DEVT Trial: Direct EVT vs Alteplase Plus EVT
In alteplase-eligible patients with proximal anterior circulation LVO presenting within 4.5 hours, is direct endovascular thrombectomy non-inferior to IV alteplase followed by thrombectomy for 90-day functional independence (mRS 0-2)?
Zi et al. (JAMA 2021) · doi:10.1001/jama.2020.23523 · 234 patients
Population
Included
- Age 18 years or older
- Acute ischemic stroke with anterior circulation proximal large vessel occlusion (ICA or M1)
- Eligible for both IV alteplase and EVT
- Treatment initiable within 4.5 hours of symptom onset
- Pre-stroke mRS 0 or 1
- Age ≥18 years
- Acute ischemic stroke eligible for IV alteplase, treated within 4.5 hours of symptom onset
- Occlusion of the intracranial internal carotid artery or first segment of the middle cerebral artery (M1) confirmed on CT angiography or MR angiography
- Randomization no later than 4 hours 15 minutes from onset (onset = last known well)
Excluded
- Contraindication to IV alteplase
- Posterior circulation or distal M2 occlusion
- Pre-existing functional disability
- Anticipated EVT initiation beyond 4.5-hour window
- Severe comorbidity limiting 90-day follow-up
- Imaging evidence of intracranial hemorrhage
- Pre-morbid modified Rankin Scale score ≥2
- Contraindication to IV recombinant tissue plasminogen activator (rt-PA)
- IV thrombolytic delivered outside the hospital
- Recent cardiac surgery
- Advanced cancer
Source: Zi et al., DEVT, JAMA 2021;325(3):234-243; reg. ChiCTR-IOR-17013568· Retrieved 2026-06-08
Primary Outcome — mRS 0-2 at 90 Days (Non-inferiority)
All randomized patients (proximal anterior circulation LVO, alteplase-eligible)
Non-inferiority design: wide margin
Non-inferiority margin was -10 percentage points, clinically wide. Met NI but cannot exclude meaningful harm. The numerical advantage for direct EVT (54.3% vs 46.6%) is not superiority. Trial was stopped early.
mRS 0-2 at 90 Days
Study Arms
- Agent
- Mechanical thrombectomy without IV thrombolysis
- Route
- Endovascular (intra-arterial)
- Co-interventions
- Rapid endovascular treatment only, no IV alteplase; operator could use stent retrievers, thromboaspiration, intra-arterial thrombolysis, balloon angioplasty, or stenting, alone or combined, at the treatment team's discretion.
Methods/Intervention, Zi JAMA 2021 p.235-236. Open-label, blinded outcome; goal eTICI 2b-3. Intra-arterial thrombolytic permitted as part of the endovascular technique.
- Agent
- IV alteplase, then mechanical thrombectomy
- Dose
- 0.9 mg/kg (10% bolus, remainder over 1-hour infusion)
- Route
- IV alteplase, then endovascular thrombectomy
- Co-interventions
- Standard-dose IV alteplase 0.9 mg/kg started as soon as possible; endovascular procedure initiated without waiting for clinical response; full alteplase dose infused even if reperfusion already established.
Methods/Intervention, Zi JAMA 2021 p.235-236. Same endovascular technique options as the direct-EVT arm.
Trial Design
Type
- Multicenter randomized noninferiority trial
- Direct thrombectomy vs alteplase plus thrombectomy
- Proximal anterior circulation occlusion within 4.5 hours
- Conducted at 33 stroke centers in China
Timeline
Enrolled 2018-2020; stopped early for efficacy
N
234
Enrollment
Multicenter Chinese noninferiority RCT enrolling 234 patients at 33 stroke centers between 2018 and 2020 (Zi JAMA 2021). Stopped early for efficacy of non-inferiority.
Bedside Pearl
DEVT met its non-inferiority threshold but at a -10 pp margin and after early stopping. The numerical 7.7 pp benefit favoring direct EVT is not superiority; treat as hypothesis-generating. Continue IV thrombolysis per guidelines unless local data and workflow specifically support omission.
See also