🧭 REBEL Rundown
📌 Key Points
- 🧠 Late-window TNK: OPTION tested tenecteplase 0.25 mg/kg for CTP-selected non-LVO stroke treated 4.5–24 hours after last-known-well.
- ✅ More excellent outcomes: mRS 0–1 at 90 days improved (43.6% vs 34.2%; RR 1.28 [1.04–1.57]; NNT ~11).
- ⚠️ Bleeding tradeoff: sICH increased (2.8% vs 0%; NNH ~35 using Heidelberg), and harms look worse with broader sICH definitions.
- 🧪 Close call: Signal is modest/fragile (FI 4; FQ ~0.7%) and likely underpowered for the observed effect size.
- 🌍 Practice-informing: Promising, but needs larger multicenter international replication before widespread adoption.
📝 Introduction
The OPTION Trial evaluated IV tenecteplase for CT perfusion–selected non-LVO ischemic stroke treated 4.5–24 hours after last-known-well. Most ischemic strokes are non–large vessel occlusions (non-LVO), yet once patients fall outside the traditional 4.5-hour thrombolysis window, options are limited and care often becomes supportive. While endovascular thrombectomy can be offered up to 24 hours for selected large vessel occlusions, an equivalent late-window strategy for non-LVO stroke has been less clear. With modern CT perfusion imaging identifying salvageable tissue well beyond 4.5 hours, OPTION asked a practical question: can late-window tenecteplase thrombolysis improve 90-day functional outcomes in non-LVO stroke without unacceptable bleeding risk?
🧾 Paper
Ma G, Mo R, Zuo Y, et al. Tenecteplase for Acute Non-Large Vessel Occlusion 4.5 to 24 Hours After Ischemic Stroke: The OPTION Randomized Clinical Trial. JAMA. Published online February 5, 2026. PMID: 41642827
⚙️ What They Did
📈 Results
Outcome | Tenecteplase | Standard medical treatment | Effect | p value / 95% CI | Notes |
Excellent functional outcome (mRS 0–1) at 90 days (primary) | 123/282 (43.6%) | 97/284 (34.2%) | RR 1.28 | 95% CI 1.04 to 1.57; P=0.02 | Post hoc risk difference 9.36%; 95% CI 1.37% to 17.46%; number needed to treat 11 |
Ordinal mRS distribution at 90 days | Shift towards lower disability | Reference | Common OR 1.39 | 95% CI 1.04 to 1.86; P=0.03 | Ordinal analysis; no multiplicity adjustment |
Functional independence (mRS 0–2) at 90 days | 177/282 (62.8%) | 157/284 (55.3%) | RR 1.14 | 95% CI 0.99 to 1.30; P=0.07 | Directionally favours tenecteplase; not statistically significant |
Reperfusion at 24 hours | 95/252 (37.7%) | 76/264 (28.8%) | RR 1.31 | 95% CI 1.02 to 1.68; P=0.03 | Assessed in patients with baseline non-LVO steno-occlusion and follow-up imaging available |
Infarct volume at 24 hours (mL), median (IQR) | 5.6 (1.2–19.8) | 6.2 (1.2–19.9) | Win ratio 1.02 | 95% CI 0.83 to 1.24; P=0.87 | Continuous non-normal outcome analysed with win ratio |
Early clinical response at 24 hours | 32/280 (11.4%) | 14/282 (5.0%) | RR 2.30 | 95% CI 1.26 to 4.22; P=0.007 | Denominators reflect available assessments |
Change in NIHSS from baseline to day 7, median (IQR) | −3 (−5 to −1) | −2 (−4 to −1) | Win ratio 1.16 | 95% CI 0.94 to 1.43; P=0.16 | Directionally favours tenecteplase |
EQ-5D-5L score at 90 days, median (IQR) | 0.9 (0.6–1.0) | 0.9 (0.6–1.0) | Win ratio 1.02 | 95% CI 0.82 to 1.27; P=0.83 | Quality-of-life measure; no clear separation |
Symptomatic intracranial haemorrhage within 36 hours (safety) | 8/281 (2.8%) | 0/284 (0%) | Risk difference 2.85% | 95% CI 1.16% to 5.54%; P=0.004 | RR not estimable due to zero events in control group |
Moderate or severe systemic bleeding within 90 days (safety) | 2/281 (0.7%) | 2/284 (0.7%) | RR 1.01 | 95% CI 0.14 to 7.12; P=0.99 | Rare events; wide confidence interval |
Death within 90 days (safety) | 14/281 (5.0%) | 9/284 (3.2%) | RR 1.57 | 95% CI 0.69 to 3.57; P=0.28 | Not statistically different; trial not powered for mortality |
- Sample size: 282 vs 284
- Age (median, IQR), y: 69 (59–75) vs 67 (59–74)
- Sex: Male 62.4% vs 68.3%; Female 37.6% vs 31.7%
- Ethnic group: Han 98.2% vs 99.7%; Manchu 0.7% vs 0%; Li 0.4% vs 0%; Mongolian 0.7% vs 0%; Zhuang 0% vs 0.4%
- Vascular risk factors / past history
- Hypertension: 74.8% vs 70.4%
- Diabetes: 25.9% vs 25.7%
- Atrial fibrillation: 12.8% vs 13.4%
- Hyperlipidemia: 7.4% vs 8.1%
- Previous stroke: 20.6% vs 21.8%
- Coronary heart disease: 9.9% vs 12.3%
- Baseline stroke severity / physiology
- NIHSS at randomization (median, IQR): 7 (5–9) vs 6 (5–9)
- Systolic BP (median, IQR), mm Hg: 153 (137–165) vs 153 (140–166.5)
- Serum glucose (median, IQR), mg/dL: 124.0 (104.6–155.5) vs 122.9 (101.0–167.0)
- Prestroke mRS: mRS 0: 94.0% vs 89.8%; mRS 1: 6.0% vs 10.2%
- Stroke location/etiology
- Circulation with steno-occlusion: Anterior 77.7% vs 76.8%; Posterior 22.3% vs 22.9%; Unknown 0% vs 0.4%
- Stroke etiology: Large-artery atherosclerosis 35.8% vs 38.0%; Cardioembolism 23.8% vs 23.2%; Undetermined/other 40.4% vs 38.7%
- Qualifying artery
- Occlusion (summary categories): M2 30.5% vs 32.8%; ACA 17.0% vs 13.4%; PCA 13.1% vs 14.4%; M3–M4 8.5% vs 6.7%; Other 6.4% vs 3.9%; M1 0% vs 1.1%
- Stenosis (any): 23.0% vs 23.6%
- No occlusion/stenosis: 1.4% vs 4.2%
- Occlusion—M2 dominance breakdown: Dominant 9.3% vs 20.4%; Codominant 9.3% vs 4.3%; Nondominant 81.4% vs 75.3%
- Occlusion—additional segments (examples): A2 12.8% vs 10.2%; P1 7.1% vs 6.3%; P2 4.6% vs 6.7%; PICA 2.5% vs 1.1%; SCA 1.8% vs 0.7% (full list in eTable 2)
- Stenosis—segment list (examples): M2 10.3% vs 10.9%; ICA 1.8% vs 2.5%; basilar 0% vs 0.4% (full list in eTable 2)
- Imaging volumes / onset category
- Ischemic core volume (median, IQR), mL: 0 (0–3.7) vs 1 (0–4.2)
- Perfusion lesion volume (median, IQR), mL: 34.8 (20.6–58.2) vs 37.6 (20.1–67.3)
- Perfusion mismatch volume (median, IQR), mL: 31.2 (18.6–56.4) vs 33.9 (18.8–63.6)
- Onset category: Known onset 67.4% vs 66.9%; Wake-up 32.3% vs 31.7%; Unwitnessed 0.4% vs 1.4%
- Timing / protocol-related baseline characteristics (supplement)
- Last-known-well → treatment (median, IQR), hours: 12.8 (9.0–17.0) vs 11.8 (8.7–16.9)
- Door-to-needle time (median, IQR), minutes: 97.5 (69.0–141.0) in TNK arm (control not applicable)
- Rescue endovascular treatment: 2.1% vs 0.4%
💥 Critical Results
- OPTION Trial Results: Tenecteplase improved mRS 0–1 at 90 days but increased symptomatic ICH.
- Primary efficacy (90-day excellent outcome):
- mRS 0–1 achieved in 43.6% (123/282) with tenecteplase vs 34.2% (97/284) with standard of care
- RR 1.28 (95% CI 1.04–1.57), p = 0.02
- Key safety (sICH within 36 hours; Heidelberg): 2.8% (8/281) with tenecteplase vs 0% (0/284) with control
- Risk difference: +2.85% (95% CI 1.16%–5.54%); p = 0.004
- Mortality (90 days): 5.0% with tenecteplase vs 3.2% with control
- Risk ratio: 1.57 (95% CI 0.69–3.57); not statistically significant
💪 Strengths
- Randomized multicenter clinical trial conducted across 48 centers, increasing internal validity and reducing single-center bias.
- Blinded outcome assessment (PROBE design: open-label treatment but blinded endpoint evaluation), which helps reduce outcome assessment bias.
- Clinically meaningful primary outcome using the modified Rankin Scale (mRS 0–1 at 90 days), a patient-centered functional endpoint commonly used in stroke trials.
- Imaging-based patient selection with CT perfusion identifying salvageable brain tissue (core–penumbra mismatch), enriching the study population most likely to benefit from reperfusion therapy.
- Central imaging core laboratory adjudication, improving consistency and reducing interpretation bias in imaging outcomes.
- Independent Data Safety Monitoring Board (DSMB) oversight to monitor safety during the trial.
- Intention-to-treat primary analysis, preserving randomization and minimizing bias from crossover or protocol deviations.
- Prespecified statistical analysis plan and sensitivity analyses, improving transparency and robustness of findings.
- Real-world extended treatment window (4.5–24 hours) addressing a clinically important gap for non-LVO stroke, where treatment options are limited.
⚠️ Limitations
- Open-label treatment allocation, which may introduce performance bias (clinicians aware of treatment assignment and may alter care based on this knowledge).
- Conducted exclusively in China, potentially limiting generalizability to other healthcare systems and stroke populations.
- Highly selected patient population requiring CT perfusion mismatch, which may limit applicability in centers without advanced imaging or in broader stroke populations. (Approx 5-7 patients per year per hospital in the included cohort)
- Moderate stroke severity (median NIHSS ~6–7), meaning results may not apply to patients with very mild or very severe strokes.
- Small absolute event rates for key safety outcomes, including symptomatic intracranial hemorrhage, leading to wide confidence intervals and fragility concerns.
- Potential detection bias in adverse events, as treating clinicians were aware of treatment allocation.
Rescue endovascular therapy was allowed after randomization and occurred more often in the TNK arm (2.1% vs 0.4%), introducing post-randomization treatment heterogeneity that could confound attribution of outcomes solely to tenecteplase.
🗣️ Discussion
Benefit vs Harm: NNT, NNH, and Likelihood to Help vs Harm (LHH)
The OPTION trial increased excellent functional outcome (mRS 0–1 at 90 days) by 9.4% (43.6% vs 34.2%), for an NNT ≈ 11. Using the trial’s primary (Heidelberg) definition, symptomatic ICH occurred in 2.8% vs 0% (risk difference +2.85%, 95% CI 1.16%–5.54%), yielding an NNH ≈ 35—about three additional excellent outcomes for every one extra symptomatic ICH event. Mortality was not statistically different (5.0% vs 3.2%; RR 1.57 [95% CI 0.69–3.57]), but the wide CI leaves room for clinically important harm—something that could become more apparent outside a tightly controlled trial setting.
Power and Precision: Sample Size Assumptions vs What Actually Happened
The trial was powered for a larger effect than it ultimately observed (planned ~12% absolute benefit; observed ~9.4%). When the true effect is smaller than expected, precision suffers—even if p < 0.05. That shows up in the primary estimate: RR 1.28 (95% CI 1.04–1.57), barely clearing the null, while harder endpoints remain uncertain (e.g., mortality RR 1.57; 95% CI 0.69–3.57). In practical terms, OPTION reads as a positive signal but there is enough uncertainty that we need replication before broad adoption—ideally in larger, more diverse cohorts across multiple countries and health systems.
Fragility Index & Fragility Quotient: How Stable Is the Signal?
OPTION is a positive trial—but it’s a close call. The fragility index (FI) is 4, meaning that if just four patients in the tenecteplase arm who were classified as mRS 0–1 were instead classified as mRS ≥2, the p-value would rise to ≥0.05 and the primary result would no longer meet conventional statistical significance. Scaled to trial size, the fragility quotient (FQ) is ~0.7% (4/566), underscoring how few outcome changes are needed to flip the headline finding.
Why does that matter here? The primary endpoint is binary (mRS 0–1 vs not), and mRS scoring is most variable at the borderline between mRS 1 and 2. If a handful of “close calls” were classified differently—especially in an open-label trial where patient reporting and follow-up care can differ—the result could plausibly shift. Fragility doesn’t invalidate OPTION, but it supports interpreting the findings as practice-informing rather than practice-changing.
Measurement and Definitions: Telephone mRS + Strict sICH Can Change the Narrative
The primary endpoint—mRS 0–1 at 90 days—is a standard, widely used stroke outcome, which makes OPTION easy to compare with prior thrombolysis trials. The tradeoff is that mRS is still partly subjective, and telephone assessment is less capable than in-person or video exams at detecting subtle disability and cleanly separating borderline categories.
In an open-label trial, that creates room for performance bias (post-randomization care differences) and response/ascertainment bias (patient reporting shaped by treatment awareness), both of which can nudge “close calls” toward better mRS categories. When the between-group difference is modest, even small shifts in classification can meaningfully influence the apparent treatment effect.
On the harm side, a strict sICH definition is pragmatic—we care most about hemorrhages that cause clinically meaningful neurologic worsening, not tiny radiographic bleeds. But stricter criteria also make events harder to qualify as “symptomatic,” which can shrink apparent harm. Using the broader NINDS definition (17/281 [6.05%] vs 5/284 [1.76%]) increases the absolute risk by 4.29%, yielding NNH ≈ 24 and tightening the likelihood-to-help-vs-harm to about 2.2:1 (24/11).
Taken together, how outcomes are measured (telephone mRS) and how harms are defined (sICH criteria) can push the narrative in opposite directions—inflating perceived benefit while minimizing perceived harm.
📘 Author's Conclusion
“Among patients with non–large vessel occlusion acute ischemic stroke and salvageable brain tissue, intravenous tenecteplase administered 4.5 to 24 hours after onset resulted in a greater likelihood of an excellent functional outcome at 90 days than standard care but had an increased risk of symptomatic intracranial hemorrhage.”
💬 Our Conclusion
The OPTION trial suggests TNK may improve excellent outcomes in CTP-selected non-LVO stroke patients treated 4.5–24 hours after last-known-well (NNT ~11 vs NNH ~35 for sICH). However, the signal is modest and fragile, and the trial was likely underpowered for the effect it ultimately observed. Given the open-label design, a partly subjective primary endpoint, and a strict sICH definition, benefit may be overstated and harm understated. Overall, OPTION is a signal toward benefit, but it needs confirmation in large, multicenter international trials to clarify efficacy and real-world bleeding risk.
🚨 Clinical Bottom Line
The OPTION trial found that in CTP-selected non-LVO stroke patients treated 4.5–24 hours after last-known-well, tenecteplase increased mRS 0–1 at 90 days (43.6% vs 34.2%; RR 1.28 [95% CI 1.04–1.57]) but increased sICH (2.8% vs 0%) Given the benefit signal is modest and fragile, these findings should be viewed as practice-informing, with larger, multicenter international trials in diverse cohorts needed before widespread adoption.
👉 FAQ
What is the OPTION Trial?
The OPTION Trial was a randomized clinical trial evaluating IV tenecteplase (TNK) 0.25 mg/kg given 4.5–24 hours after last-known-well in CT perfusion–selected non-LVO ischemic stroke patients.Who was eligible in the OPTION Trial?
Adults with non-LVO ischemic stroke, prestroke mRS 0–1, typically NIHSS 6–25 (or 4–5 with disabling deficit), and CT perfusion target mismatch (salvageable tissue). Large vessel occlusions (ICA, M1, vertebrobasilar) were excluded.What was the primary outcome in the OPTION Trial?
The primary outcome was excellent functional outcome: mRS 0–1 at 90 days.What were the main results of the OPTION Trial?
Tenecteplase increased mRS 0–1 at 90 days (43.6% vs 34.2%; RR 1.28 [95% CI 1.04–1.57]), but increased symptomatic intracranial hemorrhage (2.8% vs 0%, Heidelberg definition).What is the NNT and NNH in the OPTION Trial?
Based on the primary endpoint, NNT ≈ 11. Using the trial’s Heidelberg sICH definition, NNH ≈ 35.Should OPTION change practice?
OPTION is practice-informing, but it likely needs larger multicenter international replication before widespread adoption.
📚 References
- Rennert RC, Wali AR, Steinberg JA, et al. Epidemiology, Natural History, and Clinical Presentation of Large Vessel Ischemic Stroke. Neurosurgery. 2019;85(suppl_1):S4-S8.
- Zhou Y, He Y, Campbell BCV, et al. Alteplase for Acute Ischemic Stroke at 4.5 to 24 Hours: The HOPE Randomized Clinical Trial. JAMA. 2025;334(9):788-797.
- Ma G, Mo R, Zuo Y, et al. Tenecteplase for Acute Non-Large Vessel Occlusion 4.5 to 24 Hours After Ischemic Stroke: The OPTION Randomized Clinical Trial. JAMA. Published online February 5, 2026. PMID: 41642827
Post Peer Reviewed By: Anand Swaminathan, MD (Twitter/X: @EMSwami)
👤 Co-Editor-In-Chief
🔎 Your Deep-Dive Starts Here
MDCalc Wars: HAT vs SEDAN — Hemorrhage Risk After Stroke Thrombolysis
The HAT and SEDAN scores help estimate symptomatic intracranial hemorrhage ...
The Hope Trial: Alteplase 4.5–24 Hours After Stroke (CT Perfusion Selected)
The cornerstone of acute ischemic stroke (AIS) management involves timely ...
TNK vs tPA — The ORIGINAL Trial
The American Heart Association and the European Stroke Organization have ...
The AcT Trial: Tenecteplase vs Alteplase for Acute Ischemic Stroke
Background: Alteplase, a class of medication that converts plasminogen to ...
The ARAMIS Trial: DAPT vs Alteplase in Minor Nondisabling Acute Ischemic Stroke
Background: Current stroke guidelines recommend IV alteplase for patients with ...
Tenecteplase vs Alteplase in Acute Ischemic Stroke
Background: At my shop we have changed wholesale to the ...
