JAK inhibitors regrow hair in severe alopecia areata. About a third reach the endpoint

Alopecia areata — the autoimmune condition that causes patchy or total loss of scalp, eyebrow and eyelash hair — went from having essentially no approved treatment to having two in the space of about a year, both of them oral Janus kinase inhibitors. The trials that produced that change are worth reading closely, because the headline result is real and the fraction of patients it applies to is smaller than the coverage suggested.

Baricitinib: the two phase 3 trials

BRAVE-AA1 and BRAVE-AA2 were randomised, placebo-controlled phase 3 trials in adults with severe alopecia areata, defined as a Severity of Alopecia Tool score of 50 or higher on a scale where 0 is no scalp hair loss and 100 is complete loss [s1]. Patients were assigned in a 3:2:2 ratio to once-daily baricitinib 4 mg, baricitinib 2 mg, or placebo. The primary outcome was a SALT score of 20 or less — meaning no more than a fifth of the scalp still bare — at week 36 [s1].

BRAVE-AA1 enrolled 654 patients and BRAVE-AA2 enrolled 546 [s1]. In BRAVE-AA1, the estimated percentage reaching the endpoint was 38.8% on 4 mg, 22.8% on 2 mg and 6.2% on placebo. In BRAVE-AA2 the corresponding figures were 35.9%, 19.4% and 3.3% [s1]. Against placebo, the 4 mg dose gave a difference of 32.6 percentage points in both trials (95% confidence intervals 25.6 to 39.5 and 25.6 to 39.6), and the 2 mg dose 16.6 and 16.1 percentage points [s1].

Secondary outcomes generally favoured baricitinib at 4 mg but not at 2 mg [s1]. Acne, elevated creatine kinase and increased low- and high-density lipoprotein cholesterol were more common on baricitinib than on placebo [s1]. The trial authors' own conclusion notes that longer trials are required to assess efficacy and safety [s1].

Ritlecitinib: a different target, a similar ceiling

A phase 2b–3 trial of ritlecitinib, an oral inhibitor of JAK3 and the TEC kinase family, ran at 118 sites in 18 countries and enrolled patients aged 12 and over with at least 50% scalp hair loss [s2]. Of 1,097 screened, 718 were randomised across six regimens and two placebo arms, with the primary endpoint the same SALT score of 20 or less, assessed at week 24 rather than 36 [s2].

At week 24 the response rates were 38 of 124 (31%) on the 200 mg loading dose followed by 50 mg, 27 of 121 (22%) on 200 mg followed by 30 mg, 29 of 124 (23%) on 50 mg, 17 of 119 (14%) on 30 mg, and two of 130 (2%) on placebo [s2]. The differences against placebo were 29.1 percentage points (21.2 to 37.9), 20.8 (13.7 to 29.2), 21.9 (14.7 to 30.2) and 12.8 (6.7 to 20.4) [s2].

Two things are worth noticing across the two programmes. The placebo response is very low — 2% to 6.2% — which is what makes a 30% response rate look transformative and also what confirms that spontaneous regrowth in severe disease is uncommon [s1] [s2]. And the ceiling is similar for both drugs at their best doses, in the low-to-mid thirties, meaning roughly two-thirds of patients with severe disease did not reach the endpoint.

The ritlecitinib trial population was 62% female, 68% White, 26% Asian and 4% Black or African American [s2]. One hundred and four of the 718 randomised patients discontinued treatment [s2].

What the longer safety data show

The obvious question about an oral immunomodulator taken indefinitely for a non-life-threatening condition is what happens over years rather than months. A pooled safety analysis of the two baricitinib trials, including their long-term extensions, covered 1,303 patients and 2,789.7 patient-years of exposure, with a median of 825 days and a maximum of 1,460 days — four years [s3]. Follow-up ran through at least 152 weeks, with data cut-offs in May 2023 [s3].

Most treatment-emergent adverse events were mild to moderate [s3]. The incidence rate of serious adverse events was 2.6 per 100 patient-years and of discontinuation due to adverse events 1.7 [s3]. Over an additional year of follow-up beyond the previously reported analysis, no new cases of serious infection, opportunistic infection, major adverse cardiovascular events, deep vein thrombosis or pulmonary embolism were observed [s3]. Incidence rates for non-melanoma skin cancer (0.1) and other malignancies (0.2) remained stable, herpes zoster ran at 1.9, and no deaths were reported in either study [s3].

The authors conclude that there were no new safety concerns or signals through a maximum exposure of four years [s3]. That is genuinely reassuring within its limits, and the limits matter: 1,303 patients over four years is a dataset that can detect common problems and cannot detect rare ones, and the JAK class carries regulatory safety labelling derived from much larger studies in other diseases and older populations.

The unanswered question

Neither trial programme establishes what happens when treatment stops. Alopecia areata is a relapsing autoimmune condition, and none of the published primary trials above was designed to measure durability off-drug. A patient who reaches a SALT score of 20 or less at week 36 has demonstrated that the drug works for them; they have not demonstrated that they can stop taking it.

This article is informational and is not medical advice.

Sources

• [s1] Two Phase 3 Trials of Baricitinib for Alopecia Areata — New England Journal of Medicine, published online 26 March 2022. https://doi.org/10.1056/NEJMoa2110343
• [s2] Efficacy and safety of ritlecitinib in adults and adolescents with alopecia areata: a randomised, double-blind, multicentre, phase 2b-3 trial — The Lancet, published online 13 April 2023. https://doi.org/10.1016/S0140-6736(23)00222-2
• [s3] Safety of Baricitinib in Adults with Severe Alopecia Areata from Two Phase III Trials Over a Median of 2.3 Years and Up to 4 Years — American Journal of Clinical Dermatology, 11 April 2025. https://doi.org/10.1007/s40257-025-00932-0