Publication

• Title: Intravenous Tenecteplase Prior to Endovascular Treatment for Ischemic Stroke at 4.5 to 24 Hours: The TNK-PLUS Randomized Clinical Trial
• Acronym: TNK-PLUS
• Year: 2026
• Journal published in: JAMA
• Citation: Xiong Y, Che F, Wang H, et al; TNK-PLUS Investigators. Intravenous tenecteplase prior to endovascular treatment for ischemic stroke at 4.5 to 24 hours: the TNK-PLUS randomized clinical trial. JAMA. Published online May 8, 2026.

Context & Rationale

• Background

  Tenecteplase 0.25 mg/kg, maximum 25 mg, is now guideline-supported for selected adults with acute ischaemic stroke within 4.5 hours, and advanced-imaging-selected thrombectomy is established for selected large-vessel occlusion stroke in the late window.¹
• Background

  The role of intravenous thrombolysis immediately before thrombectomy remains unsettled. The IRIS individual-patient-data meta-analysis of 6 early-window alteplase bridging trials did not establish non-inferiority of direct EVT, whereas BRIDGE-TNK subsequently reported superiority of tenecteplase before thrombectomy within 4.5 hours.²³
• Background

  The late-window tenecteplase evidence was also divergent: TRACE-III was positive in patients with large-vessel occlusion who did not have access to EVT, whereas TIMELESS was neutral in an advanced-imaging-selected population in which most patients underwent EVT.⁴⁵
• Research Question/Hypothesis

  TNK-PLUS tested whether intravenous tenecteplase before EVT improves 90-day functional independence compared with direct EVT alone in adults presenting 4.5 to 24 hours after last-known-well with MCA-M1 or proximal M2 occlusion and salvageable tissue.
• Why This Matters

  The practical question was immediate and common: for a patient already at an EVT-capable centre in the late window, should clinicians give a thrombolytic before moving to thrombectomy, or proceed directly to EVT?

Design & Methods

• Research Question: In adults with acute ischaemic stroke due to MCA-M1 or proximal M2 occlusion, 4.5 to 24 hours after last-known-well, with perfusion-defined salvageable brain tissue and direct access to EVT, does intravenous tenecteplase before EVT improve 90-day mRS 0–2 compared with EVT alone?
• Study Type: Multicentre, prospective, phase 3, randomised, open-label, blinded-endpoint, superiority trial conducted at 40 EVT-capable centres in China.
• Population:
  • Adults aged ≥18 years.
  • Acute ischaemic stroke 4.5 to 24 hours after last-known-well, including wake-up and unwitnessed stroke.
  • CTA or MRA-confirmed MCA-M1 or proximal M2 occlusion responsible for the clinical syndrome.
  • Target mismatch on CT perfusion or MRI plus perfusion imaging: ischaemic core volume <70 mL, mismatch ratio ≥1.8, and mismatch volume ≥15 mL.
  • Prestroke mRS ≤2 and baseline NIHSS 6–25.
  • Key exclusions included declining EVT or intravenous thrombolysis, tenecteplase allergy, rapidly improving symptoms, persistent blood pressure >185/110 mmHg despite treatment, glucose <2.8 or >22.2 mmol/L, active bleeding or high bleeding risk, relevant coagulopathy, platelet count <100×10⁹/L, recent stroke or myocardial infarction within 3 months, previous intracranial haemorrhage, large infarct >1/3 MCA territory, acute or past intracerebral haemorrhage, multiple arterial occlusions, pregnancy or lactation, inability to undergo CTP/MRP, or expected survival <1 year.
• Intervention:
  • Intravenous tenecteplase 0.25 mg/kg, maximum 25 mg, administered as a single bolus over 5–10 seconds immediately after randomisation.
  • EVT was then initiated as soon as possible, without waiting for clinical response to tenecteplase.
  • Permitted EVT techniques included stent retriever thrombectomy, aspiration, balloon angioplasty, stenting, or combinations at the discretion of the interventional team.
  • Rescue intra-arterial thrombolysis was permitted.
• Comparison:
  • Direct EVT without preceding intravenous tenecteplase.
  • The same EVT techniques and rescue options were available in the control group.
  • Standard early treatment and secondary prevention were otherwise delivered according to local Chinese stroke practice.
• Blinding: Treatment was open-label, but all clinical efficacy and adverse-event endpoints were assessed by physicians unaware of allocation. Neuroimaging was adjudicated by a blinded central imaging core laboratory. The 90-day mRS was assessed by certified blinded site clinicians and double-checked by a centrally located blinded evaluator using video or telephone recordings.
• Statistics: The adaptive sample-size design planned a minimum sample size of 390, powered at 80% with two-sided α=0.05 and 5% attrition to detect an approximately 14% absolute difference in 90-day mRS 0–2; a promising-zone interim sample-size re-estimation could increase enrolment to 754 to detect a more conservative 10% difference. The interim analysis after 273 patients completed 90-day mRS follow-up showed conditional power of 0.03, so no sample-size increase was made. The primary analysis used the full analysis set and log-binomial regression with site as a random effect, adjusted for age, baseline NIHSS, and occlusion site.
• Follow-Up Period: Primary functional outcome at 90 days; symptomatic intracranial haemorrhage within 36 hours; 24-hour vascular and perfusion imaging outcomes; 90-day mortality and serious adverse events.

Key Results

This trial was not stopped early. One adaptive interim sample-size re-estimation was performed; the sample size was not increased, and final enrolment was 391 patients because the final 2 patients provided consent almost simultaneously.

• The primary result was strongly neutral: 44.2% versus 43.2% achieved mRS 0–2, with the point estimate almost exactly null.
• Prespecified subgroup analyses did not show credible heterogeneity; the lowest interaction P value was for onset-to-randomisation time, P=.09, with RR 0.82 for <10 hours and RR 1.20 for ≥10 hours.
• Safety was not definitively worse, but symptomatic intracranial haemorrhage was numerically higher with tenecteplase, 5.1% versus 2.6%, with wide confidence intervals.

Internal Validity

• Randomisation and Allocation: Randomisation used a central web-based system, 1:1 allocation, block size 4, and stratification by occlusion site, M1 versus proximal M2. The sequence was generated by an independent statistician, and block size was not known to investigators.
• Drop out or exclusions: 565 patients were screened and 391 randomised; all 391 were included in the full analysis set. Per-protocol analyses included 189 patients in the tenecteplase-before-EVT group and 178 in the EVT-alone group. Protocol violations occurred in 24 patients, 6.1%.
• Drop out or exclusions: Safety analyses used 197 patients in the tenecteplase-before-EVT group and 190 in the EVT-alone group. The published flow diagram appears inconsistent with the main text and outcome table by labelling 192 patients in the tenecteplase safety population despite 199 randomised and 2 safety exclusions; the main outcomes consistently use 197 as the tenecteplase safety denominator.
• Performance/Detection Bias: The open-label design creates plausible performance bias at the procedural and post-procedural care level. Blinded mRS assessment, blinded clinical event adjudication, and blinded imaging core laboratory assessment substantially reduced detection bias.
• Protocol Adherence: Tenecteplase was administered rapidly after randomisation: median randomisation-to-tenecteplase time 9 minutes (IQR 5–18). EVT was not delayed meaningfully: randomisation-to-arterial puncture was 38 minutes (IQR 26–50) with tenecteplase versus 34 minutes (IQR 23–47) with EVT alone.
• Baseline Characteristics: Groups were broadly balanced: median age 68.0 versus 68.5 years; female sex 38.7% versus 40.6%; median NIHSS 14 versus 13; M1 occlusion 84.4% versus 84.9%; ASPECTS 8 versus 8; intracranial atherosclerosis 66.8% versus 63.0%; cardioembolism 31.7% versus 36.5%.
• Baseline Characteristics: Median ischaemic core volume was numerically smaller in the tenecteplase group, 6.8 mL (IQR 1.0–22.5) versus 13.1 mL (IQR 2.3–24.1), which would tend to favour tenecteplase if clinically important, yet no functional benefit was observed.
• Heterogeneity: Aetiology and endovascular technique were heterogeneous. Intracranial atherosclerosis accounted for 65.0% overall, and angioplasty and/or stenting was used in 115/188 (61.2%) versus 107/185 (57.8%). This reflects Chinese EVT practice and intracranial atherosclerotic biology, but treatment-effect interaction by stroke cause was not significant, P=.17.
• Timing: This was a genuinely late-window trial: median onset-to-randomisation was 9.3 hours versus 10.8 hours, and median onset-to-revascularisation was 11.8 hours versus 12.7 hours. The key biological limitation is that tenecteplase had a median dwell time of only 26 minutes from administration to arterial puncture.
• Dose: Tenecteplase 0.25 mg/kg, maximum 25 mg, is the contemporary standard stroke dose. The trial did not test whether a higher dose, repeated dose, or different thrombolytic strategy would change late-window bridging efficacy.
• Separation of the Variable of Interest: The intervention separation was clean for intravenous thrombolysis: tenecteplase bolus versus none. Other care was similar: general anaesthesia 117/198 (59.1%) versus 122/190 (64.2%); tirofiban use 113/199 (56.8%) versus 112/192 (58.3%); total number of passes median 1 (IQR 1–2) versus 1 (IQR 1–2).
• Key Delivery Aspects: Sites were required to be certified comprehensive stroke centres with EVT capability, acute stroke trial experience, ≥30 EVT procedures per year, and routine median door-to-needle time <60 minutes and door-to-puncture time <90 minutes in the standard window.
• Key Delivery Aspects: The trial achieved high procedural quality. Post hoc reperfusion after EVT was 179/188 (95.2%) versus 168/185 (90.8%), adjusted RR 1.04; 95% CI 0.98 to 1.10; P=.19.
• Crossover: There was little conventional crossover, but important exclusions occurred: in the tenecteplase arm, 1 patient received neither intravenous thrombolysis nor EVT and 1 received intravenous alteplase before EVT; in the EVT-alone arm, 2 did not undergo EVT.
• Adjunctive therapy use: Adjunctive tirofiban was common and balanced, 56.8% versus 58.3%. This limits confounding between groups, but it also means TNK-PLUS was conducted in a pharmacologically intensive EVT environment.
• Outcome Assessment: The primary endpoint, 90-day mRS 0–2, is clinically meaningful and standard in stroke trials. The primary outcome data were complete; no imputation was needed for the primary analysis.
• Outcome Assessment: Imaging secondary outcomes were less robust because 24-hour perfusion imaging was not completed in 104 patients due to clinical status, requiring multiple imputation. The authors reported similar findings without multiple imputation and in CTA-assessed sensitivity analyses.
• Statistical Rigor: The full-analysis-set primary analysis matched the prespecified approach, used adjusted log-binomial regression, and accounted for site as a random effect. Secondary outcomes were exploratory because multiplicity was not controlled.
• Conclusion on Internal Validity: Internal validity is moderate-to-strong for the primary clinical outcome because allocation concealment, endpoint blinding, central adjudication, complete primary follow-up, and protocolised analysis were strong. It is not “high” because treatment was open-label, the safety denominator reporting is internally inconsistent in the flow diagram, imaging secondary outcomes required substantial imputation, and rare haemorrhagic harms were underpowered.

External Validity

• Population Representativeness: The trial enrolled a highly selected imaging-defined group: MCA-M1 or proximal M2 occlusion, NIHSS 6–25, prestroke mRS ≤2, core <70 mL, mismatch ratio ≥1.8, mismatch volume ≥15 mL, and presentation directly to an EVT-capable centre.
• Population Representativeness: The cohort was almost entirely Han Chinese, 198/199 (99.5%) versus 191/192 (99.5%), with high intracranial atherosclerosis prevalence and high use of angioplasty/stenting. This limits generalisability to systems with lower intracranial atherosclerosis burden and different EVT practice.
• Applicability: The findings apply best to patients already at a comprehensive stroke centre where EVT can start quickly. They should not be extrapolated to drip-and-ship patients, geographically delayed EVT pathways, or systems in which tenecteplase may have a longer pre-thrombectomy dwell time.
• Applicability: The trial does not address patients presenting before 4.5 hours, patients without EVT access, internal carotid artery occlusion, posterior-circulation occlusion, distal occlusions beyond proximal M2, large established infarct, severe coma, minor stroke outside the NIHSS range, anticoagulated patients with abnormal coagulation tests, or patients at high bleeding risk.
• Applicability: The neutral result does not negate late-window thrombolysis in non-EVT contexts; it specifically tests whether adding tenecteplase to immediate EVT provides incremental benefit.
• Conclusion on External Validity: External validity is moderate but narrow. TNK-PLUS is directly relevant to late-window proximal MCA occlusion patients presenting to EVT-capable Chinese centres with advanced imaging selection; it is less generalisable to transferred patients, non-Chinese populations, posterior circulation stroke, non-LVO stroke, and resource-limited settings without immediate EVT.

Strengths & Limitations

• Strengths: Randomised phase 3 design; multicentre recruitment; central allocation; blinded endpoint assessment; blinded imaging core laboratory; blinded clinical event adjudication; complete primary outcome follow-up.
• Strengths: Clinically precise question: tenecteplase before EVT versus direct EVT in late-window proximal MCA occlusion with salvageable tissue.
• Strengths: EVT quality was high, with final post hoc successful reperfusion after EVT of 95.2% versus 90.8%.
• Strengths: The adaptive sample-size plan was transparent; the interim conditional power result led to no sample-size expansion rather than data-driven early termination.
• Limitations: Open-label treatment could affect procedural behaviour, ancillary therapies, or thresholds for post-procedural management, although blinded outcome assessment mitigates this.
• Limitations: The achieved absolute primary outcome difference was only 0.99%, far below the trial’s 10% to 14% design effect size assumptions.
• Limitations: The study was underpowered for uncommon but critical harms such as symptomatic intracranial haemorrhage.
• Limitations: The short tenecteplase dwell time before arterial puncture, median 26 minutes, may be insufficient for meaningful thrombus dissolution in late-window clot.
• Limitations: Missing 24-hour perfusion imaging in 104 patients reduces confidence in secondary imaging endpoints.
• Limitations: Generalisability is limited by the all-China setting, almost entirely Han Chinese population, high intracranial atherosclerosis burden, and high rates of adjunctive tirofiban and angioplasty/stenting.

Interpretation & Why It Matters

• Clinical meaning

  For late-window proximal MCA occlusion patients who are already in an EVT-capable centre, TNK-PLUS does not support routine intravenous tenecteplase immediately before thrombectomy.
• Mechanistic meaning

  The trial suggests that a very short pre-EVT thrombolytic exposure in the late window may not provide enough additional reperfusion benefit to improve 90-day disability.
• Practice implication

  The priority should remain fast EVT. Giving tenecteplase should not delay arterial puncture in this population, and the data do not justify thrombolysis as an automatic add-on for direct late-window EVT patients.
• What it refines

  TNK-PLUS narrows the extended-window tenecteplase story: benefit may depend on whether EVT is unavailable, delayed, or immediate, rather than simply on the presence of salvageable tissue.

Controversies & Subsequent Evidence

• Early-window and late-window bridging should not be conflated: BRIDGE-TNK, conducted within 4.5 hours, found 90-day functional independence in 147/278 (52.9%) with tenecteplase before thrombectomy versus 120/272 (44.1%) with thrombectomy alone, unadjusted RR 1.20; 95% CI 1.01 to 1.43; P=0.04. TNK-PLUS tested a much later population, with median onset-to-randomisation around 10 hours and median tenecteplase-to-puncture time 26 minutes.³
• TIMELESS subgroup uncertainty was not confirmed: TIMELESS was neutral overall in late-window perfusion-selected patients, most of whom underwent EVT, and its MCA-M1 signal was exploratory. TNK-PLUS prospectively tested a proximal MCA EVT-capable population and still found no functional benefit.⁵
• Transfer patients remain unresolved: ETERNAL-LVO, an Australian late-window trial, was terminated early and found no primary benefit overall, mRS 0–1 or return to baseline 44/120 (37%) with tenecteplase versus 52/122 (43%) with standard care, adjusted RR 0.90; 95% CI 0.66 to 1.21; P=0.48. A planned per-protocol transfer subgroup suggested a possible mRS shift benefit, OR 2.61; 95% CI 1.07 to 6.40, supporting further study in patients with delayed access to EVT rather than direct presenters.⁶
• Non-EVT late-window thrombolysis is different: TRACE-III showed benefit where EVT was not available, and OPTION later showed benefit in non-large-vessel-occlusion stroke, with mRS 0–1 at 90 days in 123/282 (43.6%) with tenecteplase versus 97/284 (34.2%) with standard care, RR 1.28; 95% CI 1.04 to 1.57; P=.02, but sICH was 2.8% versus 0%.⁴⁷
• Time dependency is central: The JAMA IRIS time-to-treatment meta-analysis found that benefit associated with intravenous thrombolysis before thrombectomy was time dependent, statistically significant only when expected IVT administration occurred early, and no longer statistically significant after 2 hours 20 minutes from onset.⁸
• Meta-analytic context is moving quickly: A 2026 meta-analysis of tenecteplase 4.5 to 24 hours concluded that tenecteplase improves excellent functional outcome and recanalisation in selected extended-window patients, with stronger rationale when EVT is inaccessible. TNK-PLUS adds an important negative data point for the immediate-EVT subgroup rather than for all extended-window tenecteplase use.⁹
• Safety uncertainty persists: TNK-PLUS showed numerically higher sICH, 5.1% versus 2.6%, but the CI was wide. This is compatible with either a modest bleeding penalty or no true difference; larger pooled analyses will be needed to define the late-window bridging harm profile.
• Reporting precision matters: The apparent safety-population denominator inconsistency between the flow diagram and the safety outcome table is unlikely to change the interpretation, but it is an avoidable reporting weakness in an otherwise well-reported trial.

Summary

• TNK-PLUS randomised 391 late-window patients with MCA-M1 or proximal M2 occlusion, salvageable tissue, and direct access to EVT.
• Tenecteplase before EVT did not improve 90-day functional independence: 44.2% versus 43.2%, adjusted RR 1.01; 95% CI 0.83 to 1.24; P=.89.
• No secondary clinical endpoint showed meaningful benefit, and prespecified subgroup analyses did not show convincing treatment-effect heterogeneity.
• Symptomatic intracranial haemorrhage was numerically higher with tenecteplase, 5.1% versus 2.6%, but the trial was underpowered for rare safety outcomes.
• The result is best interpreted as evidence against routine late-window bridging tenecteplase for direct-to-EVT proximal MCA occlusion, not against all extended-window tenecteplase strategies.

Overall Takeaway

TNK-PLUS is an important boundary-setting trial rather than a practice-changing positive landmark. It shows that in late-window proximal MCA occlusion patients already at an EVT-capable centre, adding intravenous tenecteplase immediately before thrombectomy does not improve 90-day functional independence and may add uncertain haemorrhagic risk.

Bibliography

• 1.Prabhakaran S, Gonzalez NR, Zachrison KS, et al. 2026 Guideline for the early management of patients with acute ischemic stroke: a guideline from the American Heart Association/American Stroke Association. Stroke. Published online January 26, 2026.
• 2.Majoie CB, Cavalcante F, Gralla J, et al; IRIS Collaborators. Value of intravenous thrombolysis in endovascular treatment for large-vessel anterior circulation stroke: individual participant data meta-analysis of six randomised trials. Lancet. 2023;402(10406):965-974.
• 3.Qiu Z, Li F, Sang H, et al; BRIDGE-TNK Trial Investigators. Intravenous tenecteplase before thrombectomy in stroke. N Engl J Med. 2025;393(2):139-150.
• 4.Xiong Y, Campbell BCV, Schwamm LH, et al; TRACE-III Investigators. Tenecteplase for ischemic stroke at 4.5 to 24 hours without thrombectomy. N Engl J Med. 2024;391(3):203-212.
• 5.Albers GW, Jumaa M, Purdon B, et al; TIMELESS Investigators. Tenecteplase for stroke at 4.5 to 24 hours with perfusion-imaging selection. N Engl J Med. 2024;390(8):701-711.
• 6.Yogendrakumar V, Campbell BCV, Churilov L, et al; ETERNAL-LVO Investigators. Efficacy of tenecteplase in large vessel occlusion stroke within 24 hours of symptom onset: the ETERNAL-LVO randomized controlled trial. Stroke. 2025;56(12):3332-3341.
• 7.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. 2026;335(13):1137-1147.
• 8.Kaesmacher J, Cavalcante F, Kappelhof M, et al; IRIS Collaborators. Time to treatment with intravenous thrombolysis before thrombectomy and functional outcomes in acute ischemic stroke: a meta-analysis. JAMA. 2024;331(9):764-777.
• 9.Wang Z, Li J, Wang X, Yuan B, Li J, Ma Q. Tenecteplase for acute ischemic stroke at 4.5 to 24 hours: a meta-analysis of randomized controlled trials. Stroke. 2026;57(1):50-62.