Real-World Outcomes After Switching from Baricitinib to Ritlecitinib in Alopecia Areata: A Multicentre Retrospective Cohort Study

Why clinicians should know about it

• Picked for Dermatology (paper of the day, 1 November 2026): Real-world effectiveness of switching JAK inhibitors for alopecia areata

Abstract

BACKGROUND: Janus kinase inhibitors have reshaped the management of severe alopecia areata, but evidence for intra-class switching after an inadequate or non-durable benefit from baricitinib remains limited. OBJECTIVE: We aimed to assess real-world effectiveness after switching from baricitinib to ritlecitinib 50 mg once daily and to explore baseline factors associated with the post-switch response. METHODS: We performed a multicentre retrospective cohort study across Italian referral centres. Eligible patients had alopecia areata, had received baricitinib for at least 24 weeks and were switched to ritlecitinib 50 mg once daily. All switched patients with post-switch outcomes were included irrespective of prior baricitinib response status. The primary endpoint was a Severity of Alopecia Tool (SALT) score ≤ 20 at any post-switch assessment. Secondary endpoints were longitudinal SALT change and time to the first post-switch SALT score ≤ 20. Exploratory logistic and Cox models included age, current alopecia areata episode duration, SALT at ritlecitinib initiation (switch) and prior SALT50 response to baricitinib. RESULTS: Overall, 152 patients underwent a treatment switch and 138 had a post-switch follow-up. During baricitinib, a SALT50 score was recorded in 10/141 patients (7.1%) and a SALT score ≤ 20 at the final baricitinib assessment in 5/144 (3.5%); prior response was not an exclusion criterion. During ritlecitinib follow-up, 30/138 patients (21.7%) had a SALT score ≤ 20. Median SALT score changed from 95 at ritlecitinib initiation to 70 at week 24 and 10 at week 52 among patients observed at those visits, but a late follow-up was sparse. No baseline covariate showed clear evidence of association with the overall binary endpoint. In the exploratory Cox model, a higher SALT score at ritlecitinib initiation was modestly associated with slower achievement of a SALT score ≤ 20 (hazard ratio 0.98, 95% confidence interval 0.95-1.00; P = 0.034), with the confidence interval approaching the null. CONCLUSIONS: In this predominantly baricitinib-refractory, clinically switched cohort, a SALT score ≤ 20 was observed during the ritlecitinib follow-up in about one fifth of evaluable patients. The association between greater severity at ritlecitinib initiation and slower response was modest and should not be considered a validated predictor. Retrospective treatment selection, prior baricitinib responses in a small subgroup, late attrition, and unmeasured washout or concomitant-treatment effects limit treatment-specific attribution. These findings suggest that ritlecitinib may be a salvage option for selected patients and warrant prospective sequencing studies.

Abstract as published, via PubMed.