Highlights
- ●The association of thigh subcutaneous fat with NAFLD was first investigated based on a prospective cohort study.
- ●A higher TSFA/AFA ratio was associated with a lower risk of incident NAFLD and a higher likelihood of remitted NAFLD.
- ●ThC/WC ratio was negatively associated with incident NAFLD and positively associated with remitted NAFLD.
- ●Adiponectin, triglyceride, and HOMA-IR mediated the effects of TSFA/AFA ratio on incident and remitted NAFLD.
Abstract
Background & Aims
No prospective studies have examined the association between thigh subcutaneous fat distribution and non-alcoholic fatty liver disease (NAFLD). We investigated the associations of thigh subcutaneous fat distribution with incidence and remission of NAFLD based on a community-based prospective cohort.
Methods
We followed 1787 subjects who underwent abdominal ultrasonography, abdominal and femoral magnetic resonance imaging scans, and anthropometric assessments. Associations of thigh subcutaneous fat area (TSFA)/abdominal fat area (AFA) ratio and thigh circumference (ThC)/waist circumference (WC) ratio with incidence and remission of NAFLD were estimated using the modified Poisson regression model.
Results
Over a mean 3.6-year follow-up, 239 incident cases of NAFLD and 207 regressed cases of NAFLD were identified. Increasing TSFA/AFA ratio was associated with a lower risk of incident NAFLD and a higher likelihood of remission of NAFLD (risk ratio [RR] per standard deviation [SD]: 0.69, 95% confidence interval [CI] 0.59–0.81; 1.20, 95% CI 1.07–1.34, respectively). Each one SD increase in ThC/WC ratio was associated with a 16% lower risk of incident NAFLD (RR 0.84, 95% CI 0.76–0.94) and a 22% higher likelihood of remission of NAFLD (RR 1.22, 95% CI 1.11–1.34). Additionally, the effects of TSFA/AFA ratio on incidence and remission of NAFLD were mediated through adiponectin (14.9% and 26.6%), homeostasis model assessment of insulin resistance (9.5% and 23.9%) and triglyceride (7.5% and 19.1%).
Conclusions
These results demonstrated that a favorable fat distribution, characterized by a relatively greater ratio of thigh subcutaneous fat to lower abdominal fat, had a protective role against NAFLD.
Lay summary
The associations of thigh subcutaneous fat distribution with NAFLD incidence and remission have not been prospectively examined in a community-based cohort. Our findings added the viewpoint that greater thigh subcutaneous fat relative to a given amount of abdominal fat had a protective effect against NAFLD among middle-aged and elderly Chinese to the existing literature.
Graphical abstract
Graphical Abstract
Keywords
Conflict of interest
The authors declare no conflict of interest. Please refer to the accompanying ICMJE disclosure forms for further details.
Financial support
This work was supported by the Natural Science Foundation of Shanghai (grant number 18ZR1429000) and the Health and Family Planning Commission of Pudong New Area Joint research project (grant number PW2018D-14).
Author contributions
LY, CP, and HX made substantial contributions to the conception and design of the study. LY, CP, and HX analyzed the data. LY, CP, and HX drafted the manuscript. All the authors assisted in the acquisition and interpretation of data and contributed to the critical revision of the manuscript for important intellectual content and approved the final version.
1. Introduction
Obesity, an important risk factor for non-alcoholic liver disease (NAFLD)
1
, is a highly heterogeneous disease characterized by differences in the regional distribution of body fat. Emerging evidence indicates that visceral fat area (VFA) and abdominal subcutaneous fat area (ASFA) have different effects on the development and remission of NAFLD 2
,3
. That is to say, ASFA is not fully like VFA, which is an established risk factor of NAFLD 2
,3
. Based on only two published Korean cross-sectional studies, negative associations between thigh subcutaneous fat area (TSFA)4
or leg fat to total fat ratio5
and NAFLD are observed. The discrepancy in NAFLD risk between abdominal and thigh fat depots might originate from differences in their lipolytic activity. Femoral fat tissue might be more likely to take up non-esterified fatty acids (NEFA) from the circulating blood and thus protected the liver from high NEFA exposure, which further protected against insulin resistance 6
,7
. To our knowledge, no prospective studies regarding the associations of TSFA or its relative distribution with incidence and remission of NAFLD have been reported.Unlike the two single adiposity indicators of waist circumference (WC) and thigh circumference (ThC), the combination indicators of WC and ThC, such as waist-to-thigh ratio and its inverse, could reflect body shape and fat distribution to some extent. The waist-to-thigh ratio has been identified as a significant predictor of diabetes
8
and all-cause mortality 9
. However, little was known about the impacts of the ratio of ThC to WC (ThC/WC ratio) on incidence and remission of NAFLD.To fill these knowledge gaps, we evaluated the associations of incidence and remission of NAFLD with the ratio of TSFA to abdominal fat area (AFA)–TSFA/AFA ratio, a precise measurement, and with the ThC/WC ratio, a simple surrogate for the former, among middle-aged and elderly Chinese based on a prospective cohort dataset. Here, we used the TSFA/AFA ratio and ThC/WC ratio to more intuitively reflect the effects of higher proportion of favorable fat depots on incidence and remission of NAFLD.
2. Materials and Methods
2.1 Subjects and study design
We analyzed data from the previously described cohort–Shanghai Nicheng Cohort Study.
10
,11
. Briefly, 17212 subjects aged 45–70 years finished the baseline survey from 2013 to 2014. Of which, 2849 subjects aged 55–70 years and had complete baseline data on abdominal ultrasonography, TSFA, ASFA, VFA, ThC, and WC were invited to participate in the follow-up survey in 2018, and 2008 subjects attended (follow-up rate: 70.5%).We excluded 221 subjects according to the following criteria: (1) missing data on abdominal ultrasonography (n = 39) at follow-up or alcohol intake (n = 3) at baseline and follow-up; (2) excessive drinking (a daily alcohol consumption > 30 g/day in men and > 20 g/day in women, n = 138) at baseline and follow-up; and (3) positive hepatitis B surface antigen (n = 41) at baseline. Finally, 1787 subjects with a mean follow-up time of 3.6 (standard deviations [SD] 0.32) years were included into this study (Fig. S1).
This study conformed to the principles of the Declaration of Helsinki and was approved by the ethics committee of the Shanghai Sixth People's Hospital (Approval No: 2015-27). Written informed consent was obtained from each subject.
2.2 Clinical data collection and laboratory measurements
Data involving demographics, education levels, lifestyles (smoking status, drinking status, and leisure-time exercise), medication usage, and medical history (e.g., diabetes and hypertension) were obtained via a standardized questionnaire. Height, weight, and blood pressure were measured using an established standard protocol
12
. ThC was measured at the mid-thigh between the inguinal crease and the proximal border of the patella. WC was measured along the midline between the lower margin of the costal arch and the upper margin of the iliac crest on the mid-axillary line. ThC/WC ratio was calculated as ThC (cm) divided by WC (cm). Body mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters.Fasting plasma glucose (FPG), glycated hemoglobin (HbA1c), fasting insulin (FINS), triglyceride (TG), total cholesterol, high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol, alanine aminotransferase, aspartate aminotransferase, gamma-glutamyl transpeptidase, adiponectin, fibroblast growth factor 21 (FGF21), and retinol-binding protein-4 (RBP4) were measured using overnight fasting (at least 10 hours) venous blood samples. The laboratory measurement methods were described in Table S1. The homeostasis model assessment of insulin resistance (HOMA-IR) was used to quantify insulin resistance, calculated as FPG (mmol/L) × FINS (μU/mL) /22.5
13
.2.3 Ultrasonographic examinations
Abdominal ultrasonography (Z.One Ultra, Zonare Medical Systems, Inc, Mountain View, CA, USA) was performed by experienced ultrasonographists who were blinded to the study design and clinical data. Fatty liver was defined as present when at least two of the following three abdominal ultrasonographic features were found: diffusely increased echogenicity ("bright") liver with liver echogenicity greater than kidney or spleen, vascular blurring, and/or deep attenuation of ultrasound signal
14
.2.4 Measurement of adipose tissue areas
Abdominal and femoral magnetic resonance imaging (MRI) scans were conducted on subjects in a supine position, using a 3.0 T General Electric scanner (GE Healthcare, Milwaukee, WI, USA). Eight slices of T1 axial images centered at the navel and the mid-thigh respectively were obtained. Each slice thickness was 10.0 mm. Cross-sectional TSFA, ASFA, and VFA were measured in cm2 using the mid-thigh and umbilical slices, based on an area of 2-D pixels meeting the adipose shading threshold from the DICOM images. Two trained investigators segmented the images into different fat districts and calculated fat areas using the sliceOmatic image analysis software (version 5; Tomovision Inc., Montreal, QC, Canada). When the results differed more than 10%, a third investigator reanalyzed the images. AFA was calculated as the sum of ASFA and VFA, and TSFA/AFA ratio was created by dividing TSFA by AFA.
2.5 Outcome definitions
NAFLD was defined based on ultrasound evidence of fatty liver, in the absence of excessive drinking (alcohol consumption > 30 g/day in men and > 20 g/day in women) and other known causes of chronic liver diseases (viral hepatitis, hepatolenticular degeneration, etc.)
15
. NAFLD absent at baseline but present at follow-up was defined as incidence of NAFLD, and the reversed case was defined as remission of NAFLD. To assess the severity of NAFLD, two noninvasive indices of the Fibrosis-4 (FIB-4)16
,17
and the Fibrosis Nonalcoholic steatohepatitis Index (FNI)18
,19
were used. Advanced fibrosis was defined as FIB-4 ≥ 1.30 or FNI > 0.10, respectively.2.6 Statistical analysis
Descriptive data were expressed as means (SD), medians (25th–75th percentiles), or numbers (proportions) as appropriate. The differences between two groups were compared using the t-tests or Mann‐Whitney U test for continuous variables and the chi-squared test for categorical variables. Correlations of TSFA/AFA ratio and ThC/WC ratio with the other baseline characteristics were assessed using the Pearson and partial correlation coefficients adjusted for sex, age, and BMI.
The modified Poisson regression model with robust error variance was used to estimate the risk ratios (RRs) and 95% confidence intervals (CIs) for incidence and remission of NAFLD. TSFA/AFA ratio and ThC/WC ratio were entered into the models as per 1-SD or sex-specific per tertile increment. The regression model was used to test the linear trend of incidence or remission of NAFLD across the tertile groups of TSFA/AFA ratio or ThC/WC ratio, using the median values of each tertile to reflect the group levels. Potential interactions between TSFA/AFA ratio or ThC/WC ratio and the other analysis variables on incidence and remission of NAFLD were tested using the Wald test by adding their product terms to the regression models. In Model 1, adjustment variables included sex, age, education levels (primary school and below or middle school and above), smoking status (never, past, or current smokers), drinking status (never, past, or current drinkers), and leisure-time exercise (never, 1 – < 30 minutes/day, ≥ 30 minutes/day); in Model 2, BMI was additionally adjusted; in Model 3, hypertension (yes or no), diabetes (yes or no), TG, and HDL-C were further adjusted. Also, the associations of TSFA/AFA ratio and ThC/WC ratio with incidence and remission of NAFLD were depicted using the restricted cubic splines with 4 knots at the 5th, 35th, 65th, and 95th percentiles. The associations between TSFA/AFA ratio and incident NAFLD at different stages (non-NAFLD, NAFLD without or with advanced fibrosis) were analyzed by the multinomial logistic regression. The 'mediation' R package
20
was used to estimate the average causal mediation effect (ACME) and average direct effect (ADE), reflecting indirect and direct effects of TSFA/AFA ratio on incidence and remission of NAFLD. The mediated portion was calculated as the ratio of the ACME to the total effect (ACME plus ADE). The mediation effects were estimated using non-parametric bootstrapping (1000 simulations).Statistical analyses were carried out using SPSS, version 26.0 (SPSS Inc., Chicago, Illinois, USA), StataMP version 14.0 (StataCorp LP, Texas, USA), or R version 4.0.2 (R Foundation for Statistical Computing). A two-tailed p value < 0.05 was considered statistically significant.
3. Results
3.1 Characteristics of the study population
As shown in Table 1, among 1787 participants at baseline, 981 (54.9%) without and 806 (45.1%) with NAFLD were identified. Over a 3.6-year follow-up, 239 subjects (24.4% of 981) without NAFLD progressed to NAFLD and 207 patients (25.7% of 806) with NAFLD regressed to non-NAFLD. Compared with subjects without incident NAFLD, those with incident NAFLD had more unfavorable metabolic profiles, higher adiposity indicators and FGF21 levels, but lower TSFA/AFA ratio (mean [SD]: 0.45 [0.18] vs 0.52 [0.24], p < 0.001), ThC/WC ratio (mean [SD]: 0.59 [0.05] vs 0.62 [0.06], p < 0.001), and adiponectin. Meanwhile, subjects with regressed NAFLD had more favorable metabolic profiles, lower adiposity indicators and FGF21 levels, but higher ThC/WC ratio (mean [SD]: 0.59 [0.05] vs 0.57 [0.05], p < 0.001), adiponectin, and RBP4, and similar TSFA/AFA ratio (mean [SD]: 0.43 [0.18] vs 0.41 [0.17], p = 0.328) than those without regressed NAFLD.
Table 1Baseline characteristics of subjects by incidence and remission of NAFLD after a 3.6-year follow-up.
| Characteristic | Total (n = 1787) | Non-NAFLD at baseline (n = 981) | NAFLD at baseline (n = 806) | ||||
|---|---|---|---|---|---|---|---|
| No incidence of NAFLD (n = 742, 75.6%) | Incidence of NAFLD (n = 239, 24.4%) | p value | Remission of NAFLD (n = 207, 25.7%) | No remission of NAFLD (n = 599, 74.3%) | p value | ||
| Demographic | |||||||
| Women, no. (%) | 1131 (63.3) | 420 (56.6) | 155 (64.9) | 0.024 | 123 (59.4) | 433 (72.3) | 0.001 |
| Age, years | 62.2 (3.8) | 62.2 (3.8) | 62.7 (3.9) | 0.064 | 61.9 (3.9) | 62.2 (3.9) | 0.498 |
| Clinical | |||||||
| SBP, mmHg | 129.9 (11.7) | 127.8 (11.9) | 129.9 (12.0) | 0.018 | 131.8 (12.5) | 131.7 (10.6) | 0.915 |
| DBP, mmHg | 82.8 (6.0) | 81.6 (6.0) | 83.0 (5.6) | 0.001 | 83.5 (5.9) | 83.9 (5.8) | 0.360 |
| FPG, mmol/L | 6.0 (5.6-6.6) | 5.9 (5.5-6.4) | 6.0 (5.7-6.5) | 0.017 | 6.0 (5.7-6.6) | 6.3 (5.8-7.2) | < 0.001 |
| HbA1c, % | 5.9 (1.0) | 5.6 (0.9) | 5.8 (0.7) | 0.037 | 5.8 (0.8) | 6.1 (1.1) | < 0.001 |
| FINS, μU/mL | 6.4 (4.6-9.1) | 5.0 (3.5-6.5) | 6.7 (5.0-8.7) | < 0.001 | 6.5 (4.8-9.2) | 8.9 (6.5-12.7) | < 0.001 |
| HOMA-IR | 1.8 (1.2-2.6) | 1.3 (0.9-1.8) | 1.8 (1.3-2.4) | < 0.001 | 1.9 (1.3-2.6) | 2.6 (1.8-3.8) | < 0.001 |
| TG, mmol/L | 1.3 (0.9-1.9) | 1.1 (0.8-1.5) | 1.3 (0.9-1.8) | < 0.001 | 1.4 (1.0-2.0) | 1.8 (1.3-2.6) | < 0.001 |
| TC, mmol/L | 5.2 (1.0) | 5.1 (1.0) | 5.1 (0.8) | 0.711 | 5.3 (1.0) | 5.3 (1.0) | 0.277 |
| HDL-C, mmol/L | 1.4 (0.3) | 1.5 (0.4) | 1.4 (0.3) | < 0.001 | 1.3 (0.3) | 1.3 (0.3) | 0.101 |
| LDL-C, mmol/L | 3.1 (0.8) | 3.0 (0.8) | 3.1 (0.7) | 0.101 | 3.2 (0.8) | 3.3 (0.8) | 0.469 |
| ALT, U/L | 18.9 (10.1) | 16.7 (8.6) | 17.2 (6.6) | 0.378 | 18.0 (8.1) | 22.6 (12.3) | < 0.001 |
| AST, U/L | 23.1 (6.8) | 22.9 (6.8) | 21.8 (4.6) | 0.003 | 22.0 (5.5) | 24.3 (7.8) | < 0.001 |
| GGT, U/L | 21.0 (16.0-33.0) | 18.0 (14.0-27.0) | 20.0 (15.0-27.0) | 0.021 | 23.0 (18.0-33.0) | 27.0 (19.0-39.0) | 0.001 |
| Anthropometric | |||||||
| TSFA/AFA ratio | 0.47 (0.21) | 0.52 (0.24) | 0.45 (0.18) | < 0.001 | 0.43 (0.18) | 0.41 (0.17) | 0.328 |
| ThC/WC ratio | 0.59 (0.06) | 0.62 (0.06) | 0.59 (0.05) | < 0.001 | 0.59 (0.05) | 0.57 (0.05) | < 0.001 |
| TSFA, cm2 | 117.4 (53.7) | 104.0 (48.4) | 119.4 (50.6) | < 0.001 | 116.9 (50.9) | 133.3 (57.6) | < 0.001 |
| AFA, cm2 | 264.6 (87.8) | 207.9 (66.5) | 271.4 (68.6) | < 0.001 | 279.3 (71.3) | 327.0 (77.1) | < 0.001 |
| ASFA, cm2 | 143.9 (55.4) | 116.3 (43.4) | 154.4 (50.9) | < 0.001 | 149.6 (49.1) | 172.1 (56.3) | < 0.001 |
| VFA, cm2 | 120.6 (48.5) | 91.6 (34.6) | 116.9 (37.0) | < 0.001 | 129.7 (40.7) | 155.0 (46.5) | < 0.001 |
| ThC, cm | 48.9 (3.8) | 47.5 (3.4) | 49.2 (3.3) | < 0.001 | 49.5 (3.8) | 50.4 (3.9) | 0.005 |
| WC, cm | 82.8 (8.8) | 77.5 (7.4) | 83.3 (6.9) | < 0.001 | 84.1 (7.0) | 88.8 (7.3) | < 0.001 |
| BMI, kg/m2 | 24.9 (3.2) | 22.9 (2.4) | 25.1 (2.2) | < 0.001 | 25.3 (2.3) | 27.2 (3.0) | < 0.001 |
| Adipokine | |||||||
| Adiponectin, μg/mL | 4.5 (1.8) | 5.1 (1.9) | 4.6 (1.7) | < 0.001 | 4.2 (1.7) | 3.7 (1.3) | 0.001 |
| FGF21, pg/mL | 222.6 (129.5-332.8) | 172.3 (103.1-290.3) | 204.4 (120.1-307.0) | 0.029 | 238.5 (159.0-343.6) | 264.1 (171.5-391.6) | 0.020 |
| RBP4, mg/L | 57.3 (16.3) | 55.2 (17.1) | 56.0 (14.8) | 0.448 | 61.7 (14.4) | 58.8 (16.0) | 0.022 |
Data were presented as means (standard deviations) or medians (interquartile ranges), or numbers (proportions) as appropriate. Differences between two groups were compared using the t-tests or Mann‐Whitney U test for continuous variables and the chi-squared test for categorical variables.
AFA, abdominal fat area; ALT, alanine aminotransferase; ASFA, abdominal subcutaneous fat area; AST, aspartate aminotransferase; BMI, body mass index; DBP, diastolic blood pressure; FGF21, fibroblast growth factor 21; FINS, fasting insulin; FPG, fasting plasma glucose; GGT, gamma-glutamyl transpeptidase; HbA1c, glycated hemoglobin; HDL-C, high-density lipoprotein cholesterol; HOMA-IR, homeostasis model assessment of insulin resistance; LDL-C, low-density lipoprotein cholesterol; NAFLD, non-alcoholic fatty liver disease; RBP4, retinol-binding protein-4; SBP, systolic blood pressure; TC, total cholesterol; TG, triglyceride; ThC, thigh circumference; TSFA, thigh subcutaneous fat area; VFA, visceral fat area; WC, waist circumference.
3.2 Correlations of TSFA/AFA ratio and ThC/WC ratio with other baseline factors
Correlations of TSFA/AFA ratio and ThC/WC ratio with other baseline characteristics were provided in Table S2. After adjustment for sex, age, and BMI, TSFA/AFA ratio and ThC/WC ratio were negatively correlated with worse metabolic profiles, ASFA, VFA, FGF21, and RBP4 (r = -0.05 – -0.43, all p < 0.05), and were positively correlated with HDL-C and adiponectin (r = 0.11 – 0.30, both p < 0.001). In addition, there was a moderate correlation between TSFA/AFA ratio and ThC/WC ratio (r = 0.42, p < 0.001).
3.3 Associations of TSFA/AFA ratio and ThC/WC ratio with incident NAFLD
As shown in Table 2, TSFA/AFA ratio was negatively associated with incident NAFLD with a multivariable-adjusted RR of 0.52 (95% CI 0.38–0.70) for the highest tertile versus the lowest tertile and of 0.69 (95% CI 0.59–0.81) for each one SD increase (Model 2). The negative association was slightly attenuated after further adjustment for hypertension, diabetes, TG, and HDL-C (RR per SD 0.72, 95% CI 0.62–0.84; Model 3). Similar to the associations between TSFA/AFA ratio and NAFLD, ThC/WC ratio was negatively associated with incident NAFLD with adjusted RR being 0.65 (95% CI 0.48–0.88) for the highest tertile versus the lowest tertile and 0.84 (95% CI 0.76–0.94) for each one SD increase (Model 2). There were linearly inverse associations of TSFA/AFA ratio and ThC/WC ratio with incident NAFLD (Fig. S2A-B). Further, we assessed the associations between TSFA/AFA ratio and incident NAFLD at different stages, with or without advanced fibrosis. With the non-NAFLD as reference, in terms of FIB-4, each one SD increase in TSFA/AFA ratio was associated with a 55% lower risk of incident NAFLD without advanced fibrosis and a 35% lower risk of incident NAFLD with advanced fibrosis (odds ratio [OR] 0.45, 95% CI 0.30–0.69; OR 0.65, 95% CI 0.50–0.85, respectively; Table S3); while according to FNI, each one SD increase in TSFA/AFA ratio was associated with a 30% lower risk of incident NAFLD without advanced fibrosis and a 65% lower risk of incident NAFLD with advanced fibrosis (OR 0.70, 95% CI 0.55–0.90; OR 0.35, 95% CI 0.22–0.55, respectively; Table S3).
Table 2Associations of TSFA/AFA ratio and ThC/WC ratio with incident NAFLD after a 3.6-year follow-up.
| Variable | No. of subjects | No. of cases | Incidence rate (%) | Model 1 (RR, 95% CI) | p value | Model 2 (RR, 95% CI) | p value | Model 3 (RR, 95% CI) | p value |
|---|---|---|---|---|---|---|---|---|---|
| TSFA/AFA ratio | |||||||||
| Tertile 1 | 327 | 113 | 34.6 | Reference | < 0.001§ | Reference | < 0.001§ | Reference | < 0.001§ |
| Tertile 2 | 327 | 84 | 25.7 | 0.76 (0.60-0.96) | 0.87 (0.69-1.09) | 0.90 (0.72-1.13) | |||
| Tertile 3 | 327 | 42 | 12.8 | 0.38 (0.28-0.52) | 0.52 (0.38-0.70) | 0.55 (0.41-0.76) | |||
| TSFA/AFA ratio (per SD) | 0.61 (0.52-0.71) | < 0.001 | 0.69 (0.59-0.81) | < 0.001 | 0.72 (0.62-0.84) | < 0.001 | |||
| ThC/WC ratio | |||||||||
| Tertile 1 | 330 | 107 | 32.4 | Reference | < 0.001§ | Reference | 0.005§ | Reference | 0.045§ |
| Tertile 2 | 322 | 85 | 26.4 | 0.81 (0.64-1.04) | 0.92 (0.73-1.15) | 0.97 (0.76-1.22) | |||
| Tertile 3 | 329 | 47 | 14.3 | 0.45 (0.33-0.61) | 0.65 (0.48-0.88) | 0.72 (0.53-0.98) | |||
| ThC/WC ratio (per SD) | 0.72 (0.65-0.80) | < 0.001 | 0.84 (0.76-0.94) | 0.002 | 0.87 (0.78-0.97) | 0.016 | |||
RR (95% CI) was calculated using the modified Poisson regression model with robust error variance.
TSFA/AFA ratio: men: tertile 1, < 0.23; tertile 2, 0.23-< 0.28; tertile 3, ≥ 0.28; women: tertile 1, < 0.32; tertile 2, 0.32-< 0.40; tertile 3, ≥ 0.40. ThC/WC ratio: men: tertile 1, < 0.58; tertile 2, 0.58-< 0.62; tertile 3, ≥ 0.62; women: tertile 1, < 0.60; tertile 2, 0.60-< 0.64; tertile 3, ≥ 0.64.
§P value for trend.
AFA, abdominal fat area; BMI, body mass index; CI, confidence interval; NAFLD, non-alcoholic fatty liver disease; RR, risk ratio; SD, standard deviations; ThC, thigh circumference; TSFA, thigh subcutaneous fat area; WC, waist circumference.
∗ Model 1 was adjusted for sex, age, education levels, smoking status, drinking status, and leisure-time exercise.
† Model 2 was adjusted for variables in Model 1 and also for BMI.
‡ Model 3 was adjusted for variables in Model 2 and also for hypertension, diabetes, triglyceride, and high-density lipoprotein cholesterol.
3.4 Associations of TSFA/AFA ratio and ThC/WC ratio with remission of NAFLD
As shown in Table 3, TSFA/AFA ratio was positively associated with remission of NAFLD with a multivariable-adjusted RR of 1.40 (95% CI 1.06–1.85) for the highest tertile versus the lowest tertile and of 1.20 (95% CI 1.07–1.34) for each one SD increase (Model 2). The association was attenuated but remained in the same direction after further adjustment for metabolic risk factors (RR per SD 1.18, 95% CI 1.05–1.33; Model 3). Consistent with the associations between TSFA/AFA ratio and remission of NAFLD, ThC/WC ratio was also positively associated with remission of NAFLD with RR per SD increase ratio being 1.22 (95% CI 1.11–1.34) and 1.91 (95% CI 1.39–2.62) for the highest tertile versus the lowest tertile (Model 2). There were linearly positive associations of TSFA/AFA ratio and ThC/WC ratio with remission of NAFLD (Fig. S2C-D).
Table 3Associations of TSFA/AFA ratio and ThC/WC ratio with remission of NAFLD after a 3.6-year follow-up.
| Variable | No. of subjects | No. of cases | Remission rate (%) | Model 1 (RR, 95% CI) | p value | Model 2 (RR, 95% CI) | p value | Model 3 (RR, 95% CI) | p value |
|---|---|---|---|---|---|---|---|---|---|
| TSFA/AFA ratio | |||||||||
| Tertile 1 | 268 | 58 | 21.6 | Reference | 0.004§ | Reference | 0.016§ | Reference | 0.150§ |
| Tertile 2 | 270 | 63 | 23.3 | 1.09 (0.80-1.49) | 1.05 (0.77-1.42) | 0.98 (0.72-1.33) | |||
| Tertile 3 | 268 | 86 | 32.1 | 1.52 (1.15-2.02) | 1.40 (1.06-1.85) | 1.23 (0.92-1.64) | |||
| TSFA/AFA ratio (per SD) | 1.26 (1.11-1.42) | < 0.001 | 1.20 (1.07-1.34) | 0.001 | 1.18 (1.05-1.33) | 0.006 | |||
| ThC/WC ratio | |||||||||
| Tertile 1 | 267 | 42 | 15.7 | Reference | < 0.001§ | Reference | < 0.001§ | Reference | < 0.001§ |
| Tertile 2 | 272 | 69 | 25.4 | 1.67 (1.19-2.34) | 1.54 (1.11-2.14) | 1.47 (1.06-2.04) | |||
| Tertile 3 | 267 | 96 | 36.0 | 2.35 (1.71-3.23) | 1.91 (1.39-2.62) | 1.79 (1.30-2.46) | |||
| ThC/WC ratio (per SD) | 1.29 (1.16-1.43) | < 0.001 | 1.22 (1.11-1.34) | < 0.001 | 1.20 (1.09-1.32) | < 0.001 | |||
RR (95% CI) was calculated using the modified Poisson regression model with robust error variance.
TSFA/AFA ratio: men: tertile 1, < 0.19; tertile 2, 0.19-< 0.23; tertile 3, ≥ 0.23; women: tertile 1, < 0.28; tertile 2, 0.28-< 0.34; tertile 3, ≥ 0.34. ThC/WC ratio: men: tertile 1, < 0.55; tertile 2, 0.55-< 0.59; tertile 3, ≥ 0.59; women: tertile 1, < 0.55; tertile 2, 0.55-< 0.60; tertile 3, ≥ 0.60.
§P value for trend.
AFA, abdominal fat area; BMI, body mass index; CI, confidence interval; NAFLD, non-alcoholic fatty liver disease; RR, risk ratio; SD, standard deviations; ThC, thigh circumference; TSFA, thigh subcutaneous fat area; WC, waist circumference.
∗ Model 1 was adjusted for sex, age, education levels, smoking status, drinking status, and leisure-time exercise.
† Model 2 was adjusted for variables in Model 1 and also for BMI.
‡ Model 3 was adjusted for variables in Model 2 and also for hypertension, diabetes, triglyceride, and high-density lipoprotein cholesterol.
3.5 Mediation by HOMA-IR, TG, and adiponectin
As shown in Table 4, the effect of TSFA/AFA ratio on incident NAFLD was 9.5% (95% CI 3.9%–21.3%) mediated through HOMA-IR; 7.5% (95% CI 1.6%–16.4%) mediated through TG; and 14.9% (95% CI 2.9%–40.5%) mediated through adiponectin. As shown in Table 5, the effect of TSFA/AFA ratio on remission of NAFLD was 23.9% (95% CI 9.2%–67.0%) mediated through HOMA-IR; 19.1% (95% CI 5.6%–61.9%) mediated through TG; and 26.6% (95% CI 10.3%–84.6%) mediated through adiponectin.
Table 4Mediation analyses to estimate the indirect, direct, and total effects of TSFA/AFA ratio on incident NAFLD.
| Mediator | ACME ∗ Two multivariable-adjusted regression models (the linear regression model for the mediator and Poisson regression model for the outcome) were established to evaluate these effects. The effects of TSFA/AFA ratio on incident NAFLD were adjusted for sex, age, education levels, smoking status, drinking status, leisure-time exercise, and BMI. CIs were calculated using percentile bootstrap (replications = 1000). Estimate (95% CI) | ADE ∗ Two multivariable-adjusted regression models (the linear regression model for the mediator and Poisson regression model for the outcome) were established to evaluate these effects. The effects of TSFA/AFA ratio on incident NAFLD were adjusted for sex, age, education levels, smoking status, drinking status, leisure-time exercise, and BMI. CIs were calculated using percentile bootstrap (replications = 1000). Estimate (95% CI) | Total Effect ∗ Two multivariable-adjusted regression models (the linear regression model for the mediator and Poisson regression model for the outcome) were established to evaluate these effects. The effects of TSFA/AFA ratio on incident NAFLD were adjusted for sex, age, education levels, smoking status, drinking status, leisure-time exercise, and BMI. CIs were calculated using percentile bootstrap (replications = 1000). Estimate (95% CI) | Percentage Mediated ∗ Two multivariable-adjusted regression models (the linear regression model for the mediator and Poisson regression model for the outcome) were established to evaluate these effects. The effects of TSFA/AFA ratio on incident NAFLD were adjusted for sex, age, education levels, smoking status, drinking status, leisure-time exercise, and BMI. CIs were calculated using percentile bootstrap (replications = 1000). % (95% CI) | P value† |
|---|---|---|---|---|---|
| HOMA-IR | -0.007 (-0.01 to -0.003) | -0.06 (-0.09 to -0.04) | -0.07 (-0.09 to -0.05) | 9.49 (3.92 to 21.26) | < 0.001 |
| TG | -0.005 (-0.01 to -0.001) | -0.07 (-0.09 to -0.04) | -0.07 (-0.09 to -0.05) | 7.51 (1.59 to 16.45) | 0.016 |
| Adiponectin | -0.01 (-0.02 to -0.002) | -0.06 (-0.08 to -0.03) | -0.07 (-0.09 to -0.04) | 14.94 (2.91 to 40.52) | 0.014 |
| FGF21 | 0.000 (-0.004 to 0.003) | -0.07 (-0.09 to -0.04) | -0.07 (-0.09 to -0.04) | 0.13 (-4.90 to 6.70) | 0.992 |
| RBP4 | -0.001 (-0.004 to 0.002) | -0.07 (-0.09 to -0.05) | -0.07 (-0.09 to -0.05) | 0.86 (-2.53 to 5.24) | 0.570 |
ACME, average causal mediation effects; ADE, average direct effects; AFA, abdominal fat area; BMI, body mass index; CI, confidence interval; FGF21, fibroblast growth factor 21; HOMA-IR, homeostasis model assessment of insulin resistance; NAFLD, non-alcoholic fatty liver disease; RBP4, retinol-binding protein-4; TG, triglyceride; TSFA, thigh subcutaneous fat area.
†P value for percentage mediated.
∗ Two multivariable-adjusted regression models (the linear regression model for the mediator and Poisson regression model for the outcome) were established to evaluate these effects. The effects of TSFA/AFA ratio on incident NAFLD were adjusted for sex, age, education levels, smoking status, drinking status, leisure-time exercise, and BMI. CIs were calculated using percentile bootstrap (replications = 1000).
‡ Loge-transformed before analysis.
Table 5Mediation analyses to estimate the indirect, direct, and total effects of TSFA/AFA ratio on remission of NAFLD.
| Mediator | ACME ∗ Two multivariable-adjusted regression models (the linear regression model for the mediator and Poisson regression model for the outcome) were established to evaluate these effects. The effects of TSFA/AFA ratio on remission of NAFLD were adjusted for sex, age, education levels, smoking status, drinking status, leisure-time exercise, and BMI. CIs were calculated using percentile bootstrap (replications = 1000). Estimate (95% CI) | ADE ∗ Two multivariable-adjusted regression models (the linear regression model for the mediator and Poisson regression model for the outcome) were established to evaluate these effects. The effects of TSFA/AFA ratio on remission of NAFLD were adjusted for sex, age, education levels, smoking status, drinking status, leisure-time exercise, and BMI. CIs were calculated using percentile bootstrap (replications = 1000). Estimate (95% CI) | Total Effect ∗ Two multivariable-adjusted regression models (the linear regression model for the mediator and Poisson regression model for the outcome) were established to evaluate these effects. The effects of TSFA/AFA ratio on remission of NAFLD were adjusted for sex, age, education levels, smoking status, drinking status, leisure-time exercise, and BMI. CIs were calculated using percentile bootstrap (replications = 1000). Estimate (95% CI) | Percentage Mediated ∗ Two multivariable-adjusted regression models (the linear regression model for the mediator and Poisson regression model for the outcome) were established to evaluate these effects. The effects of TSFA/AFA ratio on remission of NAFLD were adjusted for sex, age, education levels, smoking status, drinking status, leisure-time exercise, and BMI. CIs were calculated using percentile bootstrap (replications = 1000). % (95% CI) | P value† |
|---|---|---|---|---|---|
| HOMA-IR | 0.01 (0.005 to 0.02) | 0.04 (0.007 to 0.08) | 0.05 (0.02 to 0.10) | 23.93 (9.24 to 67.04) | 0.002 |
| TG | 0.01 (0.003 to 0.02) | 0.04 (0.007 to 0.08) | 0.05 (0.02 to 0.10) | 19.12 (5.61 to 61.88) | 0.002 |
| Adiponectin | 0.01 (0.006 to 0.02) | 0.04 (0.001 to 0.07) | 0.05 (0.01 to 0.09) | 26.62 (10.31 to 84.65) | 0.014 |
| FGF21 | 0.001 (-0.001 to 0.003) | 0.05 (0.02 to 0.09) | 0.05 (0.02 to 0.09) | 0.96 (-2.31 to 6.22) | 0.524 |
| RBP4 | -0.001 (-0.004 to 0.001) | 0.05 (0.02 to 0.09) | 0.05 (0.02 to 0.09) | -1.34 (-9.36 to 2.93) | 0.526 |
ACME, average causal mediation effects; ADE, average direct effects; AFA, abdominal fat area; BMI, body mass index; CI, confidence interval; FGF21, fibroblast growth factor 21; HOMA-IR, homeostasis model assessment of insulin resistance; NAFLD, non-alcoholic fatty liver disease; RBP4, retinol-binding protein-4; TG, triglyceride; TSFA, thigh subcutaneous fat area.
†P value for percentage mediated.
∗ Two multivariable-adjusted regression models (the linear regression model for the mediator and Poisson regression model for the outcome) were established to evaluate these effects. The effects of TSFA/AFA ratio on remission of NAFLD were adjusted for sex, age, education levels, smoking status, drinking status, leisure-time exercise, and BMI. CIs were calculated using percentile bootstrap (replications = 1000).
‡ Loge-transformed before analysis.
3.6 Associations of TSFA/AFA ratio and ThC/WC ratio with incidence or remission of NAFLD among subgroups
We further divided the subjects into different subgroups according to sex, hypertension, diabetes, and obesity, respectively. As shown in Fig. S3A, the negative associations of TSFA/AFA ratio (RR per SD 0.65–0.76) and ThC/WC ratio (RR per SD 0.77–0.86) with incident NAFLD were all observed in women, in the subgroup with hypertension, in the subgroup without diabetes, and in both non-obese and obese subgroups. As shown in Fig. S3B, the positive associations of TSFA/AFA ratio (RR per SD 1.14–1.34) and ThC/WC ratio (RR per SD 1.14–1.27) with remission of NAFLD were observed in both sexes, in the subgroup with hypertension, in the subgroup without diabetes, as well as non-obese and obese subgroups, but not in the subgroup with diabetes. Additionally, a positive association between TSFA/AFA ratio and remission of NAFLD was observed in subjects with advanced fibrosis (FIB-4: RR per SD 1.24, 95% CI 1.08–1.42; FNI: RR per SD 1.28, 95% CI 1.07–1.53) but not in those without advanced fibrosis (Table S4). Moreover, there were no significant interactions between TSFA/AFA ratio or ThC/WC ratio and any one of these subgroups on incidence or remission of NAFLD (all p for interaction > 0.05; Fig. S3).
4. Discussion
4.1 Major findings
Our study demonstrated for the first time that TSFA/AFA ratio was negatively associated with incident NAFLD (RR per SD 0.69, 95% CI 0.59–0.81) and positively associated with remission of NAFLD (RR per SD 1.20, 95% CI 1.07–1.34). We also found a protective role of the simple surrogate anthropometric indicator–ThC/WC ratio against NAFLD, independent of multiple metabolic risks. Both TSFA/AFA ratio and ThC/WC ratio were adversely associated with metabolic profiles. Our findings stressed that more attention should be paid to his fat distribution and not just gross weight when considering weight control for a given individual. Our findings added the viewpoint that greater thigh subcutaneous fat relative to a given amount of abdominal fat had a protective effect against NAFLD among middle-aged and elderly Chinese to the existing literature.
4.2 TSFA/AFA ratio and metabolic risks
Several studies have reported an inverse association of lower-body adiposity with glucose and lipid levels
21
,22
. For example, one study of 2106 Americans indicated that higher TSFA (measured by computed tomography [CT]) was associated with lower levels of glucose in men and lipids in both sexes, independently of abdominal fat depots;21
another study of 108 Americans showed that leg fat mass (measured by dual-energy X-ray absorptiometry [DXA]) was, independent of the increased risk attributable to trunk fat mass, related to reduced insulin resistance and dyslipidemia.22
. In line with these previous studies, our study showed that larger TSFA/AFA ratio and its surrogate indicator (ThC/WC ratio) were associated with favorable metabolic profiles.4.3 TSFA/AFA ratio and NAFLD risks
So far, the relevant studies examining the associations of TSFA and its relative distribution with NAFLD were limited to a few cross-sectional analyses.
4
,5
. One cross-sectional study involving 408 Koreans observed a negative association of TSFA (measured by CT) with NAFLD in women after adjustment for VFA, ASFA 4
. Another recent cross-sectional study involving 14502 Koreans found that a lower leg fat to total fat ratio (measured by DXA) was associated with a higher risk of prevalent NAFLD.5
. Our prospective cohort study demonstrated that the direction of the association between TSFA (measured by MRI) and incident NAFLD reversed before and after adjustment for BMI with RR of per SD increase in TSFA originally being 1.27 (95% CI 1.12–1.43) and reversed to 0.83 (95% CI 0.71–0.95) (Table S5). This reflected different effects of absolute quantity of a given fat compartment compared to the relative amount of this compartment to the total body fat on NAFLD risks. Thus, we further analyzed the association of the relative distribution of TSFA (reflected by TSFA/AFA ratio) with NAFLD and observed a negative association between the two. In addition, this negative association was observed between TSFA/AFA ratio and both incident liver steatosis and NAFLD with advanced fibrosis.Meanwhile, our study demonstrated that a higher TSFA/AFA ratio was favorably associated with remission of NAFLD (RR per SD 1.20, 95% CI 1.07–1.34). Moreover, our study found that the odds of remission of NAFLD decreased with the presence of fibrosis: subjects with advanced fibrosis defined by FNI were associated with a 36% lower likelihood of remission of NAFLD than those without advanced fibrosis (RR 0.64, 95% CI 0.51–0.81; Table S6). However, in subjects with advanced fibrosis, a favorable effect of TSFA/AFA ratio on remission of NAFLD still existed (FIB-4: RR per SD 1.24; FNI: RR per SD 1.28, both p < 0.05; Table S4). The above significant association was not observed in people with diabetes (Table S4), which may be due to a smaller sample size. Our study was the first to provide prospective evidence that a higher TSFA/AFA ratio, reflecting the propensity to accumulate fat in the thigh subcutaneous compartment rather than in the abdominal compartment, had a protective effect against NAFLD, even against NAFLD with advanced fibrosis.
4.4 ThC/WC ratio and NAFLD risks
Simple anthropometric measurements (WC and ThC) have been widely used in studies on cardiometabolic diseases, acting as a more cost-effective and implementable substitute for accurate-measured adiposity indicators such as AFA and TSFA. The combination of WC and ThC (waist-to-thigh ratio) provides an estimate of body shape and fat distribution, which have been identified as significant predictors of diabetes
8
and all-cause mortality 9
. The Hoorn Study, which included 1357 subjects, reported that an increased waist-to-thigh ratio was associated with a higher risk of incident diabetes (OR: men 1.42; women 1.92) 8
. Another study of 10638 adults demonstrated that a larger waist-to-thigh ratio was associated with an increased risk of all-cause mortality (hazard ratio: men 1.14; women 1.21) 9
. Yet, to date, no studies have examined the association of ThC/WC ratio with NAFLD. Our study was the first to show that an increased ThC/WC ratio had a favorable effect on reducing the risk of incident NAFLD and increasing the remission probability for NAFLD. Our results suggested that ThC/WC ratio could be measured to evaluate the possibility of onset and remission of NAFLD in clinical practice.4.5 Evidence that supports the favorable effect of thigh subcutaneous fat
It has been suggested that subcutaneous fat acts as an 'energy sink', storing fat to buffer the energy surplus, protecting other tissues from lipid overflow
23
. The lipolytic rate is lower in the lower-body fat than in the upper-body fat 24
,25
. The basal blood flow, which is an important determinant of local adipose tissue fatty acid trafficking, of the lower-body fat is slower than the upper-body fat 26
. The above information indicated that the release of NEFA into the systemic circulation from lower-body fat was less than upper-body fat 25
,27
. Chronic exposure to NEFA was associated with both decreased insulin biosynthesis and impaired insulin secretion 28
. Thus, by trapping excess fatty acids 25
, thigh subcutaneous fat might protect our bodies from insulin resistance, which was a well-known pathophysiological hallmark of NAFLD 29
. This was supported by our study, where we found that TSFA/AFA ratio was negatively related to insulin resistance (r = -0.16). Meanwhile, we also found that the associations between TSFA/AFA ratio and incidence and remission of NAFLD could be mediated through insulin resistance (percentage mediated 9.5% and 23.9%, respectively).Adiponectin is exclusively secreted by adipocytes and can decrease the influx of NEFAs, increase fatty acid oxidation, and enhance insulin sensitivity in the liver
30
,31
. Higher gluteofemoral fat mass has been shown to result in higher plasma adiponectin levels and increased insulin sensitivity 32
,33
. Our results provided further prospective evidence in humans that the effect of TSFA/AFA ratio on incidence and remission of NAFLD might be partially mediated by adiponectin (percentage mediated 14.9% and 26.6%, respectively). Moreover, our results might provide some new insights into the pathogenetic pathway of NAFLD and adiponectin might be targeted as a future therapeutic for remission of NAFLD.4.6 Strengths and limitations
Our study had several strengths. First, this was the first community-based prospective cohort study to examine the associations between relative body fat distribution (TSFA/AFA ratio and ThC/WC ratio) and incidence and remission of NAFLD. Second, the adipose tissue areas were accurately measured by MRI. Third, most potential confounding factors had been considered.
We admitted that this study had limitations. First, ultrasound had limited sensitivity at 60%–94% when used to diagnose hepatic steatosis
34
. However, ultrasound was recommended as the first-choice imaging modality to detect hepatic steatosis in clinical practice and large-scale epidemiological studies 35
,36
, because it was cheap and convenient to implement. Second, although ThC and WC were simple and cost-effective anthropometric indicators, they failed to accurately reflect body fat composition. Finally, our results only applied to middle-aged and elderly Chinese.In conclusion, TSFA/AFA ratio and ThC/WC ratio were negatively associated with incident NAFLD but positively associated with remission of NAFLD among middle-aged elderly Chinese. Our findings demonstrated that a favorable fat distribution, characterized by a relatively greater ratio of thigh subcutaneous fat to lower abdominal fat, had a protective role against NAFLD. Individual obesity management should not only focus on weight but also body shape to more effectively reduce the incidence risk of obesity-related metabolic diseases.
7. Data availability statement
De-identified data in our study will not be made available publicly. For further detailed data access policy and procedure, please contact the corresponding author.
6. Acknowledgments
We are grateful to all the investigators for their contributions to this study. The illustrations of graphical abstract have been designed using assets from Freepik.com (see Supplementary materials).
Appendix A. Supplementary data
The following is/are the supplementary data to this article:
8. References
Author names in bold designate shared co-first authorship
- The diagnosis and management of nonalcoholic fatty liver disease: Practice guidance from the American Association for the Study of Liver Diseases.Hepatology. 2018; 67: 328-357
- Body Fat Distribution and Risk of Incident and Regressed Nonalcoholic Fatty Liver Disease.Clin Gastroenterol Hepatol. 2016; 14: 132-138.e4
- Effect of longitudinal changes of body fat on the incidence and regression of nonalcoholic fatty liver disease.Dig Liver Dis. 2018; 50: 389-395
- Association between low thigh fat and non-alcoholic fatty liver disease.J Gastroenterol Hepatol. 2008; 23: 888-893
- The Leg Fat to Total Fat Ratio Is Associated with Lower Risks of Non-Alcoholic Fatty Liver Disease and Less Severe Hepatic Fibrosis: Results from Nationwide Surveys (KNHANES 2008-2011).Endocrinol Metab. 2021; 36: 1232-1242
- Adipose tissue as a buffer for daily lipid flux.Diabetologia. 2002; 45: 1201-1210
- Increased fat intake, impaired fat oxidation, and failure of fat cell proliferation result in ectopic fat storage, insulin resistance, and type 2 diabetes mellitus.Ann N Y Acad Sci. 2002; 967: 363-378
- Associations of hip and thigh circumferences independent of waist circumference with the incidence of type 2 diabetes: the Hoorn Study.Am J Clin Nutr. 2003; 77: 1192-1197
- Influence of central and extremity circumferences on all-cause mortality in men and women.Obesity. 2008; 16: 2690-2695
- Abdominal subcutaneous adipose tissue: a favorable adipose depot for diabetes?.Cardiovasc Diabetol. 2018; 17: 93
- Association of MAFLD With Diabetes, Chronic Kidney Disease, and Cardiovascular Disease: A 4.6-Year Cohort Study in China.J Clin Endocrinol Metab. 2022; 107: 88-97
- Cardiovascular survey methods.Monogr Ser World Health Organ. 1968; 56: 1-188
- Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man.Diabetologia. 1985; 28: 412-419
- Asia-Pacific Working Party on NAFLD. Guidelines for the assessment and management of non-alcoholic fatty liver disease in the Asia-Pacific region: executive summary.J Gastroenterol Hepatol. 2007; 22: 775-777
- A position statement on NAFLD/NASH based on the EASL 2009 special conference.J Hepatol. 2010; 53: 372-384
- Development of a simple noninvasive index to predict significant fibrosis in patients with HIV/HCV coinfection.Hepatology. 2006; 43: 1317-1325
- Non-alcoholic fatty liver disease: A patient guideline.JHEP Rep. 2021; 3100322
- Development and Validation of a Score for Fibrotic Nonalcoholic Steatohepatitis.Clin Gastroenterol Hepatol. 2022;
- Fibrosis nonalcoholic steatohepatitis index validation and applicability considering glycaemic severity and T2D duration.Liver Int. 2022; 42: 2577-2580
- Mediation: R Package for Causal Mediation Analysis.J Stat Softw. 2014; 59
- Low subcutaneous thigh fat is a risk factor for unfavourable glucose and lipid levels, independently of high abdominal fat. The Health ABC Study.Diabetologia. 2005; 48: 301-308
- Sex differences in the association of thigh fat and metabolic risk in older adults.Obesity. 2011; 19: 422-428
- Abdominal obesity and metabolic syndrome.Nature. 2006; 444: 881-887
- Thematic review series: patient-oriented research. Free fatty acid metabolism in human obesity.J Lipid Res. 2006; 47: 1643-1650
- Gluteofemoral body fat as a determinant of metabolic health.Int J Obes. 2010; 34: 949-959
- Upper and lower body adipose tissue function: a direct comparison of fat mobilization in humans.Obes Res. 2004; 12: 114-118
- Lipolysis: contribution from regional fat.Annu Rev Nutr. 1997; 17: 127-139
- Mechanisms linking obesity to insulin resistance and type 2 diabetes.Nature. 2006; 444: 840-846
- Nonalcoholic Fatty Liver Disease and Insulin Resistance: New Insights and Potential New Treatments.Nutrients. 2017; 9: 387
- Adiponectin: more than just another fat cell hormone?.Diabetes Care. 2003; 26: 2442-2450
- The role of the novel adipocyte-derived hormone adiponectin in human disease.Eur J Endocrinol. 2003; 148: 293-300
- Possible role of adiponectin and insulin sensitivity in mediating the favorable effects of lower body fat mass on blood lipids.J Clin Endocrinol Metab. 2006; 91: 1698-1704
- Lower-body fat mass as an independent marker of insulin sensitivity--the role of adiponectin.Int J Obes. 2005; 29: 624-631
- Non-invasive assessment and quantification of liver steatosis by ultrasound, computed tomography and magnetic resonance.J Hepatol. 2009; 51: 433-445
- Low Levels of Alcohol Consumption, Obesity, and Development of Fatty Liver With and Without Evidence of Advanced Fibrosis.Hepatology. 2020; 71: 861-873
- Diagnostic accuracy and reliability of ultrasonography for the detection of fatty liver: a meta-analysis.Hepatology. 2011; 54: 1082-1090
Article info
Publication history
Accepted:
February 28,
2023
Received in revised form:
February 15,
2023
Received:
October 19,
2022
Publication stage
In Press Journal Pre-ProofIdentification
Copyright
© 2023 The Authors. Published by Elsevier B.V. on behalf of European Association for the Study of the Liver (EASL).
User license
Creative Commons Attribution – NonCommercial – NoDerivs (CC BY-NC-ND 4.0) | How you can reuse 
Elsevier's open access license policy
Creative Commons Attribution – NonCommercial – NoDerivs (CC BY-NC-ND 4.0)
Permitted
For non-commercial purposes:
- Read, print & download
- Redistribute or republish the final article
- Text & data mine
- Translate the article (private use only, not for distribution)
- Reuse portions or extracts from the article in other works
Not Permitted
- Sell or re-use for commercial purposes
- Distribute translations or adaptations of the article
Elsevier's open access license policy
00061-7&id=ga1.jpg){kind=link}