Skip to main content

Trial Summaries & Critiques

TRACE-5

Image: Shutterstock / ChaNaWiT

Publication

  • Title: Tenecteplase versus standard medical treatment for basilar artery occlusion within 24 h (TRACE-5): a multicentre, prospective, randomised, open-label, blinded-endpoint, superiority, phase 3 trial
  • Acronym: TRACE-5
  • Year: 2026
  • Journal published in: The Lancet
  • Citation: Xiong Y, Alemseged F, Cao Z, et al. Tenecteplase versus standard medical treatment for basilar artery occlusion within 24 h (TRACE-5): a multicentre, prospective, randomised, open-label, blinded-endpoint, superiority, phase 3 trial. Lancet. 2026; published online Feb 5.

Context & Rationale

  • Background
    • Basilar artery occlusion (BAO) is a rare but devastating large-vessel occlusion with very high mortality and severe disability without recanalisation.
    • Endovascular thrombectomy (EVT) improves outcomes for BAO up to 24 h in selected patients, but global access is limited; delays, transfer pathways, and affordability barriers mean many patients do not receive EVT.
    • Tenecteplase is operationally attractive (single bolus) and has accumulating evidence as an alternative to alteplase, including signals of improved early reperfusion before EVT in BAO.
    • Randomised evidence for IV thrombolysis specifically in imaging-confirmed BAO beyond 4·5 h and in pragmatic “EVT-variable” pathways has been scarce.
  • Research Question/Hypothesis
    In adults with near or complete BAO treated within 24 h, does IV tenecteplase (0·25 mg/kg; max 25 mg) improve 90-day excellent functional outcome compared with standard medical treatment (which could include alteplase within 4·5 h and usual antithrombotic strategies), with EVT permitted in both groups at clinician discretion?
  • Why This Matters
    Demonstrating effective and safe IV thrombolysis for BAO up to 24 h could broaden reperfusion access in settings with delayed, limited, or unaffordable EVT, and could potentially increase pre-EVT reperfusion in transfer pathways.

Design & Methods

  • Research Question: Does IV tenecteplase within 24 h improve 90-day excellent functional outcome after BAO compared with pragmatic standard medical treatment?
  • Study Type: Prospective, multicentre (66 stroke centres in China), randomised (1:1), open-label, blinded-endpoint, superiority, phase 3 trial.
  • Population:
    • Adults ≥18 years with posterior circulation ischaemic stroke symptoms due to near or complete BAO within 24 h of onset/last known well (or clinical deterioration/coma), confirmed on CTA or MRA.
    • BAO definition: potentially retrievable occlusion, near occlusion (99% stenosis with string sign) or complete occlusion.
    • Premorbid mRS ≤3.
    • Imaging exclusions: intracranial haemorrhage/tumour; pc-ASPECTS <6 on non-contrast CT/CTA source images or DWI; significant cerebellar mass effect or acute hydrocephalus; established frank hypodensity suggesting subacute infarction; bilateral extensive brainstem ischaemia.
  • Intervention:
    • Tenecteplase 0·25 mg/kg (maximum 25 mg) as a single IV bolus over 5–10 seconds immediately after randomisation.
  • Comparison:
    • Standard medical treatment, which could include alteplase 0·9 mg/kg (max 90 mg) within 4·5 h, anticoagulation (e.g., heparin infusion), or antiplatelets, with or without EVT at clinician discretion in both groups.
  • Blinding: Open-label treatment; outcome assessors and the Endpoint Adjudication Committee were masked to allocation.
  • Statistics: Sample size 452 to detect a 12% absolute increase in mRS 0–1 (33% vs 21%) with 80% power at one-sided α=0·025 (5% attrition). Primary and dichotomous outcomes analysed with modified Poisson regression (robust SE) adjusted for age, baseline NIHSS, and onset-to-randomisation time (<6 h vs 6–24 h). Ordinal mRS (5–6 merged) analysed via ordinal logistic regression (proportional odds assumption tested). Intention-to-treat for efficacy and safety; prespecified per-protocol analysis. No interim analyses.
  • Follow-Up Period: 90 days (mRS), with 24–36 h imaging for haemorrhage and 72 h early neurological endpoint.

Key Results

This trial was not stopped early. No interim analyses were planned or performed.

Outcome Tenecteplase (n=221) Standard medical treatment (n=231) Effect p value / 95% CI Notes
Primary: mRS 0–1 or return to baseline mRS at 90 days 83 (38%) 66 (29%) Adj RR 1.50 95% CI 1.09 to 2.08; P=0.014 Adjusted for age, baseline NIHSS, onset-to-randomisation time (<6 h vs 6–24 h)
mRS 0–2 or return to baseline mRS at 90 days 98 (44%) 99 (43%) Adj RR 1.16 95% CI 0.88 to 1.54; P=0.29 Secondary
mRS 0–3 at 90 days 114 (52%) 123 (53%) Adj RR 1.06 95% CI 0.82 to 1.37; P=0.66 Secondary
Ordinal mRS at 90 days (mRS 5–6 merged) 0: 36 (16%)
1: 43 (19%)
2: 17 (8%)
3: 18 (8%)
4: 25 (11%)
5–6: 82 (37%)		 0: 30 (13%)
1: 34 (15%)
2: 34 (15%)
3: 25 (11%)
4: 19 (8%)
5–6: 89 (39%)		 Adj common OR 1.51 95% CI 1.06 to 2.15; P=0.022 Proportional odds assumption satisfied (Brandt test P=0.40)
Early clinical improvement at 72 h (NIHSS ↓≥8 or NIHSS 0–1) 65 (29%) 64 (28%) Adj RR 1.08 95% CI 0.76 to 1.52; P=0.68 Secondary
Substantial reperfusion at initial angiogram (eTICI 2B–3)* 34/141 (24%) 16/126 (13%) Adj RR 1.90 95% CI 1.04 to 3.49; P=0.038 Among those proceeding to DSA; *complete baseline occlusion subgroup
Safety: Symptomatic intracranial haemorrhage within 36 h (SITS-MOST) 4 (2%) 7 (3%) Adj RR 0.58 95% CI 0.17 to 1.99; P=0.39 All 7 sICH in control group had EVT; 1 received alteplase before EVT
All-cause mortality within 90 days 65 (29%) 71 (31%) Adj RR 0.87 95% CI 0.62 to 1.22; P=0.41 Safety
mRS 5–6 at 90 days 82 (37%) 89 (39%) Adj RR 0.87 95% CI 0.65 to 1.18; P=0.38 Safety
Any intracranial haemorrhage (post-hoc classification) 26 (11.8%) 14 (6.1%) Adj RR 1.80 95% CI 0.93 to 3.46; P=0.079 Post-hoc; includes 4 ICH beyond 36 h1
  • Tenecteplase improved the prespecified “excellent outcome” endpoint (mRS 0–1/return to baseline) and shifted overall disability, but did not improve mRS 0–2 or mRS 0–3 dichotomies.
  • Tenecteplase increased substantial reperfusion before thrombectomy among those undergoing angiography (24% vs 13%).
  • Symptomatic intracranial haemorrhage and 90-day mortality were similar between groups.

Internal Validity

  • Randomisation and Allocation: Central web-based randomisation with allocation concealment; stratified by intention for EVT and intention for alteplase; minimisation balanced age (≤70 vs >70), NIHSS (<10 vs ≥10), and onset-to-randomisation time (<6 h vs 6–24 h).
  • Drop out or exclusions (post-randomisation): No missing primary/secondary outcomes. Protocol non-adherence occurred in 40/452 (9%), including crossover, alteplase use beyond 4·5 h, and inappropriate enrolment; prespecified per-protocol population was 210 vs 202 (tenecteplase vs control).1
  • Performance/Detection Bias: Open-label treatment introduces potential performance bias; mitigated by masked outcome assessment and masked endpoint adjudication.
  • Protocol Adherence / separation: Separation of the tested variable was clear (tenecteplase bolus vs no tenecteplase), but the comparator was intentionally heterogeneous: 80/231 (35%) received alteplase and other control therapies included antithrombotics and occasional heparin infusion.
  • Baseline Characteristics: Groups were broadly balanced; median NIHSS 13 (7–25) vs 11 (6–21), median pc-ASPECTS 8 (6–8) vs 8 (7–8); ~49% underwent EVT overall (112/221 vs 110/231).
  • Heterogeneity: Heterogeneous mechanisms (large-artery atherosclerosis predominated), time windows, and co-interventions (alteplase and EVT) likely diluted isolated drug effects but improved pragmatism.
  • Timing: Median onset-to-randomisation 6.1 h vs 6.5 h; subgroup signal for 6–24 h was larger than <6 h (RR 2.51 vs 1.16), but interaction was not statistically significant (P=0.08).
  • Separation of the Variable of Interest: EVT rates were similar (112/221 [50.7%] vs 110/231 [47.6%]); alteplase use occurred almost exclusively in control (80/231 [34.6%]) with 4 crossovers from tenecteplase group to control therapies (analysed ITT).
  • Adjunctive therapy use: Reasons for not proceeding to EVT differed: neurological improvement/recanalisation 32/109 (29.4%) vs 24/121 (19.8%); patient/family refusal 24/109 (22.0%) vs 42/121 (34.7%). This imbalance could influence functional outcomes through differential EVT exposure and is addressed in post-hoc sensitivity analyses.1
  • Statistical Rigor: Prespecified adjusted primary model was significant; post-hoc unadjusted ITT primary analysis was not (RR 1.31; 95% CI 0.95 to 1.82; P=0.10) and post-hoc adjustment for EVT refusal yielded RR 1.39; 95% CI 1.01 to 1.93; P=0.046, highlighting model sensitivity and post-randomisation complexity.1

Conclusion on Internal Validity: Moderate. Allocation concealment and blinded endpoint assessment are strengths, but open-label treatment, a heterogeneous comparator (including alteplase), and variable EVT delivery (including differential refusal) introduce performance and post-randomisation complexity that affect interpretability of a modest effect size.

External Validity

  • Population Representativeness: Chinese multicentre population; large-artery atherosclerosis was the predominant mechanism (86% vs 82%), which may differ from other settings and could influence thrombolysis responsiveness and EVT complexity.
  • Applicability: Highly relevant to systems with delayed/limited EVT access and transfer pathways, given single-bolus administration and lack of mandatory advanced imaging; generalisability to non-Asian populations and to centres with near-universal, rapid EVT is uncertain.
  • Intervention availability: Tenecteplase formulation used was the same as in TRACE-2; formal comparability across formulations is not established in the trial report.
  • Selection constraints: Excluded large baseline posterior circulation infarcts (pc-ASPECTS <6), significant mass effect/hydrocephalus, and extensive bilateral brainstem ischaemia; results do not apply to these higher-risk phenotypes.

Conclusion on External Validity: Moderate. Findings likely generalise to similar BAO populations within 24 h where EVT is delayed or variably delivered, but extrapolation beyond China, across stroke mechanisms, and into uniformly rapid EVT systems requires corroboration.

Strengths & Limitations

  • Strengths:
    • First phase 3 randomised evidence for tenecteplase in imaging-confirmed BAO up to 24 h with pragmatic co-interventions.
    • Central allocation concealment; blinded outcome assessment and adjudication; no missing primary/secondary outcomes.
    • Biologically plausible bridging signal (higher substantial reperfusion before EVT) aligned with clinical workflow relevance.
    • Safety profile reassuring for sICH and mortality in a high-risk vascular territory.
  • Limitations:
    • Open-label design and heterogeneous control arm (alteplase in 35%; variable antithrombotic strategies).
    • EVT delivered to only ~half despite EVT-capable sites; patient/family refusal was common and differed between groups, complicating causal attribution.
    • Primary benefit concentrated in “excellent outcome” and ordinal shift; broader dichotomies (mRS 0–2, 0–3) were neutral.
    • Generalisation outside China limited by differing stroke mechanisms, reimbursement structures, and tenecteplase formulations.

Interpretation & Why It Matters

  • Clinical implication
    In BAO within 24 h, tenecteplase increased the chance of excellent recovery (mRS 0–1/return to baseline) without excess sICH or mortality, supporting its use as a pragmatic reperfusion option when EVT is delayed, unavailable, or declined.
  • Bridging relevance
    Higher pre-thrombectomy substantial reperfusion supports tenecteplase as a plausible bridging therapy, particularly in transfer pathways with longer thrombolysis-to-puncture intervals.
  • Outcome nuance
    The benefit signal is strongest for “excellent outcome” and mRS shift, while functional independence (mRS 0–2) was unchanged; clinical adoption may hinge on how systems value excellent recovery vs independence thresholds in this severe-stroke syndrome.

Controversies & Subsequent Evidence

  • Comparator complexity and causal clarity: The “standard medical treatment” arm was intentionally heterogeneous (alteplase within 4·5 h, antiplatelets/anticoagulation), and EVT was discretionary in both arms; this improves pragmatism but reduces interpretability of tenecteplase versus alteplase specifically and of tenecteplase’s independent effect when EVT is reliably delivered.
  • EVT uptake and refusal: Only ~half received EVT; patient/family refusal contributed meaningfully, and was more frequent in the control group among those not proceeding to EVT (34.7% vs 22.0%), introducing a plausible post-randomisation pathway to differential outcomes and motivating post-hoc adjustment for refusal.1
  • Model dependence of primary benefit: The Lancet Comment highlighted that crude (unadjusted) ITT and per-protocol post-hoc analyses were not statistically significant for the primary and major secondary clinical endpoints, emphasising cautious interpretation of a modest effect dependent on the prespecified adjusted model.2
  • Generalisability and formulation: The Comment noted that tenecteplase was a biosimilar formulation not authorised in some regions and that ethnic/aetiologic differences may limit extrapolation outside China; it also raised concerns about variable recruitment across sites (many low-enrolling centres) and potential representativeness implications.2
  • Subsequent evidence trajectory: TRACE-5 sits alongside expanding evidence for tenecteplase in acute ischaemic stroke; ongoing BAO-focused work (e.g., POST-ETERNAL; planned IPD pooling) is intended to clarify tenecteplase’s role when EVT is intended and to refine selection in diverse populations.2

Summary

  • TRACE-5 randomised 452 patients with CTA/MRA-confirmed near or complete BAO within 24 h to tenecteplase 0·25 mg/kg bolus (max 25 mg) vs pragmatic standard medical treatment (including alteplase within 4·5 h), with EVT allowed in both arms.
  • Primary endpoint improved: mRS 0–1/return to baseline 38% vs 29% (Adj RR 1.50; 95% CI 1.09–2.08; P=0.014).
  • Overall disability shifted favourably (Adj common OR 1.51; 95% CI 1.06–2.15; P=0.022), but mRS 0–2 and mRS 0–3 dichotomies were neutral.
  • Substantial reperfusion before thrombectomy was higher with tenecteplase (24% vs 13%; Adj RR 1.90; 95% CI 1.04–3.49; P=0.038).
  • Safety was similar for sICH and mortality; post-hoc “any ICH” was numerically higher with tenecteplase (11.8% vs 6.1%; P=0.079).

Further Reading

Other Trials

Systematic Review & Meta Analysis

Observational Studies

Guidelines

Notes

  • *“Substantial reperfusion at initial angiogram” was assessed among patients with complete baseline occlusion undergoing DSA before thrombectomy; interpret as supportive mechanistic evidence rather than a whole-cohort efficacy endpoint.

Overall Takeaway

TRACE-5 is a landmark BAO thrombolysis trial because it tests single-bolus tenecteplase up to 24 h in an imaging-confirmed BAO population within a pragmatic, EVT-variable pathway. It demonstrated improved excellent functional outcome and an mRS shift without excess sICH or mortality, while highlighting the interpretive challenges of open-label care, heterogeneous comparators, and real-world EVT access and refusal.

Overall Summary

  • Tenecteplase within 24 h improved the prespecified excellent outcome endpoint (38% vs 29%) and shifted disability distribution, with similar sICH and mortality
  • Benefit did not extend to mRS 0–2 or mRS 0–3 dichotomies, and post-hoc analyses highlight sensitivity to model choice and EVT pathway factors
  • Most applicable where EVT is delayed, unavailable, or variably delivered, and where single-bolus thrombolysis can be used in transfer pathways

Bibliography