Association Between Hemoglobin Glycation Index and Coronary Artery Inflammation in Type 2 Diabetes Mellitus Based on CCTA

Summary

This study analyzed data from patients with type 2 diabetes mellitus to examine the association between hemoglobin glycation index (HGI) and coronary artery inflammation. Higher HGI levels were linked to increased perivascular inflammation and CT high-risk plaques, suggesting HGI as a potential marker for cardiovascular risk assessment.

Abstract

Hemoglobin glycation index (HGI) is used to quantify the biological variation of hemoglobin A1c (HbA1c). The present study explored the relationship between HGI and coronary artery inflammation in patients with type 2 diabetes mellitus (T2DM). A total of 360 diseased coronary arteries from 185 T2DM patients. Patients and their diseased vessels were categorized into three distinct groups according to the tertiles of HGI: H1 (low group), H2 (medium group), and H3 (high group). Clinical baseline data, perivascular fat attenuation index (FAI) within the proximal 40 mm of the three coronary arteries, quantitative plaque parameters, and the proportion of computed tomography (CT) high-risk plaque features were compared. Moreover, a multivariate logistic regression analysis was used to analyze the risk factors for CT high-risk plaques. The results showed that the perivascular FAI of the left anterior descending artery (LAD) and left circumflex artery (LCX), fibrous, and lipid component volumes, as well as their respective volume ratios, increased with an increase in HGI (P<0.05). The prevalence of CT high-risk plaques also increased with an increase in HGI (P<0.001). HGI (odds ratio (OR) = 1.764, 95% confidence interval (CI) = 1.363–2.284, P<0.001) and LAD-FAI (OR = 1.086, 95% CI = 1.032–1.143, P<0.05) were independent risk factors for CT high-risk plaques. In conclusion, HGI may serve as a potential complementary biomarker for coronary inflammation and plaque vulnerability in patients with T2DM, and warrants further validation in larger prospective studies.

Introduction

Atherosclerotic cardiovascular disease (ASCVD) remains the primary global cause of mortality and disability, with older adults (≥60 years) representing a vulnerable population at elevated cardiovascular risk¹. Type 2 diabetes mellitus (T2DM), which ranks as one of the most prevalent chronic conditions worldwide, is linked to various complications. Among individuals with T2DM, coronary artery disease (CAD) represents the leading factor contributing to unfavorable outcomes². CAD is primarily caused by a chronic inflammatory process due to coronary atherosclerosis, lipid metabolism, and vascular dysfunction³. The chronic hyperglycemic state in patients with T2DM further exacerbates vascular inflammation. Vascular inflammation not only contributes to the formation of coronary atherosclerotic plaques but is also a key trigger of plaque rupture⁴. Therefore, quantifying coronary artery inflammation may enhance risk stratification of CAD patients. However, efforts to use biomarkers to evaluate the inflammatory process have been ongoing for years; conventional circulating inflammatory markers, such as high-sensitivity C-reactive protein and interleukin-6, lack specificity for vascular inflammation. In addition, while positron emission tomography/computed tomography (PET/CT) can monitor arterial wall inflammation, it cannot directly assess the inflammatory burden of coronary arteries⁵and is expensive and limited in clinical application.

Recently, the change in pericoronary adipose tissue (PCAT) attenuation assessed by coronary computed tomography angiography (CCTA) has emerged as a new sensitive biomarker for detecting coronary inflammation⁶. This new marker, known as the perivascular fat attenuation index (FAI), has been shown to significantly predict cardiovascular events⁴. Moreover, computed tomography (CT) high-risk plaque features serve as indirect markers of vascular inflammation and also denote the vulnerability and risk of rupture of plaques (rupture is primarily caused by inflammation)⁷, which can be identified concurrently using CCTA.

Hemoglobin A1c (HbA1c) is the gold standard in clinical practice for assessing glycemic control in diabetic patients, where good glycemic control not only helps reduce microvascular complications but is also independently associated with the occurrence of major cardiovascular events in patients with T2DM and multivascular CAD⁸. However, changes in HbA1c levels can only explain 60%–80% of the mean blood glucose (MBG) levels⁹, and differences in HbA1c may be influenced by individual variability in glucose metabolism, physiological characteristics, and genetic predisposition¹⁰. Hence, the measurement of HbA1c cannot fully reflect the status of glycemic metabolism.

The hemoglobin glycation index (HGI) is used to quantify the biological variation in HbA1c independent of mean blood glucose (MBG) levels. HGI can identify individuals whose HbA1c levels are higher or lower than average compared with others with similar blood glucose concentrations¹¹. Several studies have confirmed the association between HGI and diabetes complications, such as higher risks of diabetic nephropathy or retinopathy in subjects with high HGI^12,13, and T2DM patients. A higher HGI is associated with an increased risk of chronic vascular diseases^12,13.

Currently, the association between HGI and coronary artery inflammation in T2DM patients remains elusive. Therefore, this study aimed to investigate the association between HGI and perivascular FAI, various plaque component parameters, and CT high-risk plaque features based on CCTA, to improve risk stratification in T2DM patients with concomitant coronary atherosclerosis.

Protocol

This study was approved by the Institutional Review Board of the Affiliated Hospital of Nantong University (approval No. 2025-K179-01), and the requirement for informed consent was waived. The software and the equipment used are listed in the Table of Materials.

1. Study subjects and grouping

Between January 2016 and May 2023, data were collected for 752 patients with T2DM who underwent CCTA at the Affiliated Hospital of Nantong University. T2DM was diagnosed according to the Guidelines for the Prevention and Treatment of Type 2 Diabetes in China (2020 Edition). The exclusion criteria were as follows: (1) poor CCTA image quality; (2) congenital coronary anomalies, coronary stent implantation, or post-coronary artery bypass surgery; (3) pregnancy or hemoglobinopathy; (4) cardiogenic diseases other than coronary heart disease; (5) incomplete baseline data. A total of 185 subjects were included in the final analysis. A total of 360 diseased coronary arteries were analyzed, among which the major branches with coronary plaques included the left main (LM), left anterior descending (LAD), left circumflex artery (LCX), and right coronary artery (RCA). Clinical baseline data, including fasting plasma glucose (FPG), HbA1c, total cholesterol (TC), triglycerides (TG), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), and uric acid (UA), were collected within 48 h before and after the CCTA examination.

Since HGI is a linear regression residual, the predicted HbA1c can be generated by inputting FPG into the univariate linear regression equation between HbA1c and FPG in the study population. HGI is defined as the difference between the actual HbA1c and the predicted HbA1c. Patients and their affected coronary vessels were stratified into three groups according to HGI tertiles: H1 (low group) ≤-0.91, -0.91< H2 (medium group) <0.49, H3 (high group) ≥0.49. The study design and patient selection criteria are shown in Figure 1.

2. CCTA examination

CCTA examinations were performed using a 256-slice computed tomography (CT) scanner with the following protocol: detector collimation of 160 mm × 0.5 mm, gantry rotation speed of 0.28 s/rotation, slice thickness of 0.625 mm, tube current-time product of 350–635 mAs, and tube voltage of 100–120 kVp. All patients were scanned using retrospective electrocardiogram-gated technology. The scanning was performed from 1 cm inferior to the tracheal bifurcation to the diaphragmatic surface of the heart.

Contrast-enhanced imaging was performed following the intravenous administration of a non-ionic contrast agent into the brachial vein at a flow rate of 5 mL/s, immediately followed by a 20 mL saline flush delivered at the same rate. The dosage of the contrast agent was 0.8 mL per kilogram of body weight. For enhanced scans, contrast agent tracking triggering technology was employed, automatically detecting the CT value of the aortic root, with a triggering threshold of 120 Hounsfield units (HU).

3. Analysis of FAI

The original CCTA images were uploaded to the Deepwise Vascular Analysis Software (DW-CACTAS, Deepwise Corporation, China) by a radiologist with over three years of experience, who was blind to the patient’s clinical and CT information. The software automatically delineates the PCAT around the proximal 40 mm of LAD and LCX and 10–50 mm of RCA and allows for manual adjustments to the automatic depiction of the coronary artery walls, subsequently calculating the FAI for these three coronary arteries. PCAT was characterized as the adipose tissue situated radially at a distance equal to the coronary vessel’s diameter from its outer wall, setting its threshold value between -190 and -30 HU¹⁴. FAI was defined as the mean CT attenuation value of PCAT (Figure 2).

4. Quantitative analysis of CCTA plaque

Coronary atherosclerotic plaques refer to a type of tissue structure existing within or adjacent to the coronary artery lumen, with an area exceeding 1 mm², distinct from pericardial tissue, epicardial fat tissue, and the vascular lumen¹⁵. CCTA images were transferred to a commercial workstation for plaque quantitative analysis. In CCTA images, the CT threshold values for different plaque components were as follows: lipid component CT value was set at -100–30 HU, fibrous component CT value at 31–130 HU, and calcified component CT value at 131–300 HU¹⁶, and the total plaque volume of each diseased coronary artery was measured.

The four CT vulnerability characteristics of high-risk coronary artery plaques were positive remodeling, low attenuation plaque, spotty calcification, and the napkin-ring sign. Coronary artery plaques that exhibited two or more of the above characteristics were defined as CT high-risk plaques¹⁷.

5. Statistical analysis

Categorical data were expressed as frequency and percentage [%], analyzed using the chi-square test, and further compared between groups using the row-by-column chi-square partition (corrected P = 0.0167). The Kolmogorov-Smirnov test was used to examine quantitative data normality. Normally distributed quantitative data were presented as mean ± standard deviation (SD), and statistical differences were assessed using analysis of variance (ANOVA). Quantitative data that were not normally distributed are presented as the median and interquartile range [M (Q1, Q3)], with the Kruskal-Wallis test applied for comparisons across multiple groups. Univariate and multivariate binary logistic regression analyses were conducted to identify independent risk factors associated with CT high-risk plaques. Statistical analysis was conducted using SPSS 26.0 software. A P-value of <0.05 was considered statistically significant.

6. Linear regression analysis of HGI with perivascular FAI

Multivariable linear regression was performed to assess the association between HGI (continuous) and perivascular FAI (LAD-FAI and LCX-FAI), adjusting for the same covariates as in the primary analysis.

7. Sensitivity analysis using HGI quartiles

To assess the robustness of the findings, HGI was categorized into quartiles, and its association with CT high-risk plaques was re-evaluated using multivariable logistic regression, adjusting for the same covariates as in the primary analysis.

8. Subgroup and interaction analyses

Subgroup analyses were performed according to age, gender, and glycemic control level using multivariable logistic regression, adjusting for the same covariates as in the primary analysis except the stratification variable itself. Interaction terms between HGI and each stratification variable were tested to assess effect modification.

9. Restricted cubic spline analysis

To explore potential dose-response and nonlinear relationships between HGI and the outcomes, restricted cubic spline (RCS) analyses were performed with HGI modeled as a continuous variable. For CT high-risk plaque, RCS logistic regression models were fitted, whereas for LAD-FAI, RCS linear regression models were used. All models were adjusted for the same covariates as in the primary analysis. The overall association and the nonlinear component were assessed using the Wald test.

Results

Association between HbA1c and FPG

A linear relationship was established between patients’ HbA1c and FPG measurements, deriving the regression equation: predicted HbA1c = 0.28 × FPG + 5.75, with r = 0.508 and P < 0.001 (Figure 3).

Comparison of clinical baseline data and FAI among different groups

HGI values of groups H1, H2, and H3 were -1.29 (-1.71, -1.0...

Discussion

The present study found that the volume and proportion of calcified and non-calcified plaques increased with an increase in HGI levels. Concurrently, the occurrence and proportion of CT high-risk features (low-attenuation plaques) in coronary artery atherosclerotic plaques also increased with an increase in HGI levels. Measurements of FAI values in the three main coronary arteries showed that LAD-FAI and LCX-FAI values also increased with an increase in HGI levels. After adjusting for other covariates, HGI and LAD-FAI we...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Commercial Workstation	Shukun Technology	AI-assisted diagnostic system for coronary CT imaging	Used for quantitative analysis of coronary atherosclerotic plaques.
Computed Tomography System	GE Healthcare	Revolution CT (256-slice)	Used for coronary computed tomography angiography (CCTA) examinations. Scanning parameters: detector collimation of 160 x 0.5 mm, gantry rotation speed of 0.28 second/rotation, slice thickness of 0.625 mm.
Non-ionic Contrast Agent	Bayer Health-care	Ultravist 370	Intravenous contrast agent. Dosage was 0.8 mL per kg of body weight, injection rate of 5 mL/second, followed by a 20 mL saline flush at the same rate.
Statistical Software	IBM	SPSS 26.0	Used for all statistical analyses, including ANOVA, chi-square tests, and logistic regression analysis.
Vascular Analysis Software	Deepwise Corporation	DW-CACTAS	Used to automatically delineate pericoronary adipose tissue and calculate the fat attenuation index (FAI). Operated by a radiologist blinded to patient information.