Vertebral Artery Ostium Atherosclerotic Plaque as a Potential Source of Posterior Circulation Ischemic Stroke
Result From Borgess Medical Center Vertebral Artery Ostium Stenting Registry
Background and Purpose—Although atherosclerotic plaque in the carotid and coronary arteries is accepted as a cause of ischemia, vertebral artery ostium (VAO) atherosclerotic plaque is not widely recognized as a source of ischemic stroke. We seek to demonstrate its implication in some posterior circulation ischemia.
Methods—This is a nonrandomized, prospective, single-center registry on consecutive patients presenting with posterior circulation ischemia who underwent VAO stenting for significant atherosclerotic stenosis. Diagnostic evaluation and imaging studies determined the likelihood of this lesion as the symptom source (highly likely, probable, or highly unlikely). Patients were divided into 4 groups in decreasing order of severity of clinical presentation (ischemic stroke, TIA then stroke, TIA, asymptomatic), which were compared with the morphological and hemodynamic characteristics of the VAO plaque. Clinical follow-up 1 year after stenting assessed symptom recurrence.
Results—One hundred fourteen patients underwent stenting of 127 lesions; 35% of the lesions were highly likely the source of symptoms, 53% were probable, and 12% were highly unlikely. Clinical presentation correlated directly with plaque irregularity and presence of clot at the VAO, as did bilateral lesions and presence of tandem lesions. Symptom recurrence at 1 year was 2%.
Conclusions—Thirty-five percent of the lesions were highly likely the source of the symptoms. A direct relationship between some morphological/hemodynamic characteristics and the severity of clinical presentation was also found. Finally, patients had a very low rate of symptom recurrence after treatment. These 3 observations point strongly to VAO plaque as a potential source of some posterior circulation stroke.
Early literature suggested that vertebral artery ostium (VAO) lesions do not cause artery-to-artery embolism. Rather, its sole symptomatic mechanism, it was assumed, was attributable to flow restriction when the stenosis became severe, but even this was felt to be rare because of the presence of the 2 vertebral arteries that feed the basilar artery as well as the circle of Willis. Eventually, this perspective started to be shaken and a few case series began to appear describing emboli originating from occlusive disease at the vertebral artery ostium.1–3 Even a large ulcer at the VAO was described in a case report, adding to the possibility of VAO as a source of emboli.4
The first innovative study examining the possibility of emboli originating in the vertebral artery came from the New England Medical Center Posterior Circulation Registry, in which patients with posterior circulation ischemic stroke were subjected to the same rigorous diagnostic evaluation as patients with anterior circulation ischemic stroke.5 The New England Medical Center Posterior Circulation Registry demonstrated that embolism was the most common mechanism of posterior circulation ischemic stroke (40% of the cases), 14% of which were attributable to artery-to-artery embolism, half from atherosclerotic stenosis at the VAO and half in the intracranial vertebral artery.5 In 9% of their patients, the VAO lesion was the only detected cause of ischemic stroke.6
The role of the VAO lesions in posterior circulation ischemic stroke continues to be debated and is still poorly understood. We aim to demonstrate the similar correlation between VAO plaque morphological characteristics and clinical presentation to the correlation known to exist between plaque and ischemic risk in carotid and coronary artery diseases. By doing so, we hope to shed further light on this important source of posterior circulation ischemia.
Materials and Methods
Study Design and Subject Eligibility
This Borgess Medical Center Vertebral Artery Ostium Stenting Registry is a nonrandomized, single center, single operator, investigator-initiated registry of 122 consecutive symptomatic patients presenting with posterior circulation ischemic stroke and proven to have VAO atherosclerotic plaque stenosis (≥50%). The protocol was approved by the Institutional Review Board and all patients gave written informed consent for the procedure. Data were entered prospectively and retrospectively into a secure outcome database (MD Analyze, Medtech Global, Ltd) and analyzed retrospectively. The site had Institutional Review Board approval to waive informed consent for the data collection and data analysis.
Subjects included in this study were 40 years of age or older, had TIA or ischemic stroke or both in the posterior circulation that was attributable to angiographically verified ≥50% stenosis at 1 of the VAO within the last 3 months, and underwent stenting of this lesion. Four asymptomatic subjects were included because of the severity of their stenosis, which was considered to be severe enough to merit stenting to prevent occlusion of this important vessel before the patients became symptomatic. Exclusion criteria included a coexisting condition that could limit survival, contraindication to aspirin or Plavix (Sanofi Aventis), and clear evidence that the ischemic stroke was the result of another cause, such as vertebral artery dissection or embolism from another location.
Borgess Medical Center is a high-volume acute ischemic stroke center. We see ≈300 stroke cases and 300 TIA cases each year, and ≈20% of those cases are in the posterior circulation. During the timeframe of this registry, ≈900 patients presented to Borgess Medical Center with posterior circulation ischemic stroke, and 110 symptomatic and 4 asymptomatic patients underwent stenting. All patients admitted to Borgess Medical Center with posterior circulation ischemia were evaluated by either a neurologist or an internist. Imaging studies were ordered, including MRI and MRA with or without contrast, depending on the case, and in some cases CTA was also ordered. If there was any suspicion of a cardiac cause of the symptoms (if the patient had a history of coronary artery disease or arrhythmia), then cardiac evaluation was obtained. In certain cases with high clinical suspicion in which noninvasive vascular imaging studies (MRA or CTA) were either not performed or not conclusive, we proceeded with diagnostic cerebral angiogram directly because it gives the best evaluation of the VAO and allows for treatment if significant stenosis is found. Stenting was performed if significant stenosis was found and felt to be reasonably responsible for the patient's symptoms at the discretion of the operating physician. Drug-eluting stent was used for all vessels with a diameter ≤3.5 mm, and bare metal stent was used in the other cases.
Patients were required to have a follow-up catheter angiography at the end of the dual antiplatelet therapy period (3 months for bare metal stent or 6 months for drug-eluting stent) to evaluate the rate of restenosis and to assess the need for a repeat intervention. Patients were required to have a follow-up office visit 1 year after stenting.
Following the New England Medical Center Posterior Circulation Registry, patients were divided into 4 groups in decreasing order of severity of clinical presentation. Group I included patients presenting with TIA followed by ischemic stroke, group II included patients presenting with ischemic stroke as the first symptoms, group III patients presented with TIA only, and group IV patients were asymptomatic.
MRI and CT Studies
MRI and CT scans were reviewed for signs of recent infarction. This was defined as diffusion-weighted image abnormality on the MRI scan or hypodensity with mass effect on the CT scan.
Preintervention Diagnostic Cerebral Angiogram
Cerebral angiography documented >50% stenosis on all lesions. Two other key characteristics were recorded. First, morphological changes at the plaque level (including the severity of stenosis, concentric versus eccentric, smooth versus irregular, haziness, calcification, overhanging, and the presence or absence of a clot at the VAO or intracranially) were classified according to the definitions of these terms set out in coronary literature.7–12 The second key feature noted was the hemodynamic situation of the posterior circulation blood supply, which included: bilateral VAO lesions, defined as any evidence of atherosclerotic stenosis at both VAO, including occlusion but independent of the severity; the presence of tandem lesion, defined as >50% stenosis at the ipsilateral vertebral artery; basilarization of the vertebral artery, defined as only 1 vertebral artery providing anterograde flow to the basilar artery, attributable to acquired or congenital occlusion of the other vertebral artery intracranially or extracranially; and status of the circle of Willis. Two authors (F.A.A., T.B.) recorded these morphological and hemodynamic characteristics by consensus. All of these findings were examined and determined for a second time by consensus of 3 authors (F.A.A., T.B., S.S.) 5 months later, blinded to the first classification. Agreement tests between the 2 morphological classifications were obtained by statistical analysis (percent agreement and kappa test) to document the consistency of the findings. No agreement test for the hemodynamic variables was obtained because these variables are much clearer and easier to characterize than the morphological changes.
Cardiac and Atherosclerotic Risk Factors Evaluation
Data concerning medical history and risk factors for atherosclerosis were collected and included: hypertension, defined as ≥140/80 mm Hg; history of diabetes, confirmed if the patient required insulin or was using oral hypoglycemic; hyperlipidemia, defined as low-density lipoprotein ≥3.3 mmol/L; coronary artery disease, considered present if the patient had a history of angina, myocardial infarction, or coronary artery stenosis on catheter angiography; and cardiac arrhythmia or recent cardiac echogram for search of clot in the left atrium.
VAO Lesion as the Source of the Presenting Symptoms of the Patient
Taking all of these data into consideration, the likelihood that the TIA or ischemic stroke was caused by the stenosis in the VAO was categorized as highly likely, probable, or highly unlikely. The relationship between the VAO stenosis and the patient's symptoms was classified as highly likely if the patient was symptomatic, had a cardiac evaluation showing no evidence that the embolism was cardiac in origin (demonstrated by no atrial fibrillation, no clot in the left heart, no artificial valve, and so on). Patient symptoms were also classified as highly likely if a catheter angiography demonstrated no evidence of intracranial tandem lesion if a catheter angiography indicated a clot at the VAO lesion, and if MRI or CT demonstrated ischemic stroke in the corresponding posterior inferior cerebellar artery territory (lesions in the posterior inferior cerebellar artery territory were considered localizing to the ipsilateral VAO lesion).
The relationship was considered probable if multiple causes could explain the symptoms, such as cardiac evaluation showing evidence that the embolism may have been cardiac in origin (demonstrated by atrial fibrillation, clot in the left heart, artificial valve, and so on). If no cardiac evaluation was obtained or if cerebral angiography demonstrated evidence of intracranial tandem lesion that could be causing the symptoms, then we could not definitively exclude these other possible causes of the symptoms. The relationship was determined to be highly unlikely if the patient was asymptomatic or if the patient was symptomatic with imaging findings suggesting a recent ischemic stroke in the counterlateral posterior inferior cerebellar artery territory or in the anterior circulation.
Restenosis and Reintervention Rates
All VAO stents that exhibited ≥50% stenosis on follow-up diagnostic cerebral angiogram were considered to have restenosis. Reintervention rate was calculated as the percentage of VAO stents that subsequently underwent either angioplasty or restenting. Because the risk of emboli of recurrent stenosis is relatively low, reintervention was performed only for the hemodynamically significant restenoses, defined as either ≥50% restenosis in a patient with less than optimal blood supply to the posterior circulation (including patients with occlusion of the counterlateral vertebral artery or with poor status of the circle of Willis) or ≥70% restenosis for a patient with optimal blood supply.
We first performed a percentage agreement test between the 2 separate classifications of morphological characteristics of the VAO lesion. Then, to check for a possible relationship between morphological/hemodynamic variables and clinical presentation, we used the χ2 test, Fisher exact test, logistic regression with odds ratios, and descriptive statistics. For the purpose of the analysis and because of the relatively small sample size in groups I and IV, we combined groups I and II (patients with ischemic stroke) and groups III and IV (patients without ischemic stroke). The α value used to determine statistical significance was 0.05.
From April 2002 through January 2010, a total of 114 patients underwent stenting of 127 VAO lesions. One case was technically unsuccessful, leaving 126 lesions and 113 patients for angiographic and clinical follow-up. Of the 114 patients, 110 presented with posterior circulation ischemia, whereas the other 4 patients presented with unrelated symptoms but were found to have significant, albeit asymptomatic, stenosis at the VAO. The 126 stenting cases included 45 drug-eluting stents and 81 bare metal stents. There were 78 men (68%) and 36 women (32%). Of the 110 symptomatic patients, 35 (32%) had no cardiac evaluation, presumably because the treating or consulting team or both did not suspect that the symptoms were attributable to a cardiac cause.
When dividing the symptomatic patients based on the severity of their clinical presentations, group I had 4 patients (3.5%), group II had 43 patients (39%), group III had 63 patients (57%), and group IV had 4 patients (3.5%). We tracked 9 different symptoms consistent with posterior circulation TIA, and 19 of 63 (30%) had at least 3 symptoms and 42 of 63 (67%) had at least 2 symptoms (Supplemental Table I, http://stroke.ahajournals.org).
Of the 126 lesions, 44 (35%) were found to be highly likely the source of symptoms, 67 (53%) were classified as the probable cause of symptoms, and 16 lesions (12%) were considered highly unlikely as the source of the symptoms. The morphological changes of the VAO lesion and the agreement tests between the 2 readings are shown inTable 1. The agreement tests showed overall satisfactory agreement between the 2 readings (67%–99%). The hemodynamic variables are shown in Table 2.
The second statistical analysis correlated the VAO plaque morphological changes and the hemodynamic characteristics with the clinical presentation (Table 3). Plaque irregularity was found to correlate with more severe clinical presentation (P=0.021). The presence of a clot in the posterior circulation was also found to correlate with severity of clinical presentation (P=0.023). A clot at the VAO was found only in patients who presented with ischemic stroke, but it was not statistically significant as a predictor of clinical presentation (P=0.153), which could be attributable to the small sample size. Two hemodynamic variables correlated significantly with the severity of clinical presentation: the presence of tandem lesions (P=0.002) and the presence of bilateral lesions (P=0.001). The logistic regression model showed a similar finding (Table 4). Irregularity, bilateral lesions, and the presence of a tandem lesion all correlate with more severe clinical presentation (P=0.016, P<0.001, and P=0.005 respectively).
One hundred two (89%) patients had clinical follow-up at 1 year. Two patients had ischemic stroke in the posterior circulation during that time (2%). Of the 126 stents deployed, 85 (67%) had angiographic follow-up during the appropriate timeframe; 21 of 85 stents (25%) showed ≥50% stenosis. Eleven of the 85 follow-up cases (13%) underwent reintervention. drug-eluting stent, and bare metal stent had very similar restenosis rates at 26% and 24%, respectively (Table 5).
The direct relationship between specific morphological changes at the atherosclerotic plaque level and the subsequent risk of ischemia is well-established in the literature. These include degree of stenosis, irregularity, haziness, ulceration of the plaque, and signs of active thrombosis, such as clot. The risk of ischemia increases almost in a linear fashion with the increase in degree of stenosis.11,13,14 However, this risk is not only related to the degree of stenosis but also depends on the plaque characteristics. Irregularity of the plaque, which is believed to be attributable to plaque inflammation and vulnerability, is highly associated with the risk of subsequent ischemia.11,15,16 The presence of clot next to or near a critical stenosis is the hallmark of fresh plaque thrombosis and it is also well-documented to correlate with acute ischemia.8,11 The importance of these 3 factors needs to be emphasized for the intracranial circulation.7–17
To our knowledge, the Borgess Medical Center Vertebral Artery Ostium Stenting Registry is the first large prospective study that tries to correlate morphological changes in the VAO plaque and the clinical presentation (groups I–IV) in an attempt to establish a similar relationship between atherosclerotic plaque characteristics and ischemia as seen in other vessels. As expected from atherosclerotic plaque, lesion irregularity correlated with the severity of symptoms. A clot at the VAO was found only in patients with recent ischemic stroke, which echoes the findings common during coronary angiogram for acute myocardial infarction. Despite the rarity of cases of clots in our registry (3%), the presence of clot at the VAO plaque provides the clearest indication of the cause-and-effect relationship between VAO lesions and posterior circulation ischemic stroke.
Neither eccentricity nor degree of stenosis correlated with clinical presentation. Coronary literature has demonstrated in the past that eccentric smooth lesions were not associated with severe clinical presentation.7–12 We expected the degree of stenosis to correlate with the severity of clinical presentation, but the reason for the lack of correlation is most likely attributable to the bias introduced in our study by selecting only patients with ≥50% stenosis; 83% of patients had ≥70% stenosis and 4% had ≥90% stenosis, so our data are not suited to assess the significance of the degree of stenosis at the VAO.
As in the New England Medical Center Posterior Circulation Registry, we found a correlation between the hemodynamic changes of the posterior circulation and the clinical presentation, including the presence of a tandem lesion in the vertebral artery and bilateral lesions, which may be because the presence of atherosclerotic plaque in 2 locations makes the probability of embolism higher.
When we classified the probability of the VAO lesions as the source of the symptoms, we reached almost an identical conclusion to the New England Medical Center Posterior Circulation Registry, in which the VAO stenosis was highly likely the source of ischemic stroke in 43% of the patients compared to 35% in our series. Also, 53% of the lesions in Borgess Medical Center Vertebral Artery Ostium Stenting Registry were in the probable category, whereas 57% were probable in their series7 (Table 6). Additionally, the location of the stroke was also similar in both series, with 44.1% of our subjects having stroke in the proximal region alone versus 32% in their series. The combination of proximal and distal was the second most common presentation in both series, with 55.8% versus 37%, respectively, which argues for the embolic nature of posterior circulation stroke in these subjects.
In addition, the clinical follow-up in our patients at 1 year documented the resolution of the presenting symptoms and the paucity of symptom recurrence after stenting. The 2% recurrence rate within 1 year provides additional and valid evidence for the implication of the presenting VAO plaque in the patients' symptoms.
Our overall restenosis rate of 25% is actually comparable to the restenosis rate in other vessels with similar diameter, including coronary and intracranial vessels.18,19 We believe this is an important observation because it suggests that the widely held perception that VAO lesions have a much higher restenosis rate than other vessels is most probably inaccurate, but this observation needs to be confirmed by other larger studies.
This study, however, has certain limitations. Not every patient had a cardiac evaluation to rule out a cardiac source of emboli; to counter this issue, we included those patients in the probable group, despite the lack of a clear sign of cardiac disease. Furthermore, we are suspicious that some of the patients who had a negative cardiac evaluation result and whose VAO plaque was classified as highly likely the source of the symptoms could have aortic arch plaque that went undetected. Also, the symptoms of posterior circulation TIA are vague and nonspecific, and when they lack an imaging confirmation of ischemic insult it is possible that they are attributable to other disease processes. To counter this limitation, most of these patients had multiple signs and symptoms of posterior circulation ischemia and most of the VAO lesions were classified as probable as opposed to highly likely as the source of the patients' symptoms. Our classification of the VAO plaque as the source of symptoms is not perfect, and the distinction between probable and highly likely is especially not perfect, but this is the best system we could define. One more limitation is that the determination of the morphological characteristics of the VAO plaque in this article is subjective, but the acceptable interobserver concordance suggests that these definitions are both reproducible and useful.
Finally, this is not an epidemiological study; therefore, we cannot infer from it how often the VAO plaque is the origin of posterior circulation ischemia. A much larger study with a different design is needed to examine how often VAO plaque causes posterior circulation ischemia and which plaque characteristics correlate with future risk of ischemic stroke.
In this selected group, 35% of patients with posterior circulation ischemic stroke and VAO atherosclerotic stenosis had no other valid explanation for their symptoms except the VAO lesion, and a correlation was found between certain atherosclerotic plaque morphologies and the severity of ischemic presentation similar to that in other vessels. These 2 findings argue strongly for a possible cause-and-effect relationship of the VAO atherosclerotic plaque and posterior circulation ischemic stroke in certain cases. In addition, the rarity of symptom recurrence after stenting of this lesion adds validity to this observation. It appears to us that the old concept of the benign “nonembolic” behavior of the VAO plaque is a remnant of an outdated opinion, and we think the time has come to treat atherosclerotic lesions at the VAO the same way we treat them anywhere else in the human body.
The online-only Data Supplement is available at http://stroke.ahajournals.org/cgi/content/full/STROKEAHA.110.610451/DC1.
- Received December 23, 2010.
- Accepted April 6, 2011.
- © 2011 American Heart Association, Inc.
- Caplan LR,
- Amarenco P,
- Rosengart A,
- Lafranchise EF,
- Teal PA,
- Belkin M,
- et al
- Pessin MS,
- Daneault N,
- Kwan ES,
- Eisengart MA,
- Caplan LR
- Koroshetz WJ,
- Ropper AH
- Levin DC,
- Fallon JT
- Wilson RF,
- Holida MD,
- White CW
- Kasner SE,
- Chimowitz MI,
- Lynn MJ,
- Howlett-Smith H,
- Stern BJ,
- Hertzberg VS,
- et al
- Zhang L,
- Huang QH,
- Zhang YW,
- Liu JM,
- Hong B,
- Xu Y,
- et al
- Siddiq F,
- Vazquez G,
- Memon MZ,
- Suri MF,
- Taylor RA,
- Wojak JC,
- et al