Visceral Fat and Arterial Stiffness in Youth with Healthy Weight, Obesity, and Type 2 Diabetes

Abstract

Background and Objectives:

Visceral fat is associated with increased cardiovascular risk in adults but studies in youth are limited. We assessed associations between visceral fat and arterial stiffness in youth with healthy weight, obesity, and type 2 diabetes, and determined whether relationships were independent of clinical estimates of body fatness.

Methods:

This cross-sectional sample included youth ages 10 to 23 years (67% female, 56% non-black) with healthy weight (BMI=5^th to 85^th percentile, n=236), obesity (BMI≥95^th percentile, n=224), and type 2 diabetes (BMI≥95^th percentile, n=145). Visceral fat was assessed via dual-energy X-ray absorptiometry. Carotid-femoral pulse wave velocity (PWV) was assessed via applanation tonometry. Obesity and type 2 diabetes groups were combined for final analyses. Analyses accounted for age, sex, ancestry, and mean arterial pressure.

Results:

Visceral fat and PWV were greater in youth with obesity vs. healthy weight (p<0.001). In youth with obesity, but not healthy weight, visceral fat was positively associated with PWV (p<0.001), and was predictive of PWV beyond BMI and waist circumference.

Conclusions:

Visceral fat likely contributes to subclinical cardiovascular complications in youth. Since cardiovascular health tracks from adolescence to adulthood, longitudinal studies in youth with obesity are required to define the role of visceral fat in lifelong cardiovascular disease risk.

Introduction

Cardiovascular disease (CVD) is predominantly an adult-onset condition, but cardiac and vascular changes resulting in increased CVD risk likely begin in childhood, especially among youth with obesity. These manifestations may be even more evident in youth with type 2 diabetes.¹ Stiffening of vascular structures, for example, is a predictor of cardiovascular events in adults and is attributed to a myriad of factors.² Adult studies suggest that alterations in arterial compliance are associated with the patterning of body fat deposition, and that excess abdominal visceral fat confers the greatest adverse effect on CVD risk, independent of traditional cardiometabolic risk factors.^{3, 4} Associations between visceral fat and cardiovascular outcomes in children and adolescents have been conflicting,⁵ likely due to the lack of studies in individuals with obesity. Since youth with obesity are at increased risk for CVD compared to healthy weight peers,^{1, 6, 7} this paucity of evidence represents a major gap in knowledge that requires attention. In the clinical arena, cardiometabolic risk is typically evaluated using anthropometric estimates of adiposity such as the body mass index (BMI) and waist circumference. However, the degree to which these estimates accurately represent the pathogenicity of visceral adiposity in youth with varying body weight is unknown. Accordingly, it is important to determine whether visceral fat associates with cardiovascular measures independent of standard clinical estimates of body fatness.

The ‘CVD in Adolescents with Type 2 Diabetes Study’ (i.e., “T2CVD”) was a cross-sectional case-control study aimed at identifying effects of obesity and youth-onset type 2 diabetes on risk factors for and subclinical measures of CVD. This study yielded valuable insights regarding obesity- and type 2 diabetes-related effects on CVD risk in youth,^{1, 8, 9} but visceral fat has not been studied in this cohort with respect to cardiovascular outcomes. The primary aims of this study were to assess associations between visceral fat and arterial stiffness in youth with healthy weight, obesity, and type 2 diabetes, and to determine whether relationships were independent of standard clinical estimates of body fatness (i.e., BMI and waist circumference). Since the health effects of abdominal obesity are hypothesized to differ between fat depots,¹⁰ relationships with abdominal subcutaneous fat were also evaluated.

Methods

Study participants.

This was a secondary analysis of cross-sectional data in youth with healthy weight (n=236), obesity (n=224), and type 2 diabetes (n=145). Subjects were ages 10 to 23 years (67.0% female), and classified as either black or non-black based on self-report (56.4% black). Subjects in the healthy weight or obese groups with a fasting glucose >126 mg/dL or hemoglobin A1c >6.5%, and subjects in the type 2 diabetes group with a BMI <95^th percentile were excluded from this study. Only subjects with a valid measure of visceral fat were included in this study (n=605). The Institutional Review Board for Human Subjects at Cincinnati Children’s Hospital Medical Center approved all study protocols and procedures. All subjects/guardians provided written informed consent/assent.

Anthropometrics.

Height (cm) and weight (kg) were measured using a wall-mounted stadiometer (Veeder-Rood, Elizabethtown, NC) and electronic scale (Health-O-Meter, model 770; SECA, Hanover, MD), respectively. BMI (kg/m²) was calculated, and height, weight, and BMI Z-scores were computed.¹¹ For individuals with extreme BMI values, modified BMI Z-scores were computed per recommendations from the CDC.¹² Waist circumference was assessed in duplicate using a flexible measuring tape at the level of the umbilicus according to the protocol used in the National Heart, Lung, and Blood Institute Health Institute’s Growth and Health Study,¹³ and the two measurements were averaged.

Arterial stiffness.

Pulse wave velocity (PWV; m/sec) was assessed using the Sphygmo-Cor SCOR-PVx System (Atcor Medical, Sydney, Australia) as described in detail previously.¹⁴ Arterial waveforms were assessed using a hand-held pressure tonometer at the central (carotid artery) and distal (femoral artery) arteries of interest. Additional measures were performed upon the radial artery for assessment of central mean arterial pressure (MAP). Measurements were acquired in triplicate, and averaged. Reproducibility of arterial stiffness measures was excellent.⁶

Visceral and subcutaneous fat.

Abdominal visceral and subcutaneous fat (cm²) were assessed using a Hologic QDR 4500A DXA device (Hologic, Inc., Bedford, MA). Whole body scans were acquired using instrument-specific software and procedures, and scans were analyzed using Apex software (version 5.5.3). Multiple studies have reported that DXA provides reliable and valid estimates of visceral fat in youth and adults.^15–17

Blood biochemistries.

On the morning following an overnight fast, blood draws were performed. Hemoglobin A1c (HbA1c), insulin, and glucose were assessed using standard clinical techniques.¹⁸

Statistical Analysis.

All analyses were performed using STATA v15.1. All variables were inspected for normality and implausible data points. Some variables had non-normal distributions so transformations were applied to achieve an approximately normal distribution. Carotid-femoral PWV and visceral fat were log transformed, and subcutaneous fat was square root transformed. P-values <0.05 were considered statistically significant.

Between-group comparisons of continuous variables were conducted using analysis of variance or Kruskal-Wallis/Dunn’s tests with Bonferroni post-hoc adjustment to account for multiple comparisons. Between-group comparisons of categorical variables were performed using Pearson’s chi-square tests.

Associations between abdominal fat and PWV were assessed using Pearson’s bivariate correlations and multiple linear regression. Pearson’s correlations were performed in the three groups separately, and 95% confidence intervals were computed using the ‘corrci’ command (Fisher’s Z transformation) in STATA. Linear regression analyses were performed in the total cohort. Since our study sample included three groups, preliminary regression analyses investigated group by visceral fat interactions in the prediction of PWV. There was no evidence of a group by visceral fat interaction, so subjects with type 2 diabetes were combined into the ‘obese group’. Final regression models included main and interaction effects for fat depot (visceral or subcutaneous) and group (healthy weight or obesity), as well as ancestry (black or non-black), sex (male or female), age, and MAP as additional predictor variables. Results from the main regression analyses were used to compute regression coefficients for the healthy weight and obese groups separately.

Hierarchical regression was used to determine whether associations between abdominal fat and PWV were independent of BMI or waist circumference. For all analyses, variance inflation factors were inspected, and were not indicative of excessive multi-collinearity. Since our ‘group’ variable (i.e., healthy weight vs. obese) was co-linear with BMI, analyses were performed for each group separately. The first step of the regression analysis fit age, sex, ancestry, MAP, and BMI (or waist circumference), the second step included visceral fat or subcutaneous fat, and the third step included both visceral and subcutaneous fat. Similar analyses were performed to determine whether waist circumference was additively predictive of PWV beyond BMI.

Due to the controversy regarding which BMI metric is best to use in children,¹⁹ we conducted sensitivity analyses using BMI Z-score rather than BMI for all analyses involving visceral fat. In addition, recognizing the importance of accounting for height when using body composition measures in children,²⁰ sensitivity analyses were performed adding height or height Z-score as additional model parameters.

Results

Across the three groups, there were significant differences in ancestry, height Z-score, weight Z-score, BMI Z-score, visceral fat (cm²), subcutaneous fat (cm²), PWV (m/s), MAP (mmHg), HbA1c (%), insulin (mIU/L), and glucose (mg/dL; all P<0.05; Table 1). The obese and type 2 diabetes groups had greater weight Z-score, BMI Z-score, HbA1c, insulin, glucose, visceral fat, subcutaneous fat, PWV, and MAP compared to the healthy weight group (all P<0.01). The type 2 diabetes group had greater height Z-score than the healthy weight group (P=0.013), and greater HbA1c, insulin, glucose, visceral fat, and MAP compared to the obese group (all P<0.05).

Table 1.

Sample Characteristics

	Healthy Weight (n=236)	Obese (n=224)	Type 2 Diabetes (n=145)	p value
Age (years)1	17.72 ± 3.56	18.04 ± 3.20	17.47 ± 3.17	0.175
Female (%)2	63.98	71.43	64.83	0.196
Black (%)2	51.69	66.52	48.28	<0.001
Height (Z-score)1	0.13 ± 0.99	0.21 ± 1.11	0.46 ± 1.18	0.015
Weight (Z-score)3	0.13 (−0.30 – 0.62)	2.33 (2.01 – 2.64)*	2.45 (2.07 – 2.74)*	<0.001
BMI (Z-score)3	0.12 (−0.44 – 0.52)	2.46 (1.81 – 3.18)*	2.73 (1.94 – 3.45)*	<0.001
HbA1c (%)3	5.30 (5.10 – 5.50)	5.50 (5.20 – 5.70)*	6.60 (5.80 – 9.30)*†	<0.001
Insulin (mIU/L)3	10.00 (8.30 – 13.20)	18.30 (13.90 – 27.00)*	22.20 (16.50 – 34.70)*†	<0.001
Glucose (mg/dL)3	89.10 (85.00 – 93.25)	91.35 (87.45 – 96.05)*	111.50 (90.30 – 181.50)*†	<0.001
Visceral fat (cm2)3	32.65 (23.37 – 42.51)	97.14 (77.55 – 122.81)*	122.58 (99.70 – 164.52)*†	<0.001
Subcutaneous fat (cm2)3	159.66 (84.74 – 213.40)	566.07 (454.05 – 683.29)*	567.91 (472.17 – 689.33)*	<0.001
PWV (m/s)3	5.30 (4.93 – 5.73)	6.28 (5.57 – 6.97)*	6.33 (5.63 – 7.30)*	<0.001
MAP (mmHg)3	81.00 (75.50 – 85.00)	86.00 (81.00 – 91.00)*	89.00 (83.00 – 95.00)*†	<0.001
SBP (mmHg)1	107.90 ± 10.37	117.07 ± 11.12*	121.72 ± 11.36*†	<0.001
DBP (mmHg)1	59.71 ± 12.34	66.15 ± 12.01*	66.40 ± 12.51*	<0.001

Values are written as mean ± SD or median (interquartile range), unless otherwise noted. Test of significance performed using

¹ANOVA (with Bonferroni post-hoc adjustment)

²χ² test, and

³Kruskal-Wallis/Dunn’s test (with Bonferroni post-hoc adjustment).

^*Significantly different than healthy weight group (p<0.05).

^†Significantly different than obese group (p<0.05). BMI, body mass index; HbA1c, hemoglobin A1c; PWV, pulse wave velocity; MAP, mean arterial pressure; SBP, systolic blood pressure; DBP, diastolic blood pressure. Sample sizes for PWV (n=227, 198, and 127 for the healthy weight, obese, and type 2 diabetic groups, respectively) and MAP (n=224, 210, and 137 for the healthy weight, obese, and type 2 diabetic groups, respectively).

Pearson’s bivariate correlations evaluating associations between abdominal fat measures and PWV are presented in Table S1 and visually depicted in Figure S1. Visceral fat was positively associated with PWV in the three groups (all P<0.05), and the 95% confidence intervals for the healthy weight, obesity, and type 2 diabetes groups overlapped, supporting a similar association between visceral fat and PWV regardless of group. Subcutaneous fat was positively associated with PWV in the obesity and type 2 diabetes groups (both P<0.05), but not the healthy weight group.

Multiple linear regression was performed to assess the association between visceral fat and PWV in the entire cohort, and to determine whether this relationship differed between youth with healthy weight, obesity, and type 2 diabetes, while accounting for ancestry, sex, age, and MAP. Compared to the healthy weight group, the association between visceral fat and PWV was stronger in the obesity (B=0.13, SE=0.03, P<0.001) and type 2 diabetes groups (B=0.09, SE=0.03, P=0.009). However, since the interaction effect did not differ between the obesity and type 2 diabetes groups (B=0.04, SE=0.04, P=0.245), all subsequent analyses were performed while combining the obesity and type 2 diabetes groups.

We observed a significant interaction between group and visceral fat in relation to PWV, while accounting for ancestry, sex, age, and MAP (P<0.001; Table 2). Visceral fat was positively associated with PWV in the obese group (B=0.14, SE=0.02, P<0.001), but not the healthy weight group (B=0.02, SE=0.02, P=0.283). Similar associations were observed with respect to subcutaneous fat (all P<0.001), which was positively associated with PWV in the obese group (B=0.17, SE=0.03, P<0.001), but not the healthy weight group (B=0.00, SE=0.02, P=0.835). For descriptive purposes, we present the bivariate non-linear association between visceral fat and PWV in the healthy weight and obesity groups in Figure 1.

Table 2.

Linear Regression Assessing the Relationships Between Visceral Fat and Subcutaneous Fat with Pulse Wave Velocity in Youth with Healthy Weight and Obesity

	B	SE	95% CI	p
Visceral Fat
Black race	0.080	0.011	0.059 – 0.101	<0.001
Female sex	0.004	0.011	−0.019 – 0.026	0.743
Age (years)	0.016	0.002	0.013 – 0.019	<0.001
MAP (mmHg)	0.005	0.001	0.003 – 0.006	<0.001
Obese group	−0.438	0.106	−0.646 – −0.231	<0.001
Visceral fat (cm2)	0.019	0.018	−0.016 – 0.055	0.283
Obese group by fat interaction	0.117	0.026	0.067 – 0.167	<0.001
Subcutaneous Fat
Black race	0.054	0.010	0.034 – 0.074	<0.001
Female sex	−0.016	0.014	−0.043 – 0.012	0.256
Age (years)	0.015	0.002	0.012 – 0.018	<0.001
MAP (mmHg)	0.006	0.001	0.004 – 0.007	<0.001
Obese group	−0.952	0.155	−1.256 – −0.648	<0.001
Subcutaneous fat (cm2)	0.003	0.015	−0.026 – 0.032	0.835
Obese group by fat interaction	0.171	0.026	0.120 – 0.221	<0.001

n = 538 for subcutaneous fat analyses and n = 537 for visceral fat analyses. B for the Obese group (coded as 1) represents comparison to the healthy weight group (coded as 0). B, unstandardized regression coefficients; SE, standard error; 95% CI, 95% confidence interval; MAP, mean arterial pressure.

Figure 1.

Scatter plot depicting the bivariate association between visceral fat (log) and PWV (log) in youth with healthy weight (white circles; n=227) and obesity (gray circles; n=325). A loess regression line was fitted to the data from the healthy weight and obese groups combined.

Results from the final model of the hierarchical linear regression analyses testing whether associations between abdominal fat and PWV were independent of standard clinical estimates of body fatness (i.e., BMI and waist circumference) are displayed in Table 3 and Table 4. In youth with obesity, BMI was positively associated with PWV, while accounting for age, sex, ancestry, and MAP (P<0.001). This relationship was maintained after accounting for the positive association with visceral fat (P<0.001), which accounted for an additional 1.6% explained variability in PWV. In contrast, subcutaneous fat did not account for additional variability in PWV beyond BMI. Visceral (P<0.005) but not subcutaneous fat was positively associated with PWV when both were included in the same regression model, and BMI remained positively associated with PWV (P<0.001). For models that included waist circumference rather than BMI, similar associations between visceral and subcutaneous fat and PWV were evident (Table 4). In youth with healthy weight, visceral fat, subcutaneous fat, BMI, and waist circumference were not significantly associated with PWV in any analyses.

Table 3.

Final Models from Hierarchical Linear Regression Analyses Assessing Associations Between Abdominal Fat Measures and PWV Independent of BMI in Youth with Healthy Weight and Obesity

	β	B	SE	95% CI	p	Model R2
Healthy Weight
Visceral Fat						0.356
BMI (kg/m2)	0.012	0.001	0.004	−0.008 – 0.009	0.881
Visceral fat (cm2)	0.050	0.014	0.022	−0.029 – 0.057	0.523
Subcutaneous Fat						0.355
BMI (kg/m2)	0.051	0.003	0.005	−0.007 – 0.012	0.583
Subcutaneous fat (cm2)	−0.025	−0.005	0.027	−0.058 – 0.049	0.855
Visceral Fat and Subcutaneous Fat						0.357
BMI (kg/m2)	0.035	0.002	0.005	−0.008 – 0.012	0.714
Visceral fat (cm2)	0.063	0.018	0.024	−0.029 – 0.064	0.450
Subcutaneous fat (cm2)	−0.065	−0.013	0.029	−0.070 – 0.044	0.657
Obese
Visceral Fat						0.483
BMI (kg/m2)	0.265	0.007	0.001	0.005 – 0.010	<0.001
Visceral fat (cm2)	0.173	0.074	0.024	0.026 – 0.121	0.003
Subcutaneous Fat						0.468
BMI (kg/m2)	0.359	0.010	0.002	0.006 – 0.014	<0.001
Subcutaneous fat (cm2)	−0.013	−0.007	0.047	−0.100 – 0.085	0.879
Visceral Fat and Subcutaneous Fat						0.483
BMI (kg/m2)	0.290	0.008	0.002	0.004 – 0.013	<0.001
Visceral fat (cm2)	0.175	0.074	0.024	0.027 – 0.122	0.002
Subcutaneous fat (cm2)	−0.034	−0.019	0.047	−0.111 – 0.074	0.693

All analyses included sex, ancestry, age, and MAP as additional model parameters. β, standardized regression coefficient; B, unstandardized regression coefficient; SE, standard error; 95% CI, 95% confidence interval; PWV, carotid-femoral pulse wave velocity; BMI, body mass index. n = 315 for the obese group.

Table 4.

Final Models from Hierarchical Linear Regression Analyses Assessing Associations Between Abdominal Fat Measures and PWV Independent of Waist Circumference in Youth with Healthy Weight and Obesity

	β	B	SE	95% CI	p	Model R2
Healthy Weight
Visceral Fat						0.362
Waist circumference (cm)	−0.122	−0.002	0.002	−0.006 – 0.001	0.179
Visceral fat (cm2)	0.125	0.035	0.023	−0.011 – 0.081	0.134
Subcutaneous Fat						0.358
Waist circumference (cm)	−0.118	−0.002	0.002	−0.006 – 0.002	0.272
Subcutaneous fat (cm2)	0.147	0.030	0.030	−0.029 – 0.088	0.318
Visceral Fat and Subcutaneous Fat						0.363
Waist circumference (cm)	−0.166	−0.003	0.002	−0.007 – 0.001	0.145
Visceral fat (cm2)	0.111	0.031	0.024	−0.017 – 0.079	0.198
Subcutaneous fat (cm2)	0.097	0.020	0.030	−0.040 – 0.080	0.521
Obese
Visceral Fat						0.461
Waist circumference (cm)	0.203	0.002	0.001	0.001 – 0.004	<0.001
Visceral fat (cm2)	0.196	0.083	0.026	0.032 – 0.135	0.002
Subcutaneous Fat						0.445
Waist circumference (cm)	0.257	0.003	0.001	0.001 – 0.005	0.009
Subcutaneous fat (cm2)	0.064	0.035	0.058	−0.080 – 0.150	0.546
Visceral Fat and Subcutaneous Fat						0.464
Waist circumference (cm)	0.104	0.001	0.001	−0.001 – 0.004	0.330
Visceral fat (cm2)	0.203	0.087	0.026	0.035 – 0.138	0.001
Subcutaneous fat (cm2)	0.115	0.063	0.059	−0.052 – 0.178	0.282

All analyses included sex, ancestry, age, and MAP as additional model parameters. β, standardized regression coefficient; B, unstandardized regression coefficient; SE, standard error; 95% CI, 95% confidence interval; PWV, carotid-femoral pulse wave velocity. n = 315 for the obese group.

In youth with obesity, age, sex, ancestry, MAP, and BMI accounted for 46.8% of the explained variability in PWV, and BMI was positively associated with PWV (B=0.01, SE=0.00, P<0.001). In contrast, waist circumference did not result in a significant increase in explained variability in PWV (R² change=0.001) when modelled in place of BMI. BMI remained positively associated with PWV (B=0.01, SE=0.00, P<0.001) independent of waist circumference when both estimates of body fatness were included in the model simultaneously. In youth with healthy weight, neither BMI or waist circumference were associated with PWV.

Using BMI Z-score rather than BMI as a model parameter yielded similar associations between visceral fat and PWV. Similarly, including height or height Z-score in models did not modify associations between visceral fat and PWV. Height and height Z-score were not significantly associated with PWV in any models.

Discussion

Adult studies suggest that visceral fat is associated with increased CVD risk,^{3, 21} but few studies have assessed these associations in youth with obesity and related chronic disease. The current study addressed this gap in knowledge by investigating the association between visceral fat and arterial stiffness in youth with healthy weight, obesity, and type 2 diabetes. Increased visceral fat was associated with increased PWV independent of BMI in youth with obesity, but not in youth with healthy weight. In line with the position that effects of excess abdominal fat on cardiovascular health differ between specific fat depots,¹⁰ subcutaneous fat was not independently associated with arterial stiffness when accounting for visceral fat and standard clinical estimates of body fatness (e.g., BMI or waist circumference). Although waist circumference is an easily accessible anthropometric estimate of abdominal obesity, our finding suggest that assessment of visceral fat provides more specific insights with respect to CVD risk in youth with obesity.

Childhood obesity is associated with an increased lifetime CVD risk,²² with the severity of sub-clinical cardiovascular changes being associated with the degree of excess adiposity.¹⁸ Carotid-femoral PWV is a non-invasive measure of vascular stiffness and is associated with cardiovascular event risk in adults.² Studies in adults with overweight and obesity suggest that increased visceral fat assessed via DXA and CT is associated with increased arterial stiffness measured via carotid-femoral PWV.^{3, 4} In comparison, a recent study of children ages 8 to 20 years that were predominantly healthy weight reported positive associations between visceral fat and flow mediated dilation and carotid intima media thickness, but null associations with PWV.⁵ In the current study, we also found null associations between visceral fat and PWV in youth with healthy weight. However, in youth with obesity, individuals with increased visceral fat tended to have greater arterial stiffness. One possible explanation for these differences involves the lower variability in visceral fat and generally healthier cardiometabolic profile in those with healthy weight. Although DXA-derived visceral fat has been validated against MRI in healthy weight and overweight/obese adults,²³ visceral fat stores in healthy weight youth might not be great enough to yield a detectable effect on functional measures of cardiovascular health. As displayed in Figure 1, youth with obesity have increased visceral fat compared to healthy weight peers. For this reason, we are unable to determine whether our associations between visceral fat and PWV in the obese group, but not the healthy weight group, are dependent on obesity status per se, or if these associations are only evident in the presence of excess visceral fat. Nevertheless, these collective results suggest that visceral fat might contribute to increased CVD risk in youth with excess adiposity.

BMI and waist circumference are common in research, clinical practice, and policy for cardiovascular risk assessment, but both measures are limited by their lack of specificity to distinct fat depots. Our results suggest that visceral fat is associated with arterial stiffness in youth with obesity independent of BMI or waist circumference, thereby suggesting that anthropometric estimates of body fatness are not entirely reflective of visceral fat. After accounting for BMI, visceral fat explained an additional 2% of the variability in PWV in obese youth. These cross-sectional findings underscore the need for prospective, longitudinal studies to more clearly define the development of visceral fat during growth and its effect on CVD risk. Interestingly, abdominal subcutaneous fat was not associated with PWV when accounting for other adiposity measures. This supports the position that effects of visceral fat are more pronounced than subcutaneous fat.¹⁰ Similar associations have been shown in adults with obesity, such that visceral but not subcutaneous fat cell volume assessed via abdominal biopsy predicted arterial stiffness.²⁴ Waist circumference is an easily accessible estimate of abdominal adiposity, but is more closely related to subcutaneous rather than visceral fat.²⁵ This might help explain why waist circumference was not additively predictive of PWV beyond BMI, since subcutaneous fat was also not independently associated with PWV.

Youth-onset type 2 diabetes is increasingly recognized, and is a condition characterized by accelerated onset and rapid development of comorbidities.²⁶ Obesity is associated with increased CVD risk, but prior studies using data from the current cohort have described the cardiovascular complications associated with youth-onset type 2 diabetes that are independent of obesity status.^{1, 18} We observed slightly greater visceral fat in youth with obesity and type 2 diabetes compared to youth with obesity and normal glucose control, but the association between visceral fat and arterial stiffness did not differ between these two groups. Based on these results, we suspect that differences in the relative amount of visceral fat, rather than differences in the pathogenicity of visceral fat, contribute to the independent effects of type 2 diabetes on CVD in youth. However, given previously reported associations between intrahepatic fat content and visceral adipose tissue,²⁷ it is also plausible that differences in intrahepatic fat between youth with obesity and youth with obesity and type 2 diabetes may better explain the greater CVD risk noted in youth-onset type 2 diabetes. Indeed, studies in both adults²⁸ and youth²⁹ have reported that intrahepatic fat was a better predictor of cardiometabolic risk than visceral fat alone. As we did not quantify differences in intrahepatic fat in this sample future research should look to explore the relationship between arterial stiffness and intrahepatic fat, independent of visceral fat, in obese youth with and without impaired glucose tolerance. Similarly, since the current study focused specifically on arterial stiffness assessed via carotid-femoral PWV, it is unknown whether associations between visceral fat and other CVD risk factors and subclinical measurements are modified in youth with type 2 diabetes. Additional studies are warranted to further define the cardiovascular implications of excess visceral fat that are unique to youth with diabetes.

A primary limitation of this study was the cross-sectional design, precluding inferences of causality. Prospective studies in youth are required to more clearly define normal patterns of visceral fat accrual during the formative years and subsequent risk for chronic disease. A major strength of this study was the heterogeneous study sample, which included blacks and non-blacks, males and females, and individuals with healthy weight, obesity, and type 2 diabetes. Visceral fat tends to increase during childhood and differs between males vs. females and blacks vs. non-blacks.³⁰ We found that black race, age, and MAP were positively associated with PWV, independent of visceral and subcutaneous fat, and obesity status. Additional research is needed to determine the most appropriate method of accounting for age-, sex-, and ancestry-related trends in visceral fat, as well as to delineate age-, sex-, and ancestry-dependent effects of visceral fat on CVD risk.

Conclusion

This study provides novel insights regarding associations between visceral fat and arterial stiffness, which is a subclinical marker of CVD and a predictor of future cardiovascular events in adults. That visceral fat was associated with increased arterial stiffness in youth with obesity, irrespective of standard clinical estimates of body fatness, suggests that assessment of visceral fat might complement standard clinical measures of body fatness in identifying youth at risk for obesity-related chronic disease. Longitudinal studies are required to determine the role of visceral fat in CVD pathophysiology during the formative years.

Abbreviations:

BMI

Body mass index

DXA

Dual-energy X-ray absorptiometry

PWV

Pulse wave velocity

CVD

Cardiovascular disease

MRI

Magnetic resonance imaging

CT

Computed tomography

MAP

Central mean arterial pressure

HbA1C

Hemoglobin A1c