Background Stenting of symptomatic intracranial atherosclerosis remains under investigation, yet this option to potentially avert subsequent stroke has been offered at select centers under humanitarian device exemption and off-label use for several years.
Methods Retrospective case series of consecutive patients undergoing stenting with Wingspan and balloon mounted coronary stents for symptomatic intracranial atherosclerosis at a single institution. Recurrent symptomatic ischemia in the territory of the stented artery was ascertained. Rates of recurrent ischemic stroke were calculated per patient-year of follow-up and were compared with medically treated patients in the Warfarin–Aspirin Symptomatic Intracranial Disease (WASID) trial.
Results During the 10 year study period, 41 cases of intracranial stenting were identified. Stenoses were severe (>70%) in 88% of patients. Stenting procedures occurred a median of 14 days from the most recent symptomatic event. 19 Wingspan stents and 22 balloon mounted coronary stents were deployed. Four strokes occurred within 24 h of stenting, seven within 1 month and eight within 3 months. By 3 months after stenting, no further strokes occurred during up to 2 years of follow-up. Patients had 0.194 ischemic strokes per person-year of follow-up, compared with 0.083 ischemic strokes per person-year of follow-up in the aspirin arm of WASID and 0.065 ischemic strokes per person-year of follow-up in the warfarin arm of WASID.
Conclusions Stenting of symptomatic intracranial atherosclerosis in a high risk subset of cases with advanced degree of luminal stenosis may be associated with an increased early risk of recurrent ischemic stroke.
Statistics from Altmetric.com
The Warfarin–Aspirin Symptomatic Intracranial Disease (WASID) trial demonstrated that the risk of recurrent stroke is proportional to the degree of intracranial stenosis.1 In theory, if intracranial stenosis can be reduced by stenting, the risk of stroke may be lowered. Although a randomized trial of intracranial stenting versus best medical therapy is underway,2 data are lacking at present. At least one study of patients with intracranial stents has attempted to compare the rate of recurrent stroke with an historical control—patients treated with medical therapy in the WASID trial.3 That study showed a trend toward a lower rate of recurrent stroke compared with patients in the WASID trial. We describe 41 patients with Wingspan self-expanding or balloon mounted stents and their rate of recurrent stroke up to 2 years after stenting.
Materials and methods
Patients with recurrent ischemic strokes despite best medical therapy were screened for intracranial atherosclerosis according to our standard approach with MR angiography. If MR angiography revealed a high grade stenosis that could plausibly have been the cause of an artery to artery embolism, CT angiography was obtained in patients without a contraindication. Patients with recurrent stroke despite best medical therapy and lesions with >50% stenosis were referred from their neurologist to the interventional neuroradiologist. If the patient was felt by the interventional neuroradiologist to have a lesion amenable to stenting or angioplasty, they underwent cerebral angiograms for angioplasty with or without stenting. Our stroke center logs, starting as early as 27 November 2000, were reviewed for the word ‘stent’ or ‘angioplasty’ or ‘PTA’ (for angioplasty). Patients who underwent angioplasty without stenting were excluded. Patients were also identified using a search for ‘stent’ and/or ‘angioplasty’ in a separate institutional database in order to identify more patients who received stents. A total of 133 cases of intracranial and extracranial stents were identified. To capture more patients, the words ‘thrombectomy’, ‘clot retrieval’, ‘infusion’ or ‘lysis’ were also searched in order to identify any additional patients who may have had intracranial stenting.
Cases of angioplasty for the treatment of vasospasm in the setting of subarachnoid hemorrhage were excluded. Cases of stent assisted coil embolization of aneurysms were excluded. Cases of stent placement for iatrogenic intracranial dissection were included if the stent was placed across the atherosclerotic lesion that was first treated with angioplasty.
One patient had more than one intracranial stent placed on different dates. In this case, the oldest stent that was placed was evaluated with regard to further strokes but not the subsequent stents.
The primary outcome was radiological evidence of new acute stroke on follow-up neuroimaging that was not present prior to stenting. This event was classified as a recurrent stroke after intracranial stenting. However, it was also noted whether or not the patient was symptomatic from their stroke. Most patients underwent routine follow-up imaging in the periprocedural period after stenting. However, the decision to perform further MRI or CT scans after stenting was largely determined by the clinical judgment of the treating physician.
Information on new neurological events at follow-up was assessed by looking at all electronic neurology and interventional neuroradiology follow-up notes, angiography reports and post-procedural MRI reports and inpatient discharge summaries from electronic medical records. Events were classified as strokes if there was MRI or CT evidence of acute stroke during the period of follow-up. Transient unilateral weakness, without evidence of stroke by neuroimaging, was not classified as a recurrent stroke. In some cases, follow-up outcomes were assessed by looking at the history recorded on angiogram reports.
In all 41 patients, intracranial stents were successfully deployed; thus, there was no difference in the technical success rates with the Wingspan and the other stents. Eighty-four per cent of patients with Wingspan stents had >70% stenosis and 90% of patients with other stents had >70% stenosis reports (table 1).
The per cent stenoses of the stented lesions ranged from 63% to 99%. Nine of the 41 angiogram reports did not state the per cent stenosis of the stented lesions but these were described as ‘near total occlusion’ or ‘high grade’. These cases were assigned a value of ‘>70%’ based on this description. Thirty-five patients (88%) were categorized as having >70% stenosis. Five patients (12%) were categorized as having 50–70% stenosis. In one patient, the per cent stenosis was not given in the electronic medical records.
In two patients, a precise time (in days) from stroke could not be determined from the medical records. In 39 patients this time interval was determined. Time between most recent transient ischemic attack (TIA) or stroke to stenting ranged from 1 day to 270 days. The average was 38 days and median time was 14 days (SD 59 days, IQR 33 days).
Nineteen of the stents were Wingspan stents, two were Guidant stents, 13 were AVE stents and two were ACS stents. There was one of each of the following types of stents used: Voyager, Vision cardiac, Palmatz Schatz, Multilink and Medtronic.
One of the 41 patients died before discharge from the hospital after a thrombosis in his basilar artery stent occluded. The patient was started on heparin and died of gastrointestinal bleeding 5 days after placement of the stent.
Thirty-five patients underwent follow-up imaging and six did not. Twenty-five patients had MRIs after stenting. Ten patients had a CT scan alone as follow-up imaging. The average time from the procedure to the follow-up MRI scan was 406 days. The average time from the procedure to the CT scan was 249 days (table 2).
Information about outcome at 24 h was available for 40 of the 41 patients. Of these, four had radiographic evidence of a stroke within the territory of the stent. The remaining 36 patients did not have clinical or radiographic evidence of a stroke within the territory within 24 h of the stent placement. Three additional patients had strokes between 24 h and 1 month of stent placement. Eight patients were lost to follow-up by 1 month after stent placement. One patient had a stroke between 1 and 3 months after stenting. Thus a total of eight patients had strokes within 3 months. Five additional patients were lost to follow-up by 3 months.
Between 3 months and 2 years of follow-up, no additional patients had strokes. Five more patients were lost to follow-up between 3 and 6 months. Six patients were lost to follow-up from 6 months to 1 year. Two patients were lost to follow-up from 1 to 2 years of follow-up. One subject had a stroke within the territory of the stent 5 years after stenting.
Four out of the seven recurrent strokes that occurred after stenting were symptomatic, and in three, strokes were found incidentally with imaging studies. In one patient who presented in a somnolent state due to a basilar artery occlusion and had a new stroke found on MRI within 24 h of stenting, it was not possible to know if she was symptomatic because she was in a lethargic state to begin with. No new signs of stroke were found on neurological examination. One of the symptomatic patients had a pontine infarct after stenting of a basilar artery stenosis which caused mild symptoms of a lateral rectus palsy. His diplopia improved and by 25 days after stenting, he had no further symptoms from that stroke. Another patient had occlusion of their basilar artery with recurrent strokes causing an increasingly somnolent state. The patient died of gastrointestinal bleeding after being started on a heparin drip. Another patient with a small left posterior inferior cerebellar artery infarct after stenting of a vertebral artery had symptoms of vertigo which resolved. This patient, 9 years later, had no disability. One patient with re-stenosis inside of a left middle cerebral artery Wingspan stent presented with dysarthria and aphasia due to a left middle cerebral artery territory infarct (table 3).
Twenty-two patients had no complications. There were a total of 22 complications that occurred in 19 patients (one patient had three complications) (table 4). Ten of the 22 complications were symptomatic. There were five arterial dissections of which three were symptomatic. One of the three symptomatic arterial dissections occurred in a patient who received a Wingspan stent after dissection occurred after angioplasty. Another of the symptomatic dissections occurred in a patient who suffered a dissection after angioplasty and a non-Wingspan type of stent was placed. Two of the dissections were not symptomatic and both of these were in patients who received non-Wingspan stents.
There were three patients who had groin hematomas which required surgery or antibiotics. One patient required readmission to hospital due to a contrast nephropathy that occurred after the angiogram and stent placement. Because of her kidney failure, she required a dialysis catheter in her groin which she traumatically removed while in an agitated state. The patient then developed a groin hematoma from traumatic removal of the catheter which had to be surgically explored.
Eight patients had in-stent re-stenosis (defined as any luminal narrowing within the stent seen on angiogram, regardless of whether or not the patient had symptoms) (table 5). The re-stenosis was confirmed by conventional angiogram in seven of the patients, and in one patient with a Wingspan stent, CT angiography was used to confirm the stenosis. Six of these patients were asymptomatic and three had symptoms that could be attributed to the stenosis. Wingspan stents had a higher rate of re-stenosis than the other stents (37% in the Wingspan group vs 9% in the group that received non-Wingspan stents). Wingspan stents tended to have in-stent stenosis detected longer after stenting (221 days after stenting in the Wingspan group compared with an average of 42 days after stenting in the two patients who had other stents).
One patient had hemorrhagic transformation of their infarct that was found after the stent was placed. This patient was not symptomatic from the hemorrhage. One patient had partial occlusion of a distal vessel by the stent but did not have symptoms from this occlusion. One patient had distal embolization of material during placement of the stent. The medical record does not describe any symptoms from this embolization. One patient had a small intracranial hemorrhage after stent placement but was not symptomatic.
The rate of ischemic stroke, death or hemorrhagic stroke up to 30 days was calculated as follows. There were seven ischemic strokes (two of them asymptomatic) within 1 month. One of the above patients died of gastrointestinal bleeding (thus one patient accounted for two events). There was one asymptomatic hemorrhagic transformation of an ischemic stroke that was found on a routine follow-up CT within 8 h of stenting. Thus the rate of stroke, death or hemorrhage was 22%. Counting only symptomatic strokes and hemorrhages, the rate of stroke, death or hemorrhages was 15%.
Follow-up time ranged from <1 day up to 9 years. Average follow-up time was 13.55 months. Median follow-up time was 3.6 months. The 41 patients had a total of 555.6 months of follow-up (555.6 person-months or 46.3 person-years). Nine patients had clinical and radiographic evidence of stroke during the 46.3 person-years of follow-up (this number includes the subject who had a stroke 5 years after stenting). There were 0.194 patients with stroke per patient-year of follow-up (table 6).
Kaplan–Meier survival curves were used to compare the risk of stroke over time after stenting in our study patients with the WASID group, and to a group of patients that underwent intracranial stenting for intracranial atherosclerotic disease in the National Institutes of Health (NIH) Wingspan stent registry (figure 1). This Kaplan–Meier calculation was made up to 2 years after stenting for the study patients. Thus it excluded subject No 30 who had a stroke 5 years after stent placement. Based on the curves, the risk of recurrent stroke in our patients was higher than patients in the WASID study patients up to 12 months after stenting (table 7).
The rate of stroke among patients with Wingspan stents was compared with the rate of stroke among patients with older less maneuverable cardiac stents (termed ‘non-Wingspan stents’ in this study) (table 8). Patients with Wingspan stents had 0.035 patients with stroke per person-month of follow-up while patients with non-Wingspan stents had 0.016 people with stroke per person-month of follow-up. However, patients with non-Wingspan stents had a longer period of follow-up than patients with Wingspan stents (471 months of follow-up compared with 84 months of follow-up).
Patients with Wingspan stents had fewer strokes during every time period except for the first 24 h during which there were an equal number of strokes. A log rank test did not find a significant difference in the frequency of stroke between Wingspan and non-Wingspan patients (p=0.819).
There was variability in the time between the initial stroke and treatment of the atherosclerotic lesion with stenting. For example, in one case, 270 days elapsed between the time of the patient's first stroke and the time of stenting. This patient was treated with medical therapy, then had a TIA and was referred for stenting <24 h later. In this case, the amount of time between the first stroke and the TIA created a longer ‘time to treatment’. In another case, a patient had a stroke 175 days prior to stent deployment. This patient had strokes on both clopidogrel and warfarin before their treating physicians decided to refer them for stenting. Variability in the time to treatment may be due to uncertainty about when to refer for intracranial stenting. There are no guidelines to suggest how many strokes or TIAs patients should have while on ‘best medical therapy’ before being referred for stenting. This decision is left up to the treating neurologist and interventionalist.
There were 0.194 patients with stroke per year of follow-up in the patients in our study. This is higher than the rate of stroke in the WASID group in which there were 0.083 patients with stroke per year of follow-up in the group that received aspirin and 0.0694 patients with stroke per patient- year of follow-up in the group that received warfarin (table 8).
There are several possible explanations for the higher rate of recurrent stroke seen in our study patients. The patients in our study had on average more severe intracranial stenosis; 88% of patients in our study has >70% stenosis compared with approximately 37% of patients in the WASID study. This difference was statistically significant. It was shown in WASID that the degree of intracranial stenosis is correlated with the risk of stroke. Therefore, our patients may have been at higher risk of stroke. Also, our study had significantly more ‘non-white, non black’ patients (termed in WASID and our paper as ‘other’). We do not know if intracranial stenting is less effective in these ethnicities or if the higher rate of recurrent stroke in this ethnicity was reflective of an association with more severe atherosclerotic disease.
We included patients who were asymptomatic after recurrent stroke in our study (patients who had evidence of infarction on imaging but in some cases without symptoms) while WASID and the NIH registry on the use of the Wingspan stent did not count these asymptomatic patients.1 3 It is likely that there was bias for patients with recurrent strokes and symptoms suggestive of TIA to come back to our center for further evaluation. Patients who were asymptomatic after stent placement were likely followed-by the neurologists and primary care physicians in the community and may have been less likely to be referred back to our center for follow-up. This referral bias is a second possible explanation for the higher rate of recurrent stroke seen in our study patients. A future study should call patients and review outpatient medical records from other institutions to determine if the patients who were lost to follow-up in this study continued to have ischemic strokes. We suspect that many of the patients who were not referred back to our center did not have recurrent strokes.
The patients were stratified according to whether they received a Wingspan or an older balloon mounted cardiac type of stent. The older stents were less maneuverable in patients with tortuous intracranial and extracranial arteries, and periprocedural morbidity and mortality rates of 20–30% have been reported.4 5 The purpose of stratifying the analysis was to see if the higher rate of stroke compared with WASID might be due to the fact that the older stents were included in our analysis. There were more strokes per patient-month of follow-up in the Wingspan group, which was not an expected result. However, this higher rate could be due to the longer follow-up period in the non-Wingspan group. Several patients in the non-Wingspan stent group survived many years without strokes and this may be partly because these patients have had a longer time to live with their stents than patients who more recently received Wingspan stents. Thus this longer follow-up period may make the denominator appear larger which may make the rate of stroke appear larger for Wingspan stents. Overall, a greater proportion of strokes occurred in patients with non-Wingspan stents. Although a log rank test did not find a significant difference in the frequency of stroke between the Wingspan and non-Wingspan groups, it is possible that a difference exists that was too small to detect with our sample size. Patients with Wingspan stents more frequently experienced in-stent re-stenosis, and in one patient this re-stenosis was associated with a stroke. It is possible that if a higher rate of recurrent stroke after stenting with balloon mounted coronary stents exists, it could have been negated by a higher rate of in-stent re-stenosis and associated recurrent stroke in patients with Wingspan stents.
We did not measure a difference in the technical success of stenting with Wingspan versus the other types of stents as all 41 patients were able to have stents deployed across their lesions. It is possible that some patients in whom stenting was attempted unsuccessfully were missed by the search terms used to screen the logbooks of our institution. We do not know if the lesions which were stented with Wingspan stents were technically more difficult because of lesions that were less accessible to microcatheters due to more tortuous arteries. The patients with Wingspan stents had roughly the same proportion of lesions with >70% stenosis as the other patients (84% of Wingspan patients had >70% stenoses vs 90% of other patients). Therefore, at least by this measure of difficulty, the two groups were similar.
The rate of any ischemic stroke, hemorrhagic stroke or death within 30 days in our study (22%) was higher than in the NIH Wingspan registry (9%).3 This can be partially explained by the fact that we included asymptomatic ischemic and hemorrhagic strokes in our study while the NIH Wingspan registry did not. Including death and only symptomatic stroke or hemorrhagic stroke, the rate was 15%. Our study included the less flexible older cardiac stents which in theory are more difficult to deploy intracranially. A slightly higher proportion of patients with non-Wingspan stents had strokes within 1 month (22.7% vs 15.7% in the Wingspan group) and this may be due to an increased associated risk with the older stents. More studies are required to see if this finding is replicated in the experience of other medical centers.
The rate of recurrent stroke per patient-year of follow-up in our study patients was higher than in the WASID trial. The patients in our study had a higher degree of intracranial stenosis on average and this higher degree of stenosis may have placed them at higher risk of recurrent stroke than patients in the WASID trial. The number of patients lost to follow-up in this study is a weakness of the study. There may also be bias in the data because patients with recurrent symptoms may have been followed-up more frequently than asymptomatic patients.
Stenting of symptomatic intracranial atherosclerosis remains under investigation yet this option to potentially avert subsequent stroke has been offered at select centers under humanitarian device exemption and off-label use for several years. SAMMPRIS is a randomized trial of stenting with the Wingspan stent verses medical management. SAMMPRIS investigators will soon publish the preliminary results of their trial. Until long term follow-up of the SAMMPRIS patients is complete, our study of patients with recurrent stroke due to intracranial atherosclerosis may inform clinical decisions about intracranial stenting.
Stenting of symptomatic intracranial atherosclerosis in a high risk subset of cases with advanced degree of luminal stenosis may be associated with an increased early risk of recurrent ischemic stroke. In this small series of 41 patients, there was no significant difference in the rate of recurrent stroke in patients with a variety of balloon mounted cardiac stents versus the Wingspan stent.
Competing interests None.
Ethics approval This study was conducted with the approval of the UCLA.
Provenance and peer review Not commissioned; 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.