key: cord-0881695-rpwio7ow
authors: Hörber, Sebastian; Lehmann, Rainer; Stefan, Norbert; Machann, Jürgen; Birkenfeld, Andreas L.; Wagner, Robert; Heni, Martin; Häring, Hans-Ulrich; Fritsche, Andreas; Peter, Andreas
title: Hemostatic alterations linked to body fat distribution, fatty liver and insulin resistance
date: 2021-05-31
journal: Mol Metab
DOI: 10.1016/j.molmet.2021.101262
sha: 26d5cc272630c8e69f53ce060b134ce98f2ebcf9
doc_id: 881695
cord_uid: rpwio7ow

Objective Obesity, in particular visceral obesity, and insulin resistance emerged as major risk factors for severe Coronavirus Disease 2019 (COVID-19), which is strongly associated with hemostatic alterations. Since obesity and insulin resistance predispose to thrombotic diseases, we investigated the relationship between hemostatic alterations and body fat distribution in subjects at risk for type 2 diabetes. Subjects and methods Body fat distribution (visceral and subcutaneous abdominal adipose tissue) and liver fat content of 150 subjects with impaired glucose tolerance and/or impaired fasting glucose was determined using magnetic resonance imaging and spectroscopy. Participants underwent precise metabolic characterization and major hemostasis parameters were analyzed. Results Procoagulant factors (FII, FVII, FVIII and FIX) and anticoagulant proteins (antithrombin, protein C and protein S) were significantly associated with body fat distribution. In subjects with fatty liver, fibrinogen (298 mg/dl vs. 264 mg/dl, p=0.0182), FVII (99% vs. 90%, p=0.0049), FVIII (114% vs. 90 %, p=0.0098), protein C (124% vs. 111%, p=0.0006) and protein S (109% vs. 89%, p<0.0001) were higher than in controls. In contrast, antithrombin (97% vs. 102%, p=0.0025) was higher in control subjects. In multivariate analyses controlling for insulin sensitivity, body fat compartments and genotype variants (PNPLA3 I148MM/MI/TM6SF2 E167KK/KE), only protein C and protein S remained significantly increased in fatty liver. Conclusions Body fat distribution is significantly associated with alterations of procoagulant as well as anticoagulant parameters. Liver fat plays a key role for the regulation of protein C and protein S suggesting a potential counteracting mechanism to the prothrombotic state in subjects with prediabetes and fatty liver.

During the Coronavirus Disease 2019 (COVID-19) pandemic, obesity and insulin resistance were identified as major risk factors for hospitalization and critical illness of patients with COVID-19 [1; 2] . A hallmark of COVID-19 is the association with thrombotic diseases such as deep vein thrombosis and pulmonary embolism which essentially contributes to the increased morbidity and mortality in severe COVID-19 [2; 3] . Endothelial dysfunction and alterations in the plasmatic coagulation system are potential mechanisms explaining the prothrombotic tendency in COVID-19 [4] [5] [6] .

Obesity and insulin resistance are also associated with a prothrombotic state predisposing to arterial and venous thrombosis [7; 8] . Studies found increased activities of procoagulant parameters (e.g. FVII and FVIII) and increased activities of endothelial dysfunction markers in obese patients that might explain the increased risk for thrombotic diseases [9; 10] .

The site of fat accumulation in the body is a key determinant for the cardiometabolic risk [11; 12] . Recent data suggest that visceral obesity is even a better predictor of the severity of COVID-19, compared to general obesity, but the underlying causes are unclear [13; 14] . Considering liver fat, patients with non-alcoholic fatty liver disease (NAFLD) are also at increased risk for thrombotic diseases [15] . Recently we showed that a 1-year lifestyle intervention improves the prothrombotic tendency in subjects at increased risk for type 2 diabetes [16] . This is, at least partially, mediated by a reduction in liver fat content. However, the impact of fat accumulation in specific depots on the hemostasis system has only been addressed in a few studies, so far [8; 17; 18] . Subjects at increased risk for type 2 diabetes, of which fatty liver is a J o u r n a l P r e -p r o o f 4 common comorbidity [19] , have not been studied in this context, so far. Since most of the major hemostasis parameters are synthesized in the liver and since ectopic fat has become a promising target for therapeutic interventions, it is of great interest to elucidate underlying associations between fatty liver and hemostatic alterations that may help to improve therapeutic strategies to reduce the cardiometabolic risk in these subjects [20] .

Based on findings from COVID-19 patients, we now hypothesized that body fat distribution is a key determinant for alterations in the hemostasis system. We, therefore, investigated the relationship between the amount of visceral and subcutaneous fat, liver fat content and established hemostasis parameters in subjects at risk for type 2 diabetes.

J o u r n a l P r e -p r o o f

In the present study data from 150 subjects with impaired glucose tolerance and/or impaired fasting glucose were analyzed who were randomly selected of the "Prediabetes Lifestyle Intervention Study" (PLIS) cohort (NCT01947595) [21] .

Participants used for the present analysis were exclusively recruited at the University Hospital in Tübingen. Participants (N=9) receiving vitamin-K-antagonists, direct oral anticoagulants or estrogens were excluded due to potential interferences with hemostasis assays. Further exclusion criteria included diabetes mellitus, previous thrombotic events, chronic kidney diseases, active malignant diseases, systemic infections or elevated liver transaminases (>3x of the upper reference range). The study was performed at the Department of Internal Medicine and at the Institute of Clinical Chemistry and Pathobiochemistry of the University of Tübingen. The study has been carried out in accordance with the declaration of Helsinki and was approved by the local ethics committee of the University of Tübingen (055/2012BO1).

All participants provided written informed consent before study enrollment.

All genotyping was performed using a 700k Infinium Global Screening Array from Illumina (SanDiego, CA, USA). Imputation was done on the Haplotype Reference Consortium reference panel using the Michigan Imputation Server. From the analyzed single nucleotide polymorphisms (SNPs), rs58542926 (TM6SF2) was genotyped and rs738409 (PNPLA3) has been imputed (R 2 =0.89).

J o u r n a l P r e -p r o o f

MR examinations were performed in the early morning after an overnight fast on a 3 T whole-body imager (Magnetom Vida, Siemens Healthineers, Erlangen, Germany).

MRI for quantification of visceral adipose tissue (VAT) and subcutaneous abdominal adipose tissue (SCAT) was done using fuzzy means and an extended snake algorithm for segmentation of adipose and visceral adipose tissue according to previous publications [22 ; 23] . Liver fat content was quantified in % by the ratio of lipids (methylene+methyl) and water+lipids by volume selective proton magnetic resonance spectroscopy (1H-MRS).

Determination of laboratory parameters was performed using lithium-heparin, sodium-citrate, sodium-fluoride or clot activator containing tubes (all from Sarstedt, Nümbrecht, Germany). All samples were taken after an overnight fast and immediately centrifuged after blood collection. Afterwards, plasma supernatants were transferred into tubes which were stored at -80°C until determination of laboratory parameters.

Hemostasis measurements were performed using the following reagents from Clinical chemistry parameters (creatinine, glucose, alanine amino transferase, aspartate amino transferase and C-reactive protein) were determined on an ADVIA XPT Clinical Chemistry system (Siemens Healthineers). Glycated hemoglobin (HbA1c) was determined using the Tosoh G8 HPLC Analyzer (Tosoh Bioscience, Sursee, Switzerland).

Data are presented as median and interquartile range (1 st -3 rd ). Insulin sensitivity was calculated based on the results during an oral glucose tolerance test (OGTT) according to the formula by Matsuda and DeFronzo [24] . Correlation analyses shown in figure 1 were performed using the non-parametric Spearman's rank correlation coefficient (r). Linear regression analyses were performed to analyze the relationship between hemostasis parameters and body fat distribution, liver fat content or metabolic variables. Results of clinical characteristics and laboratory parameters of subjects with increased liver fat content (>5.56%) and control subjects (liver fat content ≤5.56%) were compared using the Mann-Whitney-U-Test [25] . Multiple linear regression analysis was performed to adjust for anthropometric and metabolic variables. SCAT and VAT were used as percent (%) of total adipose tissue in linear regression models comparing control subjects and subjects with increased liver fat content. A p-value <0.05 indicates statistical significance. To test for potential multicollinearity the variation inflation factor (VIF) was used. Multicollinearity was assumed when VIF >5. All analyses were conducted using JMP software (V14.2.0, J o u r n a l P r e -p r o o f SAS Institute, Cary, United States). Figure 1 was created with GraphPad Prism Software (version 8). Figure 2 was created using Microsoft Office PowerPoint.

A total of 141 subjects with impaired fasting glucose and/or impaired glucose tolerance were included in the present study. Participants (85 women and 56 men) Investigating the relationship between liver fat content and hemostasis parameters, a positive association with liver fat content was observed for fibrinogen, FVII, FVIII, protein C and protein S (results of correlation analyses are shown in Figure 1 ). Again, antithrombin was negatively associated with liver fat content.

Next, the relationship between hemostasis parameters and several metabolic variables was investigated. Insulin sensitivity, measured as OGTT-derived insulin sensitivity, was associated with all investigated hemostasis parameters, expect PT, FII and FIX (Table 3) . 2h-glucose was related to FVII and protein C. Furthermore, glycated hemoglobin (HbA1c) was related to aPTT, FVII and protein C. Finally, Creactive protein, a marker of inflammation, was found to be associated with all investigated hemostasis parameters except of PT and aPTT.

To specifically address the impact of increased liver fat on the investigated hemostasis parameters, subjects were divided into two groups comprising subjects with increased liver fat content (>5.56%: "fatty liver"; N=60) and control subjects (liver fat content ≤5.56%: "controls"; N=66, table 4). Age and sex of subjects in both groups did not statistically differ. As expected, insulin sensitivity was significantly higher in control subjects and body weight, BMI, fasting glucose and total, visceral and subcutaneous abdominal fat were significantly higher in subjects with fatty liver J o u r n a l P r e -p r o o f (Supplemental table S2 ). Searching for differences in the hemostasis system between both groups, fibrinogen (298 mg/dl vs. 264 mg/dl, p=0.0182), FVII (99% vs. 90%, p=0.0049), FVIII (114% vs. 90%, p=0.0098), protein C (124% vs. 111%, p=0.0006) and protein S (109% vs. 89%, p<0.0001) were found to be increased in subjects with fatty liver (Table 4 ). Antithrombin (97% vs. 102%, p<0.0025) was higher in the control subjects. In multiple regression models, FVII, protein C and protein S remained significantly increased in subjects with fatty liver independent of age, sex and BMI.

Additionally, adjusting for insulin sensitivity, visceral and subcutaneous adipose tissue and genotype variants (PNPLA3 I148MM/MI and TM6SF2 E167KK/KE ), only protein C and protein S remained significantly increased in subjects with fatty liver.

J o u r n a l P r e -p r o o f

The aim of the present study was to investigate the impact of visceral and subcutaneous abdominal adipose tissue as well as liver fat content on alterations in the plasmatic coagulation system in subjects at risk for type 2 diabetes.

First, we showed that both visceral adipose tissue (VAT) and subcutaneous abdominal adipose tissue (SCAT) as well as liver fat are significantly associated with alterations in the plasmatic coagulation system (see Figure 2 ). We identified procoagulant factors (FVII, FVIII and FIX) as well as anticoagulant proteins (antithrombin, protein C and protein S) to be associated with VAT, SCAT and/or liver fat. While most of the differences in plasmatic coagulation between patients with and without fatty liver were driven by important confounders, higher protein C and protein S in case of fatty liver were independent of those. Our findings indicate that body fat distribution is an important determinant of the alterations in the hemostasis system.

In particular, liver fat was identified as an independent determinant of protein C and protein S in subjects at increased risk for type 2 diabetes. These findings suggest a potential counteracting mechanism to the prothrombotic tendency in fatty liver.

Abnormalities in the hemostasis system predisposing for a prothrombotic state are linked to obesity and insulin resistance [8] . In particular, body fat distribution is an Our study extends these findings and demonstrates that SCAT is also, to a similar degree as visceral obesity, associated with increased procoagulant factors (FVIII and FIX). In a study by Winfield et al. it was even shown that SCAT has a stronger relationship with hypercoagulability than VAT following trauma injury [28] . Consequently, fat accumulation in the two major compartments, visceral and subcutaneous abdominal adipose tissue, contributes to cardiometabolic risk via altered plasmatic hemostasis. In line with these findings, antithrombin was negatively associated with VAT and SCAT in the present study. However, the anticoagulant proteins, protein C and protein S, were found to be positively associated with VAT and protein S also with SCAT. Only few studies addressed the association of anticoagulant proteins and visceral obesity and found comparable or Since most of the investigated hemostasis parameters are mainly produced by the liver and fatty liver is closely linked to obesity and insulin resistance, the specific role of increased liver fat on hemostasis parameters needs to be elucidated [31] .

Hemostatic alterations are known to increase the risk of cardiovascular diseases in J o u r n a l P r e -p r o o f non-alcoholic fatty liver disease (NAFLD) [32; 33] . Accordingly, we found increased levels of hemostasis parameters in subjects with fatty liver compared to control subjects without fatty liver. However, for most of these parameters the differences disappeared after adjusting for important confounders. Only protein C and proteins S remained significantly increased in subjects with fatty liver.

Protein C and protein S are vitamin-K dependent proteins and belong to the anticoagulant system. Activated protein C together with protein S is able to specifically inactivate FVa and FVIIIa thereby limiting the coagulation process.

and it is speculated that this might represent a compensatory mechanism for the hypercoagulable state in obesity [34] [35] [36] . However, antithrombin, another natural anticoagulant, was negatively associated with liver fat which corresponds to the prothrombotic state in subjects with fatty liver. Since NAFLD is considered to be associated with hypercoagulability [37; 38] , further studies are required to investigate whether increased activities of protein C and protein S, as observed in our study, can counteract this hypercoagulable state. Of note, the increases in protein C and protein S are a specific feature of liver steatosis/NAFLD and cannot transferred to advanced liver diseases, as patients with liver cirrhosis have markedly decreased levels of anticoagulant proteins including protein C and protein S [37; 39] .

Few other studies investigated the role of liver steatosis/fatty liver or liver diseases in relation to single parameters of the hemostasis system and results were often limited to a small number of subjects [37; 39-42] . Assy et al. reported increased levels of protein C and protein S in subjects with fatty liver compared to control subjects and patients with chronic hepatitis [39; 43] . However, the number of patients with fatty liver was low (N=10-15) and no adjustments for metabolic variables were made. In a J o u r n a l P r e -p r o o f study by Papatheodoridis et al. protein C and protein S levels were reported to be higher in patients with NAFLD (N=60) compared to patients with chronic viral hepatitis (N=90) [41] . Patients without liver diseases were not included and results were not adjusted for metabolic variables. In our study we included subjects with increased risk for type 2 diabetes without overt liver diseases. We performed a comprehensive analysis of procoagulant and anticoagulant parameters in metabolically well characterized subjects using state-of-the-art methods for hemostasis measurements as well as for the determination of adipose tissue compartments and liver fat content. Results were adjusted for relevant variables including insulin sensitivity, body fat composition and genotype variants of important regulators of hepatic triglyceride content (PNPLA3 I148MM/MI and TM6SF2 E167KK/KE ) [44; 45] .

Therefore, we believe that our analyses may adequately assess the specific role of liver fat on the investigated hemostasis parameters.

In conclusion, our analyses identified visceral and subcutaneous abdominal adipose tissue as potential contributors to the alterations in the hemostasis system. Liver fat was independently associated with protein C and protein S suggesting a potential counteracting mechanism to the prothrombotic tendency in subjects with prediabetes and fatty liver.

We gratefully thank all the study participants and acknowledge the technical assistance of Susanne Faix, Janina Roche, Ann Kathrin Horlacher and Isolde Riedlinger. We thank Jennifer Kriebel and Harald Grallert (Institute of Epidemiology, Molecular Epidemiology, Helmholtz Center Munich) for generating and providing the Global Screening Array data.

No potential conflicts of interest relevant to this article were reported. 

The study was supported in part by a grant from the German Center for Diabetes

Research (01GI0925).

The datasets analyzed during the current study are not publicly available due to ethical regulations but are available from the corresponding author upon reasonable request.

[15]

Targher, G., Byrne, C.D., Lonardo, A., Zoppini, G., Barbui, C., 2016. Non-alcoholic fatty liver disease and risk of incident cardiovascular disease: A meta-analysis. J Hepatol 65(3):589-600. 

Global pandemics interconnectedobesity, impaired metabolic health and COVID-19

COVID-19 and diabetes mellitus: from pathophysiology to clinical management

Epidemiological and clinical characteristics of 99 cases of 2019 novel coronavirus pneumonia in Wuhan, China: a descriptive study

Endothelial Cell Contributions to COVID-19

COVID-19 and its implications for thrombosis and anticoagulation

Endothelial cell infection and endotheliitis in COVID-19

C-reactive protein, obesity, and the risk of arterial and venous thrombosis

Thrombosis in central obesity and metabolic syndrome: mechanisms and epidemiology

Mechanisms linking obesity with cardiovascular disease

Hemostatic dysfunction associated with endocrine disorders as a major risk factor and cause of human morbidity and mortality: a comprehensive meta-review

Visceral and ectopic fat, atherosclerosis, and cardiometabolic disease: a position statement

Causes, consequences, and treatment of metabolically unhealthy fat distribution

The role of visceral adiposity in the severity of COVID-19: Highlights from a unicenter cross-sectional pilot study in Germany

Visceral fat is associated to the severity of COVID-19

Standardized assessment of whole body adipose tissue topography by MRI

Insulin sensitivity indices obtained from oral glucose tolerance testing: comparison with the euglycemic insulin clamp

Magnetic resonance spectroscopy to measure hepatic triglyceride content: prevalence of hepatic steatosis in the general population

High levels of factor IX increase the risk of venous thrombosis

Role of clotting factor VIII in effect of von Willebrand factor on occurrence of deep-vein thrombosis

Adipose tissue location and contribution to postinjury hypercoagulability

Associations of haemostatic variables with body mass index: a community-based study

Activated protein C levels in obesity and weight loss influence

Non-alcoholic fatty liver disease as a cause and a consequence of metabolic syndrome

Non-alcoholic fatty liver disease and risk of cardiovascular disease

NAFLD and increased risk of cardiovascular disease: clinical associations, pathophysiological mechanisms and pharmacological implications

Insulin sensitivity and hemostatic factors in clinically healthy 58-year-old men

Coagulation and fibrinolysis abnormalities in obesity

Bariatric surgery in morbidly obese individuals affects plasma levels of protein C and thrombomodulin

Procoagulant imbalance in patients with non-alcoholic fatty liver disease

Nonalcoholic Fatty Liver Disease-Evidence for a Thrombophilic State

Association between thrombotic risk factors and extent of fibrosis in patients with nonalcoholic fatty liver diseases

Prothrombotic factors in histologically proven nonalcoholic fatty liver disease and nonalcoholic steatohepatitis

Thrombotic risk factors and liver histologic lesions in nonalcoholic fatty liver disease

Increased coagulation factor VIII, IX, XI and XII activities in non-alcoholic fatty liver disease

Elevated plasma protein C levels correlate with the presence of fatty liver (NASH and NAFLD)

Genetic variation in PNPLA3 confers susceptibility to nonalcoholic fatty liver disease

Exome-wide association study identifies a TM6SF2 variant that confers susceptibility to nonalcoholic fatty liver disease

• The link between liver fat and altered hemostasis is mainly driven by confounders like BMI and insulin sensitivity.• Only protein C and protein S were significantly and independently increased in subjects with fatty liver.

Declarations of interest: none