Biologic Monotherapies for Moderate-to-Severe Atopic Dermatitis: A Systematic Review and Bayesian Network Meta-Analysis of Established and Investigational Agents

TL;DR

This study compares biologic treatments for severe eczema, finding that dupilumab is the most effective and reliable, while newer drugs show mixed promise.

Contribution

The first unified Bayesian network meta-analysis comparing approved and investigational biologics for atopic dermatitis.

Findings

Dupilumab showed the most consistent and reliable efficacy across all key endpoints.

Rocatinlimab demonstrated promising signals for deep clinical responses but with limited evidence.

Zumilokibart, rademikibart, and temtokibart showed encouraging efficacy, while telazorlimab had modest benefits.

Abstract

Biologic therapies targeting maladaptive type 2 inflammation have transformed the management of moderate‑to‑severe atopic dermatitis (AD); however, comparative evidence integrating both approved and next‑generation investigational agents remains limited. This Bayesian network meta‑analysis (BNMA) provides the first unified evaluation of all biologic monoclonal antibodies approved as monotherapy for AD (dupilumab, lebrikizumab, tralokinumab) together with emerging immunotherapies, including amlitelimab, rademikibart, rezpegaldesleukin, rocatinlimab, telazorlimab, temtokibart, and zumilokibart, across key efficacy measures.

A PRISMA‑2020-compliant systematic review and PROSPERO‑registered protocol (CRD420251162704) identified phase 2-3 randomized, double‑blind, placebo‑controlled trials reporting week‑16 outcomes (week‑24 for rocatinlimab). A Bayesian random‑effects NMA estimated relative risks (RRs) for EASI‑75, EASI‑90, IGA‑AD 0/1, and ≥4‑point improvement in itch Numeric Rating Scale (NRS). Treatment hierarchy was evaluated using both Bayesian (SUCRA) and frequentist (P‑score) approaches.

Seventeen randomized controlled trials (RCTs) involving more than 6,000 patients were included in the analysis. Dupilumab demonstrated the most consistent and reliable efficacy across all evaluated endpoints, including improvements in EASI-75, EASI-90, IGA-AD 0/1, and itch NRS. Among investigational therapies, rocatinlimab showed a notable signal for deep clinical responses, although the available evidence remains limited and less precise. Other emerging agents, including rademikibart, temtokibart, and zumilokibart, demonstrated encouraging efficacy profiles, whereas telazorlimab showed comparatively modest clinical benefit. Ranking analyses consistently positioned dupilumab as the most reliable and highest-performing therapy overall, followed by rocatinlimab. Overall, biologic therapies were generally well tolerated in the short term; however, dupilumab remains the only agent supported by extensive long-term safety data extending up to a decade.

This analysis relied on indirect comparisons anchored to short‑term placebo‑controlled induction periods, limiting assessment of long‑term safety and durability, particularly for mechanistically distinct agents such as rezpegaldesleukin, a regulatory T cell (Treg) pathway agonist, and rocatinlimab, an OXO pathway inhibitor. Evidence for several emerging therapies was restricted to small phase 2 trials, resulting in wide credible intervals (Crls) and lower certainty in comparative rankings. Trial heterogeneity in baseline severity, ethnic composition, and geographic setting may have introduced residual confounding despite sensitivity and meta‑regression analyses. These factors should be considered when interpreting relative efficacy estimates

This NMA demonstrates that dupilumab remains the most reliable and effective biologic monotherapy for moderate‑to‑severe AD, supported by the greatest precision, reproducibility, and long‑term safety. Rocatinlimab shows promising investigational efficacy but requires efficacy and long-term safety validation. Zumilokibart, rademikibart, and temtokibart emerge as additional candidates with encouraging activity, whereas telazorlimab showed limited clinical benefit. Collectively, these findings provide a comprehensive comparative framework to inform biologic selection and therapeutic sequencing.

Figures