Fibrinogenase for injection in the treatment of acute ischaemic stroke: a randomised clinical trial — the Chinese snake-venom fibrinolytic that beat placebo on mRS at 90 days

What this is

A 235-patient, double-blind, placebo-controlled RCT run at 12 Chinese stroke centres testing a snake-venom-derived fibrinolytic — fibrinogenase for injection — against placebo in acute ischaemic stroke patients who did not receive IV thrombolysis or mechanical thrombectomy. The primary endpoint, the ordinal mRS distribution at 90 days, shifted significantly toward better outcomes with fibrinogenase (OR 2.38; p<0.001), with the largest effect in patients randomised within 48 hours and those with mild-to-moderate stroke due to large-artery atherosclerosis or small-vessel disease. Bleeding and mortality were indistinguishable from placebo.

1. Shadow Audit

The headline number — OR 2.38 on the ordinal mRS — is striking, but the clinically intuitive endpoint (mRS 0–2 at 90 days, “functional independence”) did not reach conventional significance: 80.3% vs 70.7%, OR 1.69 (95% CI 0.92–3.11; p=0.088). The entire positive result sits inside the ordinal shift: patients shifted one notch on a 7-point scale. That is a real signal, but it is not the same as “10% more patients walked out of the hospital independent.” The 90-day NIHSS, BI, and EQ-5D differences are consistent — secondary endpoints triangulate — but the binary “good outcome” measure sits at p=0.09. Frame the benefit as an ordinal shift, not as a stroke-cure.

The other quiet caveat: the population enrolled skewed toward mild-to-moderate stroke (median NIHSS 5–6) despite an inclusion window of NIHSS 4–25. The authors acknowledge this. Translation: the headline number is built on patients who were already going to do reasonably well. The trial cannot tell us whether fibrinogenase helps the NIHSS-20-plus patient who didn’t get thrombectomy.

2. Inversion Engine

For the opposite conclusion to hold — that fibrinogenase is no better than placebo — what would have to be true?

1. The ordinal mRS shift would have to be a placebo-by-chance artefact. Probability under H₀: p<0.001 means a false-positive rate of 0.1%; not impossible but unlikely. The 90-day NIHSS (OR 1.95, p<0.001) and EQ-5D/Barthel improvements are independent replications of the same signal — three endpoints in lockstep reduce the false-positive probability further.
2. However, the mRS 0–2 binary endpoint at p=0.088 is the soft underbelly. To flip the conclusion, a larger trial would need to show that the ~10 percentage-point gap on binary functional independence either narrows to <5% or widens beyond statistical significance in either direction. The current trial is underpowered to settle that.
3. The trial would have to be invalidated by baseline imbalance. The 16-percentage-point sex gap (53% male fibrinogenase vs 69% male placebo) is the largest single concern. If male sex independently predicts better 90-day mRS (it doesn’t, robustly, in stroke trials — but it’s a possible confounder), the treatment effect would be overstated. The adjusted ordinal regression partially addresses this but residual confounding remains.

The minimum trial to invert this finding: a 600+ patient, sex-stratified RCT with blinded outcome adjudication would be needed to move p=0.088 to significance on the binary endpoint. Such a trial (ChiCTR2500102258 in post-MT patients) is reportedly underway — but it’s a different population.

3. Second-Order Catalyst

If this result is real, who changes practice first?

Within China (fast): Chinese stroke neurologists working in secondary and tertiary centres without 24/7 thrombectomy capacity. Fibrinogenase for injection is already approved in China for this indication — so the regulatory barrier is gone. The trial adds a placebo-controlled evidence base to a previously empirical practice. Expect guideline updates in Chinese stroke society recommendations within 6–12 months; expect off-label use in tier-3 county hospitals where IV alteplase/tenecteplase is unavailable within 4.5 h and thrombectomy is hours away.

Outside China (slow): Western regulatory bodies (FDA, EMA) will not adopt this on a 235-patient single-country RCT. The mechanism — a snake-venom metalloproteinase acting on thrombus precursor protein — is novel and would require IND-enabling toxicology and a Western-population RCT (the ongoing ChiCTR2500102258 in post-MT patients may not satisfy that). Expect: (a) academic interest at stroke conferences in 2026–2027; (b) no practice change in the US/EU for 3–5 years; (c) potential Chinese-sponsored multinational Phase III if a Western partner emerges.

Patient subgroup that moves first: the elderly Chinese patient presenting at 24–48 h from LAA stroke to a hospital without thrombectomy, who currently receives only antiplatelet + statin + rehab. That subgroup has the largest absolute expected benefit, and the treatment is already commercially available.

4. Asymmetric Leverage

The asymmetric payoff here is mild stroke, large denominator. The enrolled population is overwhelmingly NIHSS 4–10 (median 5–6), which is the highest-volume AIS subgroup globally. A 10-percentage-point shift on the binary functional-independence endpoint across millions of mild-stroke patients per year is a large absolute population effect — even if the per-patient effect is modest. Compare this to the typical thrombectomy benefit (one additional independent outcome per 4–7 LVOs treated) — fibrinogenase has a smaller per-patient effect but applies to a denominator 5–10× larger.

Where the asymmetry bites: in low-resource settings without thrombectomy. India, sub-Saharan Africa, rural Southeast Asia, much of Latin America. There, the fibrinogenase-vs-placebo signal may be the only acute reperfusion-adjacent therapy available — and an OR 2.38 on ordinal mRS at $30–100/dose could be cost-effective even at half the observed effect size.

Where it doesn’t apply: in systems with mature thrombectomy (US, EU, Japan, Korea, urban Australia, urban China). There, fibrinogenase will be either irrelevant or, in the post-MT setting, untested.

5. Paradigm Destroyer

This paper kills the reflex that “missed the tPA window = nothing to offer acutely.” For 24–48 h presenters who don’t reach thrombectomy, the new reflex should be: “Is there an LAA or SAO mechanism? NIHSS 4–10? Within 48 h? Fibrinogenase is approved in China — order it if available.”

Two-sentence protocol change for tomorrow morning (in Chinese hospitals with access to fibrinogenase):

For acute ischaemic stroke patients presenting 4.5–48 h from onset who are not candidates for IV thrombolysis or mechanical thrombectomy, with NIHSS 4–10 and TOAST subtype large-artery atherosclerosis or small-artery occlusion, consider fibrinogenase for injection 1.0 IU IV on days 1, 3, and 5 per local protocol — pending shared decision-making and bleeding-risk review.

For US/EU practice: this paper is not yet actionable. Wait for the post-MT RCT (ChiCTR2500102258) and a Western-population trial before changing protocols.

MVP — Minimum Viable Proof

The minimum statement that, if true, would change bedside practice:

In acute ischaemic stroke patients not eligible for IV thrombolysis or mechanical thrombectomy, presenting within 48 h of onset with mild-to-moderate deficit (NIHSS 4–10) and large-artery or small-artery aetiology, fibrinogenase for injection produces a clinically meaningful improvement in 90-day functional outcome with bleeding risk indistinguishable from placebo.

This statement is directionally supported by the trial’s primary ordinal mRS result and subgroup signal, but the binary “functional independence” endpoint sits at p=0.088 — meaning the proof is suggestive, not definitive. A larger trial would be needed to lock it in.

Best Combination

Combine this finding with the WAKE-UP / EXTEND / WAKE-UP-style MRI-mismatch evidence for alteplase in the 4.5–9 h window (referenced as N Engl J Med 2019;380:1795–803 by the authors), and with tenecteplase in the 4.5–24 h window without thrombectomy (Xiong et al., NEJM 2024;391:203–12 — also referenced). The combined message: acute ischaemic stroke has an expanding menu of late-window reperfusion-adjacent therapies. Fibrinogenase sits in a unique niche: post-window, no-thrombectomy, mechanism-targeted. It complements rather than competes with tenecteplase/alteplase.

Practically: for the patient who arrives at 18 h with DWI-positive stroke, no LVO on CTA, NIHSS 6 — fibrinogenase is one of the few evidence-supported options. Combined with intensive secondary prevention (dual antiplatelet per CHANCE-2/POINT, high-intensity statin, BP control), the package is the new standard-of-care in late-presenting Chinese AIS.

Overvalue Warning

Two specific things readers might overinterpret from this paper:

1. “Fibrinogenase replaces tenecteplase/alteplase.” It does not. Tenecteplase within 4.5 h (and 4.5–24 h per Xiong 2024) is the standard. Fibrinogenase has not been tested head-to-head against either. It occupies a different clinical niche (post-window, no-thrombectomy, specific TOAST subtypes), not the front-line acute window.

2. “The 70%→80% functional-independence gain is the headline.” It isn’t. The headline is the ordinal shift (OR 2.38, p<0.001). The 10-point binary gap is real but sits at p=0.088 — i.e. one in twelve chance it’s a false positive at the conventional threshold. Don’t quote the binary number without that qualifier.