Association of insulin resistance with intra- and extra-cranial atherosclerotic burden in the nondiabetic community population

AIMS
Few population-based studies have investigated the association between insulin resistance and atherosclerotic burden in intra- and extra-cranial arteries. The purpose of this study is to explore the relationship between insulin resistance and intra- and extra-cranial atherosclerotic burden in community-based nondiabetic participants.


METHODS
This is a cross-sectional analysis from a population-based prospective cohort-PolyvasculaR Evaluation for Cognitive Impairment and vaScular Events (PRECISE) study in China. The homeostasis model assessment of insulin resistance (HOMA-IR) and insulin sensitivity indices (ISI0-120) were stratified by the quartiles, respectively. The atherosclerotic presence of plaques and burden was evaluated by high-resolution MRI. Binary or ordinal logistic regression was performed to assess the association between HOMA-IR or ISI0-120 and the presence and burden of atherosclerosis.


RESULTS
Among the 2754 participants, the mean age was 60.9 ± 6.6 years, and 1296 (47.1%) were males. Compared with the lowest quartile of HOMR-IR, the highest quartile of HOMA-IR (indicating a higher level of insulin resistance) was associated with an increased presence of plaques (OR:1.54, 95% CI:1.14-2.08), and atherosclerotic burden (OR:1.53, 95%CI:1.14-2.07) in intracranial arteries. Meanwhile, we observed a similar relationship between HOMA-IR and the presence or burden in extracranial atherosclerosis. The first (indicating a higher level of insulin resistance) quartiles of ISI0-120 were associated with the intracranial plaques (Q1, OR:1.56, 95%CI:1.16-2.11) and atherosclerotic burden (Q1, OR:1.57, 95%CI:1.17-2.12), but not extracranial plaques or atherosclerotic burden, compared with the fourth quartile of ISI0-120.


CONCLUSIONS
Insulin resistance was associated with an increased intra-and extra-cranial atherosclerotic burden in the nondiabetic elderly Chinese population.


Introduction
Atherosclerosis is characterized by progressive pathogenesis, affecting multiple vascular beds, which promotes the development of plaques, exacerbates luminal stenosis, and leads to cardiovascular diseases. (Rahman and Woollard, 2017) Early detection of relevant risk factors to identify atherosclerotic plaques in asymptomatic community populations is of great value to prevent plaque development and cerebrovascular disease. Previous studies (Cury et al., 2017;Sardu et al., 2021;Skoloudik et al., 2022;Zhang et al., 2021) have found that metabolic risk factors were associated with symptomatic intra-and extra-cranial atherosclerotic stenoses and plaques.
Insulin resistance, a comprehensive manifestation of the descent sensitivity of body tissues to insulin, was reported to be closely related to inflammatory responses, glucose and lipid metabolism, as well as endothelial dysfunction, (Baranowska-Bik and Bik, 2019;Gimbrone Jr. and Garcia-Cardena, 2016;Muniyappa and Sowers, 2013) which may contribute to the formation and progression of atherosclerosis. Two cross-sectional studies reported that insulin resistance might predict the prevalence of asymptomatic intracranial arterial stenosis (Lopez-Cancio et al., 2012;Wang et al., 2020a) whereas results from the Asymptomatic Polyvascular Abnormalities Community (APAC) showed inconsistent conclusions (Wang et al., 2021a). Moreover, previous studies revealed the positive relationships between insulin resistance and carotid plaques among the type 1 or 2 diabetes population (Ke et al., 2021;Pane et al., 2020) as well as the general population. (Ishizaka et al., 2003;Wang et al., 2021b) Conversely, a study consisting of Japanese patients with type 2 diabetes showed that metabolic syndrome, an insulin-resistant state, was not associated with carotid atherosclerosis. (Taniguchi et al., 2007) However, these investigations either used proxy index (TyG calculated with triglyceride and fasting blood glucose) to evaluate insulin resistance or used ultrasound systems to detect atherosclerotic plaques. Due to the limitations related to accuracy and reproducibility, potential bias in qualitative and quantitative measurements could have affected the results. (Lu et al., 2019) Therefore, the association between insulin resistance and intra-and extra-cranial atherosclerotic plaques in non-diabetic community populations remains controversial.
This study was conducted to evaluate the association between insulin resistance and intra-and extra-cranial atherosclerotic plaque and burden detected by high-resolution MRI in a community-based nondiabetic population.

Study setting and participants
Data were derived from the PolyvasculaR Evaluation for Cognitive Impairment and vaScular Events (PRECISE) study (ClinicalTrials. gov Registry: NCT03178448). (Pan et al., 2021)The PRECISE study was a community-based prospective cohort study of older adults (aged 50-75 years) according to cluster sampling from six villages and four communities of Lishui City in Zhejiang province, southeast China. This study involved both a cross-sectional and a longitudinal design with the primary objective to establish the prevalence of polyvascular lesions, and the progression rate of plaque formation in intracranial and carotidal arteries, with a follow-up of 4-years. Data in this analysis were obtained from the cross-sectional baseline survey of the PRECISE study, which recruited 3067 subjects from May 2017 to September 2019. The rationale, design, and baseline characteristics have been previously described. (Pan et al., 2021) The PRECISE protocol and data collection was approved by the ethics committee of Beijing Tiantan Hospital (IRB approval number: KY2017-010-01) and Lishui Hospital (IRB approval number: 2016-42). All participants signed written informed consent before data collection.
In our current analysis, participants were excluded if: 1) had a history of diabetes mellitus; 2) missing data of fasting glucose, insulin, or MRI images for intra-and extracranial arteries.

Baseline characteristics
Baseline data were collected through face-to-face interviews by trained research coordinators following the standard protocol. In our analysis, demographics, medical history, laboratory examinations, and image assessment were included. Demographics involved age, sex, body mass index (BMI = weight [kg] /the square of height in meters [m 2 ]), waist circumference, systolic blood pressure (SBP), and diastolic blood pressure (DBP). The medical history obtained included hypertension, dyslipidemia, and stroke, as well as their current smoking and alcohol consumption statuses. All participants provided blood samples to investigate the laboratory indexes including fasting plasma glucose (FPG), glycated hemoglobin (HbA1c), total cholesterol (TC), triglyceride (TG), high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C).

Insulin resistance assessment
The standard oral glucose tolerance test (75 g-OGTT) was performed in all nondiabetic subjects. Venous blood samples were obtained after an overnight fast of at least 8 h and again at 2 h after a 75-g oral glucose challenge. The fasting and 2-h post-load glucose levels were measured by an enzymatic method and competitive radioimmunoassay.

MRI acquisition and assessment
At baseline, we used a 3.0 T scanner (Ingenia 3.0 T, Philips, Best, The Netherlands) for eligible participants to detect intra-and extra-cranial atherosclerotic plaques. All the high-resolution MRI scanners were conducted on a designated machine based on a standardized protocol. (Pan et al., 2021) The imaging sequences included three-dimensional time-of-flight magnetic resonance angiography (3D-TOF MRA) and three-dimensional isotropic high-resolution black-blood T1w vessel wall imaging (3D T1w VWI). Imaging data were collected in digital imaging and communications in medicine format on discs and submitted to the imaging research center of Beijing Tiantan Hospital for central adjudication.
The atherosclerotic plaques were assessed and interpreted by two well-trained evaluators (D⋅Y and H.L), and any discrepancy was determined by a senior neurologist (J.J). The intracranial arterial segments evaluated included the bilateral internal carotid, middle cerebral artery (M1, M2), anterior cerebral arteries (A1, A2), posterior cerebral arteries (P1, P2), basilar, and vertebral arteries (V4). The extracranial arterial segments included the common carotid, proximal internal carotid artery, and vertebral artery (V1, V2, V3). The presence of atherosclerotic plaque was defined as eccentric wall thickening with or without luminal stenosis identified on the 3D T1w VWI images compared with the corresponding location on the 3D TOF MR angiogram. (eMethods 1 in the Supplement) (Qiao et al., 2016;Qiao et al., 2014) If no narrowing was identified on 3D-TOF MRA, the plaque was considered to have no detectable stenosis. (Qiao et al., 2016) The luminal stenosis in the intracranial arteries was measured by using criteria established in the Warfarin-Aspirin Symptomatic Intracranial Disease Trial (WASIDT) (Samuels et al., 2000) and the extracranial stenosis was assessed by using criteria from the North American Symptomatic Carotid Endarterectomy Trial (NASCET) (Fox, 1993). According to the degree of atherosclerotic plaques and stenosis, semi-quantitative score was used to evaluate atherosclerotic conditions (0, no atherosclerotic plaque; 1 atherosclerotic plaque without evident lumen stenosis or stenosis < 50%; 2, stenosis 50-69%; 3, stenosis 70-99%; 4, occlusion). The intracranial atherosclerotic burden was calculated by summing the intracranial atherosclerotic scores of intracranial arterial segments, which were categorized into four grades of 0, 1, 2-3, and ≥ 4 scores. Similarly, the extracranial atherosclerotic burden was counted through extracranial arterial segments, and categorized into four grades of 0, 1, 2-3, and ≥ 4 scores. (Kong et al., 2020;Sun et al., 2018;Wang et al., 2022) Meanwhile, the total atherosclerotic burden of intra-and extracranial arteries was summed and graded as 0, 1, 2-3, and ≥ 4 scores. Good interrater reproducibility was found for the presence of plaque and artery stenosis ≥50% in all territories (Cohen κ = 0.97 and 0.79 for intracranial artery and Cohen κ = 0.94 and 0.86 for extracranial artery).

Statistical analysis
Data were presented as mean±standard deviation (SD) for continuous variables with normal distribution, median and interquartile range (IQR) for skewed variables, and number and percentage for categorical variables. The baseline characteristics were compared among quartiles of HOMA-IR and ISI 0-120 using the chi-square test for categorical variables and ANOVA for continuous variables. The lowest quartile (Q1) of HOMA-IR or the highest quartile (Q4) of ISI 0-120 was used as a reference in the study.
The association of the quartiles of HOMA-IR or ISI 0-120 with the presence of atherosclerotic plaques was analyzed by binary logistic regressions with odds ratios (ORs) and their 95% confidence intervals (CIs). Ordinal logistic regression models were used to assess the association of the HOMA-IR or ISI 0-120 index with atherosclerotic burden by evaluating the common odds ratios (cORs) and their 95% CIs. Three models were performed, model 1 was adjusted for age and sex; model 2 was further adjusted for current smoking, and current drinking; model 3 was further adjusted for BMI, hypertension, and dyslipidemia. We analyzed the pattern and magnitude of association between HOMA-IR or ISI 0-120 as a continuous variable and the atherosclerotic burden using an ordinal logistic regression with restricted cubic spline adjusting for age, sex, current smoke, and current drink (model 2). The 5 knots were placed at the 5th, 25th, 50th, 75th, and 95th percentiles of HOMA-IR or ISI 0-120 .
All statistical analyses and graphs were performed with SAS software version 9.4 (SAS Institute Inc., Cary, NC). All p values were 2-sided, with p < 0.05 considered significant.

Baseline characteristics of the participants
A total of 3067 subjects were enrolled in the PRECISE study, 307 of whom were excluded due to the history of diabetes mellitus; moreover, 6 participants were further excluded owing to missing data of fasting glucose, insulin, or MRI images for intracranial arteries. A total of 2754 participants were finally included in the current analysis as shown in Fig. 1. The participants' mean age was 60.9 ± 6.6 years and 1296 (47.1%) participants were males. The median HOMA-IR was 1.55 (IQR, 1.07-2.24), and the median ISI 0-120 was 72.9 (IQR, 54.3-97.1).
The participants with higher HOMA-IR index (indicating a higher level of insulin resistance) were more likely to be younger, be female, have a higher body mass index, waist circumstance, blood pressure, as well as a higher prevalence of hypertension, dyslipidemia, but less likelihood to be current drinkers or smokers ( Table 1). Except for highdensity lipoprotein cholesterol and total cholesterol, glucose, and lipid metabolism indicators increased as the HOMA-IR elevated. Similar trends of baseline characteristics were observed when using ISI 0-120 to evaluate insulin resistance (a lower ISI 0-120 index indicates a higher level of insulin resistance) ( Table 2).

Association of HOMA-IR with the presence of plaque or atherosclerotic burden
Among the 2754 participants, 443 (16.1%) intracranial and 983 (35.8%) extracranial plaques were identified (Table 3). The higher percentage of intracranial atherosclerotic burden was in the fourth quartile of HOMA-IR (Fig. 3A). After adjustment for age and sex, current smoking, and drinking status, it was apparent that the participants with the highest quartile of the HOMA-IR (indicating a higher level of insulin resistance) were associated with an increased odds of the presence of plaques (adjusted OR,1.54; 95%CI, 1.14-2.08) or atherosclerotic burden (adjusted OR,1.53; 95%CI, 1.14-2.07) in intracranial arteries using the first quartile of HOMA-IR as the reference. A similar relationship was found that the elevated level of HOMA-IR was associated with an increased OR and cOR of extracranial atherosclerotic plaque and burden, respectively. After further adjustment of the variables, the correlation attenuated after further adjustment of BMI, hypertension, and dyslipidemia. Ordinal logistic regression with restricted cubic spline further also showed that the evaluated level of HOMA-IR was related to an increased cOR of atherosclerotic burden in intra-and extra-cranial arterial segments (Fig. 2 A-C).

Association of ISI 0-120 with the presence of plaque or atherosclerotic burden
Compared with the fourth quartile of ISI 0-120 , the first (indicating a higher insulin resistance) quartile of ISI 0-120 was associated with an increased odds of the presence of plaque (adjusted OR, 1.56, 95%CI, 1.16-2.11) or atherosclerotic burden (adjusted cOR,1.57, 95%CI, 1.17-2.12) in intracranial arteries as well as the second quartile had a similar trend in model 2. The results were not statistically different in model 3. (Table 4). Ordinal logical regression with restricted cubic spline showed an "L" shape nonlinear correlation of the ISI 0-120 level with the atherosclerotic burden (Fig. 2D). The intracranial atherosclerotic burden showed higher in the first quartile than the fourth quartile of ISI 0-120 (Fig. 3C). For extracranial atherosclerosis, the results failed to show any statistically significant association (Tables 2, 4; Figs. 2E-F, and 3D).

Discussion
In this large community-based study, we identified the association between insulin resistance and intra-and extra-cranial atherosclerosis in the elderly non-diabetic population in China. Furthermore, elevated HOMA-IR was closely related to the presence and burden of intra-and extra-cranial atherosclerotic plaque, whereas ISI 0-120 was only associated with intracranial but not extracranial atherosclerotic burden.
Previous studies indicated that insulin resistance has a positive relationship with intra- ( Barcelona-Asymptomatic Intracranial Atherosclerosis (AsIA) study enrolled 933 subjects (>50 years) and supported that insulin resistance (HOMA-IR ≥ 3.2) was an independent predictor for moderate to severe intracranial atherosclerotic disease, but was not an indicator for extracranial atherosclerosis. (Lopez-Cancio et al., 2012) The Chinese rural (Kongcun Town) residents' study confirmed the results that insulin resistance with HOMA-IR (Wang et al., 2020a) or TyG (Zhang et al., 2021) was associated with intracranial atherosclerotic stenosis. In contrast, another study in China indicated that the proxy index of TyG had an association with extracranial stenosis (detected by carotid duplex ultrasound examination), but not with intracranial stenosis (scanned by Transcranial Doppler) in the cross-sectional and longitudinal analysis. (Wang et al., 2021a) Differences in methods to assess insulin resistance (HOMA-IR or TyG), study population, and design may account for the different results between our study and previous studies ( Moreover, previous studies generally focused on macrovascular atherosclerotic stenoses but neglected early atherosclerotic plaque lesions that failed to cause luminal stenosis, which could underestimate the prevalence of atherosclerosis in the normal population. Using highresolution vascular MRI to evaluate the atherosclerotic plaques, our study found that elevated insulin resistance (higher-HOMA-IR and lower-ISI 0-120 ) increased the odds of the presence of intracranial plaques, which was in line with the Kongcun Town (Wang et al., 2020a) and the Barcelona-AsIA study (Lopez-Cancio et al., 2012). Furthermore, multiple atherosclerotic lesions were strongly associated with higher atherosclerotic burden in intracranial vessels (Lopez-Cancio et al., 2012;Sun et al., 2018), and similar results were found for the association of HOMA-IR with atherosclerotic burden in our analysis. The association between insulin resistance and intracranial atherosclerotic plaques may be influenced or neutralized by BMI, hypertension, and hyperlipidemia. Because the association between insulin resistance and BMI, hypertension, or hyperlipidemia, respectively was evident. (Barazzoni et al., 2018;Rett et al., 1994;Toh and Rader, 2008).
To our knowledge, some studies focused on the relationship between  insulin resistance and the presence of extracranial atherosclerotic plaques in diabetes. (Ke et al., 2021;Pane et al., 2020;Rett et al., 1994) However, cross-sectional studies in non-diabetic community populations are limited and conclusions remain equivocal. A cross-sectional analysis of 762 volunteers from the Cyprus Study indicated that HOMA-IR and McAuley Index were associated with both carotid plaque presence and area. (Panayiotou et al., 2015) Moreover, the Chinese investigation further found that elevated the TyG index was only significantly related to the prevalence of unstable but not stable carotid plaque in nondiabetic adults. (Sourij et al., 2008) In contrast, data from the Northern Shanghai Study (NSS) in China showed that higher-TyG index levels were not associated with carotid plaque. (Zhao et al., 2019) Interestingly, our results indicated that only a higher HOMA-IR but not lower ISI 0-120 was associated with elevated extracranial atherosclerotic plaque or burden. A similar result was presented including all the intraor extra-cranial plaques. These inconsistent findings may be attributed to the differences in the study population, design, evaluation of insulin resistance, and the accuracy of the image measurements. In this investigation, we applied high-resolution MRI to detect the presence of atherosclerotic plaques, which has superior reproducibility and accuracy to the other image techniques (ultrasound systems, MRI) adopted in previous research (Wang et al., 2020b;Young et al., 2019). Moreover, high-resolution MRI with the sequences of vessel wall magnetic resonance imaging (3D T1w VWI) directly and precisely visualized intracranial and extracranial arteries stenosis and plaques and their pathological alterations, allowing a better characterization of their pathology. Above all, the highlight of our study is not only using highresolution MRI to detect atherosclerotic plaques more precisely but also evaluating the comprehensive level of insulin resistance with two indices (HOMA-IR and ISI 0-120 ), which could provide vital evidence for the association between insulin resistance and atherosclerotic plaque and burden in large-scale population studies in Asia.
Recent studies (Baranowska-Bik and Bik, 2019; Muniyappa and Sowers, 2013) have revealed that the early pathological mechanism in the formation of atherosclerotic plaques due to insulin resistance was based on endothelial cell damage, but the exact mechanism was not fully elucidated. The balance between endothelial function and insulin metabolism is extremely important through oxidative stress and inflammatory reactions. When the balance of endothelial function and insulin metabolism is disrupted, it leads to decreased vasodilation activity, enhanced inflammatory response, and instability of lipid plaques, which seriously endangers the function and structure of blood vessels Table 3 Association of outcomes with insulin resistance states stratified by quartiles of HOMA-IR index. § The intracranial atherosclerotic burden was graded according to the sum of intracranial atherosclerosis scores, including 0, 1, 2-3 and ≥ 4 scores. || The extracranial atherosclerotic burden was graded according to the sum of extracranial atherosclerosis scores, including 0, 1, 2-3 and ≥ 4 scores. # The intra-and extra-cranial atherosclerotic burden was graded according to the sum of the intra-and extra-cranial atherosclerosis scores, including 0 score, 1 score, 2-3 scores, ≥4 scores.
and further promotes the formation and progression of plaques.However, insulin resistance and endothelial function damage involve multiple complex pathways such as glucose and lipid metabolism abnormalities. (Ormazabal et al., 2018) The ELSA-Brasil study illustrated that the relationship between HOMA-IR and carotid intima-media thickness was not mediated by glycemic levels. (Santos et al., 2017) In the current analysis, the ISI 0-120 indicator reflected the 0 min fasting and 120 min post-load level of glucose metabolism, but it was insignificantly associated with extracranial atherosclerotic plaques. We speculated that the mechanisms that were unrelated to post-load glucose homeostasis, as a direct effect of insulin on extracranial atherosclerosis, may be involved in this association. The solid line indicates adjusted cOR, and the dashed lines the 95% CI bands. Reference is the first quartile of HOMA-IR (1.07) and the third quartile of ISI 0-120 (97.1). The vertical dashed lines indicate the first, second, and third quartiles of HOMA-IR and ISI 0-120 . Ordinal logistic regression model of the restricted cubic spline (the 5th, 25th, 50th, 75th, 95th percentiles) was used for HOMA-IR and ISI 0-120 adjusting for potential covariates.
There are some limitations to our study: Firstly, our cross-sectional analysis could not conclude the causal relationship between insulin resistance and atherosclerotic burden. Our findings need to be further verified using Mendelian randomization studies. Secondly, the lack of pathological findings of the lesioned vessels in this study is related to the limited access to the plaque in the community population, which cannot provide further consistent findings of pathology and imaging. Thirdly, the assessment of the imaging was not validated by other techniques such as Doppler ultrasound and CT-Angiography. Compared with these imaging methods, high-resolution MRI scanning time is long and difficult to be tolerated by some patients. Fourthly our study only included the elderly Chinese population aged 50-to-75 years, which means the generalizability of our findings may be limited. Therefore, large-scale studies will be needed to verify the relationship between insulin resistance and atherosclerotic plaques.

Conclusions
In this study, among the non-diabetic elderly population in the Chinese community, there was a significant association between insulin resistance with a higher HOMA-IR or lower ISI 0-120 and elevated atherosclerotic burden in intra-or extra-cranial arteries. These findings suggested that insulin resistance could be a vital index to predict the presence and burden of atherosclerotic plaque in non-diabetic community-dwelling adults.

Financial support
This work was supported by the Beijing Hospitals Authority Youth Programme (QML20190501), National Natural Science Foundation of China (81971091). Key Science & Technologies R&D Program of Lishui City (2019ZDYF18), Zhejiang provincial program for the Cultivation of High-level Innovative Health talents and AstraZeneca Investment (China) Co., Ltd.

Disclosure
The authors declared they do not have anything to disclose regarding conflict of interest with respect to this manuscript.
Approval of the research protocol: The PRECISE protocol and data collection were approved by the ethics committee of Beijing Tiantan Hospital (IRB approval number: KY2017-010-01) and Lishui Hospital OR, odd ratio; cOR,common odds ratio; CI, confidence interval; Ref., reference. OR and 95% CI for the presence of intracranial and extracranial plaque by logistic regression analysis. cOR and 95% CI for intracranial, extracranial and intra-and extra-cranial atherosclerotic burden by ordinal logistic regression analysis. * Model 1 adjusted for age, sex. † Model 2 adjusted for age, sex, current smoker, current drinker. ‡ Model 3 adjusted for age, sex, current smoker, current drinker, BMI, hypertension, dyslipidemia. § The intracranial atherosclerotic burden was graded according to the sum of intracranial atherosclerosis scores, including 0, 1, 2-3 and ≥ 4 scores. | | The extracranial atherosclerotic burden was graded according to the sum of extracranial atherosclerosis scores, including 0, 1, 2-3 and ≥ 4 scores. # The intra-and extra-cranial atherosclerotic burden was graded according to the sum of the intra-and extra-cranial atherosclerosis scores, including 0 score, 1 score, 2-3 scores, ≥4 scores.
(IRB approval number: 2016-42). Informed Consent: All participants signed written informed consent before data collection.
Approval date of Registry and the Registration No. of the study/trial: June 4, 2017 ClinicalTrials. gov Registry: NCT03178448 Animal studies: N/A.

Credit author statement
This study was from the PolyvasculaR Evaluation for Cognitive Impairment and vaScular Events (PRECISE) (ClinicalTrials. gov Registry: NCT03178448).

Fig. 3.
Distribution of intra-or extra-cranial atherosclerotic burden in different quartiles of HOMA-IR or ISI 0-120 . The distribution of intracranial atherosclerotic burden in different quartiles of HOMA-IR(A). The distribution of extracranial atherosclerotic burden in different quartiles of HOMA-IR(B). The distribution of intracranial atherosclerotic burden in different quartiles of ISI 0-120 (C). The distribution of extracranial atherosclerotic burden in different quartiles of ISI 0-120 (D). The degree of intracranial and extracranial atherosclerotic burden was graded: 0 score, 1 score, >1 scores.