lable at ScienceDirect

Atherosclerosis 239 (2015) 192e202

Contents lists avai

Atherosclerosis

journal homepage: www.elsevier .com/locate/atherosclerosis

Review

A concise review of non-alcoholic fatty liver disease

Nwe Ni Than*, Philip N. NewsomeThe Centre of Liver Research and NIHR Biomedical Research Unit in Liver Diseases, University of Birmingham and University Hospital Birmingham NHSTrust, UK

a r t i c l e i n f o

Article history:Received 27 October 2014Received in revised form8 January 2015Accepted 12 January 2015Available online 13 January 2015

Keywords:Non-alcoholic fatty liver diseasePathogenesisDyslipidaemiaInsulin resistanceCardiovascular disease

* Corresponding author.E-mail addresses: nwenithan@gmail.com, sophiath

http://dx.doi.org/10.1016/j.atherosclerosis.2015.01.0010021-9150/© 2015 Elsevier Ireland Ltd. All rights rese

a b s t r a c t

Non-alcoholic fatty liver disease (NAFLD) is the hepatic manifestation of the metabolic syndrome and theincidence of which is rising rapidly due to the increasing epidemic of obesity in both adults and children.The initial accumulation of fat followed by subsequent inflammation is central to the development ofliver damage, and is critically influenced by host factors including age, gender, presence of diabetes,genetic polymorphisms and more recently by the gut microbiome. An increasing body of data suggestthat NAFLD is also an independent risk factor of cardiovascular disease, which remains the commonestcause of mortality in such patients. This review focusses on the pathogenesis of NAFLD, and the evolutionof new approaches to the management and treatment of NAFLD.

© 2015 Elsevier Ireland Ltd. All rights reserved.

1. Introduction

Non-alcoholic fatty liver disease (NAFLD) is becoming animportant public health concern due to the rising incidence ofobesity in both children and adults. NAFLD, defined by the presenceof hepatic steatosis (the presence of fat in liver parenchymawithoutinflammation) in the absence of excess alcohol consumption (lessthan 21 units in men and 14 units in women), is considered torepresent the hepatic component of metabolic syndrome [1].NAFLD represents a spectrum of disease, ranging from simplesteatosis to steatohepatitis (the presence of fat in liver parenchymawith inflammation, hepatocyte ballooning and lobular inflamma-tion) through to fibrosis and cirrhosis [2]. Simple steatosis (SS)rarely progresses to advanced disease whereas in approximately20% of patients with non-alcoholic steatohepatitis (NASH), it pro-gresses to fibrosis and cirrhosis over a 15 year time period [3].Clinical progression of NAFLD has been illustrated in Fig. 1.

It is strongly associated with insulin resistance and othermetabolic risk factors such as diabetes mellitus, central abdominalobesity and dyslipidaemia [4]. NAFLD is an independent risk factorfor cardiovascular disease (CVD) and predicts future events, inde-pendently of other risk factors such as age, gender, low density li-poprotein (LDL) cholesterol, smoking and other features of

an@ymail.com (N.N. Than).

rved.

metabolic risk factors [5e7]. NAFLD is also associated withincreased risk of all-cause mortality, contributed by liver relateddeaths as well as non-liver related causes such as malignancy,diabetes and coronary artery disease [8].

2. Epidemiology

The prevalence of NAFLD in normal weight individuals withoutthe presence of metabolic risk factors is reported to be around 16%[9] rising to 43e60% in patients with diabetes [10,11], 91% in obesepatients undergoing bariatric surgery [12], and up to 90% in pa-tients with hyperlipidaemia [4,13]. The prevalence of NAFLD alsoincreases with age from less than 20% under the age of 20 to morethan 40% in over the age of 60 [6] and indeed older age has beenshown to be an independent risk factor for hepatic steatosis andprogression to fibrosis and cirrhosis [14]. The male gender has beenregarded as a risk factor for progression to NASH and fibrosis [14].Although earlier studies suggested that ethnicity impacts on theprevalence of NAFLD [15], subsequent studies did not confirm thisonmulti-variant analysis [16]. Intriguingly though the link betweeninsulin resistance (IR) [measured by homeostasis modelassessment-estimated insulin resistance (HOMA-IR)] and the riskof NASH is seemingly different between Latino and non-Latino in-dividuals [16].

Mortality is increased in patients with steatohepatitis andadvanced fibrosis but not in patients without evidence of steato-hepatitis and fibrosis (known as bland steatosis). A long term follow

Fig. 1. Clinical progression of non-alcoholic fatty liver disease.

N.N. Than, P.N. Newsome / Atherosclerosis 239 (2015) 192e202 193

up study of 129 patients with biopsy proven NAFLD showed thatmortality was not increased in patients with simple steatosis [17],but increased in patients with NASH. The mortality was relatedmainly to cardiovascular disease although liver related deaths weremore common in patients with NASH cirrhosis [17]. A recent sys-tematic review of 221 patients with biopsy proven NASH showedthat patient's age and the degree of inflammation on initial liverbiopsy were independent predictors of progression to advancedfibrosis whilst factors such as diabetes, hypertension and obesitywere not statistically significant predictors [18]. These findingssupported that the presence of advanced fibrosis was associatedwith increased overall mortality [19], most likely from cardiovas-cular events [19].

Hamabe et al. demonstrated in a retrospective study over a 10year period that smoking was an independent risk factor for NAFLDirrespective of the presence of other metabolic risk factors [20].Notably, the rate of development of NAFLD was similar in thoseindividuals that stopped smoking to those that carried on, perhapsrelated to subsequent weight gain in individuals after cessation ofsmoking [20]. Modest alcohol consumption (one alcoholic beverageper day) had not been demonstrated to increase the prevalence ofNAFLD, and in the case of modest wine consumption it wouldappear to reduce the incidence [21].

3. Natural history and pathogenesis of NAFLD

The natural history and the pathogenesis of NAFLD is not clearlywell described. In NAFLD, simple steatosis is regarded as the pres-ence of fat in <5% of hepatocytes [22] and in about 20e25% of cases,it progresses to NASH and of these patients with NASH, 20% willdevelop fibrosis and subsequently cirrhosis [3]. The mechanismsinvolved in the progression of steatosis to NASH are complex andnot completely understood although increased visceral adiposityand insulin resistance (IR) with increased free fatty acids (FFA)release might play a role in the development of liver steatosis [22].In healthy subjects, insulin stimulates hepatic as well as peripheralglucose uptake and suppress hepatic glucose production [23]whereas in the fasting state, the liver becomes the major site ofglucose production as mediated by gluconeogenesis and glyco-genolysis [2,24]. In patients with IR, hepatic auto-regulation isdisrupted and therefore, both gluconeogenesis and glycogenolysisare increased resulting in the development of hyperglycaemia [13].

The mechanism of the liver injury in NAFLD is currently thought tobe a ‘multiple hit process’ involving IR, oxidative stress, apoptosisand perturbations of adipokines levels [25].

3.1. Two hit hypothesis

The current model of ‘two hit hypothesis in NAFD’was proposedin 1998 by Day et al. [26]. The first hit reflects the accumulation oftriglycerides (TG) and FFA in hepatocytes which is a consequence ofIR, enhanced dietary influx and increased hepatic lipogenesis [26].The second hit involves lipid peroxidation, mitochondrialdysfunction and inflammation which results in hepatocyte damageand development of liver fibrosis [26]. Activation of pro-inflammatory pathway is mediated by cytokine and patternrecognition receptors, including toll like receptors and thesepathways emerge on two main intracellular signalling pathways,known as nuclear factor-kb (NF- kb) and c-Jun N-terminal kinase(JNK) [27,28]. NF- kb activation is reported in NASH and can lead toincreased transcription of many pro-inflammatory genes, whereasJNK causes IR via direct phosphorylation and degradation of insulinreceptor substrate 1 (IRS1), reducing the intracellular signallingpathway downstream the insulin receptor [27]. Lipid peroxidationcan promote stellate cell proliferation contributing to fibrogenesis[29], whereas reactive oxygen species (ROS) induce cytokinesrelease from hepatocytes that lead to the initiation of various im-mune mediated mechanisms contributing to further liver cellinjury. The combination of hyperinsulinaemia, hepatic iron andlipid peroxidation induce oxidative stress [15], which can causemitochondrial dysfunction in NASH and contribute to TG accumu-lation and eventually cell necrosis [11].

3.2. Insulin resistance

Patients with NAFLD have reduced insulin sensitivity not only inmuscle but also in liver and adipose tissue [30], which play a majorrole in the pathogenesis of NAFLD. Due to IR, the adipose tissuebecomes resistant to the anti-lipolytic effect of the insulin and re-sults in peripheral lipolysis which causes an increase delivery of FFAto the liver as well as driving de novo lipogenesis (DNL) [6,15]. Inaddition, lipid overload in pancreatic-B cells leads to dysregulatedinsulin secretion and changes in the expression of peroxisomeproliferator-activated receptor-a (PPAR-a), glucokinase, the glucose

N.N. Than, P.N. Newsome / Atherosclerosis 239 (2015) 192e202194

transporter-2 (GLUT 2), pre-pro-insulin and pancreatix duodenalhomeoboc-1 (PDX-1), which can lead to IR as a result of FFA-induced B-cell apoptosis [12].

A study on liver insulin knock out mice (LIKO) showed that on astandard chow diet, the mice have a proatherogenic lipid profilewith low high density lipoprotein (HDL) cholesterol and very lowdensity lipoprotein (VLDL) particles that are mainly enriched incholesterol, although in the case of normal mice, VLDL is mainlycomposed of TG [31]. Within 12 weeks on an atherogenic diet (15%dairy fat, 1% cholesterol, 0.5% cholic acid), the mice showed markedhypercholesterolemia and all of LIKO mice developed severeatherosclerosis [31]. Therefore, it has been suggested that IR at thelevel of liver is sufficient enough to produce the dyslipidaemia andincreased the risk of atherosclerosis [32].

3.3. Free fatty acids

De novo lipogenesis (DNL) is an integrated metabolic pathwayconsisting of glycolysis (conversion of glucose to acetyl-CoA);biosynthesis of saturated fatty acids followed by desaturation,and the formation of TG and this pathway mainly takes place inadipose tissue and liver. Dysregulation of DNL is observed in pa-tients with obesity, metabolic syndrome or NAFLD. DNL is up-regulated after a high carbohydrate and low fat diet leading to anincrease in blood triacylglycerol levels that leads to an increase inthe synthesis and secretion of VLDL in the liver and subsequentlyhepatic hypertriglyceridemia leading to steatosis [33]. Hepatic DNLis caused as a result of fat accumulation in the liver contributed bythe increased intake of dietary FFA as well as the result of periph-eral insulin resistance. [37]. It has been demonstrated that around26% of FFAs in the liver derived from hepatic DNL in individualswith NAFLD [38] with the remaining TG content being derived fromlipolysis of adipose tissue store (59%) and from diet (15%) [38].Although the exact underlying mechanism is not completely un-derstood, it has been suggested that hepatic IR can be both a causeand/or a consequence of steatosis in the liver. Chronic hyper-insulinaemia, as found in NAFLD, promotes hepatic DNL through anup-regulation of lipogenic transcription factors [39,40].

A study [33] has demonstrated successfully that both serumconcentration of carbomoyl phosphate synthase 1 (CPS1) andglucose regulated protein (GRP78) has decreased gradually insubjects from SS to NASH. GRP78 is immunoglobulin binding pro-tein and is a central regulator of endoplasmic reticulum (ER)function and crucial for cell survival [34]. GRP78 inhibit both insulindependent and ER-stress dependent sterol regulatory element-binding protein 1c (SREBP-1c) proteocleavage that plays animportant role in DNL [34]. SREBP-1c regulates genes required forglucose metabolism, fatty acid and lipid production and its activityis regulated by insulin [35].

Recent rodent studies demonstrated that in obesity, DNL isdown regulated inwhite adipose tissue (WAT) and that its selectiverestoration in WAT reverts obesity-dependent IR [34,35]. Humanstudies demonstrate similar findings; DNL enzymes are markedlysuppressed in WAT of obese subjects, as is the glucose transportertype 4 (GLUT-4) [33]. This suppression is closely linked to impairedmetabolic control and can be reversed by weight loss throughbariatric intervention. This suggests that the level of DNL or itsrelevant products such asmonounsaturated fatty acids inWAT is animportant determinant of systemic IR and subsequent metabolicdisease [33]. In contrast to WAT, DNL in the liver has been found tobe upregulated in rodent and human obesity where it is believed topromote lipotoxicity, IR, atherogenic dyslipidaemia and NAFLD[36]. Based on this association between hepatic DNL and themetabolic syndrome, it is believed that inhibition of DNL may be aviable approach to treating obesity-related disorders such as type 2

diabetes (T2D) [37].The mechanisms of lipotoxicity as consequence of ectopic fat

accumulation in the liver are poorly described. When the energyexpenditure is less than energy intake, there is abundance of FFAresulting in overspill which are subsequently stored in muscle andliver as ectopic fat [38]. Lipid overload can also be found in otherorgans such as heart, muscle, pancreas and arterial wall where it isstored ectopically [13,39]. Adipose tissues can store large amountsof excess FFA and when the limited storage is exceeded, celldysfunction or death will result, known as lipotoxicity [39].Mechanisms of lipotoxicity involve apoptosis, decreased cardiolipincontent, increasedmembrane permeability and cytochrome releasein mitochondria, nuclear factor-kappa beta (NF-kb) activation andoxidative stress [39,40]. Previous studies from transgenic micewithover expression of lipoprotein lipase lead to lipid accumulation inthe liver and in muscle leading organ specific IR [41,42]. In NAFLD,the accumulation of hepatic fat is believed to be the consequence ofan imbalance between the uptake/production of fat in the liver andits subsequent secretion or metabolism [36].

Kupffer cells (KC) constitute the largest cellular component ofthe reticuloendothelial system, representing 80%e90% of all tissuemacrophages in the body [43] and about 10% of the resting totalliver cell population. They are located in the liver sinusoids and canmigrate into the space of Disse where they orchestrate cross-talkbetween various resident and recruited cells of the liver [44]. He-patic steatosis and inflammation is alsomediated by KC through theToll-like receptors (TLR), mainly TLR-3 and TLR-4 as well as scav-enger receptor pathway, mediating production of cytokines such asTNF-a and Interleukins-15 (IL-15) and IL-1 b [45,46]. Animal studyshowed that the effect of KC on liver TG are medicated partially byIL-1 b which suppresses PPAR-a which is a gene involved in fattyacid oxidation [46]. It has also been shown that KC in NAFLD have adefective phagocytic function which may result in an impairedclearance of toxic substances, such as endotoxin, leading to hepa-tocyte damage [47], as supported by a study showing that thenumbers of hepatic CD68þKC correlate with histological severity inpatients with NAFLD [48].

3.4. Adipose tissue

Adipose tissue is themajor source of FFA and responsible for 60%of TG accumulation. Previous studies had suggested that obeseindividuals often have enlarged adipocytes due to lipid overload.Excess lipid contents spill over from the incompetent anddysfunctional adipose tissue lead to ectopic lipid deposition in or-gans such as liver and muscle [48]. Moreover, patients with centralobesity have large amounts of visceral adipose tissue (VAT) which isdefined as intra-abdominal fat bounded by parietal peritoneum ortransversalis fascia and it is a major source of FFA, hormones, cy-tokines such as IL-6, tumour necrosis factor (TNF)-a, plasminogenactivator inhibitor-1 (PAI-1), leptin, and adipokines [49]. Thesefactors are secreted into the circulation and frequently taken up bythe liver, with the peripheral FFA flux derived from VAT being themost notable example [39]. This flux in particular is a majordeterminant of accumulation of hepatic and lipoprotein fat inNAFLD [39]. Waist circumference is highly correlated with VAT inboth genders and is used as a clinical marker for abdominal obesityand found to be better predictor than body mass index (BMI) [25].

Adiponectin is a protein that acts a protective adipokine byinhibiting liver gluconeogenesis and suppressing lipogenesis [25],an effect mediated mainly through activation of AMP mediatedprotein kinase (AMPK) and PPAR-a, thus stimulating fatty acidoxidation in liver and muscle [13]. Adiponectin levels also correlateinversely with plasma TG and positively with HDL-cholesterol andlow LDL-cholesterol levels [13]. Patients with NAFLD have lower

N.N. Than, P.N. Newsome / Atherosclerosis 239 (2015) 192e202 195

adiponectin concentration than normal subjects despite higherlipolysis and fatty acid concentrations [13]. Therefore, low adipo-nectin in NAFLD enhances FFA overload and lipid oxidation andconsequently play a role in progression from steatosis to NASH [7](Fig. 2).

3.5. Death and repair

Hepatocyte apoptosis plays a critical role in liver injury and thedevelopment of NASH [50]. Soluble Fas, a death receptormembraneof the TNF family appears to have an important role in hepatocyteapoptosis. The accumulation of FFA in hepatocytes triggeredupregulation of Fas ligands and activation of Fas receptors whichresults in apoptosis [51]. In addition, effector caspases (mainlycaspase-3) cleave various substrates inside the cell includingcytokeratin 18 (CK-18), which is a major intermediate filamentprotein in the liver, resulting in apoptosis [52]. Fragments of CK-18are detectable in the bloodstream by ELISA and it has a promisingaspect as a diagnostic tool in near future.

Hedgehog (Hh) signalling has been implicated in hepatic tissuerepair resulting in its study in NAFLD to establish whether aberrantor prolonged signalling may impact on disease progression inNAFLD [14]. Notably, the level of Hh activity seems to be propor-tional to the severity and duration of liver disease in both rodentsand humans [14]. Macrophage infiltration is also seen in patientwith NASH [24], and monocyte chemoattractant protein (MCP)-1 isalso an important factor in NASH as it can potentially mediates theprogression of disease by causing persistent inflammation as aresult of infiltration of leucocytes into the liver [24].

3.6. Genetic polymorphisms

Two independent genome-wide association studies linked thecommon rs738409 polymorphism of palatin-like phospholipasedomain-containing 3 gene (PNPLA3) with hepatic fat content and

Fig. 2. Pathogenesis of non-alc

liver enzyme levels [53,54]. This single variant (rs738409) sub-stitutes cytosine to guanine that changes codon 148 of the proteinfrom isoleucine to methionine [55]. The protein encoding PNPLA3,known as adiponutrin is particularly associated with increasedhepatocyte fat content [14]. In humans, PNPLA3 ismainly expressedin intracellular membrane fractions in hepatocytes, and is inducedin the liver after feeding and during IR [53]; the G allele has beenshown to be associated with increased risk of hepatic TG accu-mulation ad hence NAFLD [14]. The normal genotype is CC alleleand the most common found in NAFLD is CG allele and GG allelewhich have the worse outcome with rapid progression to fibrosisand cirrhosis [56,57]. Since the discovery of the association be-tween the PNPLA3 mutation with steatosis and steatohepatitis,several additional single nucleotide polymorphisms (SNPs) havebeen identified in individuals with NASH. Genome wide studieslooking at non-Hispanic, Caucasian women with biopsy provenNAFLD [58] showed that non-alcoholic steatosis (NAS) was signif-icantly associated with rs2645424 on chromosome 8 in the farnesyldiphosphate farnesyl transferase-1 gene (FDFT-1), which is a keyregulator of cholesterol biosynthesis. This pilot study also showedthat the degree of fibrosis was strongly correlated with SNPrs343062 on chromosome 7 although the exact function of this SNPis unknown. Three SNPs were associated with the lobular inflam-mation phenotype: SNP rs1227756 on chromosome 10 in theCOL13A1 (and collagen, type XIII, a 1) gene, rs6591182 on chro-mosome 11, and rs887304 on chromosome 12 in the EF-hand cal-cium binding domain 4B(EFCAB4B) gene.

Another common gene variant, glucokinase regulatory protein(GCKR) has been studied extensively. GCKR regulates glucokinase, aphosphorylating enzyme that responsible for hepatic glucosemetabolism and activates hepatic lipogenesis [59]. Polymorphismof this particular gene (rs780094 and rs1260326) was associatedwith increased risk of type 2 diabetes, especially in Asian popula-tion [60].

oholic fatty liver disease.

N.N. Than, P.N. Newsome / Atherosclerosis 239 (2015) 192e202196

3.7. Guteliver axis and the microbiome

The human gastrointestinal tract predominantly consists ofanaerobic bacteria and harbors three dominating bacterial phyla:the gram-positive Firmicutes and Actinobacteria, and the gram-negative Bacteroidetes [61]. In adults, almost 60%e80% of the gutmicrobiota consists of Firmicutes and approximately 20%e40% areBacteroidetes [62]. In healthy host, the gut microbiota plays manyimportant roles by the secretion of bioactive metabolites, protectagainst pathogens by maintaining immunity, perform digestion ofcomplex carbohydrates, synthesize vitamins and store fat. The gutepithelium acts as natural barrier and keeps bacteria, theirbyproducts and other harmful elements at bay due to the presenceof tight junctions and specialised intracellular structures.

The intestinal mucosa serves as a defence barrier that helpsprevent the entrance and the systemic spread of bacteria and en-dotoxins, most of which are lipopolysaccharides (LPS) from the cellwalls of gram-negative bacteria [63]. However, under certain con-ditions, this intestinal barrier fails, resulting in bacterial andendotoxin invasion into the gastrointestinal (GI) tract, after whichthe pathogens reach systemic organs and tissues; this process istermed bacterial translocation [63]. These microbial products exertpro-inflammatory actions mediated through TLRs [63]. TLRs detectthe pathogens enabling the host to regulate immune responsesevident on the surfaces of hepatocytes, KC and hepatic stellate cells[63]. As explained by Baldwin, the activation of TLR4 by LPS triggersan essential intracellular inflammatory cascade, includes stress-activated and mitogen-activated protein kinases, c-Jun-N- termi-nal kinase, p38 and the NFkb pathway [64]. Therefore, TLR4 has aprominent role in promoting inflammation and injury in conditionssuch as alcoholic liver disease and NASH [63]. Increasing data haveunderlined the potentially critical role of the microbiome in thepathogenesis of NAFLD [65]. Different animal mouse modelsshowed that inflammasome-deficiency-associated changes in theconfiguration of the gut microbiota are associated with increasedhepatic steatosis and inflammation through the influx of TLR-4 andTLR-9 agonists into the portal circulation, leading to the increasedexpression of TNF-a, which drives NASH progression [65]. DefectiveNLR related protein (NLPR)-2 and NLPR-6 inflammasomes alteredthe interactions between the host and the gut microbiota and theynegatively regulate NAFLD/NASH progression as well as the manyaspects of metabolic syndrome via modulation of the gut micro-biota [65].

The liver continuously receives blood from the gut through theportal system and therefore, subjected to translocation of bacteria,bacterial products, endotoxin or secreted cytokines present in thegut. Normally, endotoxaemia in the portal circulation is rapidlycleared by the liver's endothelial system, mainly KC [63]. However,in the case of liver disease, the endotoxaemic burden is higher andthe mechanisms to clear these endotoxins are poorer resulting inongoing inflammation. Miele et al. demonstrated that patients withNAFLD have increased gut permeability, possibility related to smallbacterial overgrowth (SIBO) resulting in tight junction disruption,which may impact on NAFLD progression [31]. Most of thesestudies were performed on animal models and hence, moreresearch is needed in this area of field to understand how suchmechanisms apply in the clinical setting.

A recently published prospective, cross-sectional study identi-fied important differences in intestinal microbiota (IM) in patientswith biopsy proven NAFLD (simple steatosis/SS or NASH) andhealthy controls [66]. Patients with NASH had a lower percentage ofBacteroidetes and a higher percentage of Coccoides compared toboth SS and healthy controls with no differences seen in theremaining microorganisms. Differences in levels of faecal Bacter-oidetes still remained after adjusting for BMI and dietary fat intake

between the groups suggesting that IM may play a role in thepathogenesis of NAFLD.

Recent study [67] determined the effect of gut microflora al-terations in a murine model of high-fat-diet (HFD) induced NAFLD.The study showed that when HFD-fed mice were treated withantibiotics, there was a significant increase in conjugated bile acidmetabolites, which inhibited intestinal farnesoid X receptor (FXR)signalling and as a result, there was a reduction of hepatic TGaccumulation. This was noted to be due to modulation of gutmicroflora by reduction of ceramide levels in the serum and ileumresulting in down regulation of hepatic SREPBP1C and decreasedDNL. This particular study suggested that there might be a potentialtherapeutic target for NAFLD treatment by inhibiting intestinal FXR.

3.8. NAFLD and cardiovascular risk

The pathophysiological mechanisms that link NAFLD with cor-onary heart disease (CHD), myocardial dysfunction/hypertrophy,aortic valve sclerosis (AVS) and cardiac arrhythmias are incom-pletely understood [68]. The complex interactions including IR andvisceral obesity make it extremely difficult to dissect out the precisecausal relationships responsible for the increased risk of CHD andother cardiac and arrhythmic complications observed in patientswith NAFLD [68]. The concept so far was that in patients with NASH,there was an increase in systemic and hepatic IR which in turncaused the accumulation of atherogenic dyslipidameia, charac-terised by high TG, low HDL and high VLDL. In NASH, there seemedto be increased production of many pro-inflammatory markerssuch as uric acid and C-reactive protein (CRP) [69], IL-6, TNF-a aswell as pro fibrogenic markers such as tumor growth factor-b,endothelin 1 and insulin like growth factor-1 which can lead toCHD [70]. A observational study showed that high CRP levels arehigher in biopsy proven NASH compared to SS although CRP levelsdo not reflect the degree between inflammation and fibrosis [71].

Epidemiological studies performed in United States and Japanshowed that NAFLD is associated with increased risk of cardiovas-cular disease (CVD) and is a predictor of CVD independent of thepresence of other metabolic syndrome risk factors such as hyper-tension, diabetes, dyslipidaemia, obesity and IR [21,22]. The RISC(Relationship between Insulin Sensitivity and Cardiovascular dis-ease) study showed that patients with NAFLD had an increased 10year coronary heart disease risk score even when only consideringthose at perceived low risk patients (i.e. without diabetes or hy-pertension) [23]. The study also showed that subjects with NAFLDaremore prone to early carotid atherosclerosis in the absence of co-existing metabolic syndrome risk factors [23].

Recent phase II trials assessing coronary atherosclerotic plaquein patients with NAFLD have showed that such patients are morelikely to have advanced high risk coronary plaque, independent oftraditional cardiovascular risk factors as compared with patientswithout NAFLD [72]. In a recent large cohort study of biopsy provenNAFLD such patients had an increased risk of death from cardio-vascular and liver related causes [73]. Theworst prognosis was seenin patients with stage 3 or 4 fibrosis at baseline, although surpris-ingly, in patients with high NAS score (between grade 5 and 8) withno fibrosis did not have increased mortality rate compared to thereference population [73]. This observation requires furtherconfirmation in larger cohorts to exclude the possibility of a type 2error. In another study of patients with a Framingham cardiovas-cular risk score �20%, the presence of advanced fibrosis was pre-dictive of cardiovascular events [19]. A recent study suggests thatprogression of NAFLD and risk of cardiovascular disease may not bea shared pathway since the carriers of gene TM6SF2 E167K varianthave lower serum VLDL and lower risk of cardiovascular diseasecompared to non-carriers. However, they havemore advanced fatty

Table 1Summary of treatments studied for patients with Non-alcoholic fatty liver disease(NAFLD).

Lifestyle interventions Reduction in caloric intakeExercise

Weight loss therapies OrlistatBariatric surgery

Insulin sensitisers MetforminThiazolidinedioneIncretin based therapies (Liraglutide, Exenatide)DDP4 inhibitors (Sitagliptin)

Hypertension Angiotensin II receptor blockersDyslipidaemia Statins

FibratesEzetemibe

Antioxidants/cytoprotective therapies

Vitamin EPentoxyphyllineBetaineUrsodeoxycholic acid (UDCA)

Others ProbioticsPolyunsaturated fatty acids(Both drugs have shown little beneficial effects inpatients with NAFLDa

a Sanyal, A.J., et al., No significant effects of ethyl-eicosapentanoic acid on histo-logic features of nonalcoholic steatohepatitis in a phase 2 trial. Gastroenterology,

N.N. Than, P.N. Newsome / Atherosclerosis 239 (2015) 192e202 197

liver disease with severe steatosis, necro-inflammation, ballooningand fibrosis [74].

Previous studies had shown that carotid-artery intimal medialthickness is related to the severity of NAFLD histology, indepen-dently of other cardio metabolic risk factors [75,76]. A retrospectivestudy showed that NAFLD, assessed by either CT or ultrasonogra-phy, was significantly associated with increased coronary arterycalcification (CAC) score (i.e., CAC score > 100), independently oftraditional CVD risk factors [77]. The concurrent finding ofabdominal obesity which is significantly correlated with left ven-tricular dysfunction also increases all-cause mortality [14]. Thehepatic production of PAI-1 is also elevated in NAFLD patients,which is a known risk factor for CVD [2].

A range of other pro-inflammatory molecules which aresecreted by the liver (uric acid, CRP and homocysteine concentra-tion) in response to oxidative stress may further impact on theassociated cardiovascular risk in NAFLD [11]. Increased CRP pro-motes inflammation and accelerates atherosclerosis by increasingexpression of PAI-1 and adhesion molecules in endothelial cells,inhibiting nitric oxide formation and increasing LDL uptake intomacrophages [5]. IL-6 and TNFa are the major stimuli responsiblefor increased hepatic production of CRP, fibrinogen and otheracute-phase proteins [2,10]. Uric acid, as result of increase in urea,was regulated by decrease in CPS 1 (a ligase enzyme located in themitochondria) and it has been shown to erect pro-inflammatoryand pro-oxidant effect both on adipose tissue and vascularsmooth muscles [78,79]. Uric acid may also contribute to IR byinducing local adipose tissue inflammation, which may causereduction in production of adiponectin [80].

Dyslipidaemia when associated with NAFLD tend to be pro-atherogenic in nature [81]. The characteristic findings areincreased plasma concentrations of VLDL, TG and decreased HDLcholesterol [7], is found in up to 80% of cases of NAFLD [13]. Lipo-protein is important part of underlying mechanistic action in dys-lipidaemia associated with NAFLD. Lipoprotein helped withdelivery of cholesterol and TG from the liver and intestine tomuscleand fat tissue and that action is mediated either by chylomicron andVLDL particles that contain apolipoprotein (apo) B48 and apoB100or in the case of cholesterol, it is indirectly by conversion of inter-mediate density lipoproteins (IDL) to LDL. The second function oflipoprotein is the transport of excess cholesterol from extra-hepatictissue to the liver for the elimination via bile, which is mediatedprimarily by HDL particles. Both functions are altered in NAFLD.Individuals with NAFLD have reduced hepatic insulin sensitivityand glucose clearance despite having high circulating insulin levelswhich can contribute to dyslipidemia [13,82]. Apolipoproteins areproteins that transport lipids in blood circulation. A recent studyshowed that Apo B/Apo A1 ratio are associated with the prevalenceof NAFLD and was independent of obesity and other metaboliccomponents and therefore, the ratio might be useful as a predictivemarker for cardiovascular risk in NAFLD [83].

IR has been regarded as one of the factors responsible foratherogenic dyslipidaemia, as it increases release of non-esterifiedfatty acids (NEFA) or FFAwhich stimulates hepatic TG output [7]. IRand increased lipid availability within the liver also inhibit apoli-poprotein B degradation, stabilising the formation of more VLDLwhich contains highest amount of TG [7,23]. Individuals with type 2diabetes and NAFLD have increased TG accumulation independentof their BMI [13], resulting in cytotoxicity due to increased ROSgeneration and mitochondrial dysfunction [84]. Subjects withNAFLD, despite high circulating insulin levels, have reduced hepaticinsulin sensitivity, poor postprandial glucose clearance, increasedFFA and TG concentrations. The treatment of dyslipidaemia plays acritical role in the overall management of these patients.

3.9. Diagnosis

NAFLD is commonly asymptomatic disease and often identifiedincidentally [85]. Clinicians should consider a diagnosis of NAFLD inpatients with abnormal liver tests and the presence of one or moremetabolic risk factors; indeed the likelihood of NAFLD increasesproportionately with the number of metabolic syndrome factorspresent [86]. Whilst serum aspartate transaminase (AST) andalanine transaminase (ALT) levels may be abnormal in the setting ofNAFLD, they are often not, evenwhen using more stringent cut-offsfor the upper limit of normal [87]. Several studies had demon-strated that ALT is a poor marker to predict advance fibrosis inNAFLD patients [88]. Ultrasound scanwill show an echobright liverin patients with at least 30% steatosis with a sensitivity of 64% andsensitivity of 85% rising to 90e100% [6]. Other modalities of iden-tifying hepatic steatosis include the controlled attenuationparameter (CAP software on the Fibroscan system [89,90] andmagnetic resonance spectroscopy (MRS) which are left to be moresensitive and can also provide a quantitative assessment of statuses[91,92]. For the non-invasive assessment of fibrosis in NAFLD, thereare several scoring systems described which have broadly similarability to identify those patients with little or significant liverfibrosis [93]. Variables common to many of them include the ratioof AST: ALT, age, BMI and impaired glucose tolerance. Other non-invasive tests for determination of liver fibrosis include serumEuropean Liver Fibrosis (ELF) panel (combination of 3 serummarkers: Hyaluronic acid, Pro-collagen III amino terminal peptideand Tissue inhibitor of metalloproteinase 1) and Fibro-test whichhave been reviewed in details in the literature [94]. Transientelastography such as the Fibroscan has been validated in NAFLD andis often used as an adjunctive non-invasive test to assess liverfibrosis in the outpatient setting. Currently, liver biopsy still rep-resents the gold standard for the diagnosis and staging of NAFLD/NASH and is required as the primary end-point for later stageclinical trials [7,95].

4. Treatment

The summary of treatment in patients with NAFLD are docu-mented in Table 1.

2014. 147(2): p. 377e84.e1.

N.N. Than, P.N. Newsome / Atherosclerosis 239 (2015) 192e202198

4.1. Lifestyle management

Lifestyle interventionwith diet and exercise is still the mainstayin the management of patient with NAFLD. Although lifestyleintervention when achieved is effective, it can be difficult toimplement, and thus the aim should be gradual but consistentweight loss over 6e12 months [96]. Commonly individuals arerecommended to restrict caloric intake by approximately500e1000 kcal/day in conjunction with regular interactions with adietician [97,98]. Simple carbohydrates in the diet, in particularfructose, have been linked to NAFLD. Carbohydrate consumptionaffects glucose homeostasis and free fatty acids metabolism in theliver and hence carbohydrate-restricted diet also has been studiedpreviously [99]. Several studies indicate that a reduction of be-tween 7 and 10% of body weight is associated with a reduction ininflammation in the setting of NAFLD, and thus is set as a target[100]. Whilst exercise in isolation has not been proven to beeffective, as part of dietary changes moderate intensity exercisesuch as brisk walking of 30e45 min per day can improvebiochemical and histological aspects of NAFLD [100]. Recentretrospective data from 169 obese, middle-aged men who wereenrolled in a 12-week weight reduction program through lifestylemodification consisting of dietary restriction plus aerobic exerciseshowed that moderate to vigorous physical activity(MVPA) � 250 min per week led to significant decrease in VATseverity, lipid peroxidation and a significant increase in adiponectinlevels compared to those who exercise less [101].

4.2. Weight loss therapy using pharmacological agents

Orlistat produces moderate weight loss by reducing the ab-sorption of fat by 30% by inhibition of gastric and pancreatic li-pases [96]. However, recent studies indicate the effects of Orlistatin patients with NAFLD are modest and restricted only to thosepatients that achieved 9% weight loss with lifestyle intervention[102].

Sibutramine is a serotonin and noradrenaline reuptake inhibi-tor that increases satiety and can lead to modest weight loss[103,104] and in a 6 month treatment study of obese patients withNASH, it reduced body weight by 10%, improved insulin resistanceand decreased transaminase levels [105]. However, sibutramineuse was suspended due to its risk of cardiac related mortality[106].

Rimonabant, a cannabinoid receptor antagonist, was a prom-ising agent that induced significant weight loss in association withimprovements in insulin resistance and serum lipid and adipo-nectin levels, however it has been taken off the market in 2008 dueto severe psychiatric side effects [107e110].

4.3. Bariatric surgery

Bariatric surgery, as indicated for weight loss, has been shown toreduce most of the histological features of NAFLD in several studies[111] as well as obesity associated T2DM and IR. NAFLD is not, asyet, a standalone indication for bariatric surgery although it iscommonly factored into the decision making in the management ofoverweight patients. Bariatric surgery increases insulin sensitivityin the liver, muscle and fat due to weight loss and hence, willimprove overall metabolic health. Weight loss after bariatric sur-gery increases insulin sensitivity in liver, muscle and fat. Thus, it iswell recognized that bariatric surgery can improve overall meta-bolic health, but it remains unclear how it improves insulinsensitivity.

5. Pharmacological management

5.1. Insulin sensitizers

Given the important of IR in the pathogenesis of NAFLD, insulinsensitizers such as metformin and thiazolidinedione (TZDs), havebeen extensively studied in the treatment of NAFLD [4].

5.1.1. MetforminMetformin is a biguanide insulin sensitizer widely used in

treatment of type 2 diabetes, whose major action is mediatedthrough activation of AMPK [112]. This improves peripheral glucoseintake, reduces hepatic gluconeogenesis and lipogenesis and alsoincreases beta oxidation of FFA [113]. However, clinical studies inNAFLD have not demonstrated a consistent benefit in patients withNAFLD and thus its use is reserved for the management of patientswith NAFLD and concomitant type 2 diabetes [114,115].

5.1.2. Thiazolidinediones (TZDs)TZDs are agonists for PPAR-gamma and their actions in NAFLD

are probably indirect and in large part, relate to their effects onadipose tissue. Moreover, TZDs can also upregulate adiponectinand/or sitrulin in hepatocytes which modulate central hepaticmetabolic regulators such as AMPK, Foxo1 (Forkhead box O1), LKB1(Liver Kinase B1), NAD (Nicotinamide adenine dinucleotide), NADH(Reduced nicotinamide adenine dinucleotide), PGC-1a and PPAR-1a. These result in increased fatty acid oxidation and subsequentlylead to reduction in hepatic fat accumulation [4], and indeed thesetherapeutic agents have been demonstrated to have potent benefitsin pre-clinical models of NAFLD. However, clinical concerns remainabout their effect on reducing renal excretion of sodium and in-testinal ion transport, which results in raised plasma volume/fluidretention [116] and exacerbations of pre-existing heart failure [117].TZDs also reduce hepatic steatosis and improve insulin sensitivityin muscle and liver by enhancing adipocyte insulin sensitivity andshifting ectopic lipid from muscle and the liver to subcutaneousadipose tissue [38].

A study looking at a regimen of hypo-caloric diet (500 kcal perday) plus pioglitazone versus hypo-caloric diet and placebo in pa-tients with T2DM and NAFLD showed that pioglitazone led tometabolic and histologic improvement in subjects with non-alcoholic steatohepatitis, although the sample size was small[118]. Whilst pioglitazone did not meet the primary end point in alarge, multi-centre randomised controlled trial (PIVENS - Pioglita-zone or Vitamin E for NASH study), it did improve insulin sensitivityand clearance of steatohepatitis compared to placebo [119]. Indeed,meta-analysis of clinical trials with Pioglitazone does demonstrateclinical benefit in NAFLD although the associated weight gain, pe-ripheral oedema and other cautions have limited its use in clinicalpractice [115]. Overall, the effects of TZDs on NAFLD are indirect andperhaps predominantly a result of reduction of fat deposition in theliver.

5.1.3. Incretin based therapiesIncretins, a gut derived neuro endocrine hormones, are pro-

duced by the intestinal tract in response to food ingestion and theystimulate glucose dependent insulin release, decrease glucagonrelease [120] and prolong gastric emptying [121]. Consequentlythese drugs are licensed for the management of type 2 diabeteswhere they improve glycaemic control, increased insulin sensitivityand cause significant weight loss. Circulating glucagon like peptide-1 (GLP-1) has a half-life of about 1e2 min due to rapid degradationby the enzyme dipeptdyl peptidase-4 (DPP-4) and thus, GLP-1 re-ceptor agonists with increased DPP-4 resistance (e.g., Liraglutide,Exenatide) [122e125]and DPP-4 inhibitors (e.g., Sitagliptin,

N.N. Than, P.N. Newsome / Atherosclerosis 239 (2015) 192e202 199

Saxagliptin) [126,127] have been developed as therapies. Theseagents offer promise in NAFLD and on the basis of encouragingopen-label studies controlled trials in NAFLD are underway [128].

5.1.4. Lipid lowering drugsStatins, fibrates and omega-3 polyunsaturated fatty acids

(PUFAs) are used to manage dyslipidaemia [4], which is commonlyfound in patients with NAFLD. There are no studies to supportdirect benefit of those drugs in NAFLD, but theywill be important inreducing the concomitant cardiovascular risk.

5.2. Hypertension

5.2.1. Angiotensin II receptor blockers (AIIRB)AIIRB inhibit the proliferation of stellate cells, reducing inflam-

mation and fibrosis, and although early clinical studies with Los-artan and Telmisartan have suggested improvements intransaminases and histology in patients with NASH, larger trialshave not yet been performed [105].

5.3. Anti-oxidants and cytoprotective therapies

5.3.1. Vitamin EVitamin E, a fat soluble vitamin, has an anti-oxidant property

and two recently published, large, randomized controlled trials,PIVENS and TONIC (Treatment of non-alcoholic fatty liver disease inchildren), assessed its effect on adult and paediatric NAFLD pop-ulations, respectively [119,129]. Vitamin E treatment met the pri-mary end-point in both trials [119,129] however, there remainconcerns about the long term safety at the high dose used (400 IU/day) since it has been shown to increase all-cause mortality.

5.3.2. BetaineBetaine is a naturally occurring metabolite of choline and has

been shown to increase S-adenosylmethionine levels and reduceoxidative stress [130]. Unfortunately, when compared with placeboin a randomized controlled trial, Betaine failed to improve steatosisor other histological outcomes [131].

5.3.3. Ursodeoxycholic acid (UDCA)UDCA is a hydrophilic bile acid with cytoprotective and anti-

oxidant properties [105]. Twowell-designed RCTs examining UDCAfailed to show any significant histological improvements withUDCA alone in patients with biopsy-proven NASH [132,133].

5.3.4. PentoxyphyllinePentoxyphylline is a tumour necrosis factor alpha (TNFa) in-

hibitor which has been evaluated in NASH for its anti-inflammatoryproperties [105]. A meta-analysis of randomized, double-blinded,placebo-controlled trials showed that Pentoxyphylline couldreduce the aminotransferase activities and improve the histologicalparameters in NAFLD patients [134], but the results needs to beconfirmed in larger studies.

5.4. Microbiomes

Probiotics have attracted interest given their possible ability toalter the composition of Intestinal microbiota (IM) and the subse-quent interaction with the immune system and gut epithelium[135]. Commercial probiotics commonly include lactic acid bacteria(Clostridium/Bacillus gram-positive bacteria and Actinomycetesgram-positive Bifidobacteria) and spore-forming bacteria (Clos-tridium-Bacillus gram-positive bacteria) [135]. Several probioticstain, mainly lactobacillus species, produce lactic acid and beingacid resistant, they persist in the stomach longer than other

bacteria which protect them against gastric non immunologicalbarriers such as acidity and mucosal barriers.

A recent meta-analysis of 134 NAFLD/NASH patients from 14randomised controlled trials showed probiotic therapy significantlydecreased liver aminotransferases, total-cholesterol level, TNF-aand improve insulin resistance in NAFLD patients but did not showany changes in BMI, glucose or LDL levels [136]. Of the four RCTsincluded in this meta-analysis, the studied probiotics includedlactobacillus, bifidobacterium and streptococcus [136]. Furtherdata, including histological studies are needed with probiotics. Arecent double blind randomised controlled trial looking at effect ofVSL#3 in obese children with biopsy-proven NAFLD [137]concluded there was an improvement in the USS appearances ofthe liver although there was no difference in TG levels, HOMAscores and serum ALT levels. A study which tried to evaluate theeffects of Bifidobacterium longum with fructo-oligosaccharides(Fos) with lifestyle modification vs lifestyle modification alone inthe treatment of NASH in adults [138] showed that the treatedgroup demonstrated a significant improvement in serum ASTlevels, reduction of TNF-a, CRP, HOMA-IR, serum endotoxin, stea-tosis, and the NASH activity index [138].

5.5. Liver transplantation

For those patients with end-stage liver disease due to NAFLD,liver transplantation (LT) is the only definitive treatment [139].Outcomes post-transplant for such patient are comparable to thatseen with other indications, although this likely reflects stringentcase selection. Recurrence of NASH is common post-transplant dueto the presence of existing metabolic risk factors as well as the useof immunosuppression such as corticosteroid, although it rarelyresults in allograft loss.

6. Conclusion

The incidence of NAFLD is increasing rapidly due to rising ratesof obesity in both children and adults, and can vary from simplesteatosis to inflammation all the way through to fibrosis andcirrhosis. Hepatic steatosis is recognised to be the consequence of acomplex interplay between diet, environment and the liver andadipose tissues, although a full understanding of its pathogenesishas not yet been elucidated [4]. The presence of fibrosis is impor-tant in the prognosis of the disease since it confers an increasedoverall mortality from cardiovascular, liver, and also malignancy.Whilst there are no licensed drug therapies at present, there aremany late phase clinical trials in progress so this will likely changein the near future. Given that cardiovascular disease accounts forthe majority of deaths in NAFLD, the challenge will be managingthese two components simultaneously.

Conflict of interest

None.

Disclosure

None.

Funding

N.N.Than and P.N.Newsome are funded by NIHR (NationalInstitute for Health Research).

N.N. Than, P.N. Newsome / Atherosclerosis 239 (2015) 192e202200

References

[1] N. Chalasani, et al., The diagnosis and management of non-alcoholic fattyliver disease: practice guideline by the American Association for the Study ofLiver Diseases, American College of Gastroenterology, and the AmericanGastroenterological Association, Hepatology 55 (6) (2012) 2005e2023.

[2] J.K. Dowman, J.W. Tomlinson, P.N. Newsome, Pathogenesis of non-alcoholicfatty liver disease, Qjm 103 (2) (2010) 71e83.

[3] P. Angulo, Long-term mortality in nonalcoholic fatty liver disease: is liverhistology of any prognostic significance? Hepatology 51 (2) (2010) 373e375.

[4] E. Chang, C.Y. Park, S.W. Park, Role of thiazolidinediones, insulin sensitizers,in non-alcoholic fatty liver disease, J. Diabetes Investig. 4 (6) (2013)517e524.

[5] M.H. Ahmed, S. Barakat, A.O. Almobarak, Nonalcoholic fatty liver disease andcardiovascular disease: has the time come for cardiologists to be hepatolo-gists? J. Obes. 2012 (2012) 483135.

[6] A. Brea, J. Puzo, Non-alcoholic fatty liver disease and cardiovascular risk, Int.J. Cardiol. 167 (4) (2013) 1109e1117.

[7] G. Targher, F. Marra, G. Marchesini, Increased risk of cardiovascular diseasein non-alcoholic fatty liver disease: causal effect or epiphenomenon? Dia-betologia 51 (11) (2008) 1947e1953.

[8] M.J. Armstrong, et al., Extrahepatic complications of nonalcoholic fatty liverdisease, Hepatology 59 (3) (2014) 1174e1197.

[9] G. Vernon, A. Baranova, Z.M. Younossi, Systematic review: the epidemiologyand natural history of non-alcoholic fatty liver disease and non-alcoholicsteatohepatitis in adults, Aliment. Pharmacol. Ther. 34 (3) (2011) 274e285.

[10] I.R. Wanless, J.S. Lentz, Fatty liver hepatitis (steatohepatitis) and obesity: anautopsy study with analysis of risk factors, Hepatology 12 (5) (1990)1106e1110.

[11] R.M. Williamson, et al., Prevalence of and risk factors for hepatic steatosisand nonalcoholic fatty liver disease in people with type 2 diabetes: theEdinburgh type 2 diabetes Study, Diabetes Care 34 (5) (2011) 1139e1144.

[12] M. Machado, P. Marques-Vidal, H. Cortez-Pinto, Hepatic histology in obesepatients undergoing bariatric surgery, J. Hepatol. 45 (4) (2006) 600e606.

[13] M. Gaggini, et al., Non-alcoholic fatty liver disease (NAFLD) and its connec-tion with insulin resistance, dyslipidemia, atherosclerosis and coronary heartdisease, Nutrients 5 (5) (2013) 1544e1560.

[14] B.M. Attar, D.H. Van Thiel, Current concepts and management approaches innonalcoholic fatty liver disease, Sci. World J. 2013 (2013) 481893.

[15] J.D. Browning, et al., Prevalence of hepatic steatosis in an urban population inthe United States: Impact of ethnicity, Hepatology 40 (6) (2004) 1387e1395.

[16] K. Bambha, et al., Ethnicity and nonalcoholic fatty liver disease, Hepatology55 (3) (2012) 769e780.

[17] M. Ekstedt, et al., Long-term follow-up of patients with NAFLD and elevatedliver enzymes, Hepatology 44 (4) (2006) 865e873.

[18] C.K. Argo, A.M. Al-Osaimi, S.H. Caldwell, Systematic review of risk factors forfibrosis progression in non-alcoholic steatohepatitis, J. Hepatol. 51 (2) (2009)371e379. Northup PG.

[19] H. Perazzo, et al., Prognostic value of liver fibrosis and steatosis biomarkersin type-2 diabetes and dyslipidaemia, Aliment. Pharmacol. Ther. 40 (9)(2014) 1081e1093.

[20] A. Hamabe, et al., Impact of cigarette smoking on onset of nonalcoholic fattyliver disease over a 10-year period, J. Gastroenterol. 46 (6) (2011) 769e778.

[21] W. Dunn, R. Xu, J.B. Schwimmer, Modest wine drinking and decreasedprevalence of suspected nonalcoholic fatty liver disease, Hepatology 47 (6)(2008) 1947e1954.

[22] M.N. Stankovi�c, et al., Time-dependent changes and association betweenliver free fatty acids, serum lipid profile and histological features in micemodel of nonalcoholic fatty liver disease, Arch. Med. Res. 45 (2) (2014)116e124.

[23] M.A. Edens, F. Kuipers, R.P. Stolk, Non-alcoholic fatty liver disease is asso-ciated with cardiovascular disease risk markers, Obes. Rev. 10 (4) (2009)412e419.

[24] G. Targher, C.P. Day, E. Bonora, Risk of cardiovascular disease in patients withnonalcoholic fatty liver disease, N. Engl. J. Med. 363 (14) (2010) 1341e1350.

[25] S. Treeprasertsuk, F. Lopez-Jimenez, K.D. Lindor, Nonalcoholic fatty liverdisease and the coronary artery disease, Dig. Dis. Sci. 56 (1) (2011) 35e45.

[26] C.P. Day, O.F. James, Steatohepatitis: a tale of two “hits”? Gastroenterology114 (4) (1998) 842e845.

[27] J. Hirosumi, et al., A central role for JNK in obesity and insulin resistance,Nature 420 (6913) (2002) 333e336.

[28] D. Cai, et al., Local and systemic insulin resistance resulting from hepaticactivation of IKK-beta and NF-kappaB, Nat. Med. 11 (2) (2005) 183e190.

[29] M. Del Ben, et al., Non-alcoholic fatty liver disease and cardiovascular dis-ease: epidemiological, clinical and pathophysiological evidences, InternEmerg. Med. 7 (Suppl. 3) (2012) S291eS296.

[30] G. Marchesini, et al., Association of nonalcoholic fatty liver disease with in-sulin resistance, Am. J. Med. 107 (5) (1999) 450e455.

[31] S.B. Biddinger, et al., Hepatic insulin resistance is sufficient to produce dys-lipidemia and susceptibility to atherosclerosis, Cell. Metab. 7 (2) (2008)125e134.

[32] M.S. Brown, J.L. Goldstein, Selective versus total insulin resistance: a path-ogenic paradox, Cell. Metab. 7 (2) (2008) 95e96.

[33] E. Rodriguez-Suarez, et al., Non-alcoholic fatty liver disease proteomics,

Proteomics Clin. Appl. 4 (4) (2010) 362e371.[34] J.W. Lim, J. Dillon, M. Miller, Proteomic and genomic studies of non-alcoholic

fatty liver diseaseeclues in the pathogenesis, World J. Gastroenterol. 20 (26)(2014) 8325e8340.

[35] P. Ferre, F. Foufelle, Hepatic steatosis: a role for de novo lipogenesis and thetranscription factor SREBP-1c, Diabetes Obes. Metab. 12 (Suppl. 2) (2010)83e92.

[36] L. Eissing, et al., De novo lipogenesis in human fat and liver is linked toChREBP-beta and metabolic health, Nat. Commun. 4 (2013) 1528.

[37] M. Wu, et al., Antidiabetic and antisteatotic effects of the selective fatty acidsynthase (FAS) inhibitor platensimycin in mouse models of diabetes, Proc.Natl. Acad. Sci. U. S. A. 108 (13) (2011) 5378e5383.

[38] G.I. Shulman, Ectopic fat in insulin resistance, dyslipidemia, and car-diometabolic disease, N. Engl. J. Med. 371 (12) (2014) 1131e1141.

[39] P. Loria, A. Lonardo, G. Targher, Is liver fat detrimental to vessels?: in-tersections in the pathogenesis of NAFLD and atherosclerosis, Clin. Sci. Lond.115 (1) (2008) 1e12.

[40] J.E. Schaffer, Lipotoxicity: when tissues overeat, Curr. Opin. Lipidol. 14 (3)(2003) 281e287.

[41] J.K. Kim, et al., Tissue-specific overexpression of lipoprotein lipase causestissue-specific insulin resistance, Proc. Natl. Acad. Sci. U. S. A. 98 (13) (2001)7522e7527.

[42] L.D. Ferreira, et al., Overexpressing human lipoprotein lipase in mouseskeletal muscle is associated with insulin resistance, Diabetes 50 (5) (2001)1064e1068.

[43] L. Bouwens, et al., Quantitation, tissue distribution and proliferation kineticsof Kupffer cells in normal rat liver, Hepatology 6 (4) (1986) 718e722.

[44] P.J. MacPhee, E.E. Schmidt, A.C. Groom, Evidence for Kupffer cell migrationalong liver sinusoids, from high-resolution in vivo microscopy, Am. J. Physiol.263 (1 Pt 1) (1992) G17eG23.

[45] C.A. Rivera, et al., Toll-like receptor-4 signaling and Kupffer cells play pivotalroles in the pathogenesis of non-alcoholic steatohepatitis, J. Hepatol. 47 (4)(2007) 571e579.

[46] R. Stienstra, et al., Kupffer cells promote hepatic steatosis via interleukin-1beta-dependent suppression of peroxisome proliferator-activated recep-tor alpha activity, Hepatology 51 (2) (2010) 511e522.

[47] G. Baffy, Kupffer cells in non-alcoholic fatty liver disease: the emerging view,J. Hepatol. 51 (1) (2009) 212e223.

[48] J.W. Park, et al., Predictors reflecting the pathological severity of non-alcoholic fatty liver disease: comprehensive study of clinical and immuno-histochemical findings in younger Asian patients, J. Gastroenterol. Hepatol.22 (4) (2007) 491e497.

[49] R.S. Ahima, J.S. Flier, Adipose tissue as an endocrine organ, Trends Endo-crinol. Metab. 11 (8) (2000) 327e332.

[50] A.E. Feldstein, et al., Hepatocyte apoptosis and fas expression are prominentfeatures of human nonalcoholic steatohepatitis, Gastroenterology 125 (2)(2003) 437e443.

[51] A.E. Feldstein, et al., Diet associated hepatic steatosis sensitizes to Fasmediated liver injury in mice, J. Hepatol. 39 (6) (2003) 978e983.

[52] H. Malhi, G.J. Gores, Cellular and molecular mechanisms of liver injury,Gastroenterology 134 (6) (2008) 1641e1654.

[53] P. Dongiovanni, et al., PNPLA3 I148M polymorphism and progressive liverdisease, World J. Gastroenterol. 19 (41) (2013) 6969e6978.

[54] E.K. Speliotes, et al., Genome-wide association analysis identifies variantsassociated with nonalcoholic fatty liver disease that have distinct effects onmetabolic traits, PLoS Genet. 7 (3) (2011) e1001324.

[55] S. Romeo, et al., Genetic variation in PNPLA3 confers susceptibility tononalcoholic fatty liver disease, Nat. Genet. 40 (12) (2008) 1461e1465.

[56] A. Kotronen, et al., A common variant in PNPLA3, which encodes adiponutrin,is associated with liver fat content in humans, Diabetologia 52 (6) (2009)1056e1060.

[57] S. Sookoian, C.J. Pirola, Meta-analysis of the influence of I148M variant ofpatatin-like phospholipase domain containing 3 gene (PNPLA3) on the sus-ceptibility and histological severity of nonalcoholic fatty liver disease, Hep-atology 53 (6) (2011) 1883e1894.

[58] N. Chalasani, et al., Genome-wide association study identifies variantsassociated with histologic features of nonalcoholic fatty liver disease,Gastroenterology 139 (5) (2010) 1567e1576, 1576.e1-6.

[59] N. Santoro, et al., Variant in the glucokinase regulatory protein (GCKR) geneis associated with fatty liver in obese children and adolescents, Hepatology55 (3) (2012) 781e789.

[60] D. Fu, et al., Genetic polymorphism of glucokinase on the risk of type 2diabetes and impaired glucose regulation: evidence based on 298,468 sub-jects, PLoS One 8 (2) (2013) e55727.

[61] S.M. Ferolla, et al., The role of intestinal bacteria overgrowth in obesity-related nonalcoholic fatty liver disease, Nutrients 6 (12) (2014) 5583e5599.

[62] A. Duseja, Y.K. Chawla, Obesity and NAFLD: the role of bacteria and micro-biota, Clin. Liver Dis. 18 (1) (2014) 59e71.

[63] T. Eslamparast, et al., Probiotics and nonalcoholic fatty liver disease, MiddleEast J. Dig. Dis. 5 (3) (2013) 129e136.

[64] A.S. Baldwin Jr., The NF-kappa B and I kappa B proteins: new discoveries andinsights, Annu Rev. Immunol. 14 (1996) 649e683.

[65] J. Henao-Mejia, et al., Inflammasome-mediated dysbiosis regulates pro-gression of NAFLD and obesity, Nature 482 (7384) (2012) 179e185.

[66] M. Mouzaki, et al., Intestinal microbiota in patients with nonalcoholic fatty

N.N. Than, P.N. Newsome / Atherosclerosis 239 (2015) 192e202 201

liver disease, Hepatology 58 (1) (2013) 120e127.[67] C. Jiang, et al., Intestinal farnesoid X receptor signaling promotes nonalco-

holic fatty liver disease, J. Clin. Invest (2014).[68] S. Ballestri, et al., Risk of cardiovascular, cardiac and arrhythmic complica-

tions in patients with non-alcoholic fatty liver disease, World J. Gastro-enterol. 20 (7) (2014) 1724e1745.

[69] C.E. Ndumele, et al., Hepatic steatosis, obesity, and the metabolic syndromeare independently and additively associated with increased systemicinflammation, Arterioscler. Thromb. Vasc. Biol. 31 (8) (2011) 1927e1932.

[70] A. Verrijken, et al., Prothrombotic factors in histologically proven NAFLD andNASH, Hepatology 59 (2013) 121e129.

[71] M. Yoneda, et al., High-sensitivity C-reactive protein is an independentclinical feature of nonalcoholic steatohepatitis (NASH) and also of theseverity of fibrosis in NASH, J. Gastroenterol. 42 (7) (2007) 573e582.

[72] S.B. Puchner, et al., High-risk coronary plaque at coronary CT angiography isassociated with nonalcoholic fatty liver disease, independent of coronaryplaque and Stenosis Burden: results from the ROMICAT II trial, Radiology(2014) 140933.

[73] M. Ekstedt, et al., Fibrosis stage is the strongest predictor for disease-specificmortality in NAFLD after up to 33 years of follow-up, Hepatology (2014),http://dx.doi.org/10.1002/hep.27368 (Epub ahead of print).

[74] P. Dongiovanni, et al., TM6SF2 gene variant disentangles nonalcoholic stea-tohepatitis from cardiovascular disease, Hepatology (2014), http://dx.doi.org/10.1002/hep.27490 (Epub ahead of print).

[75] G. Targher, et al., Relations between carotid artery wall thickness and liverhistology in subjects with nonalcoholic fatty liver disease, Diabetes Care 29(6) (2006) 1325e1330.

[76] Y. Colak, et al., Assessment of endothelial function in patients with nonal-coholic fatty liver disease, Endocrine 43 (1) (2013) 100e107.

[77] C.H. Chen, et al., Association between nonalcoholic fatty liver disease andcoronary artery calcification, Dig. Dis. Sci. 55 (6) (2010) 1752e1760.

[78] J.C. Sirota, et al., Elevated serum uric acid levels are associated with non-alcoholic fatty liver disease independently of metabolic syndrome featuresin the United States: liver ultrasound data from the National Health andNutrition Examination Survey, Metabolism 62 (3) (2013) 392e399.

[79] T. Nakagawa, et al., SIRT5 Deacetylates carbamoyl phosphate synthetase 1and regulates the urea cycle, Cell 137 (3) (2009) 560e570.

[80] R.J. Johnson, et al., Sugar, uric acid, and the etiology of diabetes and obesity,Diabetes 62 (10) (2013) 3307e3315.

[81] E.K. Speliotes, et al., Fatty liver is associated with dyslipidemia and dysgly-cemia independent of visceral fat: the Framingham Heart Study, Hepatology51 (6) (2010) 1979e1987.

[82] E. Bugianesi, et al., Insulin resistance in non-diabetic patients with non-alcoholic fatty liver disease: sites and mechanisms, Diabetologia 48 (4)(2005) 634e642.

[83] Y.G. Choe, et al., Apolipoprotein B/AI ratio is independently associated withnon-alcoholic fatty liver disease in nondiabetic subjects, J. Gastroenterol.Hepatol. 28 (4) (2013) 678e683.

[84] M. Maurantonio, et al., Treatment of atherogenic liver based on the patho-genesis of nonalcoholic fatty liver disease: a novel approach to reduce car-diovascular risk? Arch. Med. Res. 42 (5) (2011) 337e353.

[85] M.J. Armstrong, et al., Presence and severity of non-alcoholic fatty liverdisease in a large prospective primary care cohort, J. Hepatol. 56 (1) (2012)234e240.

[86] V.W. Wong, et al., Prevalence of non-alcoholic fatty liver disease andadvanced fibrosis in Hong Kong Chinese: a population study using proton-magnetic resonance spectroscopy and transient elastography, Gut 61 (3)(2012) 409e415.

[87] J.K. Dyson, Q.M. Anstee, S. McPherson, Non-alcoholic fatty liver disease: apractical approach to diagnosis and staging, Frontline Gastroenterol. 5 (3)(2014) 211e218.

[88] S. Verma, et al., Predictive value of ALT levels for non-alcoholic steatohe-patitis (NASH) and advanced fibrosis in non-alcoholic fatty liver disease(NAFLD), Liver Int. 33 (9) (2013) 1398e1405.

[89] T. Karlas, et al., Non-invasive assessment of hepatic steatosis in patients withNAFLD using controlled attenuation parameter and 1H-MR spectroscopy,PLoS One 9 (3) (2014) e91987.

[90] V. de Ledinghen, et al., Controlled attenuation parameter (CAP) for thediagnosis of steatosis: a prospective study of 5323 examinations, J. Hepatol.60 (5) (2014) 1026e1031.

[91] R. Banerjee, et al., Multiparametric magnetic resonance for the non-invasivediagnosis of liver disease, J. Hepatol. 60 (1) (2014) 69e77.

[92] H.S. Leitao, et al., MR fat fraction mapping: a simple biomarker for liversteatosis quantification in nonalcoholic fatty liver disease patients, Acad.Radiol. 20 (8) (2013) 957e961.

[93] S. McPherson, et al., Simple non-invasive fibrosis scoring systems can reli-ably exclude advanced fibrosis in patients with non-alcoholic fatty liverdisease, Gut 59 (9) (2010) 1265e1269.

[94] J.K. Dowman, J.W. Tomlinson, P.N. Newsome, Systematic review: the diag-nosis and staging of non-alcoholic fatty liver disease and non-alcoholicsteatohepatitis, Aliment. Pharmacol. Ther. 33 (5) (2011) 525e540.

[95] A.J. Sanyal, et al., Endpoints and clinical trial design for nonalcoholic stea-tohepatitis, Hepatology 54 (1) (2011) 344e353.

[96] Ver�onica Martín-Domínguez, R. G.-C., J. Mendoza-Jim�enez-Ridruejo,Luisa García-Buey, Ricardo Moreno-Otero, Pathogenesis, diagnosis and

treatment of non-alcoholic fatty liver disease, Rev. Esp. Enferm. Dig. 105(2013) 409e420.

[97] K. Shah, et al., Diet and exercise interventions reduce intrahepatic fat contentand improve insulin sensitivity in obese older adults, Obesity (Silver Spring)17 (12) (2009) 2162e2168.

[98] M. Palmer, F. Schaffner, Effect of weight reduction on hepatic abnormalitiesin overweight patients, Gastroenterology 99 (5) (1990) 1408e1413.

[99] W.S. Yancy Jr., et al., A low-carbohydrate, ketogenic diet versus a low-fat dietto treat obesity and hyperlipidemia: a randomized, controlled trial, Ann.Intern Med. 140 (10) (2004) 769e777.

[100] K. Promrat, et al., Randomized controlled trial testing the effects of weightloss on nonalcoholic steatohepatitis, Hepatology 51 (1) (2010) 121e129.

[101] S. Oh, et al., Moderate to vigorous physical activity volume is an importantfactor for managing non-alcoholic fatty liver disease: a retrospective study,Hepatology (2014), http://dx.doi.org/10.1002/hep.27544 (Epub ahead ofprint).

[102] S. Zelber-Sagi, et al., A double-blind randomized placebo-controlled trial oforlistat for the treatment of nonalcoholic fatty liver disease, Clin. Gastro-enterol. Hepatol. 4 (5) (2006) 639e644.

[103] G.A. Bray, et al., A double-blind randomized placebo-controlled trial ofsibutramine, Obes. Res. 4 (3) (1996) 263e270.

[104] W.P. James, et al., Effect of sibutramine on weight maintenance after weightloss: a randomised trial. STORM Study Group. Sibutramine Trial of ObesityReduction and Maintenance, Lancet 356 (9248) (2000) 2119e2125.

[105] J.K. Dowman, et al., Current therapeutic strategies in non-alcoholic fatty liverdisease, Diabetes Obes. Metab. 13 (8) (2011) 692e702.

[106] W.P. James, et al., Effect of sibutramine on cardiovascular outcomes inoverweight and obese subjects, N. Engl. J. Med. 363 (10) (2010) 905e917.

[107] A.J. Scheen, et al., Efficacy and tolerability of rimonabant in overweight orobese patients with type 2 diabetes: a randomised controlled study, Lancet368 (9548) (2006) 1660e1672.

[108] F.X. Pi-Sunyer, et al., Effect of rimonabant, a cannabinoid-1 receptor blocker,on weight and cardiometabolic risk factors in overweight or obese patients:RIO-North America: a randomized controlled trial, JAMA 295 (7) (2006)761e775.

[109] L.F. Van Gaal, et al., Effects of the cannabinoid-1 receptor blocker rimonabanton weight reduction and cardiovascular risk factors in overweight patients:1-year experience from the RIO-Europe study, Lancet 365 (9468) (2005)1389e1397.

[110] E.J. Topol, et al., Rimonabant for prevention of cardiovascular events (CRE-SCENDO): a randomised, multicentre, placebo-controlled trial, Lancet 376(9740) (2010) 517e523.

[111] R.R. Mummadi, et al., Effect of bariatric surgery on nonalcoholic fatty liverdisease: systematic review and meta-analysis, Clin. Gastroenterol. Hepatol. 6(12) (2008) 1396e1402.

[112] G. Zhou, et al., Role of AMP-activated protein kinase in mechanism of met-formin action, J. Clin. Invest 108 (8) (2001) 1167e1174.

[113] K.A. Coughlan, et al., AMPK activation: a therapeutic target for type 2 dia-betes? Diabetes Metab. Syndr. Obes. 7 (2014) 241e253.

[114] Y. Li, et al., Metformin in non-alcoholic fatty liver disease: a systematic re-view and meta-analysis, Biomed. Rep. 1 (1) (2013) 57e64.

[115] G. Musso, et al., Impact of current treatments on liver disease, glucosemetabolism and cardiovascular risk in non-alcoholic fatty liver disease(NAFLD): a systematic review and meta-analysis of randomised trials, Dia-betologia 55 (4) (2012) 885e904.

[116] A. Zanchi, et al., Effects of the peroxisomal proliferator-activated receptor-gamma agonist pioglitazone on renal and hormonal responses to salt inhealthy men, J. Clin. Endocrinol. Metab. 89 (3) (2004) 1140e1145.

[117] S. Singh, Y.K. Loke, C.D. Furberg, Thiazolidinediones and heart failure: ateleo-analysis, Diabetes Care 30 (8) (2007) 2148e2153.

[118] R. Belfort, et al., A placebo-controlled trial of pioglitazone in subjects withnonalcoholic steatohepatitis, N. Engl. J. Med. 355 (22) (2006) 2297e2307.

[119] A.J. Sanyal, et al., Pioglitazone, vitamin E, or placebo for nonalcoholic stea-tohepatitis, N. Engl. J. Med. 362 (18) (2010) 1675e1685.

[120] E. Rask, et al., Impaired incretin response after a mixed meal is associatedwith insulin resistance in nondiabetic men, Diabetes Care 24 (9) (2001)1640e1645.

[121] T. Vilsboll, et al., Both GLP-1 and GIP are insulinotropic at basal and post-prandial glucose levels and contribute nearly equally to the incretin effect ofa meal in healthy subjects, Regul. Pept. 114 (2e3) (2003) 115e121.

[122] K. Blaslov, et al., Incretin based therapies: a novel treatment approach fornon-alcoholic fatty liver disease, World J. Gastroenterol. 20 (23) (2014)7356e7365.

[123] T. Vilsboll, et al., Liraglutide, a long-acting human glucagon-like peptide-1analog, given as monotherapy significantly improves glycemic control andlowers body weight without risk of hypoglycemia in patients with type 2diabetes, Diabetes Care 30 (6) (2007) 1608e1610.

[124] O.G. Kolterman, et al., Synthetic exendin-4 (exenatide) significantly reducespostprandial and fasting plasma glucose in subjects with type 2 diabetes,J. Clin. Endocrinol. Metab. 88 (7) (2003) 3082e3089.

[125] T. Fukuhara, et al., Efficacy and safety of sitagliptin for the treatment ofnonalcoholic fatty liver disease with type 2 diabetes mellitus, Hepatogas-troenterology 61 (130) (2014) 323e328.

[126] P. Aschner, et al., Effect of the dipeptidyl peptidase-4 inhibitor sitagliptin asmonotherapy on glycemic control in patients with type 2 diabetes, Diabetes

N.N. Than, P.N. Newsome / Atherosclerosis 239 (2015) 192e202202

Care 29 (12) (2006) 2632e2637.[127] J. Rosenstock, et al., Effect of saxagliptin monotherapy in treatment-naive

patients with type 2 diabetes, Curr. Med. Res. Opin. 25 (10) (2009)2401e2411.

[128] J.B. Buse, et al., Metabolic effects of two years of exenatide treatment ondiabetes, obesity, and hepatic biomarkers in patients with type 2 diabetes:an interim analysis of data from the open-label, uncontrolled extension ofthree double-blind, placebo-controlled trials, Clin. Ther. 29 (1) (2007)139e153.

[129] J.E. Lavine, et al., Effect of vitamin E or metformin for treatment of nonal-coholic fatty liver disease in children and adolescents: the TONIC random-ized controlled trial, JAMA 305 (16) (2011) 1659e1668.

[130] M.D. Beaton, Current treatment options for nonalcoholic fatty liver diseaseand nonalcoholic steatohepatitis, Can. J. Gastroenterol. 26 (6) (2012)353e357.

[131] M.F. Abdelmalek, et al., Betaine for nonalcoholic fatty liver disease: results ofa randomized placebo-controlled trial, Hepatology 50 (6) (2009) 1818e1826.

[132] U.F. Leuschner, et al., High-dose ursodeoxycholic acid therapy for nonalco-holic steatohepatitis: a double-blind, randomized, placebo-controlled trial,

Hepatology 52 (2) (2010) 472e479.[133] K.D. Lindor, et al., Ursodeoxycholic acid for treatment of nonalcoholic stea-

tohepatitis: results of a randomized trial, Hepatology 39 (3) (2004) 770e778.[134] T. Zeng, et al., Pentoxifylline for the treatment of nonalcoholic fatty liver

disease: a meta-analysis of randomized double-blind, placebo-controlledstudies, Eur. J. Gastroenterol. Hepatol. 26 (6) (2014) 646e653.

[135] G. Paolella, et al., Gut-liver axis and probiotics: their role in non-alcoholicfatty liver disease, World J. Gastroenterol. 20 (42) (2014) 15518e15531.

[136] Y.Y. Ma, et al., Effects of probiotics on nonalcoholic fatty liver disease: ameta-analysis, World J. Gastroenterol. 19 (40) (2013) 6911e6918.

[137] A. Alisi, et al., Randomised clinical trial: the beneficial effects of VSL#3 inobese children with non-alcoholic steatohepatitis, Aliment. Pharmacol. Ther.39 (11) (2014) 1276e1285.

[138] M. Malaguarnera, et al., Bifidobacterium longum with fructo-oligosaccharides in patients with non alcoholic steatohepatitis, Dig. Dis.Sci. 57 (2) (2012) 545e553.

[139] P.N. Newsome, et al., Guidelines for liver transplantation for patients withnon-alcoholic steatohepatitis, Gut 61 (4) (2012) 484e500.