Article Text

Download PDFPDF

Original research
Risk factors for hyperperfusion-induced intracranial hemorrhage after carotid artery stenting in patients with symptomatic severe carotid stenosis evaluation
  1. Lei Zhang,
  2. Dongwei Dai,
  3. Zifu Li,
  4. Guoli Duan,
  5. Yong-wei Zhang,
  6. Pengfei Yang,
  7. Qinghai Huang,
  8. Yi Xu,
  9. Bo Hong,
  10. Jianmin Liu
  1. Department of Neurosurgery, Changhai Hospital, Second Military Medical University, Shanghai, China
  1. Correspondence to Professor Bo Hong and Dr Jianmin Liu, Department of Neurosurgery, Changhai Hospital, Shanghai 200433, China; hongbosmmu{at}, liu118{at}


Background Analyzing risk factors for hyperperfusion-induced intracranial hemorrhage (HICH) after carotid artery stenting (CAS) in patients with symptomatic severe carotid stenosis.

Methods This study retrospectively analyzed clinical data of 210 patients, who had symptomatic severe carotid stenosis (70–99%) and received CAS treatment between June 2009 and June 2015, and evaluated the relationship of HICH with patients’ clinical baseline data, imaging features, and treatment strategies.

Results Seven patients (3.3%) developed HICH after CAS. The incidence of HICH among patients with near total occlusion was significantly higher than among those without (10.1% vs 0%, P<0.001). Out of the seven, five had no development of either anterior or posterior circulations, and two had no development of anterior circulation and poor development of posterior circulation. Results showed that patients with poor compensation of Willis’ Circle were more likely to develop HICH compared with other patients (P<0.001). All patients received preoperative CT perfusion. TTP index was defined as the TTP ratio between the affected and contralateral side. The results showed that the TTP index was significantly different between the HICH group and non-HICH group (1.15±0.10 vs 1.30±0.15, P<0.001). An analysis of the ROC curve indicated that patients with TTP index >1.22 were more likely to develop HICH compared with other patients (sensitivity 100%, specificity 75.9%).

Conclusions Patients with severe unilateral carotid stenosis, the presence of near total occlusion, poor compensation of Willis’ Circle, and preoperative TTP index>1.22, have a higher risk of developing HICH after CAS.

  • atherosclerosis
  • stenosis
  • stent
  • hemorrhage
View Full Text

Statistics from

Hyperperfusion-induced intracranial hemorrhage (HICH) after carotid endarterectomy (CEA) or carotid artery stenting (CAS) can cause an extremely high rate of mortality and morbidity. This phenomenon was first reported by Wylie and Adams1 in 1964. The symptoms of hyperperfusion syndrome (HPS) mainly include ipsilateral headache, epilepsy, and intracerebral hemorrhage. Currently, it is reported that the incidence of post-CEA HICH and post-CAS HICH was 0.37% (0.2%–1%) and 0.74% (0.36%–4.5%) respectively.2 The pathogenesis of HICH is believed to be associated with the damage of autonomic regulation of cerebral vessels after long-term exposure to ischemic hypoperfusion.3 Such hemorrhage often has insidious onset, although with low incidence, the consequence can be detrimental, which makes it an urgent problem needing to be solved in clinical practice. Previous studies mainly focused on analyzing the risk factors for HPS, yet few have reported HICH directly. Certain questions with respect to whether HPS and HICH share the same risk factors, whether HICH risk factors are specific, or how the development of HICH can be avoided, are creating confusions in our clinicians. This study retrospectively analyzed the clinical data of patients who had symptomatic severe carotid stenosis and received CAS treatment. The study also discussed the risk factors of post-CAS HICH, thus helping to guide interventional treatment and avoid operational risks.



The retrospective study collected a total of 518 patients who were diagnosed with symptomatic severe carotid stenosis by digital subtraction angiography (DSA)and received CAS treatment in the Department of Neurosurgery, Changhai Hospital Affiliated to the Second Military Medical University, between June 2009 and June 2015. Before surgery, patients received cerebral CT perfusion, and we found that all the HICH patients had hypoperfusion in the affected side. Consequently, 210 out of 518 patients were included in our study. Hypoperfusion was defined as the TTP of the affected side was delayed compared to the contralateral side.4

Inclusion and exclusion criteria

Inclusion criteria: patient had had a transient ischemic attack (TIA) or stroke related to atherosclerotic carotid stenosis in the past 180 days; DSA examination suggested severe internal carotid stenosis (70%–99%) according to the North American Symptomatic Carotid Endarterectomy Trial (NASCET)5; patient had one or more risk factors for atherosclerosis (including hypertension, diabetes, coronary heart disease, hyperlipidemia, smoking history, and atherosclerosis in other parts); unilateral internal carotid artery stenosis; and preoperational cerebral CT perfusion suggested hypoperfusion.

Exclusion criteria: preoperational cerebral MRI/CT indicated hemorrhagic transformation in cerebral infarction of the operational side; patient had combined intracranial tumors, arteriovenous malformations, or aneurysms; patient had combined severe cardiovascular diseases, blood diseases, or liver/kidney diseases; patient had contraindications or allergies to treatment of anti-platelet drugs; patient had tandem lesions distal to carotid artery stenosis, or had severe stenosis in the contralateral internal carotid artery; and patient received staged treatment for severe stenosis.

Operation methods

The treatment was performed under local anesthesia. An 8F arterial sheath was punctured into the femoral artery using the Seldinger technique, and systemic heparinization (initial dose of intravenous injection (mg)=2/3 of kilogram bodyweight) was given to lengthen the time of activated clotting to 2 to 2.5 folds of the normal. During the operation, blood pressure (BP) was controlled below 140/90 mmHg. Cerebral angiography was performed to evaluate the location, severity, and length of stenosis, and also to examine the compensation of collateral circulation. When the operation angle was determined, an 8F guiding catheter was placed into the common carotid artery by the direction of the path graph, and a distal protection device (AngioGuard, Cordis; Spider; EV3; Emboshield, Abbott – all from the US) was carefully passed through the stenosis and released in the normal vessel distal to the internal carotid artery stenosis. Then, the balloon was delivered along the umbrella wire to the position of stenosis for pre-dilation (the diameter of balloon was determined according to the diameter of normal carotid artery distal to the stenosis), and if there was any consciousness disorder caused by the temporary blocking of artery blood flow in patients, blood flow was restored as soon as possible. Along the guiding micro-wire, the carotid stent (Acculink, Abbott; Xact, Abbott; Precise, Cordis; Protégé, EV3; Wallstent, Bonston Science, all from the US) was delivered to the stenosis (the length of stent should fully cover the plaque, or exceed the two ends of stenosis by 1 cm: the diameter of stent should exceed the diameter of normal proximal carotid artery by 1–2 mm). Afterwards, the stent was released, the delivery system was withdrawn, and angiography was reviewed to examine the position of the stent. The dilation of stenosis was also examined, thereby evaluating whether post-dilation was necessary. At last, the distal protection device was taken out, and craniocerebral angiography was performed again, to evaluate the condition of residual stenosis after surgery.

Perioperative medication and treatment

Before operation, patients were given aspirin (100 mg/d)+clopidogrel (75 mg/d)+atorvastatin/rosuvastatin (20 mg/d) for at least 3 days. After surgery, patients were asked to take aspirin (100 mg/d) for life and clopidogrel for 3 months. Risk factors for atherosclerosis were controlled according to related guidelines.

At 1 day post-operation, all patients were sent to neurosurgery ICU for strict monitoring of their vital signs and neurological functions. Blood pressure was strictly controlled for 2– to 3 days after surgery: BP of all patients was controlled below 140/90 mmHg, and BP of patients with ≥90% stenosis was controlled below 120/80 mmHg.6

Imaging analysis

Cerebral CT perfusion and imaging analysis were performed using 64-sliced spiral CT (Siemens Sensation Cardiac64) and a power injector (Ulrich Missouri). Axial scan was first performed taking the canthomeatal line as the scanning baseline. Then, 40 mL of iopromide (300 mgI/mL) was injected via the antecubital vein, and 20 mL of isotonic saline was power-injected at a flow rate of 5/ml. After 4 s of delay, dynamic perfusion scanning was performed at the level parallel to the supraorbital line, with the basal ganglia area as its center, and covering the upper and lower layers of Willis’ Circle. The scanning time was 40 s, the layer thickness was 7.2 mm, the covering area was 7.2 mm, and the number of slices was four. Raw images were processed by the Siemens Perfusion Analysis Package (Siemens). The superior sagittal sinus was selected as the outflow vein, and in combination with the time-density curve of the contrast medium passing through the cerebral tissue, the ROI time-concentration decay curve of cerebral blood flow could therefore be generated. Four parameters were taken for quantitative and qualitative analysis: CBF, CBV, TTP, and MTT.

Evaluation of Willis’ Circle

In reference to the study of Hoksbergen et al,7 vessels were considered to be present if the contrast agent was developed, and the ends of the artery could be seen. On the other hand, vessels without contrast agent development were considered as absent. If the diameter of vessel was <0.5 mm (or <1/2 of the diameter of the contralateral vessel), the vessel was considered to be underdeveloped. In this study, well-developed bilateral A1 and anterior communicating arteries were defined as well developed anterior circulation; underdevelopment of either bilateral A1 or anterior communicating arteries was defined as underdeveloped anterior circulation; and the absence of either bilateral A1 or anterior communicating arteries was defined as absent anterior circulation. The development of posterior circulation was defined likewise. Absent and underdeveloped Willis’ Circle were defined as poor compensation, and well-developed Willis’ Circle was defined as good compensation.

Statistical analysis

In this study, HICH was determined by a surgeon. Two or more experts, after they had reviewed the preoperative CT/MRI images and intraoperative video records, have excluded operation-related hemorrhage or hemorrhagic transformation after infarction. Data analysis was performed using the SPSS16.0 software. The measurement data were first tested for normality. Data with normal distribution were presented as mean ±SD, and data with skewed distribution were presented as median and range (M (P25, P75)). Measurement data were compared by the t-test, and counting data were compared by the Pearson chi square test, continuity-corrected chi square test, or Fisher’s exact test. P<0.05 was defined as statistical significance.


Incidence of HICH

Out of the 210 included patients, 182 were male and 28 were female. The age of the patients ranged from 36 to 82 years, and the average age was 63±9 years. Before surgery, 77 patients (36.7%) showed TIA and 133 (63.3%) had cerebral infarction. The median time interval between the last ischemic event and stenting was 22 (15–31) d.

All 210 patients received stenting treatment, and the success rate was 100%. A total of seven patients (3.3%), including six male and one female, with an average age of 64.0±6.7 years, developed HICH after surgery. Five patients had HICH on the first day post-procedure, and two patients had HICH on the second day post-surgery. Three patients were discharged from hospital after treatment, but they had varying degrees of neurological deficit. The other four patients passed away despite tefforts to rescue. The comparison of baseline data of HICH and non-HICH patients is shown in table 1

Table 1

Baseline conditon

Results of imaging examination

All patients received CT perfusion before operation. The ratio of TTP, MTT, CBV, and CBF between the affected and contralateral side was taken for statistical analysis, and the results showed that the TTP index was significantly different between the HICH and non-HICH groups (table 2). ROC curve analysis of the TTP index showed that the area under the curve was 0.89 (95% CI (0.80 to 0.97), P<0.05), the critical value was 1.22, the sensitivity was 100%, and the specificity was 75.9% (figure 1). This indicates that patients with TTP index >1.22 were more likely to develop HICH compared with other patients.

Figure 1

ROC curve analysis of the TTP index shows that the area under the curve is 0.89 (95% CI (0.80 to 0.97), P<0.05), the critical value is 1.22, the sensitivity is 100%, and the specificity is 75.9%.

Table 2

Pre-procedure CT perfusion

All patients received preoperative and postoperative angiography evaluation. The average degree of preoperative stenosis was (86.2±7.1) %, indicating that all patients had severe stenosis (≥70%), and after surgery, the average degree of stenosis decreased to (24.5±10.4) %. Based on the NASCET definition of a near total occlusion,8 patients with any two of the following would be determined as having near total occlusion: delayed filling; intracranial collaterals; ipsilateral distal ICA less than the contralateral distal ICA; and ipsilateral distal ICA equal to or less than the ipsilateral external carotid artery (ECA) (the  internal carotid artery (ICA) normally is substantially larger than the ECA). Patients were grouped based on the recognition of near total occlusion, and 69 patients developed near total occlusion. All of the seven HICH patients belonged to the near total occlusion group, and the inter-group difference was statistically significant (P<0.001).

Evaluation of Willis’ Circle

Patients were grouped based on angiographic results of Willis’ circle. Table 3 shows the relationship between the compensation of Willis’ Circle and the incidence of HICH. As indicated, patients with poor compensation of Willis’ Circle were more likely to suffer HICH after surgery, and the inter-group difference was statistically significant (out of the seven patients having HICH, five had absent anterior and posterior circulation, and two had absent anterior circulation and underdeveloped posterior circulation).

Table 3

Relationship between HICH and development of Willis’ Circle


Despite its low incidence, post-CAS HICH could cause an extremely high rate of mortality and morbidity, and this makes it a problem attracting the most clinical concern these days. In our study, the incidence of HICH was 3.3%, which was slightly higher than in previous studies, and we attribute this to the fact that our study only included patients with severe stenosis. The study has reported that severe stenosis is a risk factor for the development of HPS.9

In our study, all seven patients developed HICH within 3 days after surgery, and this was consistent with the onset time of post-CAS intracranial hemorrhage reported by previous studies. Ogasawara et al 10 suggested that the onset of post-CAS HICH was earlier than that of post-CEA HICH (1.7 d vs. 10.7 d) Some scholars attribute the different onset and incidence of HICH to different anti-platelet regimens after CAS and CEA. However, no study has yet confirmed this speculation.

Since post-CAS HICH are characterized with early onset and high mortality and disability, more attention should be paid to prevent or predict the development of HICH. Previous studies have reported many risk factors for HPS, including female gender, long-term hypertension, diabetes, age >75 years, recent history of stroke, severe stenosis, vascular malformation, bilateral carotid artery stenosis, and poor control of postoperative BP.2 9–13 However, due to the low incidence of HICH, only a few studies have reported failure of postoperative BP control and shortened time of postoperative cerebral perfusion by more than 2.7 s, as the risk factors for HICH development.10 14 This study aimed to evaluate whether HICH would be affected by compensation of Willis’ Circle, or by preoperative cerebral CT perfusion. Since bilateral carotid stenosis might cause changes in the hemodynamics and compensation of Willis’ Circle, further affecting intracranial blood perfusion, this study only included patients with unilateral carotid stenosis. Before operation, all patients received CT perfusion, and the TTP index of patients with HICH was significantly higher than that of those without HICH (P<0.05). Compared to CBV and CBF, TTP is the most sensitive factor reflecting the change of hemodynamics.15 Though CBV and CBF were insignificantly different between the two groups, an increasing trend could be seen in the HICH group, and this might indicate a loss of vascular autoregulation and the dilation of minor arteries. Therefore, even with restored perfusion pressure after CAS, the arterioles would still fail to contract properly, and this might increase the risk of HPS and HICH development.

In our study, all seven HICH patients had near total occlusion, and the results also showed that the incidence of HICH is significantly higher among patients with near total occlusion than those without. Near total occlusion is a special type of carotid artery stenosis, which often reflects severe damage of vascular autoregulation and change in intracranial hemodynamics. According to the study of Oka et al, for patients with near total occlusion, only 20% of preoperative acetazolamide load test showed a cerebrovascular reactivity >10%.16 Long-term hypoperfusion would cause decreased intracranial vasoconstriction, resulting in the loss of compensation. Based on this, the insertion of carotid stent may help to recover the diameter of proximal carotid vessels and increase both blood flow and blood velocity. However, for the capillary bed under long-term hypoperfusion, a drastic increase in perfusion pressure may damage the blood-brain barrier, and further cause angiogenic cerebral edema as well as intracranial hemorrhage.

Our study also found that patients with poor compensation of Willis’ Circle were more likely to develop HICH after surgery. Liang et al17 simulated the change of intracranial hemodynamics after CAS by building up a computational cardiovascular model and designing different development types of Willis’ Circle. The researchers found that if effective communication between bilateral carotid arteries could not be formed via Willis’ Circle, the incidence of postoperative HPS would be increased significantly. Though HICH was not evaluated, the study did, to some extent, confirm the significance of Willis’ Circle to the incidence of HICH after CAS. Previous studies have described, in detail, the influence of anterior and posterior communicating arteries on collateral circulation.18 In this study, the incidence of HICH was significantly higher among patients with poor compensation of both anterior and posterior communicating arteries, and the reason might be that Willis’ Circle failed to form an effective shunt when ipsilateral blood flow had been increased after CAS. It also should be noted that all the HICH patients had near total occlusion with poor compensation of both anterior and posterior circulation, On the other hand, there was no HICH in the patients with only poor compensation of both anterior and posterior circulation and no near total occlusion. So, patients with near total occlusion and poor compensation of both anterior and posterior circulation had higher incidence rate of HICH than others (P<0.05).

Until now, few studies have reported how HICH can be prevented. Ogasawara et al19 suggested that the postoperative use of oxygen-free radical scavenger could prevent HPS after CEA. Other studies also proposed staged treatment as an easy and effective method to prevent HICH.20 However, this study is a single-centered report with relatively small sample size, and balloon angioplasty in staged I might increase the risk of vascular dissection. In recent years, staged treatment was also performed in our hospital for some patients with severe carotid artery stenosis, but due to the small sample size, further follow-up is still needed. Ge et al21 carried out animal experiments and found that ischemic post-conditioning could reduce the incidence of HPS to some extent. However, since the study only focused on animal models, whether it could be applied in clinical practice still needs further investigation.

The limitations of these data should be noted. First, the incidence of HICH was low: the small sample sizes may cause statistical bias. Second, this was a retrospective analysis, so selection bias may be built into the data. A larger sample sizes study is required to evaluate the results above.

In conclusion, the results of our study showed that patients with severe unilateral carotid stenosis, preoperative TTP index >1.22, the presence of near occlusion, and poor compensation of both anterior and posterior circulations, have a higher risk of developing HICH after CAS. For patients with the above risk factors, strict monitoring should be given to prevent HICH onset after surgery.


View Abstract


  • LL and DD are joint co authors.

  • Contributors LZ and DD contributed equally to the preparation of the manuscript; ZL and GD contributed equally to data collection and clinical follow-up; Y-wZ, PY, QH, YX, BH, and JL contributed equally to interventional procedures.

  • Funding This work was supported by the National Natural Science Foundation of China grant number 81501008.

  • Competing interests None declared.

  • Patient consent Not required.

  • 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.