Symptomatic intracranial atherosclerosis

Methods.

We reviewed the charts of patients with the diagnosis of intracranial atherosclerosis from the Stanford Stroke Center database between August 1992 and September 1999. This database registers the clinical characteristics and the investigations performed on in- and outpatients referred to the Stanford Stroke Center with a diagnosis of TIA (21.1%) or ischemic stroke (78.9%). The database records consecutively admitted in- and outpatients from 1996 onward. The frequency of intracranial atherosclerosis was determined from the number of patients during the period 1996 to 1999. Additional nonconsecutive patients from the period 1992 to 1995 were included for all other analyses. The study was approved by the Stanford University Institutional Review Board.

Demographic characteristics, stroke risk factors, timing of cerebral ischemic events, and location and severity of intracranial atherosclerosis were determined by review of the medical records. A semistructured interview was also conducted by telephone to assess development of any new cerebral ischemic events that occurred since the last clinic or hospital visit.

Inclusion criteria for this study included TIA or stroke attributed to significant intracranial atherosclerosis of the MCA, the carotid siphon, the intracranial vertebral artery, or the basilar artery. Intracranial stenosis was defined as the presence of a significant stenosis (defined as 50 to 99% stenosis) on either angiogram (North American Symptomatic Carotid Endarterectomy Trial method), transcranial Dopper sonography (TCD), or MR angiography (MRA). For angiography or MRA reports that did not specify the percentage of stenosis, descriptive terms including “severe stenosis” or “high- or moderate-grade stenosis” or “flow gap” were accepted as evidence of ≥50% stenosis. TCD criteria for significant stenosis were mean velocity of ≥90 cm/s for vertebral and basilar artery stenosis, ≥120 cm/s for MCA, and ≥90 cm/s for the carotid siphon. TIA was defined as a transient (<24 h) focal neurologic deficit attributed to cerebral ischemia. Stroke was defined as a focal neurologic deficit lasting >24 h attributed to cerebral infarction.

The exclusion criteria for the study were as follows:

• 1. The presence of another potential cause of stroke or TIA such as cardioembolism (atrial fibrillation, prosthetic valve, myocardial infarction within 6 weeks, mitral stenosis, intracardiac clot, bacterial endocarditis, or ventricular aneurysm) or a tandem high-grade proximal stenosis (≥50%) of the vertebral or carotid artery

• 2. Uncertain diagnosis of TIA or stroke

• 3. Severity of stenosis unclear from the records

• 4. Presence of potential confounding neurologic diseases (complex partial epilepsy)

• 5. Intracranial stenosis following basilar meningitis, moyamoya disease, vasculitis, or dissection

• 6. Isolated stenosis of the anterior or posterior cerebral arteries

Failure of antithrombotic therapy was defined as the occurrence of ischemic events (either one or more TIA or strokes or both) while using therapeutic doses of aspirin (>81 mg/day), warfarin (International Normalized Ratio [INR] ≥ 2.0), ticlopidine (500 mg/day), clopidogrel (75 mg/day), or heparin (prolongation of partial thromboplastin time of >1.5 times baseline value). Patients with TIA or stroke who were taking these antithrombotic agents for unrelated reasons (i.e., coronary artery disease) at the time they had a cerebral ischemic event were also considered to fail antithrombotic therapy. Although the initial stroke or TIA was required to be attributable to an intracranial atherosclerotic lesion, all TIA or strokes that occurred during follow-up were counted as outcome events. Cerebral ischemic events related to complications of angiography or intracranial angioplasty were not considered to be qualifying or outcome events. A diagram of the types of patients included in the study is shown in the figure.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Diagram of study design. Gray boxes indicate patients who failed antithrombotic therapy.

Results.

Discussion.

Symptomatic intracranial atherosclerosis is infrequent in patients referred to the Stanford Stroke Center. It was considered to be the exclusive cause of stroke or TIA in 2.7% of consecutive TIA or stroke patients evaluated between 1996 and 1999. In another 1.3%, intracranial atherosclerosis was detected in conjunction with other etiologic mechanisms of stroke. Recent series of consecutive patients have identified symptomatic intracranial atherosclerosis as the most likely etiology of stroke in ∼8% of consecutive ischemic stroke patients.1,2 African Americans and Asians appear more likely to have intracranial atherosclerotic disease.12–14 Differences in study design and case mix may explain the lower frequency of symptomatic intracranial atherosclerosis in our cohort. Our study was retrospective in design and not population based. In addition, lesions of the anterior and posterior cerebral arteries were excluded. Most non-Caucasians were Asian in our cohort, whereas in other studies, most of the non-Caucasians were African American or Hispanic.1,2

Patients with intracranial atherosclerosis frequently had cerebral ischemic events while being treated with an antiplatelet agent or oral anticoagulant. Patients with therapeutic failure were more frequently Caucasian and were more likely to have hyperlipidemia, although these features were not independent predictors of failing antithrombotic treatment in a Cox proportional hazards model. Posterior circulation disease (versus anterior circulation disease) was not a significant predictor of failing antithrombotic therapy. The Cox model identified only older age and use of an antiplatelet agent, rather than warfarin, as predictors of therapeutic failure. Our findings are compatible with a previous retrospective study that suggested warfarin may be more effective than aspirin for treatment of patients with symptomatic intracranial atherosclerosis.5 As this was an observational study, not a randomized trial, these findings must be interpreted with caution because unknown confounding variables and biases might be responsible for the more favorable outcomes in patients treated with warfarin. For example, patients with more severe concomitant medical conditions may have been considered poor candidates for warfarin therapy and preferentially treated with antiplatelet agents; these patients might have increased stroke risk.

To our knowledge, no previous study has focused on the prognosis of patients with symptomatic intracranial atherosclerosis who fail standard antithrombotic therapy. Some of the patients enrolled in the Extracranial/Intracranial Bypass Study were likely receiving antiplatelet agents at the time of their qualifying event for the trial, but a separate analysis of these patients was not reported.3 In the retrospective WASID study, ∼30% of patients were receiving aspirin at the time of the qualifying event, but no data regarding these patients have been presented.5 The Ticlopidine Aspirin Stroke Study (TASS) appears to be the only stroke prevention study that reported results for patients who were taking aspirin or anticoagulants at the time of their qualifying TIA or stroke. In a post hoc subgroup analysis, these patients were reported to have had a higher stroke recurrence rate than patients who were not receiving antithrombotic therapy at the time of their qualifying event.15,16 The frequency of intracranial atherosclerosis in the TASS patients is not known, as intracranial imaging was not required.

As the optimal treatment for patients with symptomatic intracranial atherosclerosis is unknown and no randomized trial has convincingly demonstrated the superiority of any specific antithrombotic agent in preventing recurrent cerebral ischemic events, we considered patients to fail antithrombotic therapy if they had ischemic events while taking either warfarin therapy (INR ≥ 2.0) or an antiplatelet agent that has been shown to be effective for secondary prevention of stroke.17

Almost half of the patients who were not receiving antithrombotic therapy at the time of their initial TIA or stroke had a recurrent cerebral ischemic event. This high recurrence rate is higher than that reported in several previous retrospective and prospective studies of patients with symptomatic intracranial atherosclerosis. Recurrent annual stroke rates of 4 to 12%/year with atherosclerosis of the intracranial anterior circulation and 2.5 to 15%/year with posterior circulation atherosclerotic disease have been reported.3,4,18–26 Our very high recurrent stroke rates may be explained by the referral of high-risk patients to a stroke center and the use of different inclusion and exclusion criteria. For example, patients who had a TIA or a stroke despite treatment with heparin or warfarin were immediately eligible for inclusion in our study.

Among patients who failed antithrombotic therapy, the subsequent rate of stroke or death was extremely high (45%/year) and recurrent events occurred quickly. After 36 days, 50% of these patients had again failed antithrombotic therapy. Seven patients had intracranial angioplasty performed before any recurrent event occurred after their initial failure of antithrombotic therapy. We censored these patients at the time of angioplasty in the follow-up analysis because we wanted to study the natural history of patients who received only antithrombotic therapy.

Our group recently reported a periprocedural complication rate of 8.7% (one fatal vessel rupture, one stroke treated with intra-arterial recombinant tissue plasminogen activator) and a low annual stroke recurrence risk (4.8%) in a series of 23 patients with symptomatic intracranial atherosclerosis treated with intracranial angioplasty.7 This study, however, excluded patients who had angioplasty performed in the setting of acute stroke.

Several limitations to this study should be noted. Referral of complex cases with ongoing cerebral ischemic events may have led to selection bias, overestimating the proportion of patients who would fail antithrombotic therapy in a population-based sample. Referral bias may also have led to inclusion of younger patients, patients with a poor prognosis, or patients with unusual or very complex lesions. Angiography was not obtained in all patients, reflecting the practice at our institution of frequently relying on results of MRA or TCD. Moreover, no formal blinded review of the imaging results was performed. In addition, MRA probably overestimates the degrees of stenosis.27–29 TCD has acceptable sensitivity and specificity for detecting intracranial stenosis, but this varies with the blood flow velocity thresholds used.30–32 The technique is highly operator dependent, and false positives have been reported. If, as in extracranial atherosclerotic disease, a relationship exists between the severity of the stenosis and the occurrence of cerebral ischemic events, the inclusion of false positives might underestimate the true recurrence rate in our population.33 The retrospective design of the study did not allow calculation of ipsilateral stroke or TIA rates. Follow-up was relatively short in some patients. Some neurologic symptoms occurring during the follow-up period were difficult to localize clearly to a specific vascular territory or to verify as true cerebral ischemic events. We relied on the treating neurologist’s opinion of whether neurologic events represented ischemic phenomena.

Patients with symptomatic intracranial atherosclerosis who fail antithrombotic therapy appear to be at high risk for subsequent cerebral ischemic events. Recurrent events frequently take place within the first month after failing antithrombotic therapy. High event rates can be anticipated in the medical therapy arms in studies of new interventions that enroll patients with symptomatic intracranial stenosis who have failed antithrombotic therapy.