Introduction

Rosacea is a chronic inflammatory skin disease which typically affects the visible skin of the central face. The etiology and pathophysiology of the disease are poorly understood. Although demodex mites are common in the skin of healthy individuals, an increased amount is considered to play a pathogenic role in the development of rosacea.1

Furthermore, increased activation of keratinocytes (releasing cathelicidin and matrix metalloproteinases), macrophages (IFN-g, TNF, matrix metalloproteinases, and IL-26), and T helper type 1 (TH1) and TH17 cells indicate an immunological genesis.2 Rosacea symptoms consist of transient or persistent erythema of the face with the appearance of telangiectatic vessels, papules, and pustules, up to hypertrophy of the sebaceous glands of the nose (rhinophyma).3 Rosacea has been classified into four subtypes: erythematotelangiectatic, papulopustular, phymatous, and ocular.3, 4 In up to 50% of patients, ocular changes are involved and range from common symptoms such as mild dryness and irritation with blepharitis and conjunctivitis to rare ones like sight-threatening keratitis.2 Patients may also report an increased skin sensitivity like a burning sensation. Overall, genetic and environmental components (e.g. stress, spicy foods, and alcohol) can trigger the initiation and aggravation of rosacea.

Since rosacea affects the highly visible area of the facial skin, it has a high impact on the quality of life of those suffering, including discomfort, stigmatization, loss of self-esteem, depression, and limitations in social activities.5-7 Many patients do not receive treatment (topical, systemic, lifestyle changes) until later, which in particular in elderly men leads to the necessity of surgical correction when a rhinophyma has developed.8

Rosacea is a common condition, but the exact population-based prevalence and incidence in most countries remain largely unknown. Only a few population-based studies have been published on the prevalence of rosacea and also on its occurrence in connection with other skin diseases. An estimation of the global prevalence of rosacea based on published data postulated that 5.5% of the adult population is affected.9 However, the prevalence of rosacea reported in the literature depends on the diagnostic method, with higher estimates in questionnaire studies of rosacea symptoms and lower estimates in health registries.9 A recent population-based, cross-sectional study from China involving 10,095 participants observed rosacea in 3.5% (95% confidence interval [CI]: 3.13–3.85%) of the study population.10

The aim of the present study was to determine the point prevalence and dermatological comorbidity in a large cohort of employed persons reflecting a substantial proportion of the general working population in Germany.

Patients and methods Patients

Data of n = 161,269 employees between 16 and 70 years of age in Germany were used for this analysis. These employees were examined by dermatologists in voluntary occupational skin screenings conducted between 2001 and 2016 in more than 500 companies in Germany. The companies included a wide variety of sectors such as automobile, pharmaceuticals, banks, insurance companies, food industry, civil service, and tertiary sector. All employees working in these companies had the possibility to participate in the screening. The dermatologists documented all dermatological findings using a standardized report form. In case a need for further treatment was detected, the participants were referred to a dermatologist. Next to dermatological findings, the skin type following the Fitzpatrick classification11 as well as age and gender were documented.

The diagnosis of rosacea as well as of other skin conditions was based on the clinical findings observed by trained dermatologists. The typical criteria for diagnosis are centrofacial location of major features like erythema, inflammatory papules and pustules, telangiectasia, and/or ocular manifestations. These criteria are in accordance with the global ROSacea COnsensus (ROSCO), published later in the British Journal of Dermatology in 2017.12 Bilateral facial location was considered typical but not essential for diagnosis. In our study, all forms of rosacea were included since the main objective was to assess the overall prevalence of rosacea in a population-based study.

All procedures performed were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. The study was approved by the ethics committee of the University Medical Center Hamburg-Eppendorf (UKE).

Data and statistical analyses

The data recorded in electronic forms were migrated to the statistical program SPSS. Missing values were counted as not prevalent. Point prevalence rates were calculated for each disease. These prevalence rates were further stratified by skin type, gender, and age. Age was documented as a continuous variable and further categorized into decades. Chi-squared tests and odds ratios (OR) were calculated for each dermatological finding to occur among participants having rosacea compared to those without rosacea, resulting in crude ORs. To control for the effect of age, gender, and skin type, a logistic regression analysis was performed providing adjusted ORs for each finding. To control for multiple testing, a Bonferroni correction of P-values was conducted in this analysis.

In order to examine the association of rosacea prevalence with different professions, we analyzed a sub-data set providing information of the following occupational groups: office work, chemical or laboratory environment, metal production, warehouse workers, or other sections. This data set included data from 18,725 individuals collected between 2012 and 2014 in the same way as described above.

All statistical analyses were performed using SPSS (IBM, Armonk, NY, U.S.) for Windows version 23.

Results Rosacea prevalence

The mean age of the 161,269 participants was 43.2 ± 10.9 years, and 44.5% were female. The overall rosacea prevalence in the total sample was 2.1% and was the same for men and women (Table 1). The mean age of participants with rosacea was 50.7 ± 9.3 years and significantly higher (P < 0.001) than the mean age of participants without rosacea (43.1 ± 10.9). Stratified by age, the rosacea prevalence was significantly higher among women until the age of 49. Among the elderly (50–59 and 60–70 years), rosacea prevalence was higher for men (Fig. 1). Furthermore, the rosacea prevalence was largely age-dependent and decreased with darker skin type (Table 1).

Table 1. Rosacea prevalence in the examined groups Group n n with rosacea Rosacea point prevalence Total 161,269 3,386 2.1% Male 89,523 1,902 2.1% Female 71,746 1,484 2.1% Age groups 16–29 years 19,661 53 0.3% 30–39 years 41,332 373 0.9% 40–49 years 52,700 1,062 2.0% 50–59 years 37,419 1,324 3.5% 60–70 years 10,157 574 5.7% Skin types Skin type I 14,207 452 3.2% Skin type II 105,011 2,337 2.2% Skin type III 38,836 570 1.5% Skin type IV 2,226 10 0.4%

Point prevalence of rosacea by age and gender (total: n = 161,269; male: n = 89,523; female: n = 71,746). *P < 0.05 for the difference between male and female

Dermatological comorbidity

In the crude comparisons, participants with rosacea were more likely to have a number of further dermatological conditions. Especially vascular lesions, such as hemangioma, spider veins, and telangiectasia, were significantly more common among participants with rosacea (Fig. 2, Table 2). Besides the vascular lesions, fungal diseases and inflammatory skin diseases were the most frequent cutaneous comorbidities of rosacea (Figs. 3 and 4, Table 2).

Odds ratios and 95% confidence intervals for the occurrence of vascular skin diseases among participants with rosacea compared to those without, stratified by gender (men: n = 89,523; women: n = 71,746)

Odds ratios and 95% confidence intervals for the occurrence of inflammatory skin diseases among participants with rosacea compared to those without, stratified by gender (men: n = 89,523; women: n = 71,746)

Table 2. Odds ratios (ORs) for the occurrence of further skin conditions in people with rosacea, crude ORs, and adjusted for age, gender, and skin type (n = 161,269) Point prevalence among participants Crude OR Adjusted OR With rosacea Without rosacea OR 95% CI P-value (χ2) OR 95% CI P-value (regression) Vascular skin diseases Telangiectasia 22.8 7.4 3.7 3.4–4.0 <0.001 2.5 2.3–2.7 <0.001a Spider veins 29.4 18.9 1.8 1.7–1.9 <0.001 1.1 1.1–1.2 0.001a Hemangioma 56.1 41.5 1.8 1.7–1.9 <0.001 1.1 1.0–1.2 0.002a Nevus flammeus 6.3 4.9 1.3 1.1–1.5 <0.001 0.8 0.7–0.9 0.002a Inflammatory skin diseases Seborrheic dermatitis 6.3 3.1 2.1 1.8–2.4 <0.001 1.6 1.4–1.9 <0.001a Acne 2.8 3.3 0.8 0.7–1.0 0.088 1.6 1.3–1.9 <0.001a Psoriasis 3.3 2.0 1.7 1.4–2.0 <0.001 1.4 1.2–1.7 <0.001a Hand eczema 1.2 0.8 1.5 1.1–2.1 0.009 1.3 0.9–1.8 0.120 Exsiccation dermatosis 1.7 1.0 1.8 1.3–2.3 <0.001 1.2 0.9–1.6 0.118 Intertriginous dermatitis 1.0 0.6 1.6 1.1–2.2 0.009 1.1 0.8–1.6 0.545 Atopic dermatitis 1.4 1.4 1.0 0.7–1.3 0.842 1.0 0.7–1.3 0.975 Contact dermatitis 0.3 0.2 2.2 0.6–2.4 0.532 1.0 0.5–1.9 0.981 Pathogen-induced skin diseases (fungal, viral, and bacterial) Tinea pedis 8.8 4.5 2.1 1.8–2.3 <0.001 1.4 1.3–1.6 <0.001a Onychomycosis 11.6 6.1 2.0 1.8–2.3 <0.001 1.2 1.0–1.3 0.008 Pityriasis versicolor 1.0 1.0 1.0 0.7–1.4 0.819 1.0 0.7–1.4 0.989 Tinea corporis 0.4 0.4 1.0 0.6–1.6 0.884 0.7 0.4–1.2 0.245 Verruca vulgaris hands 0.9 0.5 1.7 1.2–2.5 0.003 1.5 1.1–2.2 0.022 Verruca vulgaris feet 2.4 2.3 1.0 0.8–1.3 0.762 1.0 0.8–1.3 0.930 Folliculitis 11.6 7.8 1.6 1.4–1.7 <0.001 1.8 1.6–2.0 <0.001a Pyoderma 0.7 0.5 1.3 0.9–2.0 0.212 0.8 0.5–1.2 0.269 χ2, chi-squared test. a Remains significant at nominal α ≤ 0.05 level after Bonferroni correction for 23 significance tests.

Odds ratios and 95% confidence intervals for the occurrence of pathogen-induced skin diseases among participants with rosacea compared to those without, stratified by gender (men: n = 89,523; women: n = 71,746)

After adjusting for age, gender, skin type, and Bonferroni correction, these positive associations remained significant for telangiectasia (OR = 2.5), folliculitis (OR = 1.8), seborrheic dermatitis (OR = 1.6), acne (OR = 1.6), tinea pedis (OR = 1.4), psoriasis (OR = 1.4), spider veins (OR = 1.1), and hemangioma (OR = 1.1). Nevus flammeus was significantly less common among participants with rosacea (OR = 0.8). Higher age and lighter skin phototype were significantly positively associated with rosacea prevalence (both P < 0.001).

Rosacea prevalence among different professions

The mean age of the 18,725 participants in the sub-data set including information on occupational groups was 42.4 ± 10.5 years; 42.6% were female (n = 7.986), and 57.4% were male (n = 10.739). Of these, 80% had an office job, 5.5% worked in a chemical or laboratory environment, 3.7% in metal production, 1.3% in warehouses, and 9.6% in other sections. Rosacea prevalence in this sample was 1.8%. Employees working in chemical or laboratory environments were significantly less likely to have rosacea (prevalence 0.9%, OR = 0.5, 95% CI: 0.2–0.9). For the other professions, no significant differences were found. Controlling for age and gender in logistic regression analyses led to no significant association of rosacea prevalence with any of the professions.

Discussion

The aim of the present study was to analyze the epidemiology and dermatological comorbidity of rosacea in the adult German working population. To do so, a large German working population of 161,269 participants was examined, revealing a mean rosacea prevalence of 2.1%, similar to the previous study of 90,880 German workers (2.3%) published in 2011.13 Yet this finding is less than half of the percentage estimated for the general population in the meta-analysis of 32 published studies with a total of 26,519,836 individuals older than 18 years (pooled proportion for rosacea 5.46%; 95% CI: 4.91–6.04).9 The meta-analysis was based on 22 populations from Europe, three from Africa, four from Asia, nine from North America, and three from South America. Among the studies of dermatological outpatients, the pooled proportion was 2.39% (95% CI: 1.56–3.39). It was stated that the numbers were highly dependent on the diagnostic method, with higher estimates in questionnaire studies of rosacea symptoms and lower estimates in health registries.9 Moreover, mostly annual prevalences were measured, whereas we obtained point prevalences. More importantly, the data of this study was derived from dermatological exams but not from claims data or patient responses. Hence, these study results are consistent and explicable to these findings. A population-based, cross-sectional Chinese study of 10,095 participants (mean age 35.5 ± 19.1 years; 50.5% female), recently published in March 2020, observed a rosacea prevalence of 3.48% (95% CI: 3.13–3.85%).10 The study’s design was comparable to ours, with skin examinations conducted by board-certified dermatologists and a more similar outcome than of the meta-analysis described above. The fact that the participants of our study were all part of the working population could have led to a lower total prevalence than found in the Chinese population-based study (3.5% vs. 2.1%). This phenomenon is described later as the ‘healthy worker effect’.

Our study showed that rosacea was most prevalent among the elderly, similar to the findings of other studies with a predominance of those aged 45–60 years.13 In our analysis, the prevalence in the age group of 60–70 years was almost 20 times as high as in the youngest age group of 16–29 years of age (5.7% vs. 0.3%). The overall higher prevalence of rosacea with higher age led to an association with also age-related cutaneous diseases in the crude comparisons. This effect has been canceled out in the multivariate regression analysis.

Interestingly, although the common consensus still describes the women of older age being most likely affected by rosacea symptoms, no difference between genders regarding the total point prevalence was found in our data (male 2.1%, female 2.1%).13 These results are in accordance with the findings in the meta-analysis of Gether et al.9 Stratified by age, however, the rosacea prevalence was significantly higher among women until the age of 49. Among the elderly (50–59 and 60–70 years), rosacea prevalence was higher for men (Fig. 1). Hence, we did find an age-dependent difference of prevalence between male and female participants after all.

As mentioned in the introduction, genetic and environmental components (e.g. stress, spicy food, and alcohol) can trigger the initiation and aggravation of rosacea by dysregulation of the innate and adaptive immune system.2 Various sociodemographic variables such as UV exposure, smoking, alcohol, family history, and a fair skin type have been positively associated with rosacea.2 In our results, positive associations were found for higher age and lighter skin type (both P < 0.001). Yet no information about other confounders such as smoking, alcohol, or family history was provided and therefore could not be analyzed in our data set.

The sub-data set analysis of occupational groups revealed that employees working in chemical or laboratory environments were significantly less likely to have rosacea.

Since this analysis is based on cross-sectional data, no statement about causality can be made. However, one explanation could be that people with rosacea deliberately do not choose a profession in this field due to the fear of potentially aggravating existing skin symptoms. However, the regression results indicate that age and gender may explain the results. Assuming that men are more likely to choose a career in the chemical or laboratory environment and they are less likely to suffer from rosacea, at least until the age of 49, this could explain why rosacea is less common in this occupational group.

Unfortunately, no data about further risk factors such as indoor or outdoor services were collected for a more detailed analysis.

The consideration of the comorbidities in our study revealed that rosacea patients were more likely to suffer from vascular lesions of the skin such as telangiectasia (OR = 2.5), spider veins (OR = 1.1), and hemangioma (OR = 1.1). Besides, a significant association was found for seborrheic dermatitis (OR = 1.6) and acne (OR = 1.6), which are closely related due to the involvement of sebaceous glands (e.g. rhinophyma) as well as follicular inflammation (e.g. papules and pustules). The clinical distinction between the three can be tricky sometimes even for trained dermatologists without further diagnostics. Therefore, there might be a slightly overestimated association due to the difficulty of making a purely clinical diagnosis. Nonetheless, psoriasis and tinea pedis were also significantly associated with the presence of rosacea (OR = 1.4). Thus, this finding underlines the necessity to examine rosacea patients for the presence of further skin diseases.

The clinical diagnostic process alone might be a limitation of the present study. However, in total, the probability of errors is kept low due to the selection of highly experienced dermatologists.

A further restriction might be a possible selection bias due to the investigation of the working population only. This could lead to an underestimation of morbidity, known as the 'healthy worker effect'. Also, the lack of data on medication and internal comorbidity is a limitation. Nevertheless, there is high consistency and validity of the data which permits conclusions for practice.

In conclusion, patients diagnosed with rosacea need to be examined and monitored closely for potential comorbidities. Further studies are needed to evaluate the epidemiology for rosacea and associated comorbidities in the general population.

Acknowledgments

The authors thank the Scientific Communication Team of the IVDP, in particular Sara Tiedemann and Mario Gehoff, for copy editing. Open Access funding enabled and organized by Projekt DEAL.