Greek dermatology clinic database study reported a rate of 4% in the youngest reported age range ( # 20 years).22 In one of the largest database surveys on rosacea, involving over 60,000 cases, approximately 10% were in the youngest age range reported (\20 years). Unfortunately, no further segregation of this range was reported.30 Gender distribution in rosacea has variably been reported as equal or female predominant. The largest prospective studies did not differentiate between subtypes and found an almost equivalent distribution between genders. In the German skin screening survey of 91,000 subjects, gender distribution for rosacea was 2.4% male and 2.1% female.14 A preliminary publication of the first 47,000 of that cohort showed a similar prevalence of 2.3% male and 2.0% female.13 In the large United Kingdomebased general practice research database study of over 60,000 rosacea cases, 62% of rosacea cases were female.30 In other prospective studies of rosacea in the general population conducted by dermatologists, a 2- to 3-fold greater prevalence in females was observed in one general population survey16 and 1.5-fold in another.17 In the former, the proportion of females was 2 and 3 times greater than males for subtypes 1 and 2, respectively; whereas the latter did not differentiate genders between subtypes. In the Irish survey of subtype 2, gender distribution for cases was 56% male and 44% female.15 Further information on the distribution of rosacea was provided in prevalence and gender distribution could be divided into 3 distinct periods: younger than 35 years, when both genders were affected at low rates; 36 to 50 years, when females predominated and frequencies increased; and older than 50 years, when frequency was highest and genders equally affected.22 Skin types Rosacea is readily visible in lighter phototypes.1,34,35 Although some cases have been diagnosed in darker skin types, such reports are infrequent.11,34 This disparity among phototypes, however, may be a result of sampling bias as large, prospective, general population surveys have not been conducted in populations with dark phototypes. UV exposure UVexposure is a commonly accepted risk factor in patients with rosacea, but its pathogenic relationship remains uncertain. It is considered a known trigger in part because of rosacea prevalence in light phototypes and the predominance of erythema on facial

J AM ACAD DERMATOL

S32 Tan and Berg

convexities.11,26,34 Although excessive UV-induced skin damage can mimic and confound the diagnosis of rosacea, UV may also induce rosacea by triggering innate immune responses.3,34 In a study analyzing potential risk factors for rosacea, a significantly greater risk for people with outdoor occupations was observed (OR 3.40; 95% CI 1.21-9.54) by multivariate analysis. This factor may be considered a proxy for greater UV exposure. However, subtype-specific analysis of this association was not performed.35 This association was corroborated in a cohort of 168 Korean patients with rosacea. However, they found that the degree of sun exposure was correlated with severity of subtype 1 but not subtype 2.26 The latter is also supported by the lack of correlation of UV exposure or photodamage with subtype 2 in an Irish study.15 These findings imply a subtype 1especific association between rosacea and UV exposure.

INFECTIONS AND INFESTATIONS Demodex The role of Demodex in rosacea has recently been extensively reviewed elsewhere and a summary is presented herein.36 Demodex mites, obligatory parasites of human pilosebaceous follicles, comprise Demodex folliculorum and D brevis. Their main food sources are epidermal cells and sebum, obtained with sharp mouth and feeding parts. In this process, the epithelial barrier may be disrupted allowing the mite to penetrate intradermally. Mite density is very low in young adults and progressively rises in the sixth decade of life, an age range associated with peak prevalence of rosacea, suggesting a temporal relationship between Demodex infestation and rosacea. A study of D folliculorum prevalence and mite densities in 100 patients with rosacea compared with healthy control subjects and patients with acne and discoid lupus showed a significantly higher prevalence and mean mite density in rosacea.27 The potential association between Demodex infestation and rosacea was also recently evaluated by means of a systematic search and meta-analysis of case-control studies analyzing this association.37 This study found 48 articles fulfilling predefined selection criteria for analysis, involving 28,527 participants. Pooled OR demonstrating a significant association between Demodex infestation and rosacea was 7.57 (95% CI 5.39-10.62). Demodex has been shown to harbor a gramnegative bacterium (Bacillus oleronius) that produces proinflammatory proteins capable of inducing peripheral blood mononuclear cell proliferation in 73% of patients with papulopustular rosacea compared with 29% of control subjects.38 Potential

DECEMBER 2013

causality of Demodex may be related to several mechanisms: mechanical pilosebaceous blockage resulting in injury of follicular distention and epithelial rupture, or elicitation of foreign body immune reaction in the presence of its chitinous exoskeleton, its waste products, or associated bacteria.36 Staphylococcus epidermidis A recent study isolated S epidermidis in 9 of 15 patients with papulopustular rosacea from pustules. In contrast, there was either no growth or mixed growth from ipsilateral unaffected areas, or from age- and sex-matched control subjects.39 S epidermidis strains isolated from patients with rosacea have previously been shown to secrete more proteins and were more consistently beta-hemolytic than those from control subjects.40 This combination of findings suggests that the initial vascular stage of rosacea and increased surface skin temperature may lead to increased bacterial peptide production and subsequent development of inflammatory papules and pustules. Helicobacter pylori Recognition of H pylori as a pathogen in gastroduodenal disease led to exploration of possible associations with various other conditions including rosacea. The latter association with rosacea has been disputed for almost 2 decades. When compared with age- and sex-matched control subjects, the prevalence of H pylori in patients with rosacea has been shown to either be elevated41-45 or not different.17,27,46-50 Although some studies have shown improvement in rosacea after eradication of H pylori,45,51,52 this outcome may be a result of the antiinflammatory effects of antibiotics used in treatment.53

FAMILY HISTORY Patients with rosacea have been found to have significantly greater odds of a positive family history compared with nonrosacea control subjects (OR 4.31; 95% CI 2.34-7.92; P \ .0001).35 This finding, however, may imply either underlying genetic mechanisms either directly (see below) or indirectly (eg, through phototype) or environmental commonalities in a kindred. One underlying genotypic mechanism that may be operative in rosacea is polymorphism in glutathione S-transferases (GSTs). These are genes that encode for enzymes that normally catalyze detoxification of various UVinduced oxidative stress products such as reactive oxygen species. Their abnormal function may explain the association between UV exposure and rosacea. GSTM1 and GSTT1 are polymorphisms of GST and have been associated with a higher risk of

J AM ACAD DERMATOL

Tan and Berg S33

VOLUME 69, NUMBER 6

rosacea compared with control subjects, with OR of 2.84 (95% CI 1.37-5.89) and 2.68 (95% CI 1.27-5.67), respectively.54 In contrast, a major role for environmental factors is implied by the observation that rosacea affected only 1 of a pair of identical (monozygotic) twins, 2 women aged 51 years. Of interest is that these siblings lived in different environments, rural for the affected twin and urban for the unaffected.55

ADDITIONAL FACTORS Other exacerbating factors that have been associated with rosacea include alcohol intake, coffee consumption, heat exposure, exercise, and smoking. Alcohol, hot beverages, exercise, and heat exposure are all thought to induce rosacea via dilation of cutaneous vessels.11,26 In multivariate analysis, patients with rosacea have been found to have significantly greater odds of smoking history (OR 2.01; 95% CI 1.07-3.80).35 The United Kingdom database study also found that current smokers had a lower risk of developing rosacea compared with nonsmokers with OR of 0.64 (95% CI 0.62-0.67). In contrast, the risk for exsmokers was slightly increased (OR 1.14; 95% CI 1.10-1.18).30 A case-control study of 103 patients with rosacea found that the odds of rosacea was not significantly different between smokers and exsmokers; however it was significant when exsmokers were defined as having stopped smoking for more than 1 year (OR 2.42; 95% CI 1.06-5.5).56 This positive effect is thought to be a result of the immunosuppressive nature of cigarettes, and may be related to similar effects in other cutaneous disorders.10,35 Although a database diagnostic code survey revealed no increased association of rosacea with alcohol abuse,57 the United Kingdomebased general practice research database study determined that the risk for rosacea increased with alcohol consumption with the highest OR of 1.51 (95% CI 1.41-1.63) for those consuming more than 25 U per week.30

COMORBIDITIES Migraine Significant associations between rosacea and migraine have been reported. In a Swedish study, migraine affected 27% of patients with rosacea compared with 13% of healthy control subjects.16 This result concurred with a case-controlled study comparing patients with rosacea to healthy control subjects, wherein migraine was observed in 60 of 137 patients with rosacea (44%) versus 21 of 161 control subjects (13%).58

Depression Psychological factors have been considered to be potential triggers in rosacea and some studies have shown higher rates of depression among patients with rosacea.57,59 The largest of these was a diagnostic code linkage analysis for rosacea and depression from databases collected by the National Ambulatory Medical Care Survey and the outpatient component of the National Hospital Ambulatory Care Survey. In that study, the OR for depression in the rosacea group was 4.81 (95% CI 1.39-16.62).57

NATURAL HISTORY There are no longitudinal studies addressing the potential for progression between rosacea subtypes. However, some evidence of progression was observed in a survey of 113 rosacea subjects based on recall of rosacea-associated clinical features. In that study, 31% fulfilled criteria for both subtype 1 and subtype 2 over the time course of their disease. Of those, 66% developed subtype 1 before subtype 2. Furthermore, for the 20% fulfilling criteria for both subtype 1 and subtype 3, 92% developed subtype 1 before subtype 3. For the 34% fulfilling criteria for both subtype 2 and subtype 3, 83% developed subtype 2 before subtype 3. Of subjects reporting ocular signs and symptoms suggestive of subtype 4, the majority developed these after the onset of cutaneous features of rosacea (75% for subtype 1; 71% subtype 2; 81% subtype 3).9 Conclusions Epidemiologic studies in rosacea provide widely varying prevalence estimates. Beyond actual differences in disease prevalence, this variation may be a result of differences in case ascertainment, study design, environmental factors, and population phototype. The largest studies with least risk of bias were conducted by dermatologists in Northern Europe among working populations with light phototypes. However, different case finding criteria were used. They excluded pediatric subjects and those who had retired. In addition, the nature of the commercial institutions surveyed and occupations evaluated may have led to gender bias. Furthermore, although prevalence of rosacea appears to be greatest in those with lighter phototypes, this may be a result of reporting bias as these studies were performed in Northern Europe. Overall, prevalence of rosacea in childhood, darker phototypes, and geographic locations beyond the US and Europe is generally lacking. Continued exploration of potential associations of rosacea with micro-organisms such as Demodex, B oleronius, S epidermidis, and H pylori may provide further insight into the cause of innate immune

S34 Tan and Berg

system up-regulation in rosacea. Furthermore, the roles of these microbial factors may be interrelated as codependent factors in rosacea. Future rosacea research requires progressive refinement of current subtype definitions. Some of these modifications are pragmaticedefining subtype 1 as diffuse central facial erythema with periodic exacerbations; distinguishing subtype 1 and subtype 2 as distinct and nonoverlappingewhereas others are iterative based on research findings. These include a time reference for papules and pustules and determination of definitive diagnostic features for ocular rosacea. Case ascertainment with particular attention to the operationalization of subtype criteria can provide a further basis on which to extend the initial objectives of the NRSEC in the conduct of research, comparison of data, and provision of optimal care for patients with rosacea. REFERENCES 1. Wilkin J, Dahl M, Detmar M, Drake L, Feinstein A, Odom R, et al. Standard classification of rosacea: report of the National Rosacea Society expert committee on the classification and staging of rosacea. J Am Acad Dermatol 2002;46:584-7. 2. Powell FC. Clinical practice: rosacea. N Engl J Med 2005;352: 793-803. 3. Steinhoff M, Buddenkotte J, Aubert J, Sulk M, Novak P, Schwab VD, et al. Clinical, cellular, and molecular aspects in the pathophysiology of rosacea. J Investig Dermatol Symp Proc 2011;15:2-11. 4. Yamasaki K, Gallo RL. Rosacea as a disease of cathelicidins and skin innate immunity. J Investig Dermatol Symp Proc 2011;15:12-5. 5. Kligman AM. An experimental critique on the state of knowledge of rosacea. J Cosmet Dermatol 2006;5:77-80. 6. Marks R. The enigma of rosacea. J Dermatolog Treat 2007;18: 326-8. 7. Crawford GH, Pelle MT, James WD. Rosacea, I: etiology, pathogenesis, and subtype classification. J Am Acad Dermatol 2004;51:327-41; quiz 42-4. 8. Rebora A. The red face: rosacea. Clin Dermatol 1993;11: 225-34. 9. Tan J, Blume-Peytavi U, Ortonne JP, Wilhelm K, Marticou L, Baltas E, et al. An observational cross-sectional survey of rosacea: Clinical associations and progression between subtypes. Br J Derm doi: 10.1111/bjd.12385. Published online April 9, 2013. 10. Chosidow O, Cribier B. Epidemiology of rosacea: updated data. Ann Dermatol Venereol 2011;138(Suppl):S179-83. 11. Elewski BE, Draelos Z, Dreno B, Jansen T, Layton A, Picardo M. Rosaceaeglobal diversity and optimized outcome: proposed international consensus from the rosacea international expert group. J Eur Acad Dermatol Venereol 2011;25:188-200. 12. Lomholt G. Prevalence of skin diseases in a population; a census study from the Faroe Islands. Dan Med Bull 1964;11: 1-7. 13. Schaefer I, Rustenbach SJ, Zimmer L, Augustin M. Prevalence of skin diseases in a cohort of 48,665 employees in Germany. Dermatology 2008;217:169-72. 14. Augustin M, Herberger K, Hintzen S, Heigel H, Franzke N, Schafer I. Prevalence of skin lesions and need for treatment in a cohort of 90,880 workers. Br J Dermatol 2011;165:865-73.

J AM ACAD DERMATOL

DECEMBER 2013

15. McAleer MA, Fitzpatrick P, Powell FC. Papulopustular rosacea: prevalence and relationship to photodamage. J Am Acad Dermatol 2010;63:33-9. 16. Berg M, Liden S. An epidemiological study of rosacea. Acta Derm Venereol 1989;69:419-23. 17. Abram K, Silm H, Oona M. Prevalence of rosacea in an Estonian working population using a standard classification. Acta Derm Venereol 2010;90:269-73. 18. Bamford JT, Gessert CE, Renier CM, Jackson MM, Laabs SB, Dahl MV, et al. Childhood stye and adult rosacea. J Am Acad Dermatol 2006;55:951-5. 19. Romanowicz M, Stephenson JJ, Del Rosso JQ, Lenhart G. Healthcare utilization and costs of patients with rosacea in an insured population. J Drugs Dermatol 2008;7:41-9. 20. Grimes DA. Epidemiologic research using administrative databases: garbage in, garbage out. Obstet Gynecol 2010; 116:1018-9. 21. Doe PT, Asiedu A, Acheampong JW, Rowland Payne CM. Skin diseases in Ghana and the UK. Int J Dermatol 2001;40:323-6. 22. Kyriakis KP, Palamaras I, Terzoudi S, Emmanuelides S, Michailides C, Pagana G. Epidemiologic aspects of rosacea. J Am Acad Dermatol 2005;53:918-9. 23. Khaled A, Hammami H, Zeglaoui F, Tounsi J, Zermani R, Kamoun MR, et al. Rosacea: 244 Tunisian cases. Tunis Med 2010;88:597-601. 24. Lal Khatri M. Spectrum of skin diseases in Yemen (Hajjah and adjacent region). Int J Dermatol 2004;43:580-5. 25. Gutierrez EL, Galarza C, Ramos W, Mendoza M, Smith ME, Ortega-Loayza AG. Influence of climatic factors on the medical attentions of dermatologic diseases in a hospital of Lima, Peru. An Bras Dermatol 2010;85:461-8. 26. Bae YI, Yun SJ, Lee JB, Kim SJ, Won YH, Lee SC. Clinical evaluation of 168 Korean patients with rosacea: the sun exposure correlates with the erythematotelangiectatic subtype. Ann Dermatol 2009;21:243-9. 27. Lazaridou E, Fotiadou C, Ziakas NG, Giannopoulou C, Apalla Z, Ioannides D. Clinical and laboratory study of ocular rosacea in northern Greece. J Eur Acad Dermatol Venereol 2011;25: 1428-31. 28. Michel JL, Cabibel F. Frequency, severity and treatment of ocular rosacea during cutaneous rosacea [in French]. Ann Dermatol Venereol 2003;130:20-4. 29. Ghanem VC, Mehra N, Wong S, Mannis MJ. The prevalence of ocular signs in acne rosacea: comparing patients from ophthalmology and dermatology clinics. Cornea 2003;22: 230-3. 30. Spoendlin J, Voegel JJ, Jick SS, Meier CR. A study on the epidemiology of rosacea in the UK. Br J Dermatol 2012;167: 598-605. 31. Kroshinsky D, Glick SA. Pediatric rosacea. Dermatol Ther 2006; 19:196-201. 32. Lacz NL, Schwartz RA. Rosacea in the pediatric population. Cutis 2004;74:99-103. 33. Drolet B, Paller AS. Childhood rosacea. Pediatr Dermatol 1992;9:22-6. 34. Del Rosso JQ. Advances in understanding and managing rosacea, part 1: connecting the dots between pathophysiological mechanisms and common clinical features of rosacea with emphasis on vascular changes and facial erythema. J Clin Aesthet Dermatol 2012;5:16-25. 35. Abram K, Silm H, Maaroos HI, Oona M. Risk factors associated with rosacea. J Eur Acad Dermatol Venereol 2010;24:565-71. 36. Jarmuda S, O’Reilly N, Zaba R, Jakubowicz O, Szkaradkiewicz A, Kavanagh K. Potential role of Demodex mites and bacteria in the induction of rosacea. J Med Microbiol 2012;61:1504-10.

J AM ACAD DERMATOL

Tan and Berg S35

VOLUME 69, NUMBER 6

37. Zhao YE, Wu LP, Peng Y, Cheng H. Retrospective analysis of the association between Demodex infestation and rosacea. Arch Dermatol 2010;146:896-902. 38. Lacey N, Delaney S, Kavanagh K, Powell FC. Mite-related bacterial antigens stimulate inflammatory cells in rosacea. Br J Dermatol 2007;157:474-81. 39. Whitfeld M, Gunasingam N, Leow LJ, Shirato K, Preda V. Staphylococcus epidermidis: a possible role in the pustules of rosacea. J Am Acad Dermatol 2011;64:49-52. 40. Dahl MV, Ross AJ, Schlievert PM. Temperature regulates bacterial protein production: possible role in rosacea. J Am Acad Dermatol 2004;50:266-72. 41. Rebora A, Drago F, Parodi A. May Helicobacter pylori be important for dermatologists? Dermatology 1995;191:6-8. 42. Rojo-Garcia JM, Munoz-Perez MA, Escudero J, Camacho F, Hergueta P, Herrerias JM. Helicobacter pylori in rosacea and chronic urticaria. Acta Derm Venereol 2000;80:156-7. 43. Erei A, Oztas M, Ilter N, Senol E, Sultan N, Grurer M. Helicobacter pylori seroprevalence in patients with acne rosacea. J Eur Acad Dermatol Venereol 1995;5(Suppl 1):S151. 44. Powell F, Daw MA, Duguid C. Positive Helicobacter pylori serology in rosacea patients. Ir J Med Sci 1992;161:75. 45. Boixeda de Miquel D, Vazquez Romero M, Vazquez Sequeiros E, Foruny Olcina JR, Boixeda de Miquel P, Lopez San Roman A, et al. Effect of Helicobacter pylori eradication therapy in rosacea patients. Rev Esp Enferm Dig 2006;98:501-9. 46. Schneider MA, Skinner R, Roserberg E. Serologic determination of Helicobacter pylori in rosacea patients and controls. Clin Res 1992;40:831. 47. Jones MP, Knable AL Jr, White MJ, Durning SJ. Helicobacter pylori in rosacea: lack of an association. Arch Dermatol 1998; 134:511. 48. Sharma VK, Lynn A, Kaminski M, Vasudeva R, Howden CW. A study of the prevalence of Helicobacter pylori infection and other markers of upper gastrointestinal tract disease

49. 50. 51.

52.

53.

54.

55. 56.

57.

58. 59.

in patients with rosacea. Am J Gastroenterol 1998;93: 220-2. Son SW, Kim IH, Oh CH, Kim JG. The response of rosacea to eradication of Helicobacter pylori. Br J Dermatol 1999;140:984-5. Herr H, You CH. Relationship between Helicobacter pylori and rosacea: it may be a myth. J Korean Med Sci 2000;15:551-4. Diaz C, O’Callaghan CJ, Khan A, Ilchyshyn A. Rosacea: a cutaneous marker of Helicobacter pylori infection? Results of a pilot study. Acta Derm Venereol 2003;83:282-6. Dakovic Z, Vesic S, Vukovic J, Milenkovic S, Jankovic-Terzic K, Dukic S, et al. Ocular rosacea and treatment of symptomatic Helicobacter pylori infection: a case series. Acta Dermatovenerol Alp Panonica Adriat 2007;16:83-6. Baz K, Cimen MY, Kokturk A, Aslan G, Ikizoglu G, Demirseren DD, et al. Plasma reactive oxygen species activity and antioxidant potential levels in rosacea patients: correlation with seropositivity to Helicobacter pylori. Int J Dermatol 2004;43:494-7. Yazici AC, Tamer L, Ikizoglu G, Kaya TI, Api H, Yildirim H, et al. GSTM1 and GSTT1 null genotypes as possible heritable factors of rosacea. Photodermatol Photoimmunol Photomed 2006;22:208-10. Palleschi GM, Torchia D. Rosacea in a monozygotic twin. Australas J Dermatol 2007;48:132-3. Breton AL, Truchetet F, Veran Y, Doumat-Batch F, Baumann C, Barbaud A, et al. Prevalence analysis of smoking in rosacea. J Eur Acad Dermatol Venereol 2011;25:1112-3. Gupta MA, Gupta AK, Chen SJ, Johnson AM. Comorbidity of rosacea and depression: an analysis of the National Ambulatory Medical Care Survey and National Hospital Ambulatory Care Surveyeoutpatient department data collected by the US National Center for Health Statistics from 1995 to 2002. Br J Dermatol 2005;153:1176-81. Tan SG, Cunliffe WJ. Rosacea and migraine. Br Med J 1976;1:21. Marks R. Concepts in the pathogenesis of rosacea. Br J Dermatol 1968;80:170-7.