*Article* **National Information Campaign Revealed Disease Characteristic and Burden in Adult Patients Suffering from Atopic Dermatitis**

**Niccolò Gori 1,2, Andrea Chiricozzi 1,2,\*, Franco Marsili 3, Silvia Mariel Ferrucci 4, Paolo Amerio 5, Vincenzo Battarra 6, Salvatore Campitiello 7, Antonio Castelli 8, Maurizio Congedo 9, Monica Corazza 10, Antonio Cristaudo 11, Gabriella Fabbrocini 12, Giampiero Girolomoni 13, Giovanna Malara 14, Giuseppe Micali 15, Giovanni Palazzo 16, Aurora Parodi 17, Annalisa Patrizi 18, Giovanni Pellacani 19, Paolo Pigatto 20, Eugenio Provenzano 21, Pietro Quaglino 22, Marco Romanelli 23, Mariateresa Rossi 24, Paola Savoia <sup>25</sup> and Ketty Peris 1,2**

	- <sup>24</sup> Department of Dermatology, University of Brescia, 25121 Brescia, Italy
	- <sup>25</sup> Department of Health Science, University of Eastern Piedmont, Via Solaroli 17, 28100 Novara, Italy
	- **\*** Correspondence: chiricozziandrea@gmail.com; Tel.: +39-339-5668320

**Abstract:** Atopic dermatitis (AD) is a common inflammatory skin disease often associated with a significant impairment in the quality of life of affected patients. The Italian Society of Dermatology and Venereology (SIDeMaST) planned a national information campaign, providing direct access to 27 dermatologic centers dedicated to the management of AD. The aim of this study aimed was to outline critical aspects related to AD in the general population. Overall, 643 adult subjects were included in this study, and in 44.2% (284/643) of cases, a diagnosis of AD was confirmed, whereas about 55% of subjects were affected by other pruritic cutaneous diseases. Higher intensity of pruritus and sleep disturbance, as well as an increased interference in sport, work, and social confidence

**Citation:** Gori, N.; Chiricozzi, A.; Marsili, F.; Ferrucci, S.M.; Amerio, P.; Battarra, V.; Campitiello, S.; Castelli, A.; Congedo, M.; Corazza, M.; et al. National Information Campaign Revealed Disease Characteristic and Burden in Adult Patients Suffering from Atopic Dermatitis. *J. Clin. Med.* **2022**, *11*, 5204. https://doi.org/ 10.3390/jcm11175204

Academic Editors: Stamatis Gregoriou and Yolanda Gilaberte

Received: 18 July 2022 Accepted: 27 August 2022 Published: 2 September 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

was reported in the AD group compared to the non-AD group. In the AD subgroup, the mean duration of disease was of 15.3 years, with a mean eczema area and severity index (EASI) score of 11.2, and investigator global assessment (IGA) score of 1.9 and an itch numeric rating scale (NRS) of 6.9. Almost 32% of patients were untreated, either with topical or systemic agents, whereas 44.3% used routine topical compounds (topical corticosteroids and calcineurin inhibitors), and only 7.0% of patients were systemically treated. Only 2.8% of patients reported complete satisfaction with the treatment received for AD to date. This study reveals a profound unmet need in AD, showing a poorly managed and undertreated patient population despite a high reported burden of disease. This suggests the usefulness of information campaigns with the goal of improving patient awareness regarding AD and facilitating early diagnosis and access to dedicated healthcare institutions.

**Keywords:** atopic dermatitis; information campaign; early diagnosis

#### **1. Introduction**

Atopic dermatitis (AD) is the most common chronic inflammatory skin disease, affecting nearly 230 million people worldwide, with a prevalence, in developed countries, ranging between 10% and 25% in children and 7–10% in adults [1]. AD is clinically characterized by intense itch, dry skin, and eczematous lesions, with the preferential involvement of flexures, head, neck, and hands in adulthood [1]. It is frequently associated with a personal and/or family history of atopic extracutaneous manifestations, such as allergic rhinitis, conjunctivitis, and asthma [1]. Several non-atopic diseases, including inflammatory, autoimmune, and mental health disorders, might be also observed in AD patients [2]. In addition, attention disturbances and poor sleep quality, likely related to itch, may occur, affecting school and work performance [3]. Treatment of mild AD is essentially based on the use of topical corticosteroids (TCSs) and calcineurin inhibitors (TCIs), whereas moderate-to-severe AD, accounting for as many as one-third of cases, is commonly treated with phototherapy and/or systemic therapies, including traditional immunosuppressants and novel immune-targeted therapies [4,5]. In addition, to improve skin dryness and reduce itch, moisturizers are usually applied daily, implying a significant economic burden for patients [6]. Notably, a recent cross-sectional study including nine European countries reported a mean annual personal extra out-of-pocket expense of EUR 927.12 for patients with AD [6].

Although they were conceived in the 1980s, the Hanifin and Rajka criteria are still the most used tool to diagnose AD in both clinical practice and research settings, whereas, considering the lack of specific diagnostic markers, diagnosis of AD is essentially based on the accurate evaluation of clinical signs, symptoms, and medical history by skilled physicians [7]. Diagnosis of AD is relatively easy in children but often challenging in adults, especially in late-onset forms, due to a broader clinical variability [7,8]. For this reason, adult AD is thought to be underdiagnosed, reflecting the highly variable prevalence of disease reported in the literature, with an estimated range varying from 0.3% to 14.3% [8–10]. To reduce the proportion of undiagnosed and undertreated cases of AD, it might be helpful to promote awareness of AD through multiple strategies, including the diffusion of disease information websites and the organization of screening campaigns. Despite the existence of numerous scientific societies and patient associations providing educational websites for AD patients, the first national information and screening campaign was organized in September 2020 by the Italian Society of Dermatology and Venereology (SIDeMaST), which provided direct access to several dermatologic centers dedicated to the management of AD, with the aim of improving patient awareness about AD and facilitating early diagnosis and access to optimal treatment management.

#### **2. Materials and Methods**

In this study, we considered data obtained from subjects referred to 27 dermatology centers homogeneously distributed in northern, central, and southern Italy for the first national AD screening program supported by the Italian Society of Dermatology (SIDeMaST). The purpose of the program was to promote the knowledge of AD among adult subjects with established disease or suspected symptoms.

In September 2020, information regarding the national AD screening program was posted on the patient-oriented AD website (www.dallapartedellatuapelle.it). In particular, the web site provided general information regarding pathogenesis and clinical presentation of AD in adults, as well as email and telephone contacts to join the AD screening program. Patients with suspected or diagnosed AD were screened by non-physician personnel on the phone or by e-mail through a brief questionnaire. As a screening program, it was not necessary to apply for ethics committee approval because patients did not furnish any sensitive data to the centers.

Inclusion criteria for prescreened patients were limited to comprehension of written Italian language and consent to compile a printed survey. In each dermatological center, patients were required to complete a 21-item questionnaire about demographic and clinical data, including age, sex, weight, height, job, disease duration, type of medical specialists previously consulted, personal and family history of AD and comorbidities, interference of AD with physical activities and work tasks, therapeutic management of disease, economic burden for supplying topical and systemic drugs, and consultations for AD. Individuals were subdivided according to working profession as white collar (intellectual jobs, including doctors, lawyers, teachers, office workers, managers, and civil servants) or blue collar (manual jobs, including craftsmen, farmers, specialized workmen, drivers of industrial machines/vehicles, armed services, and unqualified professions) [11].

Patients were evaluated by dermatologists with experience in inflammatory skin diseases to assess the diagnosis of AD and suggest the most appropriate therapeutic approach. Disease severity in patients with a confirmed diagnosis of AD was assessed using (a) the eczema area severity index (EASI), with scored ranging from 0 to 72; (b) the investigator global assessment (IGA), with scores ranging from 0 to 4; (c) the itch numeric rating scale (NRS), ranging from 0 to 10, assessing itch intensity (itch-NRS); (d) a 0–10 NRS scale rating sleeplessness (sleep-NRS); (e) a 0–10 NRS evaluating disease-induced embarrassment (e-NRS); (f) a 0–10 NRS evaluating the influence of disease on work tasks (w-NRS); and (g) a 0–10 NRS assessing the impact of disease on sporting activity (s-NRS).

#### *Statistical Analyses*

Categorical variables were analyzed as frequencies and percentages. Continuous variables were analyzed as mean and standard deviation (SD) or medians and interquartile ranges (IQRs). Variable normality was assessed by the Shapiro–Wilk W test. We compared questionnaire-obtained personal and clinical data between AD and non-AD groups, using a T test for comparison of means or Mann–Whitney test for comparison of medians, and by chi-square (or Fisher's exact test) for categorical variables. In the subgroup with confirmed AD diagnosis, clinical data were described in terms of frequencies and percentages, whereas EASI and IGA scores were analyzed as mean and SD. All statistical tests were two-tailed, and a *p*-value less than 0.05 was considered statistically significant. Analysis was performed using STATA 17 software (StataCorp, College Station, TX, USA).

#### **3. Results**

#### *3.1. Characteristic of the General Population*

Overall, 641 adult subjects were referred to the 27 outpatient dermatology centers during the open day in September 2020. Demographic and clinical data are summarized in Table 1.


**Table 1.** Characteristics of general population, and separately for atopic dermatitis (AD) and non-AD.

Legend. Categorical data expressed as n (%); continuous data expressed as mean ± standard deviation \* *p* value refers to the comparison between AD and non-AD population.

Work activity was classified as intellectual jobs (white collars), accounting for 40.6% of participants (260/641); or manual jobs (blue collars), accounting for 24.0% of participants (154/641); whereas 18.4% of participants (118/641) were students and 21.5% (138/641) were retired professionals. Proportions of 84.9% (544/641) and 34.5% (221/641) of participants reported to have visited a dermatologist an allergist, respectively, at least one time for their skin problems, whereas only 6.4% (41/641) were managed by their general practitioner.

The mean duration of skin manifestations was 10.7 ± 11.8 years. Mean itch-NRS and sleep-NRS values at the time of evaluation were 6.5 ± 2.5 and 4.4 ± 3.4, respectively. Approximately 55% (352/641) of patients reported itch as the major cause of discomfort, whereas another 17.3% (111/641) of subjects identified the presence of eczematous lesions as the main burdening factor, with 26.5% (170/641) reporting being equally disturbed by both manifestations.

*3.2. Clinical Features of the AD Subpopulation Differ from Those of the Non-AD Population* A diagnosis of AD was confirmed in 44.3% (284/641) of the screened population.

On the other hand, 55.7% of subjects were included in the non-AD group, as they reported being affected by other pruritic cutaneous disease, including seborrheic dermatitis, psoriasis, allergic contact dermatitis, scabies, prurigo nodularis, dermatitis herpetiformis, bullous pemphigoid, and pruritus.

Demographic and clinical data of each subpopulation are summarized in Table 1.

Approximately 77% (219/284) of AD patients had previously received a diagnosis of AD. In 35.3% (126/357) of non-AD patients, an erroneous diagnosis of AD had been previously made. Most AD patients (93.0%; 264/284) visited a dermatologist at least once, compared to 77.9% (278/357) of the non-AD population, whereas 48.9% (139/284) of AD compared to 20.4% (80/357) of non-AD patients had visited their skin disease evaluated by an allergologist and 2.1% (6/284) of AD versus 9.2% (33/357) of non-AD patients by a general practitioner. Mean duration of skin disease was 15.3 ± 12.6 years in the AD group and 7.0 ± 9.7 in the non-AD group (*p* < 0.0001). A family history of AD, rhinoconjunctivitis, and asthma was significantly more frequent in AD patients compared to non-AD patients (*p* < 0.0001). A higher intensity of pruritus and sleep deterioration were detected in the AD group, with mean values of itch- and sleep-NRS of 6.9 ± 2.4 and 5.0 ± 3.4, respectively, compared to 6.2 ± 2.6 and 3.9 ± 3.4 in the non-AD cohort (Table 1). Moreover, a higher grade of disease-related embarrassment and disease interference with sport and work was reported in the AD group compared with the non-AD population (Table 1). The regular use of TCS or TCI was reported in 44.3% (126/284) of AD patients compared to 37.2% (133/357) of non-AD patients (*p* < 0.0001). In addition, AD patients reported a greater use of moisturizers compared to non-AD patients (56.3% (160/284) versus 37.5% (134/357), *p* < 0.0001). On the contrary, no significant differences between the two populations were detected in terms of the use of systemic therapies. Only 2.8% (8/284) of AD patients and 4.5% (16/357) of the non-AD group reported complete satisfaction with therapy received, whereas 34.5% (98/284) of AD subjects and 18.8% (67/357) of non-AD sub-cohort reported only partial satisfaction, and 57.7% (164/284) of AD patients and 60.2% (215/357) of non-AD patients reported no satisfaction (*p* < 0.0001).

No significant difference was detected in terms of the average monthly expense for topicals, systemic drugs, and visits between the AD and non-AD groups (Table 2).

**Table 2.** Predictors of average monthly expense > 20 euros for topical therapies, systemic therapies, and visits.


Legend. \* Model adjusted for: age, gender \*\* Model adjusted for age, gender, EASI. Statistically significant results are highlighted in bold.

#### *3.3. Physician-Oriented Assessment of AD Patients*

In individuals with a confirmed diagnosis of AD (284 patients), mean EASI and mean IGA scores were 11.2 ± 12.0 and 1.9 ± 2.0, respectively. Moderate-to-severe AD, defined by an EASI score ≥16 and an IGA score ≥3, was reported in 21.5% (61/284) and 25.3% (72/284), respectively. The upper limbs were the body site more frequently affected by skin lesions (72.2%), followed by head and neck (49.6%), lower limbs (44,0%), trunk (28.2%), and back (25.7%). History of atopic comorbidities was reported in 48.2% of patients; in particular, rhinitis was described in 36.3% (103/284) of patients, conjunctivitis in 20.4% (58/284), and allergic asthma in 21.8% (62/284). In this patient population, the increase in disease severity scores was directly associated with a monthly expense for topical treatments of more than EUR 20, whereas no significant correlation between disease severity and monthly expense was observed for systemic drugs and visits (Table 2).

#### **4. Discussion**

AD is a chronic inflammatory skin disease associated with a significant deterioration of patients' quality of life [1,3]. Although it is the most common inflammatory skin disease, the current lack of specific diagnostic markers and criteria makes the identification of adult AD challenging, particularly in the adult-onset subtype [7,12].

The latest national and international guidelines suggest that diagnosis of AD in adulthood is essentially clinical, based on evaluation of morphology and distribution of lesions and the exclusion of possible differential diagnoses, including allergic contact dermatitis, scabies, dermatitis herpetiformis, and cutaneous lymphomas [7,13]. The lack of experienced general practitioners and territorial dermatologists in recognizing adult AD could result in an underestimation of disease prevalence and burden.

In this study, 44.3% (284/641) of the whole population received a clinical diagnosis of AD, and in 22.9% (65/284) of these cases, a different diagnosis was proposed during previous visits, most of which had been performed by dermatologists.

Notably, more that 35% (126/357) of patients who resulted not affected by AD in this study had received an incorrect diagnosis of AD during previous visits, thus revealing not only a low sensitivity but also a low specificity in the diagnosis of adult AD with possible overestimation of disease prevalence in some cases. This significant number of misdiagnosedcases of AD in the studied population suggests the importance of information campaigns dedicated to the general population and the relevance of scientific activities with respect to increasing knowledge and awareness of AD among physicians.

Notably, serological markers currently used by physicians to support the diagnosis of AD are limited to total and/or allergen-specific serum IgE levels and peripheral eosinophil counts, which are characterized by low sensibility and specificity [14]. The recent discovery of a new subset of T-cell cytokines and chemokines has resulted in the introduction of multiple potential biomarkers, including serum levels of CD30; macrophage-derived chemoattractant (MDC); interleukins (IL)-12, -16, -18, and -31; and thymus and activationregulated chemokine (TARC) [14–16]. Although none of these novel biomarkers have proven reliable for the diagnosis of AD in clinical practice to date, we consider further research necessary in this field with the aim of simplifying the diagnosis and management of the disease [14].

In the AD subgroup, the mean duration of disease was of 15.3 years, with a mean EASI score of 11.2, IGA score of 1.9, itch-NRS of 6.9 and sleep-NRS of 5.0. Interestingly, AD patients showed the highest values of all patient-reported outcomes in comparison with non-AD patients. Statistically higher severity in sleep disturbances was observed in the AD group vs. 3.9 in non-AD group (*p* < 0.0001). Notably, sleep disturbances are a well-known manifestation associated with AD, with a prevalence ranging between 33% and 81.7% in adults, not exclusively related to itch but also to immunological and neuroendocrine imbalance [17]. A few studies suggested a correlation between poor sleep quality and AD, regardless of disease status, suggesting that repeated flares of AD over time can lead to behavior-related sleep disorders persisting despite disease remission [18]. Thus, high prevalence of sleep disturbances detected in our AD patients could be also explained by the significantly longer duration of disease reported in the AD subgroup (15.6 years) compared to non-AD individuals (6.9 years) (*p* < 0.0001).

Furthermore, chronic sleep disorders have been identified as one of the most important risk factors for the development of several non-atopic comorbidities in AD, including mental health disorders (e.g., anxiety, depression) and cardiovascular diseases (e.g., coronary artery disease and hypertension) [1,2,19]. All these comorbidities may in turn adversely affect sleep quality and increase the disease burden of AD patients [19].

A recent survey including Irish adult patients affected by AD revealed a negative influence of disease on social and relational life, with 70% of patients reporting social anxiety, 65% avoiding sport and physical activities, 52% avoiding social activities, and 52% avoiding sexual intimacy [20]. Similarly, we detected significantly higher social embarrassment and interference with sport activities and job tasks in the AD subgroup compared to the non-AD subgroup, confirming AD as a severely debilitating cutaneous disease with multiple effects on patients' overall quality of life. The effect of AD on adults and children can currently be determined by different quality of life questionnaires, the most used of which are the Dermatology Quality of Life (DLQI), Children's Dermatology Quality of Life, and Infants Dermatology Quality of Life questionnaires; it is important to consider that all these tools are not specific for AD [21–23].

Considering the extensive and multimodal burden of AD, the development and evaluation of new specific questionnaires to evaluate the multiple domains influenced by AD would be very useful.

Almost one-third of patients did not use any compound, either topical or systemic, whereas only 7.0% of patients were treated with a systemic therapeutic agent. Notably, the economic burden of topical therapies, which are not covered by the national health care system, might negatively impact treatment access. In this study, we found a positive correlation between the monthly expense for topicals and both patient- and physicianassessed disease severity, suggesting that poorly controlled AD requires an increased use of topical agents. This increased use of topical agents could be due to undertreatment, which does not include systemic agents, which are only prescribed in a small percentage of patients (7.0%), notwithstanding the consistent number of subjects suffering from moderateto-severe AD (EASI ≥ 16 and IGA ≥ 3 reported in 21.5% and 25.3% of patients, respectively).

Importantly, only 2.8% of patients reported complete satisfaction with treatments received to date, showing profound unmet therapeutic needs among adult patients affected by AD [24]. Notwithstanding the recent introduction of novel targeted therapies approved for AD, which can be prescribed by tertiary healthcare centers only, more than 80% of patients reported lack of awareness about the existence of these therapeutic opportunities [25–28]. This indicates the necessity of creating a proactive network connecting territorial dermatologists to secondary and tertiary centers with the aim of enhancing the therapeutic management of AD patients.

In conclusion, this study underlines the utility of organizing information campaigns on AD to enhance awareness regarding disease features and management and to facilitate early diagnosis with a subsequent reduction in the burden of disease.

**Author Contributions:** Conceptualization, N.G., A.C. (Andrea Chiricozzi) and K.P.; writing—original draft preparation, N.G.; writing—review and editing, A.C. (Andrea Chiricozzi); data curation, F.M., S.M.F., P.A., V.B., S.C., A.C. (Antonio Castelli), M.C. (Maurizio Congedo), M.C. (Monica Corazza), A.C. (Antonio Cristaudo), G.F., G.G., G.M. (Giovanna Malara), G.M. (Giuseppe Micali), G.P. (Giovanni Palazzo), A.P. (Aurora Parodi), A.P. (Annalisa Patrizi), G.P. (Giovanni Pellacani), P.P., E.P., P.Q., M.R. (Marco Romanelli), M.R. (Mariateresa Rossi), P.S. and K.P.; supervision, K.P. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Informed consent was not required from subject who joined the AD screening campaign.

**Data Availability Statement:** The data relating to this study are available in an excel file.

**Acknowledgments:** We thank the following individuals for their contribution the collection of patient data: Giacomo Caldarola, Silvia Mariel Ferrucci, Teresa Grieco, Maria Mariano, Maddalena Napolitano, Michela Ortoncelli, Natale Schettini, and Federica Veronese. We thank SANOFI for the unconditional financial contribution offered for the information campaign on atopic dermatitis.

**Conflicts of Interest:** The authors declare no conflict of interest related to this study.

#### **References**


**Manuel Almenara-Blasco 1,\*,†, Jonás Carmona-Pírez 2,3,4,†, Tamara Gracia-Cazaña 1, Beatriz Poblador-Plou 2,3, Juan Blas Pérez-Gilaberte 5, Alba Navarro-Bielsa 1, Antonio Gimeno-Miguel 2,3,‡, Alexandra Prados-Torres 2,3,‡ and Yolanda Gilaberte 1,‡**


**Abstract:** Background: Atopic dermatitis (AD) is associated with different comorbidities. Methods: Retrospective, observational study based on clinical information from the individuals of the EpiChron Cohort Study (Aragon, Spain) with a diagnosis of AD between 1 January 2010 and 31 December 2018. We calculated the tetrachoric correlations of each pair of comorbidities to analyze the weight of the association between them. We used a cut-off point for statistical significance of *p*-value < 0.01. Results: The prevalence of AD in the EpiChron Cohort was 3.83%. The most frequently found comorbidities were respiratory, cardio-metabolic, cardiovascular, and mental health disorders. Comorbidities were combined into 17 disease patterns (15 in men and 11 in women), with some sex and age specificities. An infectious respiratory pattern was the most consistently described pattern across all ages and sexes, followed by a cardiometabolic pattern that appeared in patients over 18 years of age. Conclusions: Our study revealed the presence of different clinically meaningful comorbidity patterns in patients with AD. Our results can help to identify which comorbidities deserve special attention in these types of patients and to better understand the physio-pathological mechanisms underlying the disease associations identified. Further studies are encouraged to validate the results obtained in different clinical settings and populations.

**Keywords:** atopic dermatitis; comorbidities; patterns; network analysis

#### **1. Introduction**

Atopic dermatitis (AD) is a chronic inflammatory skin disease of multifactorial etiology characterized by dry skin, itching, erythema, inflammation and eczema formation [1]. It is estimated to affect 15–30% of children and 2–10% of adults [2]. Its prevalence has increased worldwide in recent years due to lifestyle and environmental changes, varying widely according to age and geographical area and being more prevalent in developed countries [2,3].

Advances in the understanding of the etiopathogenesis of AD suggest that its genesis is due to the interaction of several factors that act together to produce the onset and chronification of the disease. An alteration of the barrier function of the skin stands out, behind which there are underlying immune mechanisms, as well as genetic and environmental factors. Structural and functional abnormalities of the epidermis, together

**Citation:** Almenara-Blasco, M.; Carmona-Pírez, J.; Gracia-Cazaña, T.; Poblador-Plou, B.; Pérez-Gilaberte, J.B.; Navarro-Bielsa, A.; Gimeno-Miguel, A.; Prados-Torres, A.; Gilaberte, Y. Comorbidity Patterns in Patients with Atopic Dermatitis Using Network Analysis in the EpiChron Study. *J. Clin. Med.* **2022**, *11*, 6413. https://doi.org/10.3390/jcm11216413

Academic Editor: Stamatis Gregoriou

Received: 18 September 2022 Accepted: 26 October 2022 Published: 29 October 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

with skin inflammation due to an altered immune response, are the cornerstones of the pathogenesis of AD [4,5].

AD is considered as the cutaneous manifestation of a systemic disorder that also gives rise to other pathologies, such as asthma, allergic rhinoconjunctivitis, etc. Some patients with AD have elevated blood levels of IgE and eosinophils. These levels are related to the severity of AD, and for this reason lower values are shown on debut [4].

The immunological mechanisms involved are being investigated; Those proposed to date include antigen-presenting cutaneous dendritic cells in the pathogenesis, and also the loss of immunosuppressive capacity of CD4 +, CD25+ Treg cells [4].

As occurs in other diseases of the atopy spectrum, the predominance of Th2 cells over Th1 generates an immunological imbalance that aggravates the pathogenesis of AD, increases IgE and activates interleukins.

On the other hand, the importance of the integrity of the skin barrier should be highlighted; in recent decades, its dysfunction has been determined to be essential in the pathogenesis of AD. The structure of the skin barrier is complex. The most superficial layer of the epidermis is the stratum corneum, it is made up of proteins (filaggrin, loricin, involucrin); and by a layer of lipids composed of long-chain ceramides as the main component. The stratum corneum protects against environmental stimuli such as allergens, irritants, chemical and physical changes and infections, it also prevents trans-epidermal dehydration [5].

Family genetic studies have shown that AD is a hereditary disease. At the moment, the evidence points to chromosome 1q21 where the locus of the epidermal differentiation complex is located [4,5]. Of all the components of the skin barrier, filaggrin and its mutations are the ones that have shown the greatest association with AD. Filaggrin is a protein that interacts with intermediate filaments, producing their aggregation into macrofibrils. Filaggrin defects can lead to dysfunctions in the skin barrier, resulting in inferior protection against bacteria and allergens [5].

It is currently recognized that this disease is found, especially in children and adolescents, within the context of atopy, an entity that also includes asthma, rhino-conjunctivitis, and a significant predisposition to develop allergies [6,7]. Several studies have previously analyzed the association of AD with the presence of other diseases. A systematic review and meta-analysis by Chester et al. in 2021 [8] concluded that AD patients present an increased risk of mental and autoimmune diseases. The narrative review by Paller et al. in 2018 [9] showed that the global burden of AD is associated with mental illnesses such as depression, anxiety, and suicidal ideation, as a result of lack of sleep, itching, and stigmatization due to their skin lesions, both in children and adults. The cohort study by Mortz et al. in 2015 [10], on the other hand, showed that AD is associated with infections, neuropsychiatric disorders, metabolic syndrome, autoimmune diseases, and cancer, among others. Furthermore, the cross-sectional study by Gilaberte et al. in 2020 [11] showed that 43% of children under 18 years of age with AD in Spain have at least one additional comorbidity. The most frequent comorbidities in this study were asthma, psychosocial disorders, and visual disturbances, whereas asthma, allergic rhinitis, and irritable bowel syndrome showed the greatest strength of association with AD.

A better knowledge of the comorbidities surrounding AD could help us guide the care of these patients from a holistic perspective and better understand the etiopathogenesis of this disease. However, chronic diseases rarely appear in isolation and tend to cluster together in the form of disease patterns, which represent non-random associations among diseases. Their study could allow us to identify profiles of AD patients with specific care needs and specific preventive actions, and could also shed some light on the underlying physio-pathological mechanisms.

In this context, network science is a powerful tool that applies clustering techniques that allow us to exhaustively analyze and visualize the associations between diseases to identify disease patterns [12]. Network analysis has already been applied to study the associations among diseases in patients with specific index conditions with relevant clinical results [13,14]. However, to our knowledge, this research approach has not been applied to the study of AD comorbidity.

This study aims to explore the existence of comorbidity patterns in patients with AD using network analysis and to clinically describe the clusters obtained.

#### **2. Materials and Methods**

We conducted a retrospective, observational study in the EpiChron Cohort, which links socio-demographic and clinical data from all the users of the public health system of the Spanish region of Aragón [15]. This cohort is based on the information registered in the electronic health records (EHRs) and clinical–administrative databases of approximately 98% of the citizens of the region (reference population: 1.3 million people). For this study, we selected all the 50,801 individuals from the cohort diagnosed with AD at some point from 1 January 2010 to 31 December 2018.

The Clinical Research Ethics Committee of Aragón (CEICA) approved this study (Research protocol PI20/633) and waived the requirement to obtain informed consent from patients given the epidemiological nature of the project and the use of anonymized data.

For all patients, we studied sex, age interval (0–2, 3–10, 11–17, 18–65, and >65 years), and all chronic diseases registered in their EHRs. Diagnoses were initially coded using the International Classification of Primary Care, First Edition (ICPC–1), or the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD–9–CM). Subsequently, using the open-source algorithm Chronic Condition Indicator (CCI) [16], each ICD9 code was classified as either chronic or not. The software defines "chronic" as diseases with a duration equal to or greater than 12 months and meeting at least one of the following criteria: (a) require continuous care, that have a high risk of recurrence, and/or that continue to have implications for the management of the patient; (b) imply limitations on self-care, social interactions, and independent living. Once selected, those chronic diagnoses were grouped in 153 clinical categories through the Clinical Classifications Software (CCS) [17] based on the clinical, therapeutic and diagnostic similarities of the diseases.

First, a descriptive analysis of the demographic characteristics of the study population was performed. We summarized the results as proportions for categorical variables and as means and standard deviations for continuous variables.

Then, we performed a network analysis to study the associations between comorbidities of AD. We stratified the population by sex and age interval and built a network for each stratum, with ten networks in total. To increase the clinical interest of the study and to facilitate the interpretation of the results, only diseases with a prevalence > 1% were included in the analysis.

In a disease network, a node represents a disease, and an edge means a statistically significant correlation between a specific pair of conditions. We calculated the tetrachoric correlations of each pair of comorbidities to analyze the weight of the association between them [18]. We used a cut-off point for statistical significance of *p*-value < 0.01 to correct the family-wise error rate due to multiple comparisons [13,14].

Then, we used the network's modularity to search for clusters of diseases within each network based on the Louvain method [19], as previous disease pattern studies have done [13,14,20]. Modularity analyzes the number of edges in the network, comparing the density of edges inside a group to edges between groups [19]. The Louvain method optimizes the modularity in an iterated process, detecting communities or clusters of diseases. Community detection methods, such as the Louvain or Leiden algorithms, among others, allow the network's structure to determine the number and size of the cluster obtained [21,22] based on the density of edges and their weight (measured by the tetrachoric correlation) and not by the researcher.

Once we obtained the patterns of diseases for each age and sex stratum, all clinicians named the patterns by consensus. This last step was performed considering the prevalence and clinical relevance of the diseases, and the weight of the tetrachoric correlations, in line with the names already given in the literature.

We performed all the analysis in RStudio software (version 1.4.1106, Rstudio, Boston, MA, USA) and GEPHI software (version 0.9.2).

#### **3. Results**

#### *3.1. Characteristics of the Population*

We analyzed a population of 50,801 patients with AD (46.3% men). The demographic characteristics are shown in Table 1. The overall prevalence of AD in Aragon was 3.83%.

**Table 1.** Demographic characteristics of patients with AD.


The most prevalent diseases found in patients with AD were respiratory (i.e., upper respiratory infections, asthma, and rhinitis), cardio-metabolic (i.e., hypertension, dyslipidemia and obesity), cardiovascular (i.e., cardiac dysrhythmia and coagulation disorders), and mental health diseases (i.e., anxiety and mood disorders). Diseases were combined into seventeen patterns with some sex and age specificities, which are summarized below. The complete output of the analysis is available as Supplementary Material in which we detailed the complete pattern analysis.

#### *3.2. Comorbidity Patterns in Men*

We identified fifteen patterns in men, classified as upper respiratory infections, respiratory, otorhinolaryngological (ORL), upper respiratory infections-ORL, congenital anomaliesmental, respiratory-allergic, sensitive-digestive, sensitive-metabolic, headache-mental, cardiometabolic, mental, cardiovascular, dyslipidemic, and geriatric. Their composition, disease prevalence, and correlation between conditions are described below and in Figure 1.

In children aged 0–2, we found three patterns. The *upper respiratory infections* pattern included diseases such as upper respiratory infections, which was the most prevalent condition in this network, esophageal disorders, and congenital anomalies. We found a *respiratory* pattern that had asthma and other upper respiratory diseases, among others. An *ORL* pattern was also described with otitis as its most prevalent disease.

Four patterns in boys aged 3–10 years were identified. One that combined most of the diseases from upper respiratory infections and ORL diseases in children aged 0–2. We found a *respiratory-allergic* pattern that included asthma, other upper respiratory diseases, and rhinitis. The correlation between these two last conditions was almost perfect, with a strength of the correlation of 0.99 out of 1. The *congenital anomalies-mental* pattern associated congenital anomalies, miscellaneous mental health disorders, and other skin inflammatory conditions. The last pattern found was a *sensitive-digestive* pattern which included blindness as the most prevalent condition.

Four other patterns were identified in boys aged 11–17. The *upper respiratory infections-ORL* and *respiratory-allergic* were similar to the previous in children aged 3–10. The *headachemental* pattern included headache as the most prevalent condition; it also included anxiety and miscellaneous mental health disorders, among other diseases. The *sensitive-metabolic* pattern had diseases such as blindness and vision defects, thyroid disorders, obesity, and other nutritional/endocrine/metabolic diseases.

**Figure 1.** Comorbidity patterns in the networks of men with AD based on age. The diameter of each node and the label size are proportional to the disease prevalence. The width of each link is proportional to the correlation between disease. The colors of the nodes correspond to different patterns.

In men aged 18–65 years, three patterns were identified. One pattern combined most of the diseases from the *upper respiratory-ORL* pattern and the *respiratory-allergic* pattern from boys aged 11–17. A *cardiometabolic* pattern was the cluster with more diseases included, highlighting hypertension, obesity, other nutritional/endocrine/metabolic disorders, dyslipidemia, other inflammatory conditions of the skin, and diabetes. We also identified a *mental* pattern which included screening and history of mental health codes, anxiety, mood disorders, substance-related, and alcohol-related disorders.

In men aged 66 and older, five patterns were detected. The *upper-respiratory-allergic* pattern was similar to that found in men aged 18–65, but included other highly prevalent diseases such as osteoarthritis or cataracts. We found a *cardiometabolic* pattern with hypertension, diabetes, other nutritional/endocrine/metabolic disorders, obesity, COPD, and neoplasms as the most prevalent diseases. We identified a *cardiovascular* pattern that included cardiac dysrhythmias as its most prevalent condition. A *geriatric* pattern, with hyperplasia of the prostate, urinary incontinence, or dementia, among other diseases, was also identified. The last and less specific pattern described included dyslipidemia as its most prevalent disease.

#### *3.3. Comorbidity Patterns in Women*

We identified eleven patterns in women, which were referred as upper respiratory infections, ORL, upper respiratory infections-ORL, respiratory-allergic, sensitive-digestive, menstrual-dysphoric-metabolic, sensitive, upper respiratory-allergic, cardiometabolic, car-

diovascular, and geriatric. Their composition, disease prevalence, and correlation between conditions are described below and in Figure 2.

**Figure 2.** Comorbidity patterns in the networks of women with AD based on age. The diameter of each node and the label size are proportional to the disease prevalence. The width of each link is proportional to the correlation between disease. The colors of the nodes correspond to different patterns.

In girls aged 0–2, the two patterns found were similar to those found in boys aged 0–2 years: *upper respiratory infections* and *ORL*.

We described three patterns in girls aged 3–10, again very similar to the clusters found in boys at the same age: an *upper respiratory infections-ORL* pattern, a *respiratory-allergic* pattern, and a *sensitive-digestive* pattern.

Four other patterns were identified in girls aged 11–17: an *upper respiratory infections-ORL* pattern; a *respiratory-allergic* pattern; a *menstrual dysphoric-metabolic* pattern that included menstrual disorders as the most prevalent condition, but also anxiety, miscellaneous mental health disorders, obesity, thyroid and other metabolic disorders; and finally, the *sensitive* pattern, which included blindness as its most prevalent disease.

In women aged 18–65 years, we found three patterns. The *upper respiratory-allergic* pattern included respiratory diseases such as rhinitis, other upper respiratory diseases, and asthma, but also menstrual disorders and anxiety, among others. A *cardiometabolic* pattern was found, that was mainly characterized by thyroid diseases, obesity, and hypertension but also mood disorders. We also identified *cardiovascular* a pattern that only included three conditions: coagulation and hemorrhagic disorders, cardiac dysrhythmias, and female infertility.

In women aged 66 and older, three patterns were detected. The *upper respiratory-allergic* pattern also included osteoarthritis, osteoporosis, thyroid, and mood disorders, among other diseases. The *cardiometabolic* pattern was found, characterized by hypertension, but also including diabetes mellitus, heart failure, COPD and obstructive sleep apnea. Finally, a *geriatric* pattern that included urinary incontinence, dementia, neoplasms, and chronic skin ulcer as the most prevalent diseases was found.

#### **4. Discussion**

This study explored the comorbidity patterns of AD through the analysis and visualization of the existing disease networks. Different clusters defined as upper respiratory infections with ORL diseases, respiratory disorders with allergic conditions, and cardiovascular diseases with metabolic disorders, among others, were identified depending on age and gender. These epidemiological findings can be helpful to guide AD patients in the primary, secondary, or even tertiary prevention of their comorbidities and understand their physio-pathological mechanisms.

The present investigation shows the infectious respiratory pattern as the most consistently described pattern across all age groups and sexes. Its main component was respiratory infectious diseases, but its weight decreased in older groups in favor of diseases with an allergic component. Asthma or allergic rhinitis are some of the disorders that have been added. In the youngest groups, this allergic component has a distinct pattern by itself. In the case of children aged 0–2 years, respiratory infections are associated with genital, esophageal, or other malformations, although this is not the case in girls.

The higher incidence of infections in patients with AD has been widely described. Dysfunction of the epithelial barrier, colonization of the skin by *Staphylococcus aureus*, and the use of immunosuppressive drugs are some of their causes. In this context, the Swiss BAMSE cohort revealed a higher incidence of pneumonia, otitis media and antibiotics use in AD patients aged 0–2 years [23]. This fact is consistent with the patterns described. Although patients with AD are colonized by *S. aureus* in up to 70% of the cases and are more likely than the general population to suffer impetigo, herpetic eczema or molluscum contagiosum [24], in our analysis skin infections did not play a relevant role or were associated with extracutaneous infections.

Cardiometabolic diseases have been associated with AD in various epidemiological studies, although this association is less clear than in other diseases such as psoriasis [8,25,26]. Multifactorial etiology has been used to explain this association: insomnia, obesity, diabetes and smoking, among other variables [27,28]. Our study found a pattern of cardiometabolic comorbidities that included hypertension, obesity, and mood and thyroid disorders, among others. This pattern was common in men and women over 18 years of age, although there were differences, including mood and endocrine disorders occurring more frequently in women. As for chronic obstructive pulmonary disease, it was more frequent in the group older than 65 years old for both sexes. The existence of this pattern confirms that patients with AD present comorbidities that are cardiovascular risk factors and that tend to be associated throughout life.

Anxiety, insomnia, and mood disorders, among other mental health problems, are comorbidities with a higher incidence in patients with AD. Recent studies have shown that the earlier the disease appears, the greater the risk of suffering from psychiatric comorbidities is [11,24,29]. We found a pattern of mental comorbidities that appeared in boys older than two years and is maintained up to 65 years of age. In the case of girls, this pattern appears intermingled with a menstrual disease first and metabolic comorbidities that we call menstrual-dysphoric-metabolic pattern. The different ways of interacting between the sexes with the environment and the interpersonal relationships they establish at an early age could explain this phenomenon [30].

In patients over 65 years of age, a geriatric pattern was found in both sexes. This pattern grouped diseases such as urinary incontinence, Parkinson's disease, dementia, skin ulcers and neoplasms, among others. The association of these diseases, typical of physiological aging, with AD is complex. An increase in neoplasms has been described in patients with AD, with lymphomas being the most strongly associated [9]. Regarding the

rest of diseases of this pattern, to our knowledge no studies support a higher incidence of these diseases in patients with AD.

Regarding the limitations of this study, the fact that the clinical information obtained in the EHRs was not originally designed for research could create over- and under-diagnosis of some chronic disorders. Another limitation is the cross-sectional retrospective nature of the study, which does not allow us to know the longitudinal characteristics of the population. Additionally, we have to consider the lack of some variables that could help us explain the results obtained, such as lifestyle information, socioeconomic factors, information on functional status, and analytical variables, among others.

One of the principal strengths of our research is that it was conducted on a populationbased cohort, including 98% of the reference population. Moreover, data in the EpiChron Cohort undergo continuous quality control checkups that ensure their accuracy and reliability for research purposes. Another important strength is the innovative method applied to understand comorbidities in AD. Network analysis studies the interrelations between diseases and how patterns emerge from them. This paper shows the potentiality of applying this method to study and visualize the comorbidities of AD and achieve a more holistic understanding of these patients. In this sense, it is also important to highlight that this study exhaustively analyzed all chronic diseases obtained from the patient's EHRs created by health professionals, and not just the most relevant, prevalent or self-reported diseases.

#### **5. Conclusions**

We identified similar disease patterns in men and women with AD, with the number and complexity of such patterns increasing with age. This is the first study to analyze the comorbidity patterns of AD patients, and our results can help to guide caregivers of AD patients in the prevention of their comorbidities and to understand the physiopathological mechanisms underlying the comorbidity patterns identified. This study opens an innovative approach to analyze and help AD patients, although further studies are needed to validate the results obtained in different clinical settings and populations.

**Supplementary Materials:** The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/jcm11216413/s1.

**Author Contributions:** Conceptualization, Y.G., J.C.-P., M.A.-B., A.P.-T. and A.G.-M.; methodology, J.C.-P.; formal analysis, J.C.-P.; data curation, B.P.-P.; writing—original draft preparation, M.A.-B. and J.C.-P.; writing—review and editing, M.A.-B., J.C.-P., T.G.-C., B.P.-P., J.B.P.-G., A.N.-B., A.G.-M., A.P.-T. and Y.G.; visualization, J.C.-P.; supervision, A.G.-M., A.P.-T. and Y.G.; funding acquisition, A.P.-T., J.C.-P. and Y.G. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was co-funded by Instituto de Investigación Sanitaria Aragón and Instituto de Salud Carlos III [Río Hortega Program, grant number CM19/00164; RICAPPS, grant number RD21/0016/0019], Gobierno de Aragón (Grant numbers B01\_20R and DECRETO-LEY 3/2020, de 3 de junio, del Gobierno de Aragón) and the European Union (NextGenerationEU). The authors want to thank Sanofi-Genzyme for the financial support for the publication of this study.

**Institutional Review Board Statement:** Ethical approval for this study was obtained from the Clinical Research Ethics Committee of Aragón (CEICA) that approved the research protocol for this study (PI20/633).

**Data Availability Statement:** The data used in this study cannot be publicly shared, because of restrictions imposed by the Aragon Health Sciences Institute (IACS) and asserted by the Clinical Research Ethics Committee of Aragon (CEICA, ceica@aragon.es). The authors can establish future collaborations with other groups based on the same data. Potential collaborations should be addressed to the Principal Investigator of the EpiChron Group, Alexandra Prados-Torres, sprados.iacs@aragon.es.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

#### **References**

