Comparison of elective stenting of severe vs moderate intracranial atherosclerotic stenosis

Symptomatic intracranial atherosclerotic stenosis (SIAS) is an important cause of ischemic stroke.^1–4 The Warfarin–Aspirin Symptomatic Intracranial Disease (WASID) trial showed that patients with ≥70% stenosis are at particularly high risk of ischemic stroke in the stenotic artery territory and recommended that these patients should constitute the target group for a randomized trial comparing stenting with the best medical therapy in the future.⁵ Stenting of intracranial stenosis is technically feasible,^6–12 but it remains unclear whether ≥70% severe stenosis is associated with a higher subsequent stroke risk after stenting than 50% to 69% moderate stenosis. Currently, there has been no comparison made between stenting of severe vs moderate intracranial stenosis. In this study, we tested whether severe SIAS was associated with a higher subsequent stroke risk than moderate stenosis after elective angioplasty with a balloon-expandable stent, and we explored which factors were associated with the subsequent stroke.

Methods.

The patients analyzed herein were identified among 231 patients collected consecutively in a database of stenting for intracranial stenosis at our institute between September 5, 2001, and June 30, 2005, and followed up until end of June 2006. The database prospectively recorded the results of stenting for intracranial stenosis. To focus on knowing the effect of elective stenting on SIAS, lesions with the following characteristics were included in this study: 1) angiographically verified ≥50% stenosis of a major intracranial artery (intracranial internal carotid artery, mainstem of middle cerebral artery, intracranial vertebral artery, and basilar artery); 2) resulted in ischemic stroke or TIA as a qualifying event; 3) treated by elective stenting (i.e., interval between qualifying event and stenting was at least 24 hours for TIA, 7 days for minor stroke [NIH Stroke Scale score < 9],^13,14 or 6 weeks for major stroke [NIH Stroke Scale score ≥ 9]); and 4) in that patient was associated with one or more atherosclerotic risk factors (hypertension, diabetes, hyperlipidemia, hyperhomocysteinemia, and cigarette smoking). Lesions with the following characteristics were excluded: 1) treated at <24 hours from TIA, <7 days from minor stroke, or <6 weeks from major stroke; 2) existed concurrently with an aneurysm at the site of stenosis; and 3) in that patient was associated with no atherosclerotic risk factors. The study protocol was approved by our institutional ethics committee. Written informed consent was obtained before the procedure.

MRI and cerebral angiography were completed 1 to 7 days before the operation. Preoperative lesion-related infarct in the stenotic artery territory was evaluated blindly by a neuroradiologist (X.T.X.). The percentage of stenosis was calculated according to the WASID method,¹⁵ and measurement was completed manually and blindly by two neuroradiologists (N.M. and Q.H.W. by consensus). Target lesions were then divided into a ≥70% severe stenosis group and a 50% to 69% moderate stenosis group.

Patients were given aspirin 300 mg plus clopidogrel 75 mg daily, or aspirin 300 mg daily plus ticlopidine 250 mg twice daily, at least 7 days before the operation. The antiplatelet therapies were continued for ≥6 months after the operation. Atherosclerotic risk factors were also managed by a stroke neurologist (K.H.D.) according to the American Heart Association guidelines.¹⁶

Steps of stenting have been published previously.^9,17 In summary, the patient had either local or general anesthesia. Direct stenting with a balloon-expendable stent was then performed by one of two experienced interventional neuroradiologists (W.J.J. has 20 years of experience with endovascular therapy, and B.D. has 12 years). During the operation, the patient received heparin 2,000 to 3,000 units as an IV bolus, followed by 500 to 800 units/hour IV drip, similar to the Prolyse in Acute Cerebral Thromboembolism (PROACT) II protocol.¹⁸ Stent diameter was selected to be same as the diameter of the normal adjacent vessel or slightly smaller (1:1 or 0.9:1). The stent was then deployed by gradual balloon inflation up to 6 to 8 atm when it straddled the stenotic segment. Successful stent placement was defined when the stent completely covered a target lesion, resulting in a ≤30% residual stenosis with good anterograde blood flow. A second stent was used when the lesion was not completely covered by the first stent. Two categories of balloon-expandable stents, i.e., intracranial stent (Apollo stent) and coronary stent, were attempted in this study (table 1); BiodivYsio coronary stents and Apollo stents were mainly used. Now, the Apollo stent (MicroPort Medical [Shanghai], Co., Shanghai, China) is only available in China and is approved by the State Food and Drug Administration of our country.

Primary endpoints included lesion-related ischemic stroke, symptomatic brain hemorrhage, and symptomatic subarachnoid hemorrhage (SAH). Secondary endpoints included non–lesion-related ischemic stroke, asymptomatic brain hemorrhage and SAH, emergent cerebral revascularization (ECER), death from other vascular causes, and other major hemorrhages. Ischemic stroke was defined as a new focal neurologic deficit of sudden onset that lasted at least 24 hours and that was not caused by hemorrhage on brain CT. Ischemic stroke was considered to be 1) definitely lesion-related when the neurologic signs correlated with a new infarct on CT in the stenotic artery territory; 2) probably lesion-related when the neurologic signs were localized to the stenotic artery territory but without a new infarct on CT; 3) indeterminate when the neurologic signs were localized to two or more distinct vascular territories without new infarct on CT; or 4) definitely or probably non–lesion-related.^4,5 Lesion-related ischemic strokes included definitely and probably lesion-related ones, and non–lesion-related ischemic strokes consisted of other ischemic strokes. Brain hemorrhage was defined as evidence of brain parenchyma blood on CT, classified as symptomatic (NIH Stroke Scale score ≥ 1 greater than baseline) or asymptomatic. SAH was defined as evidence of subarachnoid space blood on CT, classified as symptomatic (meningeal irritation signs) or asymptomatic. Death from other vascular causes was defined as sudden death or death within 30 days from an acute ischemic or hemorrhagic disease other than ischemic stroke, brain hemorrhage or SAH, such as myocardial infarction, or extracranial vascular rupture. ECER was defined as an emergency therapy by IV thrombolysis or endovascular intervention for an angiographically verified acute occlusion of a cerebral artery that led to immediately complete patency and entire disappearance of a new focal neurologic deficit within 24 hours. Other major hemorrhage was defined as necessitating hospitalization, blood transfusion, or surgery.

After surgery, all patients had daily follow-up during hospitalization until discharge. After discharge, they had follow-up visits at 30 days. Subsequent follow-ups after 30 days were scheduled every 3 months until 12 months, and then every 6 months either by a clinic visit or by telephone. If a stroke was suspected, the patient would have a brain CT. Assessment of endpoint events was completed by the stroke neurologist (K.H.D.).

Follow-up angiography was scheduled after 6 months on a voluntary basis of patients, or clinically when restenosis was suspected. Restenosis was defined as an angiographically verified ≥50% stenosis within or at the edge of the stent.

Unless otherwise specified, an intention-to-treat analysis was performed. The Fisher exact or χ² test (for percentages) or the t test (for means) was used to identify differences between the groups. A hazard ratio (HR) was calculated by the Cox proportional hazards regression model (for severe stenosis as compared with moderate stenosis). The cumulative probability of primary endpoints over time was estimated by the product-limit method. Data pertaining to patients lost to follow-up were censored on the last contact date. The groups with severe and moderate stenosis were also compared using the log-rank test. To explore which factors were associated with primary endpoints in each of the two groups, univariate analyses (assessing the effect of each baseline factor) with use of the Cox proportional hazards regression model and then a stepwise regression test were performed if baseline characteristics were potentially associated with primary endpoints in the univariate analyses at the p ≤ 0.10 level. All reported p values were two-sided, and p values < 0.05 were considered significant.

Results.

A total of 220 stenoses in 213 patients were eligible for this study (figure 1). The patients’ average age was 52.8 years (range 20 to 79 years; median 53 years). The procedure was performed at a median of 30 days after the qualifying event. The severe stenosis group had 126 stenoses in 121 patients (single lesion in 117 patients, 2 lesions in 3 patients, and 3 lesions in 1 patient). The moderate stenosis group had 94 stenoses in 92 patients (single lesion in 90 patients and 2 lesions in 2 patients). There were no significant differences in baseline characteristics between the groups, but there was a significantly higher proportion of long lesion (>10 mm) or basilar stenosis in the severe stenosis group (table 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Inclusion flow chart.

The stent success rate was 92.3% (203/220). Stent failure occurred in 17 lesions due to failed positioning of the guiding catheter or microwire (n = 6) or failed navigation of the stent system through severe or moderate tortuosity despite an appropriate position of the guiding catheter and microwire (n = 11). Severe tortuosity access was associated with stent failure (10/55 severe tortuous accesses vs 7/165 mild and moderate ones; p = 0.002). In this study, we used a total of 223 stents, of which 212 were successfully placed in 203 lesions, and the other 11 were removed from the artery. A second stent was necessary in 7 lesions of the severe stenosis group and 2 lesions of the moderate stenosis group (7/126 vs 2/94; p = 0.304). There were no significant differences in procedural results between the groups (table 1).

Severe stenosis was not shown to be a high-risk feature of 30-day primary or secondary endpoints (table 3). Within 30 days, the severe stenosis group had 6 (4.8%, 6/126) primary events (1 definite lesion-related ischemic stroke and 3 probable ones, and 2 symptomatic SAHs from which 2 patients died) and 3 (2.4%, 3/126) secondary endpoints (3 ECERs). The moderate stenosis group had 4 (4.3%, 4/94) primary events (1 probable lesion-related ischemic stroke and 2 definite ones, and 1 symptomatic brain hemorrhage) and 6 (6.4%, 6/94) secondary endpoints (4 ECERs and 2 asymptomatic SAHs). All 10 primary events within 30 days occurred in the stent success subgroup of the severe (5.3%, 6/114) or moderate (4.5%, 4/89) stenosis group.

Severe stenosis was also not shown to be a high-risk feature of primary or secondary endpoints after 30 days (table 4). The average follow-up time after 30 days was 26.0 months for the severe stenosis group (121 lesions in 116 patients) and 27.6 months for the moderate stenosis group (92 lesions in 90 patients). In the severe stenosis group, 110 lesions (105 patients) belonged to the stent success subgroup, and 11 (11 patients) belonged to the stent failure subgroup. During this period, we observed 4 (3.3%, 4/121) definite lesion-related ischemic strokes (3 [27.3%, 3/11] in the stent failure subgroup and 1 [1.0%, 1/110] in the stent success subgroup) and 2 (1.7%, 2/121) definite non–lesion-related ischemic strokes. In the moderate stenosis group, 87 lesions (85 patients) belonged to the stent success subgroup, and 5 (5 patients) belonged to the stent failure subgroup. We observed 3 (3.3%, 3/92) definite lesion-related ischemic strokes (all [3.4%, 3/87] in the stent success subgroup) and 1 (1.1%, 1/92) definite non–lesion-related ischemic stroke after 30 days.

The cumulative probability of the primary endpoints (figure 2) was 7.2% (95% CI 2.6% to 11.8%) at 1 year and 8.2% (95% CI 1.9% to 14.5%) at 2 years for the severe stenosis group, and 5.3% (95% CI 0.7% to 9.9%) at 1 year and 8.3% (95% CI 1.3% to 15.3%) at 2 years for the moderate stenosis group (HR 1.09, 95% CI 0.42 to 2.87; p = 0.860) (log-rank test, p = 0.860).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Product-limit estimate of the probability of primary endpoint events vs months after the procedure, according to percent stenosis (severe stenosis shown as solid line, moderate stenosis shown as dashed line); log-rank test, p = 0.860.

Univariate analyses (table 5) showed that four factors (diabetes, no use of antithrombotic therapy at the time of the qualifying event, lesion in posterior circulation, and stent failure) were potentially associated with primary endpoints in the severe stenosis group, and no single factor was associated with primary endpoints in the moderate stenosis group. Multivariable analysis revealed that stent failure was an independent predictor of primary endpoints in the severe stenosis group (HR 5.31, 95% CI 1.35 to 20.91; p = 0.017).

Fifty-six stented vessels (51 patients) in the severe stenosis group and 43 (43 patients) in the moderate stenosis group had follow-up angiograms. There were no significant differences in baseline characteristics and clinical outcomes between lesions (having clinical follow-up after 30 days) with and without follow-up angiograms, but a significantly higher proportion of diabetes or qualifying stroke events before stenting, and a lower proportion of hyperhomocysteinemia in the angiographic follow-up group (table 6). Angiography revealed 14 (25.0%, 14/56) restenoses (11 asymptomatic and 3 symptomatic [1 ischemic stroke and 2 TIAs]) in the severe stenosis group at a mean time of 8.6 months ± SD of 5.2, and 5 (11.6%, 5/43) restenoses (3 asymptomatic and 2 symptomatic [1 ischemic stroke and 1 TIA]) in the moderate stenosis group at a mean time of 8.5 months ± SD of 4.2. Although the severe stenosis group had a higher restenosis rate than the moderate stenosis group, no significant difference was detected (HR 2.03, 95% CI 0.72 to 5.75; p = 0.18).

Other subgroup analyses were also performed. There were no significant differences between operators or between lesions treated with coronary and Apollo stents in clinical outcomes (table 5); between operators (B.D. vs W.J.J.) in stent success (44/47 vs 159/173), complication (4/47 vs 15/173), and restenosis (3/20 vs 16/79); or between lesions treated with coronary and Apollo stents in stent success (145/153 vs 58/61), complication (6/153 vs 4/61), and restenosis (13/68 vs 6/31). There was a difference between the qualifying stroke event group and the qualifying TIA event group in primary endpoints after 30 days (6/82 vs 1/138; p = 0.012), but no significant difference in 30-day primary endpoints (2/82 vs 8/138).

Discussion.

The WASID trial determined that severe stenosis is an independent predictor of lesion-related ischemic stroke. In the trial, the cumulative probability of lesion-related ischemic strokes at 1 year was ≥17% in patients with severe stenosis despite aspirin or warfarin therapy, whereas it was 7% to 8% in patients with moderate stenosis.⁵ Such a significant difference in the subsequent lesion-related ischemic stroke risk may justify the more aggressive and effective therapies such as angioplasty or stenting for patients with severe SIAS. However, there is no study yet to provide an answer about whether severe intracranial stenosis is associated with a higher subsequent stroke risk than moderate stenosis despite stenting therapy.

Indirectly compared with the WASID trial, our results suggest that patients with severe stenosis seem to receive the benefit from elective stenting, whereas patients with moderate stenosis may not. In the current study, the cumulative probability of primary endpoints after elective stenting was 7.2% at 1 year and 8.2% at 2 years in the severe stenosis group, and 5.3% at 1 year and 8.3% at 2 years in the moderate stenosis group. The similar outcomes in the two groups may mean that elective stenting has eliminated the degree of stenosis as a predictor of outcome.

A long lesion of >10 mm has been discovered to be a disadvantageous morphologic feature for intracranial angioplasty or stenting, associated with both recurrent and procedure-related morbidity.^19,20 It has also been reported that stenting of basilar stenoses is associated with a higher risk of perforator stroke than stenting of other intracranial stenoses.¹⁷ Although higher proportions of long lesion and basilar stenosis existed in the severe stenosis group of this series, severe stenosis was still not found to be associated with a higher risk of subsequent primary endpoints than moderate stenosis.

Furthermore, this study shows that stent failure is an independent predictor of the subsequent primary endpoints in the severe stenosis group. In the moderate stenosis group, there was no one particular factor (including stent failure) potentially associated with any primary endpoint. These results imply that patients with severe stenosis may benefit from successful stent placement, but patients with moderate stenosis may not. As a suggestive possible benefit, stenting of severe stenosis requires the results of a randomized trial before routine clinical use.

Procedure-related complications of stenting for intracranial stenosis are diverse, including intracranial hemorrhage, target-lesion thrombosis, perforator stroke, embolic stroke, vessel dissection, and vessel rupture.^6–12,17,19 Obviously, periprocedural stroke and death might offset the potential benefit of stenting. This study does not prove that severe stenosis is a predictor of periprocedural primary or secondary endpoints. The frequency of primary endpoints within 30 days in the severe stenosis group was 4.8%, similar to 4.3% in the moderate stenosis group. The rate of secondary endpoints within 30 days in the severe stenosis group was 2.4%, lower than 6.4% in the moderate stenosis group, but no significant difference was detected. These results imply that compared with the results of medical treatment, the procedure-related primary or secondary endpoint rate of elective stenting may be acceptable for patients with severe stenosis but may not be for patients with moderate stenosis.

Restenosis is an important topic of arterial angioplasty. The Stenting of Symptomatic Atherosclerotic Lesions in the Vertebral or Intracranial Arteries (SSYLVIA) study shows a 32.4% (12/37) restenosis rate after stenting for intracranial stenosis.⁶ In the current study, the restenosis rate was 19.2% (19/99). Although follow-up angiograms were performed in only approximately half of the stented vessels in this case series, the rate may not be underestimated, because of a higher proportion of diabetes or qualifying stroke events before stenting, a lower proportion of hyperhomocysteinemia in the angiographic follow-up group, and no differences in other characteristics. To our knowledge, diabetes is a predictor of restenosis at 6 months,⁶ and almost all available clinical trials have shown that hyperhomocysteinemia is not associated with in-stent restenosis of coronary stenting.²¹ There has been no study yet to demonstrate an association of qualifying stroke events before stenting with restenosis, though the WASID trial revealed a higher risk of recurrent lesion-related ischemic stroke in patients with stroke as a qualifying event than those with a qualifying TIA event,⁵ and the current study shows a higher rate of lesion-related stroke after 30 days of stenting in lesions with a qualifying stroke event before stenting.

There are a few limitations in this study. First, not all cases were blinded to the stroke neurologist, even though the physician did not know the stenosis degree of the target lesion. Second, there could be referral bias to the neuroradiologist for the procedure; it would have been useful to assess this bias by comparing patients not referred for stenting to those in the study cohort for any important differences in baseline characteristics and outcomes; however, we regret that this study cannot make such a comparison because the database only recorded the patients who underwent stenting. Finally, the absence of a significant difference in restenosis rate between the severe and moderate stenosis groups in this study should be cautiously explained and may be caused by a Type II error owing to low test power (0.23).