High-risk features of carotid sonography predicting cardiovascular events among the stroke patients with atrial fibrillation

Article information

J Neurosonol Neuroimag. 2025;17(1):11-19

Publication date (electronic) : 2025 June 30

doi : https://doi.org/10.31728/jnn.2025.00169

Seung-Jae Lee, MD, PhD^,*, Tae-Kyeong Lee, MD, PhD^*, Ji Eun Moon, PhD^†

^*Department of Neurology, Soonchunhyang University Bucheon Hospital, Bucheon, Korea

^†Department of Biostatistics, Clinical Trial Center, Soonchunhyang University Bucheon Hospital, Bucheon, Korea

Correspondence: Seung-Jae Lee, MD, PhD Department of Neurology, Soonchunhyang University Bucheon Hospital, 170 Jomaru-ro, Wonmi-gu, Bucheon 14584, Korea Tel: +82-32-621-6548 Fax: +82-32-621-6950 E-mail: neurosj@catholic.ac.kr

Received 2025 March 28; Revised 2025 May 5; Accepted 2025 May 23.

Abstract

INTRODUCTION

Color Doppler duplex sonography is a widely used diagnostic tool for evaluating extracranial carotid and vertebral arterial system pathologies in clinical practice. For the carotid arteries, it enables the assessment of carotid plaque and its sonographic features, such as ulceration, irregularity, echolucency, and stenotic Doppler spectra. Regarding vertebral arteries, the V2 segment is particularly accessible, allowing B-mode imaging and Doppler spectral analysis of the arterial lumen between the transverse processes of the cervical vertebrae.1

Carotid atherosclerosis is an established independent risk factor for incident stroke in the general population2,3 and is associated with an increased prevalence of atrial fibrillation (AF).4,5 It is estimated that carotid plaques are present in approximately half of AF patients.6 Furthermore, carotid stenosis and AF frequently coexist, with approximately one in ten AF patients having carotid stenosis and vice versa.6

However, whether carotid atherosclerosis is associated with incident stroke in AF patients remains uncertain due to conflicting evidence in previous studies.7 Moreover, most studies have focused primarily on carotid intima-media thickness (cIMT) or carotid plaque (presence or score), often excluding detailed sonographic findings from both the carotid and vertebral arteries in AF patients.

This study investigated high-risk features of sonography (HRFS) of the carotid and vertebral arteries that are readily identifiable and intuitive during routine examinations. This study also aimed to evaluate the association between the HRFS and future vascular outcomes in stroke patients with AF.

SUBJECTS AND METHODS

2. Carotid and vertebral Duplex sonography and HRFS

All sonographic studies were performed by a qualified sonographer (Chae Yun Eo, Registered Vascular Technologist certified by the American Registry for Diagnostic Medical Sonography) who was blinded to the clinical or angiographic data. Using a color-coded duplex machine (Affiniti 50 G; Philips, Amsterdam, the Netherlands) equipped with a 5–12-MHz linear array transducer, the carotid and vertebral (foraminal segment) arteries were examined according to previously recommended standard methods.1

HRFS were defined as the presence of ulcerative, irregular, heterogeneous, or hypoechoic plaques; >50% steno-occlusion in the carotid artery; or abnormal Doppler spectra in the V2 vertebral artery segment. Ulcerative plaques were those with a focal defect at least 2 mm deep with a well-defined back wall and a color Doppler injection.8 Irregular plaques had a depth variation of 0.4–2 mm along their contour.8 Echogenicity was standardized against the blood (anechoic), mastoid muscle (isoechogenic), or bone (hyperechogenic cervical vertebrae).8 The diagnostic criteria for >50% steno-occlusion of the internal carotid artery included a peak systolic velocity >125 cm/s and >50% diameter reduction or near or complete occlusion (Fig. 1).1,9 Abnormal Doppler spectra of the V2 vertebral artery segment indicating steno-occlusion included high-resistance flow (resistive index ≥0.9), low-resistance flow (resistive index ≤0.5), increased flow velocity (peak systolic velocity ≥137.5 cm/s), or no flow signal (Fig. 2).10-12

Fig. 1.

High-risk features of sonography in the carotid artery. (A) Ulcerative plaque, indicated by a focal defect at least 2 mm deep with a well-defined back wall at its base (white arrow). (B) Irregular plaque, defined by a depth variation of 0.4–2 mm along its contour. (C) Hypoechoic plaque, where echogenicity is lower than that of the mastoid muscle. Echogenicity is standardized against the blood (anechoic), mastoid muscle (isoechogenic), or bone (hyperechogenic cervical vertebrae). (D) Heterogeneous plaque, characterized by mixed echogenicity. (E) >50% stenosis of the internal carotid artery, defined by a peak systolic velocity (PSV) >125 cm/s and >50% diameter reduction (PSV=332 cm/s in the case). (F) Near or complete occlusion (absence of flow in the case).

Fig. 2.

High-risk features of sonography in the vertebral artery. (A) High-resistance flow (resistive index ≥0.9). (B) Low-resistance flow (resistive index ≤0.5). (C) Peak systolic velocity ≥137.5 cm/s. (D) No flow signal.

cIMT was measured in a longitudinal view strictly perpendicular to the ultrasound beam, visualizing the arterial lumen of the common carotid artery (CCA) at least 5 mm below its end.13 cIMT was automatically computed along a 10 mm length of the far wall using Q-Lab (Philips Medical Systems).14 Values from both the CCAs were averaged to express the mean cIMT.

The carotid plaque score (CPS) was measured by dividing the carotid artery into four segments bilaterally: the CCA, carotid bifurcation, internal carotid artery, and external carotid artery. The thickness of the largest plaque in each segment was measured, with ≥1.5, ≥2.5, and ≥ 3.5 mm given 1, 2, and 3 points, respectively. Points from all segments on both sides were summarized to obtain a total CPS ranging from 0 to 24.15

RESULTS

We enrolled 207 participants, including 104 women and 103 men (mean age±standard deviation, 74.2±10.6 years). Table 1 presents the basic characteristics of the study participants. Univariate and multivariate Cox proportional hazards regression analyses were performed to identify the variables associated with the outcomes.

Table 1.

Basic characteristics of the 207 study participants

For vascular events, age, CHA2DS2-VASc score, CPS, and HRFS were the univariate predictors. Adjusted analysis using these factors revealed that HRFS (aHR 4.86, 95% CI 1.41–16.77) was the only independent predictor of vascular events (Table 2).

Table 2.

Univariate and multivariate predictors of vascular events

Age, current smoking status (negative association), chronic kidney disease, vascular disease, NIHSS score, CHA2DS2-VASc score, CPS, mean cIMT, and HRFS were the univariate predictors of MACEs. Multivariate analysis using the univariate predictors demonstrated that HRFS (aHR 2.45, 95% CI 1.18–5.08), age (aHR 1.06, 95% CI 1.02–1.11), chronic kidney disease (aHR 2.03, 95% CI 1.00–4.12), and NIHSS score (aHR 1.04, 95% CI 1.00–1.08) were independently associated with MACEs (Table 3).

Table 3.

Univariate and multivariate predictors of MACEs

Kaplan–Meier curves showed that the occurrence of vascular events (p<0.001 by log-rank test) and MACEs (p<0.001 by log-rank test) was significantly more frequent in patients with HRFS than in those without HRFS (Fig. 3). Table 4 presents the discriminatory abilities of the outcomes. Among HRFS, CHA2DS2-VASc score, CPS, and mean cIMT, HRFS had the highest iAUC. HRFS demonstrated a significantly better discrimination ability than the mean cIMT in predicting outcomes. However, there was no significant difference in the discrimination ability between HRFS and the CHA2DS2-VASc score or CPS for vascular events and MACEs. Fig. 4 shows the time-dependent AUC.

Fig. 3.

Kaplan–Meier curves for vascular events (A) and MACEs (B). Outcomes were significantly more frequent in patients with HFRS than in those without HRFS (p<0.001 by log-rank test). MACEs, major adverse cardiovascular events; HRFS, high-risk features of sonography. 1: HRFS (-); 2: HRFS (+).

Table 4.

Discrimination ability for outcomes

Fig. 4.

Time-dependent receiver operating characteristic analysis for vascular events (A) and MACEs (B). HRFS, high-risk features on sonography; MACEs, major adverse cardiovascular events; cIMT, carotid intima-media thickness.

DISCUSSION

In our study, we demonstrated the following: (1) HRFS independently predicted future adverse vascular outcomes in stroke patients with AF; (2) while the CHA2DS2-VASc score, cIMT, and CPS were the univariate predictors of vascular outcomes, their associations were no longer significant after adjustment in the analysis; and (3) HRFS exhibited significantly better discriminatory ability than cIMT. Carotid atherosclerosis is closely linked to the extent of atherosclerosis in other vascular territories, including the coronary and peripheral arteries,17 and is associated with cardiovascular events.2,3 Additionally, it is correlated with the risk of incident stroke in the general population.2,3 As a marker of global atherosclerotic burden, its easy accessibility through non-invasive imaging provides valuable insights into the overall state of the circulatory system. However, in AF patients, the association between carotid atherosclerosis, as measured by cIMT, plaque presence, or stenosis, and stroke risk remains uncertain, with conflicting evidence in the literature.

In the Atherosclerosis Risk in Communities Study, increased cIMT (per 1 SD; aHR 1.23; 95% CI 1.04–1.46) and the presence of carotid plaque (aHR 1.56; 95% CI 1.00–2.45) were both associated with an elevated risk of ischemic stroke in 724 participants with AF over a mean follow-up of 8.5 years.18 Similarly, a retrospective single-center study identified the presence of carotid plaque as a risk factor for ischemic stroke (aHR 3.75; 95% CI 1.11–12.69) in 310 AF patients.19 Furthermore, a multicenter retrospective registry-based study demonstrated that carotid artery stenosis of 50% or more independently predicted stroke recurrence (aHR 2.02; 95% CI 1.37–3.01) in 899 stroke patients with AF.20 Additionally, the ARAPACIS study found that carotid plaque was more prevalent among patients who experienced stroke or transient ischemic attack (31.6% vs. 16.1%, p=0.002) in a cohort of 2027 AF patients.21

Conversely, several studies have reported no association between carotid atherosclerosis and incident stroke in AF patients with higher use of anticoagulants, suggesting that oral anticoagulants may mitigate stroke risk associated with carotid disease.7 For example, a post-hoc analysis of the SPAF II study revealed that carotid stenosis of 50% or more was not significantly associated with stroke risk (aRR 1.3; 95% CI 0.5–3.6).22 Similarly, an analysis of the ROCKET AF trial cohorts showed comparable rates of stroke or systemic embolism in patients with and without carotid artery disease (aHR 0.99; 95% CI 0.66–1.48).23 Furthermore, a prospective multicenter study of 587 AF patients referred to an anticoagulation clinic found no significant association between carotid atherosclerosis (aHR 1.19; 95% CI 0.52–2.72) or carotid stenosis of 50% or more (aHR 1.03; 95% CI 0.30–3.45) and the occurrence of stroke or TIA.24

Our research offers distinct insights compared with these studies. Previous investigations have primarily focused on pathological markers of the carotid artery, such as cIMT, carotid plaque, or stenosis. In contrast, our study analyzed sonographic findings from both the vertebral and carotid arteries, integrating the features of unstable plaques into HRFS. Additionally, in our study population, cIMT and CPS were not independently associated with vascular outcomes, which is likely attributable to the high prevalence of oral anticoagulant use at discharge (99.0% [205/207]). However, HRFS independently predicted vascular outcomes (aHR 4.86; 95% CI 1.41–16.77), indicating that oral anticoagulants may not fully mitigate the vascular risk associated with HRFS in stroke patients. These findings highlight HRFS as a potentially more vulnerable and robust marker of vascular outcomes than previously studied indicators of carotid atherosclerosis in AF-related stroke populations. Our study had several limitations. First, the findings may not be generalizable to other populations because of the study’s reliance on single-center data and its relatively small sample size. Notably, 69 of 279 AF patients were excluded from the analysis because of the lack of sonographic data, as many of these patients were unable to undergo sonographic evaluation because they required intensive care for stroke. Consequently, our results may not apply to patients in intensive care settings. Further large-scale studies are required to validate these findings. Second, causal relationships between the variables could not be established because of the observational study design. Furthermore, features such as unstable carotid plaques, carotid steno-occlusion, and abnormal Doppler findings of the vertebral artery could not be analyzed separately because of the small sample size in each category, limiting the statistical power. Nevertheless, the concept of an integrated HRFS offers a practical and intuitive framework for the clinical prognosis of stroke patients with AF.

In conclusion, our study highlights that HRFS serves as a more effective predictor of future vascular outcomes than traditional markers, such as cIMT and carotid plaque (presence or score), in AF patients. These results underscore the value of routine ultrasonographic examinations in stroke patients, as HRFS findings can be easily identified during these assessments. Consequently, our findings enhance the role of sonography in improving vascular risk prediction and optimizing the management of stroke patients with AF.

Article information Continued