Prediabetes: The Effects on Arterial Thickness and Stiffness in Obese Youth

Abstract

Objective:

Adults with prediabetes are at increased risk to develop cardiovascular disease. Whether prediabetes in adolescents is associated with early markers of cardiovascular disease is not known. We sought to 1) compare the cardiovascular risk profiles among adolescents and young adults with prediabetes with those with normal glucose tolerance and 2) determine whether prediabetes is independently associated with noninvasive measures of arterial thickness and stiffness.

Research Design and Methods:

We studied 102 obese youth with prediabetes and 139 obese youth with normal glucose tolerance in a cross-sectional study. Prediabetes or at increased diabetes risk was diagnosed by a fasting blood glucose level of ≥100 to 125 mg/dL, 2-hour oral glucose tolerance test value of ≥140 to 199 mg/dL, or a hemoglobin A1c (HbA1c) value of ≥5.7% to 6.4%. Arterial thickness and stiffness were measured by carotid intima-media thickness (IMT), augmentation index, pulse wave velocity, and brachial distensibility (BrachD).

Results:

Nearly 50% of youth were diagnosed with prediabetes by HbA1c. Youth with prediabetes had worse metabolic profiles with higher BMI z score, systolic blood pressure, and fasting insulin; increased carotid IMT; and lower BrachD compared with normal glucose-tolerant youth (P < .05). Multivariate regression analysis found prediabetes was a significant determinant of internal carotid IMT and BrachD (P < .05). After excluding youth diagnosed by HbA1c, the prediabetes group was borderline significant for internal carotid IMT (.1 > P ≥ .05) only.

Conclusions:

Youth with prediabetes have worse cardiometabolic risk factors and evidence of increased arterial thickness and stiffness. Interventions to prevent prediabetes are crucial to reduce the development of early arterial disease.

Prediabetes is the intermediate stage between normal glucose tolerance and glucose values consistent with the diagnosis of diabetes. According to the American Diabetes Association, prediabetes or at increased risk of diabetes is defined as impaired fasting glucose with a blood glucose level of ≥100 to 125 mg/dL or impaired glucose tolerance with a blood glucose value of ≥140 to 199 mg/dL at 2 hours by a 75-g oral glucose tolerance test (OGTT) (1). In 2010, the definition of prediabetes expanded to include a hemoglobin A1c (HbA1c) level of ≥5.7% to 6.4% (2–4).

In adults, prediabetes is a risk factor to develop future diabetes, cardiovascular disease (5), and subsequent adverse cardiovascular events including myocardial infarction (6), stroke (7), and death (8). Whether prediabetes in youth is associated with early markers of cardiovascular disease as measured by increased arterial thickness and stiffness is not known.

In this study we sought to 1) compare the cardiovascular risk profiles among youth with prediabetes and obese youth with normal glucose tolerance and 2) determine whether prediabetes is independently associated with early markers of atherosclerosis assessed by noninvasive measures of arterial thickness and stiffness.

Subjects and Methods

Participants

Participants in this analysis were studied as part of a larger cross-sectional study that was designed to evaluate the effects of obesity and type 2 diabetes on the vasculature in adolescents and young adults (Type 2 Diabetes and Cardiovascular Disease Study). Briefly, the Type 2 Diabetes and Cardiovascular Disease Study recruited youth ages 11 to 23 years from clinics at Cincinnati Children's Hospital Medical Center, local physicians' offices, college campuses, and health fairs. The cohort consisted of adolescents with type 2 diabetes and 2 age-, race-, and sex-matched control groups: obese controls (defined as body mass index [BMI] ≥95th percentile) and lean controls (defined as BMI <85th percentile) (9, 10)

For this analysis, only obese individuals were included. We chose to omit a lean control group and those with type 2 diabetes because we have previously published data comparing cardiovascular risk factors and arterial stiffness and thickness among those groups (10, 11). For this manuscript, we elected to focus only on the added effects of prediabetes to obesity. All obese participants underwent a 2-hour OGTT to exclude the presence of type 2 diabetes according to American Diabetes Association criteria (1). Obese youth were considered to have prediabetes if they had either 1) impaired fasting glucose defined as a fasting plasma glucose level between 100 and 126 mg/dL; 2) impaired glucose tolerance 2-hour plasma glucose concentration, measured by a 75-g OGTT, between 140 and 200 mg/dL; or 3) HbA1c value between 5.7% and 6.4%. Normal glucose tolerance obese controls had a BMI ≥95th percentile and a normal fasting glucose level defined as <100 mg/dL, normal glucose tolerance defined as 2-hour OGTT plasma glucose concentration <140 mg/dL, or an HbA1c <5.7%. Any obese participants using metformin were excluded (n = 6).

Written informed consent was obtained from subjects aged 18 years or older or from a parent or guardian with written assent for subjects younger than 18 years of age, according to the guidelines established by the institutional review board and in accordance with the Declaration of Helsinki.

Data collected

Trained personnel measured height and weight twice with the average of each used in the analyses (9, 10). BMI was calculated as weight in kilograms divided by the square of height in meters. Blood pressure (BP) was obtained manually with a sphygmomanometer (Baum Desktop model with V-Lok cuffs) according to the fourth report of the National High Blood Pressure Education Program Working Group on High Blood Pressure in Children and Adolescents (12).

Fasting plasma lipid profiles were measured in a standardized laboratory, with low-density lipoprotein cholesterol calculated by using the Friedewald equation. Fasting plasma glucose was measured by using an Hitachi model 704 glucose analyzer (Roche Hitachi). Plasma insulin was measured by RIA with an anti-insulin serum raised in guinea pigs, ¹²⁵I-labeled insulin (Linco Research, Inc), and a double-antibody method to separate bound from free tracer. HbA1c was measured in red blood cells by using HPLC. C-reactive protein (CRP) was measured with the use of a high-sensitivity ELISA, and adiponectin was measured using an RIA kit (Linco Research, Inc).

Arterial thickness and stiffness measurements

Arterial thickness was assessed in the carotid arteries using high-resolution B-mode ultrasonography (GE Vivid 7 ultrasound imaging system) and a 7.5-MHz linear array transducer. For each subject, the right and left internal, carotid, and bulb far wall segments were examined independently from continuous angles to identify the thickest area of intima-media thickness (IMT) with the average value used in the analysis. A trace technique using a Camtronic Medical System was employed to ascertain the maximum carotid thickness from leading edge (lumen-intima) to leading edge (media-adventitia). This technique has been found to be more reproducible than point-to-point measurements with coefficients of variation for repeat trace readings: 4.1% to 5.3% vs 5.4% to 7.1%, respectively.

Arterial stiffness testing used the 3 methods described below. Pulse wave velocity was measured with a SphygmoCor SCOR-PVx System (AtcorMedical). The distance from a proximal artery (carotid) to distal artery recording site (femoral artery) was measured to the nearest 0.1 cm, twice, and averaged. A tonometer was used to collect proximal and distal arterial waveforms gated by the R-wave on a simultaneously recorded electrocardiogram. Pulse wave velocity was then calculated as the distance from the carotid-to-distal path length divided by the time delay measured between the feet of the 2 waveforms reported in meters per second (13). Pulse wave velocity is based on the principle that the pressure pulse generated by the left ventricular ejection travels at a speed determined by the size, shape, and properties of the artery (14). Thus, a higher pulse wave velocity indicates increased vascular stiffness. Three measurements of pulse wave velocity were obtained on each subject and averaged. Repeat measures show a coefficient of variation of <5.2%.

Augmentation index, which provides information about arterial stiffness and pulse wave reflections (13), was also collected. Augmentation index was collected when the SphygmoCor tonometer was placed over the right radial artery. The device analyzes pulse waves using a generalized transfer function validated in the catheterization laboratory to calculate a central aortic pressure wave (15). The augmentation index is derived from the central pressure waveform by calculating the difference between the main outgoing wave and the reflected wave of the central arterial waveform, expressed as a percentage of the central pulse pressure. The magnitude of the reflected wave represents the increased afterload that the left ventricle must cope with. A higher augmentation index indicates increased vessel stiffness and has been used to predict cardiovascular events in adults (16). Because the augmentation index is affected by heart rate, all values were adjusted to a standard heart rate of 75 beats per minute. Three measurements of augmentation were obtained and averaged. Reproducibility studies in our laboratory demonstrated coefficients of variation of <7.4%.

Three measures of brachial artery distensibility were obtained with a DynaPulse pathway instrument (Pulse Metric). Brachial distensibility is used to assess resting vascular function in a medium muscular artery. It uses brachial artery pressure curves from distensibility arterial pressure signals obtained from a standard BP cuff sphygmomanometer. A lower brachial distensibility indicates increased vascular stiffness and has been shown to be a useful to assess subclinical arteriosclerotic vascular changes in young adults (17). Repeat measures in our laboratory show coefficients of variation of <9.6%.

Statistical analysis

All analyses were performed with SAS version 9.3. Nonnormal laboratory values and arterial stiffness and thickness were log transformed as needed. T tests for 2 groups, ANOVA for 3 groups, or nonparametric equivalent methods were used to compare groups, with values of P < .05 indicating significance. General linear models were constructed to elucidate independent determinants of carotid thickness and stiffness. Full models contained group (prediabetes vs controls), age, race, sex, BMI z score, lipids, BP, fasting insulin values, CRP, and adiponectin levels. Robustness of all models was assessed using the maximum R² technique.

Results

Table 1 lists how participants were classified with prediabetes. Concordance among the 3 diagnostic criteria was poor with κ-coefficients <0.10 across the various methods. Only 1 participant was classified as having prediabetes by all 3 criteria.

Table 1.

How Participants Qualified for Prediabetes^a

	IFG	IGT	HbA1c
IFG	18	4	13
IGT	4	8	9
HbA1c	13	9	52

Abbreviations: IFG, impaired fasting glucose; IGT, impaired glucose tolerance.

^aThe κ-coefficient of agreement between prediabetes diagnosed by IFG and IGT was 0.05, the κ-coefficient between IFG and HbA1c was 0.06, and the κ-coefficient between IGT and HbA1c was 0.07.

Among our obese cohort, youth with prediabetes were similar in age and sex to those with normal glucose tolerance (Table 2). A higher percentage of youth with prediabetes were African American and had higher BMI z score, systolic BP, fasting insulin and glucose, HbA1c, and 2-hour OGTT glucose values compared with youth with normal glucose tolerance (P < .05). Youth with prediabetes also had higher internal carotid artery IMT and lower brachial distensibility, indicating increased arterial thickness and stiffness (P < .05).

Table 2.

Demographic, Clinical, and Vascular Measures^a

Variable	Obese With Glucose Tolerance, n = 139	Prediabetes, n = 102	P Value
Age, y	18.3 ± 3.0	18.0 ± 3.6	.59
Sex (female), n (%)	100 (72)	65 (64)	.17
Race (African American), n (%)	84 (60)	80 (78)	<.01
Smoking (yes), n (%)	5 (4)	3 (3)	.78
Weight, kg	101.0 ± 19.6	107.0 ± 22.7	.03
BMI, kg/m2	36.2 ± 6.1	38.4 ± 7.9	.06
BMI z score	2.1 ± 0.3	2.2 ± 0.3	<.01
Systolic BP, mm Hg	114 ± 1	116 ± 1	.03
Diastolic BP, mm Hg	67 ± 12	66 ± 11	.70
Total cholesterol, mg/dL	172.0 (149.0, 194.0)	167.5 (148.0, 191.0)	.91
LDL cholesterol, mg/dL	102.5 (84.5, 119.5)	102.0 (86.0, 127.0)	.71
HDL cholesterol, mg/dL	46.5 (42.0, 54.0)	45.0 (40.0, 51.0)	.17
Triglycerides, mg/dL	86.0 (61.0, 116.0)	86.0 (62.0, 114.0)	.60
Fasting insulin, mIU/mL	16.3 (11.8, 22.5)	21.1 (15.3, 31.5)	<.01
CRP, mg/L	2.6 (1.1, 6.2)	2.8 (1.4, 7.0)	.26
Adiponectin, μg/mL	5.7 (4.0, 8.3)	5.1 (3.6, 7.1)	.06
Fasting glucose, mg/dL	90.1 (86.3, 93.4)	95.1 (89.2, 100.8)	<.01
2-h OGTT glucose, mg/dL	102.4 (91.2, 116.7)	115.3 (98.7, 134.1)	<.01
HbA1c, %	5.3 (5.1, 5.5)	5.7 (5.6, 5.9)	<.01
Common carotid, mm	0.5 (0.4, 0.5)	0.5 (0.4, 0.5)	.20
Bulb carotid, mm	0.5 (0.4, 0.6)	0.5 (0.4, 0.6)	.98
Internal carotid, mm	0.4 (0.3, 0.5)	0.4 (0.4, 0.5)	.04
Augmentation index, %	2.5 ± 11.2	2.0 ± 11.9	.76
Pulse wave velocity, m/s	6.2 (5.6, 6.8)	6.4 (5.5, 7.2)	.30
Brachial distensibility, mm/mm Hg	5.5 (4.9, 6.4)	5.3 5.2 (4.6, 5.8)	.01

Abbreviations: HDL, high-density lipoprotein; IFG, impaired fasting glucose; IGT, impaired glucose tolerance; LDL, low-density lipoprotein.

^aValues are presented as mean ± SD for t test and median (1st, 3rd quartile) for Wilcoxon rank-sum test.

Cardiometabolic variables were compared by prediabetes group in Table 3. Only individuals diagnosed by 1 criterion are compared in this table. The 3 groups differed only in triglyceride levels (P < .05). There were no differences in age, sex, BP, lipid profiles, or fasting insulin, CRP, or adiponectin levels among the 3 prediabetes groups.

Table 3.

Clinical Characteristics by Prediabetes Group^a

Variable	IFG	IGT	HbA1c	P Value
n	18	8	52
Age, y	19.3 ± 2.9	19.8 ± 2.0	17.6 ± 3.9	.09
Sex (female), n (%)	8 (44%)	6 (75%)	37 (71%)	.10
Race (African American), n (%)	11 (61%)	3 (38%)	48 (92%)	<.01
Weight, kg	110.7 ± 21	105 ± 16	105 ± 25	.67
BMI z score	2.2 ± 0.4	2.2 ± 0.4	2.2 ± 0.3	.69
Systolic BP, mm Hg	124 ± 7	114 ± 12	118 ± 11	.05
Diastolic BP, mm Hg	69 ± 8	66 ± 9	66 ± 12	.57
Total cholesterol, mg/dL	179.0 (166.0, 189.0)	171.0 (151.5, 188.0)	160.0 (147.5, 188.0)	.43
HDL cholesterol, mg/dL	45.0 (39.0, 50.0)	47.0 (38.0, 55.0)	45.5 (41.0, 51.5)	.97
LDL cholesterol, mg/dL	101.0 (90.0, 120.0)	95.5 (80.5, 107.5)	102.0 (84.0, 122.5)	.57
Triglycerides, mg/dL	92.5 (84.0, 147.0)	125.0 (77.0, 192.5)	74.0 (57.0, 93.5)	<.01
Fasting insulin, mIU/mL	16.5 (13.4, 23.4)	16.8 (15.1, 21.3)	22.1 (15.9, 31.0)	.15
CRP, mg/L	1.6 (1.0, 3.6)	4.6 (2.8, 6.1)	3.1 (1.4, 6.8)	.12
Adiponectin, μg/mL	5.1 (3.7, 7.5)	5.4 (3.5, 8.6)	5.4 (3.7, 7.5)	.88

Abbreviations: HDL, high-density lipoprotein; IFG, impaired fasting glucose; IGT, impaired glucose tolerance; LDL, low-density lipoprotein.

^aValues are presented as mean ± SD for ANOVA or median (1st, 3rd quartile) for nonparametric ANOVA.

To estimate the independent effects of prediabetes on arterial thickness and stiffness measurements, we constructed a multivariate regression model for each of the outcomes that included group (prediabetes vs normal glucose tolerance) and adjusted for differences in age, race, sex, BMI z score, lipids, BP, fasting insulin values, CRP, and adiponectin levels (Table 4). We found that prediabetes was a significant determinant of internal carotid artery (β = 0.055, P < .05) and brachial distensibility (β = −0.048, P < .05). Additional cardiovascular risk factors that were independent determinants of the internal carotid artery included older age and male sex. Age and sex explained 12% of the variance of the internal carotid wall thickness (P < .0001). The addition of prediabetes to the model increased the model R² to 14%. For brachial distensibility, Caucasian race and higher systolic BP and BMI z score explained 21% of the variance. The addition of prediabetes changed the model R² by 2%.

Table 4.

Determinants of Arterial Thickness and Stiffness Measures^a

Variables	Common Carotid	Bulb Carotid	Internal Carotid	Augmentation Index	Pulse Wave Velocity	Brachial Distensibility
R2	0.12	0.07	0.14	0.12	0.52	0.23
Intercept	−0.828	−0.966	−1.312	−0.202	0.695	2.526
Age	0.010	0.012	0.019		0.023
BMI z score	−0.065				0.127	−0.152
Systolic BP				0.002	0.002	−0.004
Diastolic BP					0.002
Sex (male)	0.085	0.088	0.094	−0.041	−0.083
Race (African American)	0.060				0.051	0.052
Prediabetes			0.055			−0.048

^aOnly significant (P < .05) parameter estimates are listed.

Additional regression models were constructed for internal carotid IMT and brachial artery distensibility using each of the 3 prediabetes diagnostic criteria as separate group variables to determine whether one specific group showed a stronger association for increased arterial thickness or stiffness. Similar adjustments were made as described above. For brachial distensibility, HbA1c was an independent predictor of a stiffer vessel (β = −0.064, P < .05). For the internal carotid artery, both HbA1c and abnormal glucose tolerance demonstrated a trend toward a thicker IMT (P < .10).

Finally, given that 52 youth in this study were classified as prediabetic by HbA1c criteria, we also reran our regression models excluding the HbA1c group to determine whether prediabetes was still an independent determinant of internal carotid IMT and/or brachial distensibility. In these models with only 50 patients, prediabetes was borderline significant for internal carotid IMT (.1 > P ≥ .05) only.

Discussion

We demonstrate that youth with prediabetes have worse cardiometabolic risk factors and increased arterial thickness and stiffness compared with their obese counterparts with normal glucose tolerance. Using multivariate regression analysis, we also show prediabetes is an independent risk factor for early atherosclerosis. This suggests prediabetes imparts some additional influence on the vasculature that is not entirely explained by traditional cardiovascular risk factors. These findings in adolescents and young adults are important given that the effects of prediabetes on the vasculature in youth have not been studied and adults with prediabetes or at increased risk to develop diabetes are known to be at high risk to develop myocardial infarction (6), stroke (7), and death (8).

Prediabetes in adults is associated with increased BMI and BP and an abnormal lipid profile (18–22). A study conducted by van Vliet et al (23) in The Netherlands found adolescents with prediabetes (mean age 11.4 ± 3.2years) also have higher insulin levels and BP and lower high-density lipoprotein cholesterol compared with normoglycemic controls. These studies support our findings. We demonstrate worse cardiometabolic profiles characterized by higher BMI z score, systolic BP, and insulin levels among youth with prediabetes compared with their peers with normal glucose tolerance. This is not surprising given that prediabetes in an intermediate stage between normal glucose tolerance and overt diabetes and that diabetes is associated with a worse metabolic profile even in adolescence (10, 24)

Adult studies have shown that prediabetes is independently associated with increased arterial thickness and stiffness (18–22, 25). Specifically, Li et al (19) and Aydin et al (18) found impaired glucose tolerance but not fasting glucose independently contributes to increased arterial thickness and stiffness as measured by mean carotid intima thickness (average of common, internal, and bulb segments) and brachial pulse wave velocity, respectively. More recently, Marini et al (20) included the HbA1c diagnostic criteria and compared the 3 methods used to diagnose prediabetes and similarly found impaired glucose tolerance, but not impaired fasting glucose or increased HbA1c, was the only criterion to be independently associated with common carotid intima media thickness.

Our study extends the effects of prediabetes on the vasculature to the adolescent and young adult population. Similar to adult studies, we demonstrate prediabetes is an independent cardiovascular risk factor for early arterial disease in youth and is associated with an increased risk to develop early arterial disease compared with obesity alone. Another important finding is that when we compared the 3 methods to diagnose prediabetes, we found HbA1c was an independent determinant of arterial stiffness (brachial distensibility) and we observed a trend in the internal carotid artery when prediabetes was diagnosed by HbA1c and impaired glucose tolerance. The discrepant results among our study in youth and previous work in adults as to which type of prediabetes has the largest effect on arterial thickness and stiffness may be due to the small sample size in our impaired fasting glucose and impaired glucose tolerance groups, differences in the vascular sites studied because none of the adult studies examined brachial distensibility or it is also possible that the prediabetes affects the vasculature in youth differently from adults.

We found prediabetes was an independent determinant of brachial distensibility and not augmentation index or pulse wave velocity. The latter 2 vascular measurements are more closely related to central arterial stiffness, while brachial distensibility is a resting measure of a medium muscular artery (16, 26). Previous work in adults has shown that atherosclerosis develops in a nonuniform fashion throughout the arterial tree with different arteries affected by cardiovascular risk factors in different fashions (27, 28). Thus, it is possible that prediabetes is affecting peripheral vasculature earlier than central vasculature.

This may similarly explain why prediabetes showed a trend to be an independent determinant for the internal carotid artery and not the common carotid or the bulb. In previous work, our group has shown that compared with lean, healthy youth, obese youth have increased internal carotid IMT but not increased common carotid thickness. It is postulated that arterial thickness may develop earlier at sites of turbulence (ie, the internal carotid and the bulb) with the common carotid being affected later (29). Given the cross-sectional nature of this study, we are unable to determine this. Longitudinal studies are needed.

We show poor concordance among the 3 methods used to diagnose prediabetes. This has been demonstrated in multiple adult studies (4, 30–33) and recently in 2 large pediatric cohorts (34, 35). The addition of HbA1c to the diagnostic criteria of prediabetes is postulated to lead to a false-positive diagnosis of prediabetes in some ethnic groups. Hispanics, American Indians, African Americans, and Asians are known to have higher HbA1c levels for any given fasting plasma glucose compared with Caucasians even after adjusting for confounding factors that influence HbA1c such as higher red blood cell turnover and differences in glycosylation (36). In contrast, in non-Hispanic whites, HbA1c is a less sensitive method to diagnose prediabetes compared with impaired fasting glucose and impaired glucose tolerance (31). In our study, nearly half of the participants were diagnosed with prediabetes only by HbA1c criteria. Ninety-two percent of these youth were African American. In this group, there may be a higher proportion of conditions that cause red blood cell turnover (ie, anemia or genetic variations) that we were unable to account for. For this reason, we excluded the HbA1c group and reran our regression models to determine whether prediabetes defined only by impaired fasting glucose and impaired glucose tolerance was a significant independent determinant of our vascular outcomes. In these models, prediabetes was borderline significant. However, the sample size was greatly reduced, which likely influenced our results. Thus, our results support the findings of others that suggest in minority ethnic groups, HbA1c alone may not be a reliable marker of prediabetes and more than 1 criterion should be used to diagnose prediabetes, or a race-specific cutoff needs to be developed.

There are some limitations that should be mentioned. First, our cross-sectional study does not allow us to determine the time sequence of the development of arterial changes in youth with diabetes. Second, we defined prediabetes and normal glucose tolerance within a preexisting cohort of obese adolescents. This limited our ability to match participants in a case/control manner and then evaluate the effects of prediabetes. However, adjustments in the regression models were performed to account for group differences. Lastly, we studied a small group of participants with only a single measure of fasting glucose, 2-hour glucose, and HbA1c; therefore, larger studies with repeated measures of glucose and HbA1c are needed. Despite these limitations, this is the first study to evaluate the effects of prediabetes on arterial stiffness and thickness in a relatively large cohort of youth.

We conclude that before the onset of overt diabetes, adolescents and young adults with prediabetes have anatomic and physiologic changes that are consistent with early arterial disease. These abnormalities are largely explained by traditional cardiovascular risk factors but prediabetes appears to impart some small additional cardiovascular risk. Continued interventions aimed at obesity and its associated risk factors before the onset of prediabetes are crucial to prevent the early arterial changes seen in prediabetes.