Recently, two landmark studies supporting
an extended time window for the treatment
of large vessel occlusion (LVO) have been
published. DAWN, published in November
2017, investigated thrombectomy between
six and 24 hours from last known well time
(LKW) and used clinical severity to imaging
mismatch for the selection of patients. In
January, DEFUSE3 was published and found
benefit of thrombectomy for patients in the six to 16 hour window
using infarcted core to ischemic penumbra mismatch on imaging
alone. Both trials used RAPID software for automated volumetric
calculations. These studies led to a change in the American Heart
Association 2018 Guidelines for the Early Management of Patients
With Acute Ischemic Stroke.
DAWN NEJM Nov 2017
Summary: DAWN is a multinational, prospective, randomized,
open-label trial with a Bayesian adaptive-enrichment design
for the use of mechanical thrombectomy with TREVO device in
patients with LVO between six and 24 hours of last well-known
(LWK) time. Patients who were > 18 years with occlusion of
intracranial ICA, the proximal MCA or both on CTA or MRA,
+/- IV tPA, baseline mRS 0-1, no infarct > 1/3 of MCA on
baseline imaging and a target mismatch were randomized to
thrombectomy plus standard medical care or standard medical
care alone. Patients were classified into three groups based on
mismatch criteria: Group A: > 80 years, NIHSS > 10, infarct < 21
ml; Group B: < 80 years, NIHSS > 10, infarct < 31 ml; Group C: <
80 years, NIHSS > 20, infarct 31- <51 ml. Baseline infarct volume
was assessed with DWI or CT perfusion (rCBF < 30 percent
indicating core) using automated RAPID software. Randomization
was 1:1, with block minimization and stratification based on
mismatch group, time from LWK (6-12 or > 12-24 hours) and site
of occlusion. Rescue interventions and stenting were not allowed,
but carotid angioplasty was.
Two-hundred six patients (107 in thrombectomy) with
median NIHSS 17 were enrolled. Median infarct was 7.6 ml in
thrombectomy group versus 8.9 ml in control and randomization
from LWK was 12.2 hours versus 13.3 hours in control.
Thrombectomy was done in 105/107 (3/105 had alternate rescue
therapies). Eighty-four percent had immediate reperfusion, with
median time from LWK to reperfusion of 13.6 hours. At 24 hours,
recanalization was 77 percent in thrombectomy versus 36 percent
in control. The mean score for disability on utility-weighted mRS
(uwMRS weights between zero, death to 10, no symptoms) was
5.5 in thrombectomy compared to 3.4 in control. Similarly, for
the traditional mRS, 49 percent versus 13 percent achieved mRS
0-2 at 90 days. No evidence of heterogeneity was present for any of
the pre-specified subgroups. Neurologic deterioration was lower in
thrombectomy (14 percent versus 26 percent, p=0.04).
Comments: Patients with acute ischemic stroke between six to 24
hours from LKW time due to an intracranial ICA or M1 occlusion,
who had clinical-imaging mismatch had significantly better
outcomes for disability and functional independence at 90 days
with thrombectomy using TREVO when compared to standard
medical care alone. The rate of functional independence (mRS
0-2) was similar to pooled outcomes from prior endovascular
trials < 6 hours. Number needed to treat for benefit was 2.8
thrombectomy patients. The number needed to save one life with
return to normal function was 4.9. Earlier treatment (6-12 hours)
did have better outcomes (54 percent versus 44 percent), but this
was not statistically significant.
DEFUSE3 NEJM Jan 2018
Summary: DEFUSE3 is a multicenter, prospective, randomized,
open-label trial for the use of mechanical thrombectomy in
patients with LVO between six and 16 hours of LKW time. Patients
who were > 18 years old with occlusion of the ICA or proximal
MCA on CTA or MRA and a core infarct volume of less than 70 ml
with a ratio of ischemic tissue to initial infarct core volume of 1.8
or more with a potentially reversible penumbra of at least 15 ml.
Candidates were randomized 1:1 to mechanical thrombectomy
plus medical management or medical management alone. Any
FDA approved thrombectomy device was allowed as well as
carotid angioplasty and stenting. Femoral puncture must occur
within 90 minutes of qualifying imaging. Intraarterial tPA was
not allowed. All medical care was based on the American Heart
Association guidelines, including administration of IV tPA in
10 (11 percent) of patients in the endovascular group and 8 (9
percent) of the medical group.
The trial was stopped after early analysis, requested by the sponsor
(NIH) after the publication of DAWN. A total of 182 patients
were randomized (92 endovascular plus medical treatment
and 90 medical treatment alone). The groups were similar in
baseline demographics. Patients who underwent endovascular
thrombectomy had better mRS at 90 days than the medically
treated group (unadjusted common OR, 2.77). After adjustment
for stratification factors, the odds ratio was 3.36. For 90 day
dichotomized functional outcome (mRS= 0-2 vs 3-6), 45 percent
of the endovascular group compared to 17 percent in the medical
group had good outcomes (RR, 2.67). Mortality at 90 days was 14
percent in the endovascular group and 26 percent in the medical
group (p=0.05). The rate of symptomatic hemorrhage was the
same in the two groups. Serious adverse reactions were reported
in 43 percent of the endovascular group and 53 percent of the
medically treated group.
The median time from LKW time to endovascular therapy was 10
hours and 29 minutes. The median time from randomization to
groin puncture was 28 minutes. Final angiography showed TICI 0
in 10 (11 percent) of patients, TICI 2a 12 (13 percent), TICI 2b 52
(57 percent), and TICI 3 17 (19 percent) of patients. The median
growth of infarct size between baseline and 24 hour imaging was
23 ml in the endovascular group and 33 ml in the medically treated
group (p=0.08), reperfusion of penumbra was more common in the
endovascular group (79 percent versus 18 percent, p<0.001), and
recanalization of the vessel was more common in the endovascular
group (78 percent versus 18 percent, p<0.001). Subgroup analyses
were underpowered for statistical significant analysis.
Comments: DEFUSE3 has demonstrated significant improvement
in functional outcome with mechanical thrombectomy within the
six to 16 hour extended window with use of automated volumetric
software for patient selection. The number needed to treat to
prevent disability is two. Patients in this trial did better clinically
than those in trials for thrombectomy within six hours. This is
likely because patients with mismatch have favorable collateral
flow and slower progression of stroke.
NEWS Review
By Aimee Aysenne, MD, MPH, and Kyle Hobbs, MD
24