Prescribing Patterns of Dupilumab for Atopic Dermatitis in Adults: Retrospective, Observational Cohort Study

Introduction

Atopic dermatitis (AD) is a chronic inflammatory disease that affects US children and adults, with a reported prevalence of 10%-13% and 7%, respectively [1-4]. Higher disease severity is associated with lower quality of life, worsened mental status, and higher use of health care resources [1,5]. This includes emergency department visits and hospitalizations as well as increased pharmaceutical and outpatient costs [6-8]. Consequently, effectively treating patients with AD has substantial clinical and economic implications.

Assessing the severity of AD is determined by the number of sites involved (eg, head and neck, upper extremities, or trunk and lower extremities), the lesion characteristics (eg, erythema, edema or papulation, and lichenification), as well as reported symptoms. Patients who have 10%-29% bodily involvement with notable signs of inflammation are classified as having moderate disease, and those with greater or equal to 30% bodily involvement are considered to have severe disease. This classification is important in the diagnosis and treatment of AD, thereby affecting therapeutic outcomes [9]. Treatment of AD with a systemic immunomodulating agent is indicated when the disease is considered moderate to severe [10]. A variety of systemic agents, including cyclosporine, azathioprine, methotrexate, mycophenolate mofetil, and systemic steroids, are used in practice without strict guidelines or recommendations to guide treatment choices [10,11]. In March 2017, dupilumab became the first biologic drug approved for the treatment of AD in adults, and in May 2020, it gained approval for use in children aged 6 years and older. Dupilumab blocks the interleukin (IL)-4 alpha receptor, inhibiting IL-4 and IL-13 signaling and preventing the release of type 2 cytokines that promote inflammation in AD [12]. Compared with the aforementioned systemic immunomodulating agents, dupilumab may be more effective as a long-term maintenance therapy and has the advantage of an overall improved side-effect profile, with no required drug-specific laboratory monitoring [11]. However, access to this immunomodulator may be limited by its novelty and cost (depending upon the dose, up to US $59,000/year for patients without insurance) [13].The disease burden of AD disproportionately affects non-Hispanic Black patients; the source of this disparity is multifactorial. Although the specific gene-environment interactions in the pathophysiology of AD are unknown, many factors—such as differences in environmental pollution, contact with tobacco smoke, hygiene practices, access to health care, diet, and exposure to disease—likely play a role [4,14,15]. Loss of function mutations in the filaggrin gene is linked with an increased risk of developing AD, oftentimes leading to persistent disease. Filaggrin loss-of-function mutations are less common in Black patients compared to White patients, yet Black patients are still more likely to experience persistent disease [16,17]. In addition, Black patients have lower skin ceramide/cholesterol ratios, attenuated T helper 1 and T helper 17 immunophenotypes, and higher serum immunoglobulin (Ig) E levels, which predisposes to skin breakdown, dysregulated immunity, and increased inflammation [18,19]. Despite higher AD disease severity and increased health care needs, Black patients are less likely to receive outpatient dermatologic care [3,20-22]. A prior investigation reporting on race- and ethnicity-related disparities in the treatment of AD found that Black patients had statistically significant lower odds of receiving dupilumab compared to White patients [23]. Previous research has also shown that Black patients with psoriasis are less likely than White patients to receive biologic treatment, independent of demographic or socioeconomic factors and comorbidities [24,25].

Data related to the prescription patterns of dupilumab for AD are needed to inform health equity and decision-making in everyday practice.

Methods

Ethical Considerations

The Colorado Multiple Institutional Review Board determined that this research did not involve human subjects, and therefore, was exempt from ethics approval.

Data Source

We performed a retrospective, observational cohort study of adult patients treated for AD, using electronic data from the University of Colorado Anschutz Medical Campus and its affiliates (hospital wards and outpatient clinics) via Health Data Compass (HDC), an electronic health data warehouse [26]. Records of patients who visited any of the institution’s facilities from January 1, 2010, to March 28, 2021, were pulled from HDC and remained in the format of the Observational Health Data Sciences and Informatics (OHDSI) Observational Medical Outcomes Partnership common data model. Data extracted included demographics, prescription history, diagnosis history, and visit details.

Study Design and Study Population

The outcome of interest was the prevalence of dupilumab prescribed for AD, stratified by patient characteristics. The study cohort, consisting of individuals with specific AD diagnoses and various dupilumab prescription types, was developed using the OHDSI Atlas cohort and concept set tools (Table 1). All study participants were between the ages of 18 and 85 years as of March 28, 2017, with a diagnosis of AD, as defined by at least two encounter diagnoses of AD (Figure 1). The drug of interest was any prescription order of dupilumab (200 mg or 300 mg syringe or pen). Records included were those of ordered medications. Information regarding fulfilment for this study was not used, nor did we include orders external to our institution. Included dupilumab prescriptions were required to have a start date on or after March 28, 2017, and to have occurred following a diagnosis of AD. Prescriber data, including their medical credential and practice setting, were not analyzed.

‎

Statistical Analysis

We assessed whether the characteristics of patients with AD who received dupilumab were different from those who did not. Each patient characteristic was gathered based on AD-associated billing and nonbilling codes. These codes were based on the International Classification of Diseases, Tenth Revision and are listed in Table 1. The characteristics were then assessed using a univariate logistic regression, with the binary outcome of receiving dupilumab or not receiving dupilumab (Table 2). The resulting P values associated with multilevel categorical characteristics (eg, race and diagnosis) were corrected for multiple testing using the False Discovery Rate (Benjamini-Hochberg correction) method. Reference levels included White race, non-Hispanic ethnicity, female sex, and AD as the first eligible diagnosis.

Results

Summary statistics gathered based on the International Classification of Diseases, Tenth Revision billing codes are provided in Table 2. There were 249 dupilumab prescriptions among 6421 patients. Our cohort had a mean age of 53.2 (SD 17.9) years and was composed of mostly non-Hispanic (n=5491, 89%), White (n=4912, 79.3%), and female (n=3532, 57%) patients. The most common recent diagnosis was a general diagnosis of atopic dermatitis (n=3263, 50.8%), followed by vesicular eczema (n=1148, 17.9%), nummular eczema (n=1126, 17.5%), flexural eczema (n=763, 11.9%), and lastly, atopic neurodermatitis (n=12, 1.9%). Among those who received dupilumab, the mean age was 51.4 (SD 16.7) years and 58.6% (n=146) were female. Every patient who received a prescription had either multiple prescriptions, prescriptions for quantities greater than 1, or received refills. The majority of patients were White (n=179, 71.9%), followed by patients identifying as Black or African American (n=34, 13.7%) and “Other” races (n=13, 5.2%). Among the most common recent diagnoses, 85.8% (n=213) were general diagnoses of AD, and the remaining were diagnoses of atopic neurodermatitis (n=14, 5.6%), flexural eczema (n=5, 2.2%), nummular eczema (n=6, 2.2%), or vesicular eczema (n=11, 4.1%).

We assessed whether the proportion of patients who received dupilumab was different based on patient characteristics. Each patient characteristic (Table 3) was assessed using logistic regression with the binary outcome of receiving dupilumab or not receiving it. The P values associated with multilevel categorical characteristics (eg, race or diagnoses) were corrected for multiple testing using the False Discovery Rate (Benjamini-Hochberg correction) method. Reference levels included white race, non-Hispanic ethnicity, female sex, and AD as the first eligible diagnosis.

In our cohort, Black patients were approximately twice as likely to have received dupilumab for AD compared to White patients (odds ratio 2.352, 95% CI 1.58-3.39). Similarly, those diagnosed with atopic neurodermatitis were about twice as likely to have received dupilumab compared to those who were diagnosed with AD (odds ratio 1.87, 95% CI 1.01-3.22). Conversely, those with other eczema diagnoses, including flexural eczema, nummular eczema, or vesicular eczema, were less likely to have received dupilumab compared to those with a most recent diagnosis of atopic neurodermatitis.

Discussion

Principal Findings

We found that patient racial characteristics as well as specific eczema diagnoses were associated with different frequencies of dupilumab prescriptions.

Specifically, our results suggest that patients who reported their race as African American or Black were more likely to have received a dupilumab prescription compared to White patients in our health care system. This result is surprising, as previous research demonstrated less outpatient dermatologic care and fewer biologic prescriptions among the population of Black patients with AD [27,28].

Our study also revealed that patients with a diagnosis of atopic neurodermatitis were twice as likely to be prescribed dupilumab compared to patients with a diagnosis of AD. The term “neurodermatitis” refers to localized, circumscribed patches of lichenified skin that are commonly associated with itching. This term is increasingly being used to describe skin eruptions associated with anxiety [29].

Our study also revealed that patients with specific AD diagnoses, such as flexural eczema, nummular eczema, and vesicular eczema, are less likely to receive dupilumab than those with general AD, without further specification. One possible explanation may be that patients with certain subtypes of AD respond well to more conservative, first-line therapies (ie, topical corticosteroids, emollients, and topical calcineurin inhibitors) [30].

Limitations

Although the study results provide novel insights pertaining to the prescribing practices of dupilumab, there are important limitations. Data capturing patients’ insurance status and type of insurance at the time of visit would have been useful to provide further clarity regarding prescribing relationships; unfortunately, our HDC data warehouse only has the current insurance status and type available from our electronic health record system. Furthermore, the type of health care provider prescribing dupilumab as well as their specific sector of medical employment were not analyzed. Additional studies evaluating the background of the clinicians prescribing dupilumab may provide a better understanding of these unique prescription patterns.

Conclusions

We found that Black patients and patients diagnosed with atopic neurodermatitis were approximately twice as likely to have received dupilumab compared to White patients and those with a diagnosis of AD, respectively. Future studies may consider exploring possible factors contributing to the dupilumab prescription patterns we discovered, such as AD disease severity as a potential contributing factor and the role of insurance type in dupilumab prescribing frequency. Investigating similar trends among other OHDSI network sites might also be beneficial in elucidating the trends noted among our study population and would contribute to the generalizability of the results. Furthermore, there is little existing data regarding the treatment of AD in the pediatric demographic—additional research would allow clinicians to provide evidence-based care to this population subset.