Background Pediatric atopic dermatitis (AD) has been linked to later metabolic dysfunction, including obesity, dyslipidemia, etc. Systemic type-2-cytokine blockade with dupilumab may confer long-term protective metabolic effects in AD patients, but longitudinal studies in children receiving modern biologic therapy compared to topicals remain scarce. We compare short and long-term incidences of 23 metabolic disorders in pediatric AD managed with dupilumab, high-potency corticosteroids (HPCS), low/medium-potency corticosteroids (LMCS), topical calcineurin inhibitors (TCI), targeted non-steroidal topicals (crisaborole or ruxolitinib), or no therapy.

Methods We queried the TriNetX Global Network for patients <18 years carrying ICD-10 code L20 for AD. The index date was the first prescription or AD-related encounter, and baseline was the 12 months prior. Cohort sizes were: dupilumab 12,871; untreated 144,655; HPCS 45,992; LMCS 514,970; TCI 40,814; targeted topicals 9,408; any topical 634,356. Each comparator was 1:1 propensity-score matched to the dupilumab cohort on age, sex, race/ethnicity, and baseline metabolic diagnoses. All standardized mean differences were <0.01.

Outcomes were assessed at 6 months, 1, 3, and 5 years, and any time after index. Logistic regression provided odds ratios (OR) with 95% confidence intervals (CI); Kaplan-Meier curves, log-rank tests, and proportional-hazards diagnostics provided time-to-event differences. Two-sided p<.05 signified significance. Analyses were conducted in R 4.3 and TriNetX; de-identified data rendered our study IRB-exempt.

Results Within 1 year of treatment-onset, dupilumab showed numerically higher—but frequently non-significant odds for metabolic dysfunction compared with topicals (e.g., overweight/obesity OR 1.225, 95% CI 0.857-1.751 at 6 months vs TCI).

However, ~30 months post-index, Kaplan-Meier curves diverged; dupilumab seemed to confer a protective metabolic effect versus all comparators. By 5 years, dupilumab exhibited lower risk for nearly all metabolic outcomes. Versus the aggregated topical cohort, for example, 5-year ORs were 0.417 for overweight/obesity, 0.435 for dyslipidemia, 0.605 for elevated glucose, and 0.554 for acanthosis nigricans (all p<.001). This protective association continued beyond 5 years.

Protective associations were similar against HPCS, LMCS, and TCI, and modestly attenuated against targeted non-steroidal topicals. No comparator showed excess late metabolic risk relative to dupilumab.

Conclusion Our analysis with this large real-world pediatric cohort suggests dupilumab may increase early metabolic dysfunction (frequently without statistical significance), while conferring pronounced long-term protective metabolic effects compared with topical-only management or no therapy, supporting the hypothesis that sustained interleukin-4/13 inhibition may favorably alter metabolic risk trajectories in pediatric AD. Prospective studies adjusting for disease severity and further mechanistic studies may help confirm this causality and elucidate the underlying mechanisms.