Background: The overwhelming benefit of endovascular stroke therapy established in recent trials fueled by rapid workflow, modern devices and favorable noninvasive imaging profiles newly question the impact of collateral grade. We analyzed the SWIFT PRIME trial to evaluate the role of collaterals with respect to advanced imaging selection, the pace of infarct growth and ultimate clinical outcomes.
Methods: Conventional angiography of the endovascular arm in SWIFT PRIME (n=98) was reviewed by 3 independent readers that scored ASITN/SIR collateral flow grade, followed by consensus adjudications. Angiographic collaterals were scored only when available, prior to use of the Solitaire stent retriever. Statistical analyses investigated the relationship of collateral grade with clinical and imaging variables in the main trial dataset.
Results: 70 endovascular arm subjects (mean age 65.2±12.1 years, 35 (50%) women) had angiographic collaterals scored by 3 independent readers (W=0.93). Pretreatment CT or MR perfusion imaging was obtained in 92% of cases, associated with a relatively wide range of angiographic collateral grades (1, n=6 (9%); 2, n=25 (36%); 3, n=37 (53%); 4, n=2 (3%)) with a preponderance (79%) of M1 occlusions. Worse collaterals were associated with prior myocardial infarction (33% (grade 1), 8% (grade 2), 3% (grades 3, 4); p=0.03) but unrelated to age, sex, or other co-morbidities or time from stroke onset. Baseline ASPECTS (median 9, r=0.20) trended higher (p=0.09) in those with more robust collaterals. Collaterals were unrelated to degree of TICI reperfusion, yet the presence of distal emboli at procedure end was linked with worse collaterals (p=0.008). More robust collaterals exhibited a potent and graded link with better day 90 mRS (r=- 0.3, p=0.03) and lower mean NIHSS at 27 hours (r=-0.3, p=0.03), whereas infarct growth at 27 hours was unrelated (r=-0.07, p=0.60).
Conclusions: Use of advanced noninvasive perfusion imaging enrolls patients with more robust collateral grades, yet a diverse range of collaterals may be seen. Even with rapid workflow, mismatch based selection and successful reperfusion, the degree of angiographic collaterals continues to be strongly associated with clinical outcomes at all timepoints.