1. Shadow Audit

What this article is not saying — and the figure does not isolate — is that the ACTISAVE trial effectively ended the glenzocimab program in acute ischemic stroke. Three data points, each individually problematic, together form a trifecta that explains the negative result and forecloses on the development path. First, the primary endpoint ran in the wrong direction: 21.6% vs 15.3% poor outcome (OR 1.51, 95% CI 0.90–2.54, P = 0.120) — a numerically worse result that did not reach significance but definitively failed to show benefit. Second, the eTICI 2b–3 recanalization in the MT subgroup: 83.9% glenzocimab vs 96.6% placebo (OR 0.16, 95% CI 0.03–0.82) — a statistically significant impairment of the very process the drug was designed to protect. Third, the bleeding signal: 55 bleeding-related events other than ICH with glenzocimab versus 26 with placebo, with 6 major events versus 2. Together, these three data points establish that glenzocimab failed to improve outcomes, may have impaired recanalization, and carried a heavier bleeding burden than placebo — a trifecta that explains why the trial was negative and why future development requires fundamental reconsideration of patient selection, timing, and mechanism.

2. Inversion Engine

For the opposite conclusion to hold — that glenzocimab is beneficial in some subgroup — you would need a pre-specified subgroup analysis that returns a clinically meaningful effect size in the protective direction with a 95% CI that excludes 1.0. The trial’s MT subgroup analysis points the wrong way (0.16 for good recanalization), and the overall primary endpoint sits at 1.51 in the harmful direction (not significant, but the wrong way). To rescue glenzocimab, you would need a future trial restricting enrollment to a population in which GPVI antagonism helps rather than hinders — possibly patients treated with thrombectomy alone (no thrombolysis) or patients with specific clot compositions where fibrin-mediated platelet activation is the dominant mechanism. ACTIMIS’s positive signal came from a smaller, less-selected population; the expansion to a confirmatory phase 2/3 was the problem, not the drug.

3. Second-Order Catalyst

If this result is real and reproducible, the second-order effect is the shutdown of one of the few novel antiplatelet programs in acute stroke. Stroke neurology has spent two decades with essentially the same recanalization toolkit; ACTISAVE is the most recent high-profile attempt to add a non-thrombolytic, non-anticoagulant, non-mechanical therapy to the acute window. The negative result narrows the pipeline and emphasizes the field’s dependency on thrombectomy access and time metrics as the only modulators of acute outcome. Phase 2 programs exploring similar targets (GPVI, PAR-1, thrombin-receptor inhibitors) will face heightened skepticism in trial design and DSMB review.

4. Asymmetric Leverage

The asymmetric payoff here is not clinical — it is methodological. ACTISAVE demonstrates that a confirmatory trial can kill a program even when the primary endpoint is “not significant.” The 95% CI for the primary endpoint (0.90–2.54) crosses 1.0 by a wide margin, but the point estimate is in the opposite direction of the hypothesis. This is exactly the scenario where sponsors should stop, and ACTISAVE is a textbook example of how a clean negative result, properly designed, ends a development program without ambiguity. The leverage is in the design choices that made this possible: the binary mRS 4–6 outcome, the ITT analysis set, the PROBE-adjacent double-blind structure, and the prespecified MT subgroup analysis that surfaced the recanalization signal.

5. Paradigm Destroyer

What does this paper kill? It kills the assumption that adding any safe GPVI-targeting agent to standard recanalization therapy in acute ischemic stroke will improve outcomes. The mechanism that made glenzocimab attractive — platelet GPVI inhibition without global bleeding risk — was supposed to combine with thrombolysis to prevent reocclusion and microvascular dysfunction. The result says: not in this population, not with this timing, not with this design. It also softens the broader assumption that “non-bleeding antiplatelet” agents are uniformly safer than P2Y12 inhibitors in the acute stroke setting — the 6 vs 2 major bleeding events and the 55 vs 26 non-ICH bleeding events in ACTISAVE are not a clean safety story.

MVP — Minimum Viable Proof

For the bedside clinician: do not use glenzocimab outside a clinical trial. The drug is investigational; ACTISAVE failed to confirm benefit, and there is a signal toward impaired recanalization in the MT subgroup. Standard recanalization therapy (IV thrombolysis ± mechanical thrombectomy) remains the standard of care.

Best Combination

What works together from this analysis: the strongest interpretive combination pairs the eTICI recanalization data with the timing data. Glenzocimab was started a median of 1.2 hours after thrombolysis (IQR 0.8–1.6), meaning GPVI blockade was active during the critical window when thrombolytic agents were dissolving clot and mechanical thrombectomy was restoring flow. The 83.9% vs 96.6% eTICI 2b–3 difference suggests that GPVI antagonism during active fibrinolysis may destabilize the platelet–fibrin interface in ways that paradoxically hinder recanalization rather than prevent reocclusion. The second powerful combination is the ACTIMIS-to-ACTISAVE comparison: ACTIMIS showed a 39.5% vs 20.7% poor outcome rate in the MT subgroup (OR ~0.4), while ACTISAVE’s placebo arm achieved only 15.3% poor outcome — a 5-point improvement in standard-of-care outcomes that evaporated the expected treatment effect.

Overvalue Warning

First, do not over-interpret the point estimate of OR 1.51 as evidence of harm. The 95% CI (0.90–2.54) crosses 1.0, and the P value of 0.120 is well above any conventional threshold. The worse direction may be driven by baseline imbalances, particularly the 3-point higher median NIHSS in the glenzocimab MT subgroup and the higher rate of diabetes. Second, the bleeding signal (55 vs 26 non-ICH events) should not be over-interpreted as a safety catastrophe — only 8 were classified as major (6 vs 2), and the authors note the overall safety profile was “favorable.” Third, the 16 vs 10 death difference (7.6% vs 4.8%) should not be over-interpreted as a mortality signal — the authors explicitly state “there were no significant differences between groups in the analyses of all-cause mortality.” Fourth, the eTICI imbalance in the MT subgroup, while striking, involves a small subgroup with known baseline imbalances and should not be taken as definitive proof of mechanism without dedicated studies. Finally, the trial’s protocol modifications — from ordinal shift to binary mRS 4–6, from 1000 to ~400 patients — while conducted blinded, introduce uncertainty about whether the trial was optimally powered for its final design.

Educational use notice. The Clotbust Review is dedicated to medical education. Figures from published articles are embedded for teaching purposes, with attribution to the original publication. We do not claim copyright over any embedded figure; all rights remain with the original publisher and authors under the article's published license. If you are a rights holder and would like a figure removed, please contact the editorial team. See our editorial notice for the full policy.