Purpose. To report plaque characteristics by virtual histology intravascular ultrasound (VH IVUS) and their association with patient characteristics. Methods. We utilized VH IVUS before stent delivery to classify the stenotic segment. We retrospectively assessed the association of plaque type, digital subtraction angiography, patient characteristics, and incidence of periprocedural myocardial infarction (MI), death, and transient ischemic attack (TIA)/stroke. Results. Among 34 patients, prevalence of pathological intimal thickening, calcified fibroatheroma, calcified thin-cap fibroatheroma, fibroatheroma, and thin-cap fibroatheroma were: 52.9%, 17.6%, 17.6%, 8.8%, and 2.9%, respectively (p < 0.0001). There was an association between hyperlipidemia and prior TIA/amaurosis fugax with necrotic core content (NC) > 10% by VH IVUS that did not achieve statistical significance (p = 0.105 and p = 0.134, respectively). Results showed 1 stroke and 2 TIAs (1.6% and 3.2%, respectively), and no MI or death. Periprocedural TIA/stroke was associated with fibroatheroma by VH IVUS (p = 0.016). The presence of a type III aortic arch was the only angiographic feature associated with TIA/stroke (p =0.031). Conclusions. Pathological intimal thickening was the most common plaque type. Fibrocalcific plaque was not observed. Fibroatheroma on VH IVUS may help identify patients at higher risk of periprocedural major and minor stroke. A larger, prospective series of VH IVUS in carotid disease management is warranted.
VASCULAR DISEASE MANAGEMENT 2011;8:E144–E150
Key words: Carotid artery stenosis, hyperlipidemia, stroke
Stroke remains a leading cause of morbidity and mortality in the United States.1,2 Experience with carotid artery stenting has grown over the past 2 decades and has become an acceptable therapy for selected patients with carotid artery stenosis to reduce the risk of stroke.3–8 An uncommon complication of carotid artery stenting (CAS) is major stroke, which occurs periprocedurally in 3% of cases.9 The utility of various noninvasive and invasive imaging modalities in risk stratifying patients with occlusive carotid disease for periprocedural stroke is the subject of much debate.10,11
VH IVUS of carotid atherosclerotic segments is an area of active exploration, and may help identify patients at risk of stroke with CAS.12–19 The CAPITAL study examined the use of VH IVUS in carotid atherosclerotic disease by comparing stenotic segments visualized in vivo with VH IVUS, with ex vivo carotid endarterectomy (CEA) specimens in 30 patients.17 The diagnostic accuracy of VH IVUS to agree with true histology from CEA specimens in different carotid plaque types was 99.4% in thin-cap fibroatheroma (TCFA), 96.1% for calcified thin-cap fibroatheroma (CaTCFA), 85.9% in fibroatheroma (FA), 85.5% for fibrocalcific (FC), 83.4% in pathological intimal thickening (PIT), and 72.4% for calcified fibroatheroma (CaFA). PIT and FC are defined as having less than 10% (NC), whereas the other plaques all have greater than 10% NC. This study further showed that patients on aspirin had plaque with significantly reduced NC content. In another series, VH IVUS from 37 CEA specimens was performed ex vivo and compared with histology, and revealed a significant association between NC content by VH IVUS and intraplaque hemorrhage by histology.15 Use of VH IVUS in a series of 18 patients undergoing CAS led to the identification of high-risk plaque features in 2 patients and resultant CEA instead of CAS, and led to changes in stent sizing or stent type in 8 patients.16 A series of 25 patients undergoing CAS reported a strong association between total plaque volume and fibro-fatty volume on VH IVUS, and the quantity of atherosclerotic debris obtained on retrieval of the distal embolic protection device.13
The relationship of plaque type by VH IVUS and patient characteristics has been investigated for coronary atherosclerotic segments. In a series comparing VH IVUS with histology obtained from atherectomy, patients with acute coronary syndrome (ACS) had significantly more NC content; diagnostic accuracy of VH IVUS was 87.1% to 96.5%,20 comparable to the CAPITAL study. A higher fibrous content has been found for stable CAD patients at low risk for coronary heart disease death by Framingham risk criteria, with a trend towards more NC in higher Framingham risk patients.21 Higher ratios of NC to dense calcification (NC/DC > 3) have been associated with low HDL, high LDL, and tobacco use.22 NC content either as TCFA or “ruptured plaque” was significantly higher with ACS in a series of 318 patients,23 but was only found to be associated with ACS in another series of 65 patients in the presence of elevated C-reactive protein (CRP).24 Stable angina patients were similarly found to have higher culprit lesion NC content in the presence of elevated high-sensitivity CRP (hs-CRP).25 Vasospastic (variant) angina patients have been shown to exhibit less NC or TCFA than unstable angina patients.26 Diabetics undergoing coronary angiography for ACS have higher NC content and more TCFA than non-diabetics.27 Finally, in a global registry of over 3,000 patients undergoing coronary VH IVUS, NC and dense calcium content have been shown to increase with age and are higher for men than women.28 This same registry has elucidated a significant positive association between NC content and individual patient characteristics of serum LDL, diabetes, hypertension and prior MI.29
Higher NC content was associated with greater elevations of creatine kinase-MB (CK-MB) and troponin and ST segment depression for patients undergoing percutaneous coronary intervention (PCI) for NSTEMI and stable angina.30–32 Among patients undergoing PCI for STEMI, re-elevation of ST segments post procedure was significantly more common in patients with lesions of higher NC volume.33 No reflow phenomenon was significantly more common with higher NC content in ACS patients undergoing PCI in 2 series.34,35
In summary, the coronary literature has demonstrated an association between risk factors for cardiovascular and cerebrovascular disease and NC content or high NC content plaque type by VH IVUS. An association between high NC content or high NC plaque type and risk of adverse events during PCI has also been demonstrated. In this retrospective series, we examine the association of VH IVUS in occlusive carotid disease with risk factors for cerebrovascular and cardiovascular disease, and incidence of periprocedural MI and stroke/TIA.
Two primary operators (MHW and BMW) performed VH IVUS in 61 cases of carotid artery stenting from December 2007 to May 2010 at a single high-volume center (>50 cases/year/operator). All patients had pre-CAS arterial Doppler and CT or MR angiography, and most also had invasive diagnostic angiography. Retrospective review of the procedure, pre-procedure invasive and noninvasive angiography, and post procedure hospital course was conducted with stored offline data and the electronic medical record in June 2010. The University of Pittsburgh Medical Center Medical Advisory Board approved this project.
Carotid artery stenting
A micropuncture needle was used to access the common femoral artery via the Seldinger technique, and an 8 Fr sheath was placed. Over 90% of patients received bivalirudin unless they arrived to the angiography lab on unfractionated heparin. In these cases, heparin was the anticoagulant and an ACT of 250–300 was achieved. A 5 Fr H1 catheter (Cook, Bloomington, Indiana) was used to selectively engage the innominate or left common carotid artery. The .035″ Wholey wire (Mallinckrodt, Hazelwood, Missouri) was next advanced to the external carotid artery and a 5 Fr multipurpose catheter used to introduce an 8 Fr multipurpose catheter to the distal common carotid. In all cases the lesion was first traversed with a distal embolic protection device (RX Accunet™ Embolic Protection System, Abbott Laboratories, Abbott Park, Illinois or SpideRX™, ev3, Irvine, California). Pretreatment with atropine was routine. Lesions were predilated with a single 4 mm balloon inflation and postdilated with a 5 mm balloon. A single “open-cell” RX Acculink® 6–8 mm (Abbott) or Protégé® RX Carotid (ev3) self-expanding, tapered stent was deployed. Digital subtraction cerebral angiography was performed poststenting in all cases.
Virtual Histology IVUS
VH IVUS was performed with the 20 MHz Eagle Eye® Gold Catheter (Volcano Corporation, Rancho Cordova, California) via manual pullback (1 mm/sec) immediately after placement of the embolic protection device. Two observers (BMW and GMS), blinded to each other and to patient characteristics and periprocedural events, analyzed all segments offline (Volcano s5 version 2.2.3, Volcano Corporation). Up to 30 frames per patient were analyzed. Plaque type was classified according to Diethrich et al17 in the CAPITAL study into PIT, FA, FC, CaFA, TCFA, and CaTCFA. CAPITAL plaque definitions are as follows: PIT: intimal media thickness was >600 mm, the fibrofatty plaque component was >10%, and there was confluent NC or calcium that amounted to <10% of the total plaque cross sectional area; FA: confluent NC >10% of the total plaque cross-sectional area; FC: confluent area of calcium >10%, with NC and fibrofatty plaque each <10% of the total plaque cross-sectional area; CaFA: FA with a confluent area of calcium; TCFA: NC >10% of the total plaque cross-sectional area, and the NC was confluent against the lumen; CaTCFA: TCFA plaque with a confluent area of calcium. Plaque components, lumen and external vessel borders, and plaque type were classified based on multiple contiguous frames spanning the minimal lumen area.
Periprocedural complications were determined through chart review of the complete hospitalization. Periprocedural major stroke was defined as persistent hemiplegia or other major deficit, or a deficit requiring hospital discharge to a rehabilitation facility. Periprocedural minor stroke or TIA was defined as a focal neurological deficit lasting less than 24 hours and discharge to home. Periprocedural MI was defined as any of the following: chest pain syndrome; any elevation of serum troponin, CK-MB; new ST segment elevations >1 mm, or ST depressions >2 mm.
The distribution of observed plaque types was calculated with a chi-square test. Association of patient characteristics and angiographic features, procedural outcomes, and VH IVUS plaque type was assessed with Fisher’s exact test (2-sided). Univariate analysis was first performed for baseline and periprocedural patient characteristics using NC content > 10% to identify candidate variables since plaque types defined in CAPITAL are classified with this cut-point. After identifying potential associations (p < 0.150), we again performed univariate analysis (Fisher’s exact test, 2-sided) for each observed plaque type and the selected preprocedural and periprocedural clinical risk factor or event, each expressed as a dichotomous variable. Interobserver agreement for plaque classification was determined using a quadratic modification of Cohen’s kappa statistic and is reported elsewhere.36 Computations were performed with SPSS version 18 (IBM Corporation, Armonk, New York).
In this series of 61 patients undergoing CAS, pre-stent VH IVUS data were available for 34 individuals. Twenty-five cases underwent CAS for symptomatic high-grade stenosis. The remainder underwent CAS for asymptomatic high-grade stenosis with an FDA-approved investigational device exemption. Filter dwell times were less than 5 minutes. All 61 had passage of the IVUS catheter across the lesion. Periprocedural outcome of TIA, stroke, major vascular complication, MI and death was ascertained for all 61 patients. In 16 patients, IVUS was performed in a re-stenosed stent, without full VH capability, or after stenting. In 11 patients VH IVUS images were available in the electronic medical record, but original data were not available for border redrawing. The baseline characteristics of the 34 individuals with complete clinical and VH IVUS data are presented in Table 1. Three periprocedural neurological events were identified among the 61 cases: 1 major stroke and 2 TIAs (1.6% and 3.2%, respectively). There were no major vascular complications, myocardial infarctions, or periprocedural deaths. All stent procedures were successful.
Representative VH IVUS plaque types are shown in Figures 1 and 2. There was 1 major ipsilateral stroke resulting in hemiplegia occurring the day after carotid artery stenting in a patient with FA plaque by VH IVUS. This patient had prior carotid endarterectomy with sequential stenoses of the common and internal carotid artery and was symptomatic before her CAS. There were 2 ipsilateral TIAs occurring immediately postprocedure. In one case, the plaque was a CaTCFA; the lesion was greater than 90%, heavily calcified, and a concomitant stenosis with dystrophic calcification of the cavernous segment was present. The second ipsilateral TIA was in an octogenarian with FA. VH IVUS plaques for these 3 patients with periprocedural TIA/stroke are presented in Figures 1D and 3.
Plaque type for 34 patients is presented in Table 2. PIT was the most common plaque type, TCFA least common, and there were no cases of FC (p < 0.0001). A non-significant association was observed with higher NC in patients with hyperlipidemia and preprocedure TIA/amaurosis fugax (p = 0.105 and p = 0.134, respectively), presented in Table 3. Analysis of individual plaque types again revealed an insignificant association between less hyperlipidemia and TIA/amaurosis in patients with PIT (p = 0.105 and 0.088, respectively). Periprocedural TIA/stroke was significantly associated with FA by VH IVUS (p = 0.016), presented in Table 4. Lesion characteristics, arch anatomy, and VH IVUS plaque type for the 3 periprocedural neurological events are presented in Table 5. A type III aortic arch was present in 2 of the 3 patients with periprocedural TIA/stroke, whereas only a single type III arch was present among the remaining patients (p = 0.031). Interobserver agreement for VH IVUS plaque classification was Κ = 0.743.
We found a trend towards higher NC content in patients with hyperlipidemia or prior TIA/amaurosis driven mainly by lower NC content and PIT in patients without hyperlipidemia or prior TIA/amaurosis fugax. These findings parallel the coronary literature showing a positive association between unstable symptoms, hyperlipidemia, and high NC content plaques.21,22,29 Only 5 of the 6 plaque types described in the CAPITAL study (n = 30) were present in this series of 34 patients (PIT, FA, CaFA, CaTCFA, and TCFA).17 The absence of FC from our series and our moderately strong agreement between observers raises the question whether this plaque type, as defined in CAPITAL, warrants distinction from the others.
The patient and angiographic characteristics we examined have been included in prior CAS trials and registries, and identified as either contributing to carotid occlusive disease or conveying higher procedural risk with CAS.6,7,9,37 The characteristics of our patients undergoing CAS are similar to those reported previously. Our incidence of periprocedural TIA and major stroke compares well to prior trials comparing CAS and CEA.8,9
We found a trend towards higher NC and periprocedural TIA and major stroke, and a statistically significant association of FA with this important periprocedural outcome. This suggests the potential for distal embolization may be greater with higher NC. This finding parallels the PCI literature indicating a higher risk of periprocedural myocardial ischemia and necrosis with higher NC in both stable and unstable coronary artery disease.26–28,30–34 Sampling of embolic debris during CAS in one study demonstrated an association between plaque fibrofatty content by VH IVUS and the mass of debris retrieved.13
Our determination of VH IVUS plaque type was based on mutually blinded assessment of plaque and periprocedural outcomes whereby each observer established their own luminal and vessel borders from the provisional images. When observations differed, the plaque type with the higher NC content was used for one TIA case, and the plaque with lower NC was used for the non-stroke cases. This assumption was justified on the basis of the aforementioned studies revealing the association of NC content and high NC plaque types with adverse events during PCI. The substantial interobserver agreement for our series argues against a systematic under or over classification with respect to NC. These findings illustrate the need for defining plaque type with VH IVUS in occlusive carotid disease in a larger series or registry.
Presence of a type III aortic arch, whereby the innominate arises from the ascending aorta, was the sole angiographic and clinical variable significantly associated with periprocedural stroke risk in this small series. We included a number of angiographic features associated with higher periprocedural stroke in our analysis in an attempt to explain these periprocedural events. A type III arch requires longer procedure time and more catheter manipulations which may result in ipsilateral or contralateral embolic events.37 This may justify direct carotid puncture as a viable means of access in selected cases with a type III or IV aortic arch, as has been reported elsewhere.38
Strengths and limitations
In this small retrospective case series of 61 patients, complete VH IVUS data were only available for 34 individuals. We were able to determine periprocedural incidence of stroke, TIA, MI, major vascular complication, and death for all 61 patients. We included all 61 patients in this series for several reasons. First, our lab used VH IVUS for the entire time represented by these 61 patients, a period during which our techniques did not change. Second, including all patients from this period allowed us a larger sample size and internal control for the purposes of monitoring our procedural success and safety. No systematic difference for the 27 patients without available VH IVUS could be identified. Our total procedure time was rarely more than 45 minutes, and did not change significantly with use of our single VH IVUS pullback (data not shown). Automated pullback could have been preferable to manual pullback for determining plaque geometry; but we believe this did not favor the imaging of 1 or more plaque types and so our determination of plaque type at the minimal lumen area should remain valid.
As this was a single-center case series with only 2 primary operators, our findings may not be generalized where other devices and techniques for CAS are employed. However, we utilized devices and techniques supported by large trials.
Accuracy of provisional border detection with the software used in this study was fair (data not shown), requiring redrawing of lumen and vessel borders in most cases. However, the substantial agreement between observers for plaque type after blinded redrawing suggests that the VH component of this software version was relatively robust for experienced viewers. This study did not reveal any FC, possibly reducing our ability to generalize our results. More likely, our substantial interobserver agreement for plaque type suggests that this plaque type is less common in occlusive carotid disease.
We do not routinely obtain EKGs or cardiac enzymes post-CAS without clinical evidence of acute coronary syndrome. It is likely that our periprocedural incidence of subclinical MI as detected by mild elevations in serum markers would be higher than the 0% incidence of MI observed in this series. The significance of these subclinical events in the context of CAS is uncertain.
We attempted to minimize bias by conducting independent assessments of plaque type, which were both blinded between interpreters and with respect to patient outcome. Blinding of outcome during assessment of angiographic features was not possible. However, arch type was reported in the angiographic procedure in most cases and was included in the analysis as such.
This small series was underpowered to examine the association of plaque type or NC alone and periprocedural TIA/stroke rate. Periprocedural TIA/stroke is an uncommon complication of CAS, and a sample size of >350 is required to examine this association a priori. In this retrospective series undertaken for the purpose of further hypothesis generation, causation between FA, a type III arch, or other patient characteristics and TIA/stroke events cannot be claimed.
In this small case series of occlusive carotid disease, PIT was the most common plaque type. There was an insignificant positive association between preprocedure risk factors of hyperlipidemia and TIA/amaurosis and higher NC by carotid VH IVUS. We observed a statistically significant association between VH IVUS fibroatheroma and major and minor stroke. We believe that a larger, prospective series of VH IVUS in carotid disease is required to explore the association and relative importance of plaque type and NC with clinical risk factors, angiographic risk factors, and periprocedural stroke/TIA.
From the 1University of Pittsburgh Medical Center, Presbyterian Hospital, Cardiovascular Institute, Pittsburgh, Pennsylvania and 2Carnegie Mellon University, Pittsburgh, Pennsylvania.
The authors report no financial relationships or conflicts of interest regarding the content herein.
Manuscript submitted March 3, 2011, provisional acceptance given April 29, 2011, final version accepted June 2, 2011.
Address for correspondence: Brion M. Winston, MD, Brigham and Women's Hospital, Interventional Cardiology, 75 Francis Street, Boston, MA 02115. Email: [email protected]