Materials and methods

Two large regional referral stroke centers participated in this retrospective analysis of consecutive patients presenting with symptomatic (transient ischemic attack (TIA) or stroke) complete occlusion of an intracranial artery of presumed atherosclerotic etiology in whom endovascular therapy was performed beyond 48 h from the time last seen well. All of the patients were considered to be in the high risk category and had symptomatic intracranial occlusions with progression of their symptoms despite maximal medical therapy. We considered maximal medical management to be therapeutically dosed anticoagulation or dual antiplatelet therapy, cholesterol lowering agent, optimal blood pressure, and glucose control, as well as smoking cessation. All patients (except patient No 12) were thought to have progression of their symptoms based on the treating physician's judgment, which for the most part included a change in the score of the National Institutes of Health Stroke Scale (NIHSS) of ≥4 points, or recurrence of their symptoms despite medical management. They also had shown signs of hemodynamic failure, either through perfusion CT/MRI imaging, history, or physical examination.

In the case of patient No 12, a poor clinical examination (NIHSS score of 28) in the setting of a large vessel occlusion and relatively small core prompted endovascular therapy. Imaging selection was based on visual inspection of MRI or CT perfusion (CTP) scans (CTP for patients with anterior circulation disease only) demonstrating mismatch between presumably irreversibly compromised brain (diffusion weighted imaging (DWI) or CTP–cerebral blood volume lesion) and inadequately perfused brain (cerebral blood flow, mean transit time, or time to peak MRI, or CTP maps). None of the patients had evidence of intracranial hemorrhage or CT evidence of hypodensity involving greater than a third of the affected territory.

The majority of the patients had vertebrobasilar occlusions. All of the patients who underwent stenting of the vertebral artery occlusion either had contralateral vertebral artery disease (occlusion or significant stenosis) or a contralateral hypoplastic vertebral artery. The criteria that were used to select patients with posterior circulation occlusion included MRI-DWI to NIHSS mismatch (eg, clinical–core mismatch) in most of the patients (based on the treating physician's judgment as no set DWI or NIHSS was used) and clinical and radiographic findings that suggested hemodynamic failure, such as worsening symptoms in the upright position, worsening symptoms with lower blood pressure, or watershed pattern infarct on MRI.

The diagnosis of intracranial atherosclerosis as the primary etiology for the treated stenosis was based on a constellation of findings including: (1) presence of at least one major vascular risks factor (eg, hypertension, diabetes mellitus, dislipidemia, or smoking); (2) presence of multifocal intracranial stenosis with angiographic appearance consistent with intracranial atherosclerotic disease; (3) lack of improvement of the stenotic area on repeat imaging (suggesting intrinsic non-reversible lesion); (4) absence of any clinical, laboratory, or imaging findings suspicious for other potential causes of intracranial stenosis, such as vasculitis, Moya–Moya syndrome, reversible cerebral vasoconstrictive syndrome, cerebral vasospasm, and intravascular lymphoma; and/or (5) evidence of calcification at the level of the stenosis.

Cases were identified through a search of prospectively acquired endovascular databases at each of the institutions. The following baseline and treatment variables were studied: age, gender, and medical comorbidities. Angiographic images were analyzed and graded at each participating center. Patients were treated between April 2005 and June 2012. Institutional review board approval at each center was obtained according to each center's retrospective studies institutional review board policies and data analysis was carried out in a de-identified manner.

After occlusion of an intracranial vessel was identified, various recanalization strategies were used at the discretion of the operator. There was one case that was treated with angioplasty only; the other 23 cases involved stent placement. Flow through the previously occluded vessel was graded using the Thrombolysis in Cerebral Ischemia (TICI) grading system at each institution. Post-treatment imaging with CT or MRI was performed in all patients according to the local standard of care. All patients were discharged on dual antiplatelets, consisting of aspirin and clopidogrel, for a minimum of 6 months. Primary endpoints included successful recanalization (defined as TICI ≥2b) and good functional outcome (defined as a modified Rankin Scale (mRS) score of ≤2 at 3 months). Secondary endpoint included rates of symptomatic postprocedure parenchymal hematomas (PH, including both PH-1 and PH-2 according to the ECASS criteria)10 and mortality. Functional outcomes were obtained in the outpatient clinic, at follow-up, or by telephone. In cases when outcomes following discharge were not available due to loss of follow-up, the discharge mRS was carried forward.