Digital Features for this article can be found at https://doi.org/10.6084/m9.figshare.22129661.

FormalPara Key Points

The oral Janus kinase inhibitor baricitinib has demonstrated efficacy for adults with severe alopecia areata (≥ 50% scalp hair loss) over 36 weeks.

In two phase III trials, efficacy of baricitinib for severe alopecia areata continued to improve over 52 weeks.

1 Introduction

Alopecia areata (AA) is an autoimmune disorder causing non-scarring hair loss on the scalp, face, and body. Severe AA (≥ 50% scalp hair loss) is unlikely to remit without treatment, with only 3.3–6.2% of patients achieving ≥ 80% scalp hair coverage with placebo in recent 36-week trials [1].

Clinical studies indicate Janus kinase (JAK) inhibitors may interrupt inflammatory pathways that contribute to the immunopathogenesis of AA [2,3,4]. The oral, selective, and reversible JAK1/JAK2 inhibitor baricitinib is the first FDA-approved treatment for adults with severe AA and is in late-stage development for pediatric patients with severe AA. There are limited data on longer-term outcomes with oral JAK inhibitors in large cohorts of patients with severe AA.

In two phase III randomized trials of adults with severe AA, baricitinib was superior to placebo in hair regrowth over 36 weeks of treatment [1]. Using data from the long-term extension periods of BRAVE-AA1 and BRAVE-AA2, we report the efficacy and safety of baricitinib 4 mg and 2 mg over 52 weeks of continuous therapy.

2 Methods

2.1 Patients and Study Design

BRAVE-AA1 (NCT03570749) and BRAVE-AA2 (NCT03899259) are ongoing, independent, randomized, double-blind, parallel-group, placebo-controlled studies evaluating the efficacy and safety of baricitinib for AA. Eligibility criteria were detailed previously [1]. Patients included adults who had a Severity of Alopecia Tool (SALT) score ≥ 50 (≥ 50% scalp hair loss) and a current AA episode lasting > 6 months to < 8 years without spontaneous improvement (i.e., no more than 10-point SALT score reduction) over the 6 months prior to screening.

The trials had identical design for the first 52 weeks. Patients were randomized 2:2:3 to receive once-daily oral placebo, baricitinib 2 mg, or baricitinib 4 mg for 36 weeks. All patients who completed the 36-week placebo-controlled period entered an extension phase, for up to 68 weeks of additional treatment. At Week 36, placebo non-responders (SALT score > 20) were rescued to baricitinib, while placebo responders (SALT score ≤ 20) remained on placebo through Week 52. The present analyses focus on patients randomized to either of the baricitinib doses at baseline, who retained their treatment allocation through Week 52 regardless of response at Week 36. Patients and investigators remained blinded to treatment assignment.

First and last patients, respectively, entered treatment in March 2019 and June 2020 in BRAVE-AA1 and July 2019 and May 2020 in BRAVE-AA2. The trials were conducted in accordance with the ethical principles of the Declaration of Helsinki and Good Clinical Practice guidelines. The research protocols were approved by each center’s Institutional Review Board (IRB) or Ethics Committee. All patients provided written informed consent.

2.2 Endpoints

Efficacy outcomes at 52 weeks included the proportion of patients achieving SALT score ≤ 20 (≤ 20% scalp hair loss), which is considered a clinically meaningful outcome for patients with ≥ 50% scalp hair loss at baseline [5]. The SALT score is a weighted sum of percentage hair loss in four areas of the scalp, ranging from 0 (no hair loss) to 100 (complete hair loss) [6]. Other endpoints included the percentage change from baseline in SALT score, the proportions of patients achieving ≥ 50% and ≥ 90% improvements from baseline in SALT score (SALT50 and SALT90, respectively), and the proportion achieving a SALT score ≤ 10 (≤ 10% scalp hair loss). Also assessed were the proportion of patients achieving Clinician-Reported Outcome (ClinRO) Measure for Eyebrow Hair Loss™ 0 or 1 (full coverage or minimal gaps) with ≥ 2-point improvement from baseline among patients with baseline scores of 2 or 3 (significant gaps or no notable eyebrows) and the proportion achieving ClinRO Measure for Eyelash Hair Loss™ 0 or 1 (no or minimal gaps) with ≥ 2-point improvement from baseline among patients with baseline scores of 2 or 3 (significant gaps or no notable eyelashes) [7].

Safety assessments included adverse events (AEs), serious AEs (SAEs), and clinical laboratory tests. An independent data and safety monitoring committee periodically reviewed unblinded efficacy and safety data. Deaths, major adverse cardiovascular events (MACEs), and arterial (ATEs) and venous thromboembolic events (VTEs) were adjudicated by an independent blinded clinical endpoint committee.

2.3 Statistical Analysis

Efficacy analysis included all patients randomized to baricitinib at baseline (intention-to-treat analysis). Data were censored after permanent study drug discontinuation or if collected remotely due to the coronavirus disease 2019 (COVID-19) pandemic. For categorical outcomes, non-responder imputation was applied to missing and censored data. For continuous outcomes, missing and censored data were imputed with modified last observation carried forward, using the most recent non-missing post-baseline assessment. Results from these prespecified analyses are presented below. Placebo-controlled outcomes at Week 36 were previously reported with both the prespecified analyses and with multiple imputation applied to missing data [1]; Week-52 results using multiple imputation are available in the electronic supplementary material (ESM) for comparison with the Week 36 results. Differences in outcomes between baricitinib doses were not analyzed.

Safety analysis included all patients randomized to baricitinib at baseline who received one or more doses of study drug and who did not discontinue for the reason ‘lost to follow-up’ at first post-baseline visit. Safety analysis included data collected through 23 August 2021 in BRAVE-AA1 and 30 August 2021 in BRAVE-AA2. Safety findings thus reflect data collected beyond 52 weeks, with up to 128 weeks in BRAVE-AA1 and 108 weeks in BRAVE-AA2. Incidence rates (IRs) were calculated based on time at risk for a patient with an event.

3 Results

3.1 Patients

Among patients randomized at baseline to baricitinib 4 mg and 2 mg, respectively, 252/281 (89.7%) and 163/184 (88.6%) in BRAVE-AA1 and 209/234 (89.3%) and 134/156 (85.9%) in BRAVE-AA2 completed 52 weeks of study treatment (ESM Fig. S1). The most common reasons for discontinuation in BRAVE-AA1 and BRAVE-AA2 were patient withdrawal (n = 21 and n = 21, respectively), AEs (n = 8 and n = 11, respectively), and loss to follow-up (n = 11 and n = 9, respectively).

Baseline characteristics were comparable across baricitinib groups in both trials (Table 1). For patients randomized to baricitinib in BRAVE-AA1 and BRAVE-AA2, respectively, mean age was 37.0 and 38.4 years, 58.9% and 63.3% were female, and mean SALT scores at baseline were 85.9 and 85.1. For ClinRO Measures for Eyebrow Hair Loss and Eyelash Hair Loss, respectively, 69.7% and 59.8% in BRAVE-AA1 and 67.9% and 58.7% in BRAVE-AA2 had scores of 2 or 3 (significant gaps or no notable hair) at baseline.

Table 1 Baseline demographics and clinical characteristics

3.2 Efficacy

The proportions of patients achieving SALT score ≤ 20 continuously increased over the treatment period, with response rates for patients treated with baricitinib 4 mg and 2 mg, respectively, reaching 40.9% and 21.2% in BRAVE-AA1 and 36.8% and 24.4% in BRAVE-AA2 at Week 52 (Fig. 1, ESM Table S1). Among patients with severe baseline disease (SALT score 50–94) treated with baricitinib 4 mg and 2 mg, respectively, a SALT score ≤ 20 was achieved in 55.6% and 36.4% in BRAVE-AA1 and 46.1% and 35.7% in BRAVE-AA2. Among patients with very severe disease at baseline (SALT score 95–100) in the baricitinib 4 mg and 2 mg groups, respectively, 27.7% and 10.3% in BRAVE-AA1 and 27.7% and 15.1% in BRAVE-AA2 achieved a SALT score ≤20.

Fig. 1
figure 1

Proportions of patients achieving SALT score ≤20 through Week 52 in a BRAVE-AA1 and b BRAVE-AA2. A SALT score ≤20 indicates ≤20% scalp hair loss. Bars represent 95% confidence intervals. Non-responder imputation was applied to missing data (prespecified analysis). Week-36 data points reflect results from the placebo-controlled period. SALT Severity of Alopecia Tool

In the baricitinib 4 mg and 2 mg groups, respectively, 29.9% and 14.1% in BRAVE-AA1 and 27.8% and 16.7% in BRAVE-AA2 achieved a SALT score ≤ 10 (ESM Table S1). With the baricitinib 4 mg and 2 mg doses, respectively, SALT50 was achieved at Week 52 in 52.0% and 31.0% in BRAVE-AA1 and 52.6% and 37.2% in BRAVE-AA2; 28.5% and 11.4% in BRAVE-AA1 and 26.1% and 14.1% in BRAVE-AA2 achieved SALT90 at Week 52 (ESM Table S1). Representative photographs of patients who responded to baricitinib in BRAVE-AA1 are presented in Fig. 2.

Fig. 2
figure 2

Patient images ©Eli Lilly and Company. SALT Severity of Alopecia Tool

Clinical photographs of patients with alopecia areata at baseline and after 36 weeks and 52 weeks of treatment with baricitinib 4 mg in BRAVE-AA1. SALT scores indicate percentage of scalp hair loss as assessed by the investigator.

Eyebrow and eyelash response rates also increased over the 52-week period. Among patients with ClinRO Measure for Eyebrow Hair Loss baseline scores of 2 or 3 who were treated with baricitinib 4 mg and 2 mg, respectively, 39.4% and 27.9% in BRAVE-AA1 and 49.7% and 16.3% in BRAVE-AA2 improved ≥ 2 points and scored 0 or 1 at Week 52 (Fig. 3, ESM Table S1). Of those with ClinRO Measure for Eyelash Hair Loss baseline scores of 2 or 3 in the baricitinib 4 mg and 2 mg groups, respectively, 40.7% and 21.6% in BRAVE-AA1 and 50.7% and 30.3% in BRAVE-AA2 had ≥ 2-point improvements and achieved scores of 0 or 1 at Week 52 (Fig. 4, ESM Table S1).

Fig. 3
figure 3

Proportion of patients achieving ClinRO Measure for Eyebrow Hair LossTM 0 or 1 with ≥2-point improvement from baseline through Week 52 among patients with a score of ≥ 2 at baseline in a BRAVE-AA1 and b BRAVE-AA2. A ClinRO score of 0 indicates full coverage and a score of 1 indicates minimal gaps in eyebrows. Bars represent 95% confidence intervals. Non-responder imputation was applied to missing data (prespecified analysis). Week-36 data points reflect results from the placebo-controlled period. ClinRO Clinician-Reported Outcome

Fig. 4
figure 4

Proportion of patients achieving ClinRO Measure for Eyelash Hair LossTM 0 or 1 with ≥ 2-point improvement from baseline through Week 52 among patients with a score of ≥ 2 at baseline in a BRAVE-AA1 and b BRAVE-AA2. A ClinRO score of 0 indicates full coverage and a score of 1 indicates minimal gaps in eyelashes. Bars represent 95% confidence intervals. Non-responder imputation was applied to missing data (prespecified analysis). Week-36 data points reflect results from the placebo-controlled period. ClinRO Clinician-Reported Outcome

Results of the post hoc analysis of efficacy outcomes using multiple imputation for missing data are reported in ESM Figs. S2–S4 and ESM Table S2.

3.3 Safety

A total of 280 (patient-years exposure [PYE] = 385.5) and 183 (PYE = 197.9) patients in BRAVE-AA1 and 233 (PYE = 284.8) and 155 (PYE = 168.7) patients in BRAVE-AA2 were exposed to baricitinib 4 mg and 2 mg, respectively. Among patients in the baricitinib 4 mg and 2 mg groups, respectively, treatment-emergent AEs (TEAEs) were reported in 69.6% (IR = 110.8) and 58.5% (IR = 94.5) in BRAVE-AA1 and 77.3% (IR = 159.1) and 74.2% (IR = 168.7) in BRAVE-AA2 (Table 2). Most TEAEs were mild or moderate in severity. The most frequently reported TEAEs were upper respiratory tract infection, headache, nasopharyngitis, acne, urinary tract infection, increased blood creatinine phosphokinase (CPK), and COVID-19 infection (Table 2). There were no patterns in types of SAEs; the most common SAEs were bone fractures due to injury (ESM Table S3). The frequency of discontinuations due to AEs was low and similar across groups (ESM Table S4).

Table 2 Summary of safety outcomes

AEs during the 36-week placebo-controlled period have been previously reported [1]. Newly reported herpes zoster infections that occurred after the placebo-controlled period were localized; all patients recovered, and none discontinued. Herpes simplex infections were all mild or moderate and there were no discontinuations. After the placebo-controlled period, serious infections during the extension phase of BRAVE-AA2 included one (0.4%, IR = 0.3) case each of appendicitis, herpes zoster, and COVID-19 with baricitinib 4 mg; all patients recovered, and none discontinued (ESM Table S3). During the extension period of BRAVE-AA2, one (0.4%, IR = 0.3) COVID-19 infection with baricitinib 4 mg led to study discontinuation. No serious infections or infections leading to discontinuation were reported in BRAVE-AA1. Malignancies reported during the trial extension periods included one (0.5%, IR = 0.5) squamous cell carcinoma after 16 months with baricitinib 2 mg, and one (0.4%, IR = 0.3) ductal carcinoma in situ after 10 months with baricitinib 4 mg in BRAVE-AA1; the latter patient discontinued per protocol. No VTEs, opportunistic infections, cases of tuberculosis, gastrointestinal perforations, or deaths were reported in either trial over the observation period (Table 2).

Most laboratory changes were balanced across baricitinib groups in both studies. Increases in CPK, low-density lipoprotein, and high-density lipoprotein were more frequent with the 4 mg dose (Table 3). CPK increases were mostly Common Terminology Criteria for Adverse Events (CTCAE) [8] grade 1 or 2, and there were no cases of rhabdomyolysis. CTCAE grade 4 neutropenia (neutrophils < 0.5 billion/L) was seen in one baricitinib 4 mg patient in the extension period of BRAVE-AA1; the patient recovered and continued the study. During the trial extension periods, thrombocytosis (platelets > 600 billion/L) occurred in one patient with baricitinib 4 mg in BRAVE-AA1 and one patient with baricitinib 2 mg in BRAVE-AA2; both patients recovered and continued the study. Frequencies of hepatic test abnormalities were consistent with those previously reported [1] and were not associated with TEAEs leading to study drug interruption or discontinuation.

Table 3 Summary of laboratory changes

4 Discussion

In BRAVE-AA1 and BRAVE-AA2, response rates in scalp hair regrowth increased over 52 weeks of baricitinib treatment in adults with severe AA, with 40.9% and 36.8%, respectively, of baricitinib 4 mg-treated patients achieving a SALT score ≤ 20 at Week 52. Eyebrow and eyelash response rates also increased over time, indicating that long-term treatment may be required to observe maximum clinical benefit. In both studies, numerically greater proportions of patients achieved efficacy endpoints with baricitinib 4 mg versus 2 mg, and these differences were sustained over time.

No new safety signals were observed versus the 36-week placebo-controlled period [1]. The most common AEs included upper respiratory tract infection, headache, nasopharyngitis, acne, urinary tract infection, CPK elevation, and COVID-19 infection. The frequency of COVID-19 infections reflects the evolution of the COVID-19 pandemic during the conduct of the trials. No VTEs were reported in either study. Newly reported malignancies included one case each of squamous cell carcinoma and ductal carcinoma in situ. Long-term safety data in rheumatoid arthritis do not suggest increased incidence of malignancies with baricitinib [9], but longer-term data are needed in AA to assess causal associations. Most increases in CPK to ≥ 5 times the upper limit of normal in baricitinib-treated patients were related to physical exercise, and no cases of rhabdomyolysis were reported.

As most patients randomized to placebo at baseline were non-responders (SALT score > 20) at Week 36 [1], they were switched to treatment with baricitinib at Week 36. Thus, there were no statistical comparisons with placebo in the present analysis. Both trials excluded patients who had current episodes lasting ≥ 8 years without any hair regrowth and patients with previous inadequate response to JAK inhibitors. Thus, efficacy in these patients remains to be established. Due to the COVID-19 pandemic, some assessments were missed or performed remotely and censored. As this may lead to an underestimation of efficacy values, analyses using multiple imputation are provided in the ESM.

Response rates for most endpoints did not plateau over the study period, indicating that maximum benefit in scalp, eyebrow, and eyelash hair regrowth may require more than 52 weeks of treatment, in particular for patients with more extensive disease. The present analysis demonstrates the benefits of continuous baricitinib treatment over 1 year for severe AA; however, longer periods of observation are required to evaluate the full extent and stability of clinical response. BRAVE-AA1 and BRAVE-AA2 are ongoing and will follow patients for up to 200 weeks. Eligible patients entered randomized withdrawal or downtitration at Week 52 to assess the durability of clinical response and long-term safety, and these data will be reported in future publications.

5 Conclusions

In two independent phase III trials of adults with severe AA, the proportions of patients achieving scalp hair, eyebrow, and eyelash regrowth increased over 52 weeks of treatment with once-daily oral baricitinib. Clinical response was numerically greater with baricitinib 4 mg than 2 mg. Safety findings were consistent with the known safety profile of baricitinib. These results confirm the potential for baricitinib in the treatment of severe AA.