Defining AD Disease Flares vs. Drug Effects: A Comparative Analysis of Phase 3 Systemic Therapy Trials

Dermatitis-related treatment-emergent adverse events (TEAEs) during the first 16 weeks of systemic therapy for atopic dermatitis (AD) vary substantially across clinical trials, according to a new structured narrative review.¹ The findings suggest that these events may reflect natural disease variability, incomplete treatment response, background topical therapy, and reporting practices rather than direct drug effects.

The review evaluated phase 3 randomized trials of approved systemic therapies used with background topical treatment, including dupilumab, lebrikizumab, tralokinumab, nemolizumab, upadacitinib, and abrocitinib.

Cross-Trial Variations and Reporting Discrepancies

Across treatment arms, dermatitis-related TEAEs ranged from 1% to 29%, while rates in placebo groups ranged from 3% to more than 41%. However, the authors cautioned against interpreting these figures as direct comparisons of flare risk because the trials differed substantially in baseline disease severity, topical treatment protocols, rescue therapy rules, and definitions of worsening.

In AD UP, dermatitis was reported in 3% of patients receiving upadacitinib 15 mg and 1% receiving 30 mg, compared with 7% with placebo. In ADhere, dermatitis-related worsening occurred in 2% of patients receiving lebrikizumab and 4% receiving placebo. JADE COMPARE reported rates of 3% with abrocitinib 100 mg, 1% with abrocitinib 200 mg, 1% with dupilumab, and 3% with placebo.

Higher rates were reported in LIBERTY AD CHRONOS, where dermatitis was recorded in 21% of patients receiving weekly dupilumab, 26% receiving dupilumab every 2 weeks, and 41% receiving placebo at week 16. A post hoc analysis using a protocol-defined worsening criterion found that dupilumab every 2 weeks plus topical corticosteroids was associated with a 78% relative reduction in annualized worsening compared with placebo.

In ECZTRA 3, protocol-defined flares occurred in 29% of patients receiving tralokinumab and 34% receiving placebo. A separate post hoc analysis found dermatitis-related adverse-event flares in 2% and 11%, respectively. In ARCADIA 1 and 2, MedDRA-coded dermatitis-related TEAEs occurred in 12% versus 11% and 7% versus 6% of nemolizumab and placebo groups, respectively.

Clinical Implications for Patient Management

The review emphasizes that “atopic dermatitis” TEAEs can encompass a range of clinical scenarios, from transient worsening to clinically meaningful exacerbations. Because most trials did not use standardized flare definitions, it is difficult to determine whether an event represented worsening of existing lesions, development of new lesions, changes in morphology or distribution, or another manifestation.

The findings have practical implications for treatment initiation. Early worsening should not automatically be interpreted as treatment failure. The authors emphasize consistent use of adjunctive topical therapy, including topical corticosteroids and other nonsteroidal topical agents as appropriate, along with regular emollient use. Counseling patients that intermittent worsening may occur during systemic treatment may help support adherence and reduce premature treatment changes. The review also notes that systemic therapies differ in their onset of action, which further complicates comparisons between JAK inhibitors and biologics. Longer-term trends may also differ as disease stabilization occurs.

Recommendations for Future Research

Future trials should use standardized definitions of AD flares incorporating objective disease-activity measures, treatment-escalation criteria, and potentially patient-reported outcomes. According to the review, greater standardization could improve interpretation of early disease activity and facilitate comparisons across systemic therapies.

Reference

1. Dasilva DR, Soto-Gonzalez A, Alkiswani H, et al. When Treatment Meets Natural History: Understanding Early Dermatitis-Related Adverse Events in Systemic Atopic Dermatitis Trials. JEADV Clinical Practice 0 (2026): 1-10, doi:10.1002/jvc2.70423.