Carotid Plaque Echolucency Increases the Risk of Stroke in Carotid Stenting
The Imaging in Carotid Angioplasty and Risk of Stroke (ICAROS) Study
Background— Carotid artery stenting (CAS) has recently emerged as a potential alternative to carotid endarterectomy. Cerebral embolization is the most devastating complication of CAS, and the echogenicity of carotid plaque has been indicated as one of the risk factors involved. This is the first study to analyze the role of a computer-assisted highly reproducible index of echogenicity, namely the gray-scale median (GSM), on the risk of stroke during CAS.
Methods and Results— The Imaging in Carotid Angioplasty and Risk of Stroke (ICAROS) registry included 418 cases of CAS collected from 11 international centers. An echographic evaluation of carotid plaque with GSM measurement was made preprocedurally. The onset of neurological deficits during the procedure and the postprocedural period was recorded. The overall rate of neurological complications was 3.6%: minor strokes, 2.2%, and major stroke, 1.4%. There were 11 of 155 strokes (7.1%) in patients with GSM ≤25 and 4 of 263 (1.5%) in patients with GSM >25 (P=0.005). Patients with severe stenosis (≥85%) had a higher rate of stroke (P=0.03). The effectiveness of brain protection devices was confirmed in those with GSM >25 (P=0.01) but not in those with GSM ≤25. Multivariate analysis revealed that GSM (OR, 7.11; P=0.002) and rate of stenosis (OR, 5.76; P=0.010) are independent predictors of stroke.
Conclusions— Carotid plaque echolucency, as measured by GSM ≤25, increases the risk of stroke in CAS. The inclusion of echolucency measured as GSM in the planning of any endovascular procedure of carotid lesions allows stratification of patients at different risks of complications in CAS.
Received August 4, 2003; de novo received December 20, 2003; revision received March 1, 2004; accepted March 17, 2004.
Stroke is a leading cause of death and disablement in Western countries. As demonstrated in multicenter randomized trials,1,2 carotid endarterectomy (CEA) is the gold standard to reduce stroke in symptomatic and asymptomatic patients with significant carotid stenosis.
Carotid artery stenting (CAS) has recently emerged as a potential alternative to CEA3 because of less invasiveness and a shorter hospitalization. The number of CAS procedures performed has steadily increased and currently makes up 9% of the total number of carotid procedures, with favorable results.4,5 The use of balloon angioplasty and stenting in carotid arteries is still limited because of concern about the lack of long-term results and neurological complications during CAS. Cerebral embolization from carotid plaque is indeed the most severe complication of CAS.6
Several randomized trials are at present comparing CEA with CAS.7 Inclusion and exclusion criteria for these trials usually consider the degree of carotid stenosis and the presence or absence of neurological symptoms, without paying any attention to the morphology of the carotid plaque.8
Early studies evaluated the relationship between the characteristics of the plaque and the presence or absence of neurological symptoms or cerebral CT lesions, with conflicting results. However, the advent of high-resolution B-mode scanners and the use of a quantitative computer-assisted index of echogenicity such as gray-scale median (GSM), introduced by our teams, have greatly improved the correlation between plaque characterization and clinical features.9,10 Only after the introduction of the image normalization did GSM become a highly-reproducible index of the echolucency of carotid plaques with low interobserver and interscanner variability.11
Several studies indicated that echogenicity is related to the histological components of carotid plaques12,13 and that carotid plaque echolucency (low echogenicity) is associated with the development of neurological events14–17 and with an increased number of emboli after CEA and CAS.18–20 On the basis of these assumptions, our groups suggested that plaque echogenicity measured by GSM can be a useful indicator of embolic potential in the carotid arteries.8,21
The aim of the Imaging in Carotid Angioplasty and Risk of Stroke (ICAROS) registry22 was to determine the preprocedural echographic criteria that may identify the carotid plaque related to a higher risk of stroke during CAS so that a better selection of candidates for CAS can be performed.
ICAROS was an international multicenter registry that collected cases of carotid angioplasty and stenting procedures between July 2000 and December 2001 from 11 participating centers. The institutional review board approved the study. Patients included received full details about the procedure and the follow-up scheme and signed an informed consent. Information on devices (guidewires, catheters, balloons, stents, brain protection devices [BPDs]) and stenting procedures was collected, together with demographic and clinical characteristics, on a data form sent by mail to the coordinating center (Bassini Teaching Hospital). Patient investigation was made in detail as previously described.22
Inclusion criteria (defined before the study) were the following: hemodynamic (≥70%) carotid stenosis, neurologically symptomatic and asymptomatic, and both NASCET/ACAS eligible and ineligible. Exclusion criteria included the following: nonhemodynamic stenosis (<70%); highly calcified, prolonged stenosis extensively involving the bifurcation beyond the internal carotid artery (ICA); an angulated ICA; intraluminal thrombus (discrete intraluminal filling defects with reduced contrast density, haziness, and irregular lesion contour); extensive aortic or brachiocephalic trunk plaque; neurologically unstable; and a life expectancy of <5 years.
A brain CT scan was considered positive for ischemic lesions if a discrete (>1 cm) subcortical or cortical infarction was present in the anterior and middle cerebral artery territories in the absence of cardioembolism (excluded on the basis of medical history and ECG).
To assess the role of GSM in different subsets, patients were subcategorized in eligible and ineligible according to NASCET and ACAS inclusion and exclusion criteria.1,2
Carotid Plaque Imaging
Degree of stenosis was calculated as the ratio of the peak systolic velocity of the ICA to that of the common carotid artery.23 This hemodynamic evaluation of carotid stenosis was based on velocity cutoff points related to angiographic evaluation of stenosis used in NASCET.1 Angiographic evaluation was performed as previously described.22
Several training courses were organized worldwide for ultrasonographers from the participating centers on how to set up the duplex scanner to standardize the collection of images. B-mode scan settings were adjusted as previously described.11
A 15-second super-VHS video or magneto-optical disk recording was performed to show the vessel and plaque from a longitudinal view. Images were then sent to the coordinating center and transferred onto a personal computer. Image normalization and calculation of the GSM were performed by the same operator, who was blinded to clinical data and outcome and used Adobe Photoshop 5.0 software as previously described (Figure 1).11
The occurrence of any stroke and transient ischemic attack (TIA) during the procedure, at discharge, and after 30 days was carefully evaluated by means of CT or MR and by an independent neurologist who was blinded to clinical data, outcome, and GSM value. A TIA was defined as a focal neurological deficit lasting <24 hours. A stroke was considered disabling (major) if patients had a Rankin score of ≥3 at 90 days.
A Wilcoxon rank-sum test was used for nonparametric comparisons of the medians of GSM. Comparisons between event rates in different subgroups (univariate analysis) were performed by means of Fisher’s exact test.
A logistic regression model was applied to investigate the impact of various factors on the odds of any stroke and death (primary end-point analysis) or of any stroke and death plus TIA (secondary end-point analysis) as occurred during the procedure or within 30 days. The regression factors considered were related to both relevant clinical aspects and the procedure, namely GSM, degree of stenosis, neurological symptoms, preprocedural brain CT, type of lesion, use of BPD, age, sex, and NASCET/ACAS eligibility.
A backward procedure was performed (α=0.10) to identify the final model, and ORs with 95% CIs were estimated. The presence of possible interactions was also considered when applicable. All tests were performed with a significance level of α=0.05 (2 sided).
A receiver-operating characteristic (ROC) curve was used to choose the GSM cutoff value with the best sensitivity and specificity. The logistic model was used to analyze the role of different GSM values on the basis of maximization of the likelihood of stroke and minimization of probability value.
ICAROS reported 496 cases of CAS worldwide, of which 78 (15.7%) were excluded because of low-quality images or incomplete information on data forms.
The risk factors and clinical features of the 418 patients analyzed are summarized in Table 1. In particular, the GSM distribution is shown in Figure 2. There was no correlation between GSM and degree of stenosis (r=0.09).
Stenting procedures were performed in 415 cases (99.3%). A BPD was applied in 219 patients (52.4%): 210 distal (95.9%) and 9 proximal BPDs (4.1%).
There were 13 TIAs (3.1%; 10 ipsilateral), 9 minor strokes (2.2%; 7 ipsilateral), and 6 major strokes (1.4%; 6 ipsilateral); no deaths were observed. The 30-day combined death and any stroke rate was 3.6%.
The GSM value in complicated patients was significantly lower than in uncomplicated patients in both the stroke (20.80±17.43 versus 35.07±19.60; P=0.0036; Figure 3A) and the stroke plus TIA (20.93±15.47 versus 35.53±19.61; P=0.0001) subsets (Figure 3B).
Figure 4A shows an ROC curve used to determine the cutoff value of GSM to identify patients at higher risk of complications during CAS; a GSM value of 25 was used as the threshold. A comparison of the significance and risk of complications at different GSM values using the logistic regression model was used to analyze the role of different GSM cutoff points (Figure 4B).
Eleven of 155 patients with a GSM ≤25 had complications compared with 4 of 263 patients with GSM >25 (7.1±2.1% versus 1.5±0.8%; P=0.005; Figure 5A). The event rates increased to 12.9% and 3.0%, respectively, when both stroke and TIA were counted (12.9±2.7% versus 3.0±1.1%; P=0.002; Figure 5B).
Detailed descriptions of the outcome by subgroups defined according to each factor considered and GSM levels are reported in Table 2.
There were 5 of 219 complications (2.3%) in protected and 10 of 199 (5.0%) in unprotected procedures (P=0.188). However, protection gave different results in the GSM subgroups. In patients with GSM ≤25, the use of a BPD tended to increase the risk of complications (P=0.153), whereas it had a protective value in the GSM >25 subgroup (P=0.010).
The overall complication rate was higher in primitive lesions than in restenosis (5.2% versus 2.2%; P=0.117). This difference was also observed in GSM >25 patients (4.0% versus 0%; P=0.020) but not in GSM ≤25 patients (6.6% versus 7.8%; P=0.762).
Multivariate regression analysis revealed that GSM (OR, 7.11; P=0.002) and degree of stenosis (OR, 5.76; P=0.010) are significant independent predictors of stroke alone, whereas preprocedural symptomatology and preprocedural brain CT are borderline significant (Table 3). Similar results were found in the analysis of stroke plus TIA as end points (data not shown). The logistic regression model confirmed the role of GSM as a predictor of stroke in both NASCET/ACAS-eligible (P=0.024) and -ineligible (P=0.031) patients.
The ICAROS study was the first large study that was aimed at evaluating the relationship between the echolucency of carotid plaque, as measured by GSM, and the risk of stroke during CAS.
In the ICAROS study, the rate of stroke was higher in GSM ≤25 patients than in those with GSM >25, which confirms that echolucent plaques have a significantly higher embolic potential when treated with angioplasty and stenting. This finding was also confirmed in a multivariate analysis in which degree of stenosis was the only other relevant independent predictor of stroke.
The role of GSM as predictor of stroke was confirmed in both NASCET/ACAS-eligible and -ineligible patients.
The higher neurological complication rate in patients with severe stenosis could be explained by a higher number of embolic particles released from the crossing of endovascular devices through a tighter lesion.18
The complication rate was 5.3% for symptomatic and 2.8% for asymptomatic patients. These rates are within limits set by the American Heart Association (6% for symptomatic, 3% for asymptomatic patients) for CEA.24 In the GSM >25 subset, results were far better, with a rate of 3.3% for symptomatic and 0.6% for asymptomatic patients. The selection of patients according to their risk factors (echolucency may be one of these, but many unknown factors remain) could be very important for improving the outcome of CAS.
The use of BPDs reduced neurological complications (although not significantly) in the overall population, as previously described.20,25 Nevertheless, in the GSM ≤25 subset, BPDs did not seem to maintain their efficacy. It should be noted that the BPDs used in the different centers were almost exclusively distal BPDs (96% of the total), requiring the passage of the device across the lesion with the inherent risk of embolization. A proximal BPD based on reverse flow (with no crossing of the plaque) could eventually be a better option in this kind of plaque. On the other hand, the use of BPDs in GSM >25 plaques provided significant protection.
Restenosis, frequently with a homogenous fibrous content and a high GSM value, is associated with a low-embolic-risk endovascular procedure, as confirmed by our study. Nevertheless, the analysis of restenosis with a GSM value <25, although less frequent, showed for the first time a rate of stroke similar to that for primitive lesions.
Unlike a previous study, age was not a predictor of stroke in the present study, probably because of the smaller number of patients >80 years of age.5
The first limitation of the present study is related to the observational and less rigorous design of a multicenter registry compared with a randomized trial in which the selection of patients is more accurate and the procedural protocol is more homogenous in all the centers, which is less likely to create bias. Second, in all the evaluations made, we discriminated between hypoechoic and hyperechoic plaques using a cutoff value of 25. The clinical relevance of this cutoff should be validated in future studies. Third, we should remember that the reproducibility of the GSM after a training program is very high11 but not absolute. Finally, a long-term analysis of the results of CAS should be performed that considers the change in the morphology of the carotid plaque over time.26
The high-resolution B-mode scanners provided a reliable analysis of echogenicity, allowing us to recognize the role of echolucency in predicting neurological events. Several studies analyzing >6000 patients showed that echolucent plaques have an increased risk of ischemic cerebrovascular events, with a relative risk of 1.9 to 4.6.14–17
Our research teams separately evolved a similar computerized method for analyzing the echolucency of the carotid plaque. We demonstrated for the first time that patients with neurological symptoms and brain CT scans positive for stroke had an increased incidence of echolucent low-GSM-value plaques, which are correlated with high lipid and hemorrhagic content.9,10,12,21 Only after the introduction of the image normalization did GSM become a highly reproducible index, allowing us to compare images from different scanners by different ultrasonographers and through different peripherals.11
To establish an indication to treat patients with an echolucent plaque, we should compare the long-term results of the CAS (as emerging from the follow-up of the ICAROS patients) with those of medical therapy and CEA. The ongoing Asymptomatic Carotid Stenosis and Risk of Stroke (ACSRS) study27 is analyzing the rate of neurological events in patients with echolucent carotid plaques, as measured by GSM, treated with the best medical therapy alone. A study analyzing the relationship between GSM and the rate of complications in CEA is under consideration. Only by comparing these results will it be possible to choose the best treatment for patients with echolucent plaques stratified according to clinical characteristics (neurological symptomatology, degree of stenosis, etc). A GSM value <25 would not be a contraindication to endovascular treatment if the rate of complications after the best medical therapy or CEA was higher than CAS.
It should be noted that CAS is to date an experimental procedure. For patients at high risk or unsuitable for surgery, CAS could be considered only if physicians with high-level skills and techniques are available and if at least equivalent (if not better) results for CAS than CEA can be shown. Low-risk patients should be treated with CAS only within a clinical trial.28
Further studies are required to assess the difference between bare, drug-eluting, and coated stents for the reduction of embolization to the brain.
In our study, asymptomatic patients were included. Nevertheless, the indication for both surgical and endovascular treatment for these patients remains controversial.29 Selecting asymptomatic patients according to their risk factors (such as echolucency in the ACSRS study)27 may allow only subsets at higher risk to be treated, with maximum advantage for patients and a reduction in expense for the community through the avoidance of unnecessary procedures.
The clinical impact of GSM relies on the ability to identify a wide number of patients (155 of 418, 37%) at higher risk of stroke during CAS and to distinguish subsets of patients (with restenosis or with protected procedure) in whom the rate of complications is different from the overall population.
In conclusion, carotid plaque echolucency, as measured by GSM <25, increases the risk of stroke in CAS. Measuring echolucency is a simple method to identify preprocedurally carotid plaques at higher risk of stroke. Therefore, in addition to neurological symptomatology and degree of stenosis, GSM should always be considered in the planning of any endovascular treatment of carotid lesions.
Dr Biasi was supported by a grant from the Associazione per la Ricerca, la Prevenzione e il Trattamento delle Malattie Vascolari.
Ferguson GG, Eliasziw M, Barr HW, et al. The North American Symptomatic Carotid Endarterectomy Trial: surgical results in 1415 patients. Stroke. 1999; 30: 1751–1758.
Roubin GS, New G, Iyer SS, et al. Immediate and late clinical outcomes of carotid artery stenting in patients with symptomatic and asymptomatic carotid artery stenosis: a 5-year prospective analysis. Circulation. 2001; 103: 532–537.
Markus HS, Clifton A, Buckenham T, et al. Carotid angioplasty: detection of embolic signals during and after the procedure. Stroke. 1994; 25: 2403–2406.
Biasi GM, Mingazzini PM, Sampaolo A, et al. Echographic characterization of carotid plaque and risk for cerebral ischemia. In: Castellani L, ed. Progress in Angiology and Vascular Surgery. Torino, Italy: Minerva Medica; 1995: 59–65.
Sabetai MM, Tegos TJ, Nicolaides AN, et al. Reproducibility of computer-quantified carotid plaque echogenicity: can we overcome the subjectivity? Stroke. 2000; 31: 2189–2196.
Gronholdt ML, Nordestgaard BG, Wiebe BM, et al. Echo-lucency of computerized ultrasound images of carotid atherosclerotic plaques are associated with increased levels of triglyceride-rich lipoproteins as well as increased plaque lipid content. Circulation. 1998; 97: 34–40.
Mathiesen EB, Bonaa KH, Joakimsen O. Echolucent plaques are associated with high risk of ischemic cerebrovascular events in carotid stenosis: the Tromso Study. Circulation. 2001; 103: 2171–2175.
Gronholdt ML, Nordestgaard BG, Schroeder TV, et al. Ultrasonic echolucent carotid plaques predict future strokes. Circulation. 2001; 104: 68–73.
Liapis CD, Kakisis JD, Kostakis AG. Carotid stenosis: factors affecting symptomatology. Stroke. 2001; 32: 2782–2786.
Moneta GL, Edwards JM, Chitwood RW, et al. Correlation of North American Symptomatic Carotid Endarterectomy Trial (NASCET) angiographic definition of 70% to 99% internal carotid artery stenosis with duplex scanning. J Vasc Surg. 1993; 17: 152–157;discussion 157–159.
Biller J, Feinberg WM, Castaldo JE, et al. Guidelines for carotid endarterectomy: a statement for healthcare professionals from a Special Writing Group of the Stroke Council, American Heart Association. Circulation. 1998; 97: 501–509.
Kastrup A, Groschel K, Krapf H, et al. Early outcome of carotid angioplasty and stenting with and without cerebral protection devices: a systematic review of the literature. Stroke. 2003; 34: 813–819.
Barr JD, Connors JJ 3rd, Sacks D, et al. Quality improvement guidelines for the performance of cervical carotid angioplasty and stent placement. AJNR Am J Neuroradiol. 2003; 24: 2020–2034.
Goldstein LB. Extracranial carotid artery stenosis. Stroke. 2003; 34: 2767–2773.