AI-generated research brief — verify at source
Direct Angiography Transfer Raises Haemorrhage Risk With No Functional Benefit in Severe Stroke
In a randomised trial stopped early for safety reasons, bypassing pre-treatment imaging and transferring suspected large vessel occlusion stroke patients directly to the angiography suite was associated with an approximately elevenfold higher risk of symptomatic intracranial haemorrhage compared with the conventional imaging-first pathway (15% vs 0%; adjusted OR 11.0, 95% CI 1.28–1406). Functional independence at 90 days was not improved and trended in the wrong direction (36% vs 42%; adjusted OR 0.73, 95% CI 0.32–1.69), making this a clearly negative trial for the direct transfer strategy.
What Was Studied
The trial examined whether transferring patients with acute severe neurological deficits highly suggestive of large vessel occlusion ischaemic stroke directly to the angiography suite—without prior cross-sectional imaging—could accelerate in-hospital workflow sufficiently to improve functional outcomes. The primary endpoint was functional independence, defined as a modified Rankin Scale score of 0 to 2, assessed at 90 days post-stroke.
How It Was Studied
DIRECT ANGIO was an open-label, blinded-endpoint, multicentre randomised controlled trial conducted across ten comprehensive stroke centres in France. Eligible patients were adults aged 85 years or younger presenting within five hours of symptom onset with a unilateral motor deficit scoring 5 or above plus a cortical symptom scoring 1 or above on the National Institutes of Health Stroke Scale, a profile considered highly predictive of large vessel occlusion. A total of 115 patients were randomly assigned 1:1 to either direct transfer to the angiography suite (n=57) or the conventional pathway—imaging first, followed by angiography suite transfer if eligible (n=58). Randomisation used a web-based system, and all outcomes were assessed on an intention-to-treat basis. Enrolment ran from July 2020 to April 2023, with the trial steering committee voting to halt the study permanently in December 2023 after a pre-planned interim analysis revealed a safety signal.
What Was Observed
- Symptomatic intracranial haemorrhage was substantially more frequent in the direct transfer group. Five of 34 patients (15%) in the DTAS group experienced symptomatic intracranial haemorrhage compared with none of 42 patients (0%) in the conventional group—roughly an elevenfold higher adjusted odds (adjusted OR 11.0, 95% CI 1.28–1406). The very wide confidence interval reflects the small number of events and the trial’s premature closure.
- Functional independence at 90 days was not improved by direct transfer and showed a non-significant trend toward worse outcomes. Twenty of 56 DTAS participants (36%) achieved a modified Rankin Scale score of 0–2, versus 22 of 53 conventional pathway participants (42%), representing about a 27% lower adjusted odds of independence that did not reach statistical significance (adjusted OR 0.73, 95% CI 0.32–1.69).
- All-cause mortality at 90 days did not differ significantly between groups. Death occurred in 10 of 56 DTAS patients (18%) versus 6 of 53 conventional patients (11%), an approximately 65% higher adjusted odds in the DTAS group that was not statistically significant given the imprecise estimate (adjusted OR 1.65, 95% CI 0.52–5.55).
Why This Matters
The hypothesis underlying DTAS was that eliminating the imaging step would compress door-to-puncture times and translate into better neurological recovery. These findings challenge that premise directly, suggesting that omitting pre-procedural imaging may instead introduce clinically serious haemorrhagic risk—potentially by treating patients who do not have ischaemic stroke or who harbour contraindications only detectable through imaging. The trial’s early termination means definitive conclusions cannot yet be drawn, and the authors themselves call for further clinical trials before the safety and efficacy of DTAS can be firmly established.
How to Read This Result
Although the haemorrhage signal is statistically significant, this trial was stopped early with only 115 participants enrolled, which substantially limits the precision of all effect estimates and the ability to generalise the findings; the result should be treated as an important safety warning requiring replication rather than a conclusive verdict on the DTAS strategy.
Limitations
The most consequential limitation is early termination for safety reasons, which left the trial with a sample size of only 115 patients—far below what would be needed to precisely estimate effects on the primary functional outcome or secondary endpoints. Confidence intervals are correspondingly wide, particularly for the haemorrhage odds ratio, where the upper bound exceeds 1400. The open-label design introduces the possibility of performance bias, as clinical teams were aware of group assignment during treatment delivery, even though outcome assessment was blinded.