Association between the Plasma-Glycosylated Hemoglobin A1c/High-Density Lipoprotein Cholesterol Ratio and Carotid Atherosclerosis: A Retrospective Study

Background Diabetes mellitus (DM) and dyslipidemia are the main risk factors for atherosclerosis. Elevated glycosylated hemoglobin A1c (HbA1c) and reduced high-density lipoprotein cholesterol (HDL-C) are associated with the progression of atherosclerosis. The aim of this study is at exploring the relationship between the HbA1c/HDL-C ratio and atherosclerosis evaluated using carotid artery intima-media thickness (cIMT) and carotid artery plaque. Methods In this retrospective study, we enrolled 1304 patients who had multiple cardiovascular risk factors or symptoms of suspected coronary artery disease. cIMT and carotid artery plaque were measured using ultrasonography. Logistic regression was used to explore the correlation between the HbA1c/HDL-C ratio and cIMT or carotid artery plaque. We used restricted cubic spline curves to assess nonlinear relationships between the HbA1c/HDL-C ratio and cIMT or carotid artery plaque. Results With increased quartiles of HbA1c/HDL-C, patients had higher cIMT and a greater carotid plaque burden. After adjusting for other relevant clinical covariates, patients with the highest HbA1c/HDL-C ratio (quartile 4 (Q4)) had a 2.88 times (95% confidence interval (CI): 2.02–4.10, P < 0.001) more abnormal mean cIMT, 3.72 times (95% CI: 2.55–5.44, P < 0.001) more abnormal maximum cIMT, and 2.58 times (95% CI: 1.70–3.91, P < 0.001) greater carotid artery plaque burden compared with patients who had the lowest HbA1c/HDL-C ratio (Q1). Moreover, the association of HbA1c/HDL-C with atherosclerosis remained significant in a subsample of patients with and without DM. Conclusion As a novel compound indicator for evaluating blood glucose homeostasis and dyslipidemia, the HbA1c/HDL-C ratio was positively correlated with carotid atherosclerosis evaluated using the mean and maximum cIMT as well as the carotid artery plaque burden.


Introduction
Atherosclerosis, a multifactorial and systemic disease process [1,2], can lead to cardiovascular diseases (CVD) such as coronary artery disease, ischemic stroke, and peripheral artery disease. Carotid artery intima-media thickness (cIMT) and carotid artery plaque are simple, noninvasive, and clinically accessible ultrasound markers that have been widely regarded as a surrogate for carotid atherosclerosis and a strong predictor of future CVD [3][4][5][6]. Therefore, the influence of risk factors on the development of carotid atherosclerosis can be measured by cIMT and carotid artery plaque.
The risk factors contributing to atherosclerosis are diverse, such as diabetes mellitus (DM) and dyslipidemia [7]. The relationship between DM and atherosclerosis has been widely established in clinical research and molecular research [8,9]. cIMT gradually increases during the deterioration of glucose homeostasis from normal glucose tolerance to DM [10]. Glycosylated hemoglobin A1c (HbA1c) plays an important role in monitoring glucose control, even in patients without DM [11,12]. Many clinical studies have described the positive relationship between HbA1c and cIMT, emphasizing the recognition of atherosclerosis development at an early stage using blood glucose monitoring [13][14][15]. As for dyslipidemia, in the past few decades, clinical studies and meta-analyses have consistently reported that high-density lipoprotein cholesterol (HDL-c) levels are negatively correlated with the risk of atherosclerotic CVD [16,17]. The role of HDL-C in lipid metabolism (promoting reverse cholesterol transport), inflammation (tempering inflammatory mediator), and antithrombosis determines the close connection between low HDL-C and progression of atherosclerosis [18,19]. However, the metabolic processes of blood glucose and lipids interact with each other. A post hoc analysis in the Stop Atherosclerosis in Native Diabetics Study (SANDS) suggested a negative correlation between HbA1c and the achievement of blood lipid goals [20]. The presence of DM eliminates the vascular protective effect of HDL-C, and poor blood glucose control is an independent risk factor for low HDL-C in DM [21,22]. In turn, HDL-C can reduce blood glucose in patients with DM by increasing plasma insulin [23][24][25].
As a single index of DM or lipids for the evaluation of atherosclerosis status may be insufficient, comprehensively considering the effects of both blood glucose control and HDL-C can avoid the inadequacy of a single index. The HbA1c/HDL-C ratio is proposed as an easily available index related to atherosclerosis. To our knowledge, however, there is no research focusing on this comprehensive ratio and atherosclerosis. We hypothesized that the HbA1c/HDL-C ratio is related to subclinical atherosclerosis measured using the cIMT and carotid artery plaque via ultrasonography.

Study
Population. This retrospective study included 1654 consecutive patients with multiple cardiovascular risk factors or symptoms of suspected coronary artery disease in Guangdong Provincial People's Hospital from September 2014 to September 2015. We collected information on demographic features, health behaviors, medical history (stroke, hypertension, diabetes, and chronic kidney disease), laboratory tests, and carotid ultrasound examinations from the electronic medical records. We excluded 244 patients who had used statins within the past 3 months, 84 patients who had missing data of HbA1c or HDL-C, and 22 patients without information of carotid ultrasonography examination ( Figure 1).
This study was conducted in accordance with the Declaration of Helsinki. The study was approved by the Ethics Committee of Guangdong Provincial People's Hospital, and all patients provided their verbal informed consent.
2.2. Measurement. Carotid ultrasonography examination was performed manually by experienced sonographers, who were unaware of participants' baseline characteristics and laboratory results, using a GE Vivid E95 (GE Healthcare, Milwaukee, WI, USA) and 7.5-12 MHz phased array probe. Two sonographers measured cIMT and carotid artery plaque and determined the final results together. If there was any discrepancy, a third sonographer would be consulted and the results would be determined through discussion. Carotid atherosclerosis was defined as abnormal cIMT and/or presence of carotid artery plaque [4][5][6].
The region of interest for cIMT measurement was located at the far wall of the bilateral carotid arteries proximal to the bifurcation, with ≥10 mm of plaque-free lesions on each side. cIMT was defined as the distance between the interface of the lumen-intima and media-adventitia. The mean cIMT was taken as the average of the cIMT values of the left and right carotid arteries. Maximum cIMT was taken as the larger cIMT value between the left and right carotid arteries. We defined abnormal cIMT as mean cIMT or maximum cIMT values ≥ 1 mm [26,27]. We also assessed the carotid artery plaque at three different locations: the common carotid artery, carotid bifurcation, and internal carotid artery on both the left and right sides. Carotid artery plaque was defined as a focal structure that encroaches into the arterial lumen at least 0.5 mm or 50% of the surrounding cIMT value or demonstrates a thickness > 1:5 mm as measured from the media-adventitia interface to the intimalumen interface [28]. The carotid plaque score was calculated as the sum of the thickest plaque values on both the left and right sides.

Covariates.
Hypertension was diagnosed according to the European Society of Cardiology guidelines as systolic blood pressure ≥ 140 mmHg and/or diastolic blood pressure ≥ 90 mmHg, which is equivalent to a 24 h ambulatory blood pressure monitoring average of ≥130/80 mmHg or a home blood pressure monitoring average of ≥135/85 mmHg for two measurements over at least 3 days [29]. DM was diagnosed based on the presence of diabetes. If the patient had a negative history of diabetes, fasting blood glucose ≥ 7:0 mmol/L (126 mg/dL) and hemoglobin A1c ≥ 6:5% or positive oral glucose tolerance test (2 h plasma glucose ≥ 11:1 mmol/L (200 mg/dL)) or random plasma glucose ≥ 11:1 mmol/L (200 mg/dL) were adopted to define DM according to the European Society of Cardiology guidelines [30]. Smoking was defined as previous smoking, and alcohol consumption was defined as a previous drinking habit.

Statistical
Analysis. Descriptive statistics were conducted using Kruskal-Wallis H tests and chi-square or Fisher's exact tests for continuous and dichotomous variables, respectively. We explored the relationship between the HbA1c/HDL-C ratio and abnormal cIMT or carotid artery plaque using logistic regression. The regression results are reported according to HbA1c/HDL-C ratio quartiles (as a categorical variable) using the lowest quartile (Q1) as reference. Odds ratios (ORs) with 95% confidence intervals (CIs) were calculated for the associations. We adjusted personal characteristics and healthy behaviors (age, sex, smoking, alcohol consumption, and history of hypertension) in different models to further evaluate the relationship. LDL-C, white blood cell counts, and creatinine levels were also adjusted in another model, because lipid levels, chronic low-grade inflammatory responses, and renal function affect glycometabolism and atherosclerosis. We used restricted cubic spline curves to visually assess the relationships between the HbA1c/HDL-C ratio (as a continuous variable) and abnormal cIMT or carotid artery plaque. ORs and 95% CIs were derived from restricted cubic spline regression, with knots placed at the 25th, 50th, 75th, and 95th percentiles of the distribution of the HbA1c/HDL-C ratio. The reference value of the HbA1c/HDL-C ratio was the median of the reference group.
The proportion of missing data in the sample for analysis was not greater than 5%; missing data were interpolated using the mean value. Associations with P < 0:05 (two sided) were considered to be statistically significant. All of the analyses were performed using Stata 15.0 (StataCorp LLC, College Station, TX, USA), R version 3.6.1 (The R Project for Statistical Computing, Vienna, Austria), and EmpowerStats (X&Y Solutions Inc., Boston, MA, USA).

Patient Population.
A total of 1304 patients with complete data were included in this retrospective study (including 897 men and 407 women). The mean age (standard deviation) was 63 (11) years. The median HbA1c/HDL-C ratio was 5.85 (interquartile range (IQR) 4.78-7.30). Baseline information of the study population is shown in Table 1. According to quartiles of the HbA1c/HDL-C ratio, patients were divided into four groups (Q1-Q4). Patients with the highest HbA1c/HDL-C ratio (Q4) mostly comprised men and had higher values of HbA1c, fasting blood glucose, cIMT, and prevalence of carotid artery plaque, as well as lower HDL-C and ApoA. The HbA1c/HDL-C ratio was positively correlated with hypertension and white blood cells. Baseline information of the study population grouped by the presence or absence of DM is shown in Supplemental Table 1 and Supplemental Table 2.

HbA1c/HDL-C Ratio and cIMT or Carotid Artery
Plaque. As shown in Figure 2, the mean cIMT and maximum cIMT showed an upward trend with an increased HbA1c/HDL-C ratio (P < 0:001). Table 2 presents results for the relationship between the HbA1c/HDL-C ratio and abnormal cIMT or carotid artery plaque. In the unadjusted model, the OR (95% CI) for quartile 4 (Q4) of the HbA1c/HDL-C ratio was 2.55 (1.86-3.51) for abnormal mean cIMT, 3

Subgroup Analysis.
Subgroup analysis showed that the above covariates (age, sex, smoking, alcohol consumption, and hypertension history) did not change the relationship between the HbA1c/HDL-C ratio and abnormal cIMT or carotid artery plaque (P for interaction > 0:05; Supplemental Table 3).

Discussion
In the present study, we proposed a composite indicator of blood glucose homeostasis and dyslipidemia, the HbA1c/HDL-C ratio, and demonstrated the relationship between this ratio and subclinical carotid atherosclerosis. We observed a positive correlation between HbA1c/HDL-C levels and abnormal mean/maximum cIMT as well as carotid artery plaque burden in patients with a high risk of CVD, independent of traditional clinical risk factors. More importantly, the relationship between HbA1c/HDL-C and abnormal cIMT remained in a subsample of patients with and without DM. HbA1c/HDL-C may be an important biomarker for clinical evaluation of atherosclerosis progression. This article was previously reported in the Preprint platform (2021, Research Square) [31].
Atherosclerosis is a slow, gradually progressing disease that begins in childhood, but there are potential risk factors that can accelerate its development such as poor glycemic control and low HDL-C [17,32]. DM patients showed increased arterial stiffness and degenerated artery compliance, which worsened carotid atherosclerosis [33]. Amelioration of postprandial hyperglycemia in patients with type 2 DM could prevent the progression of carotid atherosclerosis, which indicates the importance of blood glucose homeostasis [34]. The extent of glycemic control, sensitively reflected by HbA1c, is related to CVD complications in patients with DM [35][36][37]. Observational studies have shown that HbA1c levels are significantly related to cIMT and carotid plaques, and long-term stable HbA1c levels are of great significance for preventing subclinical coronary atherosclerosis in patients with type 2 DM [13,38]. Moreover, using HbA1c as an evaluation indicator, a randomized controlled trial compared intensive blood glucose control with conventional treatment, thereby proving that intensive blood glucose control can reduce the progression of CVD in DM patients [39]. Low HDL-C is an important manifestation of dyslipidemia. Clinical studies have consistently shown that low HDL-C is an independent risk factor for atherosclerotic CVD [17,40]. A meta-analysis involving 21,000  Journal of Diabetes Research people concluded that low HDL-C levels is associated with elevated cIMT, independent of LDL-C levels. Even in patients who are treated aggressively with statins to decrease LDL-C, HDL-C is still a major predictor of CVD development [41]. In addition, a study synthesizing three cohorts has verified that ApoA levels cannot provide additional predictive information over HDL-C for CVD, highlighting the important and independent role of HDL-C [42]. Therefore, HbA1c and HDL-C are useful indicators that reflect DM and dyslipidemia, respectively, which are closely related to the progression of atherosclerosis. Our study puts forth a new marker, the HbA1c/HDL-C ratio, which was found to be significantly associated with abnormal cIMT and carotid artery plaque after adjustment for other relevant clinical covariates. The present findings were consistent with those of previous studies on HbA1c and HDL-C [38][39][40][41]. Furthermore, the HbA1c/HDL-C ratio takes into account the relative state of blood glucose control and lipid levels. In recent years, studies have found that blood glucose homeostasis has an impact on blood lipid metabolism and vice versa. The presence of DM eliminated the protective effect of HDL-C on the progression of atherosclerosis, and furthermore, HbA1c was found to be negatively correlated with the efficacy of lipid-lowering therapy [20,22]. In contrast, HDL-C has potent antidiabetic properties via increasing insulin sensitivity and β-cell function, reflecting its role in regulating blood glucose homeostasis [23][24][25]. In this study, the HbA1c/HDL-C ratio marker could comprehensively reflect blood glucose homeostasis and dyslipidemia, which can be used to monitor progression of atherosclerosis. The mechanisms of deterioration of blood glucose homeostasis and dyslipidemia leading to atherosclerosis are multifaceted. Hyperglycemia increases inflammatory response and mitochondrial oxidative stress, directly resulting in vascular endothelial dysfunction [9]. A reduction in HDL-C activates local inflammation and endothelial  Data were shown as OR (95% CI) which was evaluated using logistic regression models. † Model 1: unadjusted; model 2: adjusted for age and sex; model 3: further adjusted for smoking, alcohol consumption, and history of hypertension; model 4: further adjusted for LDL-C, white blood cell count, and creatinine level; HbA1c: hemoglobin Alc; HDL-C: high-density lipoprotein cholesterol; cIMT: carotid intima-media thickness; Ref: reference; LDL-C: lowdensity lipoprotein cholesterol; CI: confidence interval. 5 Journal of Diabetes Research thrombosis, increases endothelial cell apoptosis, and slows down vascular repair [43]. Inflammation caused by hyperglycemia and low HDL-C mediates the immune system disorder, disrupting the immune balance [44]. With the double attack of increasing HbA1c and decreasing HDL-C, the progression of atherosclerosis will rapidly accelerate. From this point of view, the HbA1c/HDL-C ratio could reflect the relative status of blood glucose and serum lipids as well as the extent of damage to the arterial endothelium and inflammation, which in turn can explain the rationality of the HbA1c/HDL-C ratio as a marker of atherosclerosis progression through the above pathophysiological mechanism. This shows that comprehensive management of comorbidities of glycometabolic disorder and dyslipidemia is an effective way to prevent atherosclerosis [45].
Another finding of our study was that in a subsample with and without DM, the HbA1c/HDL-C ratio was still associated with abnormal cIMT. Blood glucose homeostasis as reflected by HbA1c is not only applicable in the population with DM but also useful in those without diabetes. A recent cohort study showed that the use of HbA1c can better identify subclinical atherosclerosis in individuals without DM, on the top of traditional CVD prediction models [12]. Bobbert  (h) abnormal maximum cIMT in patients without DM; (i) carotid artery plaque in patients without DM. Odds ratios and 95% confidence intervals (CIs) were adjusted for the same variables as in model 4 ( Table 2). The reference value of the HbA1c/HDL-C ratio is the median of the reference group. The area between the two dotted lines represents the 95% CI. Owing to the small sample and large 95% CI, the highest 4% of patients are not shown in the figure. HbA1c: hemoglobin Alc; HDL-C: high-density lipoprotein cholesterol; cIMT: carotid intima-media thickness; DM: diabetes mellitus. 6 Journal of Diabetes Research had a stronger correlation with cIMT than fasting or 2 h glucose levels [46]. Because blood glucose homeostasis in people without DM will also affect subclinical atherosclerosis, HbA1c levels also require attention. Therefore, the application of the proposed new marker, the HbA1c/HDL-C ratio, for individuals without DM may have an important role in primary prevention among this population. The present research has several limitations. First, this was a single-center study; our findings should be extrapo-lated carefully to other populations with different demographic characteristics. Second, the cross-sectional study design prevented us from prospectively identifying a causal relationship between the long-term HbA1c/HDL-C ratio and development of atherosclerosis. Third, in patients with DM, the dose-response relationship between the HbA1c/HDL-C ratio and carotid plaque score showed that with increased HbA1c/HDL-C ratio, total carotid artery plaque thickness was increased. Although there was no  7 Journal of Diabetes Research significant correlation between the HbA1c/HDL-C ratio and the presence of carotid artery plaque when the HbA1c/HDL-C ratio was at high level, these still tended to be positively correlated. The reason may be an insufficient sample size in both the DM subsample and quartile groups.

Conclusions
As a new indicator for evaluating blood glucose homeostasis and dyslipidemia, the elevated HbA1c/HDL-C ratio was found to be associated with the mean and the maximum cIMT, as well as the occurrence of carotid artery plaque, independent of conventional risk factors. These associations remained in patients with and without DM. Our study findings suggest that the HbA1c/HDL-C ratio may be useful for monitoring high-risk patients in the primary prevention of atherosclerosis.

Data Availability
The data analyzed in this study can be obtained from the corresponding author with a reasonable request.

Conflicts of Interest
The authors declare no conflicts of interest related to this article.