Prothrombotic factors and nonalcoholic fatty liver disease: An additional link to cardiovascular risk?


  • Francisco Barrera M.D.,

    1. Storr Liver Unit, Westmead Millennium Institute, University of Sydney at Westmead Hospital, Sydney, New South Wales, Australia
    2. Departamento de Gastroenterología, Pontificia Universidad Católica, Santiago, Chile
    Search for more papers by this author
  • Jacob George MBBS, Ph.D., FRACP

    Corresponding author
    1. Storr Liver Unit, Westmead Millennium Institute, University of Sydney at Westmead Hospital, Sydney, New South Wales, Australia
    • Address reprint requests to: Professor Jacob George, Darcy Road Westmead NSW 2145, Australia. E-mail:; fax: +61-2-96357582.

    Search for more papers by this author

  • Potential conflict of interest: Nothing to report.

  • Supported by a Robert W. Storr Bequest to the Sydney Medical Foundation and National Health and Medical Research Council (NHMRC 632630 and 1049857).

  • See Article on Page 121


extracellular matrix




matrix metalloproteinase


plasminogen activator inhibitor-1


tissue-type plasminogen activator


urokinase plasminogen activator.

One of the vexing questions in clinical hepatology is defining the specific and independent contribution of the liver to systemic metabolic dysregulation, defined operationally by the term metabolic syndrome. The latter comprises a spectrum of disorders including obesity, insulin resistance, hypertension, and dyslipidemia. Clarifying the conundrum is difficult, as we tend to practice in silos as hepatologists or diabetic specialists, rather than as physicians. Ideally, defining the role of the liver to cardiometabolic risk requires prospective, well-defined, large patient cohorts with baseline liver histology, determinations of fat depots, an assessment of endocrine function and insulin sensitivity, together with long-term follow-up and regular liver assessments.

Nevertheless, some progress has been made. Several cross-sectional studies have described an association between the presence of nonalcoholic fatty liver disease (NAFLD) and markers of atherosclerosis such as carotid artery thickness, endothelial dysfunction, coronary artery calcification, and stenosis.[1] Epidemiological studies have also demonstrated an association between an imaging-based diagnosis of NAFLD and an increased risk of coronary, cerebrovascular and peripheral vascular disease, and mortality. While some of these associations persist after adjusting for traditional cardiovascular risk factors,[2, 3] in others the association is lost.[4, 5] The latter, however, does not negate a role for the liver since, for example, atherogenic dyslipidemia is, in large part, liver-dependent.

Numerous mechanisms have been proposed to explain the contribution of NAFLD to cardiovascular risk, including hepatic insulin resistance, atherogenic dyslipidemia, hepatic inflammation, and a prothrombotic milieu.[6] In fatty liver, the accumulation of diacylglycerol and sphingolipids enhances hepatic insulin resistance.[7] Likewise, hepatic cholesterol synthesis and very-low-density lipoprotein (VLDL) export are increased, catabolism of triglyceride-rich HDL is augmented, and LDL receptors are reduced, favoring the presence of cholesterol-rich VLDL remnants, small atherogenic LDL, and reduced HDL.[8, 9] In addition, hepatocytes, macrophages, and stellate cells are sensitized to endo- and exotoxins, increasing inflammatory cytokines and promoting a systemic proinflammatory atherogenic state.[6]

Since the liver is the site of production of most coagulation factors, clotting factor abnormalities are expected in liver disease. In this issue, Verrijken et al.[10] clarify this aspect of the NAFLD-metabolic syndrome-cardiovascular risk puzzle. In a large well-characterized group of NAFLD subjects (n = 273), serum levels of five procoagulant factors (factors VII, VIII, XI, von Willebrand, and fibrinogen), two anticoagulant factors (protein C and AT III), activated protein C resistance (APC-R) and plasminogen activator inhibitor-1 (PAI-1) were quantified and platelet function assessed. In accordance with prior data, a correlation was observed between components of the metabolic syndrome and elevated fibrinogen, factor VIII, von Willebrand factor and PAI-1, and decreased ATIII.[11, 12] Interestingly, PAI-1 was the only factor associated with hepatic histology, namely, steatosis, inflammation, ballooning, and fibrosis. In multivariate analysis, steatosis was an independent predictor of PAI-1 levels, after adjusting for metabolic factors. However, only 12% of its variance was explained by hepatic histology, probably a consequence of the ubiquitous expression of PAI-1. These findings align with prior reports in NAFLD where PAI-1 was elevated and in which the association persisted after adjusting for metabolic factors.[13] Importantly, in a subgroup who had available liver tissue, PAI-1 expression was higher in those with nonalcoholic steatohepatitis (NASH) than those without, suggesting that the increased PAI-1 derives, at least partially, from liver.

PAI-1 is a member of the serine protease inhibitor proteins family that inhibits tissue-type plasminogen activator (tPA) and urokinase plasminogen activator (uPA) (Fig. 1), the major enzymes involved in activating plasmin and inducing fibrinolysis after clot formation. PAI-1 synthesis is ubiquitous, including by vascular endothelium, platelets, adrenals, and liver. Elevated PAI-1 levels have been extensively reported as a risk factor for thrombosis and cardiovascular events.[14] PAI-1 levels are increased in metabolic disease by various stimuli including insulin, angiotensin, renin, tumor necrosis factor alpha, transforming growth factor beta, and lipopolysaccharide (LPS). Notably, PAI-1 plays a role in fibrosis in liver and other organs. The mechanism involves matrix metalloproteinases (MMPs), a group of plasmin-activated enzymes implicated in the degradation of extracellular matrix (ECM). By reducing plasminogen activation to plasmin, PAI-1 shifts the balance towards ECM deposition and fibrosis[15] (Fig. 1). Mouse models with abrogated PAI-1 have elevated MMP-9 activity and are resistant to liver fibrosis following bile duct ligation.[16] In addition, the reduction in fibrinolysis increases deposition of fibrin in liver parenchyma and sensitizes it to LPS-induced necrosis and inflammation.[17] Thus, the present findings represent a potential link between NAFLD and cardiovascular risk and liver fibrosis.

Figure 1.

NAFLD and cardiovascular disease. NAFLD is associated with insulin resistance, systemic inflammation, dyslipidemia, and, more recently, a prothrombotic state. These factors combine to increase endothelial dysfunction, atherogenesis, plaque rupture, thrombosis, and vascular occlusion. Hepatic inflammation in NAFLD activates stellate cells, with ECM deposition, and hepatocyte lipoapoptosis resulting in further amplification of insulin resistance, systemic inflammation, dyslipidemia, and the prothrombotic state. In this issue, Verrijken et al.[10] demonstrate that PAI-1 levels are independently associated with NAFLD histologic severity. Elevated PAI-1 promotes a prothrombotic milieu and cardiovascular risk by reducing tPA and uPA-mediated plasmin activation, consequently reducing fibrin degradation and thrombolysis. Increased PAI-1 may also induce tissue fibrosis by reducing plasmin mediated activation of MMPs, resulting in reduced matrix degradation.

Some interesting questions arise from this study. The contribution of gender, the PAI-1 4G/5G polymorphism, and ethnicity to PAI-1 variance in NAFLD remains to be elucidated.[14] The low prevalence of overweight (7.9% with body mass index [BMI] <30), advanced fibrosis (9.1%), and the exclusion of diabetic subjects (9%) limits the applicability of the results to these subgroups that are typical of “office” cases of NAFLD. However, perhaps the most important but unanswered question from a cross-sectional study is whether NAFLD-associated increases in PAI-1 promotes cardiovascular disease or liver fibrosis progression. Conceivably, such considerations are more academic than of clinical consequence, and not easily answered without carefully designed longitudinal studies. For example, PAI-1 is consistently associated with obesity, insulin resistance, diabetes mellitus, and a sedentary lifestyle, all predictors for the development of both NASH and cardiovascular disease.

But will defining an independent link of NAFLD to cardiovascular risk change NAFLD treatment? The benefit of pharmacological strategies for primary prevention of cardiovascular disease in NAFLD patients (e.g., antiplatelet agents) has not been demonstrated. The cost-effectiveness of this measure depends on demonstrating that NAFLD poses a significant additional cardiovascular mortality risk compared to traditional factors. On the other hand, implementing specific therapy with vitamin E or pioglitazone in NAFLD could theoretically be an attractive intervention to reduce cardiovascular risk. However, properly designed prospective studies and validation of new interventions need to be performed before recommending their use for this specific indication, considering their adverse effects and costs. In the meantime, the “simplest” approach would be early initiation of lifestyle intervention therapies. Although long-term compliance continues as its major drawback, the weight-independent reduction of PAI-1 observed in obese diabetic subjects undergoing lifestyle intervention should be an additional incentive to promote it, and hopefully modify cardiovascular risk and adverse liver-related outcomes.[18]

  • Francisco Barrera, M.D.1,2

  • Jacob George, MBBS, Ph.D., FRACP1

  • 1Storr Liver Unit

  • Westmead Millennium Institute

  • University of Sydney at Westmead Hospital

  • Sydney, New South Wales, Australia

  • 2Departamento de Gastroenterología

  • Pontificia Universidad Católica

  • Santiago, Chile