Prediction of patterns of persistent problematic atopic dermatitis in childhood

Summary: Abstract Body: Patterns of atopic dermatitis (AD) activity and severity vary throughout childhood, and clinicians have limited tools to counsel patients and their families on the likely course of the disease. We developed a clinical risk prediction algorithm for persistent problematic AD through age 14 among 7,406 children utilizing previously identified disease subgroups from a latent class analysis from the Avon Longitudinal Study of Parents and Children birth cohort. A multivariable logistic regression model was fitted to identify variables associated with the outcome of persistent problematic AD (a combination of the ‘Severe-Frequent’ and ‘Moderate-Frequent’ AD subtype compared to the ‘Moderate-Declining’, ‘Mild-Intermittent’, and ‘Unaffected/Rare’ subtypes). Known risk factors from early life (up to 42 months) were included, and a stepwise variable selection method identified the most relevant predictors. Fifteen candidate predictors were considered, and seven were retained for analysis. Predictors that were significant in the multivariable analysis (p < 0.05) were retained in the final scoring model. The final predictors were child sex, parental AD, food allergy, parent-reported severity of flexural dermatitis (FD) or convexity dermatitis between birth and 18 months, FD severity between 30 and 42 months, and age ≤ 6 months at the first report of dermatitis. Regression coefficients were used to develop a risk score (0–26 points), with children categorized into low (0–4), moderate (5–15 points), and high (16–26) risk groups. The predicted risk of persistent problematic AD ranged from 0.4% (score of 0) to 85.7% (score of 26). The area under the receiver operating characteristic curve was 0.89 (95% CI; 0.88–0.90). This easy-to-use risk score could help predict persistent problematic AD through early adolesence among children up to 42 months and may improve counseling and targeted identification of disease-modifying treatments. Ahnna Lee¹, Morgan Ye¹, Mi-Ok Kim¹, Lucy Pembrey², Charlotte Rutter², Sinéad M. Langan², Katrina Abuabara¹ 1. University of California San Francisco, San Francisco, CA, United States. 2. London School of Hygiene & Tropical Medicine, London, England, United Kingdom. Clinical Research: Epidemiology and Observational Research