Data from patients admitted to our tertiary medical center for acute ischemic stroke during the study period were retrospectively analyzed. Symptomatic BVAO was defined as the presence of neurological symptoms associated with posterior circulation stroke along with an ischemic lesion on computed tomography (CT) or magnetic resonance imaging (MRI) that could be attributed to BVAO. BVAO was confirmed by CT angiography (CTA) and MR angiography (MRA). Cases with stenosis or hypoplasia of at least 1 vertebral artery were excluded.

Clinical data were obtained from patients’ medical records, including demographic features, vascular risk factors, time from onset to door, National Institutes of Health Stroke Scale (NIHSS) score during admission and at discharge, modified Rankin Scale (mRS) score at discharge, 2-week, 3-month follow-up, revascularization therapy such as intravenous thrombolysis and EVT, medical treatment, and imaging. The time of stroke onset was recorded if witnessed, or from the last known normal time if the exact time was not known. Stroke severity was assessed using the NIHSS. Imaging data included brain CT, CTA, MRI, MRA, and digital subtraction angiography (DSA) images obtained during the admission period. In our institutional imaging protocol, patients who visited the emergency room initially underwent brain CT, CTA from the aortic arch to the vertex with or without perfusion imaging, and brain MRI and MRA within 24 hours after admission, 5 days after admission, or after neurological deterioration. Collateral flow state was assessed on the basis of initial CTA, MRA, and DSA.

On admission, a stroke neurologist performed a neurological assessment based on the NIHSS score. Inclusion criteria for rescue EVT were as follows: (1) procedure initiated within 24 hours of symptom onset or the last known time the patient was normal, (2) baseline NIHSS score >5, (3) BVAO, (4) no intracranial hemorrhage on brain CT. However, during the study of patients with ischemic stroke due to BVAO, it was found that patients who initially presented with mild stroke symptoms (NIHSS≤5) experienced neurological deterioration during their hospitalization. Therefore, we performed rescue EVT even in cases classified as showing mild stroke symptoms (NIHSS≤5). Patients who presented to the hospital >24 hours after symptom onset were not considered for rescue EVT. Written informed consent for EVT was obtained from each patient’s family. Initially, DSA was performed to evaluate the occlusion sites and the status of the collateral circulation. The choice of EVT procedure was at the discretion of the interventional neurologist. After confirming ease of catheter access to the vessel, the interventional neurologist proceeded with EVT. A stent retriever (Solitaire^TM; Medtronic) with simultaneous aspiration (5Fr Sofia^TM; MicroVention) was used as the first-line EVT. After thrombectomy, balloon angioplasty (Gateway^TM; Stryker) and stent angioplasty (Enterprise^TM2; Cerenovus) were performed. If restenosis or occlusion was detected, tirofiban (0.5–1.0 mg) was injected through the guiding catheter, and additional tirofiban was continuously infused intravenously for 12 hours. After revascularization therapy, the systolic blood pressure was maintained at <180 mmHg. Brain MR and MRA were performed within the first 24 hours after the procedure and 5 days after admission. Antithrombotic agents were initiated after confirmation of hemorrhage or hemorrhagic transformation on brain CT or MRI.

Inclusion criteria of conventional medical treatment were as follows: (1) patients who do not meet the inclusion criteria for rescue EVT and (2) patients who attempted rescue EVT but underwent only DSA due to difficulty in vertebral artery selection. Patients without a cardiogenic embolic source, such as those with atrial fibrillation, were administered dual antiplatelet agents (aspirin 100 mg and clopidogrel 75 mg) with high-intensity statins (atorvastatin 40 mg or rosuvastatin 20 mg). The systolic blood pressure was allowed up to 220 mmHg during the acute phase. Comorbidities such as diabetes mellitus and hyperlipidemia were also treated concomitantly.

The NIHSS score was used to assess neurological outcomes. Neurological improvement was defined as the difference between the discharge NIHSS score and the baseline NIHSS score. Neurological deterioration was defined as an incremental change in total NIHSS score with new ischemic lesions on brain MRI and was assessed using the difference between the severe stage NIHSS score and the baseline NIHSS score. Functional outcome was assessed by mRS at discharge, 2 weeks, and 3 months after discharge. Good clinical outcome was defined as no stroke recurrence and an mRS score ≤2 at 3 months after discharge.

Ten patients received conventional treatment. Two of them underwent DSA followed by an attempted EVT, but due to difficulty in selecting vertebral artery orifice, the procedure was terminated, and they received conventional treatment. Dual-antiplatelet agents (aspirin 100 mg and clopidogrel 75 mg) and high-intensity statins (atorvastatin 40 mg or rosuvastatin 20 mg) were initially administered as conventional treatment. The median onset to door time was 49 hours (range 6–144 hours). The median initial NIHSS score was 5 points (range 0–18 points). The median increase in NIHSS score at the most severe stage from admission was 9 points (range 1–16). The median time between admission and the highest NIHSS score was day 9 of hospitalization (range 3–14 days). Nine of the 10 patients who received medical treatment without EVT presented an increase in the NIHSS score with additional ischemic lesions in the pons and/or midbrain (Fig. 1). Five patients presented bilateral peduncular ischemic stroke accompanied by a lesion in the bilateral pons that progressed to locked-in syndrome during hospitalization (Fig. 2). At discharge, 2 patients died due to multi-organ failure associated with systemic infection and cardiogenic shock following acute myocardial ischemia, and 7 patients had severe disability with mRS 4 and mRS 5 (progression to locked-in syndrome). During follow-up, 3 of the 4 patients discharged with mRS 5 died and 1 patient was discharged with an mRS of 4 and had recurrent stroke in the bilateral pons and progressed to mRS 5. Only 1 patient had a favorable clinical outcome with an mRS of 1.

Seven patients initially underwent rescue EVT. The median onset to door time was 9.5 hours (range 5–16 hours). The median initial NIHSS score was 8 points (range 3–17 points). After mechanical thrombectomy, balloon angioplasty and stent placement were performed in the V4 segment of the vertebral artery (Fig. 3). There were no complications, such as intracerebral hemorrhage, hemorrhagic transformation, dissection, or re-occlusion, after the procedure. No patient who received EVT presented with neurological deterioration. The median decrease in NIHSS score at discharge was 5 points (range 2–10 points) compared with that at admission. At the 3 months follow-up, 5 patients presented with favorable clinical outcomes. Of the 2 patients with unfavorable outcomes at 3 months, 1 died of an unknown cause and the other who also had basilar artery occlusion initially presented locked-in syndrome with an mRS of 5.