Introduction

Rosacea is a chronic and inflammatory skin disease affecting the central face with common clinical presentations including flushing, erythema, telangiectasia, papules, pustules, rhinophyma, and ocular involvement. The prevalence of rosacea in the general population ranges from <1% to 22%.[1,2] In the absence of histological or serological markers, the diagnosis and classification of this disease are mainly established on observable characteristics that are derived from the dermatologists’ experiences. However, the clinical discrimination of non-typical cases from other similar facial diseases, such as acne vulgaris, seborrheic dermatitis, lupus vulgaris, and perioral dermatitis, might be challenging.[3]

Currently, two classifications systems of rosacea are available. The most updated one is based on patient-tailored analyses of the presented phenotypes and has been extensively used to assess and treat rosacea.[4] Nevertheless, guidelines for the management of rosacea produced by the British Association of Dermatologists that recommended the older rosacea classification system, containing erythematotelangiectatic (ETR), papulopustular (PPR), phymatous (PHR), and ocular characterized by clinical signs, should also be taken into account.[5] Moreover, in clinical practice guidelines and consensus of several countries, therapeutic regimens are still guided by the four main rosacea subtypes.[3,6,7] Therefore, an appropriate classification is essential to improve the patients’ prognoses. Unfortunately, the classification can be difficult sometimes for the naked eyes, especially for atypical and overlapped cases.

Dermoscopy is a useful non-invasive diagnostic tool for various melanocytic lesions and inflammatory diseases, which can increase diagnostic accuracy.[8–10] Generally, the diagnosis of inflammatory skin diseases depends mainly on clinical appearances. Moreover, in ambiguous cases, dermoscopy can lead to accurate diagnoses and avoid unnecessary biopsy by providing discriminative clues. Recently, the dermoscopic characteristics of ETR and PPR have been investigated, and successful attempts to apply dermoscopy beyond diagnosis have also been reported.[10–13] For example, dermoscopy has been considered an additional assessment tool to record therapeutic effects.[13] However, these reported cases focused only on one rosacea subtype, primarily ETR or PPR, with relatively small numbers of patients. Therefore, the dermoscopic features in these studies are incomplete, and little is known about PHR dermoscopic characteristics.

In the present study, we summarized the dermoscopic features and patterns of three rosacea subtypes (ETR, PPR, and PHR) in the Chinese Han population to improve the diagnostic accuracy, perform reasonable classification, and guide optimal therapeutic schedules. Furthermore, we compared the differences in dermoscopic features among subtypes, gender, ages, and durations.

Methods Study design

This retrospective morphological study was carried out at the China-Japan Friendship Hospital from August 1st, 2020 to October 31st, 2021. All subjects were diagnosed and classified by two experienced associate chief or chief physicians based on diagnostic criteria developed and published by the National Rosacea Society Expert Committee. If there was any disagreement between the two experts, the case was eliminated. The exclusion criteria were: overlapping cases (hard to fit into a certain category); and individuals who had received rosacea treatment within 3 months before enrollment. This study was approved by the Research Ethics Committee of China-Japan Friendship Hospital (No. 2020-130-K83) and followed the Declaration of Helsinki. Informed consent was obtained from all patients.

Imaging procedure and evaluation

Dermoscopic images were obtained with a digital dermoscopy system (Medicam 800HD, FotoFinder Systems GmbH, Birbach, Germany) at a 20-fold magnification. Both non-polarized and polarized contact modes were utilized for each case. Minimal pressure was applied to acquire better visualization, and ultrasound gel was used to preserve vessels’ morphology when the non-polarized contact mode was employed. Dermoscopic images were taken in areas where lesions were significant and then evaluated by two experienced dermoscopy experts. The two experts were asked to independently complete a pre-designed list with various dermoscopic rosacea features. We developed this list based on a review of the published literature[10–12,14–19] and unified terminology according to an expert consensus announced by the International Dermoscopy Society.[20] If necessary, new findings beyond the list could be included. During this process, any discrepancy between the two experts’ opinions was settled by a consensus meeting with other experts.

Dermoscopic features evaluated

The following dermoscopic features were evaluated in each subject: vessels including reticular linear vessels, unspecific linear vessels, unspecific dotted vessels, and linear vessels with branches; scales (mainly yellow scales); follicular findings such as follicular plugs, follicular red dots, and perifollicular white color; other structures containing red/pink/orange diffuse structureless areas, brown/orange/red focal structureless areas, white lines, follicular pustules, and white/black vellus hairs. The definitions of dermoscopic features were the same as previously described.[10–12,14–20]

Division of age, gender, and duration

General information including age, gender, and duration was also collected for further analyses. According to the latest age classification methods in China and the World Health Organization, we divided the patients into two groups: ≤40 years and >40 years. Moreover, individuals were classified into two groups based on the course of their diseases: ≤24 months and >24 months.

Statistical analyses

Statistical Product and Service Solutions version 21.0 (IBM Corp., Armonk, NY, USA) was used for statistical analyses. The continuous data are expressed as means (M) ± standard deviations (SD), and the categorical data as numbers (N) and percentages (%). Categorical variables were compared using the χ2 test. Fisher exact test and continuity correction in the χ2 test were also used when appropriate. A two-sided P value < 0.05 was considered statistically significant for the χ2 and Fisher exact tests. A Bonferroni correction adjusted P value < 0.0167 was considered statistically significant for multiple statistical tests within three different subtypes.

Results General information of the studied population

A total of 87 patients, 29 men and 58 women (mean age 40.0 ± 11.9 years, ranging from 21.0 to 65.0 years), contributed to our investigation. The courses of their diseases lasted from 3.0 to 336.0 months, with an average duration of 42.9 ± 61.1 months. The detailed information regarding the demographic and clinical characteristics of patients is presented in [Table 1].

Table 1 - General information of patients. General information ETR PPR PHR Total N 40 30 17 87 Gender  Male (n) 6 10 13 29  Female (n) 34 20 4 58 Age (years)  Mean 38.7 38.8 45.2 40.0  SD 10.4 13.0 12.5 11.9 Duration (months)  Mean 18.3 24.6 132.8 42.9  SD 20.9 9.00 90.3 61.1

ETR: Erythematotelangiectatic; PHR: Phymatous; PPR: Papulopustular; SD: Standard deviations.

Dermoscopic features of ETR

The most prominent ETR characteristic was reticular linear vessels (frequency of 85.0%), followed by red diffuse structureless areas (n = 24, 60.0%). Unspecific linear vessels were detected in 14 (35.0%) patients and dotted vessels in 10 (25.0%). Yellow scales were observed in 9 (22.5%) cases. Follicular findings included follicular plugs (n = 10, 25.0%), follicular red dots (n = 4, 10.0%), and perifollicular white color (n = 5, 12.5%). Other structures and their frequencies were as follows: pink diffuse structureless areas (n = 8, 20.0%), orange diffuse structureless areas (n = 8, 20.0%), brown focal structureless areas (n = 13, 32.5%), red focal structureless areas (n = 9, 22.5%), orange focal structureless areas (n = 3, 7.5%), white lines (n = 7, 17.5%), follicular pustules (n = 4, 10.0%), white vellus hairs (n = 20, 50.0%), and black vellus hairs (n = 5, 12.5%).

Dermoscopic features of PPR

Regarding PPR, typical reticular linear vessels were identified in 29 (97.7%) subjects, unspecific linear vessels in 3 (10.0%), and dotted vessels in 6 (20.0%). Yellow scales were detected in 19 (63.3%) cases. Follicular findings included follicular plugs (n = 27, 90.0%), follicular red dots (n = 10, 33.3%), and perifollicular white color (n = 10, 33.3%). For other structures, the most common feature was follicular pustules (n = 20, 66.7%). Other features included red diffuse structureless areas (n = 21, 70.0%), pink diffuse structureless areas (n = 5, 16.7%), orange diffuse structureless areas (n = 4, 13.3%), brown focal structureless areas (n = 13, 43.3%), red focal structureless areas (n = 15, 50.0%), orange focal structureless areas (n = 4, 13.3%), white lines (n = 13, 43.3%), white vellus hairs (n = 15, 50.0%), and black vellus hairs (n = 17, 56.7%).

Dermoscopic features of PHR

In the PHR subtype, vessels structures were not primary. In this subtype, the most salient vessel disturbances were linear vessels with branches (n = 10, 10/17), followed by unspecific linear vessels (n = 9, 9/17), reticular linear vessels, and unspecific dotted vessels (both n = 7, 7/17). Yellow scales were detected in 4 (4/17) cases. For follicular findings, both follicular plugs and perifollicular white color were found in 12 (12/17) cases, and follicular red dots in 3 (3/17). The other structures were as follows: red diffuse structureless areas (n = 5, 5/17), orange diffuse structureless areas (n = 12, 12/17), brown focal structureless areas (n = 3, 3/17), red focal structureless areas (n = 1, 1/17), orange focal structureless areas (n = 8, 8/17), white lines (n = 13, 13/17), follicular pustules (n = 3, 3/17), white vellus hairs (n = 8, 8/17), and black vellus hairs (n = 4, 4/17). Typical dermoscopic images are shown in [Figures 1–3].

Figure 1:

Common vessels of rosacea in dermoscopy. (A) Reticular linear vessels characteristically arranged in polygonal networks; (B) unspecific linear vessels; (C) unspecific dotted vessels in a patchy distribution; and (D) linear vessels with branches.

Figure 2:

Typical scales and follicular findings of rosacea in dermoscopy. (A) Yellow scales; (B) follicular plugs; (C) follicular red dots; and (D) perifollicular white color.

Figure 3:

Other structures of rosacea in dermoscopy. (A) Red focal structureless areas; (B) orange focal structureless areas; (C) white lines; and (D) follicular pustules.

Comparisons of dermoscopic features among the three rosacea subtypes

The differences in reticular linear vessels (P < 0.001), unspecific linear vessels (P = 0.005), linear vessels with branches (P < 0.001), yellow scales (P = 0.001), follicular plugs (P < 0.001), perifollicular white color (P < 0.001), red diffuse structureless areas (P = 0.022), orange diffuse structureless areas (P < 0.001), red focal structureless areas (P = 0.002), orange focal structureless areas (P = 0.003), white lines (P < 0.001), follicular pustules (P < 0.001), and black vellus hairs (P < 0.001) were statistically significant among the three subtypes. The multiple comparisons revealed that reticular linear vessels were more common in ETR and PPR, unspecific linear vessels were more common in ETR and PHR, while linear vessels with branches were distinctive for PHR. Yellow scales were more frequent in PPR. On the other hand, follicular plugs in PPR and PHR did not differ (P = 0.118), but the percentages in these two types were higher than in ETR (P < 0.001 and P = 0.002, respectively). Orange diffuse structureless areas presented the highest percentage in PHR, and the differences between PHR and ETR, and PHR and PPR were statistically significant (both P < 0.001). Follicular pustules were more frequent in PPR, the differences between ETR and PPR (P < 0.001), and PHR and PPR (P = 0.002) were statistically significant [Table 2].

Table 2 - Summarization and comparison of dermoscopic features in different rosacea subtypes, genders, ages, and durations. Subtypes Gender Age (years) Duration (months) Dermoscopic features ETR (N = 40) PPR (N = 30) PHR (N = 17) P value Male (N = 29) Female (N = 58) P value 95% CI ≤40 (N = 48) >40 (N = 39) P value 95% CI ≤24 (N = 49) >24 (N = 38) P-value 95% CI Vessels  Reticular linear vessels 34 (85.0)∗ 29 (97.7)† 7 (7/17) <0.001 19 (65.5) 51 (87.9) 0.021 0.087–0.784 38 (79.2) 32 (82.1) 0.791 0.284–2.434 43 (87.8) 27 (71.1) 0.061 0.967–8.816  Unspecific linear vessels 14 (35.0)‡ 3 (10.0)† 9 (9/17) 0.005 13 (44.8) 13 (22.4) 0.046 1.080–7.326 16 (33.3) 10 (25.6) 0.487 0.568–3.698 12 (24.5) 14 (36.8) 0.244 0.220–1.404  Unspecific dotted vessels 10 (25.0) 6 (20.0) 7 (7/17) 0.306 11 (37.9) 12 (20.7) 0.121 0.877–6.260 11 (22.9) 12 (30.8) 0.468 0.257–1.741 12 (24.5) 11 (29.0) 0.807 0.306–2.072  Linear vessels with branches 0∗ 0‡ 10 (10/17) <0.001 8 (27.6) 2 (3.5) 0.002 2.093–54.367 3 (6.3) 7 (17.9) 0.173 0.073–1.269 0 10 (26.3) <0.001 1.122–1.641 Scales  Yellow scales 9 (22.5)‡ 19 (63.3)† 4 (4/17) 0.001 10 (34.5) 22 (37.9) 0.817 0.339–2.186 18 (37.5) 14 (35.9) >0.999 0.446–2.576 16 (32.7) 16 (42.1) 0.380 0.277–1.604 Follicular findings  Follicular plugs 10 (25.0)∗‡ 27 (90.0) 12 (12/17) <0.001 17 (58.6) 32 (55.2) 0.821 0.467–2.837 27 (56.3) 22 (56.4) >0.999 0.424–0.329 22 (44.9) 27 (71.1) 0.018 0.135–0.816  Follicular red dots 4 (10.0) 10 (33.3) 3 (3/17) 0.056 4 (13.8) 13 (22.4) 0.403 0.163–1.881 14 (29.2) 3 (7.7) 0.015 1.304–18.723 10 (20.4) 7 (18.4) >0.999 0.388–3.327  Perifollicular white color 5 (12.5)∗ 10 (33.3) 12 (12/17) <0.001 12 (41.4) 15 (25.9) 0.219 0.787–5.202 13 (27.1) 14 (35.9) 0.485 0.266–1.652 12 (24.5) 15 (39.5) 0.164 0.198–1.248 Other structures  Red diffuse structureless areas 24 (60.0) 21 (70.0)‡ 5 (5/17) 0.022 13 (44.8) 37 (63.8) 0.111 0.186–1.142 25 (52.1) 25 (64.1) 0.284 0.256–1.446 31 (63.3) 19 (50.0) 0.275 0.728–4.075  Pink diffuse structureless areas 8 (20.0) 5 (16.7) 0 0.13 3 (10.3) 10 (17.2) 0.53 0.140–2.192 9 (18.8) 4 (10.3) 0.368 0.571–7.141 10 (20.4) 3 (7.9) 0.135 0.761–11.755  Orange diffuse structureless areas 8 (20.0)∗ 4 (13.3)‡ 12 (12/17) <0.001 13 (44.8) 11 (19.0) 0.021 1.299–9.279 14 (29.2) 10 (25.6) 0.811 0.461–3.090 8 (16.3) 16 (42.1) 0.009 0.099–0.725  Brown focal structureless areas 13 (32.5) 13 (43.3) 3 (3/17) 0.207 8 (27.9) 21 (36.2) 0.477 0.253–1.779 14 (29.2) 15 (38.5) 0.493 0.269–1.615 15 (30.6) 14 (36.8) 0.648 0.309–1.854  Red focal structureless areas 9 (22.5) 15 (50.0)† 1 (1/17) 0.002 11 (37.9) 14 (24.1) 0.213 0.734–5.023 11 (22.9) 14 (35.9) 0.235 0.208–1.357 13 (26.5) 12 (31.6) 0.639 0.308–1.989  Orange focal structureless areas 3 (7.5)∗ 4 (13.3) 8 (8/17) 0.003 9 (31.0) 6 (10.3) 0.032 1.229–12.373 5 (10.4) 10 (25.6) 0.087 0.104–1.089 5 (10.2) 10 (26.3) 0.084 0.098–1.029  White lines 7 (17.5)∗ 13 (43.3) 13 (13/17) <0.001 16 (55.2) 17 (29.3) 0.034 1.177–7.484 17 (35.4) 16 (41.0) 0.66 0.330–1.882 15 (30.6) 18 (47.4) 0.124 0.203–1.182  Follicular pustules 4 (10.0)‡ 20 (66.7)‡ 3 (3/17) <0.001 13 (44.8) 14 (24.1) 0.084 0.990–6.585 10 (20.8) 17 (43.6) 0.035 0.133–0.873 11 (22.4) 16 (42.1) 0.063 0.157–1.009  White vellus hairs 20 (50.0) 15 (50.0) 8 (8/17) >0.999 11 (37.9) 32 (55.2) 0.173 0.200–1.235 21 (43.8) 22 (56.4) 0.284 0.256–1.409 26 (53.1) 17 (44.7) 0.519 0.596–3.269  Black vellus hairs 5 (12.5)‡ 17 (56.7) 4 (4/17) <0.001 11 (37.9) 15 (25.9) 0.321 0.676–4.543 13 (27.1) 13 (33.3) 0.639 0.296–1.886 10 (20.4) 16 (42.1) 0.035 0.137–0.909

∗P < 0.0167, comparison between ETR and PHR.

†P < 0.0167, comparison between PPR and PHR.

‡P < 0.0167, comparison between ETR and PPR. CI: Confidence interval; ETR: Erythematotelangiectatic; PHR: Phymatous; PPR: Papulopustular.

Comparisons of dermoscopic features between different gender, age, and duration groups

Between females and males, we detected statistical differences for reticular linear vessels (P = 0.021), unspecific linear vessels (P = 0.046), linear vessels with branches (P = 0.002), orange diffuse structureless areas (P = 0.021), orange focal structureless areas (P = 0.032), and white lines (P = 0.034). Only reticular linear vessels were common in females while other features were more found in males. Follicular red dots (P = 0.015) were more frequent in patients with ≤40 years, and follicular pustules (P = 0.035) were more frequent in ones with >40 years. Linear vessels with branches (P < 0.001), follicular plugs (P = 0.018), orange diffuse structureless areas (P = 0.009), and black vellus hairs (P = 0.035) were more common in patients with a disease course of >2 years [Table 2].

Summarization of the dermoscopic patterns of the three rosacea subtypes

The representative dermoscopic pattern of ETR was red diffuse structureless areas and reticular linear vessels. For PPR, the typical dermoscopic pattern was the combination of red diffuse structureless areas, reticular linear vessels, yellow scales, follicular plugs, and follicular pustules. Regarding PHR, the prominent dermoscopic pattern included orange diffuse structureless areas, linear vessels with branches, perifollicular white color, orange focal structureless areas, and white lines [Table 3].

Table 3 - Dermoscopic patterns of the three rosacea subtypes and typical dermoscopic features of related differential diagnosis. Diseases Typical dermoscopic features Three subtypes of rosacea  ETR Red diffuse structureless areas, reticular linear vessels  PPR Red diffuse structureless areas, reticular linear vessels, yellow scales, follicular plugs, and follicular pustules  PHR Orange diffuse structureless areas, linear vessels with branches, perifollicular white color, orange focal structureless areas, and white lines Other differential skin diseases  Seborrheic dermatitis Dotted vessels in a patchy distribution and fine yellowish/whitish scales[10]  Discoid lupus erythematosus Early lesions: follicular plugging (yellow clods) perifollicular whitish halo, and white scalesMature lesions: blurred, telangiectatic, arborizing vessels; white structureless areas; and hyperpigmentation[29]  Acne vulgaris Comedones: Erythematous periphery; dilated, roundish, central pore filled with a brown-yellow plug[30]Inflammatory lesions: erythematous roundish areas with central white-yellowish structure[31]  Lupus vulgaris Yellow to golden-colored background, fine focused telangiectasias, milia-like cysts, and whitish reticular streaks[32] Malar lesion of SLE Reddish/salmon-colored follicular dots surrounded by white halos, branched vessels, white scaling, dotted and network-like vessels[18]  Facial psoriasis Evenly distributed red dots or globules over a pale red erythematous background along with white scales[33]  Contact dermatitis Allergic: intense erythema, vesicles or pustules, orange-yellowish patchy areas and crusts, dense dotted and linear vessels;Irritant: moderate erythema, sporadic vesicles, sparse vesse