Accepted Manuscript HbA1c is significantly associated with arterial stiffness but not with carotid atherosclerosis in a community-based population without type 2 diabetes: the Donggu Study Young-Hoon Lee, Min-Ho Shin, Jin-Su Choi, Jung-Ae Rhee, Hae-Sung Nam, Seul-Ki Jeong, Kyeong-Soo Park, So-Yeon Ryu, Seong-Woo Choi, Bok-Hee Kim, Gyung-Jae Oh, Sun-Seog Kweon PII:
S0021-9150(16)30031-4
DOI:
10.1016/j.atherosclerosis.2016.01.032
Reference:
ATH 14464
To appear in:
Atherosclerosis
Received Date: 4 June 2015 Revised Date:
15 December 2015
Accepted Date: 19 January 2016
Please cite this article as: Lee Y-H, Shin M-H, Choi J-S, Rhee J-A, Nam H-S, Jeong S-K, Park K-S, Ryu S-Y, Choi S-W, Kim B-H, Oh G-J, Kweon S-S, HbA1c is significantly associated with arterial stiffness but not with carotid atherosclerosis in a community-based population without type 2 diabetes: the Dong-gu Study, Atherosclerosis (2016), doi: 10.1016/j.atherosclerosis.2016.01.032. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT HbA1c is significantly associated with arterial stiffness but not with carotid atherosclerosis in a community-based population without type 2 diabetes: the Dong-gu Study
RI PT
Running title: HbA1c and subclinical atherosclerosis
Young-Hoon Lee a,b, Min-Ho Shin c, Jin-Su Choi c, Jung-Ae Rhee c, Hae-Sung Nam d, Seul-Ki Jeong e, Kyeong-Soo Park f, So-Yeon Ryu g, Seong-Woo Choi g, Bok-Hee Kim
SC
a
, Gyung-Jae Oh a,b, Sun-Seog Kweon c,i,*
M AN U
h
Department of Preventive Medicine & Institute of Wonkwang Medical Science,
Wonkwang University School of Medicine, Iksan, Jeonbuk, South Korea b
Regional Cardiocerebrovascular Center, Wonkwang University Hospital, Iksan,
Jeonbuk, South Korea
Department of Preventive Medicine, Chonnam National University Medical School,
TE D
c
Gwangju, South Korea d
Department of Preventive Medicine, Chungnam National University College of
e
EP
Medicine, Daejeon, South Korea
Department of Neurology, Chonbuk National University Medical School-Chonbuk
f
AC C
National University Hospital, Jeonju, Jeonbuk, South Korea Department of Preventive Medicine, Seonam University College of Medicine,
Namwon, Jeonbuk, South Korea g
Department of Preventive Medicine, Chosun University Medical School, Gwangju,
South Korea h
Department of Food & Nutrition, Chosun University College of Natural Sciences,
Gwangju, South Korea 1
ACCEPTED MANUSCRIPT I
Jeonnam Regional Cancer Center, Chonnam National University Hwasun Hospital,
Hwasun, South Korea
*
Corresponding author
322 Seoyang-ro, Hwasun 519-763, South Korea
RI PT
Jeonnam Regional Cancer Center, Chonnam National University Hwasun Hospital,
AC C
EP
TE D
M AN U
SC
Tel: +82.61-379-7723, Fax: +82.61-379-7880, E-mail address: [email protected]
2
ACCEPTED MANUSCRIPT Abstract Objectives: We examined the associations between HbA1c levels and various atherosclerotic vascular parameters among adults without diabetes from the general population.
RI PT
Methods: A total of 6,500 community-dwelling adults, who were free of type 2 diabetes and ≥ 50 years of age, were included. High-resolution B-mode ultrasound was used to evaluate carotid artery structure, including intima-media thickness (IMT),
SC
plaque, and luminal diameter. Brachial-ankle pulse wave velocity (baPWV), which is
waveform analysis device.
M AN U
a useful indicator of systemic arterial stiffness, was determined using an automatic
Results: No significant associations were observed between HbA1c, carotid IMT, plaque, or luminal diameter in a fully adjusted model. However, the odds ratio (95% confidence interval) for high baPWV (defined as the highest quartile) increased by
TE D
1.43 (1.19–1.71) per 1% HbA1c increase after adjusting for conventional risk factors in a multivariate logistic regression analysis. In addition, HbA1c was independently associated with baPWV in a multivariate linear regression analysis.
EP
Conclusions: High-normal HbA1c level was independently associated with arterial
AC C
stiffness, but not with carotid atherosclerotic parameters, in the general population without diabetes. Our results suggest that the functional atherosclerotic process may already be accelerated according to HbA1c level, even at a level below the diagnostic threshold for diabetes.
Keywords: Hemoglobin A, Glycosylated; Carotid atherosclerosis; Arterial stiffness; Cross-sectional studies
3
ACCEPTED MANUSCRIPT 1. Introduction HbA1c is a measure of long-term glycemic control, and has been recommended for screening and diagnosing diabetes mellitus [1]. An increase in HbA1c is associated with cardiovascular disease (CVD) in patients with diabetes [2,3]. Studies
RI PT
have shown significant associations between HbA1c and subclinical atherosclerosis factors, such as carotid atherosclerosis [4] and arterial stiffness [5] in patients with diabetes. Epidemiological studies have reported a positive association between
SC
HbA1c and increased risk of CVDs in individuals without diabetes [6,7], whereas other studies have not observed such an association [8,9]. Moreover, it is unclear
population without diabetes.
M AN U
whether HbA1c is associated with subclinical atherosclerotic parameters in the
Several subclinical atherosclerotic markers have been developed. Carotid intimamedia thickness (IMT) and plaque are surrogate markers for subclinical carotid
TE D
atherosclerosis and are strong predictors of future CVD [10,11]. Increased IMT may represent an early phase of atherosclerosis, whereas plaque formation may reflect a later stage of atherosclerotic progression [12]. Carotid arterial diameter correlates
EP
with conventional cardiovascular risk factors, and enlargement of the carotid arteries
AC C
is a good indicator of carotid atherosclerosis [13]. Arterial stiffness is a vascular measure that represents arterial wall thickening and loss of elasticity, and is associated with an increased risk of cardiovascular events [14]. Brachial-ankle pulse wave velocity (baPWV), which is a simple, reproducible, and noninvasive measure of systemic arterial stiffness, has been widely used in community settings and by primary care physicians [15,16]. Until now, few studies have simultaneously investigated the associations between HbA1c and various atherosclerotic vascular disease factors in a large community4
ACCEPTED MANUSCRIPT dwelling population without diabetes. Therefore, we investigated the associations between HbA1c as a marker of long-term glycemic exposure and subclinical atherosclerotic vascular parameters, including carotid atherosclerosis (IMT, plaque, and luminal diameter) and baPWV among adults without diabetes from the general
RI PT
population.
2. Subjects and methods
SC
2.1. Study population
The study population consisted of community-dwelling adults ≥ 50 years of age
M AN U
who participated in the baseline Dong-gu Study survey. The Dong-gu Study is an ongoing prospective population-based cohort study designed to investigate the prevalence, incidence, and risk factors for chronic diseases in urban communities [17]. A total of 9,260 subjects were enrolled in the baseline survey from 2007 to 2010
TE D
(3,711 males and 5,549 females). Of them, 848 subjects with a history of coronary heart disease or stroke were excluded. Subjects taking anti-diabetes medication (n = 975) and subjects with a fasting blood glucose (FBG) ≥ 126 mg/dL or HbA1c ≥ 6.5%
EP
(n = 702) were also excluded. After excluding subjects whose records lacked
AC C
information regarding anthropometric or laboratory measures or poor atherosclerotic parameters (n = 235), 6,500 subjects were included in the final analyses. This study complied with Declaration of Helsinki guidelines. All of the subjects were fully informed of the study content and provided informed consent. This study protocol was approved by the Institutional Review Board of Chonnam National University Hospital [No. I-2008–05–056].
2.2. Measurements and assessments 5
ACCEPTED MANUSCRIPT All of the participants underwent a standardized physical examination performed by a well-trained research staff. Anthropometric measurements were measured while the subjects were lightly dressed without shoes. Body mass index (BMI) was calculated by dividing weight (kg) by height squared (m2). Blood pressure was
RI PT
measured using a customized cuff on the right upper arm and a mercury sphygmomanometer after a 5 min rest in the sitting position. Three consecutive systolic blood pressure (SBP) and diastolic blood pressure (DBP) readings were
SC
recorded at 1 min intervals, and the mean was used for the analysis. Information regarding demographics, smoking habits, alcohol consumption, medical history, and
M AN U
medication was collected from each subject by well-trained interviewers using a standardized questionnaire. A venous blood sample was drawn from all participants after at least a 12 h overnight fast. Lipid profiles and FBG were measured using enzymatic techniques with an automated analyzer (Hitachi-7600, Hitachi Ltd., Tokyo,
TE D
Japan). HbA1c levels were analyzed by high-performance liquid chromatography using the VARIANT II system (Bio-Rad; Hercules, CA, USA). Two well-trained physicians used high-resolution B-mode ultrasonography
EP
(Sonoace 9900; Medison, Seoul, Korea) to identify changes in the carotid arterial wall,
AC C
including IMT, atherosclerotic plaques, and luminal diameter. IMT was determined as the distance from the media–adventitia interface to the intima–lumen interface on the far wall in a region free of plaque. Common carotid IMT (CC-IMT) was measured between the origin of the carotid bulb and a point 10 mm proximal to the common carotid artery, and the carotid bulb IMT (CB-IMT) was measured in the carotid bulb region. Carotid diameter was determined as the mean of the luminal diameter of the left and right common carotid arteries at a point 10 mm proximal to the carotid bulb origin. Carotid plaques were defined as focal structures that encroached into the 6
ACCEPTED MANUSCRIPT lumen by at least 100% of the surrounding IMT value. The CC-IMT, CB-IMT, and carotid diameter values were determined as the mean of the maximum IMT of the left and right, respectively. The highest quartile of CC-IMT, CB-IMT, and carotid diameter was classified as high CC-IMT, high CB-IMT, and high carotid diameter,
RI PT
respectively. The correlation coefficients for between- and within-examiner variability for IMT were 0.86 and 0.90, respectively. The kappa coefficients for between-examiner agreement and within-examiner agreement for carotid plaques
SC
were 0.76 and 0.85, respectively.
M AN U
baPWV was assessed using a noninvasive, automatic oscillometric device (VP1000, Colin, Tokyo, Japan). The distance between baPWV sampling points was calculated automatically according to participant height. We used the mean of the baPWV values for the right and left ankles. The highest baPWV quartile was
TE D
classified as high baPWV.
2.3. Statistical analysis
EP
Statistical analyses were conducted with SPSS ver. 20.0 software (SPSS Inc.,
AC C
Chicago, IL, USA), and a P value < 0.05 was considered statistically significant. Baseline characteristics of the study population according to HbA1c quartiles were compared using analysis of variance for continuous variables and the χ² test for categorical variables. Associations between atherosclerotic indices and various parameters were analyzed using Pearson’s correlation analysis after adjusting for age and sex. Multiple linear regression analysis was used to evaluate the associations between HbA1c and carotid IMT (CC- and CB-), diameter, and baPWV, and the regression coefficient (β) and P value were assessed. In addition, multiple logistic 7
ACCEPTED MANUSCRIPT regression analysis was used to evaluate the associations between HbA1c, carotid plaques, and other atherosclerotic findings, which were defined as the highest quartile (versus other quartiles). Adjusted odds ratios (ORs) and 95% confidence intervals (CIs) were assessed according to HbA1c (per percent). All of the analyses were
RI PT
performed in sequential order in three models: model 1 was adjusted for sex and age; model 2 was further adjusted for BMI, current smoking, and current drinking plus model 1; model 3 was further adjusted for SBP, total cholesterol (TC), high-density
SC
lipoprotein (HDL) cholesterol, and use of anti-hypertensive and anti-dyslipidemic
M AN U
medications plus model 2.
3. Results
3.1. Baseline characteristics of the study population
The study population was composed of 6,500 subjects, of whom 4,103 (63.1%)
TE D
were women. The mean ages of the men and women were 65.9 ± 8.1 and 63.7±8.1 years, respectively (P < 0.001). The clinical and biochemical characteristics of the study population according to HbA1c quartiles are shown in Table 1. Age, BMI, waist
EP
circumference (WC), SBP, DBP, total cholesterol, triglycerides (TGs), FBG,
AC C
proportion of women, current smoking, and use of anti-hypertensive and antidyslipidemic medications were positively correlated with HbA1c quartiles. In contrast, HDL cholesterol and the proportion of current drinkers were negatively associated with HbA1c quartiles. CC-IMT, CB-IMT, carotid diameter, and baPWV significantly increased with increasing HbA1c quartile (Ps < 0.001), whereas no difference and no linear trend was observed in the association between HbA1c quartiles and the proportion of carotid plaque (Table 1).
8
ACCEPTED MANUSCRIPT 3.2. Correlations Correlation analyses were performed after adjusting for sex and age (Table 2). HbA1c was positively correlated with CC-IMT, CB-IMT, carotid diameter, and baPWV. CC-IMT was positively correlated with BMI, WC, SBP, DBP, TC, and TGs,
RI PT
but was negatively correlated with HDL cholesterol. CB-IMT was positively correlated with BMI, WC, and SBP, but was negatively correlated with FBG and HDL cholesterol. Carotid diameter was positively correlated with FBG, BMI, WC,
SC
SBP, DBP, and TG, but was negatively correlated with TC and HDL cholesterol.
M AN U
baPWV was positively correlated with FBG, SBP, DBP, TC, and TG.
3.3. Associations between HbA1c and carotid atherosclerosis indices The associations between HbA1c and carotid atherosclerotic indices were evaluated using multiple linear regression models (Table 3). Although HbA1c was
TE D
positively associated with CC-IMT in model 1, the significant association disappeared after further adjustment (models 2 and 3). Similar patterns were found for the associations between HbA1c and CB-IMT and carotid diameter. HbA1c was
EP
positively associated with CB-IMT in models 1 and 2, and the association between
AC C
HbA1c and CB-IMT was no longer significant after further adjustment (model 3). In addition, although HbA1c was positively associated with carotid diameter in model 1, the significant association disappeared after further adjustment (models 2 and 3). The associations between HbA1c and carotid atherosclerotic findings, which were
defined as carotid plaque and the highest CC-IMT, CB-IMT, and carotid diameter quartiles, were evaluated using multiple logistic regression models (Table 4). Although the OR (95% CI) for high CC-IMT increased by 1.23 (1.06–1.44) per percent of HbA1c, the significant association disappeared in models 2 and 3. In 9
ACCEPTED MANUSCRIPT addition, the OR (95% CI) for high carotid diameter increased by 1.20 (1.03–1.40) per percent of HbA1c, but the significant association disappeared in models 2 and 3. However, no significant association was observed between HbA1c and high CB-IMT in any of the models. The ORs (95% CI) for carotid plaque increased by 1.18 (1.02–
RI PT
1.35) in model 1 and 1.17 (1.02–1.34) in model 2 per percent of HbA1c, respectively. However, the significant association between HbA1c and carotid plaque disappeared in model 3. As a result, none of the carotid atherosclerotic indices was associated with
SC
HbA1c in the fully adjusted model.
M AN U
3.4. Associations between HbA1c and arterial stiffness
The associations between HbA1c and baPWV were evaluated using multiple linear regression models (Table 3). HbA1c was independently and positively associated with baPWV after adjusting for sex and age (P < 0.001). The associations
TE D
between HbA1c and baPWV remained significant after further adjustment in models 2 and 3 (P < 0.001). In the multiple logistic regression analysis, the OR (95% CI) for high baPWV increased by 1.39 (1.18–1.64) per 1% increase in HbA1c after adjusting
EP
for sex and age. The significant association remained after further adjustment; the
AC C
ORs (95% CI) for high baPWV increased by 1.42 (1.20–1.67) in model 2 and 1.43 (1.19–1.71) in model 3 per percent of HbA1c, respectively (Table 4).
4. Discussion
Epidemiological studies have shown that increased HbA1c is associated with CVD in patients with diabetes [2,3] and in individuals without diabetes [6,7]. However, little is known about the underlying mechanisms for the association between HbA1c and CVD. The relationship between HbA1c and subclinical 10
ACCEPTED MANUSCRIPT atherosclerosis might explain the potential mechanism underlying the association between HbA1c and CVD. Many studies have reported a positive relationship between HbA1c and carotid IMT in the population without diabetes, but the results are controversial. HbA1c is
RI PT
significantly associated with carotid IMT among subjects with normal glucose tolerance [18,19]. Similarly, Hung et al. [20] found a significant association between HbA1c and carotid IMT in Chinese subjects without known diabetes. A multi-ethnic
SC
cohort study reported positive associations between HbA1c, common carotid IMT, and internal carotid IMT in a population without diabetes [21]. In addition, a
M AN U
population-based study reported a positive association between HbA1c and carotid IMT in cross-sectional analyses, but no significant association in longitudinal analyses was observed [22]. In contrast, Kowall et al. [23] found that none of the glycemic measures (HbA1c, FBG, or 2 h plasma glucose) were independently
TE D
associated with carotid IMT in a population-based study. We examined the associations between HbA1c within normal range (
S0021-9150(16)30031-4
DOI:
10.1016/j.atherosclerosis.2016.01.032
Reference:
ATH 14464
To appear in:
Atherosclerosis
Received Date: 4 June 2015 Revised Date:
15 December 2015
Accepted Date: 19 January 2016
Please cite this article as: Lee Y-H, Shin M-H, Choi J-S, Rhee J-A, Nam H-S, Jeong S-K, Park K-S, Ryu S-Y, Choi S-W, Kim B-H, Oh G-J, Kweon S-S, HbA1c is significantly associated with arterial stiffness but not with carotid atherosclerosis in a community-based population without type 2 diabetes: the Dong-gu Study, Atherosclerosis (2016), doi: 10.1016/j.atherosclerosis.2016.01.032. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT HbA1c is significantly associated with arterial stiffness but not with carotid atherosclerosis in a community-based population without type 2 diabetes: the Dong-gu Study
RI PT
Running title: HbA1c and subclinical atherosclerosis
Young-Hoon Lee a,b, Min-Ho Shin c, Jin-Su Choi c, Jung-Ae Rhee c, Hae-Sung Nam d, Seul-Ki Jeong e, Kyeong-Soo Park f, So-Yeon Ryu g, Seong-Woo Choi g, Bok-Hee Kim
SC
a
, Gyung-Jae Oh a,b, Sun-Seog Kweon c,i,*
M AN U
h
Department of Preventive Medicine & Institute of Wonkwang Medical Science,
Wonkwang University School of Medicine, Iksan, Jeonbuk, South Korea b
Regional Cardiocerebrovascular Center, Wonkwang University Hospital, Iksan,
Jeonbuk, South Korea
Department of Preventive Medicine, Chonnam National University Medical School,
TE D
c
Gwangju, South Korea d
Department of Preventive Medicine, Chungnam National University College of
e
EP
Medicine, Daejeon, South Korea
Department of Neurology, Chonbuk National University Medical School-Chonbuk
f
AC C
National University Hospital, Jeonju, Jeonbuk, South Korea Department of Preventive Medicine, Seonam University College of Medicine,
Namwon, Jeonbuk, South Korea g
Department of Preventive Medicine, Chosun University Medical School, Gwangju,
South Korea h
Department of Food & Nutrition, Chosun University College of Natural Sciences,
Gwangju, South Korea 1
ACCEPTED MANUSCRIPT I
Jeonnam Regional Cancer Center, Chonnam National University Hwasun Hospital,
Hwasun, South Korea
*
Corresponding author
322 Seoyang-ro, Hwasun 519-763, South Korea
RI PT
Jeonnam Regional Cancer Center, Chonnam National University Hwasun Hospital,
AC C
EP
TE D
M AN U
SC
Tel: +82.61-379-7723, Fax: +82.61-379-7880, E-mail address: [email protected]
2
ACCEPTED MANUSCRIPT Abstract Objectives: We examined the associations between HbA1c levels and various atherosclerotic vascular parameters among adults without diabetes from the general population.
RI PT
Methods: A total of 6,500 community-dwelling adults, who were free of type 2 diabetes and ≥ 50 years of age, were included. High-resolution B-mode ultrasound was used to evaluate carotid artery structure, including intima-media thickness (IMT),
SC
plaque, and luminal diameter. Brachial-ankle pulse wave velocity (baPWV), which is
waveform analysis device.
M AN U
a useful indicator of systemic arterial stiffness, was determined using an automatic
Results: No significant associations were observed between HbA1c, carotid IMT, plaque, or luminal diameter in a fully adjusted model. However, the odds ratio (95% confidence interval) for high baPWV (defined as the highest quartile) increased by
TE D
1.43 (1.19–1.71) per 1% HbA1c increase after adjusting for conventional risk factors in a multivariate logistic regression analysis. In addition, HbA1c was independently associated with baPWV in a multivariate linear regression analysis.
EP
Conclusions: High-normal HbA1c level was independently associated with arterial
AC C
stiffness, but not with carotid atherosclerotic parameters, in the general population without diabetes. Our results suggest that the functional atherosclerotic process may already be accelerated according to HbA1c level, even at a level below the diagnostic threshold for diabetes.
Keywords: Hemoglobin A, Glycosylated; Carotid atherosclerosis; Arterial stiffness; Cross-sectional studies
3
ACCEPTED MANUSCRIPT 1. Introduction HbA1c is a measure of long-term glycemic control, and has been recommended for screening and diagnosing diabetes mellitus [1]. An increase in HbA1c is associated with cardiovascular disease (CVD) in patients with diabetes [2,3]. Studies
RI PT
have shown significant associations between HbA1c and subclinical atherosclerosis factors, such as carotid atherosclerosis [4] and arterial stiffness [5] in patients with diabetes. Epidemiological studies have reported a positive association between
SC
HbA1c and increased risk of CVDs in individuals without diabetes [6,7], whereas other studies have not observed such an association [8,9]. Moreover, it is unclear
population without diabetes.
M AN U
whether HbA1c is associated with subclinical atherosclerotic parameters in the
Several subclinical atherosclerotic markers have been developed. Carotid intimamedia thickness (IMT) and plaque are surrogate markers for subclinical carotid
TE D
atherosclerosis and are strong predictors of future CVD [10,11]. Increased IMT may represent an early phase of atherosclerosis, whereas plaque formation may reflect a later stage of atherosclerotic progression [12]. Carotid arterial diameter correlates
EP
with conventional cardiovascular risk factors, and enlargement of the carotid arteries
AC C
is a good indicator of carotid atherosclerosis [13]. Arterial stiffness is a vascular measure that represents arterial wall thickening and loss of elasticity, and is associated with an increased risk of cardiovascular events [14]. Brachial-ankle pulse wave velocity (baPWV), which is a simple, reproducible, and noninvasive measure of systemic arterial stiffness, has been widely used in community settings and by primary care physicians [15,16]. Until now, few studies have simultaneously investigated the associations between HbA1c and various atherosclerotic vascular disease factors in a large community4
ACCEPTED MANUSCRIPT dwelling population without diabetes. Therefore, we investigated the associations between HbA1c as a marker of long-term glycemic exposure and subclinical atherosclerotic vascular parameters, including carotid atherosclerosis (IMT, plaque, and luminal diameter) and baPWV among adults without diabetes from the general
RI PT
population.
2. Subjects and methods
SC
2.1. Study population
The study population consisted of community-dwelling adults ≥ 50 years of age
M AN U
who participated in the baseline Dong-gu Study survey. The Dong-gu Study is an ongoing prospective population-based cohort study designed to investigate the prevalence, incidence, and risk factors for chronic diseases in urban communities [17]. A total of 9,260 subjects were enrolled in the baseline survey from 2007 to 2010
TE D
(3,711 males and 5,549 females). Of them, 848 subjects with a history of coronary heart disease or stroke were excluded. Subjects taking anti-diabetes medication (n = 975) and subjects with a fasting blood glucose (FBG) ≥ 126 mg/dL or HbA1c ≥ 6.5%
EP
(n = 702) were also excluded. After excluding subjects whose records lacked
AC C
information regarding anthropometric or laboratory measures or poor atherosclerotic parameters (n = 235), 6,500 subjects were included in the final analyses. This study complied with Declaration of Helsinki guidelines. All of the subjects were fully informed of the study content and provided informed consent. This study protocol was approved by the Institutional Review Board of Chonnam National University Hospital [No. I-2008–05–056].
2.2. Measurements and assessments 5
ACCEPTED MANUSCRIPT All of the participants underwent a standardized physical examination performed by a well-trained research staff. Anthropometric measurements were measured while the subjects were lightly dressed without shoes. Body mass index (BMI) was calculated by dividing weight (kg) by height squared (m2). Blood pressure was
RI PT
measured using a customized cuff on the right upper arm and a mercury sphygmomanometer after a 5 min rest in the sitting position. Three consecutive systolic blood pressure (SBP) and diastolic blood pressure (DBP) readings were
SC
recorded at 1 min intervals, and the mean was used for the analysis. Information regarding demographics, smoking habits, alcohol consumption, medical history, and
M AN U
medication was collected from each subject by well-trained interviewers using a standardized questionnaire. A venous blood sample was drawn from all participants after at least a 12 h overnight fast. Lipid profiles and FBG were measured using enzymatic techniques with an automated analyzer (Hitachi-7600, Hitachi Ltd., Tokyo,
TE D
Japan). HbA1c levels were analyzed by high-performance liquid chromatography using the VARIANT II system (Bio-Rad; Hercules, CA, USA). Two well-trained physicians used high-resolution B-mode ultrasonography
EP
(Sonoace 9900; Medison, Seoul, Korea) to identify changes in the carotid arterial wall,
AC C
including IMT, atherosclerotic plaques, and luminal diameter. IMT was determined as the distance from the media–adventitia interface to the intima–lumen interface on the far wall in a region free of plaque. Common carotid IMT (CC-IMT) was measured between the origin of the carotid bulb and a point 10 mm proximal to the common carotid artery, and the carotid bulb IMT (CB-IMT) was measured in the carotid bulb region. Carotid diameter was determined as the mean of the luminal diameter of the left and right common carotid arteries at a point 10 mm proximal to the carotid bulb origin. Carotid plaques were defined as focal structures that encroached into the 6
ACCEPTED MANUSCRIPT lumen by at least 100% of the surrounding IMT value. The CC-IMT, CB-IMT, and carotid diameter values were determined as the mean of the maximum IMT of the left and right, respectively. The highest quartile of CC-IMT, CB-IMT, and carotid diameter was classified as high CC-IMT, high CB-IMT, and high carotid diameter,
RI PT
respectively. The correlation coefficients for between- and within-examiner variability for IMT were 0.86 and 0.90, respectively. The kappa coefficients for between-examiner agreement and within-examiner agreement for carotid plaques
SC
were 0.76 and 0.85, respectively.
M AN U
baPWV was assessed using a noninvasive, automatic oscillometric device (VP1000, Colin, Tokyo, Japan). The distance between baPWV sampling points was calculated automatically according to participant height. We used the mean of the baPWV values for the right and left ankles. The highest baPWV quartile was
TE D
classified as high baPWV.
2.3. Statistical analysis
EP
Statistical analyses were conducted with SPSS ver. 20.0 software (SPSS Inc.,
AC C
Chicago, IL, USA), and a P value < 0.05 was considered statistically significant. Baseline characteristics of the study population according to HbA1c quartiles were compared using analysis of variance for continuous variables and the χ² test for categorical variables. Associations between atherosclerotic indices and various parameters were analyzed using Pearson’s correlation analysis after adjusting for age and sex. Multiple linear regression analysis was used to evaluate the associations between HbA1c and carotid IMT (CC- and CB-), diameter, and baPWV, and the regression coefficient (β) and P value were assessed. In addition, multiple logistic 7
ACCEPTED MANUSCRIPT regression analysis was used to evaluate the associations between HbA1c, carotid plaques, and other atherosclerotic findings, which were defined as the highest quartile (versus other quartiles). Adjusted odds ratios (ORs) and 95% confidence intervals (CIs) were assessed according to HbA1c (per percent). All of the analyses were
RI PT
performed in sequential order in three models: model 1 was adjusted for sex and age; model 2 was further adjusted for BMI, current smoking, and current drinking plus model 1; model 3 was further adjusted for SBP, total cholesterol (TC), high-density
SC
lipoprotein (HDL) cholesterol, and use of anti-hypertensive and anti-dyslipidemic
M AN U
medications plus model 2.
3. Results
3.1. Baseline characteristics of the study population
The study population was composed of 6,500 subjects, of whom 4,103 (63.1%)
TE D
were women. The mean ages of the men and women were 65.9 ± 8.1 and 63.7±8.1 years, respectively (P < 0.001). The clinical and biochemical characteristics of the study population according to HbA1c quartiles are shown in Table 1. Age, BMI, waist
EP
circumference (WC), SBP, DBP, total cholesterol, triglycerides (TGs), FBG,
AC C
proportion of women, current smoking, and use of anti-hypertensive and antidyslipidemic medications were positively correlated with HbA1c quartiles. In contrast, HDL cholesterol and the proportion of current drinkers were negatively associated with HbA1c quartiles. CC-IMT, CB-IMT, carotid diameter, and baPWV significantly increased with increasing HbA1c quartile (Ps < 0.001), whereas no difference and no linear trend was observed in the association between HbA1c quartiles and the proportion of carotid plaque (Table 1).
8
ACCEPTED MANUSCRIPT 3.2. Correlations Correlation analyses were performed after adjusting for sex and age (Table 2). HbA1c was positively correlated with CC-IMT, CB-IMT, carotid diameter, and baPWV. CC-IMT was positively correlated with BMI, WC, SBP, DBP, TC, and TGs,
RI PT
but was negatively correlated with HDL cholesterol. CB-IMT was positively correlated with BMI, WC, and SBP, but was negatively correlated with FBG and HDL cholesterol. Carotid diameter was positively correlated with FBG, BMI, WC,
SC
SBP, DBP, and TG, but was negatively correlated with TC and HDL cholesterol.
M AN U
baPWV was positively correlated with FBG, SBP, DBP, TC, and TG.
3.3. Associations between HbA1c and carotid atherosclerosis indices The associations between HbA1c and carotid atherosclerotic indices were evaluated using multiple linear regression models (Table 3). Although HbA1c was
TE D
positively associated with CC-IMT in model 1, the significant association disappeared after further adjustment (models 2 and 3). Similar patterns were found for the associations between HbA1c and CB-IMT and carotid diameter. HbA1c was
EP
positively associated with CB-IMT in models 1 and 2, and the association between
AC C
HbA1c and CB-IMT was no longer significant after further adjustment (model 3). In addition, although HbA1c was positively associated with carotid diameter in model 1, the significant association disappeared after further adjustment (models 2 and 3). The associations between HbA1c and carotid atherosclerotic findings, which were
defined as carotid plaque and the highest CC-IMT, CB-IMT, and carotid diameter quartiles, were evaluated using multiple logistic regression models (Table 4). Although the OR (95% CI) for high CC-IMT increased by 1.23 (1.06–1.44) per percent of HbA1c, the significant association disappeared in models 2 and 3. In 9
ACCEPTED MANUSCRIPT addition, the OR (95% CI) for high carotid diameter increased by 1.20 (1.03–1.40) per percent of HbA1c, but the significant association disappeared in models 2 and 3. However, no significant association was observed between HbA1c and high CB-IMT in any of the models. The ORs (95% CI) for carotid plaque increased by 1.18 (1.02–
RI PT
1.35) in model 1 and 1.17 (1.02–1.34) in model 2 per percent of HbA1c, respectively. However, the significant association between HbA1c and carotid plaque disappeared in model 3. As a result, none of the carotid atherosclerotic indices was associated with
SC
HbA1c in the fully adjusted model.
M AN U
3.4. Associations between HbA1c and arterial stiffness
The associations between HbA1c and baPWV were evaluated using multiple linear regression models (Table 3). HbA1c was independently and positively associated with baPWV after adjusting for sex and age (P < 0.001). The associations
TE D
between HbA1c and baPWV remained significant after further adjustment in models 2 and 3 (P < 0.001). In the multiple logistic regression analysis, the OR (95% CI) for high baPWV increased by 1.39 (1.18–1.64) per 1% increase in HbA1c after adjusting
EP
for sex and age. The significant association remained after further adjustment; the
AC C
ORs (95% CI) for high baPWV increased by 1.42 (1.20–1.67) in model 2 and 1.43 (1.19–1.71) in model 3 per percent of HbA1c, respectively (Table 4).
4. Discussion
Epidemiological studies have shown that increased HbA1c is associated with CVD in patients with diabetes [2,3] and in individuals without diabetes [6,7]. However, little is known about the underlying mechanisms for the association between HbA1c and CVD. The relationship between HbA1c and subclinical 10
ACCEPTED MANUSCRIPT atherosclerosis might explain the potential mechanism underlying the association between HbA1c and CVD. Many studies have reported a positive relationship between HbA1c and carotid IMT in the population without diabetes, but the results are controversial. HbA1c is
RI PT
significantly associated with carotid IMT among subjects with normal glucose tolerance [18,19]. Similarly, Hung et al. [20] found a significant association between HbA1c and carotid IMT in Chinese subjects without known diabetes. A multi-ethnic
SC
cohort study reported positive associations between HbA1c, common carotid IMT, and internal carotid IMT in a population without diabetes [21]. In addition, a
M AN U
population-based study reported a positive association between HbA1c and carotid IMT in cross-sectional analyses, but no significant association in longitudinal analyses was observed [22]. In contrast, Kowall et al. [23] found that none of the glycemic measures (HbA1c, FBG, or 2 h plasma glucose) were independently
TE D
associated with carotid IMT in a population-based study. We examined the associations between HbA1c within normal range (
