Association between triglyceride glucose index and atherosclerotic plaques and Burden: findings from a community-based study

The PolyvasculaR Evaluation for Cognitive Impairment and vaScular Events (PRECISE) study recruited 3067 participants aged 50 to 75 from a cluster sample of 6 villages and 4 communities in Lishui city between May 2017 and September 2019. Details of the study have been published elsewhere[11]. Participants with contraindications to computed tomography angiography (CTA) and magnetic resonance imaging (MRI), mental illness, and life expectancy of less than 4 years were excluded. All participants were comprehensively evaluated for stenosis and plaques in coronary, intra- and extracranial arteries with advanced vascular imaging techniques. The Ethics Committee approved the study protocol of Beijing Tiantan Hospital (IRB Approval No. Ky2017-010-01) and Lishui Hospital (IRB Approval No. 2016-42), and all participants have signed informed consents before enrollment. Participants with self-reported diabetes previously diagnosed by a physician or using of glucose-lowering or insulin medications were excluded from this analysis. Non-diabetic participants who underwent CTA examination, MRI examination, and blood sampling were included in the analysis.

All anthropometric, clinical and socio-demographic data were collected face to face by trained interviewers following the standard protocols, which included age, sex, education level, body mass index (BMI), smoking, drinking, relevant medical history, and self-reported medication use. BMI was calculated as weighted (kg) divided by the square of height (m²) [6]. Hypertension was defined as hypertension previously diagnosed by physicians or current use of antihypertensive agents or systolic blood pressure ≥ 140 mmHg or diastolic blood pressure ≥ 90 mmHg [12]. History of dyslipidemia was determined by self-reported dyslipidemia previously diagnosed by a physician usually according to Chinese guidelines for the prevention and treatment of adult dyslipidemia[13] or current use of lipid-lowering drugs (e.g., statins, fibrates).

Venous blood samples were collected in the morning after fasting for at least 8 h at night and transported to Tiantan Hospital biobank for preservation by the cold chain. Biological measurements, including fasting glucose, fasting insulin and triglyceride, were validated and measured in the Central Laboratory of Beijing Tiantan Hospital under strict quality control standards. Insulin resistance was assessed by using HOMA-IR model and the TyG index. The TyG index was determined as ln (fasting triglycerides [mg/dL] × fasting glucose [mg/dL]/2)] [8] and the HOMA-IR was calculated as (fasting insulin [µU/ mL] × fasting glucose [mmol/L])/22.5 [6]. Participants were categorized into 4 concordance/discordance groups: high TyG index/ high HOMA-IR, high TyG index/ low HOMA-IR, low TyG index/ high HOMA-IR, low TyG index/ low HOMA-IR [14]. A TyG index equal to or greater than the median and less than the median were defined as high TyG index and low TyG index, respectively, and similar definition was applied to HOMA-IR. The high TyG index /low HOMA-IR and high HOMA-IR/ low TyG index groups were considered as discordance.

The CTA was performed on a dual-source CT scanner (SOMATOM Force, Siemens Healthineers, Forchheim, Germany) to examine coronary atherosclerosis burden. The presence of coronary plaque was defined as a structure > 1mm² within and/or adjacent to the coronary artery lumen, which could be clearly distinguished from the vessel lumen and the surrounding pericardial/myocardial tissue [15]. The evaluated coronary artery segments included the left main trunk, anterior descending branch, diagonal branch, circumflex branch, obtuse margin branch, middle branch, right coronary artery, posterior descending branch and posterior lateral branch. Coronary stenosis was quantified by visual estimation, and each artery segments graded as following: score 0, minimal (< 30%); score 1, mild (30–49%); score 2, moderate (50–69%); score 3, severe (> 70%) [16]. The segment stenosis score (SSS) was calculated as a measure of the overall coronary artery plaque burden, and was graded and summed for each coronary segment as none to severe plaque (0–3) based on the extent of obstruction (total score ranging: 0 to 27) [15]. The SSS was further categorized into three grades (grade 0, 0 score; grade 1, 1–3 scores; grade 2, ≥ 4 scores). We calculated the Cohens κ to evaluated interobserver agreement for the presence of plaque and artery stenosis (Cohen κ = 0.96 and 0.92).

The 3.0T scanner (Ingenia 3.0T, Philips, Best, The Netherlands) was used to detect intra-/extra-cranial atherosclerotic plaques and stenosis. 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). The presence of intra- and extracranial atherosclerosis plaques was defined as eccentric wall thickening with or without luminal stenosis identified on the 3D T1w VWI images in comparison to the corresponding location on the 3D TOF MR angiogram at MR angiography [17, 18]. The evaluated intracranial arterial segments included the bilateral internal carotid, anterior cerebral arteries (A1, A2), the middle cerebral artery (M1, M2), posterior cerebral arteries (P1, P2), and basal vertebral arteries (V4). The extracranial arterial segments included the common carotid, proximal internal carotid artery, and vertebral artery (V1, V2, V3) [17]. The Warfarin-Aspirin Symptomatic Intracranial Disease Trial (WASIDT) [19] was used to assess the luminal stenosis in the intracranial arteries, and the North American Symptomatic Carotid Endarterectomy Trial (NASCET) [20] was used to measure the luminal stenosis in the extracranial arteries. 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) [21]. The intracranial atherosclerotic burden was calculated through summing the intracranial atherosclerotic scores of intracranial arterial segments, which were categorized into four grades of 0, 1, 2–3, and ≥ 4 scores [21]. Similarly, the extracranial atherosclerotic burden was counted through extracranial arterial segments, and categorized into four grades of 0, 1, 2–3, and ≥ 4 scores[21]. We calculated the Cohens κ to assess the interrater reliability for the presence of plaques and artery stenosis ≥ 50% in intra- and extracranial artery (Cohen κ = 0.97 and 0.79 for intracranial artery and Cohen κ = 0.94 and 0.86 for extracranial artery).

The characteristics of the participants were displayed according to the tertiles of insulin resistance index. Data are presented as mean ± standard deviation (SD) or median [interquartile range (IQR)] for continuous variables and as the frequency (%) for categorical variables. We used ANOVA or Kruskal-Wallis test for continuous variables and χ2 test or Fisher’s exact for categorical variables. The association of TyG index or HOMA-IR with the presence of atherosclerotic plaques was analyzed by binary logistic regressions and with the atherosclerotic burden by ordinal logistic regression models. The odds ratio (OR) or common odds ratio (cOR) with 95% confidence intervals (CI) were estimated. Logistic regression models were also applied to estimate odds for atherosclerosis associated with concordance/discordance groups. We used 2 models: model 1, unadjusted; model 2, adjusted for age, sex, current smoking, current drinking, BMI, estimated glomerular filtration rate, low-density lipoprotein levels, medical history of transient ischemic attack, stroke, coronary artery disease, hypertension, dyslipidemia, and use of antihypertensive and antiplatelet medications. Restricted cubic splines with 5 knots were constructed to model a non-linear relationship between HOMA-IR or TyG and the presence of atherosclerotic plaques. The area under the receiver operating characteristic curve (AUC), the integrated discrimination improvement (IDI), and the net reclassification improvement (NRI) were determined to assess the predictive performance of insulin resistance index [7].

Two-tailed P value < 0.05 was considered statistically significant, and all analyses were performed using SAS version 9.4 (SAS Institute Inc, Cary, NC).