Abstract
Introduction
In clinical trials of patients with atopic dermatitis (AD), conjunctivitis and keratitis occurred more frequently with dupilumab than placebo. Studies using real-world data have also shown a higher incidence with dupilumab but did not use validated algorithms to identify the population and outcomes. The objective of this study was to investigate the incidence of conjunctivitis and keratitis among patients with moderate-to-severe AD treated with dupilumab relative to dupilumab-naïve patients in a real-world setting using validated algorithms.
Methods
A retrospective observational cohort study was conducted in an insurance claims database using validated algorithms for moderate-to-severe AD and ocular outcomes. Initiators of dupilumab were identified and propensity score (PS) matched with dupilumab-naïve patients with moderate-to-severe AD. Incidence rate ratios (IRRs) with 95% confidence intervals (CIs) for conjunctivitis and keratitis during follow-up were estimated using Poisson regression.
Results
Among 13,790 patients in the moderate-to-severe AD study population, 2175 dupilumab initiators and 2189 dupilumab-naïve patients were included in the analysis. Dupilumab was associated with an increased risk of conjunctivitis (IRR = 1.86 [95% CI 1.51–2.30]) and keratitis (IRR = 4.06 [95% CI 1.70–9.68]) compared to patients with moderate-to-severe AD not receiving dupilumab. The cumulative 1-year risk of conjunctivitis was 15.8% and 8.4% in the dupilumab and dupilumab-naïve cohorts, respectively; the cumulative 1-year risk of keratitis was 1.3% and 0.2%, respectively.
Conclusion
Study findings are consistent with safety data from clinical trials and existing literature. However, a few keratitis events were observed, and post hoc analyses suggested residual confounding might be present. The known benefit-risk profile for dupilumab remains unchanged.
Supplementary Information
The online version contains supplementary material available at 10.1007/s13555-025-01367-5.
Keywords: Atopic dermatitis, Adverse events, Conjunctivitis, Dupilumab, Keratitis, Real-world evidence
Key Summary Points
| Clinical trials and observational studies have shown increased ophthalmic adverse events with dupilumab for atopic dermatitis (AD), although most have not resulted in permanent discontinuation. |
| Prior studies that evaluated dupilumab in routine care defined the study population (moderate-to severe AD) and study outcomes (conjunctivitis and keratitis) using only healthcare claims without validation against patient medical records. |
| This study is the most robust to date evaluating dupilumab use and the risk of ocular outcomes in routine care. |
| Findings are consistent with the known safety profile of dupilumab. |
Introduction
Atopic dermatitis (AD) is a common chronic inflammatory skin condition characterized by pruritis and skin lesions [1]. The estimated overall prevalence of AD in US adults is 10.2%, of which approximately 40% can be classified as moderate to severe [2]. AD affects quality of life and has psychological impacts on patients [3–6]. Treatment of moderate-to-severe AD (MTSAD) in adults includes topical and systemic medications (e.g., oral corticosteroids, immunomodulators) and/or phototherapy. Dupilumab is a human monoclonal immunoglobulin G-4 antibody that inhibits interleukin-4 and interleukin-13 signaling [7]. Initially approved by the US Food and Drug Administration (FDA) for use in adults with MTSAD in 2017, the label has since been expanded to include patients 6 months and older and for the treatment of other conditions [8].
Although the etiology remains unclear, conjunctivitis and keratitis have been reported among patients with MTSAD at higher rates than in the general population, and increased risk of ocular disorders has been linked with increasing severity of AD [9]. Conjunctivitis, or inflammation of the conjunctiva covering the eye, may be caused by bacterial or viral infection or an allergic reaction, and conjunctivitis of any type is the most common ophthalmic complication in patients with MTSAD [10]. Keratitis, or inflammation of the cornea of the eye, can similarly be caused by infection or allergic reaction but is typically more severe than conjunctivitis, possibly resulting in more intense pain or impaired eyesight [11]. Clinical trials in patients with AD found that conjunctivitis and keratitis occurred more frequently with dupilumab compared to placebo [12]. This finding has not been observed among patients treated with dupilumab for other indications, including asthma, eosinophilic esophagitis, chronic rhinosinusitis with nasal polyposis, chronic spontaneous urticaria, or COPD [12–14]. Conjunctivitis in patients with AD treated with dupilumab has also been described in case reports and single-arm studies [15–18].
To our knowledge, only two studies have quantified the incidence of conjunctivitis by type of AD therapy in a population-based cohort [19, 20]. These studies used claims-based codes without medical record confirmation or validation to identify MTSAD, conjunctivitis, or keratitis, which could result in unmeasured confounding due to the inclusion of patients with milder AD in the comparator group or in bias due to the misclassification of outcomes. Thus, there was a need to investigate the incidence of conjunctivitis and keratitis among individuals with MTSAD in a real-world setting using validated algorithms. Also needed are studies to describe the characteristics, risk factors, and management of conjunctivitis and keratitis among patients with MTSAD. To this end, an observational cohort study was conducted within a large US insurance claims database using validated algorithms to separately assess conjunctivitis and keratitis among patients with MTSAD, with and without dupilumab use.
Methods
Study Populations
This study utilized the Optum Research Database (ORD), a claims database from a large, geographically diverse US health insurer. Patients 12 years of age or older with at least one International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM) diagnosis code for AD (Supplementary Table 1) and a minimum of 12 months of continuous health plan enrollment were accrued into the source population from 28 March 2017, the date of US FDA approval of dupilumab, through 30 November 2019, which allowed for a minimum of 2 months of follow-up for all accrued patients prior to the beginning of the COVID-19 pandemic.
Optum authors received approval to use the Optum Research Database for the purposes of this study. The study was also reviewed and approved by BRANY IRB, approval #21–10-294–856.
From this source population, patients with MTSAD were identified using a validated algorithm [21] (Supplementary Table 2). This algorithm assessed diagnosis codes and treatments for AD available during the study period and did not include biologics or Janus kinase inhibitors that were approved after the follow-up period (2019). Among those with MTSAD, cohorts of patients initiating dupilumab and patients naïve to dupilumab receiving a non-dupilumab MTSAD treatment were identified. Cohort entry was the date of the first qualifying dispensing or administration meeting the MTSAD algorithm. Both cohorts were required to have no use of dupilumab in the prior 12 months but could have prior use of other AD therapies. Therefore, the dupilumab-naïve cohort included prevalent users of MTSAD therapies, which, based on validation analyses, were expected to be more comparable with dupilumab initiators with respect to AD disease severity than patients newly initiating these therapies. The baseline period was the 12 months prior to cohort entry. Patients could be members of both cohorts at different times if they met inclusion criteria for both. Within each cohort, only one cohort entry per patient was allowed.
Outcomes and Follow-up
The outcomes were conjunctivitis and keratitis, reported individually and defined based on validated algorithms [22] (Supplementary Tables 3–5). Follow-up for outcomes began the day after entry into the dupilumab initiator or dupilumab-naïve cohorts and continued until the earliest of a conjunctivitis event (in the analysis of conjunctivitis only), a keratitis event (in both analyses), discontinuation of dupilumab (among dupilumab initiators), start of dupilumab (among dupilumab-naïve patients), disenrollment from the health plan, or the end of follow-up (31 January 2020). Discontinuation of dupilumab was defined as a gap in dispensings of at least 60 days following the end of the days’ supply from the prior dispensing. The other AD therapy cohort was not censored because of discontinuation since this cohort was intended to represent routine care in the absence of dupilumab rather than treatment with any specific AD therapy.
Covariates and Matching
Members of the study cohorts were described with respect to baseline demographics, health plan enrollment, healthcare utilization, prior AD therapy, ocular conditions, prescription medications and presence of relevant comorbidities. These covariates were also included as dependent variables in a propensity score (PS) logistic regression model intended to balance the dupilumab initiators and dupilumab-naïve patients regarding the measured baseline covariates. The PS value for each patient in the dupilumab initiator and the dupilumab-naïve cohorts was estimated as the predicted probability of dupilumab initiation, given that patient’s observed covariate values. Dupilumab initiators were then matched on PS value 1:1 to dupilumab-naïve patients using a greedy digit-based matching algorithm [23]. After PS matching, patients with ICD-10 codes for conjunctivitis or keratitis in the 30 days prior to cohort entry were excluded to remove patients who may have had ongoing events at the beginning of follow-up.
Analysis
The incidence rates of conjunctivitis and keratitis within the PS-matched cohorts were calculated with 95% confidence intervals (CIs) [24]. Incidence rate ratios (IRRs) comparing dupilumab initiators with dupilumab-naïve patients were estimated using Poisson regression models, adjusting for variables that remained imbalanced after PS matching. IRRs were presented overall and separately, during and after the first 8 weeks of follow-up, as trials found increased risks after 8 weeks of dupilumab use [12]. Kaplan-Meier curves were produced to assess outcome timing and cumulative incidence.
To better understand patients with conjunctivitis or keratitis, those with the study outcomes during follow-up were described with respect to demographics and risk factors (using all available data prior to the event), specialty of the diagnosing provider, changes to dupilumab dose or frequency, and diagnostic tests or treatments used (during the 30 days before and after the event). Risk factors under consideration were selected based on a review of available literature [25–28].
To assess whether dupilumab initiators may have been more likely to discontinue corticosteroids near cohort entry compared with dupilumab-naïve patients, a post hoc analysis was conducted to describe systemic corticosteroid (SCS) use from 90 days before through 180 days after cohort entry.
All data management and analyses were performed using Statistical Analysis System (SAS) version 9.4 (SAS Institute Inc., Cary, NC). This study received approval by an Institutional Review Board, and all data access conformed to applicable Health Insurance Portability and Accountability Act policies.
Results
A study population of 13,790 patients with MTSAD was identified (Supplementary Fig. 1). Within this population, there were 2446 dupilumab initiators and 12,117 dupilumab-naïve patients meeting cohort inclusion criteria. After PS matching and exclusion of prevalent conjunctivitis and keratitis cases, 2175 dupilumab initiators and 2189 dupilumab-naïve patients were available for analysis.
Table 1 summarizes selected characteristics of patients before and after PS matching; the full list is provided in Supplementary Table 6. Before matching, the distributions of several baseline covariates differed between cohorts, with absolute standardized mean differences > 0.1 [29]. After matching, only a few differences remained. Specifically, the number of AD diagnosis codes on different days, number of dermatologist visits, and number of unique AD therapies were higher on average in the dupilumab cohort, and these were therefore included as adjustment variables in outcome models estimating IRRs.
Table 1.
Selected covariate distributions before and after propensity score matching among the dupilumab and dupilumab-naïve cohorts
| Before matching | After matchinga | |||
|---|---|---|---|---|
| Dupilumab Initiator | Dupilumab-naïve | Dupilumab Initiator | Dupilumab-naïve | |
| N = 2446 | N = 12,117 | N = 2175 | N = 2189 | |
| Age (years), median (IQR) | 40 (24–54) | 44 (25–57) | 40 (24–54) | 41 (25–54) |
| Female sex, N (%) | 1296 (53.0%) | 7191 (59.3%) | 1166 (53.6%) | 1151 (52.6%) |
| Healthcare utilization, median (IQR) | ||||
| Years of health plan enrollment | 3 (2–4) | 3 (2–4) | 3 (2–4) | 3 (2–4) |
| No. AD diagnosis codes on unique days | 3 (2–5) | 2 (1–3) | 3 (2–4) | 2 (2–3) |
| No. allergist visits | 0 (0–1) | 0 (0–0) | 0 (0–1) | 0 (0–1) |
| No. dermatologist visits | 2 (1–4) | 2 (1–3) | 2 (1–4) | 2 (1–4) |
| No. emergency department visits | 0 (0–1) | 0 (0–1) | 0 (0–1) | 0 (0–1) |
| No. hospitalizations | 0 (0–0) | 0 (0–0) | 0 (0–0) | 0 (0–0) |
| No. ophthalmologist visits | 0 (0–0) | 0 (0–0) | 0 (0–0) | 0 (0–0) |
| No. physician visits | 9 (5–14) | 8 (5–13) | 9 (5–13) | 9 (5–14) |
| No. unique AD therapies | 4 (3–6) | 3 (2–4) | 4 (3–6) | 3 (3–4) |
| Prior conditions, N (%) | ||||
| Allergic contact dermatitis | 424 (17.3%) | 1410 (11.6%) | 362 (16.6%) | 373 (17.0%) |
| Allergic rhinitis | 893 (36.5%) | 3727 (30.8%) | 751 (34.5%) | 767 (35.0%) |
| Asthma | 708 (28.9%) | 2545 (21.0%) | 587 (27.0%) | 577 (26.4%) |
| Cardiovascular disease | 182 (7.4%) | 1274 (10.5%) | 172 (7.9%) | 170 (7.8%) |
| Cerebrovascular disease | 29 (1.2%) | 249 (2.1%) | 25 (1.1%) | 26 (1.2%) |
| Conjunctivitis | 312 (12.8%) | 1284 (10.6%) | 212 (9.7%) | 216 (9.9%) |
| Crohn's disease/ulcerative colitis | 19 (0.8%) | 214 (1.8%) | 17 (0.8%) | 15 (0.7%) |
| Food allergies | 92 (3.8%) | 251 (2.1%) | 72 (3.3%) | 70 (3.2%) |
| Herpes simplex virus | 92 (3.8%) | 372 (3.1%) | 81 (3.7%) | 84 (3.8%) |
| Hypertension | 558 (22.8%) | 2985 (24.6%) | 506 (23.3%) | 530 (24.2%) |
| Hypothyroidism | 189 (7.7%) | 1273 (10.5%) | 170 (7.8%) | 181 (8.3%) |
| Keratitis/keratoconjunctivitis | 29 (1.2%) | 138 (1.1%) | 23 (1.1%) | 31 (1.4%) |
| Pruritus | 679 (27.8%) | 2203 (18.2%) | 578 (26.6%) | 601 (27.5%) |
| Psoriasis | 244 (10.0%) | 787 (6.5%) | 213 (9.8%) | 196 (9.0%) |
| Rheumatoid arthritis | 23 (0.9%) | 291 (2.4%) | 22 (1.0%) | 17 (0.8%) |
| Sinusitis | 359 (14.7%) | 2284 (18.8%) | 322 (14.8%) | 330 (15.1%) |
| Type 1 diabetes | 17 (0.7%) | 90 (0.7%) | 17 (0.8%) | 12 (0.5%) |
| Type 2 diabetes | 153 (6.3%) | 960 (7.9%) | 137 (6.3%) | 142 (6.5%) |
| Urticaria | 270 (11.0%) | 1009 (8.3%) | 235 (10.8%) | 248 (11.3%) |
| Systemic corticosteroid use before and after cohort entry, % | ||||
| 60 to 31 days prior | – | – | 23.7% | 19.5% |
| 30 days prior | – | – | 28.0% | 21.7% |
| 30 days after | – | – | 14.4% | 33.0% |
| 31 to 60 days after | – | – | 9.3% | 14.0% |
AD atopic dermatitis, IQR interquartile range, No. number
aThe number of patients in each cohort is unequal because patients with conjunctivitis or keratitis codes in the 30 days prior to cohort entry were dropped from the matched cohorts
Table 1 also displays a subset of results from the post hoc analysis of SCS use in the matched cohorts (full results available in Supplementary Table 7). Although use of SCS during the 12-month baseline period was balanced between cohorts after PS matching (Supplementary Table 6), their use near cohort entry remained imbalanced. Specifically, during each of the months prior to cohort entry, SCS use was more common among patients in the dupilumab initiator cohort, while after cohort entry, use was more common in the dupilumab-naïve cohort, indicating that patients may be stopping SCS at the time of dupilumab initiation.
In the overall cohorts before matching, 325 and 1030 conjunctivitis events were identified during follow-up among dupilumab initiators (N = 2446) and dupilumab-naïve patients (N = 12,117), respectively (Table 2). After PS matching (dupilumab initiators = 2175; dupilumab-naïve patients = 2189), the numbers of events were reduced to 260 and 148, respectively. After adjusting for the three variables that remained imbalanced in the PS-matched cohorts (number of AD diagnosis codes on different days, number of dermatologist visits, and number of unique AD therapies), dupilumab was associated with an estimated 86% increase in the incidence of conjunctivitis (IRR = 1.86 [95% CI 1.51–2.30]). Most of the excess incidence in the dupilumab cohort occurred after 8 weeks of dupilumab use, as the estimated IRR was 1.23 (95% CI 0.82–1.84) during the first 8 weeks of follow-up and 2.14 (95% CI 1.67–2.75) after 8 weeks of follow-up.
Table 2.
Relative incidence of conjunctivitis and keratitis among the dupilumab and dupilumab-naïve cohorts before and after propensity score matching
| Dupilumab initiator | Dupilumab-naïve | Adjustedb IRR (95% CI) |
|||||
|---|---|---|---|---|---|---|---|
| N | PY | IRa | N | PY | IRa | ||
| Conjunctivitis | |||||||
| Before matching | 325 | 1768.0 | 183.8 | 1030 | 11,481.9 | 89.7 | 1.78 (1.56, 2.04) |
| After matching | 260 | 1612.7 | 161.2 | 148 | 1789.7 | 82.7 | 1.86 (1.51, 2.30) |
| Up to 8 weeks follow-up | 63 | 323.7 | 194.7 | 45 | 312.8 | 143.9 | 1.23 (0.82, 1.84) |
| After 8 weeks follow-up | 197 | 1289.0 | 152.8 | 103 | 1476.9 | 69.7 | 2.14 (1.67, 2.75) |
| Keratitis | |||||||
| Before matching | 27 | 1995.7 | 13.5 | 58 | 12,320.4 | 4.7 | 2.72 (1.66, 4.48) |
| After matching | 26 | 1789.5 | 14.5 | 7 | 1892.7 | 3.7 | 4.06 (1.70, 9.68) |
| Up to 8 weeks follow-up | 3 | 328.2 | 9.1 | 0 | 316.6 | 0.0 | ** (**, **)c |
| After 8 weeks follow-up | 23 | 1461.3 | 15.7 | 7 | 1576.1 | 4.4 | 3.90 (1.62, 9.37) |
CI confidence interval, IR incidence rate, IRR incidence rate ratio, PY person-years
aIR calculated per 1000 person-years
bAdjusted for number of AD diagnosis codes on unique days, number of dermatologist visits, and number of unique AD therapies
cBecause there were 0 keratitis events during the first 8 weeks of follow-up in the other AD therapy cohort, it was not possible to estimate the keratitis IRR during this period
Few keratitis events were observed during follow-up, including 27 and 58 before matching and 26 and 7 after matching in the dupilumab and dupilumab-naïve cohorts, respectively. After matching and adjustment, the estimated IRR for keratitis comparing dupilumab initiators to dupilumab-naïve patients was 4.06 (95% CI 1.70–9.68). Due to the small numbers of events, an IRR could not be estimated during the first 8 weeks of follow-up. The IRR after 8 weeks was 3.90 (95% CI 1.62–9.37). Survival curves show that the cumulative 1-year risk of conjunctivitis was 15.8% and 8.4% in the dupilumab and dupilumab-naïve cohorts, respectively; the cumulative 1-year risk of keratitis was 1.3% and 0.2%, respectively. (Figs. 1a and b).
Fig. 1.
a Conjunctivitis events among propensity-score matched cohorts of dupilumab initiators and dupilumab-naïve patients. b Keratitis events among propensity-score matched cohorts of dupilumab initiators and dupilumab-naïve patients
Selected characteristics of observed conjunctivitis events in the dupilumab and dupilumab-naïve cohorts are provided in Table 3; the full list of characteristics evaluated before and after matching are available in Supplementary Tables 8 and 9 for conjunctivitis and keratitis events, respectively. Conjunctivitis was more commonly diagnosed by a dermatologist, ophthalmologist, or optometrist among dupilumab initiators, while diagnosis by an allergist or emergency medicine provider was more common among dupilumab-naïve patients. These patterns remained after PS matching. Among patients in the dupilumab cohort with a conjunctivitis event during follow-up, the median duration of dupilumab use before the event was 99 days. Only a small subset of dupilumab patients (8.6%) discontinued treatment within 30 days of the event, and nearly all remaining patients continued the dose that they were receiving prior to the event. As the study is based on claims data, the reasons for discontinuation are unknown.
Table 3.
Diagnosis and management of conjunctivitis events among the dupilumab and dupilumab-naïve cohorts before propensity score matching
| Dupilumab initiator | Dupilumab-naïve | |
|---|---|---|
| N = 325 | N = 1030 | |
| Specialty of the diagnosing provider | ||
| Allergist | 74 (22.8%) | 354 (34.4%) |
| Dermatologist | 29 (8.9%) | 6 (0.6%) |
| Emergency medicine | 3 (0.9%) | 22 (2.1%) |
| Ophthalmologist/optometrist | 102 (31.4%) | 254 (24.7%) |
| Primary care provider | 100 (30.8%) | 320 (31.1%) |
| Demographics | ||
| Age (years), median (IQR) | 41 (25–54) | 41 (22–56) |
| Female sex | 169 (52.0%) | 662 (64.3%) |
| Days of dupilumab use prior to outcome, median (IQR) | 99 (50–194) | – |
| Conjunctivitis management and treatmenta | ||
| Dupilumab discontinuationb | 28 (8.6%) | – |
| Dupilumab dose changeb | 4 (1.2%) | – |
| Diagnostic tests | ||
| Allergen specific IgE testing | 8 (2.5%) | 24 (2.3%) |
| Bacterial, fungal, or viral culture | 25 (7.7%) | 72 (7.0%) |
| Routine ophthalmic examination | 94 (28.9%) | 209 (20.3%) |
| Conjunctivitis and keratitis treatments | ||
| Punctal plug procedure | 4 (1.2%) | 5 (0.5%) |
| Mast cell stabilizers and combinations | 5 (1.5%) | 5 (0.5%) |
| Ophthalmic antibiotics (including steroid combinations) | 97 (29.8%) | 283 (27.5%) |
| Ophthalmic antihistamines | 39 (12.0%) | 100 (9.7%) |
| Ophthalmic corticosteroids | 65 (20.0%) | 71 (6.9%) |
| Ophthalmic antivirals | 22 (6.8%) | 45 (4.4%) |
| Topical immunomodulators | 20 (6.2%) | 55 (5.3%) |
| Topical NSAIDs | 3 (0.9%) | 9 (0.9%) |
| Systemic immunosuppressants | 55 (16.9%) | 191 (18.5%) |
| Systemic antibiotic therapy without other infection codes | 20 (6.2%) | 111 (10.8%) |
| Risk factors for conjunctivitisc | ||
| Allergic rhinitis (hay fever) | 242 (74.5%) | 748 (72.6%) |
| Asthma | 182 (56.0%) | 460 (44.7%) |
| History of conjunctivitis | 169 (52.0%) | 547 (53.1%) |
| History of keratitis | 22 (6.8%) | 38 (3.7%) |
AD atopic dermatitis, IgE immunoglobulin E, IQR interquartile range, NSAID non-steroidal anti-inflammatory drug
aTreatment and management were evaluated within ± 30 days of conjunctivitis event
bDupilumab discontinuation and dose change were evaluated only among the 83.1% of patients who had at least 30 days of dupilumab use at the time of the conjunctivitis event
cRisk factors were evaluated using all available data from October 2015 to the conjunctivitis event
The most common diagnostic test occurring within 30 days of a conjunctivitis event was a routine ophthalmic examination, occurring more frequently in the dupilumab cohort (28.9%) than in the dupilumab-naïve cohort (20.3%). Ophthalmic treatments were also more common in the dupilumab cohort, particularly ophthalmic corticosteroids (20.0% versus 6.9%). In contrast, use of systemic immunosuppressants (which were included in the cohort-defining drugs for the dupilumab-naïve cohort) and systemic antibiotics without any other infectious disease codes were more common in the dupilumab-naïve cohort. Treatment patterns were similar for conjunctivitis events occurring in the matched cohorts. Asthma and history of keratitis were more common among dupilumab cohort members with conjunctivitis (Table 3).
Due to the small number of observed keratitis events, a comparison of characteristics between study cohorts was difficult. Dupilumab discontinuation (7.4%) or dose change (0.0%) were infrequent following keratitis events (Supplementary Table 9).
Discussion
The results from this analysis for conjunctivitis (IRR = 1.86 [95% CI 1.51–2.30]) and keratitis (IRR = 4.06 [95% CI 1.70–9.68]) are consistent with the results of randomized trials, which estimated hazard ratios for conjunctivitis in the range of 2.0–4.1 for dupilumab compared with placebo [12]. This study also found that the increased risk occurred primarily after 8 weeks of dupilumab use and outcome events were treated with standard ophthalmic treatments (ophthalmic antibiotics and antibiotic steroid combinations) and did not typically lead to discontinuation of dupilumab. The 1-year risk of conjunctivitis observed in the current study is similar to the 1-year risk observed in clinical trials of patients with MTSAD [30] and does not increase with time as reported in a 5-year open-label extension study of dupilumab [31]. This consistency across different study designs reinforces the reliability of our findings.
The findings are additionally in line with two previous real-world studies of conjunctivitis risk among dupilumab versus methotrexate initiators, which estimated risk ratios of 2.45 (95% CI 1.47–4.08) [19] and 2.12 (95% CI 1.56–2.91) [20]. In the same studies, comparisons with cyclosporine and mycophenolate produced similar increased risks. However, sensitivity analyses that investigated potential detection bias indicated no increased surveillance for conjunctivitis events among the dupilumab group relative to the cyclosporine and mycophenolate groups, while this study provides evidence that detection bias may be present. Specifically, the increased prevalence of conjunctivitis and keratitis events diagnosed by ophthalmologists and dermatologists in the dupilumab group relative to the dupilumab-naïve group may indicate that these patients were monitored more closely for ocular events because of dupilumab’s known safety signal.
This study controlled for underlying conjunctivitis and keratitis risk and AD severity through the inclusion of relevant baseline variables in the PS model. Three variables (number of AD diagnosis codes on unique days, number of dermatologist visits, and number of unique AD therapies) intended to capture AD severity remained imbalanced after PS matching, but they were adjusted for in outcome models estimating IRRs. Nevertheless, there may be residual confounding present, including residual differences in underlying AD severity due to incomplete capture of AD severity in claims. It was observed that dupilumab initiators who developed conjunctivitis had a higher prevalence of a conjunctivitis risk factor, i.e., asthma, compared to dupilumab-naïve patients who developed conjunctivitis [32]. Dupilumab initiators may have had greater uncontrolled AD symptoms at cohort entry compared to their dupilumab-naïve counterparts, who were not initiating a new treatment. Indeed, in a post hoc analysis, dupilumab initiators were more likely than dupilumab-naïve patients to be using SCS immediately prior to cohort entry, and studies have reported higher usage of SCS among patients with severe AD compared to patients with moderate AD. Additionally, the data cannot capture patient compliance with medications, although repeated dispensings are typically associated with high compliance [29].
A significant strength of this study is the use of validated algorithms for the identification of both the MTSAD study population and the conjunctivitis and keratitis outcomes, reducing the potential for confounding by indication and outcome misclassification. Additional strengths were the large size of the study population and the robust PS-matched analysis, which controlled for a wide range of potential confounders and covariates. However, the algorithms for conjunctivitis and keratitis may not fully capture the milder forms of these conditions treated at home or with over the counter medications without consulting a healthcare provider, and residual confounding due to unmeasured/incomplete control for confounders might exist. AD is associated with several ocular events, and the risk increases with increasing AD severity [10, 32–34]. The current study evaluated only conjunctivitis and keratitis due to the availability of validated claims-based algorithms for these outcomes and does not include all other ocular events associated with having AD. Providers should consider monitoring all patients with MTSAD for ocular risks, given the incidence of conjunctivitis in these patients [10].
Conclusion
This study represents the largest and most rigorous claims-based investigation to date of the association between dupilumab and conjunctivitis/keratitis, including description of the management of, and risk factors for, these events in a US real-world setting. While increased risks for conjunctivitis and keratitis were observed among dupilumab initiators versus dupilumab-naïve patients, the study findings suggest that dupilumab initiators may have had more severe AD, greater surveillance for conjunctivitis and keratitis, and a higher prevalence of risk factors for the conditions. Notably, most patients with conjunctivitis and keratitis continued to receive dupilumab following these events. The findings of this study are consistent with the existing literature and do not alter the benefit-risk profile of dupilumab.
Supplementary Information
Below is the link to the electronic supplementary material.
Author Contributions
Jessica M. Franklin, Andrea F. Marcus, Ihtisham Sultan, Ashley Howell, Rachel E. Sobel, Sarah J. Sinnott, and Florence T. Wang contributed to the study conception and design. Data collection and analysis were performed by Stephen Ezzy and Robert Gately. The first draft of the manuscript was written by Jessica M. Franklin and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Funding
This work was completed under a research contract between Optum and Regeneron Pharmaceuticals, Inc. and funded by Regeneron Pharmaceuticals, Inc (Tarrytown, NY, USA) and Sanofi (Cambridge, MA, USA), including the journal’s Rapid Service Fee.
Declarations
Conflict of Interest
Jessica M. Franklin, Stephen Ezzy, Robert Gately, and Florence T. Wang are employees of Optum and may own stock in UnitedHealth Group. Andrea F. Marcus, Ihtisham Sultan, Ashley Howell, Jeannette Green, and Rachel E. Sobel are employees of and may own stock in Regeneron Pharmaceuticals, Inc. Rachel E. Sobel is a stockholder of Pfizer. Sarah J. Sinnott is an employee of and may own stock in Sanofi.
Ethical Approval
Optum authors received approval to use the Optum Research Database for the purposes of this study. The study was also reviewed and approved by BRANY IRB; approval #21–10-294–856.
References
- 1.Alexander T, Maxim E, Cardwell LA, et al. Prescriptions for atopic dermatitis: oral corticosteroids remain commonplace. J Dermatolog Treat. 2017;9:1–10. [DOI] [PubMed] [Google Scholar]
- 2.Fuxench ZCC, Block JK, Boguniewicz M, et al. Atopic dermatitis in america study: a cross-sectional study examining the prevalence and disease burden of atopic dermatitis in the US adult population. J Invest Dermatol. 2019;139(3):583–90. [DOI] [PubMed] [Google Scholar]
- 3.Birdi G, Cooke R, Knibb RC. Impact of atopic dermatitis on quality of life in adults: a systematic review and meta-analysis. Int J Dermatol. 2020;59(4):e75–91. 10.1111/ijd.14763. [DOI] [PubMed] [Google Scholar]
- 4.Chiricozzi A, Esposito M, Gisondi P, Valenti M, Gori N, Giovanardi G, Bellinato F, De Simone C, Costanzo A, Fargnoli MC, Peris K. Disease severity is associated with alexithymia in patients with atopic dermatitis. Dermatology. 2020;236(4):329–35. 10.1159/000507246. [DOI] [PubMed] [Google Scholar]
- 5.Courtney A, Su JC. The psychology of atopic dermatitis. J Clin Med. 2024;13(6):1602. 10.3390/jcm13061602. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Alessandrello C, Sanfilippo S, Minciullo PL, Gangemi S. An overview on atopic dermatitis, oxidative stress, and psychological stress: possible role of nutraceuticals as an additional therapeutic strategy. Int J Mol Sci. 2024;25(9):5020. 10.3390/ijms25095020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Wenzel S, Ford L, Pearlman D, et al. Dupilumab in persistent asthma with elevated eosinophil levels. N Engl J Med. 2013;368:2455–66. [DOI] [PubMed] [Google Scholar]
- 8.Dupixent (dupilumab). Full Prescribing Information. Regeneron. Tarrytown, NY. 2022.
- 9.Ravn NH, Ahmadzay ZF, Christensen TA, et al. Bidirectional association between atopic dermatitis, conjunctivitis and other ocular surface diseases: a systemic review and meta-analysis. J Am Acad Dermatol. 2021;85(2):453–61. 10.1016/j.jaad.2020.11.037. [DOI] [PubMed] [Google Scholar]
- 10.Thyssen JP, Toft PB, Halling-Overgaard A-S, Gislason GH, Skov L, Egeberg A. Incidence, prevalence, and risk of selected ocular disease in adults with atopic dermatitis. J Am Acad Dermatol. 2017;77(2):280–6. 10.1016/j.jaad.2017.03.003. [DOI] [PubMed] [Google Scholar]
- 11.Sharma S. Keratitis. Biosci Rep. 2001;21:419–44. 10.1023/A:1017939725776. [DOI] [PubMed] [Google Scholar]
- 12.Akinlade B, Guttman-Yassky E, de Bruin-Weller M, et al. Conjunctivitis in dupilumab clinical trials. Br J Dermatol. 2019;181(3):459–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Zirwas M, Chovatiya R, Gold LS, Shumel B, Praestgaard A, Chen Z, Gonzalez T, Gherardi G, Lawless E. 564-conjunctivitis adverse events in dupilumab clinical trials. Br J Dermatol. 2024. 10.1093/bjd/ljad498.060. [Google Scholar]
- 14.Bhatt SP, Rabe KF, Hanania NA, Vogelmeier CF, Cole J, Bafadhel M, Christenson SA, Papi A, Singh D, Laws E, Mannent LP, Patel N, Staudinger HW, Yancopoulos GD, Mortensen ER, Akinlade B, Maloney J, Lu X, Bauer D, Bansal A, Robinson LB, Abdulai RM, BOREAS Investigators. Dupilumab for COPD with type 2 inflammation indicated by eosinophil counts. N Engl J Med. 2023;389(3):205–14. 10.1056/NEJMoa2303951. [DOI] [PubMed] [Google Scholar]
- 15.Maudinet A, Law-Koune S, Duretz C, et al. Ocular surface disease induced by dupilumab in severe atopic dermatitis. Opthamol Ther. 2019;8(3):485–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Nettis E, Bonzano L, Patella V, et al. Dupilumab-associated conjunctivitis in patients with atopic dermatitis: a multicenter real-life experience. J Investig Allergol Clin Immunol. 2020;30(3):201–4. [DOI] [PubMed] [Google Scholar]
- 17.Liberman P, Shifera AS, Berkenstock M. Dupilumab-associated conjunctivitis in patients with atopic dermatitis. Cornea. 2020;39(6):784–6. [DOI] [PubMed] [Google Scholar]
- 18.Achten R, Thijs J, van der Wal M, et al. Dupilumab-associated ocular surface disease in atopic dermatitis patients: clinical characteristics, ophthalmic treatment response and conjunctival goblet cell analysis. Allergy. 2023;78:2266–76. 10.1111/all.15717. [DOI] [PubMed] [Google Scholar]
- 19.Schneeweiss M, Kim SC, Wyss R, et al. Dupilumab and the risk of conjunctivitis and serious infection in patients with atopic dermatitis: a propensity score-matched cohort study. J Am Acad Dermatol. 2021;84(2):300–11. [DOI] [PubMed] [Google Scholar]
- 20.Schneeweiss M, Wyss R, Chin K, et al. Incidence of Bacterial and Nonbacterial Conjunctivitis in Patients with Atopic Dermatitis Treated With Dupilumab: A US Multidatabase Cohort Study. Dermatitis. 2022;33(6S1):S73–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Chomistek A, Franklin JM, Sobel R, et al. 348 development and validation of a claims-based algorithm for moderate-to-severe atopic dermatitis using U.S. real-world data. Br J Dermatol. 2023. 10.1093/bjd/ljac140.041. [Google Scholar]
- 22.Chomistek A, Franklin JM, Sobel R, et al. Development and validation of claims-based algorithms for conjunctivitis and keratitis. Halifax, Nova Scotia, Canada: ICPE; 2023. [DOI] [PubMed] [Google Scholar]
- 23.Parsons LS. Reducing Bias in a Propensity Score Matched-pair Sample Using Greedy Matching Techniques. Paper 214–26 in Proceedings of the Twenty-Sixth Annual SAS Users Group International Conference. Long Beach, California: SAS Institute, Inc; 2001. http://www2.sas.com/proceedings/sugi26/p214-26.pdf. Accessed 18 November 2020.
- 24.Rothman KJ. Epidemiology: An Introduction. New York, New York: Oxford University Press; 2002. [Google Scholar]
- 25.Chen Y-H, Lin C-L, Bau D-T, et al. Risk of allergic conjunctivitis in patients with type 1 diabetes mellitus: a population-based retrospective cohort study. BMJ Open. 2017;7(6): e015795. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Zhu L, Liu B, Zhong J. Post-traumatic right carotid-cavernous fistula resulting in symptoms in the contralateral eye: a case report and literature review. BMC Ophthalmol. 2018;18:1–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Azari AA, Barney NP. Conjunctivitis: a systematic review of diagnosis and treatment. JAMA. 2013;310(16):1721–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Treister AD, Kraff-Cooper C, Lio PA. Risk factors for dupilumab-associated conjunctivitis in patients with atopic dermatitis. JAMA Dermatol. 2018;154(10):1208–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Steiner JF, Porchazka AV. The assessment of refill compliance using pharmacy records: methods, validity, and applications. J Clin Epidemiol. 1997;50(1):105–16. [DOI] [PubMed] [Google Scholar]
- 30.Blauvelt A, de Bruin-Weller M, Gooderham M, Cather JC, Weisman J, Pariser D, Simpson EL, Papp KA, Hong HC, Rubel D, Foley P, Prens E, Griffiths CEM, Etoh T, Pinto PH, Pujol RM, Szepietowski JC, Ettler K, Kemény L, Zhu X, Akinlade B, Hultsch T, Mastey V, Gadkari A, Eckert L, Amin N, Graham NMH, Pirozzi G, Stahl N, Yancopoulos GD, Shumel B. Long-term management of moderate-to-severe atopic dermatitis with dupilumab and concomitant topical corticosteroids (LIBERTY AD CHRONOS): a 1-year, randomised, double-blinded, placebo-controlled, phase 3 trial. Lancet. 2017;389(10086):2287–303. 10.1016/S0140-6736(17)31191-1. [DOI] [PubMed] [Google Scholar]
- 31.Beck LA, Bissonnette R, Deleuran M, et al. Dupilumab in adults with moderate to severe atopic dermatitis: a 5-year open-label extension study. JAMA Dermatol. 2024;160(8):805–12. 10.1001/jamadermatol.2024.1536. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Rønnstad ATM, Hansen PM, Halling AS, Egeberg A, Kolko M, Heegaard S, Thyssen JP. Factors associated with ocular surface disease and severity in adults with atopic dermatitis: a nationwide survey. J Eur Acad Dermatol Venereol. 2022;36(4):592–601. 10.1111/jdv.17832. [DOI] [PubMed] [Google Scholar]
- 33.Govind K, Whang K, Khanna R, Scott AW, Kwatra SG. Atopic dermatitis is associated with increased prevalence of multiple ocular comorbidities. J Allergy Clin Immunol Pract. 2019;7(1):298–9. 10.1016/j.jaip.2018.10.009. [DOI] [PubMed] [Google Scholar]
- 34.Achten RE, Van Luijk C, Van der Rijst L, Bakker D, Spekhorst L, Zuithoff N, Schuttelaar M, Romeijn G, Voorberg A, Kamsteeg M, Haeck I, De Graaf M, Thijs J, De Boer J, De Bruin-Weller M. Identification of risk factors for dupilumab-associated ocular surface disease in patients with atopic dermatitis. Acta Derm Venereol. 2022. 10.2340/actadv.v102.1128. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.