Eyelid dermatitis in patch-tested adult patients: a systematic review with a meta-analysis

Abstract

Eyelid dermatitis (ED) affects a cosmetically significant area and leads to patients’ distress. Despite ongoing and recent research efforts, ED remains a multidisciplinary problem that needs further characterization. We aimed to evaluate the atopic eyelid dermatitis (AED) frequency in ED patients and to perform their clinical profiling. PubMed databases were searched from 01.01.1980 till 01.02.2024 to PRISMA guidelines using a search strategy: (eyelid OR periorbital OR periocular) AND (dermatitis or eczema). Studies with patch-tested ED patients were included. Proportional meta-analysis was performed using JBI SUMARI software. We included 65 studies across Europe, North America, Asia and Australia, with a total of 21,793 patch-tested ED patients. AED was reported in 27.5% (95% CI 0.177, 0.384) of patch-tested ED patients. Isolated ED was noted in 51.6% (95% CI 0.408, 0.623) of 8453 ED patients with reported lesion distribution, including 430 patients with isolated AED. Our meta-analysis demonstrated that the AED frequency in patch-tested ED patients exceeded the previous estimate of 10%. Isolated AED was noted in adult patients, attending contact allergy clinics. Future studies are needed to elucidate the global prevalence and natural history of isolated AED in adults.

Introduction

Eyelid dermatitis (ED) (Fig. 1) affects the periorbital skin¹. Although ED has been a focus of large multi-center studies^2–7, many aspects of ED, and particularly atopic eyelid dermatitis (AED), are only partially understood, highlighting important clinical needs and gaps of knowledge (Table 1).

Figure 1.

Clinical presentations in patients with atopic eyelid dermatitis. Legend: The representative images of three patients with adult-onset atopic eyelid dermatitis (AED) (Patients 1–3), who are under care at the Department of Dermatology and Venereology (Head—Prof. Olga Olisova), I.M. Sechenov First Moscow State Medical University: patient 1 (A,D,H), aged 32 (an AED onset at 30 years), patient 2 (B,E,I), aged 37 (an AED onset at 37 years), patient 3, (C,F,J) aged 31 (an AED onset at 21 years). Facial skin imaging was carried out using the AI face recognition technology, 8-spectrum imaging technology and deep quantitative analysis by Capricorn Intelligent Imager ZMLH01 (Beijing Sincoheren S&T Development, Beijing, China). The high-definition images were acquired in a negatively polarized UV light (365 nm, A-F), white light (RGB, range 420–620 nm, H-J). AED patients presented with symmetrical, scaly erythematous periorbital areas with a hyperpigmentation and a lichenification, particularly pronounced at the medial aspects of the upper and lower eyelids. Informed consents from the AED patients for a publication of these images were obtained.

Table 1.

Unresolved questions related to atopic eyelid dermatitis.

1	Is there a genetic predisposition to the AED?
2	What are the pathophysiological mechanisms underlying AED?
3	What are the AED biomarkers in peripheral blood, affected skin and tear fluid?
4	What are the transcriptomic biomarkers for AED endotyping?
5	What is the eyelid microbiome in AED?
6	What is microbial diversity of the eyelid surface in AED?
7	What are the clinical and demographic characteristics of AED in adult AD patients?
8	What is the age at disease onset and what is the frequency of AED across age groups?
9	What are the AED risk factors?
10	What are the allergen sensitisation profiles in AED patients?
11	Is there an increased risk of developing a comorbid ACD in AED patients?
12	Which allergens (contact allergens and/or aeroallergens) mostly contribute to AED development?
13	What is the role of individual allergens in AED? Particularly HDM?
14	What are the differential diagnostic pointers in ED patients? Can eyelid distribution of skin lesions predict the diagnosis in ED patients?
15	Are there any predictive biomarkers for AED development?
16	What are microbiome associations in AED with or without ocular comorbidities?
17	What is the optimal clinical grading of AED severity? What indices can be used to assess the AED severity?
18.	What are the optimal diagnostic approaches to the patients with AED? Particularly, in the context of personalized management of AED patients?
19	Which skin or tear biomarkers could indicate the AED activity?
20	How severe is the QoL impairment in AED patients?
21	What are treatment strategies, limitations in therapy, recommendations for eyelid skin care?
22	How the treatment strategy in AED patients may vary depending on genetic or transcriptomic signature?
23	What are predictive biomarkers for treatment efficacy in AED?
24	Which biomarkers can be used for monitoring the AED activity during in the follow-up of AED patients?
25	Which factors contribute to the flare-ups in AED patients?
26	Can we prevent the development of AED flare-ups?
27	How effective the ASIT in AED patients?
28	Can severe isolated AED be an indication for biological therapy?
29	Can steroid dependency in AED patients be avoided? Which patient education initiatives are needed?
30	What is the optimal treatment strategy for resistant AED?

Importantly, eyelids are a visible and cosmetically significant area⁸, which is crucial for social interactions. Conditions affecting the face, especially the eye area, may contribute to a patient’s distress and social disadaptation⁸. For example, the impact on patients’ psychological well-being in AED is thought to be considerable⁹, and patient’s quality of life improves with skincare¹⁰ and topical treatment. However, the ED burden needs to be better characterized as the existing dermatological QoL instruments may only partially apply to ED. Currently, there is a need for evidence-based guidelines for the ED treatment because skin symptoms, their duration, resistance to topical treatment and severe disease burden may require special considerations in treatment decision-making in ED patients.

The eyelids are characterized by thin skin¹¹ with fewer stratum corneum cell layers¹², low surface lipid content¹³, and high drug permeability¹⁴. A close proximity of eyelids to ocular surface immunity^15,16 creates a unique eye microenvironment that can be disturbed in skin conditions, affecting eyelids. For example, AD patients with comorbid ocular surface disease were shown to have decreased goblet cell density^17,18, increased mucin (MUC5AC) production¹⁹, and elevated concentrations of proinflammatory mediators (histamine, LTB₄)²⁰ and cytokines (TNF-α, IL-4 and IL-5)²¹ in the tear fluid, thus opening the possibility of the biomarker development in the tear fluid in ED patients. These anatomical features, an immunological microenvironment, together with an exposure to allergens/irritants and chronic rubbing, may contribute to ED in predisposed individuals.

In clinical practice, ED may present a difficult challenge in a differential diagnosis with ED of various aetiology.²² In many cases, the ED etiology is difficult to establish²³. Given a persistent course, frequent comorbidities and a broad differential diagnosis, ED is a multidisciplinary clinical problem, which is of interest for dermatologists, allergists, and ophthalmologists, justifying the set-up of multidisciplinary clinics for ED patients as implemented in some specialized centers.

In AED, ocular/periorbital changes (e.g. recurrent conjunctivitis, Dennie-Morgan infraorbital fold, keratoconus, anterior subcapsular cataracts and orbital darkening) were included in the 1980 Hanifin and Rajka diagnostic criteria (H&R)²⁴ but not in the U.K. Working Party’s diagnostic criteria (UKWP) for AD²⁵, which may lead to an underdiagnosis of isolated AED. AED may last up to 10 years or longer²⁶. Patients may repeatedly use topical corticosteroids (CS) around the eyes, with a limited effect²⁷. AED may co-occur with common atopic diseases³ or ocular comorbidities, including allergic conjunctivitis, atopic keratoconjunctivitis, vernal keratoconjunctivitis, and keratoconus²⁸. Known as a localized AD clinical phenotype, AED can be the only clinical manifestation in AD patients. The global prevalence of AED is unknown, although regional variations are likely to occur.

We hypothesized that AED may have phenotypic differences from ED in general. Here, we aimed (1) to carry out clinical profiling of ED patients, in general; (2) to systematically evaluate the AED frequency in patch-tested ED patients and (3) to carry out clinical profiling of adult AED patients. For this analysis, we focused on a patch-tested ED population due to two important considerations. Firstly, AED patients are often referred to contact allergy clinics and may not be under care as AD patients in specialized departments²⁹. Secondly, using ‘flexural eczema’ as an inclusion criterion in epidemiological studies of AD may miss patients with isolated AED. This work will raise the awareness of medical professionals of an ED burden, its broad differential diagnosis, and may prompt further clinical and mechanistic studies in ED, particularly in AED.

Results

Study design

In total, 65 studies, published between 1986 and 2024, were included (Table S2)^{2–7,23,27,29–85}. In selected studies, there were 21,793 patch-tested ED patients. There were 25 large studies with 100 ED patients or more, ranging from 100⁶⁵ to 4779 patients⁶ per study.

There were 12 prospective and 31 retrospective studies, with the remaining studies with an unspecified design (Table S2). Thirteen studies were multi-centered. Most studies were cross‐sectional (Table S2) and four studies were longitudinal^20,34,41,58, with the follow-up period from 3 to 120 months in the study by Rapaport & Rapaport²⁷, a follow-up of 12 weeks or longer in the study by Hong and colleagues⁴¹, the observation period of 60 days in the interventional study by Katsarou and colleagues⁵⁸ and an unspecified follow-up period in the study by Ornek and coworkers³⁴.

Setting

The studies were conducted various clinical settings, including dermatology or ophthalmology departments at the hospitals^{30,32,41,44,75}, universities^{23,33,35,36,39,43,47,50–53,56,57,59,62,66,68,71,73,76,78}, or institutes^48,67, contact allergy clinics at the medical centers^31,40, and a private practice^39,44. Four large studies were conducted at 153 research centers of the German Information Network of Departments of Dermatology (IVDK)^2,5,6,38. A large retrospective study of the North American Contact Dermatitis Group (NACDG)⁴ collected data from 13 sites across the US and Canada. A large prospective study was carried out by the members of Dermato-Allergology Group of the French Society of Dermatology in 14 centers across France⁷ whereas another large multi-center study—by 12 SIDEV-GIRDCA research centers in Italy³. Selected studies were carried out across Europe, North America, Asia and Australia.

Study population

In 47 studies with gender reporting, there were 17,309 women, with the pooled proportion of 82.7% [95% CI 0.795; 0.856] (Fig. 2A). There was no difference in the racial distribution between ED patients and a general patch-tested patient population⁵³ and no association between skin phototype and any specific diagnosis in ED patients³.

Figure 2.

(A-Q) Forest plots for demographic and clinical characteristics of patients with eyelid dermatitis. Legend: The proportional meta-analyses [random-effects] were carried out for female patients (A) and patients aged 40 years or older (B) in the studies including the patients with eyelid dermatitis. The proportion of atopic eyelid dermatitis in the studies with patients with eyelid dermatitis (C) was estimated at 27.5% (95% CI 0.177, 0.384). The differential diagnosis of eyelid dermatitis includes eyelid allergic contact dermatitis (D), seborrheic dermatitis (E), psoriasis (F), and irritant contact dermatitis (G). H–N represent proportional meta-analyses for contact sensitizations to nickel (positive patch tests and clinically relevant reactions), gold, thimerosal, neomycin, and corticosteroids (budesonide and toxicortol), respectively. Clinically, the proportion of isolated eyelid dermatitis (O) of 8453 ED patients was 51.6% [95% CI 0.408, 0.623] whereas the proportion of isolated atopic eyelid dermatitis (P) of the patients with atopic dermatitis with an eyelid involvement was 26.9% [95% CI 0.117, 0.449].

In 13 studies, 9470 ED patients were aged 40 years or older, with the pooled proportion of 65.2% [95% CI 0.568; 0.731] (Fig. 2B). Isolated ED was more common in patients aged 30–70 years⁵³.

Nethercott and colleagues³² noted that AED patients were younger than patients with ACD, ICD or other dermatoses. In the study by Svensson and Möller²³, AED patients had a mean age of 42.2 years compared to that of 50.2 years (19–83) in ED patients. Of interest, 5/15 AED patients with previous or present flexural dermatitis had a mean age of 30 years²³. As reported by Besra and coworkers⁶⁷, AED patients were aged 28.7 ± 17.3. In the study by Hong and coworkers, 94.5% of 127 AED patients were under 30 years old⁶⁶.

AED frequency

AED was reported in 27.5% [95% CI 0.177, 0.384] of ED patients across 29 studies (Fig. 2C). In two largest retrospective studies, AED was reported in 16.6% of 4779 patients⁶ and 13.4% of 3955 patients⁴. AD was more frequent in ED patients (including ED only⁴) than non-ED patients^4,32, however, this was not confirmed in other studies^53,71.

AED duration

ED duration can vary from a few weeks to several years^3,7,40,75. The mean ED duration was 31.4 months (1–360 months) in a French prospective multicenter study (n = 264)⁷, which was in keeping with that of 29.5 months (1–180 months) in a retrospective single-centered study from Turkey (n = 60)⁷⁵.

According to Bosco and coworkers⁷¹, 73.3% (140/191) of ED patients had a chronic/chronic recurring course (> 2 months). Assier and colleagues⁷ reported a chronic ED in 26.9% (71/264) cases.

Long-lasting AED (> 6 months) was noted in 69.1% (41/60) of patients by Ayala and colleagues³. In a follow-up by Ornek and coworkers⁷⁵, 80% of AED patients had a prolonged disease. In the study by Hong and coauthors⁶⁶, 26.3% of 80 AED patients had a relapse during a follow-up.

Risk factors

Feser and colleagues² identified female gender, age ≥ 40 years and atopic skin diathesis as ED risk factors and a past history of AD was more frequent in ED patients compared to all patch-tested patients in Erlangen. In the prospective study of 447 ED patients³, a stepwise logistic regression model suggested a disease onset over 6 months before the examination [OR 4.9 [95% CI 1.8–13.9] and a personal history of atopy [26.9 [95% CI 11.0–65.4]] as the main AED risk factors.

Lesion distribution

In 26 studies, 51.6% [95% CI 0.408, 0.623] of 8453 ED patients presented with an eyelid involvement only, including in total 430 patients with isolated AED (Fig. 2O,P). Of 33 adult-onset AED patients, 14 had isolated AED, 12 AED with facial dermatitis and seven patients—AED with an involvement on other body sites⁴².

The lesional distribution in ED patients is of great clinical interest but the number of studies for each localization (face, head/neck, hands and other body parts) did not allow for meta-analyses. There is limited data on the distribution of eczematous lesions on the eyelids in patch-tested ED patients. In the study by Ayala and coworkers³, all patients with four eyelid involvement had ED only. In the French prospective study,⁷ four eyelids were involved in 185/238 patch-tested ED patients, upper eyelids—in 47/238, lower eyelids in 18/238 and unilateral involvement in four patients only. Crouse and coworkers⁷⁶ reported that there were five AD patients out of 18 ED patients with a bilateral eyelid involvement. Whereas, ACD was diagnosed in patients with unilateral ED⁷⁶, bilateral⁷⁶ or four-eyelid involvement^3,76. Notably, drug-induced ACD was noted to affect lower eyelids⁶⁷. Additionally, Dennie‐Morgan fold was reported in 18.2% (8/44) of ED patients⁶⁷.

Comorbidities

Atopic diseases

A history of atopy (defined as a personal or family history of eczema, asthma, or hay fever with or without positive prick tests to aeroallergens) in ED patients was reported in up to 50% of ED patients³⁷. Ayala and colleagues reported a personal history of atopy in 80% of 55 AED patients³. In the study by Bosco and coworkers⁷¹, there was no difference in concomitant atopy between ED and non-ED.

Ocular diseases

Ocular symptoms were common in ED patients. Temesvári and coworkers⁵⁷ reported that 30.4% of 401 ED patients had ocular symptoms, including chronic conjunctivitis (119/151 patients), blepharoconjunctivitis (31/151 patients) and corneal ulcus (1 patient). Assier and colleagues⁷ found conjunctivitis in 74 of 264 ED patients. Ocular symptoms were noted in 24.73% of 93 ED patients by Rojo-España and coworkers²⁹.

Allergic contact dermatitis

In the meta-analysis of 38 selected studies, ACD was reported with a pooled proportion of 57.8% [0.477; 0.677] of 18,155 patch-tested ED patients, being the most common diagnosis in ED patients (Fig. 2D). Despite not being the focus of this study, these data provide important insights into relevant contact allergens in ED patients.

The most frequent contact allergens in ED patients were metals, fragrances and preservatives (Tables S3A-C). Nickel sulfate was the leading contact allergen in 23.2% (95% CI 0.167; 0.303) of 14,273 ED patients (Fig. 2H), with a wide variation in its clinical relevance ( Fig. 2I). The data on contact sensitization to nickel sulfate in ED patients compared to non-ED patients is conflicting (Table S5)^7,37,49,61. Gold sensitization was noted in 13.5% [95% CI 0.037; 0.273] of 1845 ED patients (Fig. 2J) and was more frequent in ED than non-ED patients in the study by Warshaw and colleagues.⁴

Herbst and colleagues⁵ noted a downward trend in contact sensitization to fragrances over time. Importantly, contact sensitization to fragrances may occur by direct contact or through airborne exposure⁷. There was an upward trend in contact sensitization to MI/MCI after 2013 due to a global outbreak^4,7,71. ED patients had more frequently positive patch tests to Kanthon CG than those without ED³⁴. In the study by Bosco and colleagues⁷¹, MCI/MI sensitization was significantly more frequent in ED patients with an eyelid involvement only compared to ED patients with associated dermatitis at other body areas, highlighting an exclusive palpebral involvement for MI/MCI contact allergy. The cumulative frequency of contact sensitization to thiomersal in ED patients was estimated at 6.2% [0.046; 0.080] (Fig. 2K), although this may be of past clinical relevance due to thimerosal-containing vaccines. In the study by Feser and coauthors², ophthalmic therapeutics were clinically relevant in older ED patients compared with all patch-tested patients. Contact sensitization to neomycin was reported in 1.43% [95% CI 0.053; 0.266] of ED patients (Fig. 2L). By contrast, occupational cases were less frequent in ED patients than non-ED patients⁶.

Given multiple contact sensitizations in some patients, the relevance of contact sensitizations to certain allergens may well have been overestimated³². Clinical relevance was significantly higher in ED patients with multiple contact sensitizations than in monosensitized individuals²⁹.

Atopic eyelid dermatitis with superimposed allergic contact dermatitis

There were 164 AED patients with superimposed ACD in nine studies (Fig. 2Q). In the study by Ochenfels and colleagues³⁸, positive patch tests were found in every third AD patient. Contact sensitizations in AED patients were reported only by Ayala and coauthors³: nickel sulfate—29.6%, Dermatophagoides mix—44.4%, perfume mix—11.1% and cobalt chloride—11.1%. However, the clinical relevance of these sensitizations in AED needs further research.

Diagnostic work-up

Questionnaires

Standardized questionnaires were used in all multicenter studies^3,5–7,38. The Erlangen Atopy questionnaire and the MOAHLFA (male, occupational dermatoses, atopic eczema, hand dermatitis, leg dermatitis, facial dermatitis, age ≥ 40 years) was used in ED patients in Erlangen and IDVK (excluding Erlangen) cohorts². Ocular questionnaires [e.g. Visual Functioning Questionnaire 25 (VQF-25)] or standardized measurement instruments (e.g. Total Ocular Symptom Score (TOSS) or Ocular Surface Disease Index) were not used in the selected studies.

Patch testing

From 1986 to 2024, 21,793 patch-tested ED patients across 65 studies were patch tested to exclude suspected ACD, using NACDG panel^{4,32,40,53,76}, European standard series^{7,32,37,41,42}, standard ICDRG panel²³, and country-specific baseline series^{2,3,29,38,61,75}. According to Amin and Belsito⁵³, most culprit contact allergens in ED patients were not included in standardized panels in the US.

In the selected studies, specialized series included fragrance, ophthalmic drug, eyelid series, preservative, antiseptic, facial series, medicament series, and parfum series. The use of specific eyelid/ophthalmic patch testing series yielded positive results in 4/264 ED patients in the study by Assier and coworkers⁷, and in 9/133 ED patients as reported by Temesváry and colleagues⁵⁷.

Importantly, positive reactions were found in patients tested with their own products, which included facial moisturizers, cosmetics, ophthalmic medications, eye drops, nail vanishes and glue products. Assier and coworkers⁷ reported that 19% (16/82) of the allergic patients were only identified by testing with their personal products. In the study by Katz and colleagues⁴⁰, 11 of 32 patients, who reacted to their own products, had negative patch tests with standard series. Rojo-Espana and colleagues²⁹ carried out open tests on the periorbital area in 36/93 patients, using the products as patches on one eyelid or as drops in one eye. In a French study²⁸, repeated open application test or a use test was positive in 4/16 patients with negative patch testing.

Positive patch tests to dust mite mix were reported in 32 of 58 ED patients in Guin’s clinical practice⁴⁵ and in 44.4% of AED patients in the study by Ayala and coworkers³. Pónyai and coworkers⁵⁶ reported positive atopy patch tests to house dust mites in six AD patients with eyelid symptoms.

Allergy testing with aeroallergens

As reported by Ornek and colleagues⁷⁵, in 15% (9/60) ED patients with positive prick tests, the most prevalent allergens were house dust mites (HDM) (D. pterynissinus—11.6% (7/60); D. farina—11.6% (7/60), grass mix—5% (3/60), and cat dander—3.3% (2/60). In four studies^7,42,45,49, prick tests with aeroallergens were performed without reporting the results. Component-resolved diagnosis has not been used in selected studies, therefore, the molecular patterns of HDM sensitization in AED patients remains to be established.

Microbiome studies

There were no microbiome data in AED patients in the selected studies.

AED clinical diagnosis

Several diagnostic criteria were used across selected studies, including the H&R diagnostic criteria^29,34,56,61 and the UKWP criteria⁷. Some authors relied on a personal or family history of atopy, personal history of eczema in childhood/early adulthood, lesion distribution typical for AD with⁴² or without positive prick testing to aeroallergens^45,71, but lacking contact allergy. Others diagnosed AD based on present or past history of flexural or popliteal eczema^23,42. Feser and colleagues² used the Erlanger Atopy Score (a mean score ≥ 12.5 points for AED patients). Svensson and Möller²³ used their own scoring system of patient’s history and clinical data.

Differential diagnosis

The most common cause of ED was ACD (Table 2). Common differential diagnoses in ED patients include ICD (Fig. 2G), seborrheic dermatitis (Fig. 2E) and airborne dermatitis. Warshaw and colleagues⁴ reported ACD to be more common in ED patients with or without a head/neck involvement, possibly due to allergen spreading to the adjacent areas. Rare causes of ED were psoriasis (Fig. 2F) and rosacea. Guin⁴⁹ reported tinea involving the eyelids (2 patients) and viral epidemic keratoconjunctivitis (1 patient) as rare ED causes. There were three patients with periorbital discoid lupus erythematosus^57,75, one patient—with progressive systemic sclerosis⁵⁷ and one patient—with mucosal pemphigoid².

Table 2.

Differential diagnosis in patients with eyelid dermatitis.

Common diseases presenting as ED (> 5% in selected series)	Rare diseases presenting as ED (< 5% in selected series)
Allergic contact dermatitis	Seborrheic dermatitis
Atopic eyelid dermatitis	Airborne dermatitis
Irritant contact dermatitis	Psoriasis
	Rosacea

Other diagnoses (< 1% in selected series): discoid lupus erythematosus, progressive systemic sclerosis, mucosal pemphigoid, dermatomyositis.

Treatment

Rappaport and Rappaport²⁷ observed 67 patch-tested AED patients, who chronically used topical and/or systemic CS. The addiction to topical CS may occur within two months of continuous use. In AED patients, the duration of the previous CS use was 3–480 months, mostly 1–2 years. Burning was noted in 43%, telangiectasia in 31%, and skin atrophy—in 26% of AED patients. Noteworthy, contact sensitization to budesonide was reported in 1.5% [0.007; 0.026] of 1795 ED patients (Fig. 2M) and to tixocortol—in 1.4% [0.005; 0.025] of 1840 patch-tested ED patients (Fig. 2N). On stopping the CS use, distant eczematous rashes (often on antecubital areas, or the legs or upper chest) occurred in 20/67 AED patients. Occasionally, UVB phototherapy was used upon a CS withdrawal. Clearing, without AED flare-ups for 4–6 weeks, was achieved in 2–30 months of a CS discontinuation²⁷.

Discussion

Our main finding was the AED frequency of 27.5% (95% CI0.177, 0.384) in patch-tested ED patients, exceeding the previous estimate of over 10%²². Importantly, isolated AED, as a distinct localized AD phenotype in adults, was reported in 430 AD patients, although its frequency in adult AD patients remains unknown. Patients with isolated AED may be more likely to attend contact allergy clinics and, therefore, may not have been included in large AD cohorts. Flexural AD is frequently included as diagnostic criterion in epidemiological studies⁸⁶, and thus may overlook patients with isolated AED. From a global perspective, published studies with AED patients were mostly carried out in Europe and the US, were underrepresented in Asia and Australia, and were lacking in South America and Africa. Future studies are needed to elucidate the frequency of isolated AED in adult AD patients, taking into account a global variation in AD phenotypic presentations across races and ethnicities⁸⁷.

High proportion of ED patients aged 40 years or older may reflect age-dependent variations in stratum corneum thickness and hydration in the eyelid area. Limited evidence suggests that AED patients may be younger than ED patients. This is in keeping with a recent prospective study⁸⁸, a meta-analysis⁸⁹ and our clinical experience suggesting higher AED likelihood in children and young adults. Future studies into gender- and age-related differences in eyelid stratum corneum morphology and function in ED/AED patients may provide an explanation for their demographic characteristics.

An ED duration may exceed 6 months in some patients^3,7,27. This warrants further studies into an AED natural history. Published AED case reports²⁶ suggest that AED may last up to 10 years or longer, thus raising the questions as to which factors (the country of residence, airborne pollution, climatic differences, allergen sensitization, genetic and socioeconomic factors, etc.) may underlie the AED persistence in adults. It remains unclear whether chronic periorbital changes (chronic lichenification, Dennie-Morgan infraorbital fold(s)) in AED can be prevented.

Clinical scoring of AED severity has been proposed⁹⁰. Transepidermal water loss (TEWL) can be measured on the eyelids, using closed-chamber devices with goggle-adaptors^90,91, although eyelid TEWL can be affected by tear evaporation. Comorbidities may increase the AED burden. In a meta-analysis²⁸, conjunctivitis was reported in 31.7% (95% CI 27.7–35.9) of AD patients, although the data for AD phenotypes (including AED) was limited. In a prospective study¹⁷, moderate-to-severe ocular surface disease was significantly more frequent in patients with AED or AD facial involvement in the past year. Future studies into ocular surface disease in AED patients may help elucidate whether AED patients should have a baseline ophthalmic examination. Until then, a referral to an eye care specialist is warranted for patients with conjunctival scarring or swelling, pain, watery discharge, blepharitis, dryness, irritation and light sensitivity or vision problems⁹² or symptomatic AD patients, whose ocular symptoms do not improve within 2 weeks⁹³. However, some authors advocate a baseline ophthalmological examination for at-risk individuals, including ED patients⁹².

The link between ACD and AD has been a subject of controversy and intense research⁹⁴. Our meta-analysis suggested that superimposed ACD may occur in AED patients. In our analysis, ED patients were frequently sensitized to metals, fragrances and preservatives which is in keeping with previous studies of eyelid ACD^64,95, and general patch-tested population⁹⁶. Some contact allergens, however, were more frequent in ED patients^4,32,53,71, particularly, with a predilection to an exclusive eyelid involvement for MI/MCI⁷¹. An individualized approach with baseline and specialized panels and, importantly, patient’s own products², related to patient’s occupational, personal care products and/or topical medications, can be useful in AED patients. Albeit being used by some authors^29,97, patch testing or repeated open application test on eyelid(s) has to be further evaluated. A conjunctival challenge test was proposed for ED patients with conjunctival symptoms to topical ophthalmic drugs⁶.

In clinical practice, it may be difficult to distinguish isolated AED and ICD on clinical grounds or by using available diagnostic approaches (patch testing, repeated open application test). Skin testing with irritants and repeated open application tests may be helpful to diagnose ICD in healthy individuals but not in atopic patients as these two conditions may co-exist. Skin biopsy on the eyelid area is rarely used for differential diagnosis of ED in dermatological practice. In research, transcriptomic biomarkers by skin tape stripping have been successfully used for differential diagnosis of AD, psoriasis, ACD and ICD^98,99 but not specifically on the eyelid area. It remains to be established whether transcriptomic biomarkers in the skin and the tear fluid can be used for differential diagnosis of AED and other skin conditions affecting the eyelid area.

Non-pharmacological options include allergen avoidance, including relevant contact allergens and aeroallergens if possible. Until further evidence, the avoidance of certain contact allergens is not warranted in all AED patients¹⁰⁰. Some ED patients sensitized to gold sodium thiosulfate may benefit from a 2-to-3 month trial of gold jewelry avoidance to establish clinical relevance¹⁰¹. HDM avoidance was shown to be effective in reducing the AD severity due to the reduction of Der p 1 concentration in the domestic environment in a landmark UK-based double-blind controlled study¹⁰². In illustrative cases by Beltrani¹⁰¹, AED patients may benefit from HDM elimination measures, although this requires further research in larger cohorts. The use of artificial tears can be helpful.

A short trial with low-potency topical CS may be required. Unfortunately, AED patients may over-use topical CS, with CS dependency occurring within two months of a continuous use.²⁷ A prolonged use of topical CS in the eyelid area is associated with the risk of increased intraocular pressure in at-risk individuals¹⁰³. Contact allergy to CS (Fig. 2M) may occur in ED patients^{5,7,29,41,45,53,71,75,76} but its frequency in AED is currently unknown. Topical calcineurin inhibitors are preferred for the eyelid areas. The phosphodiesterase 4 inhibitor crisaborole can be used around eyes¹⁰⁴. Narrowband UVB phototherapy should be used with caution for AED patients¹⁰⁵, because keratinopathy was reported in an AED patient¹⁰⁶. Allergen-specific immunotherapy has not been evaluated in AED patients with or without atopic comorbidities. The use of biological therapies (omalizumab and dupilumab) in AED patients is of clinical interest. Janus kinase inhibitors (Table 3) were shown to be effective in an ED patient¹⁰⁷ and in AD patients with severe blepharitis¹⁰⁸.

Table 3.

Management of atopic eyelid dermatitis

First line treatments

Eyelid hygiene, artificial tear *

Ice or cool water compresses, 4 times a day for 15 min2*

House dust mite (HDM) avoidance in sensitized AED individuals102*

Avoidance of contact allergens or foods, containing relevant allergens, in sensitized AED patients.

A 2–3 month trial of gold jewelry avoidance in sensitized ED patients43*

Emollients

Topical calcineurin inhibitors ****

Short-term use of mild topical corticosteroids (with caution due to the risk of cutaneous atrophy and elevated intraocular pressure).

?Facial masks with soft zinc paste9*

Second line treatments

Eyelid phototherapy105*** (with caution due to the risk of phototherapy-induced keratopathy and an early onset of cataract106).

? Topical Janus kinase inhibitors107,108**

Third line treatments.

? Allergen-specific immunotherapy (ASIT) with HDM allergens.

? Omalizumab.

? Dupilumab.

AD atopic dermatitis, ASIT allergen-specific immunotherapy, HDM house dust mites, JAK Janus kinase.

*Expert opinion.

** A case report(s).

*** Patient series.

**** Systematic review.

In our comprehensive analysis of 65 studies with adult ED patients, the strengths rely mostly on high-quality studies, including large multi-center studies^2–7. and the use of patch testing in ED patients. However, our analysis revealed certain limitations. Firstly, the studies with ED patients were carried out with different objectives⁷, study designs and study populations, thus resulting in a high heterogeneity between the included studies which is common in the proportional meta-analyses in the prevalence studies¹⁰⁹. Secondly, different AED criteria across the studies may not be equally sensitive to adult-onset AD or phenotypic differences in patients of various descent⁵⁷. Furthermore, a referral bias to contact allergy clinics⁴² may have led to an ACD overrepresentation. Whereas, AED patients may have been under-reported due to a frequent recall bias for AD²⁸. Most studies were carried out in the academic settings, thus reducing the generalizability of the results to the real-life clinical practice. Our inclusion criteria and the search period (1980–2024) did not include some large ED series without patch testing¹¹⁰. Nevertheless, we believe that our systematic review allowed us to glean important insights into ED/AED demographic and clinical features which can be useful for clinical practice.

Many unresolved questions related to AED clinical profiling remain. In the future, the AED disease burden requires a closer attention, particularly in AED patients with all skin types worldwide, given notable AD phenotypic differences in various geographical regions. Age-related differences in AED clinical presentations may provide further insights into adult-onset AED. An ongoing multicenter observational prospective PAUPIAD study (NCT04826744) in France will characterize a palpebral involvement in adult AD patients.

Much work needs to be done to further understand the genetic background, environmental factors (aeroallergens and pollutants) and gene-environmental interactions that may influence the AED development. Further research is needed on AED endotyping in adults based on allergen sensitizations, FLG mutations, microbiome and possibly by the transcriptome in the tear fluid and in the eyelid skin.

In conclusion, there is a need for an optimized diagnostic algorithm, better treatments and prevention strategies in AED patients. It remains to be seen whether patients with severe treatment-resistant AED may be eligible for topical (eg. JAK inhibitors, topical Syk inhibitor) or systemic biological agents (eg. omalizumab, dupilumab). In the future, precision medicine based on AED clinical profiling and endotyping, together with the AI use for digital eyelid images, may advance the AED management, thereby reducing patient’s disease burden, improving their outcome and preventing long-term sequelae.

Methods

PubMed/Medline database was searched from 1980, since the 1980 H&R diagnostic criteria²⁴, till 1 September 2022, using a search strategy: ((eyelid OR periorbital OR periocular) AND (dermatitis or eczema)) in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement (Appendix 1, Figs. S1–S2)¹¹¹. There were no language restrictions applied to the search strategy.

The inclusion criteria were: (1) the ED diagnosis; and (2) patch testing. The exclusion criteria were: (1) reviews/editorials/guidelines/correspondence; (2) abstracts, conference posters/oral presentations, (3) studies not in humans; 4) diagnoses other than ED (5) patients aged < 18 years old; (6) drug-related side-effects with an eyelid involvement; (7) unclassified ED and (8) dupilumab-associated ocular conditions in AD patients. The reference lists were screened for additional studies. Case reports with 1–3 cases were excluded.

Initial title and abstract screening and full-text reviews were carried out by three reviewers (EB, ES and AB) (Tables S2, S4). Disagreements between reviewers were resolved via consensus. The methodological quality of selected studies was assessed using Newcastle–Ottawa scale modified for AD cross-sectional studies (Tables S1)¹¹². A proportional meta-analysis was carried out using JBI SUMARI software. Between-study heterogeneity was assessed by the I² statistic.

Abbreviations

ACD

Allergic contact dermatitis

AD

Atopic dermatitis

AED

Atopic eyelid dermatitis

CS

Corticosteroids

DKG

German Contact Dermatitis Research Group

ED

Eyelid dermatitis

FLG

Filaggrin gene

GIRDCA

Gruppo Italiano Ricerca Dermatitida Contatto Ambientali

HDM

House dust mites

H&R

The 1980 Hanifin and Rajka criteria

ICDRG

International Contact Dermatitis Research Group

IgE

Immunoglobulin E

IL-4

Interleukin—4

IL-5

Interleukin—5

IVDK

Information Network of Departments of Dermatology

LTB4

Leukotriene B4

MUC5AC

Mucin—5AC

NACDG

North American Contact Dermatitis Group

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

RCT

Randomized controlled study

QoL

Quality of life

SIDAPA

Italian Society of Allergological Dermatology

TEWL

Transepidermal water loss

TNF-α

Tumor necrosis factor—alpha

UKWP

UK Working Party’s diagnostic criteria

Author contributions

E.B.: Conceptualization, Methodology, Software, Visualization, Formal analysis, Supervision, Resources, Data curation, Writing—Original draft preparation. E.S.: Methodology, Formal analysis, Visualization, Resources, Supervision, Writing—Reviewing and Editing. A.B.: Formal analysis, Data curation, Visualization, Writing—Reviewing and Editing.

Data availability

All data used in the study are from publicly available sources that are listed in the tables and cited in the references. The tables and the study protocol and all datasets generated and analysed, including the search strategy, data extracted and quality assessment, are available in the Article and on request from the corresponding author.

Competing interests

EB serves on the WAO Committee on Atopic Dermatitis, received honoraria for educational lectures from Novartis and Sanofi and research funding from GSK, outside the submitted work. EB provided pro bono consulting to Kimialys (France) outside the submitted work. EB received research funding from the British Skin Foundation (2009) for her research, outside the submitted work. ES received honoraria for lectures from Dr. Reddy’s Laboratories Ltd., as well as support for attending congresses from Galderma and Drs. Reddy’s Laboratories Ltd. as a lecturer. AB is working on her PhD research project on atopic eyelid dermatitis supervised by ES and EB.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-024-69612-z.