Ectopic pancreatic fat as a risk factor for hypertension in children and adolescents with nonalcoholic fatty liver disease

Abstract Although nonalcoholic fatty liver disease (NAFLD) is known to be a risk factor for cardiovascular diseases, few studies have reported an association between ectopic fat deposition and metabolic complications, including hypertension, in children with NAFLD. The present study evaluated the risk factors for hypertension in children with NAFLD from the aspect of ectopic fat. This cross‐sectional retrospective study investigated 65 children with NAFLD (49 boys, mean age 13.0 ± 3.2 years, mean body mass index z‐score [BMI‐z] 2.5 ± 1.2), who underwent liver biopsy and magnetic resonance imaging‐based fat fraction measurement for ectopic hepatic and pancreatic fats, as well as anthropometry, blood pressure, laboratory tests, and body composition analysis. A logistic regression model was used to identify the risk factors for hypertension. Through a simple logistic regression analysis, age (OR 1.392), BMI‐z (OR 3.971), waist circumference‐to‐height ratio (OR 1.136), fat‐free mass index (OR 1.444), γ‐glutamyl transferase (OR 1.021), quantitative insulin sensitivity check index (OR 0.743), dyslipidemia (OR 5.357), and pancreatic fat fraction (PFF) (OR 1.205) were associated with hypertension. The optimal cut‐off of PFF to divide children with NAFLD into two groups with and without hypertension was 4.39% (area under the curve 0.754, p = .001, sensitivity 82.4%, specificity 73.9%). Multiple logistic regression analysis in the fully adjusted model revealed both BMI‐z (OR 4.912, 95% CI, 1.463–16.497) and PFF (OR 1.279, 95% CI, 1.007–1.624) were independent risk factors for hypertension. In conclusions, in addition to BMI‐z, ectopic pancreatic fat is an important risk factor for hypertension in children with NAFLD.

Mounting evidence highlights that NAFLD is an independent risk factor for CVD in adults. However, to date, this association has not been fully substantiated in children with NAFLD. 6 Recently several studies on children with NAFLD suggest that in children, as well as in adults, NAFLD might be associated with potential cardiovascular complications. 6 This is independent of the coexistence of well-known risk factors and features of MetS. 6 The association between NAFLD and various aspects of CVD, including cardiac dysfunctions, atherosclerosis, and hypertension (HTN) in children has been investigated. [6][7][8][9] In particular, children with NAFLD are at an increased risk for HTN compared with obese children without NAFLD, a risk that persists over time. 8,10,11 When circulating triglycerides (TG) and free fatty acid levels exceed the metabolic capacity through adipocyte hypertrophy and hyperplasia, ectopic fat accumulation in non-adipose tissues such as the heart, liver, pancreas, and skeletal muscles in patients with obesity becomes an increasing risk. 12,13 This accumulation of dysfunctional visceral and ectopic fat causes low-grade inflammation, oxidative stress, endothelial dysfunction, atherogenic dyslipidemia, and impaired glucose metabolism, ultimately leading to the development of CVD in patients with NAFLD. 14,15 Along with ectopic hepatic fat in NAFLD, pancreatic fat is also considered as an obesity-induced ectopic fat depot, which may contribute to cardio-metabolic disturbances such as NAFLD, HTN, DM, dyslipidemia, and MetS. 16,17 The pancreas is divided into pancreatic islets that secrete endocrine hormones such as insulin and glucagon and exocrine regions that secrete digestive enzymes. 18 It is presumed that pancreatic fat accumulation might cause pancreatic islet inflammation and β-cell dysfunction, resulting in deterioration of the insulinsecreting capacity. 18 However, the association between pancreatic steatosis and cardiovascular disturbances along with impaired glucose metabolism represented by β-cell dysfunction and IR remains unclear. 12,13,16 Furthermore, little is known about the association between magnetic resonance imaging (MRI)-based ectopic fat accumulation in the liver and pancreas and metabolic components of obesity-related complications including HTN, particularly in children with biopsy-proven NAFLD.
Therefore, the present study aimed to investigate the risk factors associated with HTN in pediatric patients with NAFLD from the aspect of ectopic fat.

Anthropometric measurements and body composition measurements
Anthropometric parameters, including height, weight, and waist circumference (WC), were measured in all children using standardized methods; we calculated the waist circumference-to-height ratio (WHtR), body mass index (BMI), and BMI standard deviation score (zscore). In the Korean national growth charts, WC z-score (WC-z) has not yet been developed; therefore, we used WHtR instead of WCz. Obesity was defined as BMI ≥95 th percentile and overweight as BMI between the 85 th and 95 th percentiles, adjusted for age and sex, according to the Korean National Growth Charts 2017. 25 Central obesity was defined as WC ≥90 th percentile, adjusted for age and sex, according to the Korean National Growth Charts 2007. 26 Body composition was measured using bioelectrical impedance analysis (InBody J10, Biospace Co., Ltd., Seoul, South Korea), and the total body fat mass and fat-free mass were recorded. The fat mass index (FMI) and fat-free mass index (FFMI) were calculated as the fat mass (kg) and fat-free mass (kg), respectively, divided by the square of height (m 2 ).

BP measurements
BP was measured at every hospital visit using a standard measurement method. 20   The homeostasis model assessment of insulin resistance (HOMA-IR) was calculated using the following formula: FPG (mg/dl) × fasting insulin (μU/ml)/405, 27

Liver biopsy
Liver biopsy was performed by an experienced pediatric radiologist with ultrasound guidance, and the findings were interpreted by an expert liver pathologist who was blinded to the patients' clinical data.
All biopsy specimens were evaluated based on the NAFLD Clinical Research Network criteria, and the NAFLD activity score (NAS) was assessed. 29 The degree of steatosis, lobular inflammation, portal inflammation, and hepatocyte ballooning (steatosis 0-3, lobular inflammation 0-3, portal inflammation 0-2, hepatocyte ballooning 0-2) were graded accordingly. Moreover, hepatic fibrosis was staged as 0-4. The NAS was calculated using an 8-point scale as the sum of the scores for steatosis, lobular inflammation, and hepatocyte ballooning.  30 HFF ≥5.0% on MRI-PDFF is defined as a fatty liver because MRI-PDFF is highly accurate compared to the histological steatosis. 30 ROIs of pancreatic fat fraction (PFF) measurements were generated in the head, body, and tail of the pancreas. The normal range of PFF to define fatty pancreas has not yet been established.

Statistical analysis
Descriptive characteristics are presented as mean ± SD for normally distributed variables or medians and ranges for non-normally distributed variables. Categorical measurements are expressed as absolute numbers and percentages. Intergroup differences were evaluated using an independent t-test for parametric variables and the Mann-

Association between metabolic components and MRI-measured ectopic fat fraction of the liver and pancreas
As for ectopic fat on MRI, the median HFF was 24.3% (range, 4.2%-49.9%) and the median PFF was 3.8% (range, 0.4%-26.9%). HFF was not significantly different according to the status of obesity, HTN, central obesity, dyslipidemia, prediabetes/DM, and MetS status. Moreover, PFF did not differ according to obesity, central obesity, dyslipidemia, prediabetes/DM, and MetS status. However, PFF was significantly different according to HTN status. The median PFF of the HTN group was 6.7% (range 1.0%-26.9%) and that of the non-HTN group was 3.2% (range 0.4%-15.8%) (p = .002).

Multiple logistic regression analysis for hypertension in children with NAFLD
Multiple logistic regression analysis was performed after selecting noncollinear covariates through the stepwise selection method to predict the risk factors associated with HTN. The remaining significant variables were BMI-z and PFF. For the selection of variables to be adjusted in the multiple logistic regression, we examined the multicollinearity among the variables. Because children and adolescents are growing, the z-score form is more appropriate than the anthropometric values. As BMI-z implies a concept for adjusting for sex and age, it is reasonable that sex and age were not included in the multiple logis-

Optimal cut-off value of PFF and BMI-z for hypertension in children with NAFLD
The ROC curve was constructed to determine the optimal cut-off value for PFF to divide the patients into two groups: those with and those

F I G U R E 1
Optimal cut-off value for PFF and BMI-z for hypertension in children with NAFLD. Receiver operating characteristic (ROC) curves were constructed to examine the optimal cut-off value for pancreatic fat fraction (PFF) and body mass index standard deviation score (BMI-z) to divide children with non-alcoholic fatty liver disease into two groups, with and without hypertension.

DISCUSSION
To our knowledge, this is the first study to investigate the risk factors for developing HTN using MRI-measured HFF, PFF, and obesity-related metabolic components in children with liver biopsy-proven NAFLD.
In our study, multiple logistic regression analysis showed that BMIz and PFF were independent risk factors for HTN in children with biopsy-proven NAFLD after adjusting for FFMI, GGT, dyslipidemia, and QUICKI, which were associated with HTN through simple logistic regression analysis. Especially, MRI-measured PFF was significantly associated with HTN in the fully adjusted model, including BMI-z. Furthermore, PFF was not inferior to BMI-z in the suggestion of HTN when confirmed through a pairwise comparison of the ROC curves of BMI-z and PFF.
NAFLD is known to be a strong cardiovascular risk factor, independent of traditional cardiovascular risk factors including obesity. 14 A systematic review and meta-analysis incorporating almost 165 000 participants in 34 studies showed the association of NAFLD with both prevalent and incident CVD, including coronary artery disease, atherosclerosis, and HTN. 31 Furthermore, increasing scientific studies on children with NAFLD suggest that NAFLD in children as well as in adults might be an independent risk factor for CVD. 6,7,9,11 The recent clinical practice guidelines for the management of NAFLD in children recommend screening of the cardiovascular system for all patients with NAFLD. 6,10 At least a detailed risk factor evaluation and regular monitoring is recommended. 6,10 Possible mechanisms leading to CVD in patients with NAFLD might originate from the expanded and inflamed visceral fat. 5 NASH may play a part in the pathogenesis of CVD through the systemic release of several inflammatory, hemostatic, and oxidative-stress mediators, or through the contribution of NAFLD to IR and atherogenic dyslipidemia. 5,14 IR is a shared pathologic condition supporting several dysmetabolic status of obesity including prediabetes/type 2 DM, dyslipidemia, atherosclerosis, and NAFLD. 32 In children and adolescents with obesity, a strong association between IR and a higher prevalence of MetS components has been observed; thus, a higher cardiovascular risk is predicted in these patients. 32 IR has been demonstrated to be a reliable marker in the prediction of cardiovascular risk. 32 It has been suggested that IR may be involved in the pathogenesis of atherosclerosis, according to the evidence that the more IR is increased in youths, the more circulating biomarkers of endothelial dysfunction are elevated, while adiponectin, which plays an antiatherogenic role, is reduced. 32 The relevant association between IR and cardiovascular risk in the pediatric population is well known, especially in obese children, 32 even though further research is needed to clarify this association.
In our study, QUICKI, which is indicative of IR, was significantly associated with HTN in a simple logistic regression analysis, but not TA B L E 2 Simple logistic regression analysis for predicting risk factors associated with hypertension in 65 children with non-alcoholic fatty liver disease Note: Odds ratios of WHtR and QUICK were odds ratios for hypertension by every 0.01 increment of WHtR and QUICKI. Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI z-score, body mass index standard deviation score; CI, confidence interval; FF, fat fraction.; FFMI, fat-free mass index; FMI, fat free mass index; GGT, γ-glutamyl transferase; HbA1c, glycosylated hemoglobin; HDL-C, high density lipoprotein cholesterol; HOMA-IR, homeostatic model assessment for insulin resistance; LDL-C, low density lipoprotein cholesterol; MRI, magnetic resonance imaging; OR, odds ratio; QUICKI, quantitative insulin-sensitivity check index; WHtR, waist circumference to height ratio.
with HOMA-IR. We were not able to clearly explain why QUICKI, but not HOMA-IR, was significantly associated with HTN. The gold standard for the assessment of IR was the hyperinsulinemic-euglycemic clamp study; however, its costs and difficult management in clinical and research practice have determined the need of surrogate markers. 32 It is known that HOMA-IR and QUICKI present a favorable correlation with the hyperinsulinemic-euglycemic clamp, 32 and display identical diagnostic accuracy. 33 Even though a previous study insisted that HOMA-IR was more reliable than QUICKI as a measure of IR among children and adolescents, 34 there were some studies showing that QUICKI displayed better reproducibility than HOMA-IR. 35,36 Moreover, in hypertensive obese patients, QUICKI was well-correlated with BP and with brain natriuretic peptide deficiency. 37 A relative natriuretic peptide deficiency, probably related to the IR found in obese individuals, may represent one of the reasons for inducing HTN by IR. 32 As it is presumed that IR may be associated with ectopic fat accumulation in the pancreas, we hypothesized that pancreatic steatosis may also affect the development of CVD through IR and undiscovered mechanisms. According to a meta-analysis on pancreatic ectopic fat, non-alcoholic fatty pancreas disease was associated with a significantly increased risk of arterial HTN (risk ratio 1.67, 95% CI, 1.32-2.10, p < .0001). 17 Several studies have shown that MRI-evaluated PFF was highly correlated with HTN. 38,39 One previous study also showed that fatty pancreas is a contributing factor for the development of atherosclerosis in adults with biopsy-proven NAFLD. 40 The present study additionally revealed that FFMI was significantly associated with HTN. A British study demonstrated that body fat was significantly higher and skeletal muscle mass was significantly lower in any subcategory of HTN in both men and women; thus, the estimation of both body fat and skeletal muscle mass may be important in clinical approach. 41 In another adult study, patients with sarcopenic obesity had a greater risk of HTN than those with simple obesity or sarcopenia alone. 42 According to a pediatric study on healthy Chinese children, a high fat-free mass percentage was associated with a low BP and low HTN risk in the fully adjusted model. 43 In our study, the GGT level was associated with the development of HTN. Longitudinal studies that assessed the association between NAFLD and the incidence of HTN in adults showed that GGT levels were significantly related to the incidence of HTN. 44 Moreover, a study on children with NAFLD confirmed that GGT might be a potentially reliable, simple, and non-invasive biochemical marker for the estimation of cardiovascular risk in obese children with NAFLD. 6,45 Dyslipidemia is known to increase the risk of developing HTN after adjusting for age, BMI, DM, alcohol, smoking, exercise, and parental history of HTN. 46 The mechanisms through which obesity causes HTN are complex, including sympathetic nervous system over-activation, stimulation of the renin-angiotensin-aldosterone system, alterations in adipose-derived cytokines, structural functional renal changes, and IR. 47 Therefore, more studies are needed to prove the association between PFF and HTN in the future. is recognized as a defense mechanism to prevent ectopic fat accumulation and, therefore, prevents metabolic degeneration and dysfunction of β-cells. 12,50 Third, as this was a retrospective study with a relatively small sample size, we caution against over-interpreting the significance of the results. In our center, liver biopsy and abdominal MRI are performed in patients at increased risk for severe liver disease.
These tests are performed when persistently elevated liver enzymes and associated dyslipidemia, type 2 DM, and HTN are observed. This suggests that the results of the present study may not be generalizable to patients with mild NAFLD and/or no metabolic risk factors.
However, it is critical to identify which pediatric patients with NAFLD are at greatest risk for cardiometabolic dysregulation, including HTN.
Providing active intervention early to prevent disease progression is essential. Lastly, although we found an association between PFF and HTN, we did not analyze the data of carotid intima-media thickness or carotid-femoral pulse wave velocity to reflect atherosclerosis and pro-inflammatory adipokines as a link between PFF and HTN, because these values were not measured in all study patients.
In conclusions, in addition to BMI-z, ectopic pancreatic fat deposi- together with HFF and PFF to confirm the association between HFF, PFF, and metabolic components. More research is needed in the future to study the association between PFF and atherosclerosis parameters in children with NAFLD to confirm PFF as a risk factor for HTN.