Skip to main contentSkip to main content
HsiaoEye
HsiaoEye
蕭閔謙醫師 · 眼科衛教筆記
You are reading the English-mode interface. Some article body content is currently Chinese-only; full translation is in progress. Switch to Chinese
📑 Multicentre Bayesian RCT · Saadoun et al. 2026 · Am J Ophthalmol · 27 centres · 112 patients
Latest Research · Uveitis · Biologics

Biologics for Refractory Uveitis
3 Drugs Compared — 2026 RUBI Trial

A May 2026 trial in American Journal of Ophthalmology by Saadoun et al. — the RUBI trial (Refractory Uveitis Biotherapies) — is the first head-to-head, multicentre, Bayesian, open-label, 1:1:1 randomised comparison of three biologics for refractory noninfectious uveitis (NIU). Across 27 French centres, 112 patients with active refractory non-anterior uveitis received adalimumab (anti-TNF-α, n=44), anakinra (anti-IL-1, n=18), or tocilizumab (anti-IL-6 receptor, n=50) for 16 weeks. Key findings: the composite primary endpoint (≥2-step drop in Miami 9-step vitreous haze AND prednisone tapered to ≤0.1 mg/kg/day) was met by 16% on adalimumab and 14% on tocilizumab (difference −2.0%, 95% CrI −16.8 to +12.3) — comparable, and both modest. Anakinra was stopped early by the DSMB for insufficient efficacy. Corticosteroid-taper success: adalimumab 59% vs tocilizumab 74%. An informative exploratory study — not a copy-paste drug-selection formula for every patient.

⚠️ Disclaimer: General medical education, summarising the May 2026 RUBI trial (Saadoun D et al.) in Am J Ophthalmol. Uveitis has many causes; infectious and noninfectious disease require entirely different approaches. Biologic choice, initiation, cessation, and switching must be evaluated by an ophthalmologist experienced in uveitis together with relevant specialists (rheumatology/internal medicine). This article does not replace your discussion with your treating physician — do not stop, add, or switch medication on your own. This site does not engage in medical advertising or endorse any specific drug, clinic, or physician. For NHI coverage, refer to the latest NHIA announcements and your care team at the time of treatment.
RUBI 試驗一張圖看懂：難治型葡萄膜炎 + 三藥對照 + 行動卡 Diagram: noninfectious uveitis anatomy and inflammation targets, a 3-drug comparison table (adalimumab vs tocilizumab vs anakinra), and three patient-action cards. Refractory Noninfectious Uveitis: 3 Biologics Saadoun et al. 2026 · RUBI · 27 centres · 112 patients · 16 weeks ① What is «refractory» noninfectious uveitis? Uveal + vitreous inflammation Vitreous haze, macular edema, vasculitis «Refractory» = relapse despite: ① Stable corticosteroid (10–80 mg/d) ② ≥1 conventional immunosuppressant (MTX, azathioprine, MMF) → then biologics ② Three biologics at 16 weeks Measure Adalimumab anti-TNF-α Tocilizumab anti-IL-6R Anakinra anti-IL-1 Primary endpoint haze ↓≥2 + steroid taper 16% 14% stopped Steroid taper met ≤0.1 mg/kg/d 59% 74% — No macular edema CRT < 300 µm 48% 54% — Adverse events mostly injection-site 43% 54% — Adalimumab ≈ Tocilizumab on primary endpoint (diff −2.0%, NS) ✱ A strict composite endpoint — both rates are low; read in context ✓ Both reasonable ADA ≈ TCZ efficacy Choose by comorbidity, route, safety profile TCZ tapered slightly more (74% vs 59%) ! Anakinra ineffective here anti-IL-1 stopped early DSMB halted for futility Only this population/endpoint Don't extrapolate broadly or other indications ⚠ Read with care 16wk, open-label Small anakinra arm Noninfectious, non-anterior only Individualise the choice with your care team Refractory NIU: ADA ≈ TCZ, anakinra ineffective here · still individualise
RUBI trial at a glance. ① Refractory noninfectious uveitis (NIU): uveal and vitreous inflammation causing vitreous haze, macular edema, retinal vasculitis; «refractory» means relapse despite a stable corticosteroid dose (10–80 mg/d prednisone) plus ≥1 conventional immunosuppressant (MTX, azathioprine, MMF) before biologics are considered. ② Three biologics at 16 weeks: primary composite endpoint adalimumab (anti-TNF-α) 16%, tocilizumab (anti-IL-6R) 14% (difference −2.0%, NS); steroid-taper success ADA 59% vs TCZ 74%; no macular edema ADA 48% vs TCZ 54%; adverse events ADA 43% vs TCZ 54%; anakinra (anti-IL-1) stopped early for insufficient efficacy. ③ Three take-homes: (green) ADA and TCZ are both reasonable, individualised choices; (yellow) anakinra was ineffective in this population/endpoint but cannot be extrapolated broadly; (red) 16-week, open-label, small anakinra arm, only for noninfectious non-anterior uveitis — drug choice still needs individual assessment with your care team. Source: Saadoun et al. 2026 AJO (RUBI, NCT02929251, 112 patients, 27 centres).

1. Common clinic scenarios

'Doctor, my uveitis has flared for years; I can't taper off steroids — every time I do, it relapses — and even with an added immunosuppressant it's unstable. I heard there are «biologic» injections; will switching to one fix it? Which one?'

'Doctor, online I saw adalimumab and tocilizumab, and also anakinra. Which is most effective? Is the newer, pricier one better?'

'Doctor, my daughter (a child) also has uveitis and her biologic is NHI-covered. I'm an adult — why does mine need a self-pay assessment?'

The core question behind all three: once refractory noninfectious uveitis (NIU) reaches the «time for a biologic» stage, is one of the three drugs clearly better? The May 2026 RUBI trial in Am J Ophthalmol is the first head-to-head randomised comparison of adalimumab, anakinra, and tocilizumab. This article distils what it tells us — and what it cannot yet say.

2. Common patient Q&As

Q1: Has tocilizumab been proven better than adalimumab?
A: No. They were comparable on the primary endpoint, with no clear winner. RUBI's composite primary endpoint (≥2-step drop in Miami 9-step vitreous haze AND prednisone ≤0.1 mg/kg/d): adalimumab 16% (7/44), tocilizumab 14% (7/50), difference only −2.0% (95% CrI −16.8% to +12.3%, crossing 0). Tocilizumab's numbers looked slightly higher on some secondary measures (e.g. steroid-taper success 74% vs 59%), but these are exploratory with wide intervals — not proof of superiority.
Q2: Only 14–16% met the endpoint — does that mean biologics don't work?
A: No — the key is that the primary endpoint bar was set very high. It required BOTH «marked vitreous-haze improvement» AND «steroid tapered to very low (≤0.1 mg/kg/d) within 16 weeks» — a demanding dual threshold. Taken separately: steroid-taper success was actually 59% (adalimumab) to 74% (tocilizumab); macular edema resolved in 48–54%. So «14–16%» reflects the difficulty of hitting both strict conditions at once, not overall drug failure. It's a reminder that refractory uveitis is, by definition, hard to treat — full success within 16 weeks is not easy.
Q3: Anakinra was stopped early — does that mean it's useless or dangerous?
A: Don't extrapolate. It was stopped for futility (insufficient efficacy), not for safety. RUBI used a Bayesian multi-arm multi-stage design with a pre-specified «minimum 30% response rate» futility rule. At interim analysis the anakinra (anti-IL-1) arm's response was too low, and the independent DSMB recommended halting accrual. This only shows that in this specific setting — adult, refractory, noninfectious, non-anterior uveitis, this strict primary endpoint — anakinra was insufficiently effective. It does not mean IL-1 blockade is useless across all uveitis or other autoinflammatory diseases (e.g. adult-onset Still's, juvenile idiopathic arthritis), which have their own evidence.
Q4: So should I choose adalimumab or tocilizumab?
A: There's no «standard answer» — your care team decides based on your individual situation. Since efficacy is comparable, the choice weighs: (1) comorbid/systemic disease — e.g. with IBD, psoriasis, or ankylosing spondylitis, the anti-TNF-α adalimumab may serve double duty; (2) route and frequency — adalimumab subcutaneous every 2 weeks, tocilizumab weekly subcutaneous in this trial; (3) safety/contraindications — e.g. demyelinating disease, specific infection risks; (4) prior drug history — whether a biologic class already failed; (5) availability and cost. «Newer/pricier ≠ better» — RUBI shows the well-established adalimumab and tocilizumab perform comparably.
Q5: Are biologics safe? Will they make me prone to infection?
A: Most adverse events within RUBI's 16 weeks were mild to moderate. Overall 52% of patients had ≥1 treatment-related mild-to-moderate AE; the most common were injection-site reactions (75%), then outpatient-managed infections (16%) and GI effects (6%). AE rate was ~43% (adalimumab) and ~54% (tocilizumab). There were 13 serious AEs. But note: 16 weeks is short-term. Long-term biologic infection risk (including TB and hepatitis B reactivation) requires pre-treatment screening and ongoing monitoring — that's for your care team to manage.
Q6: For adult uveitis, is biologic therapy always NHI-covered?
A: Not necessarily — and adult vs paediatric coverage conditions differ. Rely on the current NHI rules and your care team at the time of treatment. NHI coverage for uveitis biologics has specific indications, age criteria, and prior-authorisation conditions that change over time; different drugs and underlying causes (e.g. with a specific systemic disease) have different conditions. This article makes no claim about your individual coverage — whether you qualify, whether self-pay applies, and the application process should be confirmed with your ophthalmologist or rheumatologist, your hospital's pre-service notice, and the latest NHIA announcements.
Q7: Can these results apply to my anterior or infectious uveitis?
A: No. RUBI enrolled only «noninfectious» and «non-anterior» (intermediate, posterior, panuveitis) patients. (1) Infectious uveitis (e.g. herpetic, toxoplasmic, TB-related) needs an entirely different approach — treat the infection first; immunosuppression/biologics can worsen it. (2) Isolated anterior uveitis follows different management logic and wasn't included. So RUBI's drug-selection conclusions apply only to confirmed-noninfectious, intermediate/posterior refractory disease. Pinning down which type of uveitis you have is the first step of treatment.

3. 30-second takeaway: 8 things to remember

📋 RUBI trial: 8 things to remember

  • First head-to-head 3-drug RCT: 27 French centres, 112 patients, Bayesian, open-label, 1:1:1, 16 weeks
  • ADA ≈ TCZ on primary endpoint: 16% vs 14% (difference −2.0%, 95% CrI −16.8 to +12.3, NS)
  • Strict composite endpoint: haze ↓≥2 steps AND prednisone ≤0.1 mg/kg/d; component rates are higher
  • Anakinra stopped early for futility: DSMB halted per futility rule, not safety; don't extrapolate
  • Steroid-taper success: ADA 59% vs TCZ 74% (exploratory, wide intervals)
  • Acceptable 16-week safety: 52% mild-moderate AEs, mostly injection-site (75%); long-term risk needs monitoring
  • Noninfectious, non-anterior only: not for infectious or isolated anterior uveitis; classify type first
  • Still individualise the choice: by comorbidity, route, safety, drug history, cost — decided by your care team; newer/pricier ≠ better

4. Noninfectious uveitis — and why it gets «refractory»

Uveitis is a broad group of intraocular inflammatory diseases, classified by location into anterior (iris/ciliary body), intermediate, posterior (choroid/retina), and panuveitis. Noninfectious uveitis (NIU) is not caused by a pathogen but by immune dysregulation — it may be isolated to the eye or part of a systemic disease (e.g. sarcoidosis, Behçet, Vogt-Koyanagi-Harada, birdshot chorioretinopathy). In RUBI, the commonest causes were idiopathic (48%), birdshot (25%), sarcoidosis (16%), Behçet (6%). NIU accounts for 10–15% of permanent blindness in industrialised countries and needs active control.

Treatment ladder: first-line is corticosteroids (fast-acting but with long-term side effects); then add a conventional immunosuppressant (methotrexate, azathioprine, mycophenolate mofetil) to spare steroids. When disease still relapses despite a stable corticosteroid dose (10–80 mg/d prednisone) plus ≥1 immunosuppressant, it is refractory — and only then are biologics considered. RUBI enrolled exactly this hardest-to-treat group, especially those with macular edema or retinal vasculitis, the lesions most likely to cause irreversible vision loss.

5. How was RUBI designed? (Bayesian 1:1:1)

RUBI (Refractory Uveitis Biotherapies) was a multicentre, open-label, 1:1:1 RCT across 27 centres of the French Uveitis Network, registered on ClinicalTrials.gov (NCT02929251) and following CONSORT. Inclusion: aged ≥18, active and refractory noninfectious, non-anterior uveitis (intermediate/posterior/panuveitis). «Active» meant vitreous haze >1 (Miami 9-step), and/or macular edema (central retinal thickness CRT ≥300 µm), and/or active retinal vascular lesions. The three arms:

BiologicTargetNRegimen
AdalimumabTNF-α44SC, 80 mg then 40 mg q2w
TocilizumabIL-6 receptor50SC, 162 mg weekly
AnakinraIL-118SC, 100 mg daily (stopped early)

All patients followed a standardised steroid protocol: starting prednisone 0.5 mg/kg/d (max 40 mg/d) for 4 weeks, then tapering toward ≤0.1 mg/kg/d by weeks 12–16. The Bayesian multi-arm multi-stage design used staged interim analyses with a pre-specified «minimum 30% response rate» futility rule — the basis for halting the anakinra arm.

6. Why was the primary endpoint so strict?

RUBI's primary endpoint was a composite: at week 16, a ≥2-step drop in vitreous haze on the Miami 9-step scale, AND prednisone tapered to ≤0.1 mg/kg/d. Why both? Because the real goal in uveitis isn't just «suppress inflammation» but «control it without depending on high-dose steroids» — inflammation controlled only by heavy steroids leads to cataract, glaucoma, osteoporosis, and glucose problems over time. So this composite tests efficacy AND steroid-sparing together — a pragmatic but high bar. The 14–16% success must therefore be read in the context of a demanding dual condition, not taken at face value as «the drug doesn't work.»

7. Results: adalimumab and tocilizumab are comparable

From June 2017 to August 2022, 114 were randomised and 112 analysed (55% female, median age 49; 49% panuveitis, 71% bilateral). Week-16 primary and key secondary outcomes (adalimumab vs tocilizumab):

Outcome (week 16)AdalimumabTocilizumabDifference (95% CrI)
Primary composite16% (7/44)14% (7/50)−2.0% (−16.8 to +12.3), NS
Steroid ≤0.1 mg/kg/d59%74%+14% (−4 to +33)
No macular edema (CRT<300)48%54%+6.1% (−13.5 to +25.5)
Vasculitis resolved34%26%−8.2% (−31 to +15)
BCVA change≈ stable≈ stableNo significant difference

Three observations: (1) Comparable on the primary endpoint — the difference crosses 0 with wide intervals, no clear winner. (2) Tocilizumab's numbers were slightly higher for steroid-tapering and macular edema, but these are exploratory secondary outcomes with wide intervals — signals worth studying, not proven advantages. (3) Vision was roughly stable over 16 weeks — for a refractory group, «holding steady, not worsening» is itself meaningful, but large short-term vision gains are uncommon. Overall: in refractory NIU, adalimumab and tocilizumab are comparable options.

8. Three cytokine pathways — why did IL-1 fail?

The three drugs block different inflammatory signalling «cytokines», which helps explain the differing results:

🔬 Three cytokine pathways

  • TNF-α ← Adalimumab: TNF-α is a core driver of the uveitic immune response; anti-TNF-α is already proven for non-anterior NIU in trials and is the only biologic approved for this indication. RUBI reaffirms it as a robust option.
  • IL-6 ← Tocilizumab: elevated aqueous IL-6 is found across Behçet, VKH, sarcoidosis, and idiopathic uveitis; anti-IL-6R improved acuity, reduced haze, and spared steroids in prior small trials and retrospective work. In RUBI it matched adalimumab.
  • IL-1 ← Anakinra: IL-1 blockade has a role in autoinflammatory diseases (adult-onset Still's, JIA) and small studies showed reduced macular thickness in some uveitis. But in RUBI's setting — adult, refractory, non-anterior NIU, strict composite endpoint — anakinra's response was too low and it was stopped early. Possible reasons: IL-1 may not be the dominant driver in this uveitis network, its short half-life requires daily injection, or this population's pathology is TNF/IL-6-led.

The clinical message: different cytokines play different roles in different uveitis, and «which one you block» affects efficacy. RUBI supports prioritising TNF-α- or IL-6-directed biologics, while IL-1 blockade underperformed in this specific population yet retains value in other autoinflammatory settings.

9. Safety: what was seen within 16 weeks?

Within 16 weeks, 146 treatment-related mild-to-moderate adverse events occurred in 52% of patients:

Adverse eventShare
Injection-site reactions75%（109 件）
Infections (mostly outpatient)16%（23 件）
Gastrointestinal6%（9 件）

Overall AE rate was ~43% (adalimumab) and ~54% (tocilizumab); plus 13 serious AEs. Remember two things: (1) RUBI observed only 16 weeks — this is short-term safety. The long-term vigilance for biologics is around TB reactivation, hepatitis B reactivation, serious infection, and malignancy, all needing thorough pre-treatment screening and regular on-treatment monitoring. (2) Injection-site reactions are common but usually mild. Safety oversight belongs with a care team experienced in biologics.

10. What this study cannot yet answer

⚠️ 5 limitations to remember

  • 16 weeks is short: refractory uveitis is a chronic course; 16 weeks can't show relapse rate, long-term vision preservation, or long-term safety (infection, malignancy).
  • Open-label (not double-blind): both sides knew the drug; subjective grading (e.g. vitreous haze) may be influenced by expectation, though imaging endpoints (CRT, vasculitis) are more objective.
  • Small anakinra arm: only 18 after early stop; can't finely characterise any subgroup role. «Ineffective» applies to this endpoint and population.
  • Strict composite lowers success rate: 14–16% reflects the bar, not «overall efficacy»; conclusions differ by endpoint (e.g. steroid-tapering).
  • Population-limited: adults, noninfectious, non-anterior; not generalisable to children, infectious, isolated anterior uveitis, or causes not enrolled.

11. Taiwan context and NHI notes

⚠️ Costs & coverage — confirm with your care team and the latest NHIA rules

  • NHI coverage for biologics: for uveitis, adalimumab/tocilizumab coverage has specific indications, age, severity, and prior-authorisation conditions that differ for adults vs children and change over time. This article makes no claim about your individual coverage — refer to your ophthalmologist/rheumatologist, your hospital's pre-service notice, and the latest NHIA announcements.
  • Pre-treatment screening: before biologics, screening for TB and hepatitis B/C is typically needed to reduce reactivation risk; arrange these and confirm costs with your care team.
  • Multidisciplinary care: refractory NIU often needs joint ophthalmology + rheumatology/internal-medicine care, especially with systemic disease (Behçet, sarcoidosis); drug choice can address both eye and systemic disease.

12. Red flags: when uveitis needs urgent care

🚨 Red flags for uveitis patients

  • Sudden significant vision drop, or a surge of new floaters/shadows — possible acute flare or complication
  • Severe eye pain, marked redness, worsening photophobia — with headache/nausea, rule out acute IOP rise
  • Central distortion, warping, or a scotoma — possible worsening macular edema
  • Fever, persistent cough, weight loss, night sweats while on biologics/immunosuppressants — warning for serious infection or TB reactivation; seek care promptly

If any of the above, return promptly to your ophthalmology or retina/uveitis specialist; also report drug-related systemic symptoms to your prescribing internist/rheumatologist.

13. Conclusion: putting the trial in its proper place

RUBI is a rare head-to-head randomised trial in refractory NIU. Its biggest contribution isn't «finding the strongest drug» but providing the direct comparative evidence that was missing: in this hardest-to-treat group, adalimumab and tocilizumab are comparable, both reasonable options, while anakinra was insufficiently effective here.

Three practical messages: (1) Don't chase «newest, priciest» — RUBI shows the established adalimumab and tocilizumab perform comparably; «right for you» matters more than «newest». (2) Drug choice is a team decision — integrating your uveitis type, comorbidities, drug history, safety, and cost, with ophthalmology and relevant specialists. (3) Ongoing follow-up beats one-off success — refractory uveitis is a long campaign; regular visits, guided adjustments, and pre/on-treatment screening are what preserve vision.

📚 HsiaoEye Related Articles — Retina / Inflammation

References

  1. Saadoun D, Ghembaza A, Touhami S, Girszyn N, Bielefeld P, Seve P, Pugnet G, André M, Leclercq M, Rogier T, et al; on behalf of the RUBI study group. Adalimumab, Anakinra, and Tocilizumab in Patients With Noninfectious Uveitis: A Multicenter Randomized Controlled Trial. Am J Ophthalmol. 2026;285:202–212. doi:10.1016/j.ajo.2026.01.037
  2. Jaffe GJ, Dick AD, Brézin AP, et al. Adalimumab in patients with active noninfectious uveitis. N Engl J Med. 2016;375(10):932–943.
  3. Sepah YJ, Sadiq MA, Chu DS, et al. Primary (Month-6) outcomes of the STOP-Uveitis study: evaluating the safety, tolerability, and efficacy of tocilizumab in patients with noninfectious uveitis. Am J Ophthalmol. 2017;183:71–80.
  4. Dick AD, Tugal-Tutkun I, Foster S, et al. Secukinumab in the treatment of noninfectious uveitis: results of three randomized, controlled clinical trials. Ophthalmology. 2013;120(4):777–787.
  5. Fabiani C, Vitale A, Emmi G, et al. Interleukin-1 inhibition with anakinra and canakinumab in Behçet's disease–related uveitis. Clin Rheumatol. 2017;36(1):191–197.
  6. Dick AD, Rosenbaum JT, Al-Dhibi HA, et al; Fundamentals of Care for Uveitis International Consensus Group. Guidance on noncorticosteroid systemic immunomodulatory therapy in noninfectious uveitis: Fundamentals Of Care for UveitiS (FOCUS) initiative. Ophthalmology. 2018;125(5):757–773.
  7. Levy-Clarke G, Jabs DA, Read RW, et al. Expert panel recommendations for the use of anti–tumor necrosis factor biologic agents in patients with ocular inflammatory disorders. Ophthalmology. 2014;121(3):785–796.
  8. Nussenblatt RB, Palestine AG, Chan CC, Roberge F. Standardization of vitreal inflammatory activity in intermediate and posterior uveitis. Ophthalmology. 1985;92(4):467–471.