Introduction

Carotid artery stenting (CAS) has gradually become an alternative to carotid endarterectomy (CEA) for the treatment of carotid artery stenosis [1]. With the development of interventional devices and advances in minimally invasive surgery [2], patients with carotid artery stenosis are increasingly willing to undergo CAS. Diffusion-weighted imaging (DWI) may represent a surrogate tool for optimizing diagnostic and therapeutic vascular procedures [3]. Indeed, DWI could be performed to easily detect post-CAS cerebral emboli, which are asymptomatic and termed silent new ischemic cerebral lesions (sNICL) [4].

Embolization from plaque fragment mobilization is a common complication of endovascular procedures [5], and recent reports indicate sNICL rates ranging from 18 to 57% [6,7,8,9,10]. Even without a corresponding focal deficit, sNICL might lead to clinical consequences in the long term, including cognitive decline and dementia [11]. In addition, NICL could lead to an elevated risk of future stroke events [12].

Hence, evaluating risk factors associated with post-CAS sNICL has recently attracted increasing attention, but findings by various reports were limited and inconsistent [6, 7, 13]. Meanwhile, identifying potential risk factors for new cerebral lesions is essential in clinical practice. Therefore, the present work aimed to determine risk factors for sNICL detected by DWI following CAS. The associations of clinical and morphological variables with the occurrence of sNICL detected by DWI after CAS were assessed.

Materials and methods

Study design and patients

The present study retrospectively enrolled patients eligible for Revascularization of Extracranial Carotid Artery Stenosis (RECAS) study, which was a multicenter, prospective cohort study involving both CAS and CEA in Chinese patients with extracranial carotid artery stenosis. The RECAS study enrolled patients from 36 centers in China between December 2013 and February 2016. For the RECAS study, inclusion criteria were (1) age ≥ 18 years old and (2) symptomatic internal carotid artery (ICA) stenosis ≥ 50% or asymptomatic ICA stenosis ≥ 70% by computed tomography angiography (CTA) or digital subtraction angiography (DSA). Exclusion criteria were (1) stenosis combined with intracranial arteriovenous malformation or aneurysm; (2) severe stenosis or occlusion of the ipsilateral intracranial artery; (3) unstable angina, myocardial infarction (MI), or congestive heart failure within the previous 6 months; (4) uncontrolled diabetes mellitus (DM) defined as glucose > 300 mg/dL; (5) uncorrectable coagulation abnormalities; (6) pregnancy or being in the perinatal period for women; (7) severe concomitant disease with poor prognosis (life expectancy < 2 years); or (8) intolerance or allergies to any of the study medications, including aspirin and clopidogrel. Data of patients in our Hospital were extracted to perform the present single-center retrospective analysis. Patients without post-procedural symptomatic stroke were included, and those without pre-procedural or post-procedural DWI for any reason were excluded.

Carotid stenosis measurement was performed according to the North American Symptomatic Carotid Endarterectomy Trial (NASCET) [14], stenosis of 70–99% was defined as severe, and that of 50–70% as moderate. Patients with transient ischemic attack, retinal ischemic event, or ischemic stroke resulting from the narrowed carotid artery within the previous 6 months were considered to be recently symptomatic [15]. The degree of stenosis was morphologically evaluated by CTA of the supra-aortic trunks, agreeing with Doppler ultrasound examination results.

The RECAS study, registered on clinicaltrials.gov (NCT01994187), was approved by the ethics committees of all participating centers and performed according to the Declaration of Helsinki.

Preoperative data

The clinical variables evaluated included demographic data, stroke history, and cardiovascular risk factors such as atrial fibrillation (AF), hypertension, and DM. Stenosis characteristics, arch type, target carotid artery, and stenting procedure were also analyzed as potential risk factors for sNICL.

A neurological examination was performed the day before CAS. Neurological status recorded over the last 6 months before stenting was analyzed. The degree of artery stenosis and plaque characteristics in the ipsilateral and contralateral carotids, as well as vertebral arteries, were evaluated by preoperative ultrasound and/or CTA. Both common carotid arteries (CCA) and the carotid bifurcation were also assessed. All images were evaluated by qualified technicians blinded to clinical data. Plaques were classified as calcified, ulcerated, or other. Target carotid lesions were evaluated by diagnostic DSA during the stenting procedure. All angiograms were evaluated by qualified interventional technicians and neuroradiologists blinded to clinical data.

Stenting procedure

Patients received aspirin (100 mg/day) and clopidogrel (75 mg/day) for at least 3 days preoperatively. Heparin (1 mg/kg) was administered at the beginning of the intervention to maintain an activated clotting time at approximately twice the base value. After the procedure, the patients had both aspirin (100 mg/day) and clopidogrel (75 mg/day) as daily maintenance medications for 3 months. Those who could not tolerate aspirin or clopidogrel for any reason received cilostazol 0.1 g BID instead. Femoral access with an 8-French introducer and catheter was achieved under local anesthesia to introduce a guidewire up to the origin of the external carotid artery (ECA) in all patients. Carotid and cerebral angiography was subsequently performed to confirm the level and the degree of arterial stenosis. All stenting procedures were performed by surgeons highly experienced in angiographic procedures. Embolic protection devices (EPDs) were used in all patients. Pre-dilatation, if necessary, was performed after EPD placement. Above work was followed by stent deployment. Stent size was based on the estimated diameters of the CCA and ICA. In case of insufficient stent expansion, post-dilatation was performed. Atropine sulfate (0.5 mg) was administered intravenously during angioplasty balloon insufflation to prevent carotid sinus stimulation and bradycardia. The choice of catheters and guidewires was at the discretion of the surgeon. The stenting protocol in this study was similar to that reported previously [16].

DWI

The first magnetic resonance imaging (MRI) was performed before stenting to assess the pre-procedure condition and establish baseline information. The second MRI was performed within 3 days after stenting to assess post-procedural new ischemic cerebral lesions and at any time thereafter in case of neurological deterioration. The protocol included isotropic DWI sequences. MRI was performed on a 3.0T system (MAGNETOM Verio, Siemens Healthcare, Erlangen, Germany). The following brain MR sequences were applied, assessing apparent diffusion coefficients: axial plane T1-weighted spin-echo (TR = 135, TE = 2.55, slice thickness = 5 mm, and interslice gap = 1.2 mm); FLAIR (TR = 8800, TE = 84, TI = 8800, slice thickness = 5.0 mm, and interslice gap = 1.2 mm); and DWI (TR = 5500, TE = 90, b value = 1000, slice thickness = 5 mm, and interslice gap = 1.2 mm). The occurrence of new high-signal DWI lesions within any brain territory after stenting with corresponding low apparent diffusion coefficient indicated the existence of a new ischemic cerebral lesion [17]. MR images were evaluated by experienced neuroradiologists blinded to clinical data, angiograms, and ultrasound results.

Statistical analysis

Continuous data were given as mean ± standard deviation (SD). Categorical variables were presented as numbers and percentages. Differences between patients with and those without DWI lesions were analyzed by the chi-square test. Proportions were tested by the chi-square test or Fisher’s exact test. To determine the univariate associations of lesion occurrence with potential risk factors, odds ratios (ORs) and 95% confidence intervals (CIs) were calculated. Multiple binary logistic regression using the backward and forward variable selection procedures was employed to determine risk factors for sNICL. Two-tailed P < 0.05 was considered statistically significant. Data were analyzed with SPSS 23.0 (IBM, NY, USA).

Results

Patient baseline characteristics

A total of 894 consecutive patients eligible for the RECAS study were initially enrolled. Of those, 19 patients with post-procedural complications, and 181 patients lacked DWI images were excluded. A total of 694 asymptomatic patients with complete imaging data were finally included in this study (Fig. 1). The patients were 66.03 ± 8.03 years old, and 83.3% of them were men. Symptomatic lesions before CAS accounted for 45.5%. Detailed patient characteristics are shown in Table 1.

Fig. 1
figure 1

Flowchart of the study population

Full size image
Table 1 Patient demographics and stenosis characteristics
Full size table

Postoperative analysis

Stent placement and balloon angioplasty were successfully performed in all cases. Fifty-seven patients (8.2%) had unfavorable arch anatomy, which was defined as a type III or bovine arch. Meanwhile, 352 patients (50.7%) had right CAS, and the remaining 342 cases (49.3%) showed left CAS. A total of 414 patients (59.7%) were treated with open-cell stents. In this study, 524 (75.5%) and 203 (29.3%) patients had pre-dilatation and post-dilatation, respectively, with 79 of them having both pre-dilatation and post-dilatation.

We performed stenting without dilatation in 46 patients (6.6%); the stent delivery device had no difficulty in passing through the stenotic lesion, so there was no pre-dilatation; the stent was implanted satisfactorily, and the residual stenosis was < 30%, and post-dilatation was deemed unnecessary. At the end of the procedure, all patients had successful stenting, with ≤ 30% residual carotid stenosis on angiography (NASCET criteria). Postoperative DWI detected sNICL in 356 patients without persistent clinical deterioration, including 95 (26.7%) mono-infarct and 261 (73.3%) multiple infarct cases. Among these patients, 150 (42.1%) had lesions ipsilateral to the treated artery, 56 (15.8%) showed contralateral lesions, and 150 (42.1%) had ipsilateral and contralateral lesions. DWI lesions were mainly located in anterior vascular territories (54.8%). New infarcts were found in both anterior and posterior vascular territories in 122 patients (34.3%). A total of 19.7% of the patients had some scattered infarcts involving multiple sites, and 16 patients (4.5%) showed infarcts distributed mainly in the watershed districts (Table 2).

Table 2 Surgical data and distribution of the new lesions
Full size table

Univariate analysis

Univariate analysis identified significant factors associated with sNICL occurrence, including DM (P = 0.008), ipsilateral calcified plaques (P = 0.036), ipsilateral ulcerated plaques (P = 0.026), pre-dilatation (P = 0.003), and open-cell stent use (P < 0.001) (Table 3).

Table 3 Associations of demographic parameters with the occurrence of new ischemic brain lesions, assessed by univariate regression analysis
Full size table

Multivariate analysis

Of the parameters significantly associated with sNICL occurrence in univariate analysis, multivariate analysis revealed four risk factors for sNICL, including DM (OR = 1.579, 95%CI 1.142–2.184; P = 0.006), ipsilateral calcified plaques (OR = 1.726, 95%CI 1.073–2.775; P = 0.024), ipsilateral ulcerated plaques (OR = 1.912, 95%CI 1.102–3.316; P = 0.021), and open-cell stent use (OR = 3.316, 95%CI 2.321–4.738; P < 0.001). In multivariate analysis, pre-dilatation was not a significant risk factor for sNICL (Table 4).

Table 4 Risk factors for postoperative new ischemic brain lesions assessed by multivariate analysis
Full size table

Discussion

In this study, DM, ipsilateral calcified plaques, ipsilateral ulcerated plaques, and open-cell stent use were found to be independent risk factors for sNICL after CAS.

A total of 356 patients (51.3%) had postoperative new DWI-positive cerebral ischemic lesions. This was in line with previous reports revealing rates of ischemic lesions post-CAS of 18–57% [6,7,8,9,10]. In this study, sNICL occurrence had no associations with pre-procedural parameters, including symptomatic and asymptomatic stenoses, and the location of new lesions. This contradicted previous findings that symptomatic patients harbor a higher risk of post-procedural ipsilateral ischemic events [18], but corroborated Cosottini et al. who reported that symptomatic cases show no association with the incidence of silent ischemic lesions detectable by DWI [4].

As shown above, four risk factors for sNICL occurrence after CAS were identified, including DM, ipsilateral calcified plaques, ulcerated plaques, and open-cell stent use. In previous trials, sex, stenosis severity, side of intervention, hypertension, DM, dyslipidemia, and smoking were not associated with periprocedural ischemic events in CAS [15]. We found a relationship between DM and new cerebral ischemic lesions after stenting. Interestingly, Henry et al. demonstrated that DM is associated with a specific pattern of compensatory remodeling based on ultrasound examination [19]. Thus, we hypothesized that differences in plaque consistency and surface appearance could result in variations of plaque vulnerability during invasive therapy, with DM-associated plaques being particularly prone to rupture and dispersal during treatment. However, further studies are required for confirmation.

Although unfavorable arch anatomy increases the difficulties of DSA and CAS, no relationship between arch anatomy and sNICL after CAS was found. In agreement, a previous report showed that neither arch type nor bovine arch is an independent risk factor for ischemic lesions associated with CAS [20]. Carotid plaque surface irregularity reflects plaque instability and causes solid cerebral embolism during CAS in symptomatic patients. Our results corroborated the aforementioned findings, demonstrating that ulcerated plaques detected on pre-procedural DSA increase the risk of DWI lesions after stenting [21]. However, other authors hold a different view [22]. Such discrepancy could be explained by the fact that plaque characteristics evaluated by different imaging methods may vary, which results in errors. Moreover, Ichinose et al. reported that proximal calcification is an important predictor of new DWI lesions after CAS [23], which is consistent with our above results.

One of the main findings of this study was that, besides DM and plaque characteristics, open-cell stent use independently predicted post-procedural ischemic lesions after CAS. In a previous study, patients treated with open-cell stents had a significantly higher stroke and death rates after CAS compared with those treated with closed cell stents [24], indicating that the open-cell stent option has an intrinsically greater potential to cause cerebral embolism of plaque debris. Recently, a meta-analysis based on 46,728 procedures by de Vries et al. [25] demonstrated that open-cell stenting resulted in a significantly higher number of subclinical post-procedural new ischemic lesions detected on DWI compared with closed cell stenting. These findings were supported by the present study.

Khan et al. [26] suggested that the use of EPD may not be meaningful. At the beginning, 4 patients accepted CAS without successful EPD use for the badly tortuous carotid artery. However, these 4 patients were excluded out of the study for complications. Hence, we cannot evaluate our data in the same way as Khan did.

The rate of pre-dilatation was relatively high in our center. Pre-dilatation was usually performed after EPD placement, and individualized dilatation planning was performed based on information such as calcified plaque detected by ultrasound before surgery, residual stenosis, and so on.

This study had several limitations. The RECAS study was not originally designed to answer the question of the present post hoc study, and the data are limited because of the retrospective nature of the study. During its design, no consensus was reached on the platelet inhibition test or platelet function evaluation for guiding the medication plan in stenting patients. It is necessary to include this in follow-up studies. In addition, we did not evaluate the circle of Willis although several studies have demonstrated that the contralateral carotid artery becomes the primary collateral vessel supplying blood to it in patients with severe carotid stenosis, and vertebral or contralateral carotid severe stenosis or occlusion is positively associated with infarct volume [27, 28]. Some scholars proposed that unstable carotid plaques confer a high risk of embolism [29]. A meta-analysis demonstrated that patients with intraplaque hemorrhage on pre-procedural MRI have elevated rates of silent ischemia [30], while Chung et al. considered it the latter parameter is not a significant risk factor for cerebral embolism during CAS [18]. However, this study did not assess plaque details by high-resolution MRI or pathological examination. Therefore, it is plausible that not all unstable plaques were identified, especially in the case of intraplaque hemorrhage. ICA tortuosity, which may increase the risk of cerebral ischemia during CAS, should also be taken into account in future studies [31]. Serum biochemical or physiological indices, such as TNF-α levels, intra-arterial oxidative stress, and white matter damage, may be associated with new cerebral ischemic lesions [6, 7]. The plaque properties referred to in this article refer to the plaque properties of the stenotic lesions during surgery. Due to the limitations of the design of RECAS study, the present retrospective study could not complete the statistical evaluation of plaque morphology and stent type within the entire group of sNICL or within the group of sNICL located ipsilateral territory.

In conclusion, the rate of postoperative sNICL after CAS was 51.3% in this study. We also found that DM patients with calcified or ulcerated plaques who undergo CAS with open-cell stents are likely at a high risk of post-procedural ischemic events. This is a rare study involving the risk factors of sNICL after CAS in Chinese individuals, which may provide insights into patient selection for carotid stenosis. Nevertheless, large-scale prospective randomized controlled trials are needed to confirm these findings, which may be used to identify high-risk patients for CAS to prevent postoperative ischemic lesions.