Article Text


Original research
Arterial occlusions increase the risk of in-stent restenosis after vertebral artery ostium stenting
  1. Jingzhi Li1,2,
  2. Yang Hua1,
  3. Laurence Needleman2,
  4. Flemming Forsberg2,
  5. John R Eisenbray2,
  6. Zhaojun Li3,
  7. Ran Liu1,
  8. Xiaojie Tian1,
  9. Liqun Jiao4,
  10. Ji-Bin Liu2
  1. 1Department of Vascular Ultrasonography, Xuanwu Hospital, Capital Medical University, Beijing, China
  2. 2Department of Radiology, Thomas Jefferson University, Philadelphia, Pennsylvania, USA
  3. 3Department of Ultrasound, Shanghai General Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China
  4. 4Department of Neurosurgery, Xuanwu Hospital, Capital Medical University, Beijing, China
  1. Correspondence to Professor Yang Hua, Department of Vascular Ultrasonography, Xuanwu Hospital, Beijing 100053, China; dryanghua99{at}


Objective The study was designed to investigate if vascular occlusion in the internal carotid artery (ICA) or the contralateral vertebral artery (VA) contribute to developing in-stent restenosis (ISR) in patients with vertebral artery ostium stenosis (VAOS).

Methods 420 consecutive patients treated with VAOS stents (from a population of 8145 patients with VAOS) from January 2013 to December 2014 were analyzed in this retrospective study; 216 with drug eluted stents and 204 with bare metal stents. Based on pre-stent DSA findings, patients were divided into four groups: both carotid and vertebral arteries patent (PAT), ICA occlusion (ICA-OCC), contralateral VA occlusion (CVA-OCC), and combined occlusions (C-OCC). The incidence of ISR (stenosis >50%) was compared between groups using Cox regression analysis.

Results Of the 420 patients, the mean incidence of ISR was 36.4%, with a median 12 months of follow-up (IQR 3–12). Logistic regression analysis showed that drug eluting stent had less ISR than bare metal stent (OR=0.38, 95% CI 0.19 to 0.75, P=0.01). Cox regression analysis showed that CVA-OCC (HR=1.63, P=0.02) and C-OCC (HR=3.30, P=0.001) were risk factors for ISR but not ICA-OCC (P=0.31). In the CVA-OCC and C-OCC groups, in-stent peak systolic velocity (PSV) ≥140 cm/s, 1 day after successful stenting, was associated with subsequent development of ISR (OR=2.81, 95% CI 1.06 to 7.43, P=0.04).

Conclusion Contralateral VA occlusion at the time of stenting increased the risk of ISR, especially if stent PSV on day 1 was >140 cm/s. Bare metal stents had more ISR than drug eluting stents.

  • stroke
  • ultrasound
  • stent
  • stenosis

Statistics from


According to recent epidemiological data, approximately 16.9 million people have a stroke each year, equivalent to a global incidence of 258/100 000/year.1 Of these strokes, approximately 20% are from the posterior circulation.2 Compared with an anterior circulation stroke, a posterior circulation stroke is more difficult to diagnose and refractory to medical treatment, resulting in a higher incidence of recurring strokes.3 Vertebral artery ostium stenosis (VAOS) is one of the major factors contributing to posterior circulation ischemic events4 and thus VAOS stenting has become an alternative to standard medical management. Multiple clinical studies have demonstrated the safety and efficacy of this stenting procedure5 6 but with an incidence of in-stent restenosis (ISR) ranging from 12% to 48%.5 7–9

The risk of an ISR is one of the main limitations of widespread application of VAOS therapy. Previous investigations have identified factors that increase the likelihood of ISR, such as arterial anatomy, medical management, and stent type,10 11 while also illustrating that drug eluting stents have less ISR compared with bare metal stents.12 However, the effect associated with carotid artery stenosis or occlusion in VAOS stent patients is poorly understood. The aim of this study was to determine if coexistent carotid vascular occlusions is associated with ISR following stenting for symptomatic VAOS.


Patient selection

This retrospective cohort study was approved by the ethics committee of Xuanwu Hospital, Capital Medical University, and no informed consent was required. The study reviewed 8145 patients with VAOS from January 2013 to December 2014. The inclusion criteria for our study were as follows: (1) patients with severe VAOS (stenosis ≥70%) according to the vertebral stenosis criteria13 (ostium site peak systolic velocity (PSV) ≥210 cm/s, and PSV ratio (vertebral ostium stenosis PSV divided by intervertebral segment PSV) ≥4.0), (2) vertebral stenosis confirmed by DSA, (3) vertebral artery (VA) stent placed in our center, and (4) no prior VA or other stent. Exclusion criteria included: (1) residual stenosis (≥50% stenosis after stenting), (2) severe ipsilateral subclavian artery stenosis or occlusion, and (3) history of neck irradiation therapy.

Based on their DSA results, patients were divided into four groups: (1) patent group (PAT)—bilateral internal carotid artery (ICA) and contralateral VA were patent, stenosis may be present; (2) ICA occlusion group (ICA-OCC)—with ipsilateral or bilateral ICA occlusion; (3) contralateral VA occlusion group (CVA-OCC), or (4) combined group (C-OCC)—with ipsilateral, contralateral, or bilateral ICA occlusion and contralateral VA occlusion.

DSA and stenting

All patients were fully informed of the angiography and stent procedure and signed a consent form. All procedures were performed by experienced interventional neurosurgeons (performed at least 30 cases annually). The indications for stent procedure included: (a) the target VA was responsible for patients with posterior transient cerebral ischemia (TIA) or stroke, (b) anterior TIA or stroke and the target stenosed VA was potentially a collateral through the posterior communication artery, (c) during angiography, the carotid artery system was assessed, including the common carotid artery, ICA, and external carotid artery, and both VAs. If necessary, superselective angiography, rotating DSA, or three-dimensional DSA was used to evaluate suspicious stenoses. All DSA and VAOS procedures were performed using the Neurostar Plus/TOP dual C-arm angiography system (Siemens, Munich, Germany). According to the North American symptomatic carotid endarterectomy test method (NASCET),14 the degree of vascular stenosis was classified as <50%, 50–69%, or 70–99% stenosis, or occlusion.

All patients underwent balloon angioplasty prior to stent implanting. Stent type was selected according to standard of care as well as the specific characteristics of the stenotic lesion, and included balloon expanded stents (BES) (Cipher (Cordis Corp. Bridgewater, New Jersey, USA), Apollo (MicroPort Scientific Corp, Shanghai, China), Endeavor (Medtronic Inc, Galway, Ireland), Firebird (MicroPort Scientific Corp, Shanghai, China), Palmaz Blue (Cordis Corp, Oosteinde LJ Roden, The Netherlands), and Xience V (Abbott Laboratories, AbbottPark, Illinois, USA)), and self-expanded stents (SES) (Protégé (EV3 Endovascular, Plymouth, Minnesota, USA), Acculink (Abbott Laboratories, Abbott Park, Illinois, USA), and Precise (Cordis Endovascular, Bridgewater, New Jersey, USA)). All patients were administered aspirin 100 mg/day and clopidogrel 75 mg/day for at least 3 months after the procedure. In addition, patients underwent treatment for hypertension, hyperlipidemia, diabetes, and other factors to minimize atherosclerotic risk.

Ultrasound imaging and follow-up

All scanning was conducted by physicians who had extensive experience in vascular ultrasound (at least 3 years of clinical experience of more than 3000 vascular cases per year). Color Doppler examinations were performed with an IU22 ultrasound system (Philips Medical Systems, Amsterdam, The Netherlands) using a 4.0–8.0 MHz or 3.0–9.0 MHz linear array probe or a 2.0–5.0 MHz convex array probe. Grayscale, color, and spectral Doppler images were obtained from the entire accessible segments of the ICAs and VAs. All spectral sampling angles were adjusted to ≤60°. Representative common carotid artery and external carotid artery images were obtained. All acquisition images were stored in the PACS system for subsequent analysis.

VA parameters included residual and unobstructed diameters, peak systolic velocity (PSVos) and end diastolic velocity (EDVos) of the VA ostium, and peak systolic peak velocity (PSViv) and diastolic velocity (EDViv) of the intervertebral segment (C4–5 or C3–4). The ultrasound examination was carried out before surgery and on day 1, and then approximately 3, 6, 12, and 24 months after stent placement. All cases were reviewed twice in the PACS system by two different physicians unaware of the prior readings.


Event occurrence was ISR >50% during follow-up; all patients ended follow-up at 24 months. All color Doppler results were diagnosed according to published criteria for ISR,15 defined as PSV ≥170 cm/s, EDV ≥45 cm/s, and PSV ratio ≥2.7.

Reproducibility and variability

For evaluation of intra-observer variability of PSV measurements, one of the two investigators repeated the measurement in 50 randomly selected patients a week after they had performed the first set of measurements. For inter-observer variability, we randomly selected 50 cases, and the measurements were repeated by the observer who had not performed the previous set of measurements.

Statistical analysis

Data analysis was performed using SPSS Statistical Software V.22.0 (IBM Corporation, Armonk, New York, USA). All continuous variables were compared using the one sample Kolmogorov–Smirnov test. Normally distributed continuous variables are expressed as mean Embedded Image±SD, and variables that were not normally distributed are expressed as median (IQR). The χ2 test or ANOVA test was used to compare age, gender, hypertension, hyperlipidemia, diabetes mellitus, smoking, stent type, stent length, stent diameter, and procedure side. Logistical regression was used to identify the significant parameters and then calculated ORs for each variable. The Student’s t test or Kolmogorov–Smirnov test was used to compare differences in Doppler parameters between each group in terms of pre- and post-stenting and restenosis. Doppler velocity parameters were analyzed by ANOVA or the Kruskal–Wallis test among occlusion groups. Kaplan–Meier curves and the log rank test were used to show differences in ISR incidence between the groups. The hazard ratio (HR) and 95% CI of the different occlusion groups were assessed by Cox regression. Statistical tests were two sided, and a P value < 0.05 indicated a significant difference.


Patient enrollment and clinical characteristics

A total of 420 patients (353 men and 77 women, mean age 65 years) met the inclusion criteria (see online supplementary figure 1 listed in the data supplement). We found that 401 patients had posterior TIA or stroke, the other 19 patients had anterior TIA or stroke, and the posterior communication artery was shown to be patent on DSA before the stenting procedure. The average length of the stent used was 15.5±5.4 mm, and the average diameter was 4.5±1.1 mm. There were 216 drug eluting stents (DES, 51.4%) and 204 bare metal stents (BMS, 48.6%) implanted. An example of a case with pre-stenting and post-stenting DSA images is shown in figure 1. The clinical data, along with group information, are summarized in table 1. Logistic regression results showed no effect on ISR based on age, gender, diabetes, hypertension, hyperlipidemia, smoking, stent type II (SES vs BES), stent length, stent diameter, or procedure side (P>0.05). Stent type I (DES vs BMS, OR=0.31, 95% CI 0.14 to 0.16, P=0.00) did have an effect on ISR (see online supplementary table 1 in data supplement).

Figure 1

DSA imaging before (A) and after (B) stenting of vertebral artery ostium stenosis showed that the stenosis was alleviated with flow perfusion of the vertebral artery. R-SUB, right subclavian; RVA, right vertebral. Arrow in (A), stenosis site; arrow in (B), stent site.

Table 1

Summary of baseline clinical characteristics and stent information in the four patient groups

Findings of ISR

In our population, 90 patients were lost to follow-up after day 1. Among the remaining subjects, 330 patients had clinical and color Doppler follows-up with a median of 12 months (IQR 3, 12). ISR occurred in 120 patients (36.4%), a median of 6 months (IQR 3, 12) after stenting. Symptomatic restenosis appeared in 4 cases (3 with 70–99% and 1 with occlusion), and all presented with symptoms of dizziness as the main complaint. The remaining 116 ISR cases were asymptomatic.

Multiple regression analysis showed that patients implanted with a DES were less likely to have ISR compared with those who received a BMS (28.6% vs 44.4%, OR=0.38, 95% CI 0.19 to 0.75, P=0.01). Details of ISR patients by stent type and occlusion group are shown in the supplementary figure 2 in the data supplement.

The log rank (Mantel–Cox) test and disease free curves (figure 2) indicated that the C-OCC group had the worst ISR (P<0.05). Cox regression analysis (table 2) was performed to obtain HRs for each occlusion group compared with the PAT group. The results indicated that ICA-OCC was not a risk factor (HR=1.34, P=0.31) for ISR, but CVA-OCC (HR=1.66, P=0.02) and C-OCC (HR=3.34, P=0.001) were risk factors for ISR.

Figure 2

Kaplan–Meier curves. PAT, patent group; ICA-OCC, internal carotid artery occlusion group; CVA-OCC, contralateral VA occlusion group; C-OCC, combined occlusions group. ×Case censored.

Table 2

Stent type adjusted Cox regression analysis of in-stent restenosis risk of occlusion lesions

Doppler parameters

The results showed that post-stenting velocity was decreased compared with pre-stenting, and increased when ISR occurred (see online supplementary table 2 listed in the data supplement). An example of an ISR case is shown in the online supplementary figure 3 in the data supplement.

On day 1, patients in the CVA-OCC and C-OCC groups had higher velocities than the PAT group (P<0.05, see online supplementary table 3 listed in the data supplement), and C-OCC velocities were higher than in the CVA-OCC group (P=0.01). PSVos was not associated with ISR in the total population. However, logistic regression analysis showed that PSVos >140 cm/s was a risk factor (OR=2.81, 95% CI 1.06 to 7.43, P=0.04) for ISR in the CVA-OCC and C-OCC groups.

We observed good agreement between measurements taken by the same observer and by two independent observers for PSV values. The mean (±SD) difference was −0.52 (±8.47) for repeated measurements of PSV ratio taken by the same observer and 2.22 (±11.45) for those taken by two independent observers (see online supplementary figure 4 in the data supplement).


This study investigated 420 stented VAOS patients, which is, to the best of our knowledge, the largest group reported to date. Our results demonstrated three important findings. First, DES reduces ISR compared with BMS, confirming the results of previous analyses.12 16 Second, a VAOS stent associated with contralateral VA occlusion is a significant risk factor for ISR. Third, this is especially true if PSVos on day 1 post-stenting is >140 cm/s.

Vertebral stenting has become an alternative to medical therapy to treat VAOS.5 16 However, the best treatment for extracranial VAOS is still under investigation. Compared with the ICA, endovascular therapy in VA stenosis is less established, and medical management with antiplatelet and anticoagulant therapy is the usual therapeutic option.17 The American Heart Association stroke prevention guidelines recommend angioplasty limited to symptomatic extracranial VA stenoses18 19 which are refractory to medical therapy. Generally, patients with persistent symptoms despite maximal medical therapy with dual antiplatelet therapy, including aspirin and clopidogrel for at least 3 months, are considered as refractory to medical therapy in our center. A major critique of VAOS stenting is a relatively high incidence of restenosis (range 12–48%).5 7–9 The incidence of ISR in this study (36.4%) was consistent with these results.

PSVos was the main Doppler parameter evaluated in this study based on results of a previously published study, where ROC analysis showed that PSV was the optimal parameter in identifying ISR and in differentiating the severity of ISR. The sensitivity and specificity for 50–69% ISR were 100% and 97.4%, respectively, while for 70–99% ISR they were 92.9% and 94.6%.15

Although the factors that lead to ISR are incompletely established, endothelial hyperplasia is felt to be a crucial pathological factor.20 A DES may prevent the migration and proliferation of smooth muscle cells and reduce the incidence of ISR accordingly.16 21 A systematic meta-analysis for VAOS angioplasty with DES and BMS illustrated that DES had a significantly lower rate of ISR (8.2%, 14/170) compared with BMS (23.7%, 68/287), and other factors (such as length and diameter) had no significant impact on ISR.22 Our current study confirmed that there was a difference in ISR between DES (28.6%) and BMS (44.4%), and that utilization of DES was a significant protective factor for ISR (OR=0.38, 95% CI 0.21 to 0.82, P=0.01). Our results for ISR were higher than this meta-analysis as we set ISR at a more conservative level (50% vs 70%). Another meta-analysis showed that DES not only reduced the incidence of ISR but also alleviated recurrent symptoms on long term observation.12 However, we had too small numbers of symptomatic ISR patients (four cases) to analyze this. Therefore, the protective effect of DES on controlling recurrent symptoms is unclear in this population.

A previous study demonstrated that contralateral carotid occlusion is an independent predictor of ISR after carotid artery stenting.23 VAOS has concomitant stenotic or occlusive lesions in other vessels (4.8, 12.9%).24 The stented group in this study had a higher incidence of occlusion lesions (36.4%); we demonstrated that contralateral VA occlusion with or without ICA occlusion is a risk factor for ISR. In particular, our results showed that combined occlusion had a higher incidence compared with all other groups at any time during follow-up.

PSVos on day 1 post-stenting varied among groups. CVA-OCC exhibited a higher velocity compared with the PAT group. We believe this is due to the stented VA acting as a collateral to the contralateral VA territory. The C-OCC group had even higher velocity than the CVA-OCC group, perhaps representing collateralization to posterior and anterior circulations simultaneously for the patient who had an incomplete circle of Wills.25–27 The anterior communicating artery is the most collateral, more than the posterior communicating artery in patients with an ICA occlusion.26 The current study found no differences in PSVos between the PAT and ICA-OCC groups on day 1 post-stenting, which is consistent with this previous finding that VA was not the primary collateral to the isolated ICA occlusion. Moreover, previous studies demonstrated that high velocity could increase restenosis after renal and ICA stenting.23 28 This study supports this result, as Cox regression results showed that CVA-OCC was a risk factor for ISR, and C-OCC had an even higher HR value compared with CVA-OCC (3.34 vs 1.66).

It had been proven that PSV ≥120 cm/s measured with postprocedural carotid duplex ultrasound is an independent predictor of carotid stenting ISR.23 Although PSVos was not an independent risk factor for ISR in our studied population, when the population only included the CVA-OCC and C-OCC groups, logistic regression showed that PSVos ≥140 cm/s (OR=2.81, 95% CI 1.01 to 7.43, P=0. 04) was a significant risk factor for ISR. PSVos >140 cm/s indicates that the stented artery acts as a collateral as it is higher than an unstenosed native vessel (PSV <140 cm/s) but lower than a 50% or higher stenotic stent (>170 cm/s).

Although DES reduced ISR, the incidence was still relatively high in this study, and most of the ISR occurred within 1 year. Moreover, our results demonstrated that VAOS patients associated with contralateral vertebral occlusion and combined occlusions were at high risk for developing ISR in the future, especially patients with a higher velocity post-stenting.

Study limitation

The main limitation of this study was the use of a retrospective cohort, with a relatively high rate of subjects lost to follow-up (21.4%). All data were acquired from a single institution, and findings should be repeated in a multicenter setting. Second, ultrasound was used to diagnose ISR in this study and a small number of patients had this confirmed by angiography. However, based on a previous cohort of stented patients, the specificity and sensitivity of diagnostic ultrasound criteria were satisfactory compared with DSA.15 Third, although the study discussed the impact of compensation status in developing ISR, transcranial Doppler was not used in characterizing collaboration further. In light of this, transcranial Doppler should be considered for future trials investigating this hypothesis.


The results of this study confirmed that DES had less ISR than BMS. In addition, VAOS stenting associated with contralateral VA occlusion, and combined ICA and contralateral VA occlusions had a higher incidence of ISR. VAOS patients were especially associated with contralateral VA occlusion, demonstrating a higher in-stent velocity >140 cm/s post-stenting. Physicians and patients should be aware of these risk factors and carefully consider implanting stents under these circumstances. If these stents are used, different interval follow-up may be required.


The authors thank Dr Yue Zhao, Dr Duo Xu, Dr Ke Zhang, Dr Na Li, Dr Na Lei, and Dr Yumeng Luo for study management, and all the investigators and clinical staff for their outstanding contributions.


View Abstract


  • Contributors JL, YH, LN, and LJ: substantial contributions to the conception or design of the work. JL, YH, and LN: analysis and interpretation of the data. JL, LN, FF, JRE, and J-BL: drafting the work or revising it critically for important intellectual content. JL, FF, JRE, and ZL: data analysis and statistics. JL, RL, and XT: acquisition of the data. LJ and YH: agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. YH: final approval of the version to be published.

  • Funding The work was supported by the National Natural Science Foundation of China Grant, grant No 81070924, and by Beijing Municipal Administration of Hospitals’ Youth Programme, grant No QML20150803.

  • Competing interests None declared.

  • Patient consent Not required.

  • Ethics approval The research protocol was reviewed and approved by the ethics committee of Xuanwu Hospital, Capital Medical University.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

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.