Background and aim Posterior circulation stenosis may be a risk factor associated with stroke after intracranial stenting as compared with anterior circulation stenosis. Our aim was to test our hypothesis that there was no difference in clinical outcome poststenting between patients with severe stenosis of the basilar artery (BA) and intracranial vertebral artery (VA).
Methods Using the Cox proportional hazards regression model adjusted for prespecified factors (qualifying event, and timing of stenting after the qualifying event), we compared primary endpoint (ischemic stroke in the vertebrobasilar territory, including any stroke or death within 30 days of stenting) between patients with severe symptomatic atherosclerotic BA and VA stenosis who underwent elective stenting in our prospective database. Analysis was by intention-to-treat principle.
Results Primary endpoint event occurred in 13 (18.8%) of 69 patients with BA stenosis during a mean 23.4 months (9 within 30 days and 4 afterward) and 3 (4.3%) of 70 patients with VA stenosis during a mean 26.4 months (2 within 30 days and 1 afterward). Patients with BA stenosis had a significantly higher risk of the primary endpoint (adjusted HR=4.87, 95% CI 1.37 to 17.29; p=0.014) or any stroke or death within 30 days of stenting (adjusted HR=5.13, 95% CI 1.10 to 23.96; p=0.038) than those with VA stenosis.
Conclusion A significantly higher stroke risk poststenting exists in patients with severe BA stenosis than those with VA stenosis. The discrepancy in clinical outcome after stenting between patients with BA and VA stenosis should be considered in clinical practice and stenting trials.
Statistics from Altmetric.com
Severe symptomatic atherosclerotic intracranial stenosis (SAIS) is an important cause of ischemic stroke.1–5 There has been increased enthusiasm for the use of stents to treat SAIS patients in recent years, but the reported outcomes show wide variability in periprocedural stroke or death within 30 days of stenting, ranging from 0% to 36%.6–23 Some characteristics of patient or site may be risk factors associated with stroke after intracranial stenting. These include intracranial vertebrobasilar arterial (VBA) stenosis,24 25 low stenting-procedure volume sites (<10 patients each), stroke as a qualifying event and stenting soon (within 10 days) after the qualifying event.25 Among them, intracranial VBA stenosis was found to be the factor with the highest hazard ratio (HR) for ischemic stroke in the territory of the stented artery (including any stroke or death within 30 days of stenting) as compared with anterior circulation stenosis.25
It is necessary to ascertain which has the higher risk of subsequent stroke, basilar artery (BA) stenting or intracranial vertebral artery (VA) stenting. From 2001 to 2008, we performed, on average, ≥60 stenting procedures for symptomatic intracranial stenosis per year. This provided a unique opportunity to test our hypothesis that there was no difference in clinical outcome between patients with severe SAIS of BA and intracranial VA who underwent elective stenting at a high-volume site.
The study protocol was approved by our institutional ethics committee, and written informed consent was obtained before each stenting procedure.
Four hundred and forty-four consecutive patients with 468 symptomatic intracranial stenoses of ≥50% were registered in our stenting database between September 2001 and December 2008. The database prospectively collected patient demographics, atherosclerotic risk factors, qualifying events, laboratory and imaging findings, procedure results and clinical outcomes. Patients with the following characteristics were eligible for this study: (1) ischemic stroke or transient ischemic attack (TIA) within 6 months as a qualifying event, attributable to an angiographically verified ≥70% stenosis of BA or intracranial VA; (2) treated by elective stenting with the use of either a self-expanding Wingspan stent (Boston Scientific, Fremont, California, USA) or a balloon-expandable stent (Apollo stent, MicroPort Medical, Shanghai,China; or coronary stent), after an interval (between the qualifying event and stenting) of ≥24 h for TIA, ≥7 days for minor stroke (NIH Stroke Scale score <9) or ≥6 weeks for major stroke (NIH Stroke Scale score ≥9); (3) with one or more atherosclerotic risk factors (hypertension, diabetes, hyperlipidemia, hyperhomocysteinemia, and cigarette smoking). Patients with the following conditions were excluded: (1) treated by emergency stenting due to an acute ischemic stroke, or by restenting due to a restenosis; (2) treated by balloon angioplasty alone, or by Neuroform stent (a self-expanding stent approved for assistance of coil embolization of aneurysms (Boston Scientific)); (3) with a concurrent aneurysm at the stenotic site, or without one atherosclerotic risk factor. On the basis of these criteria, 139 patients with 140 lesions (one with severe, bilateral intracranial VA stenoses) were enrolled into this study (figure 1), including 50 patients whose long-term outcome has been described previously.18
Stenting was performed by one of two experienced interventional neuroradiologists (WJJ and BD). Balloon-expandable stents were used between September 2001 and December 2006, and Wingspan stents had been used since January 2007. The peri-operative preparation and management, and the balloon-expandable stenting procedure were the same as those of the studies described previously.10 17 The Wingspan stenting procedure was similar to that of the balloon-expandable stent placement in many ways. However, during the Wingspan stent placement, there was predilation of the stenotic site with an undersized Gateway balloon (Boston Scientific) at low pressure before deployment of the stent. The length of the balloon was approximated to the lesion's length, and the inflated diameter was sized to approximately 80% of the normal vessel size. The diameter of the Wingspan stent was sized to exceed the diameter of the normal parent vessel by 0.5–1.0 mm, and the length was estimated according to that of the lesion plus 3 mm on either side. In this study, successful stent placement was defined as complete coverage of the target lesion with the stent, resulting in <50% residual stenosis, with good anterograde blood flow. The degree of stenosis was independently evaluated by a radiologist (XTX) according to the Warfarin–Aspirin Symptomatic Intracranial Disease (WASID) trial criteria.26
Clinical follow-up information was derived from daily patient exams until discharge, 30-day follow-up visits and subsequent clinic visits or telephone conversations until the end of March 2009. Follow-up angiography was scheduled after 6 months on a voluntary basis, or when restenosis was suspected clinically. Restenosis was defined as an angiographically verified ≥50% stenosis within the stent or at the edge of the stent, which was also assessed independently by the radiologist (XTX).
Primary endpoint was ischemic stroke in the VBA territory after 30 days, or any stroke or death within 30 days of stenting. It was independently evaluated by an experienced stroke neurologist (KHD). Intracranial hemorrhage (ICH) was defined as CT evidence of blood in the brain parenchyma or subarachnoid space. Ischemic stroke was defined as the sudden onset of a new focal neurological deficit persisting for at least 24 h that was not caused by ICH. The stroke was further classified into perforator related and non-perforator related,16 and was considered (1) definite posterior circulation stroke when the neurological signs correlated with a new infarct in the VBA territory on CT or MRI; (2) probable posterior circulation stroke when the neurological signs were localized to the VBA territory but without a new infarct; (3) indeterminate when the neurological signs were localized to two or more distinct vascular territories in the anterior and posterior circulation without a new infarct; or (4) definite or probable anterior circulation stroke.17 In the current study, ischemic stroke in the VBA territory was definite or probable.
We determined the test power of the study with the use of the method of Schoenfeld for the proportional hazards regression model.27 Assuming that the primary endpoint would occur in 10.1% of patients in this cohort before the end of the study on the basis of a pilot study,18 we calculated that the power was 80% to detect an HR of ≥e1.5 (regression coefficient 1.5) for the primary endpoint between the BA vs VA group at the 5% level of significance (two-tailed).
All data were analyzed according to the intention-to-treat principle. Continuous variables were presented as means±SD, and categorical variables were presented as percentages. To identify a difference in patient and procedure characteristics, and in clinical outcome and restenosis between the BA and VA group, Fisher's exact test (when the expected cell frequency was <5) or the χ2 test was used for categorical variables, Student's t test was used for continuous variables and the Mann–Whitney U test was used for continuous variables that had skewed distributions. Cumulative probability of the primary endpoint over time was estimated by the product-limit method. The probabilities of the two groups were compared using the log-rank test. Data pertaining to patients lost to follow-up were censored on the last contact date. HR for the primary endpoint between the BA vs VA group was calculated by the Cox proportionalhazards regression model. Adjusted HR for the primary endpoint between the two groups was calculated by the multivariable Cox proportional hazards regression model adjusted for two prespecified factors (qualifying event, stroke versus TIA; and timing of stenting after the qualifying event, < versus ≥ median time in this cohort), which were potentially associated with the primary endpoint.25 All reported p values were two-sided, and p values <0.05 were considered significant.
Patient and procedure characteristics
The stenting was performed at a median of 26 days after the qualifying event. The rate of successful stent placement was 94.2% (131/139). There was no significant difference in patient and procedure characteristics between the BA (n=69) and VA (n=70) group (table 1).
All patients had clinical follow-up after 30 days, except for 4 patients (3 in the BA group: 2 lost to follow-up at 30 days after stenting and 1 died from ICH within 30 days; and 1 in the VA group, who was lost at 30 days). Mean follow-up time was 23.4 months±20.6 in the BA group versus 26.4 months±21.2 in the VA group (p=0.397).
Sixteen primary endpoint events occurred in this cohort (table 2): 13 (18.8%) in the BA group (9 within 30 days, and 4 afterward), and 3 (4.3%) in the VA group (2 within 30 days, and 1 beyond 30 days) (p=0.006).
Additionally, 6 patients had ischemic stroke in the anterior circulation or ICH after 30 days: 2 (2.9%) in the BA group (2 ischemic strokes) versus 4 (5.7%) in the VA group (3 ischemic strokes and 1 non-fatal ICH) (p=0.681). There were three deaths from non-stroke cause after 30 days, all of which occurred in the BA group.
Follow-up angiography was performed in 19 patients in the BA group at 13.4 months±12.1 versus 33 patients in the VA group at 10.0 months±7.7 (p=0.221). Four (21.1%) restenoses were detected in the BA group versus 12 (36.4%) in the VA group (p=0.249). The symptomatic restenosis rate was 15.8% (3/19; 2 led to TIA, and 1 to ischemic stroke in the VBA territory) in the BA group versus 21.2% (7/33; 6 led to TIA, and 1 to ischemic stroke in the VBA territory) in the VA group (p=0.729).
Primary endpoint between the BA and the VA groups
Figure 2 shows cumulative probabilities of the primary endpoint between the two groups (log-rank statistic=7.16; p=0.008). The probability at 1 year was 22.5% (95% CI 9.7% to 35.3%) in the BA group versus 2.9% (95% CI 0.0% to 7.9%) in the VA group.
HR and adjusted HR for the primary endpoint for the BA versus the VA group (table 2) was 4.63 (95% CI 1.32 to 16.24; p=0.017) and 4.87 (95% CI 1.37 to 17.28; p=0.014), respectively. The significant difference in the primary endpoint between the two groups was noted early within 30 days (13.0% (9/69) vs 2.9% (2/70); p=0.026) (HR=4.64, 95% CI 1.00 to 21.47; p=0.050) (adjusted HR=5.13, 95% CI 1.10 to 23.96; p=0.038).
We did not find that the primary endpoint was associated with stroke as a qualifying event (adjusted HR=3.08, 95% CI 0.47 to 20.06; p=0.240) or stenting on <26 days from the qualifying event (adjusted HR=0.45, 95% CI 0.05 to 4.54; p=0.499) in our multivariable Cox proportional hazards regression model.
This study shows that BA (compared with VA) stenosis is associated with a higher risk of ischemic stroke in the territory of the stented artery, including any stroke or death within 30 days of stenting. Of note is that the significant difference in the primary endpoint between the two groups arose very early, beginning within 30 days of stenting (table 2 and figure 2). The higher peri-procedural stroke risk in the BA group is presumably due to the fact that BA has richer perforators than VA. Stenting may cause occlusion of one or more perforators by squashing plaque into ostia, or injury to fragile perforators by the stent system-induced parent vessel shift.18
This finding should be reliable for the following reasons. The current study mitigated the confounding effect of stenting-procedure volume sites, as the data were derived from a single high-volume center. Additionally, there was no difference in restenosis or symptomatic restenosis on follow-up angiography between the two groups, which could influence clinical outcome.
This finding has important implications. The discrepancy of clinical outcome after stenting between patients with BA and VA stenosis should be considered in clinical practice and future trials comparing stenting with best medical treatment. For example, SAIS patients should be randomized into treatment and control arms of a trial according to specific lesion locus.
Recently, two reports on risk factors for stroke after SAIS stenting have been published. One was the NIH-funded multicenter Wingspan registry. It revealed that patients (n=64) with posterior circulation stenosis had a significantly higher risk of ischemic stroke in the territory of the stented artery (including any stroke or death within 30 days of stenting) than patients (n=96) with anterior circulation stenosis (HR=3.4, 95% CI 1.2 to 9.3; p=0.018).25 The other was a systematic review of the literature. It analyzed the data extracted from 31 articles with a total of 1177 stenting procedures for intracranial stenosis, showing that peri-procedural stroke or death risk within 30 days of stenting for the posterior versus anterior circulation lesion was 12.1% vs 6.6% (OR=1.94, 95% CI 1.21 to 3.10; p<0.01).24 While both showed intracranial VBA stenosis to be the most significant risk factor for stroke after SAIS stenting, neither indicated whether BA or VA lesion was more predictive of stroke after stenting.
The NIH-funded multicenter Wingspan registry study also showed low stenting-procedure volume centers to be associated with higher risk of a subsequent stroke (HR=2.8, 95% CI 1.1 to 7.6; p=0.038); and stroke as a qualifying event (HR=3.2, 95% CI 0.9 to 11.2; p=0.064) or stenting soon (within 10 days) after the qualifying event (HR=2.7, 95% CI 1.0 to 7.8; p=0.058) to be a risk factor with a trend toward significance.25
The current study does not show that the primary endpoint is associated with stroke as a qualifying event or stenting within 26 days after the qualifying event. The former should be explained cautiously. It might be a type II error due to a lower test power (53%), which is related to a lower proportion of patients with stroke as a qualifying event in this cohort. However, the latter should be realistic because the power was large enough (80%). This could be explained by a longer interval from the qualifying event to stenting in this cohort. It is possible that a suitable delay in stenting for patients with recent stroke or TIA event may reduce periprocedural stroke risk, because the recent event may be associated with unstable atherosclerotic plaque that may increase procedural risk.25
Although this is the largest case series of patients with severe intracranial VBA stenosis undergoing elective stenting in the literature to date, the data were derived from a single center so the results might not be generalizable due to differences in neurointerventionist experience and multidisciplinary care. This is not a randomized trial comparing stenting with medical treatment, so it could not directly address the efficacy of stenting. Additionally, we have not compared the clinical outcome of Wingspan stenting with that of balloon-expandable stenting due to the low power of the test. We will do it when the sample size is large enough.
In conclusion, a significantly higher stroke risk poststenting exists in patients with severe BA stenosis than those with VA stenosis. The discrepancy in clinical outcome after stenting between patients with BA and VA stenosis should be considered in clinical practice and stenting trials.
We thank Mr Jing-Qiao Lu, PhD (at Research Center of Clinical Epidemiology, Peking University Third Hospital, Beijing, China) for completing the data analysis.
Funding This study was supported by the Ministry of Health of The People's Republic of China (2004BA714B-7) (20080527) to Dr Wei-Jian Jiang.
Competing interests None.
Ethics approval This study was conducted with the approval of the Beijing Tiantan Hospital Ethics Committee.
Patient consent Obtained.
Provenance and peer review Commissioned; not externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.