Utility of carotid duplex ultrasonography in a general inner-city hospital

Introduction

Carotid artery stenosis is an important cause of cervicalbruit, amaurosis fugax, acute ischemic stroke and transient ischemic attack (TIA). These patients can be evaluatedwith carotid duplex ultrasonography. Prevalence of carotidartery stenosis in the general population is 0.2% to 7.5% inmoderate (>50%) stenosis and lower 0% to 3.1% in severe(>70%) stenosis [1]. However, only selected asymptomaticindividuals with severe carotid artery stenosis greater than70% are shown to benefit from carotid endarterectomy(CEA) [2]. While CEA reduces incidence of stroke [2],routine use of CDUS has not been shown to reduce perioperative stroke or mortality [3]. CDUS is often used inpatients with planned Coronary Artery Bypass Grafting(CABG), known atherosclerosis (coronary or peripheralartery disease), and syncope. Although studies [4,5] haveshown correlation of carotid artery disease and coronaryartery disease, its utility in asymptomatic patients isunclear [6]. In 2011 American College of CardiologyFoundation (ACCF) delineated recommendations forappropriate use of CDUS [7]. Our objective is to determine the yield of CDUS and identify risk factors of significant cerebrovascular disease (sCBVD). We also seekto calculate prevalence of significant carotid artery diseaseamong appropriate and uncertain indications as per ACCFappropriate use criteria task force. Our hypothesis is thatCDUS is over utilized for uncertain or inappropriate indications at our institution. We hope that our study helpsproviders in recognizing this and thereby promote optimalutilization of carotid ultrasound.

Materials and methods

Study design and patients

After approval of Institutional Review Board at NewarkBeth Israel Medical Center, we included 827 carotid ultrasounds consecutively ordered between March 1st, 2013and August 31st, 2013 at Newark Beth Israel MedicalCenter. Baseline characteristics such as age, sex, diabetesmellitus, systemic hypertension, coronary artery disease(CAD), peripheral arterial disease (PAD), smoking history,carotid bruit and indications for CDUS were abstractedusing retrospective chart review. Importantly, our dataalso included a cohort of patients who underwent CDUSbefore CABG (91), valvular replacement/repair (34) andOrthotopic Heart Transplantation (25).

Study protocol

All studies were performed in vascular laboratory usingduplex scanner GE Logiq E-9 by credentialed vasculartechnologists. Patients were asked to lay supine with chinraised and head slightly turned. Imaging started with common carotid artery (CCA) in transverse position at clavicular level followed by examination in a sagittal planeutilizing gray scale. Then, spectral waveform analysis wasobtained with transducer in sagittal plane and dopplerangle kept between 45 and 60 degrees positioned in proximal CCA. This procedure was repeated along entirelength of CCA and a representative spectral waveformwas recorded of the proximal, mid and distal CCA andbulb area. Doppler samples and representative spectralwaveforms were recorded of proximal, mid and distal internal carotid artery (ICA) and proximal external carotidartery (ECA). If any plaque was noted, waveforms were recorded in area proximal to plaque, at the area of plaqueand distal to the plaque. Similarly, if there was significantstenosis waveform were recorded proximal to stenosis, atthe stenosis and distal to stenosis if they were not alreadydocumented as per stated protocol. Finally, waveforms ofvertebral and subclavian arteries were recorded. Peak systolic velocities, end diastolic velocities, plaque descriptionusing B-mode was also documented.

Stenosis in vertebrobasilar system or greater than 50%ICA stenosis was considered significant cerebrovasculardisease (sCBVD) as per the Intersocietal AccreditationCommission (IAC) vascular testing carotid stenosis interpretation criteria based on Society of Radiologists inUltrasound Consensus Conference [8]. Based on theseguidelines less than 50% ICA stenosis was diagnosed withpeak systolic velocity (PSV) less than 125 cm/sec and novisible plaque or intimal thickening, 50%–69% stenosiswas identified when ICA PSV was 125–230 cm/sec withvisible plaque and ICA PSV greater than 230 cm/sec alongwith visible plaque and lumen narrowing was comparablewith 70% stenosis to near occlusion. Markedly narrowedlumen was suggestive of near occlusion and no detectable lumen or flow at gray-scale was indicative of total occlusion. These generalized thresholds were used to minimizevariability in reporting of degree of stenosis. All resultswere read and confirmed by 2 board certified vascular surgeons to reduce inter-observer variability.

Statistical analysis

Statistical analysis was completed using SPSS software,version 22.0. Continuous data was shown using medianand interquartile range (IQR) while categorical data wasshown in frequencies and percentages. Age, sex, smokingstatus, CAD, PAD, systemic hypertension, diabetes mellitus, TIA, acute ischemic stroke, carotid bruit, cardiac surgery, syncope, and prior carotid endarterectomy (CEA) orcarotid artery stenting (CAS) were fitted as independentvariables and sCBVD was fitted as a dependent variable. Amultinomial step-wise logistic regression was then performed along with likelihood ratio, odds ratio and 95%confidence intervals to determine association of variousrisk factors with sCBVD. A P-value of less than 0.05 wasconsidered statistically significant.

Results

Out of 827 CDUS performed, significant cerebrovasculardisease (sCBVD) was found in 88 (10.6%) cases. Baselinecharacteristics and demographics are shown in Table 1. 62(7.5%) patients were found to have 50-69% ICA occlusion,11 (1.3%) had 70-99% ICA occlusion, 3 (0.4%) patientshad near total occlusion, 10 (1.2%) patients had total ICAocclusion and 2 (0.2%) had vertebral artery occlusion. 7individuals with severe ICA stenosis (70-99%) underwentcarotid endarterectomy while 4 were treated with medicaltherapy. Thus, 11 out of 827 (1.3%) patients had 70-99%stenosis in which a carotid intervention was deemedbeneficial.

Yield of carotid ultrasound was different among variousindications (Table 2 and Figure 1). Prevalence of sCBVDwas higher in all appropriate use indications whencompared to uncertain use indications except in patients undergoing heart transplant (Table 2). Patients withage≥65 (OR 2.1, 95% CI 1.2-3.7; P = 0.006), carotid bruit(OR 7.7, 95% CI 3.6-16.6; P <0.001) and prior carotid endarterectomy (CEA) or prior carotid artery stenting (CAS)(OR 5.8, 95% CI 2.3-14.8; P <0.001) were significantlymore likely to have sCBVD (Table 3). Out of 26 patientswith previous carotid endarterectomy, sCBVD was seen in14 patients. Among 14 cases, contralateral ICA stenosiswas seen in 11 patients with 50-69% stenosis in 3 patients,70-99% stenosis in 5 patients and absent flow in 3 patients. Remaining 3 out of 14 patients had stenosis in ipsilateral ICA. Mean peak systolic velocities (PSV) are listedin Table 4. Average PSV was 228.9 cm/s and 556.1 cm/sin patients with 50-69% stenosis and 70-99% stenosisrespectively.

CDUS had lowest yield in patients younger than 65 years(6.4%), presenting with syncope (6.5%) non-focal neurologic symptoms (6.1%) such as delirium, gait imbalance,paresthesias and patients undergoing valvular repair or replacement (5.9%). Of note, sCBVD was higher (16%) inpatients undergoing Orthotopic Heart Transplantation orLeft Ventricular Assist Device placement but didn’t reachstatistical significance.

Discussion

Interestingly, prevalence of severe carotid artery stenosis(≥70%) was lower (1.3%) in our study considering it included patients with symptomatic carotid artery disease.This is similar to previous studies albeit in asymptomaticpatients [1,9]. Recent guidelines adopted by the ACCF[7] and USPSTF [9] indicate that performing“screening”CDUS in asymptomatic patients is not as useful compared to symptomatic patients. Evaluation of ischemicstroke, TIA, focal neurological symptoms, cervical bruitas well as asymptomatic patients with known CAD, PADor abdominal aortic aneurysm was considered appropriate indications. However syncope, planned CABG andvalvular surgery were labeled as an“uncertain”indication. Carotid intima-media thickness [6] and computerized tomography guided coronary artery calcium scoringare two of the most commonly used techniques to detectsubclinical atherosclerosis [10,11]. However, carotid duplex ultrasound exam does not include formal measurement of carotid intimal medial thickness. In addition,several studies support the association of CAD and carotid artery disease [4,10] as well as between PAD andcarotid artery disease [12,13]. Thus, patients with CADor PAD may warrant assessment of carotid artery diseaseeven in the absence of any focal neurological symptoms.Carotid bruit can also signify systemic atherosclerosis[14] and thus impart increased risk of stroke [15], TIAs

and mortality [16]. Some studies indicate the risk ofstroke after cardiac surgery to be about 2% in patientsundergoing CABG [17,18]. Factors such as advanced age,presence of PAD, left main disease, diabetes mellitus,systemic hypertension, stroke and history of previouscerebrovascular disease could confer higher risk of perioperative stroke and thus suggest the need for screeningCDUS [17]. While a few studies have demonstrated correlation between these risk factors and carotid artery stenosis≥50% [19,20] these risk prediction models are notwell validated. Moreover, there are no randomized controlled trials to support performing CEA in asymptomaticindividuals prior to vascular surgeries such as CABG and valvular repair/replacement. Routine use of CDUS beforeheart transplantation is also unclear.

Several factors can lead to overutilization of carotidultrasound such as ease of availability, lack of conclusiveevidence supporting CDUS in asymptomatic individualsand concerns for peri-operative stroke. To date there is noconclusive evidence that routinely screening asymptomatic individuals for carotid artery stenosis reduces perioperative stroke. In symptomatic patients, CDUS candetect significant carotid artery stenosis in order to reduce risk of peri-operative stroke via medical therapyor surgical intervention. Three RCTs comparing carotidendarterectomy with medical therapy showed 27% reduction (RR 0.72, 95% CI 0.58-0.90) of stroke 30 days postCEA [21]. Recent advances in medical therapy with higherstatin use has also reduced incidence of ipsilateral stroketo about 1.13% per year [21,22]. ACST-1 trial included patients without stroke, TIA or any other relevant neurological symptoms in past 6 months [2]. However, wefound that non-focal neurological symptoms were a common indication to perform CDUS. There is no strong evidence supporting such rationale.

There are several limitations in our study. There couldbe inter-observer variability among technologists. Severity of CAD could affect degree of atherosclerosis in carotid vasculature and thereby could be a significantpredictor of cerebrovascular disease. It is likely that bruiton physical exam was under-documented or examinedonly in patients with high risk for atherosclerotic cardiovascular disease. This could explain high yield of carotidstenosis in such patients. Moreover, patient with lessthan 50% stenosis of carotid arteries could have heterogeneous unstable plaque and thus could be consideredhigh risk for stroke. High PSV in patients with previousCEA could be attributed to CEA however most patientswith sCBVD in this group had contralateral ICA stenosis. Being a single center study results need to be confirmed in a multi-center prospective randomized study in order to apply it in a much broader population. Although our study shows few risk predictors of sCBVD, itsaccuracy needs to be studied in a validation cohort. Onthe other hand, our data is different from previous studies[17,20] as it includes a cohort of patients awaiting hearttransplantation.

Conclusions

Yield of carotid ultrasound in patients less than 65 yearsage, syncope, non-focal neurological symptoms and priorto cardiac surgeries was low. Reducing its use in low riskpatients can eliminate unnecessary use of CDUS and thusoptimize efficient utilization. On the contrary, certainpatients may have higher risk of stroke despite nonsignificant carotid artery stenosis such as those with unstable carotid plaque. Hence, a case-by-case approach maybe prudent and CDUS may be needed in such scenarios. Finally, randomized prospective study and a costeffectiveness analysis can provide better risk predictionmodel for detecting significant carotid artery stenosis inboth high and low risk patients.

Consent

Since this was a retrospective chart review, all health information was de-identified in compliance with IRB andU.S. Department of Health and Human Services PrivacyRule. We obtained a waiver of HIPAA (Health InsurancePortability and Accountability Act) privacy authorization,as it was impractical to conduct research without thewaiver.

Competing interests

The authors declare that they have no competing interests.

Authors’contributions

MB conceived and designed the study and drafted the manuscript. DS wasinvolved in literature review and data interpretation. MD was responsible fordata abstraction and revision of manuscript. AJ performed statistical analysis.MC examined accuracy of carotid ultrasound reports as well as final approvalof manuscript. NW coordinated with statistician and vascular surgeons. Allauthors read and approved the final manuscript.

Acknowledgements

We appreciate the assistance of our Vascular Laboratory manager JeanetteFlannery in providing the database.

Funding sources

All funding was supported by Department of Medicine and Department ofCardiology at Newark Beth Israel Medical Center, Newark, NJ.

Received: 4 September 2014 Accepted: 19 November 2014

Published: 25 November 2014