Carotid artery stenosis accounts for 10 to 15% of ischemic stroke (1). Carotid endarterectomy (CEA) remains a reference treatment for carotid stenosis. Carotid artery stenting (CAS) became an important alternative to CEA (2). In CAS, the periprocedural risk and the in-stent restenosis (ISR) risk in the long term are major concerns (3, 4). Technical advances, including the development of embolic protection devices and stents, have helped to reduce the periprocedural risk. A meta-analysis of five randomized trials that exclusively used embolic protection devices showed that CAS and CEA were associated with a similar risk of a composite of periprocedural stroke, myocardial infarction, death, or non-periprocedural ipsilateral stroke (5). However, ISR after the CAS is still an issue. A recent Cochrane meta-analysis showed that CAS had a higher risk of moderate or higher restenosis (>50%) or occlusion than CEA and a similar risk of severe restenosis (>70%) with CEA (6). A systematic review showed that the >50% restenosis rate at 6 months was 3.9%, 12 months was 5.7% (7). The second analysis of the International Carotid Stenting Study (ICSS) showed that the cumulative 5-year risk of >50% restenosis was 40.7% and the cumulative 5-year risk of >70% restenosis was 10.6% (8). Therefore, the in-stent restenosis after CAS is an important problem to address. Some options, including percutaneous transluminal angiography (PTA) with balloon, repeat CAS (rCAS), CEA and artery bypass, have been proposed and compared in previous reviews (9–11). Recently, new techniques, such as drug-eluting balloons and stents, were proposed (12, 13). Here, we provided an updated systematic review of the current literature to evaluate the therapeutic options for ISR after CAS.

Herein, a systematic review of 35 selected studies, including 1,359 patients and 1,374 treatments, of ISR after CAS demonstrated that the most used treatment was rCAS, followed by PTA and CEA. Techniques including carotid artery bypass and EBR were applied in few cases. New PTA techniques, including CB-PTA and DEB-PTA, were widely employed. DUS was the most common method in detecting ISR, and was usually accompanied with CTA, MRA, or DSA. PTA was associated with the highest long-term overall stroke & TIA and recurrent restenosis rates. The rates of postoperative and long-term stroke & TIA and recurrent restenosis showed no significant difference between the rCAS and CEA groups.

DUS is a non-invasive technique, which was used in many randomized controlled trials (RCTs), including CREST and EVA-3s (49–51). Previous data indicated that a PSV of 300–350 cm/s may be a good predictor of ≥ 70% ISR (52). The present study confirmed the applicability of a PSV threshold of 300–330 cm/s. Of note, DUS is an indirect method to detect the stenosis and results are operator- and stent-dependent (53). Previous studies also showed that restenosis detecting by DUS were usually overestimated in the stented artery (54, 55). Therefore, direct methods of detecting lumen stenosis, including DSA, CTA, and MRA, are important in confirmation of the diagnosis, especially DSA usually regarded as the gold standard. In addition, new image techniques, such as dual-source CTA (can address the effects of beam hardening in regular CTA) (56) and 3D black-blood MRI (can assess artery plaques) (57, 58), have been gradually applied to evaluate the artery stenosis. The present study found that criteria for restenosis treatment varied. The criteria for asymptomatic patients were stricter than that for symptomatic patients. This is consistent with intervention decisions for carotid artery stenosis (2). Although stenosis remains a primary consideration, treatment decisions may benefit from consideration of other parameters. One recent study showed that for asymptomatic carotid stenosis, high-risk plaques were common and associated with higher risk of ipsilateral ischemic cerebrovascular event, suggesting the degree of stenosis alone may not fully capture the clinical situation (59). For ISR, evaluation of the plaque, the stent and the vessel are important. Some new techniques, such as the intravascular optical coherence tomography (OCT) (60), make it possible to access the plaque, stent and vessel in vivo (61). OCT had been used to assess the carotid plaque characteristics in carotid disease (62). A study of coronary disease showed that OCT could help to identify high risk patients and guide the treatment of ISR (63). OCT as also applied to evaluate tissue prolapse after CAS (64). Given this, the use of OCT to evaluate restenosis after CAS should be considered in future. It should also be noted that almost one out of ten studies in the review did not report diagnostic modality to verify severity of the stenosis. All these limitations, especially the drawbacks of DUS, should be kept in mind when one intends to translate the current knowledge into clinical practice.

PTA was the most common method in a previous review (10), but in the present study, rCAS was the most frequently used treatment. With respect to the postoperative risk, PTA had the lowest risk. It should be noted that some studies, especially the study involving PTA, did not report the postoperative outcomes in detail, which could confound the results of our analysis. In the review, compared to rCAS, CEA was associated with higher rates of death and other complications, including bleeding and cranial nerve injury. A previous meta-analysis comparing the rCAS and CEA for the ISR after CAS found no difference in the mortality rate, stroke-free rate, and even the rate of cranial nerve injury and haematoma (14). Given the discrepancies in these findings, future studies should compare the postoperative risks after rCAS and CEA. With respect to the long-term overall risk, PTA was associated with the highest rates of stroke & TIA and recurrent restenosis. Our review also confirmed the high restenosis rate associated with PTA (10, 15). The death rate was relatively high in all subgroups. This may be attributed to the long duration of the follow up which ranged from several months and many years. The death causes included trauma, cancer, and the procedure. However, many studies didn't provide detailed clinical information of the cause and time of death. Thus, we should be cautious when interpreting the results.

CEA remains the gold standard for the treatment of carotid artery stenosis. CEA is also suitable for ISR, especially for the patients difficult to conduct rCAS and PTA (36, 65). A recent study comparing CAS and CEA in De Novo carotid stenosis and postintervention restenosis suggested that for post-CAS restenosis, CAS and angioplasty were recommended for moderate stenosis and CEA for severe stenosis (66). In the present analysis, CEA was associated with a relatively high rate of postoperative death but similar long-term overall risks of stroke & TIA, death, and recurrent restenosis as rCAS. These findings suggest that CEA is an acceptable alternative to rCAS, especially in difficult situations. The proportion of symptomatic ISR after PTA was significantly lower than that after rCAS and CEA (PTA 32%, rCAS 65%, CEA 66%), and the contralateral occlusion rate associated with PTA was higher than those associated with rCAS and CEA (PTA 50%, rCAS 14%, CEA 11%) (Supplementary Table VIII). This may reflect the preferred treatment choice in the real world.

New PTA techniques are seeing application including DEB-PTA, CB-PTA, and paclitaxel-coated balloon PTA. All these techniques are aimed at decrease the relatively high recurrent restenosis in PTA. A meta-analysis of RCTs comparing plain balloon angioplasty, DEB-PTA, and DES for the treatment of ISR in coronary intervention showed that DEB-PTA and DES were associated with a significantly lower restenosis risk than plain balloon angioplasty (67). In contrary to good effect of DEB-PTA in coronary artery disease, however, the present study demonstrated that the Re-ISR rates of these techniques were also high. As shown in Supplementary Table V (online only), the stent types employed in the studies included the Wallstent, Acculink, and Palmaz stents, and the drug-eluting stents were very few. These new treatment techniques should be further evaluated in the future.

There are some important limitations of the present study. First, some studies included in the review did not report basic information on patient symptoms or gender, indications, postoperative complications, or long-term outcomes in detail, which may cause publication bias. We also separately analyzed studies with intact basic information and found that the tendency of the comparisons between the outcomes of those interventions is similar to our original analysis but without statistical significance (Supplementary Table VI and VII). Second, there was heterogeneity in the imaging standards of ISR, the follow-up time, and clinical outcome identification among these studies. The sample size varied from 3 to 645 and the primary results were inevitably influenced by the studies with large sample size. Generally, pronounced heterogeneity within and between studies, a meta-analysis could not be conducted. Third, no study in the review was a RCT. Because of the heterogeneity and the data quality of these studies, we must be cautious when interpreting the results regarding the statistical evaluation of the postoperative and long-term outcome. We therefore cannot conclude which treatment for ISR is optimal based on the current evidence, and randomized controlled studies are needed to further assess these treatments.

The indication for ISR depended on the degree of stenosis but the criteria for symptomatic and asymptomatic patients were different. DUS was the most common method used for detecting ISR and was often accompanied with CTA, MRA, or DSA. A PSV threshold of 300–330 cm/s correlated with high-grade ISR. Historically, rCAS is the most common treatment for ISR with low postoperative risk and low long-term risk. CEA is an important alternative for rCAS. PTA is less recommended due to its high long-term risks of stroke & TIA and recurrent restenosis. Based on the present analysis, it was not possible to ascertain which treatment for ISR is optimal and large, multi-center, blinded analysis may help stratify the optimum therapeutic options.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

HH put forward the idea of this article, worked with LW to complete literature retrieval and screening, data extraction, and manuscript writing. HH and LW contributed equally to this work. XL participated in the guidance of experimental design, etc. YG, YZ, and JZ participated in the verification of the data included in these literatures. ZY participated in the language modification work of the article.

Key Research and Development Program of Hubei Province (2020BCA070 to XL), the Application Foundation Frontier Special Project of Wuhan Science and Technology Bureau (2020020601012226 to XL), the Flagship Program of Tongji Hospital (2019CR106 to XL), and the second batch of clinician research projects of HUST to XL.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.