Comment
Testing early haemostatic therapy for acute intracerebral haemorrhage
Craig S Anderson, Chen Chen, Rustam Al-Shahi Salman
Lancet Available online 4 February 2026
https://doi.org/10.1016/S0140-6736(26)00146-7
The clinical severity, narrow time window to reduce bleeding, and variable care pathways for patients with spontaneous (non-traumatic) intracerebral haemorrhage (ICH) are enormous challenges to be overcome by treatments for this serious condition. In this issue of The Lancet, Joseph P Broderick and colleagues1 present the results of the FASTEST trial, a monumental effort to establish the safety and efficacy of the potent haemostatic agent, intravenous recombinant factor VIIa, administered to patients within 2 h of the onset of ICH. This multicentre, double-blind, randomised, placebo-controlled, adaptive, phase 3 trial was conducted through an international network of hospitals and mobile stroke units across six countries between 2021 and 2025. Unfortunately, after the first 626 of a planned 860 patients were randomly assigned to either recombinant factor VIIa or an identical placebo at the time of the second pre-planned interim analysis, futility signals in the first 434 patients with available data on the primary outcome of functional recovery at 180 days prompted the steering committee to terminate the trial. Overall, there was no benefit of recombinant factor VIIa on either the primary outcome (odds ratio 1·09 [95% CI 0·79–1·51]) or secondary outcomes, despite a significant reduction in ICH expansion with recombinant factor VIIa (–3·68 mL [95% CI –5·40 to –1·94]; p=0·0011). However, non-significant signals emerged in small, prespecified subgroups of patients defined by time to treatment (<90 min) and the presence of a spot sign indicating active bleeding on CT angiography. FASTEST, which was itself based on post-hoc subgroup analyses of the earlier FAST trial,2, 3 reminds us to be cautious when seeking to substantiate the effects seen in patient subgroups and to be wary when hoping that significant effects on plausible mechanistic surrogates will improve functional outcomes.
The rationale behind the use of recombinant factor VIIa for patients with ICH was based on established safety and efficacy in patients with pre-existing coagulation disorders, with an initial proof-of-concept trial showing that early treatment reduced ICH haematoma expansion.4 However, a subsequent international, phase 3, randomised controlled trial showed no effect of recombinant factor VIIa on the functional outcomes of patients when administered within 4 h after the onset of ICH, despite further evidence of a reduction in ICH expansion.2 Furthermore, recombinant factor VIIa produced an excess of serious thromboembolic complications that, alongside the neutral effects on all outcomes in a meta-analysis of all trials to date,5 prompted guidelines to recommend against the use of this agent for ICH.6 Despite these setbacks, post-hoc modelling of data from the FAST trial identified several subgroups of patients who might benefit from recombinant factor VIIa,3 constituting a patient population of potential responders to target in the FASTEST trial.1
Given that clinical trials are a major investment by investigators, sponsors, and participants, it is appropriate to gain as much information as possible from them. This information typically involves analyses to establish whether the average effect of the treatment varies more (or less) to the benefit (or harm) of patients. However, careful thought must be applied to the interpretation of such data because it is difficult to disentangle the influence of low statistical power, multiple testing, exposure interactions, confounding, and chance. One solution is to only accept such results as hypothesis-generating from analyses of prespecified subgroups, and where there are biological explanations and data relevant for guiding clinical decisions. Ultimately, however, large clinical trials and meta-analyses of individual patient data are required to provide reliable estimates of effects overall and by patient subgroups. For the use of early intensive blood pressure lowering in patients with acute ICH, a clear interaction between time from symptom onset to randomisation and the effect of recombinant factor VIIa only became apparent after data were pooled from 11 312 patients who participated in five clinical trials conducted over the past 20 years.7
FASTEST continues to investigate the benefit of ultra-early recombinant factor VIIa by recruiting its patient population of potential responders within 90 min of symptom onset or within 2 h if a spot sign is identified. If ultra-early recombinant factor VIIa is shown to benefit individuals with active bleeding, we will finally have evidence that limiting haematoma expansion with a haemostatic agent improves functional outcomes in this patient group. However, the generalisability of the population will be minimal worldwide unless code ICH care pathways can identify people with hyperacute ICH in the ambulance and convey them rapidly to hospital, where CT angiography will need to be available (as it is for acute ischaemic stroke due to a large-vessel occlusion), as well as fast-care bundles to effectively control elevated blood pressure and other physiological parameters.8
