Quantitative evaluations of time-course and effectiveness of systemic treatments for atopic dermatitis: protocol for...

Treatment responses to systemic therapies for atopic dermatitis (AD) may vary significantly based on medication types, treatment duration, concomitant topical corticosteroid use and specific patient subgroups. Evidence regarding the optimal medication selection and predictors of treatment response is warranted. A systematic approach to studying the effectiveness of systemic treatments from multiple perspectives may generate previously unreported evidence. We aim to quantitatively evaluate the time-course and effectiveness of various systemic treatments for AD, and to identify associated factors that significantly affect their effectiveness, using a pharmacodynamic model-based meta-analysis (MBMA). This protocol describes the methods for this analysis.

Data for the analysis will mainly originate from the database of a living network meta-analysis. We will include randomised controlled trials examining systemic treatments against placebo or any active comparator in patients with moderate-to-severe AD. The Investigator Global Assessment Scale, the Eczema Area and Severity Index Scale, the Patient Oriented Eczema Measure Scale and scales measuring quality of life or itch will be used as outcome measures. Study screening, data extraction and quality assessment will be conducted by two investigators independently. Where available, we will extract efficacy outcome data at each follow-up time point. We will use MBMA, the core principle of which involves developing a pharmacodynamic model to synthesise these longitudinal trial-level data. Whenever possible, modelling and simulation will be conducted for both binary and continuous outcomes. This MBMA will be performed using hierarchical models with non-linear mixed-effects methods including structural, random-effects and covariate models with the maximum likelihood estimation method. Response to systemic treatments will be analysed by conducting 1000 Monte Carlo simulations. Potential influencing factors will be tested using covariate modelling and subgroup analysis.

Ethical approval is not required for this study. Results from our analyses will be published in a peer-reviewed journal.

CRD420251068535.