Organ fat in Latino youth at risk for type 2 diabetes

Abstract

PURPOSE:

Obesity in youth increases the risk for type 2 diabetes (T2D) and elevated abdominal adipose tissue and organ fat may be particularly deleterious. The purpose of this study was to examine associations among measures of adiposity including total, visceral, and organ fat (hepatic and pancreatic) and whether these measures were independently associated with glycemia in Latino youth at risk for diabetes.

METHODS:

Latino adolescents (47 boys and 32 girls, 13.7±1.4 years) with obesity (BMIz 2.3±0.3) were assessed for total fat by DXA and visceral and organ fat by 3 Tesla magnetic resonance imaging. Glycemic indicators included HbA1c, fasting glucose (FG), and 2-hour glucose (2-HrG) following an oral glucose tolerance test. Pearson correlations and stepwise linear regression analyses controlling for age and sex were used to examine independent associations between adiposity and glycemia.

RESULTS:

Total fat was associated with visceral (r=0.66, p=0.001) and hepatic fat (r=0.34, p<0.01) while visceral fat was associated with hepatic (r=0.42, p<0.001) and pancreatic fat (r=0.36, p<0.001). In stepwise linear regression analysis, hepatic and pancreatic fat were significant predictors of FG, explaining 4.7% and 5.2% of the variance, respectively (total R²=0.14, p=0.02). Hepatic fat was the only significant predictor of 2-HrG explaining 9.9% of the variance in the model (total R²=0.12, p=0.03). No measure of adiposity was retained as a significant predictor of HbA1c.

CONCLUSION:

Hepatic and pancreatic fat were the only adiposity measures independently associated with glycemia but the small amount of variance explained underscores the need for additional T2D biomarkers in high risk youth.

Background

Latino youth are disproportionately impacted by obesity contributing to an increased risk for type 2 diabetes (T2D).¹ In addition to total adiposity, elevated regional fat distribution and ectopic fat are thought to exacerbate T2D risk in Latino youth.²

Among Latino adolescents with obesity, visceral fat is independently associated with insulin resistance³ and prospectively associated with prediabetes.⁴ Visceral fat has also been associated with hepatic fat in this population but visceral and hepatic fat may contribute to obesity-related disease risk through different mechansisms.⁵ In addition to visceral and hepatic fat, pancreatic fat is an emerging fat depot that is associated with various measures of adiposity⁶ as well as insulin and glucose metabolism in youth.⁷ Further, both hepatic and pancreatic fat have been associated with prediabetes in adolescents with obesity after controlling for visceral fat.⁸ Interestingly, the associations between organ fat and hyperglycemia varied by ethnicity with hepatic fat being a significant predictor of prediabetes in Latino youth whereas pancreatic fat was a significant predictor of prediabetes among African-American youth.⁸ Ethnic differences in total, regional, and organ fat may contribute to ethnic-specific pathways underpinning T2D-related disparities,² therefore, it is important to understand if and how various adipose depots are related to each other and contribute to hyperglycemia within sub-populations at high risk for T2D such as Latino youth with obesity.

The purpose of this study was to examine the associations between measures of adiposity including total, visceral, and organ fat (hepatic and pancreatic) and whether these measures were independently associated with glycemic indicators among high-risk Latino youth. We hypothesized that these fat depots would be related to each other and would be independently associated with measures of hyperglycemia.

Methods

Participants:

Seventy-nine Latino youth ages 12-16 with obesity (BMI≥95^th percentile for age and sex) were recruited as part of a larger randomized, controlled lifestyle intervention trial with baseline data used for the present analysis.⁹ Informed consent was obtained from a parent or legal guardian and the youth provided assent. The Institutional Review Boards at the Arizona State University and Phoenix Children’s Hospital approved study procedures.

Clinical assessments:

Participants reported to the Arizona State University Clinical Research Center after an overnight fast (>8 hours) for assessment of anthropometrics and glycemic measures. The evening prior to the visit, participants were contacted to remind them not to eat or drink after 10:00 pm and fasting was confirmed verbally upon arrival. Height and weight were measured to the nearest 0.1 cm and 0.1 kg and used to calculate BMI, BMI percentile, and BMIz. Waist circumference (WC) was measured in triplicate to the nearest 0.1cm at the level of the umbilicus. A blood sample was collected to assess HbA1c and fasting glucose (FG) followed by a standard, 75-gram Oral Glucose Tolerance Test to assess 2-hour glucose (2-HrG). Participants with T2D according to the American Diabetes Association criteria were excluded from the study.

Adiposity measures:

Total fat mass was assessed using Dual-energy X-Ray Absorptiometry (DXA) using the GE Lunar iDXA (GE Lunar, Madison, WI). Visceral fat volume as well as hepatic and pancreatic fat fraction were measured by 3 Tesla magnetic resonance imaging using a Philips Ingenia® platform. A 3D quantitative chemical shift-encoded water-fat (mDIXON) technique was used to generate voxel-wise fat fraction maps across the abdomen. The scans were acquired with breath-holds using a multi-channel torso coil array 5mm contiguous slices. Post-processing of the subsequent data was completed by a single analyst with 15 years of experience (HHH) using SliceOmatic (Tomovision, Inc., Montreal, Québec, Canada) software. Region-of-interest (ROIs) analysis was performed manually to determine hepatic and pancreatic fat fractions. For the liver, ROIs on every slice that exhibited the liver were drawn, using co-registered water-only images as a guide. ROIs avoided major hepatic vessels and biliary structures. The mean fat fraction of all ROIs across the liver was computed. Similarly, for the pancreas, ROIs were manually delineated across the organ on each individual 5mm slice, and the average of all slices was computed.

Statistical analysis

Descriptive data are presented as mean±SD. Pearson correlation coefficients were used to examine bivariate associations among adiposity measures (including anthropometrics) as well as associations between adiposity and glycemia. Stepwise linear regression analyses accounting for age and sex were performed to examine which measures of adiposity were significant independent predictors of FG, 2-HrG, and HbA1c. Statistical analyses were performed using SPSS Version 27.0 and a p-value ≤0.05 was considered significant.

Results

The sample included 47 boys and 32 girls with a mean age of 13.7±1.4 years, BMI of 34.1±5.5 kg/m² (BMIz=2.3±0.3), waist circumference 107.9±13.7 cm, total fat mass 40.6±12.4 kg, hepatic fat fraction 9.2±7.9%, pancreatic fat fraction, 4.6±4.2%, HbA1c of 5.7±0.3%, FG of 102±7 mg/dl, and 2-HrG of 144±29 mg/dl. Table 1 summarizes correlations among adiposity measures and with glycemic indicators. BMIz and waist circumference were significantly correlated with total fat mass, visceral fat volume, and hepatic fat fraction. Total fat was correlated with visceral fat volume and hepatic fat fraction while visceral fat volume was correlated with hepatic and pancreatic fat fraction. In terms of associations between adiposity and glycemia, FG was significantly correlated with hepatic and pancreatic fat fraction whereas 2-HrG was significantly correlated with hepatic fat fraction. HbA1c was not significantly correlated with any measure of adiposity.

Table 1.

Pearson correlations among various measures of adiposity and glycemia

	Waist Circumference	Total Fat Mass	Visceral Fat	Hepatic Fat	Pancreatic Fat	Fasting Glucose	2-Hour Glucose	HbA1c
BMIz	r=0.84, p<0.001	r=0.88, p<0.001	r=0.64, p<0.001	r=0.38, p=0.001	r=0.20, p=0.08	r=0.09, p=0.42	r=−.01, p=0.96	r=.03, p=0.78
Waist Circumference	----	r=0.92, p<0.001	r=0.68, p<0.001	r=0.40, p<0.001	r=0.21, p=0.05	r=0.07, p=0.53	r-0.07, p=0.53	r=0.08, p=0.49
Total Fat Mass		----	r=0.66, p<0.01	r=0.34, p<0.01	r=0.15, P=0.19	r=0.06, p=0.59	r=−0.09, p=0.46	r=0.06, p=0.62
Visceral Fat Volume			----	r=0.42, p<0.001	r=0.36, p<0.001	r=0.18, p=0.12	r=0.11, p=0.35	r=−0.01, p=0.94
Hepatic Fat Fraction				-----	r=0.15, P=0.2	r=0.29, p<0.01	r=0.29, p=0.01	r=0.19, p=0.1
Pancreatic Fat Fraction					-----	r=0.27, p=0.02	r=0.16, p=0.15	r=0.03, p=0.78

Significant correlations noted in bold

Table 2 summarizes the results of stepwise linear regression models to examine the independent contributions of adiposity measures to glycemic indicators. Hepatic and pancreatic fat fraction were significant independent predictors of FG explaining 4.7% and 5.2% of the variance, respectively (total R²=0.14, p=0.02). Hepatic fat fraction was the only independent predictor of 2-HrG explaining 9.9% of the variance in the model (total R²=0.12, p=0.02). No measure of adiposity was retained in the model to predict HbA1c.

Table 2.

Stepwise linear regression analysis examining independent associations between adiposity and measures of glycemia

	Fasting Glucose		2-Hour Glucose		HbA1c
	B±SE	p-value	B±SE	p-value	B±SE	p-value
Hepatic fat	0.21±0.2	0.04	1.26±0.43	0.005	----	----
Pancreatic fat	0.37±0.1	0.04	----	----	----	----
R2	0.14		0.12		0.006
F	3.10	0.02	1.26	0.03	0.24	0.7

Models controlled for age and sex. Significance was deemed at P≤0.05. B=Beta, SE=Standard Error

Discussion

This study aimed to examine the relationship between various measures of adiposity and whether they were associated with glycemia in a cohort of Latino youth with obesity. Our data support an association among measures of adiposity but suggest that only hepatic and pancreatic fat fraction are independent predictors of hyperglycemia.

Hepatic fat was found to be a significant independent predictor of both FG and 2-HrG in our study. The association between hepatic steatosis and T2D is well-described in adults¹⁰ with emerging evidence in children suggesting that nearly 30% of youth with non-alcoholic fatty liver disease (NAFLD) exhibit concomitant hyperglycemia.¹¹ In adults, NAFLD can precede the development of T2D where the proposed mechanisms linking hepatic fat with hyperglycemia include insulin resistance mediated lipogenesis leading to increased free fatty acid (FFA) influx into the liver, as well as inflammation.¹² Subsequent β-oxidation of FFA leads to inhibition of insulin-signaling within hepatocytes contributing to disruptions in glycogen synthesis, thereby increasing hepatic glucose production and hyperglycemia.¹³ However, insulin resistance alone does not completely explain the link between liver fat and hyperglycemia as NAFLD precedes and predicts the development of T2D independent of baseline insulin resistance in adults.¹⁴ Therefore, there are likely additional contributory mechanisms that link hepatic fat to T2D risk, and export of VLDL-triglycerides from the liver to the pancreas has been previously proposed.¹⁵

In our study, pancreatic fat fraction significantly predicted FG but not 2-HrG or HbA1c. Although the pathogenic association of pancreatic fat and glucose metabolism remains unclear, pancreatic fatty acid accumulation has been linked to cell injury and loss of β-cell mass and function.¹⁶ Heni et al. found that pancreatic fat was negatively correlated with insulin secretion in adults with prediabetes but not in those with normal glycemic profiles.¹⁷ Similarly, Tushuizen et al. reported significant associations between pancreatic fat and β-cell dysfunction in non-diabetic adults; however, the association did not hold for those with T2D.¹⁸ The lack of association in those with T2D may indicate that once the β-cell has failed and overt disease is established, other factors may further exacerbate β-cell dysfunction.¹⁹ Much less is known about pancreatic fat and T2D risk in the pediatric population; however, a link between insulin secretion and pancreatic fat was observed in African American adolescents with prediabetes where a modest increase in pancreatic fat was hypothesized to play a role in β-cell dysfunction early in the pathogenesis of hyperglycemia.⁸

In our study, no adiposity measure predicted HbA1c but this null finding may be due to the small sample size in conjunction with limited variability in HbA1c in the cohort. Analysis from the Hispanic Community Children’s Health Study/Study of Latino Youth demonstrated a lack of association between various measures of adiposity including BMI, waist circumference, % body fat and HbA1c in over 1200 Latino adolescents.²⁰ Collectively, these and the current findings suggest that other factors may be drivers of elevated HbA1c in Latino youth.

Strengths of the study include the focus on a vulnerable population of Latino youth that experiences significant obesity-related disparities, the application of sophisticated imaging measures to quantify total, visceral, liver and pancreatic fat, and the inclusion of robust measures of hyperglycemia. Despite these strengths, limitations include the cross-sectional design that prohibits cause and effect conclusions, the relatively small sample size, the inability to control for pubertal development, the limited variability in HbA1c, the lack of data from other ethnic groups which limits generalizability, and the inability to include genetic or behavioral factors which may shed additional light T2D risk in this population. In addition to these limitations, the assessment and quantification of pancreatic fat by MRI can be technically challenging due to its size, shape, and location and, therefore, further research is needed to establish the clinical importance of this fat depot.

In conclusion, our data suggest that hepatic and pancreatic fat may be important predictors of T2D risk among Latino youth with obesity. However, the total amount of variance explained across the various measures of glycemia was relatively small. Given the complex nature of pediatric T2D, future studies that integrate physiologic markers with social determinants across diverse, high-risk population sub-groups are needed to better inform prevention and treatment efforts in youth.

DATA AVAILABILITY:

The data that support the findings of this study may be made available by contacting the corresponding author.