sICAM-1 concentrations are associated with inflammation in contralateral carotid plaque in patients with ischemic stroke

Abstract

Background

Atherosclerotic plaques in the internal carotid artery are responsible for more than 15% of ischemic strokes. Carotid ¹⁸F-fluorodeoxyglucose positron emission tomography (¹⁸F-FDG PET) detects plaque inflammation. Plasma ICAM-1 and LRP1 concentrations have been associated with inflammation in ipsilateral carotid plaque. The aim of the present study was to test the association between the soluble (s) form of these biomarkers and contralateral carotid plaques.

Methods

Prospective study conducted in 53 patients with a recent ischemic stroke and at least one atherosclerotic plaque in both carotid arteries. All of the patients underwent an early carotid ¹⁸F-FDG PET, and a blood sample was obtained at 7 ± 1 days. Several plasma inflammatory markers were evaluated by Multiplex and sLRP1 levels were measured by commercial ELISA. Bivariate and multivariable linear regression was used to assess the association between inflammatory markers and the clinical variables, including contralateral maximum standardized uptake value (SUVmax) and mean SUVmax (mean of contralateral and ipsilateral SUVmax) of ¹⁸F-FDG uptake. Hazard ratio (HR) was estimated with Cox models adjusted for potential confounding factors to evaluate recurrence.

Results

Multivariable linear regression analysis showed an independent association between sICAM-1 and sVCAM-1 and mean SUVmax (CI = −0.064–0.325, p = 0.004; CI = 0.079–0.554, p = 0.010). In addition, in bivariate regression analysis, sICAM-1 was associated with contralateral SUVmax (CI = 0.049–0.382, p = 0.012). Cox regression showed that mean SUVmax was associated with stroke recurrence (HR = 5.604, p = 0.044).

Conclusions

sICAM-1 was independently associated with mean carotid plaque inflammation and with inflammation in contralateral plaque. sICAM-1 could be an indicator of plaque inflammation even in asymptomatic plaques.

Resumen

Antecedentes

Las placas ateroscleróticas en la arteria carótida causan más del 15% de ictus isquémicos. La tomografía carotídea por emisión de positrones con ¹⁸F-fluorodesoxiglucosa (¹⁸F-FDG PET) detecta inflamación de la placa. La concentración plasmática de ICAM-1 y LRP1 se ha asociado con inflamación en placa ipsilateral. El objetivo del estudio fue estudiar la asociación entre la forma soluble (s) de estos biomarcadores y las placas carotídeas contralaterales.

Métodos

Estudio prospectivo de 53 pacientes con ictus isquémico reciente y al menos una placa aterosclerótica en ambas carótidas. A los pacientes se les realizó ¹⁸F-FDG PET y se obtuvo sangre entre los 7 ± 1 días. Se evaluaron biomarcadores inflamatorios por Multiplex y los niveles de sLRP1 mediante ELISA. Se evaluó mediante regresión lineal bivariable y multivariable la asociación entre los biomarcadores y las variables clínicas, incluido el valor máximo de captación estandarizado contralateral (SUVmáx) y el SUVmáx medio (media decontralateral e ipsilateral). El índice de riesgo de recurrencia (HR) se estimó con modelos Cox.

Resultados

La regresión multivariable mostró la asociación independiente entre sICAM-1 y sVCAM-1 y SUVmáx medio (IC = −0,064-0,325; p = 0,004; IC = 0,079-0,554; p = 0,010). Además, en análisis de regresión bivariado, sICAM-1 se asoció con SUVmáx contralateral (IC = 0,049-0,382; p = 0,012). La regresión de Cox mostró que el SUVmáx medio se asoció con recurrencia del ictus (HR = 5,604; p = 0,044).

Conclusión

sICAM-1 se asoció de forma independiente con la inflamación media de la placa carotídea y con la inflamación de la placa contralateral. sICAM-1 podría ser un indicador de inflamación de la placa incluso en placas asintomáticas.

Introduction

Atherosclerosis is an inflammatory and chronic process characterized by the deposition of lipids into the artery wall. A plethora of inflammatory molecules and cells within the artery plays a key role in the progression of the lesion.¹ Eventually, this slow but progressive process of lipid accumulation and inflammation leads to plaque vulnerability, promoting that, when a certain threshold is reached, the plaque evolves to rupture, triggering the onset of vascular diseases.

Atherosclerotic plaques in the internal carotid artery are responsible for more than 15% of ischemic strokes. Ischemic stroke is a main cause of death and disability, and shows a high rate of recurrence, mainly in the atherothrombotic subtype. In order to prevent the high risk of recurrence associated with symptomatic carotid plaques, surgical revascularization is applied. These procedures, mostly carotid stenting or carotid endarterectomy (CEA), are conducted in carotid plaques causing a high degree of stenosis.^2,3 However, other vulnerability plaque characteristics should be considered in some common clinical situations, such as patients with vulnerable plaques causing less than 50% of stenosis or female patients with moderate carotid stenosis.^4,5

Carotid ¹⁸F-fluorodeoxyglucose positron emission tomography (¹⁸F-FDG PET) is a promising imaging biomarker that quantifies carotid plaque inflammation, a feature leading to plaque vulnerability. The predictive role of ¹⁸F-FDG PET is based on a large multicenter study conducted in stroke patients with carotid stenosis, in which an independent association between maximal standardized uptake value (SUVmax) within the plaque and early stroke recurrence was found.⁶ The prediction of recurrence improved when SUVmax and carotid stenosis were included in a risk score.⁷ Both, early and late stroke recurrences were predicted with high sensitivity by a value of SUVmax ≥ 2.85 g/mL,^6,8 even in patients with uncertain benefits from revascularization according to previous clinical trials.⁹

Contralateral Carotid Stenosis (CCS) has been proposed as a predictor of long-term adverse events in CEA.¹⁰ In addition, in the Controlled Clinical Trial SPACE-2, CCS together with echolucent carotid plaque has been associated with elevated stroke risk in asymptomatic patients with carotid stenosis.¹¹ A significant correlation was found between FDG-uptake in carotids tributary for CEA and in contralateral carotid plaques with less than 50% stenosis.¹² In other study, ¹⁸F-FDG uptake in the contralateral carotid artery was reported to be detectable but at a lower degree than in symptomatic atherosclerotic carotid arteries ipsilateral to a cerebrovascular event.¹³

¹⁸F-FDG PET is a reliable tool; however, it shows a high cost and limited availability. For this reason, several inflammatory molecules have been evaluated as plausible markers of carotid plaque vulnerability with the final aim of determining the risk of ischemic stroke. Therefore, several studies have been conducted to find new plasma biomarkers with diagnostic potential to detect carotid plaque vulnerability.

The clinical usefulness of determining blood inflammatory markers for the prediction of first-ever or recurrent strokes is controversial.¹⁴ Fractalkine (FKN), soluble vascular cell adhesion molecule-1 (sVCAM-1), and soluble intracellular adhesion molecule-1 (sICAM-1) concentrations were elevated in plasma from ischemic stroke patients and independently associated with the degree of inflammation in ipsilateral carotid plaque assessed by ¹⁸F-FDG PET. Among them, the concentration of sICAM-1 predicted the risk of finding highly inflamed carotid plaques and the recurrence of stroke within one year.¹⁵ Interestingly, such study also showed an association between BMI category and SUVmax, probably owing to the inflammatory state ascribed to obesity. In other study in the same cohort of patients with recent ischemic stroke and carotid atherosclerosis, soluble LDL receptor-related protein 1 (sLRP1) concentrations were also independently associated with the degree of plaque inflammation, as measured by ¹⁸F-FDG PET, and with the presence of highly inflamed carotid plaque. Moreover, those patients with stroke recurrence showed higher sLRP1 levels at baseline.¹⁶ These findings suggest that sICAM-1 and sLRP1 may be considered as surrogate markers of inflammation in symptomatic ‘culprit’ carotid plaque. However, it is unknown whether these molecules are related to inflammation in asymptomatic contralateral carotid plaque.

The detection of biomarkers indicating inflammation in the contralateral carotid artery, which is still asymptomatic, could be of great interest to early detect plaques prone to progress or become vulnerable in the future. Therefore, the aim of this study was to establish the association between contralateral inflammation assessed by ¹⁸F-FDG PET and inflammatory molecules previously found to be indicators of carotid plaque vulnerability in ischemic stroke patients.

Methods

Study design and population

This is an observational cohort study (NCT03218527) of patients who had had a recent anterior circulation ischemic stroke and present carotid atherosclerosis. The study was conducted in the Hospital Santa Creu i Sant Pau between January 2016 and March 2019. It was approved by the Ethics Committee of the Hospital (IIBSP-LRB-2017-54, June 26th 2017), and in accordance with the Helsinki Declaration. The patients gave written consent to participate.

The patients included in the study were those belonging to the same cohort than in our previous publications.^15,17 They were included when the following criteria were fulfilled: (1) age ≥ 50 years; (2) anterior circulation ischemic stroke or transient ischemic attack (TIA) within the previous 7 days; (3) at least one atherosclerotic plaque in the internal carotid artery (ICA) on the side consistent with stroke symptoms. Carotid stenosis was graded using the NASCET approach¹⁸ with computed tomography (CT) or magnetic resonance imaging-angiography, and based on hemodynamic criteria by using ultrasound¹⁹; (4) at least one atherosclerotic plaque in the ICA on the opposite site (contralateral) of the stroke symptoms ≥ 30% of stenosis; and (5) previous modified Rankin Scale (mRS) score < 4. The exclusion criteria were: (1) presence of cardioembolic, lacunar, or unusual stroke etiology according to the TOAST criteria²⁰; (2) presence of a hemodynamic stroke/TIA; (3) prior carotid surgery/stenting; (4) presence of comorbidities conditioning a life-expectancy <1 year; (5) concomitant infections, and (6) total artery occlusion.

The following clinical variables were recorded for all of the patients: (1) age and sex; (2) past medical history; (3) previous treatments; (4) National Institutes of Health Stroke Scale (NIHSS) score, as a surrogate of infarct size; (5) body mass index (BMI); (6) Regular physical exercise according to the physician-based assessment and counseling for exercise (PACE) scale²¹; (7) Mediterranean diet adherence according to the PREDIMED score²²; (8) mRS score at inclusion; (9) stroke etiology according to the TOAST criteria²⁰; and (10) results from the admission blood test including renal function, hemogram, hemostasis, and lipid profile. After the stroke, the treating clinicians provided medical and revascularization treatments according to guidelines.²³

Carotid plaque imaging

All stroke patients included in the study underwent a carotid ultrasound, including duplex examination and a Contrast-Enhanced Ultrasound Study (CEUS), a CT-angiography or MR-angiography, and an ¹⁸F-FDG PET/CT within 15 days from the index stroke. Inflammation of the carotid plaques was assessed by ¹⁸F-FDG PET according to the maximum Standardized Uptake Value (SUV) within the plaque, as detailed as follows. The other techniques are detailed in previous publications.^15,17

Carotid ¹⁸F-FDG PET was performed in a Philips Gemini TF TOF 64 PET/CT (Philips Medical System, Eindhoven, Netherland). The examinations were performed after a fast that lasted a minimum of six hours. PET scans were not performed if pre-PET blood glucose exceeded 10 mmol/L. Two hours before image acquisition, 320 MBq of ¹⁸F-FDG was administered. The uptake phase was standardized with the patient resting. PET images were acquired in a 3-dimensional mode in 2-bed positions for 10 min each. Images from CT angiography and PET were co-registered afterwards to assess the slice of maximal plaque stenosis.

¹⁸F-FDG activity was measured in 10 regions of interest, which were defined relative to the slice of maximal stenosis on the co-registered CT angiography, corresponding to a 1 mm axial plaque slice (5 distal and 5 proximal). ¹⁸F-FDG was quantified using standardized uptake values (SUV g/mL, defined as measured uptake [MBq/mL]/injected dose [MBq] per patient weight [g]). We defined the single hottest slice as the axial slice with maximal SUV uptake (SUVmax).⁶ The measurement was performed in each ICA and thereafter we calculated the mean SUVmax between both (mean SUVmax)

Plasma determinations

Peripheral blood samples from the stroke patient were collected at day 7 ± 1 from the stroke. Plasma was collected in ethylenediaminetetraacetic acid (EDTA)-containing Vacutainers and serum in Serum Separator Tubes with clot activator. The tubes were centrifuged at 1500 × g for 5 min at 4 °C, and the aliquots were frozen at −80 °C.

The plasma concentrations of sICAM-1, sVCAM-1, FKN, IL-6, IL-1β, and IL-8 were analyzed in a Luminex using xMAP® technology with a MILLIPLEX® MAP multiplexed assay kit (Merck Millipore, Burlington, MA, United States). sLRP1 concentrations were measured using commercially available enzyme-linked immunosorbent assay (ELISA) (Uscn Life Science Inc., Wuhan, China) according to the manufacturer's recommendations.

Statistical analysis

Continuous descriptive variables were reported as means and standard deviations (SD) or medians (md) and interquartile ranges (IQR) if they were not-normally distributed. Categorical variables were expressed as counts and percentages.

Bivariate linear regression analyses, after logarithmic transformation of the variables if they were not-normally distributed, were used to determine plasma molecules associated with carotid plaque ¹⁸F-FDG uptake (SUVmax in ICA contralateral and mean SUVmax), expressed as β coefficient with 95% confidence interval (CI). Backward stepwise multivariable linear regression modeling was performed considering each biomarker individually, and including independent clinical variables based on a p < 0.1 on bivariate analyses.

Bivariate Cox regression analyses were used to determine the association between SUVmax in the contralateral ICA and mean SUVmax with stroke recurrence during the follow-up, which was expressed as hazard ratio (HR) with 95% CI. Backward stepwise multivariable Cox regression analysis was performed, including independent clinical variables based on a p < 0.1 on bivariate analyses.

Spearmen's correlation was performed to analyze the correlation between contralateral carotid SUVmax and ipsilateral carotid SUVmax and contralateral SUVmax and plasma sICAM-1 concentration.

Statistical significance for all the analyses was set at p < 0.05 (two-sided). Analyses were performed using RStudio v. 4.3.2 (Vienna, Austria).

Results

Parameters associated with mean SUVmax

The study population included 53 patients, the clinical characteristics of whom are detailed in Table 1. The mean age was 75.5 ± 9.7 years, and 26.4% of female were included in the study. In this cohort, 54.7% showed a degree of stenosis ≥ 50% in the ipsilateral ICA, whereas only 11.3% did in the contralateral ICA. Contrariwise, contralateral SUVmax value did not differ from ipsilateral SUVmax value or mean SUVmax. A correlation was observed between contralateral SUVmax and ipsilateral SUVmax value, altough the correlation coefficient is rather low (p-value: 0.016; correlation coefficient = 0.33).

Table 1.

Clinical characteristics and biochemical parameters of the patients.

	Stroke patients (n = 53)
Age, mean (SD)	75.5 (9.7)
Sex (female), n (%)	14 (26.4)
BMI, mean (SD)	25.4 (3.4)
Regular exercise (PACE ≥ 4), n (%)	25 (45.3)
PREDIMED score, md (IQR)	9 (7–9)
Current Smoking, n (%)	13 (24.5)
Hypertension, n (%)	46 (86.8)
Diabetes, n (%)	23 (43.4)
Dyslipidemia, n (%)	37 (69.8)
Active or recent cancer (<5years), n (%)	4 (7.5)
Coronary artery disease, n (%)	17 (26.)
Prior stroke, n (%)	9 (17.0)
Prior antiplatelet therapy, n (%)	22 (41.5)
Prior statin therapy, n (%)	22 (41.5)
Carotid stenosis ≥ 50% in the ipsilateral ICA, n (%)	29 (54.7)
Carotid stenosis ≥ 50% in the contralateral ICA, n (%)	6 (11.3)
Mean SUVmax, md (IQR)	2.6 (2.5–2.7)
Ipsilateral SUVmax, md (IQR)	2.6 (2.4–2.8)
Contralateral SUVmax, md (IQR)	2.5 (2.4–2.7)
Baseline NIHSS, md (IQR)	2 (1–3)
Intravenous fibrinolysis, n (%)	11 (20.8)
Acute lesion on neuroimaging, n (%)	38 (71.7)
Atherothrombotic	20 (37.7)
Lacunar	7 (13.2)
Cryptogenic	20 (37.7)
Undetermined (two causes)	6 (11.3)
Creatinin (mg/dL), md (IQR)	0.96 (0.84–1.04)
Hemoglobin (g/L), md (IQR)	135 (130–143)
HbA1c (%), md (IQR)	5.9 (5.8–6.2)
Triglycerides (mg/dL), md (IQR)	104 (93–127)
Total cholesterol (mg/dL), md (IQR)	149 (137–170)
LDLc (mg/dL), md (IQR)	84 (64–101)
HDLc (mg/dL), md (IQR)	40 (36–45)
hsCRP (mg/L), md (IQR)	4.9 (3.1–7.1)

BMI (body mass index); PACE (physician-based assessment and counseling for exercise); NHISS (National Institutes of Health Stroke Scale); HbA1c (glycated hemoglobin); LDLc (low-density lipoprotein cholesterol); HDLc (high-density lipoprotein cholesterol); hsCRP (high-sensitivity C-reactive protein).

Analysis of Variance (ANOVA) or the Kruskal–Wallis rank-sum test (when a non-parametric test was required) for continuous variables, and the χ² test for categorical variables were used; p < 0.05 indicates significant differences between groups.

In a previous study, plasma concentration of sICAM-1, sVCAM-1, and FKN were found increased in patients versus healthy controls, and their blood levels were associated with ipsilateral SUVmax. Accordingly, in the present study, among the inflammatory molecules evaluated, the concentrations of sICAM-1, sVCAM-1, and FKN were also associated with mean SUVmax (Table 2). The strongest association was observed for sICAM-1 (p < 0.001). Among clinical parameters, BMI and current smoking were associated with mean SUVmax (Table 3). As shown in Table 3, the multivariable linear regression analysis showed an independent association with mean SUVmax for sICAM-1 and sVCAM-1.

Table 2.

Bivariate linear regression analyses of the association between each inflammatory biomarker and carotid inflammation (mean SUVmax).

	β coefficient	95% CI	β standardized	p
sICAM-1 (ng/mL)	0.272	0.139–0.405	0.523	<0.001
sVCAM-1 (ng/mL)	0.331	0.043–0.619	0.326	0.025
FKN (pg/mL)	0.311	0.094–0.528	0.391	0.006
Il-1β (pg/mL)	0.106	−0.065–0.278	0.181	0.219
IL-6 (pg/mL)	0.036	−0.037–0.197	0.145	0.324
IL-8 (pg/mL)	0.068	−0.065–0.278	0.154	0.219
sLRP1 (μg/mL)	0.158	−0.074–0.390	0.190	0.176

sICAM-1 (soluble intercellular adhesion molecule-1); sVCAM-1 (soluble vascular cell adhesion molecule-1); FKN (fractalkine); IL (interleukin); sLRP-1 (soluble low-density lipoprotein receptor-related protein 1).

Logarithmic transformation of the variables was used in bivariate linear regression analysis; n = 53; p < 0.05 indicates significant differences between groups.

Table 3.

Bivariate linear regression analyses of the association between clinical variables and mean SUVmax, and multivariable linear regression analyses of predictors of carotid plaque inflammation measured by ¹⁸F-FDG PET (mean SUVmax).

Bivariate analysis
	β coefficient	95% CI	β standardized	p
Age	0.007	−0.005–0.019	0.170	0.230
Sex (female)	0.101	−0.162–0.364	0.108	0.444
BMI	0.052	0.022–0.082	0.439	0.001
PACE score	−0.058	−0.119–0.003	−0.103	0.062
PREDIMED score	0.010	−0.040–0.061	0.057	0.688
Current smoking	−0.427	−0.662–(−0.192)	−0.459	<0.001
Hypertension	0.292	−0.033–0.617	0.247	0.077
Diabetes	0.055	−0.175–0.285	0.067	0.635
Dyslipidemia	0.108	−0.138–0.354	0.124	0.383
Active or recent cancer (<5 years)	0.405	−0.009–0.819	0.268	0.055
Coronary artery disease	0.009	−0.250–0.267	0.009	0.947
Prior stroke	−0.128	−0.429–0.172	−0.121	0.395
Prior antiplatelet therapy	0.004	−0.228–0.236	0.005	0.972
Prior statin therapy	0.007	−0.225–0.238	0.008	0.955
Carotid stenosis ≥ 50% in ipsilateral ICA	0.146	−0.081–0.373	0.180	0.202
Carotid stenosis ≥ 50% in contralateral ICA	0.290	−0.075–0.655	0.243	0.116
Triglycerides (mg/dL)	0.001	−0.001–0.004	0.169	0.232
Total cholesterol (mg/dL)	−0.000	−0.002–0.003	0.044	0.756
LDLc (mg/dL)	−0.000	−0.003–0.003	−0.012	0.931
HDLc (mg/dL)	0.001	−0.008–0.009	0.017	0.905

Multivariable analysis
	β coefficient	95% CI	β standardized	p
Model 1 (including sICAM-1) R^2 = 0.445
sICAM-1 (ng/mL)	0.195	−0.064–0.325	0.374	0.004
BMI	0.034	0.004–0.063	0.283	0.027
Current smoking	−0.225	−0.474–0.024	−0.242	0.075
Model 2 (including sVCAM-1) R^2 = 0.424
sVCAM-1 (ng/mL)	0.318	0.079–0.554	0.312	0.010
BMI	0.037	0.007–0.067	0.312	0.018
Current smoking	−0.336	−0.573–(−0.099)	−0.362	0.006
Model 3 (including FKN) R^2 = 0.379
FKN (ng/mL)	0.193	−0.006–0.392	0.243	0.057
BMI	0.032	0.001–0.063	0.269	0.044
Current smoking	−0.303	−0.549–(−0.056)	−0.327	0.017

BMI (body mass index); PACE (physician-based assessment and counseling for exercise); LDLc (low-density lipoprotein cholesterol); HDLc (high-density lipoprotein cholesterol); sICAM-1 (soluble intercellular adhesion molecule-1); sVCAM-1 (soluble vascular cell adhesion molecule-1); FKN (fractalkine).

Logarithmic transformation of the variables, if they were not normally distributed, was used in bivariate and multivariable linear regression analysis; backward stepwise multivariable linear regression modeling was performed individually for each biomarker; n = 53; p < 0.05 indicates significant differences between groups.

Parameters associated with contralateral SUV max

We repeated the association analyses considering specifically contralateral SUVmax. Table 4 shows that only sICAM-1 showed an association with contralateral SUVmax. Accordingly, in Fig. 1, a statistically significant correlation between plasma sICAM-1 values and contralateral SUV max was observed (p = 0.0122).

Table 4.

Bivariate linear regression analyses of the association between each inflammatory biomarker and contralateral carotid inflammation (contralateral SUVmax).

	β coefficient	95% CI	β standardized	p
sICAM-1 (ng/mL)	0.215	0.049–0.382	0.362	0.012
sVCAM-1 (ng/mL)	0.189	−0.153–0.532	0.164	0.272
FKN (pg/mL)	0.219	−0.042–0.480	0.241	0.098
Il-1β (pg/mL)	0.092	−0.106–0.289	0.136	0.356
IL-6 (pg/mL)	0.063	−0.019–0.144	0.222	0.129
IL-8 (pg/mL)	0.061	−0.087–0.209	0.122	0.409
sLRP1 (μg/mL)	−0.041	−0.311–0.228	-0.044	0.759

Contralateral SUVmax (contralateral maximal standardized uptake value); sICAM-1 (soluble intercellular adhesion molecule-1); sVCAM-1 (soluble vascular cell adhesion molecule-1); FKN (fractalkine); IL (interleukin); sLRP-1 (soluble low-density lipoprotein receptor-related protein 1).

Logarithmic transformation of the variables was used in bivariate linear regression analysis; n = 53; p < 0.05 indicates significant differences between groups.

Figure 1.

Correlation between plasma sICAM-1 (soluble intercellular adhesion molecule-1) concentration and contralateral SUVmax (contralateral maximal standardized uptake value) by spearmen's analysis.

Regarding clinical parameters, BMI, current smoking, and PREDIMED score were found associated with contralateral SUVmax (Table 5). However, according to the multivariable models, no independent association was found.

Table 5.

Bivariate linear regression analyses of the association between clinical variables and contralateral SUVmax, and multivariable linear regression analyses of predictors of contralateral carotid plaque inflammation measured by ¹⁸F-FDG PET (contralateral SUVmax).

Bivariate analysis
	β coefficient	95% CI	β standardized	p
Age	0.009	−0.004–0.022	0.187	0.184
Sex (female)	0.051	−0.251–0.352	0.048	0.737
BMI	0.059	0.025–0.094	0.440	0.001
PACE score	−0.026	−0.098–0.046	−0.103	0.468
PREDIMED score	0.056	0.000–0.111	0.273	0.049
Current smoking	−0.482	−0.751–(−0.212)	−0.453	<0.001
Hypertension	0.319	−0.054–0.170	0.236	0.092
Diabetes	−0.092	−0.354–0.170	−0.099	0.485
Dyslipidemia	0.173	−0.106–0.452	0.173	0.219
Active or recent cancer (<5years)	0.144	−0.345–0.633	0.083	0.558
Coronary artery disease	0.185	−0.105–0.476	0.178	0.205
Prior stroke	−0.110	−0.454–0.235	−0.090	0.526
Prior antiplatelet therapy	0.180	−0.080–0.439	0.193	0.171
Prior statin therapy	0.043	−0.221–0.308	0.076	0.743
Carotid stenosis ≥ 50% in contralateral ICA	0.104	−0.330–0.539	0.076	0.630
Triglycerides (mg/dL)	0.000	−0.003–0.004	0.040	0.777
Total cholesterol (mg/dL)	−0.001	−0.004–0.002	−0.080	0.575
LDLc (mg/dL)	−0.001	−0.005–0.003	−0.087	0.546
HDLc (mg/dL)	−0.005	−0.015–0.005	−0.136	0.336

Multivariable analysis
	β coefficient	95% CI	β standardized	p
Model 1 (including sICAM-1)R^2 = 0.264
sICAM-1 (ng/mL)	0.128	−0.022–0.279	0.216	0.093
BMI	0.064	0.029–0.099	0.473	0.001
PREDIMED score	0.074	0.018–0.129	0.338	0.011

SUVmax contralateral (contralateral maximal standardized uptake value); BMI (body mass index); PACE (physician-based assessment and counseling for exercise); LDLc (low-density lipoprotein cholesterol); HDLc (high-density lipoprotein cholesterol); sICAM-1 (soluble intercellular adhesion molecule-1).

Logarithmic transformation of the variables, if they were not normally distributed, was used in bivariate and multivariable linear regression analysis; backward stepwise multivariable linear regression modeling was performed individually for each biomarker; n = 53; p < 0.05 indicates significant differences between groups.

Parameters associated with outcome

In bivariate Cox regression analyses, mean SUVmax, but not contralateral SUV max, was associated with stroke recurrence within 1 year. However, mean SUVmax showed no independent association in multivariable Cox regression (Supplementary Table 1). There was not any recurrent event associated with contralateral plaque.

Discussion

The present study suggests that sICAM-1 is a good indicator of the inflammatory state of the carotid arteries. Among several inflammatory markers previously found increased in ischemic stroke patients with carotid atherosclerosis, we have found that: (1) sVCAM-1 and sICAM-1 are independently associated with global inflammation in both carotid plaques, ipsilateral and contralateral; and (2) Only sICAM-1 was specifically associated with inflammation in contralateral carotid plaque.

In a previous publication, we found that FKN, sVCAM-1, and sICAM-1 concentrations were independently associated with the degree of inflammation in ipsilateral carotid plaque assessed by ¹⁸F-FDG PET.¹⁵ Conversely, in a study conducted in non-stroke patients, but with coronary heart disease or at high risk of heart disease, sICAM-1, sVCAM-1 and other inflammatory mediators did not correlate with carotid target background ratio, another parameter of plaque inflammation.²⁴ FKN, VCAM-1, and ICAM-1 are membrane proteins that are released to the circulation as soluble forms by activated cells in inflammatory states. Temporal dynamics of biomarkers after stroke may depend on infarct size, as found for FKN in the study of Grosse et al., in which a pronounced decrease of FKN associated with worse clinical outcome after the stroke.²⁵ In our population, neurological damage as a source of inflammatory biomarkers is discarded by the low Rankin and NIHSS scores, corroborating that the patients had minor strokes and low brain injure.

Of the candidate molecules, sICAM-1 additionally predicted the risk of finding highly inflamed carotid plaques (SUVmax ≥ 2.85 g/ml) and the recurrence of stroke within one year.¹⁵ Likewise, sLRP1 concentrations were independently associated with the degree of plaque inflammation, highly inflamed plaque, and stroke recurrence post-PET.¹⁶ These studies suggest that sICAM-1 and sLRP1 were good candidate biomarkers for carotid plaque inflammation. However, the association was established in “symptomatic” ipsilateral carotid plaque, and hence in asymptomatic plaques the results could be rather different. In this context, it would be very useful to identify biomarkers that mirror the global inflammatory state of the carotid arteries and/or of asymptomatic carotid arteries in order to early detect those plaques prone to progress or become vulnerable, leading eventually to a future ischemic stroke.

In this study, mean SUVmax, which is the mean of the inflammation degree of ipsilateral and contralateral carotid plaques, was associated with plasma concentrations of FKN, sVCAM-1, and sICAM-1; of them, sVCAM-1 and particularly sICAM-1 showed an independent statistical association with mean SUVmax. Mean SUVmax was found associated with ischemic stroke recurrence, in agreement with that found for SUVmax in ipsilateral carotid plaque.¹⁵ Unexpectedly, sLRP1 concentration was not associated with mean SUVmax. These observations indicate that sICAM-1 seems to be a more global marker of the inflammatory state of carotid plaques, and sLRP1 could rather be an indicator of local injury. A previous study in an experimental rabbit model showed that vascular cholesteryl ester (CE) accumulation was closely related to LRP1 expression.²⁶ In smooth muscle cells, LRP1-mediated CE accumulation induced tissue factor (TF) overexpression and release.²⁷ Owing to this role of LRP1 in mediating an exarcebated release of TF, a crucial mediator of thrombosis, it seems logical that sLRP1 associates with SUVmax in those carotid plaques that are more susceptible to thrombus formation. Moreover, lack of correlation with SUVmax does not discard an association of sLRP1 with the global inflammatory state of the contralateral plaque quantified by parameters such as whole-vessel SUVmean. Dynamics of sLRP1 and other biomarkers after the stroke onset and the association with carotid inflammation and clinical outcome also deserve further investigations.

When the association of the putative biomarker with the degree of inflammation specifically in the contraleral carotid artery was evaluated, only sICAM-1 showed a statistically significant association. However, in the multivariable analysis, sICAM-1 was not independently associated with contralateral SUVmax, although it showed a strong trend to reach the signification (p = 0.093). CCS has been described to predict long-term adverse events in CEA¹⁰ and risk of stroke in asymptomatic patients with carotid stenosis.¹¹ In the present study, the degree of contralateral inflammation was not associated with ischemic stroke recurrence in our patients within one year.

In an interesting study, a correlation between ¹⁸F-FDG uptake in symptomatic carotids scheduled for CEA and contralateral carotid plaques with moderate stenosis was found.¹² Accordingly, we also showed a correlation between SUVmax in ipsilateral and contralateral carotid plaques, finding similar ¹⁸F-FDG uptake in both despite contralateral plaques showing lower stenosis degree. On the other hand, other study revealed lower ¹⁸F-FDG uptake in the contralateral carotid artery than in symptomatic ipsilateral carotid arteries.¹³ Altogether, these observations suggest that, as inflammation is a systemic process, the evaluation of the inflammatory state of carotid plaques would provide useful information to stratify the risk of ischemic stroke. However, owing to the high cost and limited availability of PET, finding a plasma biomarker such as sICAM-1 as surrogate of carotid plaque inflammation is of great importance.

Of note, in our study, BMI was independently associated with mean SUVmax and with contralateral SUVmax. This finding is in line with the association found between BMI and ipsilateral SUVmax in our cohort.¹⁵ It is widely described that not only obese but also overweight patients have a low-grade chronic inflammatory state and cardiovascular risk.²⁸ The association between sICAM-1 and SUVmax does not seem to depend on this risk factor, as it was found to be independent of BMI in multivariable analysis, and it was also corroborated by the lack of correlation between sICAM-1 and BMI (p = 0.1517, data not shown). Some studies have described a relationship of BMI with carotid atherosclerosis, with BMI ≥ 30 predicting the presence of unstable carotid plaques²⁹ and plaque progression.³⁰ However, the relationship between BMI (or alterations in adiposity) and atherothrombotic ischemic stroke remains elusive and deserves further investigations.

In summary, as inflammation is a systemic process, the evaluation of the inflammatory state of carotid plaques would provide useful information to stratify the risk of ischemic stroke. However, owing to the high cost and limited availability of PET, finding plasma biomarkers as surrogates of carotid plaque inflammation is essential. The present study suggests that sICAM-1 is a good indicator of the global inflammatory state of the carotid arteries, also of contralateral asymptomatic plaques. In this context, it would be potentially useful for the identification of asymptomatic carotid arteries prone to progress or become vulnerable, leading eventually to a future ischemic stroke. This knowledge could be of huge importance in future clinical practice in order to early predict inflamed plaques with risk of stroke in asymptomatic patients.

Funding

This study has received funding from Fundación Española de Arteriosclerosis (FEA/SEA 2021 Basic Research Award to SB), Fundació La Marato TV3 (201716.10), FIS PI19/00421 (to PC and SB) and FISPI21/01523 (to VLl-C), from the Instituto de Salud Carlos III (ISCIII) and co-financed by the European Regional Development Fund). SB and JLS-Q are part of CIBER of Diabetes and Metabolic Diseases (CIBERDEM, CB07/08/0016) and VLl-C is part of CIBER Enfermedades Cardiovasculares (CIBERCV; CB16/11/00276) by ISCIII. PC-R and SB are members of RICORS-ICTUS by ISCIII. EG and VLl-C are members of Redes de investigación (Enfermedades Metabólicas y Cáncer RED2018-102799-T), a project run by MINECO. VLl-C is a member of the Quality Research Group 2017 SGR 946 and SB, NP, and JLSQ of the 2017-SGR-1149 group from Generalitat de Catalunya. VLL-C, EG, SB and JLS-Q are members of the Spanish Atherosclerosis Society Vascular Biology Group.

Conflict of interests

The authors declare that they have no conflict of interest.