PDF: MR WITNESS
(2018)Objective
MR WITNESS: To assess the safety of IV tPA in patients with wake-up stroke who were selected using MRI criteria (DWI-positive, FLAIR-negative) to approximate stroke onset within 4.5 hours.
Study Summary
• sICH occurred in only 1.9% of treated patients
• 39% achieved mRS 0–1 at 90 days, consistent with standard tPA outcomes
Intervention
Open-label, single-arm, phase 2a trial of 80 wake-up stroke patients treated with IV alteplase within 4.5 hours of DWI lesion detection, based on MRI mismatch (DWI-positive, FLAIR-negative).
Inclusion Criteria
• Wake-up stroke with last known well >4.5 hours
• Age ≥18 years
• NIHSS 2–25
• DWI-positive, FLAIR-negative lesion ≤70 mL
• Within 4.5 hours of MRI
Study Design
Arms: Single-arm tPA group (no control)
Patients per Arm: 80
Outcome
• 90-day mRS 0–1: 39% (95% CI 28.3–50.3%)
• Mortality at 90 days: 4%
Bottom Line
IV alteplase in qDFM-selected unwitnessed stroke was safe: sICH rate 1.3% (1/80; 95% CI 0.0-6.8%), not significantly different from ECASS-3 benchmark of 5.3% (RR=0.24; P=0.07). At 90 days, 38.8% achieved mRS 0-1 overall, 43.5% without pre-stroke disability, and 48% in non-LVO patients. Median time from last known well to treatment was 11.24 hours. This phase 2a safety trial supported larger RCTs (WAKE-UP) of MRI-guided thrombolysis.
Major Points
- sICH rate 1.3% (1/80) — well below the 5.3% ECASS-3 benchmark (RR=0.24; P=0.07).
- Median time from last known well to treatment was 11.24h (IQR 9.46-13.26), far beyond the standard 4.5h guideline window.
- qDFM doubled potential enrollment vs qualitative DFM: 50% of enrolled were FLAIR-positive but SIR <1.15 (quantitative threshold), which would have been excluded by qualitative assessment.
- 38.8% achieved mRS 0-1 at 90 days; 43.5% among those without pre-stroke disability; 48.1% in non-LVO subgroup — comparable to witnessed stroke thrombolysis trials.
- 71.3% had wakeup stroke (most common unwitnessed subtype); wakeup-to-non-wakeup ratio 2.5:1.
- No lacunar patients developed aICH at 24h (0/11 vs 21/69 non-lacunar; P<0.001).
- Higher NIHSS independently predicted both worse outcome (OR 0.76/point; P<0.001) and aICH (OR 1.22/point; P=0.001).
- 22.9% had ICA/M1 LVO; non-LVO patients had significantly better outcomes (48.1% vs 18.8% mRS 0-1; P=0.045).
- Symptomatic brain edema 3.8% (3/80), less than ECASS-3 rate of 6.9% (RR=0.54; P=0.19).
- Phase 2a single-arm safety trial (n=80) across 14 US sites; supported rationale for WAKE-UP and subsequent MRI-guided thrombolysis trials.
Study Design
- Study Type
- Phase 2a, open-label, single-arm safety trial
- Randomization
- No
- Blinding
- Open-label; independent medical monitor and neuroradiology core adjudicated all ICH and brain edema events.
- Sample Size
- 80
- Follow-up
- 90 ± 14 days
- Centers
- 14
- Countries
- United States
Primary Outcome
Definition: Symptomatic ICH (ECASS-2 definition: any brain blood + NIHSS ≥4 point increase or death)
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| ECASS-3 benchmark: 22/418 (5.3%) | 1/80 (1.3%; 95% CI 0.0-6.8%) | - (0.0-6.8%) | 0.07 (RR=0.24 vs ECASS-3) |
Limitations & Criticisms
- No concurrent control arm — single-arm phase 2a safety study; cannot conclude efficacy.
- Small sample size (n=80) — powered for safety only.
- Open-label design with potential for assessment bias.
- MRI-based selection limits generalizability — MRI less available than CT.
- Slow enrollment (~3 patients/year per site) — feasibility questions.
- LVO not excluded; majority (77.1%) were non-LVO — limits applicability to EVT-eligible patients.
- Comparison to ECASS-3 is indirect (different population: witnessed, known onset 3-4.5h).
- Median NIHSS 7 (lower than ECASS-3 median 9) — qDFM may exclude more severe strokes, improving safety profile.
- No perfusion imaging for selection — cannot assess mismatch influence.
Citation
Ann Neurol. 2018;83(5):980-993.