An Inconvenient Relationship of Hemoglobin A1c Level with Endothelial Function in Type 2 Diabetes

Background Diabetes is associated with endothelial dysfunction. However, there is little information on the relationship between hemoglobin A1c (HbA1c) level and endothelial function. This study evaluated the relationship between HbA1c level and flow-mediated vasodilation (FMD). Methods We measured FMD in 1215 patients with type 2 diabetes including 349 patients not taking antidiabetic drugs and 866 patients taking antidiabetic drugs. The patients were divided into four groups based on HbA1c levels: <48 mmol/mol, 48–52 mmol/mol, 53–63 mmol/mol, and ≥ 64 mmol/mol. (< 6.5%, 6.5–6.9%, 7.0-7.9%, ≥ 8.0%) Results An inverted U-shaped pattern of association between HbA1c level and FMD was observed at the peak of HbA1c of about 53 mmol/mol (7%). FMD was significantly smaller in the HbA1c < 48 mmol/mol (6.5%) group than in the HbA1c 48–52 mmol/mol (6.5–6.9%) group and HbA1c 53– 63 mmol/mol (7.0-7.9%) group (p < 0.001 and p < 0.001), and FMD values were similar in the HbA1c < 48 mmol/mol (6.5%) group and HbA1c ≥ 64 mmol/mol (8.0%) group. There were no significant differences in nitroglycerine-induced vasodilation (NID) values among the four groups. After adjustments for confounding factors, FMD was significantly smaller in the HbA1c < 48 mmol/mol (6.5%) group than in the HbA1c 48–52 mmol/mol (6.5–6.9%) and HbA1c 53–63 mmol/mol (7.0-7.9%) group (p = 0.002 and p = 0.04). In patients not taking antidiabetic drugs, FMD was also significantly smaller in the HbA1c < 48 mmol/mol (6.5%) group than in the HbA1c 48–52 mmol/mol (6.5–6.9%) group and HbA1c 53–63 mmol/mol (7.0-7.9%) group (p < 0.001 and p = 0.02), and there were no significant differences in NID values among the four groups. Conclusions These findings suggest that a low HbA1c level of < 48 mmol/mol (6.5%) is associated with endothelial dysfunction. An HbA1c level of 48–52 mmol/mol (6.5–6.9%) may be appropriate for maintenance of endothelial function. HDL-C indicates high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; CVD, cardiovascular disease; FMD, NID, nitroglycerine-induced vasodilation.


KEYWORDS
hemoglobin A1c, type 2 diabetes, endothelial function, flow-mediated vasodilation, nitroglycerine-induced vasodilation Abstract Background Diabetes is associated with endothelial dysfunction. However, there is little information on the relationship between hemoglobin A1c (HbA1c) level and endothelial function. This study evaluated the relationship between HbA1c level and flow-mediated vasodilation (FMD).
Conclusions These findings suggest that a low HbA1c level of < 48 mmol/mol (6.5%) is associated with endothelial dysfunction. An HbA1c level of 48-52 mmol/mol (6.5-6.9%) may be appropriate for maintenance of endothelial function.

Background
Diabetes is a risk factor for atherosclerosis and subsequent cardiovascular disease (CVD) and CV events.
[1] Previous studies showed that adults with diabetes have 2-4-fold higher rates of all-cause mortality and CVD mortality than those in subjects without diabetes. [2] [3] Therefore, prevention of CVD in patients with diabetes is clinically important. HbA1c level, an index of glycemic control, is usually checked in patients with diabetes. However, HbA1c-guided diabetes treatment is still controversial.

Previous large clinical trials, including the Veterans Affairs Diabetes Trial (VADT), the Action in
Diabetes and Vascular Disease: Preter Ax and Diamicron MR Controlled Evaluation (ADVANCE) trial, and the Kumamoto study, have shown that intensive glucose control reduces the incidences of microvascular diseases such as retinopathy and nephropathy but not the incidence of macrovascular diseases such as myocardial infarction and stroke in patients with type 2 diabetes. [4][5][6][7] The Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial showed that intensive therapy increased allcause mortality in patients with type 2 diabetes. [8] The VADT and the ADVANCE trial showed that severe hypoglycemia increases death from cardiovascular disease and any-cause of death. [5,7] Unfortunately, the optimal target level of HbA1c in diabetes is unclear, and it is still controversial whether intensive glucose control by HbA1c-guided therapy reduces the incidence of CV events.
Endothelial dysfunction is well known as the initial step of atherosclerosis, and it plays a critical role in the development of atherosclerosis, leading to CVD. [9] Measurement of flow-mediated vasodilation (FMD) in the brachial artery is an established tool for assessment of endothelial function [10] and it is well known as an independent predictor of cardiovascular events. [11] Endothelial function assessed by FMD is impaired by traditional cardiovascular risk factors such as hypertension, dyslipidemia, smoking, chronic alcohol drinking and also diabetes. FMD is reversible by several interventions such as life-style modifications and pharmacological treatment. [12,13] Therefore, FMD is a very useful tool for assessing the current vascular function and cardiovascular risk.
Diabetes is associated with endothelial dysfunction. [14] Chronic hyperglycemia is a major contributor to increased oxidative stress and causes endothelial dysfunction through inactivation of nitric oxide. [15] Several studies have shown that endothelial function is improved by antidiabetic therapy including use of antidiabetic drugs. [12,16] However, there is little information on the relationship between HbA1c level and endothelial function.
Therefore, we evaluated the relationship between HbA1c level and endothelial function assessed by FMD in patients with type 2 diabetes.

Study patients
A total of 10260 subjects (7385 patients from the FMD-J study and 2875 patients who underwent a health checkup at Hiroshima University Hospital between August 2007 and August 2016) were recruited in this study. The FMD-J study was a prospective multicenter registry. The design of FMD-J study has been described in detail previously. [17] The protocol used for measurement of FMD was the same in the FMD-J study and at Hiroshima University Hospital. Exclusion criteria was listed on Online Figure S1. Finally, we enrolled 1215 subjects in this study. Hypertension was defined as the use of antihypertensive drugs or systolic blood pressure of more than 140 mmHg or diastolic blood pressure of more than 90 mmHg measured in a sitting position on at least 3 occasions. Written informed consent for participation in this study was obtained from all participants.

Study 1. HbA1c levels and vascular function in patients with type 2 diabetes
In study 1, we assessed the relationships between HbA1c level and vascular function as assessed by measurement of FMD, an index of endothelium-dependent vasodilation, and by measurement of nitroglycerine-induced vasodilation (NID), an index of endothelium-independent vasodilation, in 1215 patients with type 2 diabetes. First, we divided the patients into two groups based on their HbA1c levels: <48 mmol/mol (6.5%) and ≥48 mmol/mol (6.5%). Multivariate regression analysis was performed to identify independent variables associated with vascular function. Next, we divided the patients into four groups according to HbA1c levels: <48 mmol/mol (6.5%), 48-52 mmol/mol (6.5-6.9%), 53-63 mmol/mol (7.0-7.9%), and ≥64 mmol/mol (8.0%). We next assessed the relationships of HbA1c levels with FMD and NID using propensity score matching.

Study 2. HbA1c levels and vascular function in patients with type 2 diabetes not taking antidiabetic drugs
We evaluated the relationships of HbA1c levels with FMD and NID in 349 patients with type 2 diabetes who were not taking antidiabetic drugs by using the same protocol as that used in study 1.

Measurements of FMD and NID
A high-resolution ultrasonography equipment specialized to measure FMD (UNEXEF18G, UNEX Co., Nagoya, Japan) was used to evaluate FMD. Additional details are available in the Online methods. The intraclass correlation coefficient between each participating institutions and the core laboratory has been previously described. [20]

Statistical analysis
Results are presented as means±SD. All reported probability values were 2-sided, and a probability value of <0.05 was considered statistically significant. An association between FMD and HbA1c level was explored visually using a locally weighted regression smoothing (Lowess) plot. Categorical values were compared by means of the chi-square test. Continuous variables were compared by using ANOVA multiple groups. Comparisons between the groups categorized according to HbA1c levels were carried out using repeated measures ANOVA with Tukey's post hoc test. Univariate linear regression analyses were performed to assess the relationships among the variables. Multivariate logistic regression analysis was performed to identify independent variables associated with lower quartiles of FMD (<2.1%) and NID (<6.2%). Age, gender, body mass index, creatinine levels, current smoking, and the presence of hypertension, dyslipidemia and CVD were entered into the multivariate logistic regression analysis. As a sensitivity analysis, propensity score analysis was used to minimize the selection bias for evaluation of the relationship between HbA1c level and vascular function. The propensity score was calculated for each patient on the basis of logistic regression analysis of the probability of not taking antidiabetic drugs within groups stratified by HbA1c levels ((< 48 mmol/mol (6.5%), 48-52 mmol/mol (6.5-6.9%), 53-63 mmol/mol (7.0-7.9%), and ≥64 mmol/mol (8.0%)) using clinical variables including age, sex, body mass index (BMI), systolic blood pressure, diastolic blood pressure, heart rate, total cholesterol, triglycerides, high-density lipoprotein (HDL-C), uric acid levels, Results Study 1.

Baseline characteristics of patients with type 2 Diabetes
The baseline characteristics of the 1215 patients are summarized in

Univariate analysis of relationships among FMD, NID, HbA1c level and variables in patients with type 2 Diabetes
Online  Figure 2A). A scatter plot between FMD and HbA1c level with a Lowess smoothed curve is shown in Online Figure   S2B. FMD gradually increased with increase in HbA1c level to about 48-52 mmol/mol (6.5-6.9%) and the decreased with increase in HbA1c level above 53 mmol/mol (7.0%).

HbA1c level and variables
The division points for the lowest quartile and second quartile were 2.1% FMD and 6.2% NID.

Relationships among FMD, NID and HbA1c levels in patients with type 2 Diabetes
determined by using propensity score matching analysis.

Univariate analysis of relationships among FMD, NID, HbA1c level and variables in patients with type 2 Diabetes who were not taking antidiabetic drugs
Online  Figure S4A). Scatter plots between FMD and HbA1c with a Lowess smoothed curve are shown in Online Figure S4B. FMD gradually increased with increase in HbA1c level to about 48-52 mmol/mol (6.5-6.9%) and then decreased with increase in HbA1c level above 53 mmol/mol (7.0%) (Online Figure S4B).

HbA1c level and variables in patients with type 2 Diabetes who were not taking antidiabetic drugs
Multiple logistic regression analysis revealed that after adjustments for age, gender, BMI, current smoking, creatine, and presence of hypertension, dyslipidemia and CVD, HbA1c level of <48 mmol/mol (6.5%) was independently associated with a lower quartile of FMD (OR: 2.57, 95% CI: 1.45-4.54; p=0.001) but was not associated with a lower quartile of NID (OR: 1.29, 95% CI: 0.43-3.91; p=0.65) (Online Table S7).

Discussion
In the present study, we demonstrated that a low HbA1c level of < 48 mmol/mol (6.5%) was independently associated with small FMD in patients with type 2 diabetes. After adjustments for confounding factors, FMD was significantly smaller in the HbA1c < 48 mmol/mol (6.5%) group than in the HbA1c 48-52 mmol/mol (6.5-6.9%) group and HbA1c 53-63 mmol/mol (7.0-7.9%) group. In patients who were not taking antidiabetic drugs, FMD was also significantly smaller in the HbA1c < 48 mmol/mol (6.5%) group than in the HbA1c 48-52 mmol/mol (6.5-6.9%) group and HbA1c 53-63 mmol/mol (7.0-7.9%) group. In addition, we confirmed that FMD was significantly smaller in the low HbA1c group than in the HbA1c 48-52 mmol/mol (6.5-6.9%) group by using propensity score matching analysis. To our knowledge, the present study is the first study showing the detailed relationships between HbA1c levels and endothelial function in patients with type 2 diabetes including patients not taking antidiabetic drugs.
Interestingly, in the present study, HbA1c levels were not corelated with NID. There were no significant differences in NID values among the HbA1c groups of < 48 mmol/mol (6.5%), 48-52 mmol/mol (6.5-6.9%), 53-63 mmol/mol (7.0-7.9%), and ≥ 64 mmol/mol (8.0%). In patients with type 2 diabetes who were not taking antidiabetic drugs, there were also no significant differences in NID values among the four groups. These findings suggest that HbA1c level is not correlated with vascular smooth muscle function.
It is well known that the incidence of myocardial infarction increases in relation to HbA1c level. [21] It is thought that FMD, an index of endothelial function, decreases with increase in HbA1c level.
However, in the present study, a low HbA1c level of < 48 mmol/mol (6.5%) was found to be independently associated with endothelial dysfunction in patients with type 2 diabetes. To avoid the effects of antidiabetic drugs on HbA1c levels and to minimize the effect of hypoglycemia, we evaluated the relationships between HbA1c levels and FMD in patients with type 2 diabetes who were not taking antidiabetic drugs, and we found that the results were similar for patients taking and those not taking antidiabetic drugs.
The key finding of this study was that an inverted U-shaped pattern of association between HbA1c and FMD was observed at the peak of HbA1c of about 53 mmol/mol (7%) in patients with type 2 diabetes. This result may reflect the existence of a J-curve pattern of association between HbA1c and all causes of mortality. Diabetes is well known as a risk factor for endothelial function as well as for CVD. [14,22,23] However, the effect of intensive glucose control therapy on all causes of mortality is still controversial. Previous studies focused on the relationship between HbA1c and all causes of mortality. Some studies showed a positive linear relationship between HbA1c and all causes of mortality [24,25], while other studies showed a J-shaped relationship between HbA1c and all causes of mortality [26,27]. The effects of intensive glucose control therapy on morbidity and mortality of CV events are also controversial. [27,28] The UKPDS 73 study showed that the frequency of hypoglycemia in patients not taking antidiabetic drugs was 0.1%. [29] Hypoglycemia during intensive glucose control is probably a predictor of morbidity and mortality of CV events. It has been shown that the hazard ratios for all causes of mortality including CV events in patients with severe hypoglycemia episodes are between 1.74 and 3.27. [30,31] It has been postulated that hypoglycemia activates the sympathetic nervous system, releases catecholamines that cause increase heart rate and myocardial contractility [32], and activates platelet aggregation, leading to acute coronary syndrome and fatal arrythmia. [33] Although the precise mechanisms by which a low HbA1c level impairs endothelial function is uncertain, activation of the sympathetic nervous system may play a critical role in endothelial dysfunction. We cannot deny the possibility that factors other than hypoglycemia contribute to low HbA1c-induced endothelial dysfunction.

Study Limitations
This study has some limitations. First, this study was a cross-sectional study, although the study was conducted in multiple centers and had a large sample size. Therefore, we were able to evaluate the association but not causality between low HbA1c level and FMD. Second, unfortunately, we did not have information on the duration of diabetes from onset. The UKPDS80 study has shown that CVD risk reduction was observed after 10 years of follow up of intensive glucose therapy in patients with newly diagnosed type 2 diabetes. Third, this study was conducted in Japan, and our results for the association between HbA1c and FMD might not be applicable to other races. However, the ACCORD trial was conducted in North America, and the ADVANCE trial was conducted in 20 countries including countries in Asia and Europe and in North America and Australia. The results of those studies suggest that an inverted U-shaped pattern of the association of FMD with HbA1c that found in the present study is observed in all races. It is well known that HbA1c levels do not accurately reflect mean glucose values in patients with end-stage chronic kidney disease and in patients with dialysis. In the present study, we excluded those patients and we adjusted serum creatinine levels using propensity score matching analysis. In addition, since elderly patients often have malnutrition due to anorexia that leads to low HbA1c, we excluded patients over 80 years of age. Even after excluding these confounding factors, a low HbA1c level was associated with endothelial dysfunction in patients with type 2 diabetes.

Conclusions
A Low HbA1c level (< 48 mmol/mol (6.5%)) is associated with endothelial dysfunction in patients with The ethical committees of the participating institutions approved the study protocol. All participants provide written informed consent before date collection.

Consent for publication
Not applicable

Availability of date and material
Not applicable

Competing interests
The authors declared that they do not have anything to disclose regarding conflict of interest with respect to this manuscript    Low quartile of FMD indicates less than 2.1%. Low quartile of NID indicates less than 6.2%.  in patients with HbA1c of <48 mmol/mol (6.5%) and patients with HbA1c of ≥48 mmol/mol (6.5%).

Supplementary Files
This is a list of supplementary files associated with this preprint. Click to download.