Haemoglobin A1c variability is independently associated with carotid intima media thickness and plaque in type 2 diabetes: a retrospective, cross-sectional study

Abstract

Haemoglobin A1c (HbA1c) variability, a metric reflecting long-term glycaemic fluctuation, is associated with macrovascular events in type 2 diabetes. We aimed to investigate the impacts of HbA1c variability on preclinical atherosclerosis in patients without prior cardiovascular disease. We conducted a cross-sectional study on 564 participants with diabetes who underwent general health checkups from 2016–2022. At least three HbA1c measurements were conducted for each patient. Carotid intima-media thickness (CIMT) and plaque were evaluated by B-mode ultrasonography on bilateral carotid common arteries. The standard deviation (SD) and coefficient of variance (CV) of HbA1c were calculated. We found that each doubling in CV-HbA1c and SD-HbA1c was associated with a significant increment in CIMT. The effects were more pronounced in the groups with higher mean-HbA1c (mean-HbA1c ≥6.5%). The odds ratio (95% confidence interval) for the carotid plaque was 2.68 (1.57, 4.56) and 2.88 (1.16, 5.13) in the third tertile of CV-HbA1c and SD-HbA1c respectively after fully adjusting for all the conventional risk factors in the multivariable logistic regression analysis. High mean-HbA1c plus the third tertile of HbA1c variability sharply increased the prevalence of carotid plaques. In conclusion, HbA1c variability was independently associated with CIMT and plaques in populations with diabetes. CV-HbA1c and SD-HbA1c had more effects on subclinical atherosclerosis in patients with poorly-controlled blood glucose levels.

THE CAROTID INTIMA-MEDIA THICKNESS (CIMT) and carotid plaque are widely considered surrogate markers for asymptomatic and preclinical atherosclerosis [1]. CIMT is associated with regional alteration in shear stress and tensile stress, which often happens in the very early stages of atherosclerosis. A carotid plaque often represents the later stage of atherogenesis [2]. Both CIMT and carotid plaque are widely used as predictors of future atherosclerotic cardiovascular disease (CVD), especially in patients with diabetes [3, 4]. Therefore, ultrasound imaging of the carotid artery is recommended for populations with diabetes to evaluate the conditions of the artery and assess the risk of future CVD [5].

Several factors contribute to the progression of atherosclerosis. Glycaemic fluctuation is an independent risk factor for macrovascular complications in patients with diabetes [6-8]. Some studies have attempted to explore the impact of glycaemic fluctuation on CIMT [9, 10] but the results were contradictory.

Haemoglobin A1c (HbA1c) is the recommended golden standard for glucose control over decades. It can reflect chronic glucose exposure over the past three months [11]. Recently, HbA1c variability was also found to be associated with all-cause mortality and CVD outcomes in several studies [12-15], independently of HbA1c. Visit-to-visit HbA1c variability, usually calculated as intraindividual standard deviation (SD) or the coefficient of variance (CV) of HbA1c, reflects long-term variability in glucose homoeostasis [16]. However, the potential mechanism underlying the association between HbA1c variability and macrovascular complications remains unknown. A few studies have investigated the relationship between HbA1c variability and preclinical atherosclerotic vascular metrics.

Therefore, our study aimed to investigate the association between HbA1c variability and CIMT or carotid plaque in populations with type 2 diabetes, who are not diagnosed with prior CVD. We further estimated the effect of mean HbA1c on this association.

Materials and Methods

Study participants

Study subjects were recruited from the populations who underwent physical examination in Zhongda Hospital from 2016 to 2022. The inclusion criteria were as follows: 1. diagnosed with type 2 diabetes; 2. at least three measurements of HbA1c with intervals less than one year; 3. had complete data of carotid ultrasound examination. Exclusion criteria were as follows: 1. history of CVD including myocardial infarction, stroke, and coronary heart disease affirmed by radiography; 2. severe liver or renal dysfunction or malignancy; 3. missing data. The study protocol was approved by the ethics committee of Southeast University. Written informed consent was obtained from each participant.

Clinical and laboratory variables

Anthropometric parameters were collected, including heart rate, blood pressure, height, and weight. Blood samples were collected after overnight fasting. Fasting plasma glucose (FPG), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), total cholesterol (TC), and triglycerides were measured using AU5821 automatic analyzer (Beckman Coulter Corporation, USA). Levels of HbA1c were measured by high-performance liquid chromatography with a D-10 haemoglobin testing program (Bio-Rad, USA). The intra-assay and inter-assay CVs were <3%. Body mass index (BMI) was calculated as weight (kg) divided by the square of height (m). Triglyceride glucose index (Tyg) [17] was calculated as ln [triglyceride (mg/dL) × FPG (mg/dL)/2]. The estimated glomerular filtration rate (eGFR) was calculated with the Modification of Diet in Renal Disease formula [18]. The mean value and SD of HbA1c were calculated individually. CV of HbA1c (CV-HbA1c) was calculated as SD-HbA1c divided by mean-HbA1c. Both SD-HbA1c and CV-HbA1c were used as the indicators of HbA1c variability in this study. Type 2 diabetes was defined as (1) self-reported type 2 diabetes; (2) FPG ≥7.0 mmol/L, and/or HbA1c ≥6.5%. Dyslipidemia was defined as meeting any one of the following criteria: TC ≥6.2 mmol/L, LDL-C ≥4.1 mmol/L, HDL-C <1 mmol/L, triglycerides ≥2.3 mmol/L. Hypertension was defined as blood pressure ≥140/90 mmHg or concomitant use of anti-hypertensive medications. A questionnaire was collected from each participant concerning, smoking habits, duration of diabetes, history of chronic disease, and concomitant use of medication.

Carotid artery measurement

Carotid ultrasound images were taken from the left and right common carotid arteries by two professional medical doctors using real-time B-mode ultrasonography with a 10-MHz linear transducer (XH40, SAMSUNG, Korea). CIMT was measured at 1.0 cm proximal to the dilatation of the carotid bulb and was defined as the distance between the lumen-intima inter-face and media-adventitia interface. The value of CIMT analyzed in our research was the mean CIMT value of both left and right carotid arteries. A plaque was defined as a CIMT value greater than 1.5 mm or a focal intimal medial thickening of greater than 50% of the surrounding area [5]. When plaque existed, its thickness was considered as a part of CIMT according to the guidelines of the American Society of Echocardiography [5], because the thickness of the plaque may also provide important information. The interobserver correlation coefficient was 0.81 (p < 0.01), reflecting good consistency.

Statistical analysis

Normality was tested using the Kolmogorov–Smirnov test. All variables were expressed as mean ± SD or percentages. T-test or ANOVA was used to compare differences between groups. Correlations between HbA1c variability and CIMT were evaluated using the Spearman correlation coefficient. Since CV-HbA1c and SD-HbA1c were not normally distributed, log₂-transformed values were used in analyses. Multivariable regression was used to examine the relationship between HbA1c variability and CIMT. Age and gender were adjusted in model 1; hypertension, BMI, dyslipidemia, Tyg and eGFR were additionally in model 2; mean HbA1c level was adjusted in model 3 in addition to those in model 2; model 4 was additionally adjusted for smoking habits, duration of diabetes, medication (including insulin and oral drugs). HbA1c variability was both investigated as continuous and categorical variables using multivariable logistic regression to evaluate its relationship with carotid plaque. The results were presented as odds ratios (ORs) and the corresponding 95% confidence interval (95% CI). Subgroup analysis stratified by traditional CVD risk factors according to previous studies [19] was also conducted. Statistical significance was considered at p-value <0.05. Analyses were performed using the SPSS 17.0 software (SPSS, Chicago, USA).

Results

564 participants were finally enrolled into the study. The average number of HbA1c measurements per participant was 4.7 ± 1.1. Flow diagram of the population is represented in Fig. 1.

Fig. 1

Flow chart of cohort participants

The baseline characteristics according to CV-HbA1c tertiles are described in Table 1. The group with higher HbA1c variability was more likely to comprise men and had higher FPG, higher Tyg, higher mean-HbA1c, longer duration of diabetes, and more insulin usage. Elevated CIMT and greater prevalence of carotid plaque were also found.

Table 1 Baseline characteristics of the study population according to the CV-HbA1c tertile

	T1 (0.0074–0.0476)	T2 (0.0476–0.0920)	T3 (0.0920–0.3504)	p
Age, year	63.2 ± 10.8	61.8 ± 10.7	61.3 ± 10.2	0.203
Male (%)	69.7%	79.8%	80.9%	0.021
SBP, mmHg	138.8 ± 18.6	140.1 ± 20.1	143.7 ± 21.4	0.053
DBP, mmHg	77.7 ± 11.9	78.5 ± 12.5	79.8 ± 11.6	0.249
FPG, mmol/L	7.10 ± 1.34	7.54 ± 2.07	8.90 ± 3.21	<0.001
TC, mmol/L	5.12 ± 0.99	5.04 ± 1.12	5.10 ± 1.18	0.742
Triglyceride, mmol/L	1.80 ± 1.40	1.98 ± 1.42	2.06 ± 1.76	0.287
HDL-C, mmol/L	1.36 ± 0.30	1.31 ± 0.29	1.29 ± 0.27	0.055
LDL-C, mmol/L	3.08 ± 0.81	3.05 ± 0.83	3.06 ± 0.89	0.962
BMI	25.1 ± 3.3	25.7 ± 3.1	25.5 ± 3.4	0.265
Tyg	7.44 ± 0.60	7.57 ± 0.65	7.71 ± 0.74	<0.001
eGFR, mL/min/1.73 m2	79.6 ± 23.5	83.4 ± 28.4	85.8 ± 30.8	0.100
Mean-HbA1c, %	6.4 ± 0.7	6.9 ± 1.1	7.7 ± 1.2	<0.001
Presence of plaque (%)	36.2%	52.1%	67.0%	<0.001
Duration of diabetes, years	4.3 ± 4.8	4.8 ± 5.6	5.7 ± 5.2	0.026
Current smoker (%)	31.4%	34.0%	37.8%	0.434
Use of anti-platelet agent (%)	11.2%	19.7%	18.1%	0.062
Use of statins (%)	36.2%	39.9%	41.0%	0.638
Use of ARB/ACEI (%)	12.8%	17.0%	16.5%	0.544
Use of insulin (%)	6.9%	14.9%	18.1%	0.001
Other antidiabetic drugs (%)
Sulfonylureas	28.2%	25.5%	30.3%	0.606
Metformin	53.7%	62.8%	64.4%	0.059
Thiazolidinediones	20.7%	21.8%	18.1%	0.682
DPP-4 inhibitors/GLP-1 RA	28.7%	31.4%	25.0%	0.395
SGLT-2 inhibitors	17.6%	21.9%	20.7%	0.644
CIMT, mm	1.22 ± 0.81	1.49 ± 0.83	1.73 ± 0.77	<0.001

Data are given as percentage or Mean ± standard deviation.

Variables with significant difference among groups are marked in bold.

SBP, systolic blood pressure; DBP, diastolic blood pressure; FPG, fasting plasma glucose; TC, total cholesterol; HDL-C, high density lipoprotein; LDL-C, low density lipoprotein; BMI, body mass index; Tyg, triglyceride glucose index; eGFR, estimated glomerular filtration rate; ARB, angiotension II receptor blockers; ACEI, angiotensin-converting enzyme inhibitors; DPP-4, dipeptidyl peptidase-4; GLP-1RA, glucagon-like peptide-1 receptor agonist; SGLT-2, sodium glucose cotransporter 2; CIMT, carotid intima-media thickness

We explored the associations between CIMT and HbA1c variability. Fig. 2 shows that both CV-HbA1c and SD-HbA1c are positively correlated with CIMT; the correlation coefficients are 0.272 and 0.306, respectively (Fig. 2a, 2b). The correlation remained after excluding participants with carotid plaques (Fig. 2c, 2d). Multiple linear regression also revealed the associations between CV-HbA1c, SD-HbA1c and CIMT (Table 2). Each doubling in CV-HbA1c and SD-HbA1c was associated with a significant increment in CIMT (0.177 mm and 0.134 mm respectively). Doubling in CV-HbA1c and SD-HbA1c contributed to 0.066 mm and 0.088 mm increments in CIMT, respectively, in participants without carotid plaques. We further stratified the patients into two subgroups according to mean-HbA1c <6.5% and ≥6.5%, HbA1c variability remained independently associated with CIMT in each subgroup (Table 3) and had more impact on CIMT in groups with higher mean-HbA1c.

Fig. 2

Relationship between CV-HbA1c, SD-HbA1c and CIMT in all populations (a, b) and in subjects without carotid plaques (c, d)

Table 2 Multiple linear regression analysis for CV-HbA1c and SD-HbA1c with CIMT

	Model 1	Model 2	Model 3	Model 4
	Non-standardized β (95% CI), p
Log2 (CV-HbA1c)	0.310 (0.220, 0.400), <0.001	0.302 (0.200, 0.405), <0.001	0.181 (0.075, 0.287), 0.001	0.177 (0.073, 0.281), 0.001
Log2 (SD-HbA1c)	0.225 (0.171, 0.279), <0.001	0.241 (0.175, 0.308), <0.001	0.136 (0.056, 0.215), 0.001	0.134 (0.056, 0.213), 0.001
After excluding participants with plaque
Log2 (CV-HbA1c)	0.091 (0.058, 0.124), <0.001	0.092 (0.049, 0.134), <0.001	0.065 (0.021, 0.108), 0.004	0.066 (0.023, 0.109), 0.003
Log2 (SD-HbA1c)	0.089 (0.065, 0.113), <0.001	0.107 (0.073, 0.140), <0.001	0.085 (0.046, 0.123), <0.001	0.088 (0.049, 0.126), <0.001

Model 1: age, gender

Model 2: model 1 + hypertension + BMI + dyslipidemia + Tyg + eGFR

Model 3: model 2 + mean-HbA1c

Model 4: model 3 + current smoking + duration of diabetes + medications

Table 3 Multiple linear regression analysis for CV-HbA1c and SD-HbA1c with CIMT under HbA1c stratification

		All participants	Participants without plaque
		Non-standardized β (95% CI), p
mean-HbA1c <6.5%	Log2 (CV-HbA1c)	0.149 (0.052, 0.246), 0.003	0.067 (0.029, 0.105), <0.001
	Log2 (SD-HbA1c)	0.146 (0.048, 0.243), 0.003	0.067 (0.029, 0.104), <0.001
mean-HbA1c ≥6.5%	Log2 (CV-HbA1c)	0.195 (0.103, 0.287), <0.001	0.111 (0.055, 0.166), <0.001
	Log2 (SD-HbA1c)	0.194 (0.102, 0.285), <0.001	0.110 (0.055, 0.165), <0.001

All the risk factors were adjusted in the model.

Next, we explored the relationship between HbA1c variability and carotid plaques. Table 4 shows that HbA1c variability is independently associated with the presence of carotid plaque after fully adjusting for all other risk factors. Doubling in CV-HbA1c was associated with a 2.5-fold higher risk of carotid plaques (OR 2.50, 95%CI 1.73–3.62). Doubling in SD-HbA1c contributed to a 1.7-fold higher risk of carotid plaque (OR 1.71, 95%CI 1.37–2.22). We then treated the parameters as categorical variables. Compared to the reference group (first tertile), ORs for carotid plaque were 2.68 (95%CI 1.57, 4.56) and 2.88 (95%CI 1.61, 5.13) in the third tertile of CV-HbA1c and SD-HbA1c, respectively.

Table 4 Multivariate logistic regression analysis for CV-HbA1c and SD-HbA1c with carotid plaques

	Model 1	Model 2	Model 3	Model 4
Log2 (CV-HbA1c)	3.26 (2.50, 4.98), <0.001	3.51 (2.49, 4.96), <0.001	2.50 (1.743, 3.59), <0.001	2.50 (1.73, 3.62), <0.001
CV-HbA1c as categorical variables
T2	2.23 (1.44, 3.46), <0.001	2.27 (1.46, 3.52), <0.001	1.82 (1.16, 2.86), 0.010	1.85 (1.17, 2.93), 0.009
T3	5.24 (3.29, 8.33), <0.001	4.80 (3.02, 7.63), <0.001	2.66 (1.58, 4.48), <0.001	2.68 (1.57, 4.56), <0.001
Log2 (SD-HbA1c)	2.02 (1.69, 2.40), <0.001	2.01 (1.69, 2.40), <0.001	1.67 (1.35, 2.07), <0.001	1.71 (1.37, 2.22), <0.001
SD-HbA1c as categorical variables
T2	2.17 (1.40, 3.36), <0.001	2.16 (1.39, 3.35), <0.001	1.75 (1.11, 2.75), 0.015	1.79 (1.13, 2.84), 0.013
T3	5.85 (3.67, 9.34), <0.001	5.46 (3.42, 8.70), <0.001	2.77 (1.81, 5.74), <0.001	2.88 (1.61, 5.13), <0.001

Data are presented as Odds ratio (95% confidence interval), p

T2, tertile 2; T3, tertile 3

Tertile 1 was chosen as the reference group.

Model 1: age, gender

Model 2: model 1 + hypertension + BMI + dyslipidemia + Tyg + eGFR

Model 3: model 2 + mean-HbA1c

Model 4: model 3 + current smoking + duration of diabetes + medications

To explore how mean-HbA1c would affect the relationship between HbA1c variability and carotid plaques, a cross-tabulation analysis was conducted. We chose the group with low mean-HbA1c and the first tertile of CV or SD as the reference group. Fig. 3 shows that in the groups with relatively well-controlled diabetes (mean-HbA1c <6.5%), elevated HbA1c variability was associated with increased prevalence of carotid plaque, though the results were statistically insignificant. High mean-HbA1c and high HbA1c variability had additive effects on carotid plaque. The fully-adjusted ORs for carotid plaque were 4.68 (95%CI 2.71, 8.10) and 4.86 (95%CI 2.84, 8.31) in the third tertile of CV-HbA1c and SD-HbA1c with mean-HbA1c ≥6.5%, respectively.

Fig. 3

Associations between HbA1c variability tertile, mean HbA1c and incidence of carotid plaque. All risk factors were adjusted in the model.

Table 5 shows the relationship between HbA1c variability and carotid plaques in different subgroups. HbA1c variability was associated with plaques in most subgroups. However, the interactions disappeared in populations with normal serum lipid levels and those who were current smokers.

Table 5 Multivariate logistic regression analysis for CV-HbA1c and SD-HbA1c with carotid plaques in subgroups

		CV-HbA1c		SD-HbA1c
		T2	T3	T2	T3
Elderly	yes	1.57 (1.13, 3.07), 0.001	2.52 (1.15, 4.73), <0.001	1.68 (1.13, 2.97), 0.032	2.40 (1.13, 6.32), 0.001
	no	1.67 (1.16, 3.15), 0.039	2.67 (1.54, 4.43), 0.006	1.72 (1.01, 2.86), 0.049	2.63 (1.19, 4.83), 0.006
Dyslipidemia	yes	1.72 (0.88, 3.11), 0.069	2.72 (1.38, 5.36), 0.004	1.79 (1.02, 3.23), 0.042	3.37 (1.48, 6.46), 0.001
	no	3.22 (1.27, 8.06), 0.037	2.55 (0.76, 8.591), 0.130	2.01 (0.67, 6.76), 0.179	1.42 (0.36, 4.18), 0.562
Hypertension	yes	1.80 (0.96, 3.36), 0.065	2.59 (1.23, 5.45), 0.012	2.09 (0.89, 4.58), 0.081	4.61 (1.63, 6.13), 0.002
	no	1.93 (0.86, 4.33), 0.112	3.82 (1.43, 11.07), 0.005	1.82 (0.95, 3.10), 0.093	2.83 (1.27, 5.97), 0.014
Obesity	yes	1.41 (0.53, 3.91), 0.518	6.83 (1.79, 20.38), 0.003	1.72 (0.51, 5.33), 0.311	5.66 (2.71, 10.12), <0.001
	no	1.57 (0.87, 3.31), 0.571	2.21 (1.12, 4.27), 0.015	1.91 (1.22, 3.17), 0.015	2.17 (1.41, 5.22), 0.007
Current smoker	yes	0.44 (0.39, 1.14), 0.507	1.82 (0.90, 2.10), 0.166	0.66 (0.54, 1.11), 0.531	1.82 (0.72, 2.13), 0.160
	no	1.73 (1.01, 3.26), 0.049	2.72 (1.41, 5.37), 0.003	1.72 (0.92, 2.81), 0.069	2.61 (1.41, 5.67), <0.001
Tyg	<7.5	2.41 (1.17, 5.21), 0.008	2.74 (1.38, 6.77), 0.012	2.09 (0.88, 5.36), 0.087	2.67 (1.51, 5.68), 0.001
	≥7.5	1.17 (0.59, 3.12), 0.477	2.50 (1.09, 5.13), 0.033	1.32 (0.62, 2.46), 0.517	3.13 (1.32, 8.17), 0.002

Data are presented as Odds ratio (95% confidence interval), p

Results with statistical significance are marked in bold.

T2, tertile 2; T3, tertile 3

Tertile 1 was chosen as the reference group.

All risk factors were adjusted.

Discussion

Our study found that visit-to-visit HbA1c variability was associated with CIMT and plaques in populations with diabetes. CV-HbA1c and SD-HbA1c had more impacts on subclinical atherosclerosis in patients with poorly-controlled blood glucose. We also found men with more severe insulin resistance, longer duration of diabetes, and poorer glycaemic control tended to have higher HbA1c fluctuation.

The development of atherosclerosis is complex, and an increase in intimal thickness often happens in the early stage of the disease [20]. Blood fluctuations play an important role in the development of atherosclerosis in addition to mean glucose levels. Some short-term glycaemic variability metrics including the mean amplitude of glycaemic excursion [21] and time in range (TIR) [9] are positively associated with CIMT. However, Cutruzzola et al. [10] found the percentage of TIR did not correlate with preclinical atherosclerosis. The authors attributed inconsistent results to the deficiency of TIR to identify postprandial hyperglycaemia, which is an independent risk factor for CVD and endothelial damage [22]. The glucose profiles provided by CGM represent blood excursions over a short period, while atherosclerosis is a chronic disease.

HbA1c variability not only reflects glucose fluctuations but also ambient hyperglycaemia [23]. A positive correlation was found between HbA1c variability and mean HbA1c in our study. Several large-scale cohort studies [12-15, 24, 25] proved that HbA1c variability could predict CVD and all-cause mortality in populations with or without diabetes. Our study further investigated the impact of HbA1c variability on artery walls prior to CVD events.

Each doubling in CV-HbA1c or SD-HbA1c would increase the thickness of the carotid intima media and prevalence of plaque regardless of mean-HbA1c, as the beginning of atherogenic progression. We also found HbA1c variability had more impacts on subclinical atherosclerosis in patients with HbA1c ≥6.5%. Our results were inconsistent with some previous studies [25, 26] in which HbA1c variability had larger effects on CVD events in patients with lower mean HbA1c values. Shen et al. [27] found that high HbA1c variability was associated with more severe and recurrent hypoglycaemic events. High HbA1c variability with low mean HbA1c often means more hypoglycaemic events, which are risk factors for CVD events. In this study, the information of hypoglycaemia is lacking. But we noticed the usage rates of insulin and sulfonylureas were much lower than previous studies [26, 27]. Thus we hypothesized our participants were less likely to suffer from hypoglycaemia in group with mean-HbA1c <6.5%.

Although both CIMT and plaque are indicators of subclinical atherosclerosis, there are some differences between the two metrics. Plaques behave superior to CIMT in the early detection of CVD [28], and are strongly associated with CVD risk factors [29]. CIMT is influenced more by some physiological factors like ageing [30]. High HbA1c variability and high HbA1c mean together increased the risk of asymptomatic atherosclerosis, resulting in potential CVD outcomes in the future [25]. However, the associations between HbA1c variability and carotid plaque disappeared in subgroups with a smoking history or lipids in the normal range. We suppose that smoking itself is a strong risk factor for atherosclerosis [31], covering the influence of HbA1c variability. The small sample size in the group without dyslipidemia may result in bias.

Several previous studies have proved that short-term glycaemic fluctuation results in the overproduction of reactive oxygen species [32], resulting in the secretion of inflammatory factors [33]. Chronic inflammation not only damages endothelial cells directly [34], but also leads to dyslipidemia and insulin resistance [35, 36]. However, few studies have focused on the mechanisms of chronic glycaemic fluctuation on the pathogenesis of atherosclerosis.

Several indirect clues are hinting that chronic glycaemic variability may impair vascular endothelium through insulin resistance. Tyg index, a new metric involving levels of fasting plasma glucose and triglyceride, is suggested to be a reliable marker for insulin resistance according to recent research [17]. Tyg levels are higher in individuals with high HbA1c variability, suggesting HbA1c variability is positively associated with insulin resistance. However, our study failed to find an association between Tyg and CIMT or plaques. We supposed it was because using Tyg to detect CVD is biased by diabetes and hyperlipidemia [37] and is less accurate compared to the Homeostasis Model Assessment of Insulin Resistance (HOMA-IR). A previous study [38] found that visit-to-visit fasting plasma glucose variability (another metric reflecting long-term glucose variability) increased the risk of incident diabetes mediated by HOMA-IR. A meta-analysis [39] found an overall stronger HR for HbA1c variability linked to macrovascular complications than microvascular complications. The former is more vulnerable to lipid abnormalities and insulin resistance. Therefore, future research is needed to explore the impact of insulin resistance on the association between HbA1c variability and endothelial dysfunction.

To our knowledge, this is the first study to explore the relationship between HbA1c variability and CIMT or carotid plaques in populations with diabetes. There are inevitably some limitations to our study. First, it was a cross-sectional study, so we could not determine the causal relationship between HbA1c variability and CIMT or carotid plaque. Second, it was a single-center study with a relatively small sample size, which may cause some bias in our results. Third, the carotid ultrasound was performed only at the common carotid artery but CIMT in different segments may have different clinical meanings and specific associations with cardiovascular risk factors [40]. Forth, Some other confounding factors such as dietary habit, physical activity and diabetic complications were not assessed in this study. Lastly, there was heterogeneity in times and intervals of the HbA1c measurements for each individual, which may influence the interpretation of the metrics.

In conclusion, our research found HbA1c variability was independently associated with CIMT and carotid plaques even in patients with relatively well-controlled blood glucose levels. High mean-HbA1c plus high HbA1c variability significantly elevates the prevalence of carotid plaques. We suggest that the oscillation of HbA1c should be monitored in clinics in addition to the single HbA1c value.

Acknowledgements

We acknowledge all the involved participants and technicians in Health Management Center of Zhongda Hospital for their efforts to complete our research.

Disclosure

None of the authors have any conflicts of interest associated with this research.

Reference

• 1   Milutinović  A,  Šuput  D,  Zorc-Pleskovič  R (2020) Pathogenesis of atherosclerosis in the tunica intima, media, and adventitia of coronary arteries: an updated review. Bosn J Basic Med Sci 20: 21–30.
• 2   Johnsen  SH,  Mathiesen  EB (2009) Carotid plaque compared with intima-media thickness as a predictor of coronary and cerebrovascular disease. Curr Cardiol Rep 11: 21–27.
• 3   Cao  JJ,  Arnold  AM,  Manolio  TA,  Polak  JF,  Psaty  BM, et al. (2007) Association of carotid artery intima-media thickness, plaques, and C-reactive protein with future cardiovascular disease and all-cause mortality: the Cardiovascular Health Study. Circulation 116: 32–38.
• 4   Lorenz  MW,  Markus  HS,  Bots  ML,  Rosvall  M,  Sitzer  M (2007) Prediction of clinical cardiovascular events with carotid intima-media thickness: a systematic review and meta-analysis. Circulation 115: 459–467.
• 5   Stein  JH,  Korcarz  CE,  Hurst  RT,  Lonn  E,  Kendall  CB, et al. (2008) Use of carotid ultrasound to identify subclinical vascular disease and evaluate cardiovascular disease risk: a consensus statement from the American Society of Echocardiography Carotid Intima-Media Thickness Task Force. Endorsed by the Society for Vascular Medicine. J Am Soc Echocardiogr 21: 93–111.
• 6   Tang  X,  Li  S,  Wang  Y,  Wang  M,  Yin  Q, et al. (2016) Glycemic variability evaluated by continuous glucose monitoring system is associated with the 10-y cardiovascular risk of diabetic patients with well-controlled HbA1c. Clin Chim Acta 461: 146–150.
• 7   Gohbara  M,  Hibi  K,  Mitsuhashi  T,  Maejima  N,  Iwahashi  N, et al. (2016) Glycemic variability on continuous glucose monitoring system correlates with non-culprit vessel coronary plaque vulnerability in patients with first-episode acute coronary syndrome—Optical Coherence Tomography Study. Circ J 80: 202–210.
• 8   Yoon  JE,  Sunwoo  JS,  Kim  JS,  Roh  H,  Ahn  MY, et al. (2017) Poststroke glycemic variability increased recurrent cardiovascular events in diabetic patients. J Diabetes Complications 31: 390–394.
• 9   Lu  J,  Ma  X,  Shen  Y,  Wu  Q,  Wang  R, et al. (2020) Time in range is associated with carotid intima-media thickness in type 2 diabetes. Diabetes Technol Ther 22: 72–78.
• 10   Cutruzzolà  A,  Parise  M,  Scavelli  FB,  Barone  M,  Gnasso  A, et al. (2022) Time in range does not associate with carotid artery wall thickness and endothelial function in type 1 diabetes. J Diabetes Sci Technol 16: 904–911.
• 11   American Diabetes  Association (2013) Diagnosis and classification of diabetes mellitus. Diabetes Care 36 Suppl 1: S67–S74.
• 12   Scott  ES,  Januszewski  AS,  O’Connell  R,  Fulcher  G,  Scott  R, et al. (2020) Long-term glycemic variability and vascular complications in type 2 diabetes: post hoc analysis of the FIELD study. J Clin Endocrinol Metab 105: e3638–e3649.
• 13   Forbes  A,  Murrells  T,  Mulnier  H,  Sinclair  AJ (2018) Mean HbA1c, HbA1c variability, and mortality in people with diabetes aged 70 years and older: a retrospective cohort study. Lancet Diabetes Endocrinol 6: 476–486.
• 14   Li  S,  Nemeth  I,  Donnelly  L,  Hapca  S,  Zhou  K, et al. (2020) Visit-to-visit HbA1c variability is associated with cardiovascular disease and microvascular complications in patients with newly diagnosed type 2 diabetes. Diabetes Care 43: 426–432.
• 15   Wan  E,  Yu  E,  Chin  W,  Ng  F,  Chia  S, et al. (2020) Age-specific associations of glycated haemoglobin variability with cardiovascular disease and mortality in patients with type 2 diabetes mellitus: a 10- year cohort study. Diabetes Obes Metab 22: 1316–1327.
• 16   Bonke  FC,  Donnachie  E,  Schneider  A,  Mehring  M (2016) Association of the average rate of change in HbA1c with severe adverse events: a longitudinal evaluation of audit data from the Bavarian disease management program for patients with type 2 diabetes mellitus. Diabetologia 59: 286–293.
• 17   Selvi  NMK,  Nandhini  S,  Sakthivadivel  V,  Lokesh  S,  Srinivasan  AR, et al. (2021) Association of triglyceride-glucose index (TyG index) with HbA1c and insulin resistance in type 2 diabetes mellitus. Maedica (Bucur) 16: 375–381.
• 18   Poggio  ED,  Wang  X,  Greene  T,  Van Lente  F,  Hall  PM (2005) Performance of the modification of diet in renal disease and Cockcroft-Gault equations in the estimation of GFR in health and in chronic kidney disease. J Am Soc Nephrol 16: 459–466.
• 19   Takase  M,  Nakaya  N,  Nakamura  T,  Kogure  M,  Hatanaka  R, et al. (2023) Carotid intima media thickness and risk factor for atherosclerosis: Tohoku Medical Megabank Community-Based Cohort Study. J Atheroscler Thromb. doi: 10.5551/jat.64039. Online ahead of print.
• 20   Falk  E (2006) Pathogenesis of atherosclerosis. J Am Coll Cardiol 47: C7–C12.
• 21   Chen  XM,  Zhang  Y,  Shen  XP,  Huang  Q,  Ma  H, et al. (2010) Correlation between glucose fluctuations and carotid intima-media thickness in type 2 diabetes. Diabetes Res Clin Pract 90: 95–99.
• 22   Ceriello  A (2010) Hyperglycaemia and the vessel wall: the pathophysiological aspects on the atherosclerotic burden in patients with diabetes. Eur J Cardiovasc Prev Rehabil 17: S15–S19.
• 23   Kilpatrick  ES,  Rigby  AS,  Atkin  SL (2008) A1c variability and the risk of microvascular complications in type 1 diabetes: data from the diabetes control and complications trial. Diabetes Care 31: 2198–2202.
• 24   Ghouse  J,  Skov  MW,  Kanters  JK,  Lind  B,  Isaksen  JL, et al. (2019) Visit-to-visit variability of hemoglobin A1c in people without diabetes and risk of major adverse cardiovascular events and all-cause mortality. Diabetes Care 42: 134–141.
• 25   Sheng  CS,  Tian  J,  Miao  Y,  Cheng  Y,  Yang  Y, et al. (2020) Prognostic significance of long-term HbA1c variability for all-cause mortality in the ACCORD trial. Diabetes Care 43: 1185–1190.
• 26   Sato  M,  Inaishi  J,  Saisho  Y,  Sato  Y,  Komuro  I, et al. (2021) Association of visit-to-visit glycemic variability with risk of cardiovascular diseases in high-risk Japanese patients with type 2 diabetes: a subanalysis of the EMPATHY trial. J Diabetes Investig 12: 2190–2196.
• 27   Shen  Y,  Zhou  J,  Shi  L,  Nauman  E,  Katzmarzyk  PT, et al. (2021) Association between visit-to-visit HbA1c variability and the risk of cardiovascular disease in patients with type 2 diabetes. Diabetes Obes Metab 23: 125–135.
• 28   Sillesen  H,  Sartori  S,  Sandholt  B,  Baber  U,  Mehran  R, et al. (2017) Carotid plaque thickness and carotid plaque burden predict future cardiovascular events in asymptomatic adult Americans. Eur Heart J Cardiovasc Imaging 19: 1042–1050.
• 29   Azcui Aparicio  RE,  Carrington  MJ,  Ball  J,  Abhayaratna  W,  Stewart  S, et al. (2021) Association of traditional risk factors with carotid intima-media thickness and carotid plaque in asymptomatic individuals with a family history of premature cardiovascular disease. Int J Cardiovasc Imaging. doi: 10.1007/s10554-021-02459-x.
• 30   Baroncini  LAV,  de Castro Sylvestre  L,  Filho  RP (2015) Carotid intima- media thickness and carotid plaque represent different adaptive responses to traditional cardiovascular risk factors. Int J Cardiol heart Vasc 9: 48–51.
• 31   Tattersall  MC,  Gassett  A,  Korcarz  CE,  Gepner  AD,  Kaufman  JD, et al. (2014) Predictors of carotid thickness and plaque progression during a decade: the Multi-Ethnic Study of Atherosclerosis. Stroke 45: 3257–3262.
• 32   Ceriello  A,  Esposito  K,  Piconi  L,  Ihnat  MA,  Thorpe  JE, et al. (2008) Oscillating glucose is more deleterious to endothelial function and oxidative stress than mean glucose in normal and type 2 diabetic patients. Diabetes 57: 1349–1354.
• 33   Tanaka  A,  Azuma  K,  Toyofuku  Y,  Kurokawa  A,  Otsuka  A, et al. (2006) Insulin and nateglinide reduce monocyte adhesion to endothelial cells in Goto-Kakizaki rats exhibiting repetitive blood glucose fluctuation. Biochem Biophys Res Commun 350: 195–201.
• 34   Piconi  L,  Quagliaro  L,  Assaloni  R,  Da Ros  R,  Maier  A, et al. (2006) Constant and intermittent high glucose enhances endothelial cell apoptosis through mitochondrial superoxide overproduction. Diabetes Metab Res Rev 22: 198–203.
• 35   Lauterbach  MA,  Wunderlich  FT (2017) Macrophage function in obesity-induced inflammation and insulin resistance. Pflugers Arch 469: 385–396.
• 36   Monnier  L,  Mas  E,  Ginet  C,  Michel  F,  Villon  L, et al. (2006) Activation of oxidative stress by acute glucose fluctuations compared with sustained chronic hyperglycemia in patients with type 2 diabetes. JAMA 295: 1681–1687.
• 37   Alizargar  J,  Bai  CH,  Hsieh  NC,  Wu  SV (2020) Use of the triglyceride-glucose index (TyG) in cardiovascular disease patients. Cardiovasc Diabetol 19: 8.
• 38   Xu  J,  Li  L,  Huang  S,  Song  H,  Gao  J, et al. (2022) Impact of visit-to-visit fasting plasma glucose variability on the development of diabetes: the mediation by insulin resistance. J Diabetes 14: 205–215.
• 39   Sartore  G,  Ragazzi  E,  Caprino  R,  Lapolla  A (2023) Long-term HbA1c variability and macro-/micro-vascular complications in type 2 diabetes mellitus: a meta-analysis update. Acta Diabetol 60: 721–738.
• 40   Schott  LL,  Wildman  RP,  Brockwell  S,  Simkin-Silverman  LR,  Kuller  LH, et al. (2004) Segment-specific effects of cardiovascular risk factors on carotid artery intima-medial thickness in women at midlife. Arterioscler Thromb Vasc Biol 24: 1951–1956.