Introduction
Interleukins 4 (IL-4) and 13 (IL-13) are key drivers of type 2 immune responses that contribute to the pathogenesis of atopic dermatitis (AD) [
1]. Dupilumab is a fully human monoclonal antibody that blocks the shared receptor component for interleukin (IL)-4 and IL-13 and modulates the downstream pathways regulated by these cytokines [
2,
3]. Dupilumab is approved in the USA (Dupixent) for patients aged ≥ 6 years with moderate-to-severe AD not adequately controlled with topical therapies, as add-on maintenance treatment in patients ≥ 12 years with moderate-to-severe asthma with an eosinophilic phenotype or oral corticosteroid-dependent asthma, and as add-on maintenance treatment in adults with inadequately controlled chronic rhinosinusitis with nasal polyposis [
4]. The efficacy of dupilumab for the treatment of AD has been demonstrated in several clinical trials in patients with moderate-to-severe AD. Those trials showed that dupilumab not only reduced the clinical signs and symptoms of AD and improved patient-reported outcomes relative to placebo [
5‐
11] but also resulted in clinically relevant effects on disease biomarkers and reversal of AD-associated epidermal abnormalities [
12,
13]. However, patients in clinical trials may not necessarily reflect the patient population likely to be treated in routine clinical practice. Thus, it is important to evaluate the real-world effectiveness of treatment and to determine whether the effects observed in the clinical setting support those reported in clinical trials. Several real-world studies have reported on the effectiveness of dupilumab [
14‐
27], but these studies either evaluated small populations or were from a limited number of study sites.
Electronic medical records (EMR) are increasingly being utilized as a source of real-world data for large populations. The purpose of the current analysis is to evaluate the real-world effectiveness of dupilumab in adults with moderate-to-severe AD over a time period comparable to that of dupilumab clinical trials, using a large, dermatology-specific structured EMR database and clinically relevant outcomes of clinician assessment of global AD severity, percentage body surface area affected, and patient-reported itch severity.
Methods
Data source
Data for this retrospective observational study were derived from Modernizing Medicine’s Electronic Medical Assistant (EMA) dermatology-specific EMR database. This database is the most widely used dermatology-specific EMR platform in the USA, containing structured, real-world data from over 30% of US dermatologists. All patient-level data were fully anonymized to ensure confidentiality and compliance with the Health Insurance Portability and Accountability Act of 1996 (HIPAA).
Use of dermatology-specific EMR enables extraction of AD assessments, where available, including severity using the six-point Investigator Global Assessment scale (IGA; 0 = clear to 5 = very severe), and itch severity, which was assessed using a numerical rating scale (NRS) based on the question “How intense is your itching, 0 being no itch, 10 being the most severe itch possible?” Information on percent of body surface area (BSA) affected is also available in the database. These measures were used for both identification of cohorts and for evaluation of effectiveness.
Study populations
From the EMR database, patients were identified who had a diagnosis of AD (ICD-10 codes L20.0, L20.81, L20.82, L20.83, L20.84, L20.89, L20.9) and received ≥ 1 dupilumab prescription(s) between 1 April 2017 and 31 January 2019 (representing the initial approval in adults); the date of the first dupilumab prescription was defined as the index date, and patients were required to be ≥ 18 years of age at index. The study period was from 1 April 2016 to 31 May 2019 to capture outcomes assessment at 12 months pre-index (baseline) and 4 months post-index, defined as 120–149 days after the initial prescription. For inclusion in the analysis, patients were required to have IGA, itch severity NRS, and BSA recorded in the EMR within 3 months pre-index and any time during the 120–149-day post-index period. Inclusion criteria were based on availability of the outcomes independent of whether the patient discontinued treatment, since they had to have at least one treatment with dupilumab, i.e., treatment was initiated during the specified time frame. From this population, three cohorts were identified for analysis based on the most recently recorded IGA, itch severity NRS, and BSA scores within the 3-month pre-index period: (1) patients with moderate-to-severe AD defined as IGA ≥ 3, (2) patients with moderate-to-severe itch, defined as NRS score ≥ 3 [
28], and (3) patients with BSA ≥ 10%.
Effectiveness
Effectiveness was evaluated based on changes from baseline at 4 months post-index on the IGA, itch severity NRS, and BSA scores in the three populations, respectively. The proportion of patients with post-index IGA 0/1 was determined, as was the proportion of patients with IGA scores ≥ 1 and ≥ 2 points lower than baseline; a change ≥ 1 point has been considered to be clinically relevant [
29,
30]. Similarly, for the itch severity NRS, in addition to the mean change in score, the proportion of patients who improved by a ≥ 3-point reduction in score from baseline was determined, as were the proportions of patients with post-index scores stratified by itch severity strata of 0–3 (mild), 4–6 (moderate), and 7–10 (severe) [
28]; severity strata were used rather than NRS scores because data cells with ≤ 5 patients in the Modernizing Medicine’s EMR database are masked to ensure patient privacy. A ≥ 3-point reduction can be considered clinically meaningful based on empirically derived thresholds from a similar NRS [
31].
Effectiveness in the cohorts was further evaluated among patients stratified by treatment history during the baseline period, and sex and age. Two distinct treatment history strata were defined: one stratum consisted of patients with either a history of topical therapy only (topical corticosteroids, topical calcineurin inhibitors, and phosphodiesterase 4 inhibitor) or no treatment, and the other stratum consisted of those who had history of use of any systemic therapy (systemic corticosteroids, immunosuppressants, and phototherapy). Age strata were 18–34 years, 35–54 years, and ≥ 55 years.
In the BSA cohort, patients were stratified by quartiles of BSA affected (10%–25%, 26%–50%, 51%–75%, and 76%–100%).
Statistical analysis
Results were analyzed descriptively, with mean and standard deviation (SD) generated as measures for continuous variables, and count (frequency) and percentages generated for categorical variables. Statistical comparisons of changes in severity of AD and itch outcomes from pre- and post-index were performed using two-sided paired t-tests or Wilcoxon signed-rank tests at 0.05 significance level. The mean percent change in itch severity NRS score was derived from the percent change calculated across the individual patients. For BSA quartiles, least-squares (LS) mean change from baseline and 95% confidence intervals (95% CI) were estimated using analysis of covariance (ANCOVA) adjusted for age, gender, baseline IGA, and binary baseline therapies of interest (topical corticosteroids; topical calcineurin inhibitors; PDE-4 inhibitors; systemic steroids; immunosuppressants); confidence intervals that do not include 0 would map to p < 0.05.
All analyses were conducted using SAS version 9.4 (SAS Institute, Cary, NC).
Discussion
This study, through its use of a dermatology-specific EMR database to identify cohorts of patients defined as having moderate-to-severe AD based on established severity thresholds, adds to the expanding body of evidence supporting the effectiveness of dupilumab in the real-world clinical setting. Evaluation of dupilumab treatment outcomes was from both the clinician’s perspective using a global severity assessment (IGA) and an objective assessment of BSA, and the patient’s perspective based on the hallmark symptom of itch. Furthermore, the analyses included stratification by treatment history, with evaluation conducted at a time point that allowed for comparison with clinical trial data. On all outcomes, the results showed that the majority of patients achieved benefits after initiation of dupilumab therapy that were statistically significant, clinically meaningful, and consistent with improvements in AD observed at similar time points in the dupilumab clinical trials [
5,
7,
9‐
11] and other real-world studies [
14‐
24,
26].
IGA is a relevant efficacy endpoint in clinical trials as it considers global physician assessment of disease and may be less cumbersome than multi-item measures [
32]. In the IGA cohort, 59.1–66.2% of patients had a ≥ 2-point reduction in score, which resulted in substantial proportions of the patients achieving an IGA score of clear or minimal AD (39.1–47.6%) regardless of strata and 42.8% overall. These proportions may be considered comparable to what was observed in phase 3 clinical trials, which used a combined endpoint of IGA 0/1 plus a ≥ 2-point reduction in score, and reported that 36–40% of patients met this endpoint after 16 weeks of dupilumab treatment [
7,
9,
10]. It should be noted that the IGA in dupilumab clinical trials was based on a 0–4-point scale rather than the 0–5-point scale in this study. However, the additional discrimination associated with the wider scale was related to expansion of scores representing severe (4) and very severe (5), and thus patients at higher severity required a greater point reduction to achieve IGA 0/1 in this analysis compared with clinical trials. The IGA has been reported in few real-world studies, with proportions of patients who achieved a clear/minimal score ranging from 38% to 60% within 3–4 months after dupilumab initiation [
18,
33,
34], although the study populations were small.
Both the absolute and percent reduction in itch severity were substantial, and 68.1–72.5% of patients across age, sex, and treatment history strata had a clinically meaningful improvement of ≥ 3 points. These effects of dupilumab on itch are consistent with what has been reported after 16 weeks of dupilumab treatment in phase 3 clinical trials with regard to magnitude of reduction (44.3–56.2%) and clinical relevance (47–66% of patients with reductions ≥ 3 points) [
7,
9,
10]. Importantly, the observed effects of dupilumab in reducing AD severity and itch were independent of sex and age, and appeared to be comparable regardless of treatment history.
The BSA affected by AD was also significantly reduced in the post-index period after initiation of dupilumab therapy. Notably, patients who had a higher percentage of BSA affected at baseline were characterized by greater absolute reductions. However, it should also be noted that the change in the lowest quartile may reflect a floor effect with regard to baseline, since these patients may have had less room for improvement. As with the other outcomes, improvements in BSA were similar regardless of demographic characteristics and treatment history.
Interpretation of these results should consider several study limitations, including that certain data elements, such as those evaluated in this study, are optional fields in EMRs and may not be populated at each visit for all patients. Consequently, only a small proportion of patients had these outcomes recorded in the EMR prior to and after the initiation of dupilumab therapy. In this regard, the absence of multi-item measures such as the Eczema Area and Severity Index (EASI) or Scoring of Atopic Dermatitis (SCORAD) may also be considered a limitation, but since these measures are time consuming, they are infrequently used in daily clinical practice and are even less likely to be captured in EMR than the ones reported in this study. These limitations further suggest that more routine recording in the EMR of AD severity from the clinician and patient perspectives should be encouraged as part of regular clinical visits. Assessment of itch severity in dupilumab clinical trials was the average of daily scores for a week, whereas a single time point was used in the current analysis. Treatment exposure was based on prescription orders, and whether the prescription was actually filled and appropriately used by the patient could not be confirmed. However, the use of an intention-to-treat approach provided a conservative estimate of real-world dupilumab effectiveness. While information on treatment history was available, concomitant medication use throughout the study period was not captured. Nevertheless, this analysis showed that substantial and clinically relevant improvements from baseline were observed at the post-index assessment after initiation of dupilumab.
Acknowledgements
Disclosures
Lawrence F. Eichenfield has received honoraria for consulting services from Almirall, Celgene, Dermira, Dermavant, Eli Lilly and Company, Forté Biosciences, Galderma, Incyte, Glenmark, Leo Pharmaceuticals, Otsuka, Novartis, Pfizer, Regeneron Pharmaceuticals, Inc., Sanofi Genzyme, and Valeant/Ortho Dermatologics; and study support (to institution) from AbbVie, Dermira, Dermavant, Eli Lilly and Company, Galderma, Incyte, Pfizer, Regeneron Pharmaceuticals, Inc., Sanofi Genzyme, and Ortho Dermatologics. April Armstrong has received consultancy honoraria from AbbVie, Amgen, Eli Lilly, Janssen, Merck, Novartis, and Pfizer; research funding from AbbVie, Eli Lilly and Janssen; and speaker honoraria from AbbVie. Emma Guttman-Yassky has acted as a consultant for and received grants/honoraria from AbbVie, Anacor, Celgene, Celsus Therapeutics, Dermira, Galderma, Glenmark, Janssen Biotech, LEO Pharmaceuticals MedImmune, Novartis, Pfizer, Regeneron Pharmaceuticals, Inc., Sanofi, Stiefel/GlaxoSmithKline, Vitae, Mitsubishi Tanabe, Eli Lilly, Asana Biosciences, Kiowa Kirin; has acted as an investigator for Celgene, Glenmark, Leo Pharmaceuticals, MedImmune, Regeneron Pharmaceuticals, Inc., Eli Lilly; and has participated in advisory boards for Celgene, Celsus Therapeutics, Dermira, Galderma, Glenmark, MedImmune, Novartis, Pfizer, Regeneron Pharmaceuticals, Inc., Sanofi, Stiefel/GlaxoSmithKline, Vitae and Asana Biosciences. Peter A. Lio reports research grants/funding from Regeneron/Sanofi Genzyme, and Abbvie; is on the speaker's bureau for Regeneron/Sanofi Genzyme, Pfizer, and Galderma; reports consulting/advisory boards for UCB, Dermavant, Regeneron/Sanofi Genzyme, Dermira, Pfizer, LEO Pharmaceuticals, AbbVie, Kiniksa, Eli Lilly, Menlo Therapeutics, Galderma, IntraDerm, Exeltis, Realm Therapeutics. Chi-Chang Chen, Dionne M. Hines, and Catherine B. McGuiness are employees of IQVIA, which received funding from Regeneron and Sanofi to conduct the study. Debra Sierka is an employee and stockholder of Sanofi. Sohini Ganguli and Usha G. Mallya were employees and stockholders of Sanofi at the time of study. Dimittri Delevry is an employee and stockholder of Regeneron.