30-second key takeaways
- Study: JAAD June 2026, multicenter retrospective from 6 US alopecia clinics (Martin et al.), 108 severe AA patients switching to a second oral JAK inhibitor after ≥6 months on the first. Baseline mean SALT 78 (85% with ≥50% loss, 48% total scalp loss).
- Second JAKi overall: 48.8% reached SALT≤20 (≤20% scalp loss — a clinically meaningful response), 32.6% reached SALT≤10. A failed first JAKi does not mean the JAKi route is exhausted.
- Third JAKi also worthwhile: of 21 who reached a third JAKi, 75% had any improvement, 52.4% reached SALT≤20, 38.1% reached SALT≤10. Most common path: tofacitinib → baricitinib → ritlecitinib.
- Strongest predictor: patients who responded to the first JAKi (SALT≤20) had 3.3× higher odds of responding to a second (adjusted OR 3.33; 95% CI 1.22-9.68; P=.022). This suggests responsiveness reflects an individual's disease biology that persists across different JAKi agents.
- Switch-path differences: tofacitinib → ritlecitinib (n=9) reached SALT≤20 in 87.5%, tofacitinib → baricitinib (n=59) 51.0%, baricitinib → ritlecitinib (n=20) 42.9%; switching to upadacitinib (n=3) showed no improvement. Group sizes vary widely — interpret cautiously.
- Safety: adverse events during switching were consistent with single-JAKi use, no new unexpected events. Most common: hypercholesterolemia 7.4%, acne 6.5%, LDL elevation 4.6%, URI 4.6%, headache 3.7%, scalp folliculitis 3.7%. No de-escalation required.
- Taiwan status (per our main alopecia areata article): baricitinib (Olumiant) is approved for adults ≥18 with severe AA, ritlecitinib (Litfulo) for ages ≥12 with severe AA; tofacitinib (Xeljanz) and upadacitinib (Rinvoq) are off-label for AA. Reimbursement follows current NHIA announcements.
This is an advanced topic that assumes you already know what alopecia areata is, how SALT is scored, how JAK inhibitors work, and which drugs are available in Taiwan. If not, start with our complete alopecia areata guide:
→ Alopecia areata complete guide: diagnosis, SALT severity, JAK inhibitors, treatment ladder
Why the switching question matters
JAK inhibitors are the biggest recent breakthrough for severe AA. But clinical practice faces a question the pivotal trials never answered: when the first JAK inhibitor is inadequate, loses efficacy over time, or must be stopped for side effects / insurance / other reasons — does switching to another drug of the same class (still a JAK inhibitor) still help? Or should you abandon this route?
Previously there were only scattered small studies and case reports (e.g., small tofacitinib-to-baricitinib cohorts) — small, single-center, and without assessing who is more likely to succeed when switching. Martin et al.'s 2026 study matters because it is the largest real-world series on this topic to date (108 patients), spanning 6 specialty centers, including under-studied third-line switches and less common agents (ritlecitinib, ruxolitinib, upadacitinib), and it used statistical modeling to identify predictors of switching success.
Study design
| Item | Detail |
|---|---|
| Design | Multicenter, retrospective chart review (no control group, no randomization) |
| Sites / period | 6 US specialty alopecia clinics; data collected Oct 2024 – Jan 2025 |
| Inclusion | Confirmed severe AA; ≥6 months on the first oral JAKi before switching (each subsequent JAKi also required ≥6 months before assessing response); excluded if <6 months on first JAKi or concurrent other systemic immunosuppressants / biologics |
| N / age | 108 patients; mean age 38.5 (SD 16.5); 66.7% female; 66.7% White, 7.4% Black, 7.4% Asian |
| Severity | Baseline mean SALT 78.0 (range 25-100); 85% SALT ≥50, 48% SALT=100 (total scalp loss); subtypes: universalis 45.4%, patchy 16.7%, totalis 16.7%, ophiasis 12.0% |
| Disease course | Mean AA duration 12.1 y; current episode mean 65.8 months (~5.5 y); family history 15.7%, atopic comorbidity 15.7%, other autoimmune 17.6% |
| Concurrent tx | 54.6% concurrent oral minoxidil (mean 2.97 mg/day), 28.7% intralesional corticosteroids |
| Outcome | SALT (Severity of Alopecia Tool, % scalp loss, 0 = none, 100 = complete); clinically meaningful response defined as SALT ≤20, with SALT ≤10 as a higher bar |
Two features to keep in mind about this cohort: first, it is hard-to-treat — mean disease duration >10 years, nearly half with total scalp loss, drawn from referral centers, inherently a poorer-responding group. Second, the gap between drugs was short: mean 2.7 months from stopping the first to starting the second, and the second was used for a mean of 17.9 months.
Response to the first JAK inhibitor
The most common first JAKi was tofacitinib (71.3%, mean 12.0 mg/day), then baricitinib (25.0%, mean 3.9 mg/day); deuruxolitinib / ritlecitinib / ruxolitinib ≤2 each. Overall, 66.7% had any improvement on the first, 35.8% reached SALT≤20, 27.4% reached SALT≤10.
The two main drugs differed widely: tofacitinib 75.8% improved, 47.8% SALT≤20; baricitinib only 41.7% improved, 8.3% SALT≤20, 0% SALT≤10. But this should NOT be read as "baricitinib is worse" — the baricitinib group happened to have a much longer current episode (mean 53.8 months), and longer episodes are less likely to reach SALT≤20, introducing selection bias.
Response to the second JAK inhibitor
Overall second JAKi: 48.8% reached SALT≤20, 32.6% reached SALT≤10. For a hard-to-treat group (first JAKi failed, very long disease, nearly half with total loss), this is encouraging — almost half got scalp loss under 20% after switching.
The third JAK inhibitor (few, but worthwhile)
21 patients reached a third JAKi, mean duration 12 months. Overall 75% had any improvement, 52.4% reached SALT≤20, 38.1% reached SALT≤10. The most common path was tofacitinib → baricitinib → ritlecitinib (n=9), of whom 55.6% reached SALT≤20 and 33.3% reached SALT≤10. Separately, all 3 who switched from ritlecitinib back to tofacitinib as the third agent reached SALT≤10.
There were failures too: 3 patients maintained SALT ≥50 (more than half the scalp bald) throughout even on a third agent. Switching is not a panacea, but for patients willing to keep trying, a third line still offers a better-than-even chance.
The key finding: who is more likely to respond to a switch?
The real clinical value of this study is identifying predictors of switching success. The authors entered all plausible factors into a multivariable logistic regression. The results:
The implication of "responded to first → likely to respond to second": responsiveness may reflect whether an individual's AA biology is sensitive to JAK-STAT blockade, a trait that carries across different JAKi agents. Practically: if you had partial regrowth on the first JAKi (even if it later failed or was stopped), a second is worth trying; if you had no response at all to the first, a second can still work (some initial non-responders succeeded), but the odds are lower — set expectations up front.
Outcomes by switch path
The study reports response rates for different "from-which-to-which" paths. First, a caveat: group sizes vary enormously (from 3 to 59), so small-group percentages are unstable — read the following as trends, not as proof that any path is definitively better.
A secondary observation: within the tofacitinib → baricitinib path, those previously on lower-dose tofacitinib (10 mg/day) reached SALT≤20 after switching to baricitinib 4 mg more often (59.5%) than those previously on higher-dose tofacitinib (20 mg/day) (12.5%) (P=.022). The authors speculate that disease uncontrolled even on high-dose tofacitinib may simply be more refractory. This is again a small-subgroup, exploratory finding — insufficient to support any "reduce dose before switching" rule.
Safety of switching
Adverse events across the whole switching course (first, second, third) were consistent with prior single-JAKi studies — no unexpected new events, and no de-escalation or discontinuation due to side effects. Most common:
- Hypercholesterolemia 7.4%, LDL elevation 4.6%, triglyceride elevation 2.8% — known JAKi lipid changes; monitor lipids periodically during treatment.
- Acne 6.5%, scalp folliculitis 3.7% — JAKi-related skin effects, usually manageable, rarely requiring discontinuation.
- URI 4.6%, GI symptoms 4.6%, headache 3.7% — mostly mild.
- Events at <1%: lip HSV reactivation, leukopenia, weight gain, mild LFT / creatinine elevation, esophageal candidiasis, UTI, lower-extremity tingling, epistaxis, night sweats, cold intolerance.
JAK inhibitors available in Taiwan and their brand names
- baricitinib (Olumiant): JAK1/2 inhibitor, approved in Taiwan for adults ≥18 with severe AA (SALT >50%), usually 4 mg/day (2 mg/day if ≥75 or infection history). The most common "second" agent in this study.
- ritlecitinib (Litfulo): JAK3/TEC inhibitor, approved in Taiwan for severe AA aged ≥12 (including adolescents), 50 mg/day. A common second/third-line switch target in this study.
- tofacitinib (Xeljanz): JAK1/3 inhibitor, approved in Taiwan for RA, psoriatic arthritis, ulcerative colitis, etc., but off-label for AA. The most common "first" agent in this study.
- upadacitinib (Rinvoq): JAK1 inhibitor, approved in Taiwan for atopic dermatitis, RA, etc.; off-label for AA. Only 3 patients switched to it (none responded), too few to evaluate.
- deuruxolitinib (Leqselvi): US FDA-approved (2024) for adult severe AA, but not currently available in Taiwan.
Reimbursement / out-of-pocket terms, the latest indications, and age limits should follow current TFDA labels and NHIA reimbursement policy; for the detailed Taiwan coverage and cost breakdown, see our main AA article.
Frequently asked questions
Q1. The first JAK inhibitor grew no hair in 6 months — is a second worth trying?
There's a real chance. This study specifically looked at switching after ≥6 months on the first JAKi, and 48.8% reached SALT≤20 on the second. Be realistic, though: if you had NO response at all to the first, your odds on a second are lower than for those with prior partial response (responders had 3.3× higher odds). Whether to switch, and to what, is a dermatologist's decision based on your overall status, drug access, and safety.
Q2. How long before deciding the first one isn't working and switching?
This study required ≥6 months before assessing and switching. AA regrowth is inherently slow; most guidelines suggest trialing a JAKi for at least 6 months (sometimes 9-12) before judging efficacy, because some patients are "late responders" who only start regrowing after 6 months. The authors note their 6-month threshold may underestimate late responders. So the timing of "switch if it fails" should be individualized — don't give up too early, but don't wait indefinitely either.
Q3. Which JAK inhibitor is the best to switch to?
This study cannot answer "which is best." Some paths look great (baricitinib → tofacitinib 100%, tofacitinib → ritlecitinib 87.5%), but those are single-digit groups with huge variance. The largest, most reliable is tofacitinib → baricitinib (59 patients, 51.0%). The multivariable analysis found the identity of the second drug was NOT significantly associated with response — what matters is whether you responded to the previous one, not the brand. In practice the choice weighs age (ritlecitinib for ≥12), indication / reimbursement, safety, and drug access.
Q4. Does it matter whether the first one worked then failed, versus never worked?
This study's sample is too small to cleanly separate "responded then relapsed" from "never responded" (primary non-response) — a limitation the authors themselves note. But given that "responded to the first → 3.3× higher odds on the second," those who ever responded (even if it later failed) likely have a better switching prognosis than primary non-responders. Larger studies are needed to confirm.
Q5. How long a gap between drugs? Do I need a washout?
In this study the mean gap from stopping the first to starting the second was just 2.7 months, with no deliberate washout. JAK inhibitors have short half-lives and wash out quickly, so in practice switching usually means "stop the previous one, confirm no active infection or other contraindication, then start the next." The exact interval is the physician's call based on infection, vaccination, surgery, and lab values — not something to extend or shorten on your own.
Q6. Does repeated switching lose efficacy or cause resistance?
The data don't support the "diminishing returns" worry — the third-line response rate (52.4%) was not lower than the second (48.8%). JAK inhibitors are small-molecule targeted drugs and, unlike biologics, don't provoke anti-drug antibodies, so the "resistance" mechanism differs. Still, those reaching a third line are a selected group, so the figures can't be applied to everyone.
Q7. Can I use minoxidil or steroid injections alongside switching?
In this study, over half (54.6%) used low-dose oral minoxidil and 28.7% used intralesional corticosteroids — common adjuncts. The multivariable analysis found concurrent minoxidil / intralesional steroids were NOT significantly associated with switching outcomes — i.e., they aren't decisive for success, but are widely accepted adjuncts. Let your physician arrange any combination.
Q8. Do these results apply to me in Taiwan?
Partly, with caveats. This is a retrospective study from 6 US centers, predominantly White (66.7%), only 7.4% Asian; and tofacitinib was often first-line (off-label) in the US, whereas Taiwan's first-line may lean toward the approved baricitinib / ritlecitinib, changing the switch sequence. The biology of AA response to JAKi should be broadly similar across ethnicities, but the exact figures, drug access, and reimbursement differ. Treat it as directional ("switching is worth a try, and prior response is a good sign"), not as precise probabilities for Taiwan.
Common myths
Myth 1: "If the first JAK inhibitor fails, this route is dead."
False. Martin 2026 shows 48.8% reach SALT≤20 on a second JAKi and 52.4% on a third. A failed first does not mean the JAKi mechanism won't work for you — switching within the class still offers a substantial chance.
Myth 2: "You must switch to the newest / strongest one."
The multivariable analysis found the identity of the switched-to drug was NOT significantly associated with response. The real predictor is whether you responded to the previous drug, not the newness or claimed potency. Some paths look great (87.5%, 100%) only because of tiny group sizes and high variance — not because that drug is stronger.
Myth 3: "Switching is dangerous — side effects stack up."
The adverse events seen here matched single-JAKi use, with no unexpected new events and no de-escalation needed. Switching doesn't "stack" side effects (you stop one before starting the next, not take both at once). But the JAKi class warnings (infection, thrombosis, cardiovascular, malignancy) don't disappear with switching — monitoring is needed throughout.
Advanced: notes for clinical colleagues
First, the "response carries across agents" mechanistic hypothesis. SALT≤20 on the first → adjusted OR 3.33 on the second. The most practical reading: sensitivity to JAK-STAT blockade is a patient-level trait (perhaps reflecting peri-follicular cytokine profile, IFN-γ / IL-15 axis activity, or JAK-isoform dependence) that persists across different JAKi agents, rather than an idiosyncratic single-drug response. Clinically: anchoring switch decisions on the degree of first-drug response is more evidence-based than agonizing over which isoform selectivity to pick next.
Second, why upadacitinib looks worse but can't be read literally. Only 3 patients switched to upadacitinib here, all failing to reach target. That's both a tiny sample and selection-biased (those reaching a rarer third/fourth agent are the most refractory who failed everything prior). Upadacitinib is a highly selective JAK1 inhibitor, potent in atopic dermatitis but with limited AA experience; there's no adequate evidence it's inferior for AA — only that this study cannot evaluate it.
Third, the dosing-signal interpretation trap. Tofacitinib 10 mg → baricitinib 4 mg outperformed tofacitinib 20 mg → baricitinib 4 mg (59.5% vs 12.5%, P=.022). The intuitive misread is "lower dose first is better," but the more plausible reading is reverse causation / confounding by indication: those escalated to tofacitinib 20 mg and still uncontrolled had more severe, refractory disease with worse prognosis. This is an exploratory subgroup analysis and cannot guide a "reduce dose before switching" strategy.
Fourth, limitations to face honestly. Retrospective, no control group, no standardized switching criteria, referral-center population (limited generalizability), sample too small to separate relapse from primary non-response, the 6-month assessment threshold may underestimate late responders, and selection bias inflates second/third-line response rates. All this means figures like "48.8% / 52.4%" should be seen as an optimistic upper bound, not a general-population expectation. Next steps: prospective studies with standardized switching protocols and isoform-specific / biomarker stratification.
Takeaway: what you can do
If you or a family member has severe AA and is using or considering a JAK inhibitor:
- If the first doesn't reach the desired effect, don't lose heart and don't stop on your own. A second — even a third — still offers nearly half to better-than-half odds of meaningful regrowth. Whether, when, and what to switch to is a dermatologist's call.
- Remember that "the first one once responded" is a good sign — even if it later failed, switching success is notably more likely. At follow-up, proactively review with your doctor how much you responded to the previous drug.
- Give it enough time: AA regrowth is slow; usually at least 6 months (sometimes 9-12) is needed to judge a JAKi, so don't rush to switch after 1-2 months of no growth.
- Stay with physician-led safety monitoring: JAKi carry class warnings — pre-treatment evaluation, periodic blood tests (counts, liver/kidney, lipids), and watching for infection / zoster signs remain just as important after switching.
- Don't buy, adjust, or switch drugs on your own: JAKi are prescription drugs; dosing, indications, age limits, and reimbursement all require physician judgment and the latest Taiwan regulations.
For colleagues: the single takeaway is that switching JAKi is a reasonable strategy for refractory AA, especially in prior responders, and prior response can help guide sequencing. But all figures carry selection bias and small-sample limits — individualized decision-making remains central.
References
- Martin A, Chen LC, Kreytak C, et al. Switching between Janus kinase inhibitors for treatment of alopecia areata: A multicenter retrospective review. J Am Acad Dermatol. 2026;94(6):1662-1670.
- King BA, Mesinkovska NA, Craiglow B, et al. Development of the alopecia areata scale for clinical use: results of an academic-industry collaborative effort. J Am Acad Dermatol. 2022;86(2):359-364.
- Katamanin O, Ch'en PY, Song EJ. Exploring the efficacy of baricitinib in treating alopecia areata after failed janus kinase inhibitor therapy. JAAD Case Rep. 2024;43:36-39.
- Ehsani A, Razavi Z, Rahimnia A, et al. Switching JAK inhibitors: evaluating baricitinib's effectiveness in alopecia areata after tofacitinib failure. Arch Dermatol Res. 2025;317(1):491.
- Kazmi A, Moussa A, Bokhari L, et al. Switching between tofacitinib and baricitinib in alopecia areata: a review of clinical response. J Am Acad Dermatol. 2023;89(6):1248-1250.
- Triyangkulsri K, Suchonwanit P. Role of janus kinase inhibitors in the treatment of alopecia areata. Drug Des Devel Ther. 2018;12:2323-2335.
- Wyrwich KW, Kitchen H, Knight S, et al. The alopecia areata investigator global assessment scale: a measure for evaluating clinically meaningful success in clinical trials. Br J Dermatol. 2020;183(4):702-709.
- Taiwan Food and Drug Administration. Baricitinib (Olumiant) and Ritlecitinib (Litfulo) approval and labeling information. https://www.fda.gov.tw/