Association of Subclavian Steal Phenomenon with Prevalence of Contralateral Vertebral Artery Atherosclerotic Stenosis: A Hospital-Based Cohort Study

Abstract

Aims: It is uncertain if there is a connection between subclavian steal phenomenon (SSP) and atherosclerotic stenosis in the opposite vertebral artery (VA). We aimed to explore the association between SSP and the incidence of contralateral vertebral artery stenosis (VAS) in vivo.

Methods: In this prospective registry study, we included patients diagnosed with ＞50% stenosis of proximal subclavian artery (SA) or innominate artery (INA) by digital subtraction angiography (DSA) from our comprehensive stroke center between 2011 and 2022. VAS and SSP was diagnosed by DSA in the resting state. Propensity score matching (PSM) was conducted among all participants and subgroups with a 1:1 ratio according to the presence of SSP. We further conducted sensitivity analysis by dividing all participants into subgroups according to the degree of stenosis and type of SSP. Binomial logistic regression analysis was applied to investigate the association of SSP with contralateral VAS.

Results: A total of 774 patients were included in this study and 309 (39.9%) were found with SSP. After PSM, presence of SSP was associated with lower prevalence of contralateral VAS among all participants (OR 0.45; 95% CI 0.31−0.65; p＜0.001). In subgroup analysis, the association was respectively found within left subclavian (LSA) stenosis group (OR 0.43; 95% CI 0.29−0.65; P＜0.001) and right subclavian artery (RSA) / INA stenosis group (OR 0.36; 95% CI 0.19−0.69; P=0.002).

Conclusions: SSP is associated with lower prevalence of contralateral VAS.

Li He and Muke Zhou are joint senior authors.

Introduction

The subclavian steal phenomenon (SSP) is usually characterized by retrograde blood flow (via collateral pathways commonly originating from the contralateral vertebral artery as well as the circle of Willis) within the ipsilateral vertebral artery (VA) to feed the upper limb attributed to hemodynamically significant lesion or obstruction present in the proximal subclavian artery (SA) and the innominate artery (INA), which is usually caused by atherosclerotic stenosis or occlusion^{1, 2)}. SSP often performs with alterations in the hemodynamic patterns within collateral circulation from the contralateral vertebral artery (e.g., blood flow velocity and volume)^{3, 4)}, which are correlated with wall shear stress (WSS) changes^5-7). Emerging evidence from case reports has shown that the formation and rupture of vertebrobasilar junction aneurysm might be associated with abnormal hemodynamic stress secondary to increased blood flow through VA induced by SSP^{8, 9)}. Furthermore, whether SSP is associated with atherosclerotic stenosis development of contralateral VA deserves further exploration.

The present study aimed to explore whether SSP was associated with the incidence of contralateral vertebral artery stenosis (VAS). We hypothesized that proximal subclavian artery stenosis (SAS) or innominate artery stenosis (INAS) with ipsilateral SSP was associated with the incidence of contralateral VAS.

Material and Methods

Study Design

The dataset utilized in this study was available from the secure DSA registry of our comprehensive stroke center at West China Hospital, Sichuan University, China. All patients who underwent DSA observed the following criteria: (1) suspected carotid or/and cerebral arterial lesions by other non-invasive examinations, such as arterial stenosis or occlusion, aneurysm, moyamoya disease, vascular malformation, etc.; (2) acute cerebrovascular disease requiring arterial thrombolysis or other endovascular treatments; (3) suspected cerebral venous lesions, such as venous stenosis, venous thrombosis, arteriovenous malformation; (4) examination before implementing vascular intervention surgery treatment; (5) examination of the etiology of intracerebral or subarachnoid hemorrhage. Data was prospectively collected between January 2011 and December 2022. Consecutively, all patients who received DSA while being hospitalized were duly registered. Comprehensive information acquired from all participants (e.g., patients’ demographics, risk factors, diagnostic test outcomes, neuroimaging findings, administration of treatments, and any complications during hospitalization) can be accessible in the secure database.

Participants

Participants who underwent DSA diagnosed with proximal SAS and INAS (stenosis degree ＞50%) were included in the study. Exclusion criteria at baseline included: (1) missed clinical information; (2) presented with ipsilateral VAS ＞30%; (3) with basilar artery stenosis ＞30%; (4) with bilateral subclavian steal phenomenon or bilateral SAS (or both LSA and INA stenosis) ＞50%; (5) with vertebral artery hypoplasia (VAH); (6) with fetal type posterior cerebral artery; (7) SAS or VAS due to other causes (e.g., arterial dissection, thrombosis, Takayasu’s and other forms of arteritis, cerebral arteriovenous malformation, fibromuscular dysplasia and mechanical causes); (8) SSP due to other etiologies of non-luminal stenotic lesion (e.g., arteriovenous fistula).

Imaging Protocol and Diagnosis of SAS, INAS, VAS and SSP

DSA is the method of choice to diagnose proximal SAS, INAS, VAS and SSP. The consultant neuroradiologists reviewed all DSA images to identify the presence of SSP and atherosclerotic stenosis of SA, INA and VA, and stenosis degree was determined using the WASID method¹⁰⁾. During the imaging process, the diagnosis of SSP was in the resting state based on the following criteria: (1) evidence of significant proximal SAS or INAS (stenosis degree ＞50%); (2) patency of the ipsilateral vertebral artery and the basilar artery (stenosis degree ＜30%); (3) angiographic flow characteristics within the ipsilateral VA: complete SSP is defined as SAS/INAS with permanent reversal flow in ipsilateral VA; latent SSP is defined as SAS/INAS with delayed opacification sign or to-and-fro flow in ipsilateral VA^11-13). Definition of VAH was based on the criteria that diameter of VA is less than 2mm in the V4 segment and with radiographic signs of slimness or absence of the whole VA on DSA, or a concomitant diameter asymmetry ratio of ≤ 1:1.7 in all of the 4 vertebral segments (V1–V4)¹⁴⁾. VAS was defined as vertebral stenosis of at least 30% resulting from presumed atheromatous disease determined by DSA¹⁵⁾.

Statistical Analysis

Data analysis and management included in present research were conducted with R software (version 4.0.2). The study population were further classified into two subgroups: left subclavian artery (LSA) stenosis group and right subclavian artery (RSA) / INA stenosis group. Firstly, we analyzed group characteristics and differences about baseline demographic and clinical features among all participants and subgroups according to the presence of SSP. Independent t tests were used to compare continuous variables (expressed as means and SDs) with normal distributions, while Mann-Whitney U tests were used for variables with skewed distributions. At the same time, Chi-square tests were employed to compare categorical variables (represented by frequencies and percentages). To balance the baselines of both groups and minimize potential confounding factors, a 1:1 propensity score matching (PSM) ratio was conducted among all participants and specific subgroups (i.e., LSA stenosis group and RSA / INA stenosis group) according to the presence of SSP with the method of greedy nearest-neighbor algorithm. We computed the propensity score based on age and sex. Binomial logistic regression models were performed in univariate and multivariate analyses to identify the association of SSP with the incidence of contralateral VAS. Finally, we further conducted sensitivity analysis according to the stenosis degree and grade of SSP. The findings were presented in the form of odds ratios (ORs) along with 95% confidence intervals (CIs). Statistical significance was defined as p＜0.05 when two-sided tests.

Ethics Approval and Consent to Participate

The approval for this research protocol was granted by the ethics committee of the West China Hospital, Sichuan University (2023(619)). All patients or their families gave written consent for study procedure.

Data Availability

Data not provided in the article because of space limitations may be shared (anonymized) at the request of any qualified investigator for purposes of replicating procedures and results.

Results

Participants Demographics before Matching

The flow chart for participants included in this study is shown in Fig.1. A total of 10144 patients (age ＞18 years) were consecutively recruited between January 2011 and December 2023 in our database. There were 812 (8.0%) patients present with LSAS, 288 (2.8%) patients with RSAS and 86 (0.8%) patients with INAS, respectively. After excluding patients according to exclusion criteria, a total of 774 participants diagnosed with proximal SAS and INAS were included in our analysis, there were 546 (70.5%) participants presented with LSA stenosis. On the right lateral, there were 207 (26.7%) patients present with RSA stenosis, and 21 (2.7%) individuals diagnosed with INA stenosis. Among 774 individuals diagnosed with SAS and INAS, 309 (39.9%) were present with SSP. In participants with LSA stenosis, the prevalence of SSP was 41.0% (224 / 546). Among patients with RSA and INA stenosis, the prevalence of SSP was 39.1% (81 / 207) and 19.0% (4 / 21), respectively. Among 309 participants diagnosed with SSP, there were 102 (33.0%) patients diagnosed with neurologic symptoms.

Fig.1. Flow chart of participants recruitment

Abbreviation: SAS, subclavian artery stenosis; INAS, innominate artery stenosis; DSA, digital subtraction angiography; VAS, vertebral artery stenosis; BAS, basilar artery stenosis; SSP, subclavian steal phenomenon; VAH, vertebral artery hypoplasia; FTP, fetal type posterior cerebral artery; LSA, left subclavian artery; INA, innominate artery, RSA, right subclavian artery.

Comparations of clinical characteristics between participants with and without SSP were presented in Table1 before PSM. Among all participants, individuals with SSP had lower prevalence of contralateral VAS (30.1% vs 49.7%, p＜0.001) in comparison to those without SSP before PSM. The same effects were respectively found in LSA stenosis group (30.4% vs 51.6%, p＜0.001) and RSA / INA stenosis group (29.4% vs 45.5%, p=0.017) (Table 1).

Table 1.Baseline characteristics of all participants before PSM

Characteristics	All participants (n=774)			LSA stenosis (n=546)			RSA / INA stenosis (n=228)
	With SSP (N=309)	Without SSP (N=465)	P	With SSP (N=224)	Without SSP (N=322)	P	With SSP (N=85)	Without SSP (N=143)	P
Age (y)	65.2±9.3	67.3±9.5	0.003	64.6±9.8	66.8±9.7	0.011	66.8±7.7	68.4±9.0	0.167
Male	244 (79.0)	367 (78.9)	0.989	176 (78.6)	250 (77.6)	0.796	68 (80.0)	117 (81.8)	0.734
BMI (kg/m2)	23.8±2.9	23.7±3.0	0.890	23.8±2.9	23.7±3.3	0.905	23.8±2.7	23.7±2.4	0.851
Smoking	153 (49.6)	197 (42.4)	0.198	109 (48.8)	144 (44.8)	0.560	43 (51.2)	53 (37.1)	0.147
Hypertension	181 (58.6)	318 (68.4)	0.005	134 (59.8)	216 (67.1)	0.082	47 (55.3)	102 (71.3)	0.014
Diabetes mellitus	110 (35.6)	204 (43.9)	0.022	82 (36.6)	152 (47.2)	0.014	28 (32.9)	52 (36.4)	0.601
Hyperlipidemia	59 (19.1)	81 (17.4)	0.553	48 (21.4)	54 (16.8)	0.170	11 (12.9)	27 (18.9)	0.245
Coronary heart disease	40 (12.9)	83 (17.9)	0.068	29 (13.0)	64 (19.9)	0.034	11 (12.9)	19 (13.3)	0.941
Contralateral VAS	93 (30.1)	231 (49.7)	＜0.001	68 (30.4)	166 (51.6)	＜0.001	25 (29.4)	65 (45.5)	0.017
＜50%	29 (9.4)	81 (17.4)	0.010	23 (10.3)	54 (16.8)	0.034	6 (7.1)	27 (18.9)	0.019
50-69%	43 (13.9)	115 (24.7)	0.001	34 (15.2)	83 (25.8)	0.003	9 (10.6)	32 (22.4)	0.032
70-99%	21 (6.8)	35 (7.5)	0.102	11 (4.9)	29 (9.0)	0.094	10 (11.7)	6 (7.7)	0.057

Abbreviations: PSM, propensity score matching; LSA, left subclavian artery; RSA, right subclavian artery; INA, innominate artery; SSP, subclavian steal phenomenon; BMI, body mass index; VAS, vertebral artery stenosis. Data given as mean±SD or n (%).

After Propensity Score Matching

Demographic Information

Group characteristics and comparisons after PSM were shown in Table 2. The 1:1 ratio PSM were conducted to 465 participants without SSP and 309 participants with SSP. Next, we further applied PSM for individuals with and without SSP in LSA stenosis group and RSA / INA stenosis group, respectively (Fig.1).

Table 2.Baseline characteristics of participants among LSA stenosis and RSA / INA stenosis group after PSM

Characteristics	LSA stenosis (N= 412)			RSA / INA stenosis (N= 156)
	With SSP (N= 206)	Without SSP (N= 206)	P	With SSP (N= 78)	Without SSP (N= 78)	P
Age (y)	66.0±8.9	65.9±9.1	0.952	67.6±7.4	67.6±7.5	0.983
Male	161 (78.2)	157 (76.2)	0.639	62 (79.5)	64 (82.1)	0.685
BMI (kg/m2)	23.8±2.9	23.7±3.3	0.905	23.9±2.7	23.5±2.4	0.416
Smoking	100 (48.7)	94 (45.8)	0.710	38 (48.7)	28 (35.9)	0.260
Hypertension	127 (61.7)	137 (66.5)	0.304	44 (56.4)	57 (73.1)	0.029
Diabetes mellitus	76 (36.9)	101 (49.0)	0.013	24 (30.8)	28 (35.9)	0.497
Hyperlipidemia	41 (19.9)	37 (18.0)	0.615	9 (11.5)	14 (18.0)	0.259
Contralateral VAS	65 (31.6)	106 (51.5)	＜0.001	23 (29.5)	42 (53.9)	0.002

Abbreviations: PSM, propensity score matching; LSA, left subclavian artery; RSA, right subclavian artery; INA, innominate artery; SSP, subclavian steal phenomenon; BMI, body mass index; VAS, vertebral artery stenosis. Data given as mean±SD or n (%).

Relationship between SSP and Contralateral VAS

After PSM, contralateral VAS was less observed in patients with SSP than those without (30.5% vs 54.5%, p＜0.001) among all participants. In subgroup analysis, the same results were respectively observed in LSA stenosis group (31.6% vs 51.5%, p＜0.001) and RSA/INA stenosis group (29.5% vs 53.9%, p=0.002) (Table 2).

Univariate logistic regression among all participants showed SSP was associated with lower frequency of contralateral VAS (OR, 0.45 [95% CI, 0.31−0.65]; p＜0.001). In the multivariate model adjusted for hypertension, diabetes, hyperlipidemia, smoking and BMI, SSP was independently associated with contralateral VAS (OR, 0.49 [95% CI, 0.26−0.93]; p=0.028) (Table 3).

Table 3.Association between SSP and prevalence of contralateral VAS among all participants after matching

Variable	Univariate analysis		Multivariate analysis
	OR (95% CI)	P	OR (95% CI)	P
BMI	0.98 (0.91-1.07)	0.714	0.95 (0.85-1.06)	0.332
Smoking	2.10 (1.20-3.67)	0.010＊	2.44 (1.31-4.55)	0.005＊
Hypertension	1.92 (1.30-2.82)	＜0.001＊	2.64 (1.39-5.01)	0.003＊
Diabetes mellitus	0.91 (0.63-1.33)	0.637	0.59 (0.31-1.10)	0.097
Hyperlipidemia	0.55 (0.33-0.92)	0.024＊	0.74 (0.32-1.72)	0.491
With SSP	0.45 (0.31-0.65)	＜0.001＊	0.49 (0.26-0.93)	0.028＊

Abbreviations: SSP, subclavian steal phenomenon; VAS, vertebral artery stenosis; OR, odd ratio; CI, confidence level; BMI, body mass index.

^＊ Statistical significance.

In the LSA stenosis group, univariate logistic regression revealed that left SSP was associated with lower prevalence of contralateral VAS (OR, 0.43 [95%CI, 0.29−0.65]; p＜0.001). In the multivariate regression model adjusted for hypertension, diabetes and hyperlipidemia, Smoking and BMI, left SSP was still independently associated with prevalence of contralateral VAS (OR, 0.40 [95%CI, 0.19−0.81]; p=0.011) (Table 4).

Table 4.Association between SSP and prevalence of contralateral VAS among LSA stenosis and RSA / INA stenosis group respectively after matching

Variable	LSA stenosis				RSA / INA stenosis
	Univariate analysis		Multivariate analysis		Univariate analysis		Multivariate analysis
	OR (95% CI)	P	OR (95% CI)	P	OR (95% CI)	P	OR (95% CI)	P
BMI	1.04 (0.95-1.14)	0.373	1.02 (0.91-1.14)	0.751	0.91 (0.77-1.07)	0.234	0.74 (0.57-0.96)	0.021＊
Smoking	2.97 (1.58-5.59)	＜0.001＊	3.12 (1.56-6.25)	0.001＊	1.80 (0.72-4.53)	0.212	3.47 (0.99-6.52)	0.052
Hypertension	2.09 (1.37-3.20)	＜0.001＊	2.02 (0.98-4.17)	0.056	3.96 (1.87-8.37)	＜0.001＊	3.52 (1.91-5.13)	0.004＊
Diabetes mellitus	0.83 (0.56-1.24)	0.367	0.71 (0.35-1.43)	0.333	1.32 (0.67-2.58)	0.422	0.33 (0.10-1.13)	0.079
Hyperlipidemia	0.80 (0.48-1.33)	0.390	0.99 (0.42-2.33)	0.986	1.65 (0.68-4.00)	0.271	3.46 (0.73-6.46)	0.118
With SSP	0.43 (0.29-0.65)	＜0.001＊	0.40 (0.19-0.81)	0.011＊	0.36 (0.19-0.69)	0.002＊	0.26 (0.08-0.84)	0.025＊

Abbreviations: SSP, subclavian steal phenomenon; VAS, vertebral artery stenosis; LSA, left subclavian artery; RSA, right subclavian artery; INA, innominate artery; OR, odd ratio; CI, confidence level; BMI, body mass index.

^＊Statistical significance.

In the RSA/INA stenosis group, univariate logistic regression showed that right SSP was associated with incidence of contralateral VAS (OR, 0.36 [95%CI, 0.19−0.69]; p=0.002). The independent association between SSP and prevalence of contralateral VAS still exist (OR, 0.26 [95%CI, 0.08−0.84]; p=0.025) after adjusting for hypertension, diabetes and hyperlipidemia, smoking, and BMI (Table 4).

Finally, we further conducted sensitivity analysis by dividing all participants into subgroups based on the degree of stenosis and type of SSP. Group characteristics and the association between SSP and contralateral VAS were shown respectively (Table 5 and Table 6).

Table 5.Baseline characteristics of participants in different group according to stenosis degree

Variables	LSA stenosis (n= 546)		P	RSA/INA stenosis (n= 228)		P
	50-69% (n= 142)	70-99% (n= 404)		50-69% (n= 77)	70-99% (n= 151)
Age (y)	65.6±10.3	66.0±9.7	0.658	65.6±10.3	66.0±9.7	0.658
Sex, male	117 (82.4)	309 (76.5)	0.252	65 (84.4)	120 (79.5)	0.619
BMI (kg/m2)	23.0±2.8	23.9±3.2	0.078	23.4±2.4	23.9±2.6	0.540
Smoking	73 (51.2)	182 (45.1)	0.661	36 (47.2)	60 (40.0)	0.570
Hypertension	93 (65.5)	257 (63.6)	0.619	57 (74.0)	92 (60.9)	0.120
Diabetes mellitus	67 (47.2)	167 (41.3)	0.034	30 (39.0)	50 (33.1)	0.557
Hyperlipidemia	27 (19.0)	75 (18.6)	0.911	11 (14.3)	27 (17.9)	0.698
Coronary heart disease	28 (19.7)	65 (16.1)	0.333	11 (14.3)	19 (12.6)	0.606
With SSP	54 (38.0)	170 (42.1)	＜0.001	20 (26.0)	65 (43.0)	＜0.001
With latent SSP	54 (38.0)	32 (7.9)	＜0.001	20 (26.0)	26 (17.2)	＜0.001
With complete SSP	0 (0.00)	138 (34.2)	＜0.001	0 (0.00)	39 (25.8)	＜0.001

Abbreviations: LSA, left subclavian artery; RSA, right subclavian artery; INA, innominate artery; BMI, body mass index; SSP, subclavian steal phenomenon. Data given as mean±SD or n (%).

Table 6.Association between different grade of SSP and prevalence of contralateral VAS among LSA stenosis and RSA / INA stenosis group respectively

		Contralateral VAS (stenosis degree)
		＜50%		50-69%		70-99%
		OR (95% CI)	P	OR (95% CI)	P	OR (95% CI)	P
LSA stenosis
50-69%	With latent SSP	0.14 (0.03-0.63)	0.010＊	0.17 (0.05-0.58)	0.005＊	2.22 (0.91-5.39)	0.078
	With complete SSP	-	-	-	-	-	-
70-99%	With latent SSP	0.21 (0.05-0.89)	0.035＊	0.23 (0.05-0.98)	0.048＊	0.37 (0.08-1.60)	0.182
	With complete SSP	0.34 (0.13-0.93)	0.022＊	0.46 (0.22-0.96)	0.039＊	0.45 (0.25-0.81)	0.037＊
RSA/INA stenosis
50-69%	With latent SSP	0.37 (0.25-0.78)	0.027＊	0.78 (0.45-0.96)	0.039＊	0.51 (0.15-2.77)	0.233
	With complete SSP	-	-	-	-	-	-
70-99%	With latent SSP	0.65 (0.15-1.55)	0.561	0.79 (0.35-3.76)	0.225	0.21(0.01-1.19)	0.457
	With complete SSP	0.47 (0.21-1.78)	0.322	0.19 (0.06-0.69)	0.047＊	0.43 (0.25-0.98)	0.050＊

adjustment for age, sex, body mass index, smoking, hypertension, diabetes and hyperlipidemia.

Abbreviations: SSP, subclavian steal phenomenon; VAS, vertebral artery stenosis; LSA, left subclavian artery; RSA, right subclavian artery; INA, innominate artery; OR, odd ratio; CI, confidence level. ^＊ Statistical significance.

Discussion

We investigated the association between SSP and prevalence of contralateral VAS using DSA at rest state in this study and found that SSP is associated with lower prevalence of contralateral VAS.

In present study, the SSP was found to be more prevalent in the LSA, which is in accordance with previous study¹⁶⁾. According to previous literature, LSA is more susceptible to atherosclerosis than the right one in patients with SSP¹⁷⁾. At the origin of LSA, the angle between LSA and aortic is usually more acute, causing turbulence flow and a higher likelihood of atherogenesis. This has been suggested as a possible explanation for this finding¹⁸⁾. The RSA typically arises from the INA and scarcely from the arch. The anatomical foundation has the potential to alter the hemodynamic connection.

The potential mechanism linking SSP to the incidence of contralateral VAS is incompletely explicit. The speculations and illustration of potential mechanism about the effects of SSP on the contralateral VAS is shown in Fig.2. Collateral circulation in SSP plays an important role in the compensation of upper limb ischemia. The establishment of collateral circulation in turn replaces the original blood flow from contralateral VA and causes a prolonged alteration in flow pattern in this region^{3, 4)}.

Fig.2. Illustration of potential mechanism about the effects of SSP on the development of contralateral VAS

The left diagram shows SAS without SSP (A), and when SSP happens, the blood flow velocity and volume increased in the contralateral vertebral artery (B). According to the simplified equation, increased blood flow velocity is associated with high WSS. The possible biological factors associated with anti-atherosclerotic role of high WSS are shown on the right. Abbreviation: SSP, subclavian steal phenomenon; SAS, subclavian artery stenosis; VAS, vertebral artery stenosis; WSS, wall shear stress; µ, viscosity; v, velocity; r, arterial radius; MMPs, matrix metalloproteinases; VCAM-1, vascular cell adhesion molecule-1; SMCs, smooth muscle cells; ATP, adenosine triphosphate; NO, Nitric Oxide.

The occurrence and progression of atherosclerosis is directly linked to disruption in blood flow which plays a significant role in the biomechanics of atherosclerosis^{5, 7, 19)}. Based on the findings of previous studies, it has been indicated that changes in wall shear stress (WSS) are a crucial factor in the regulation of endothelial function and the development of atherosclerosis^{20, 21)}. The alteration in WSS depends on factors such as blood flow rate, arterial diameter, and blood viscosity²²⁾. Changes in blood flow velocity or volume which are commonly observed on the contralateral VA among patients with SSP can influence WSS and decide the trend of WSS alteration, which would contribute to our understanding about the effects of SSP on the development of contralateral VAS. Furthermore, according to the Bernoulli equation, there exists an inverse relationship between the potential energy and the kinetic energy within a hemodynamic system²³⁾. For participants with SSP, increase blood flow velocity present in the contralateral VA means higher kinetic energy, which correspondingly leads to the decrease of pressure (potential energy). The mechanism could potentially explain the relationship between SSP and prevalence of contralateral VAS. However, this assumption still awaits further conclusion through computational fluid dynamics and requires additional validation of vessel wall imaging.

It is noteworthy that radiographic signs indicative of SSP can be observed even in those without SAS or INAS, typically in individuals with VAH (with or without concomitant FTP) and significant proximal stenosis of the VA^{24, 25)}. Patients with VAH or VAS (stenosis degree ＞30%) ipsilateral to SAS or INAS were excluded in this research to reduce misdiagnosis and control confounds²⁴⁾. Retrograde blood flow in the VA which is ipsilateral to the stenosis or obstruction of the proximal SA or INA may have different types of collateral pathways in the process of whole cardiac cycle^{26, 27)}. Among all observed collateral circulation present in SSP, the reversed blood flow most commonly arises from the contralateral VA²⁶⁾. The observed relationship between SSP and contralateral VAS confirmed the important role of the collateral pathway. When the compensatory effect of the contralateral VA is weakened by the lesion of stenosis or occlusion, collateral circulation between the carotid artery and the basilar artery can be able to further develop, resulting in reversal basilar artery blood flow²⁸⁾. In present study, to control the confounds between the incidence of contralateral VAS and SSP, we excluded patients diagnosed with basilar artery stenosis (stenosis degree ＞30%), VAH and fetal type posterior cerebral artery to keep the steal pathway patency.

There are certain strengths in our study. Firstly, to our best knowledge, few studies have been conducted to analyze the independent effect of SSP on the development of contralateral VAS in a cross-sectional or in a longitudinal fashion, and SSP performs as a compensatory state replacing original blood flow pattern, which is an optimal choice to explore the effects of a long-term alteration in hemodynamic patterns (e.g., increased blood flow velocity and volume) on development of atherosclerosis in vivo. Secondly, the data used in this analysis were prospectively collected and participants included in this study were based on DSA which is still the widely accepted standard for detecting artery stenosis accurately. Finally, it is appropriate and suitable to conduct PSM design for a hospital-based cohort to minimize confounding bias for exploring the association of SSP with the development of contralateral VAS.

Despite above strengths, this study had certain limitations of systematic selection bias and cross-section nature. First, we conducted the analysis at a single-center. It should be noted that our study cohort may not be representative of the worldwide general population, because there exist racial differences for cerebral artery atherosclerosis^{29, 30)}. Therefore, further longitudinal studies are needed to identify and determine the causal relationship between SSP and development of contralateral VAS. Second, While DSA is the preferred diagnostic method for accurately identifying artery stenosis and evaluating potential collateral patterns associated with SSP, it has poor sensitivity for the detection of incomplete blood flow reversal in comparison to Doppler sonography^{13, 31)}, which is not combined in this study. Third, according to previous study, it has been documented that individuals who are diagnosed with proximal SAS (stenosis degree ＞50%) often have a permanent reversal blood flow^{2, 4)}. Therefore, to minimize missed diagnosis and misdiagnosis, we only included participants diagnosed with proximal SAS (stenosis degree ＞50%) to improve the sensitivity and reliability of diagnoses of SSP by DSA in this study, which may cause selection bias of study population. Fourth, since we did not directly measure the hemodynamic patterns in the contralateral VA, the alteration levels (i.e., blood flow velocity and other hemodynamic parameters) in cases might be different, we included cases diagnosed with SAS (stenosis degree＞50%) and conducted PSM to minimize the difference and bias. Finally, grading of SSP was just based on angiographic flow characteristics and the physical patterns of blood flow (e.g., WSS distribution, oscillatory shear index, and wall pressure) were not analyzed in in present study. Further studies incorporating these factors could potentially contribute to our understanding of this phenomenon, which may further improve the performance of this model for exploring the development of atherosclerosis.

Conclusion

Our findings indicate that the pathological process of SSP is associated with lower prevalence of contralateral VAS. Meticulous attention to the pathophysiological process of SSP might help explore the effect of hemodynamic changes on development of atherosclerosis in vivo.

Conflict of Interest

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

This study is supported by grant from the China National Key Research and Development Project, 2021YFC2500506.

Author Contribution

Dr. Zhou: conceptualization and design of the study, revision of the manuscript. Zhao Zhang: collection and analysis of the data, drafting and revision of the manuscript, and prepared all the figures. Anling Luo, Yujia Yang, Xuzi Li, Yiting Deng: collection of the data. Dr He: revision of the manuscript.

Acknowledgements

We would like to express our appreciation to Dr. Hongbo Zheng, Dr. Fayun Hu, Dr. Lizhang Chen and Dr. Jian Wang for their willingness to share their expertise and resources.

References

• 1)  Reivich M, Holling HE, Roberts B, Toole JF. Reversal of blood flow through the vertebral artery and its effect on cerebral circulation. N Engl J Med, 1961; 265: 878-885
• 2)  Potter BJ, Pinto DS. Subclavian steal syndrome. Circulation, 2014; 129: 2320-2323
• 3)  Tan TY, Schminke U, Chen TY. Hemodynamic effects of subclavian steal phenomenon on contralateral vertebral artery. J Clin Ultrasound, 2006; 34: 77-81
• 4)  Harper C, Cardullo PA, Weyman AK, Patterson RB. Transcranial Doppler ultrasonography of the basilar artery in patients with retrograde vertebral artery flow. J Vasc Surg, 2008; 48: 859-864
• 5)  J M, L N, A V, E C, V L, K BS, L L, T G, M P, P M. Wall Shear Stress Alteration: a Local Risk Factor of Atherosclerosis. Curr Atheroscler Rep, 2022; 24: 143-151
• 6)  Aizawa K, Ramalli A, Sbragi S, Tortoli P, Casanova F, Morizzo C, Thorn CE, Shore AC, Gates PE, Palombo C. Arterial wall shear rate response to reactive hyperaemia is markedly different between young and older humans. J Physiol, 2019; 597: 4151-4163
• 7)  Malek AM, Alper SL, Izumo S. Hemodynamic shear stress and its role in atherosclerosis. JAMA, 1999; 282: 2035-2042
• 8)  Okuhara T, Hashimoto K, Kanemaru K, Yoshioka H, Yagi T, Obata JE, Kugiyama K, Kinouchi H. A Case of Ruptured Vertebrobasilar Junction Aneurysm Associated with Subclavian Steal Phenomenon. J Stroke Cerebrovasc Dis, 2017; 26: e160-e164
• 9)  Tonetti DA, Jankowitz BT. Subclavian Steal Flow-Related Aneurysm Formation. World Neurosurg, 2019; 125: 101-103
• 10)  Prognosis of patients with symptomatic vertebral or basilar artery stenosis. The Warfarin-Aspirin Symptomatic Intracranial Disease (WASID) Study Group. Stroke, 1998; 29: 1389-1392
• 11)  Fields WS, Lemak NA. Joint Study of extracranial arterial occlusion. VII. Subclavian steal--a review of 168 cases. JAMA, 1972; 222: 1139-1143
• 12)  Prasad S, Kasner SE. Teaching NeuroImage: subclavian steal syndrome. Neurology, 2007; 69: E1
• 13)  Yip PK, Liu HM, Hwang BS, Chen RC. Subclavian steal phenomenon: a correlation between duplex sonographic and angiographic findings. Neuroradiology, 1992; 34: 279-282
• 14)  Thierfelder KM, Baumann AB, Sommer WH, Armbruster M, Opherk C, Janssen H, Reiser MF, Straube A, von Baumgarten L. Vertebral artery hypoplasia: frequency and effect on cerebellar blood flow characteristics. Stroke, 2014; 45: 1363-1368
• 15)  Gulli G, Khan S, Markus HS. Vertebrobasilar stenosis predicts high early recurrent stroke risk in posterior circulation stroke and TIA. Stroke, 2009; 40: 2732-2737
• 16)  Osiro S, Zurada A, Gielecki J, Shoja MM, Tubbs RS, Loukas M. A review of subclavian steal syndrome with clinical correlation. Med Sci Monit, 2012; 18: RA57-63
• 17)  Tan TY, Schminke U, Lien LM, Tegeler CH. Subclavian steal syndrome: can the blood pressure difference between arms predict the severity of steal? J Neuroimaging, 2002; 12: 131-135
• 18)  Nicholls SC, Koutlas TC, Strandness DE. Clinical significance of retrograde flow in the vertebral artery. Ann Vasc Surg, 1991; 5: 331-336
• 19)  Kwak BR, Back M, Bochaton-Piallat ML, Caligiuri G, Daemen MJ, Davies PF, Hoefer IE, Holvoet P, Jo H, Krams R, Lehoux S, Monaco C, Steffens S, Virmani R, Weber C, Wentzel JJ, Evans PC. Biomechanical factors in atherosclerosis: mechanisms and clinical implications. Eur Heart J, 2014; 35: 3013-3020, 3020a-3020d
• 20)  Gnasso A, Irace C, Carallo C, De Franceschi MS, Motti C, Mattioli PL, Pujia A. In vivo association between low wall shear stress and plaque in subjects with asymmetrical carotid atherosclerosis. Stroke, 1997; 28: 993-998
• 21)  Nigro P, Abe J, Berk BC. Flow shear stress and atherosclerosis: a matter of site specificity. Antioxid Redox Signal, 2011; 15: 1405-1414
• 22)  Cunningham KS, Gotlieb AI. The role of shear stress in the pathogenesis of atherosclerosis. Lab Invest, 2005; 85: 9-23
• 23)  Kliewer MA, Hertzberg BS, Kim DH, Bowie JD, Courneya DL, Carroll BA. Vertebral artery Doppler waveform changes indicating subclavian steal physiology. AJR Am J Roentgenol, 2000; 174: 815-819
• 24)  Chen SP, Hu YP, Fan LH, Zhu XL. Bidirectional flow in the vertebral artery is not always indicative of the subclavian steal phenomenon. J Ultrasound Med, 2013; 32: 1945-1950
• 25)  Kargiotis O, Siahos S, Safouris A, Feleskouras A, Magoufis G, Tsivgoulis G. Subclavian Steal Syndrome with or without Arterial Stenosis: A Review. J Neuroimaging, 2016; 26: 473-480
• 26)  Toole JF, McGraw CP. The steal syndromes. Annu Rev Med, 1975; 26: 321-329
• 27)  Lord RS, Adar R, Stein RL. Contribution of the circle of Willis to the subclavian steal syndrome. Circulation, 1969; 40: 871-878
• 28)  Klingelhofer J, Conrad B, Benecke R, Frank B. Transcranial Doppler ultrasonography of carotid-basilar collateral circulation in subclavian steal. Stroke, 1988; 19: 1036-1042
• 29)  Qureshi AI, Caplan LR. Intracranial atherosclerosis. Lancet, 2014; 383: 984-998
• 30)  Gorelick PB, Caplan LR, Hier DB, Parker SL, Patel D. Racial differences in the distribution of anterior circulation occlusive disease. Neurology, 1984; 34: 54-59
• 31)  Saia A, Manzoni S, Iaia V. Subclavian steal diagnosis: Doppler sonography vs. angiography. Ital J Neurol Sci, 1983; 4: 439-442