OPENS
(2026)Objective
Assess whether adjunctive normobaric hyperoxia (NBO) added to endovascular thrombectomy (EVT) improves early neurological outcomes in patients with anterior-circulation large-vessel occlusion presenting 6 to 24 hours after stroke onset.
Study Summary
• Median 24–48h infarct volume was significantly smaller with NBO+EVT (20.5 vs 32.3 mL; adjusted β 17.27, 95% CI 6.99–27.55; P=0.001).
• 90-day mRS numerically favored NBO+EVT (median 2 vs 3; adjusted common OR 1.52, 95% CI 0.87–2.63; P=0.154) but was not statistically significant; mortality, symptomatic ICH, and other safety events were similar between groups.
Intervention
Normobaric hyperoxia (100% O2 via face mask at 10 L/min for 4 hours starting before recanalization) as adjunct to endovascular thrombectomy.
Inclusion Criteria
Adults ≥18 y with anterior circulation LVO (terminal ICA or M1/M2), stroke onset 6–24 h, NIHSS ≥6, ASPECTS 6–10, imaging-confirmed ischemic penumbra, prestroke mRS ≤1.
Study Design
Arms: EVT + Normobaric Hyperoxia (100% O2 at 10 L/min x 4 h) vs EVT alone (room air)
Patients per Arm: 60 per arm (total N=120)
Outcome
• Secondary: 24–48h infarct volume smaller with NBO+EVT (20.5 vs 32.3 mL; P=0.001); 90-day mRS 0–2 53% vs 41% (adjusted OR 1.68, 95% CI 0.75–3.71; P=0.209)
• Safety: Mortality 7% vs 10% (P=0.416); sICH 5% vs 7% (P=0.321); pneumonia 37% vs 42% — no significant differences
Clinical Question
In patients with acute ischemic stroke due to anterior-circulation large-vessel occlusion treated 6–24 hours after onset, does adjunctive normobaric hyperoxia added to endovascular thrombectomy improve early neurological outcomes and reduce infarct volume compared with EVT alone?
Bottom Line
In this phase IIb trial of 120 patients with LVO stroke in the 6–24-hour window, 4 hours of 100% oxygen added to EVT significantly increased early neurological improvement (35% vs 19%; adjusted OR 2.86) and reduced 24–48h infarct volume (20.5 vs 32.3 mL; P=0.001) without safety concerns, though 90-day functional outcomes only trended toward benefit.
Major Points
- Early neurological improvement (≥30% NIHSS reduction at 24 h) was significantly higher with NBO+EVT: 35% (20/57) vs 19% (11/58); adjusted odds ratio 2.86 (95% CI 1.12–7.45); P=0.031.
- Median 24–48h infarct volume was significantly smaller with NBO+EVT (20.5 mL, IQR 13.6–31.8) vs EVT alone (32.3 mL, IQR 22.7–44.5); adjusted β 17.27 (95% CI 6.99–27.55); P=0.001.
- Post-therapy arterial pO2 was markedly higher with NBO (205 vs 99 mm Hg; P<0.001), confirming hyperoxia delivery, without a significant rise in pCO2.
- 90-day modified Rankin Scale distribution numerically favored NBO+EVT (median 2 vs 3; adjusted common OR 1.52, 95% CI 0.87–2.63; P=0.154) and mRS 0–2 was achieved by 53% vs 41% (adjusted OR 1.68, 95% CI 0.75–3.71; P=0.209), but neither reached statistical significance.
- Safety was comparable: mortality 7% vs 10%, symptomatic intracranial hemorrhage 5% vs 7%, any ICH 33% vs 35%, early neurological deterioration 5% vs 9%, and pneumonia 37% vs 42% — no significant differences.
- Consistent direction of benefit across prespecified subgroups (age, sex, ASPECTS, IV thrombolysis, occlusion location, baseline NIHSS) with no significant interactions.
- Findings extend prior OPENS-1 and OPENS-2 evidence (early window) to the 6–24-hour late window, supporting a larger multicenter trial (AN-O2-Trans; NCT06666764) currently underway.
Study Design
- Study Type
- Randomized Controlled Trial (Phase IIb)
- Randomization
- Yes
- Blinding
- Assessor-blinded (open-label treatment)
- Sample Size
- 120
- Follow-up
- 90 days
- Centers
- 2
- Countries
- China
Primary Outcome
Definition: Early neurological improvement (≥30% reduction in NIHSS from baseline to 24 hours)
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| 19% (11/58) | 35% (20/57) | 2.86 (1.12–7.45) | 0.031 |
Limitations & Criticisms
- Small phase IIb trial (n=120) at only 2 Chinese academic stroke centers, limiting generalizability and statistical power for functional outcomes.
- Assessor-blinded but not double-blind or sham-controlled; patients and treating clinicians could not be blinded to oxygen administration.
- Randomization was performed at the subcenter level with sealed envelopes rather than a centralized system, and center-specific quotas may have influenced enrollment dynamics.
- Patients with ASPECTS <6 were excluded, so results do not extend to those with larger baseline ischemic cores.
- Primary endpoint was changed mid-trial (from imaging-based infarct volume to early neurological improvement) after ~one-third of patients had enrolled, though this occurred before database lock and unblinding.
- 90-day functional outcomes (mRS distribution, mRS 0–2) only trended toward benefit and were not statistically significant, so clinical translation of the 24-hour benefit remains unproven.
- Findings from Chinese stroke centers with high thrombectomy volumes may not generalize to systems with different late-window imaging selection or workflow.
Citation
Stroke 2026;57(8):2265-2275