Nonalcoholic fatty liver in Asia: Firmly entrenched and rapidly gaining ground


  • Conflict of interest
    The authors have declared that they have no conflicts of interest.

Dr Shivakumar Chitturi, Senior Lecturer in Medicine, Australian National University Medical School, Gastroenterology and Hepatology Unit, The Canberra Hospital, Yamba Drive, Garran, ACT 2605, Australia. Email:


Nonalcoholic fatty liver disease (NAFLD) is becoming an important chronic liver disorder in Asia. Prevalence figures show regional variations but at least 10% of the general population in Asia have fatty liver. Fatty liver can develop with relatively small changes in weight (2–3 kg), often with increasing central adiposity. The metabolic syndrome may precede or follow NAFLD. Overt diabetes is present in one-third of cases but when oral glucose tolerance tests are performed, a further third of individuals have impaired glucose tolerance or diabetes. Natural history data are still scarce but cases of advanced hepatic fibrosis and hepatocellular carcinoma are now regularly reported. Many cases of cryptogenic cirrhosis are also attributable to NAFLD. Histological progression has been demonstrated for patients with NASH as well as for those with hepatic steatosis alone. Genetic factors may in part contribute to the rise in NAFLD. Polymorphisms within apolipoprotein C3 (APOC3) gene have been linked to NAFLD in lean Indian men. Although a number of other polymorphisms involving genes controlling adipose distribution, insulin signalling, adipokine responses and hepatic fibrosis have been reported, these studies have been underpowered. Transient elastography could help in detecting and monitoring hepatic fibrosis but further refinements in technique are necessary for obese individuals. Of the biomarkers, hyaluronic acid and cytokeratin-18 fragment testing show promise as markers of hepatic fibrosis and NASH, respectively. Lifestyle alterations including dietary changes and increased physical activity remain the cornerstone of management. Attention should be paid to prevention through public education of campaigns addressing the increase in both adult and childhood obesity.

Nonalcoholic fatty liver disease (NAFLD) encompasses a spectrum of liver disorders characterized by intrahepatic fat accumulation (simple steatosis) accompanied by varying degrees of hepatic necroinflammatory activity and hepatic fibrosis (non-alcoholic steatohepatitis (NASH)) through to cirrhosis.1 With prevalence figures of up to 30%, fatty liver has become the pre-eminent chronic liver disorder in the general population of industrialized North American, European and Australasian countries.2 The prevalence of fatty liver is even higher in persons with type 2 diabetes (50%), obesity (76%) and morbid obesity (nearly 100%).3

Individual case reports and small case series of Asian patients with NAFLD were first published in the 1970s and 1980s.4 However, interest in this disorder across the Asia-Pacific region gathered momentum only in the last decade, culminating in the inaugural Okuda lecture of the Asia Pacific Association for the Study of the Liver in 2003 with its central theme being the emergence of fatty liver in Asia.5 In the following year, the potential hepatic and metabolic implications of NAFLD were further explored in an article entitled “Nonalcoholic fatty liver disease in the Asia-Pacific: Future shock” published in this Journal.6 In 2007, a Working Party of regional experts convened in Hong Kong to evaluate the epidemiological and other aspects of fatty liver. Consensus guidelines on how NAFLD should be diagnosed and managed in Asia were drawn and published.7–11 In this review, we provide an updated account of progress in this field since 2007. We discuss whether the dire predictions of future disease burden are still valid, focus on emerging trends and finally, we examine possible strategies to deal with this growing problem.

Lessons from population studies

In the last decade, the results of a number of population-based studies and large surveys have become available (Table 1).12–17 These studies are more representative of the prevalence of NAFLD than data collated from tertiary centres. Broadly, they reaffirm that the prevalence of fatty liver across the region is at least 10%, and in some regions as many as one-third of individuals could be affected. However, regional variations within individual countries can be striking. These likely represent the impact of urbanization and affluence. For instance, the prevalence of NAFLD in China varies nearly two-fold between Chengdu (12.5%) to Central China (24.5%).12 The urban-rural divide is also becoming more apparent as seen in Guangdong province, China where the prevalence of NAFLD varies two-fold between urban (23%) and rural (13%) areas.18 Men outnumber women in most Western case series. Similar trends are observed in Asia. Beyond the age of 50 years, there is a sharp increase in the prevalence of NAFLD among women. Parallels can be drawn with the relative “protection” from cardiovascular disease for women in their pre-menopausal years but not beyond.

Table 1.  Prevalence of non-alcoholic fatty liver disease in Asia: population-based studies and health surveys from 2007
AuthorYearCountrynPrevalence of NAFLDAge (mean) in yearsMen (%)Obesity (%)Central obesity (%)Type 2 diabetes (%)Predictors of NAFLD
  • BMI ≥ 30 kg/m2.

  • Raised fasting blood glucose.

  • BMI, body mass index; DM, diabetes; F, female; FBG, fasting blood glucose; GGT, gamma glutamyl transferase; HDL-C, high-density lipoprotein-cholesterol; IR, insulin resistance; LDL-C; low-density lipoprotein cholesterol; M, male; TC, total cholesterol; TG, triglycerides; ULN, upper limit of normal.

Li122009China (Ningbo)8 92512%5766Not stated; mean BMI, 25.9 kg/m2277Female, age, BMI, waist circumference, GGT, TG, LDL-C, FBG
Lee132009South Korea13 62135% (M), 16% (F) 3552 60Obesity, male gender
Mohan142009India (Chennai)54135% (M), 29% (F)464736–7764–10038–48DM, obesity, central obesity, IR, serum lipids, metabolic syndrome
Das152010India (West Bengal)1 9118.7%3954253926Central obesity, raised FBG, increasing income
Jamali162008Iran2 0492%413522 4.2Male gender, overweight, obesity, DM, urban location
Dassanayake172009Sri Lanka2 98533%534547 (F); 30 (M)70 (F); 30 (M)35 (F); 37(M)Obesity, acanthosis nigricans, IR, FBG, TG, diastolic blood pressure, ALT x 2 ULN

Anthropometric profiles

A welcome feature of recent Asian publications has been the near universal adoption of regional anthropometric criteria. As expected, most individuals with fatty liver are overweight or obese (the latter defined by body mass index (BMI) ≥ 25 kg/m2). Small changes in weight (2–3 kg) can increase the risk of fatty liver. Even more alarmingly, this may occur within the “normal” range. In a South Korean study, a weight gain of > 2.3 kg was associated with hepatic steatosis even in individuals within the non-obese range (BMI 18.5–22.9 kg/m2).19 It is difficult to interpret this study without details of body fat distribution because central (truncal, visceral) obesity is a better measure of insulin resistance and is even more closely associated with NAFLD than an increased BMI.20–22 On the other hand, persons with a pronounced subcutaneous fat distribution and peripheral adiposity (typically in the thigh region) are less likely to have significant hepatic steatosis.

Insulin resistance, metabolic syndrome, future metabolic risk

Cross-sectional and longitudinal Asian studies reaffirm the strong links of fatty liver with insulin resistance and the metabolic syndrome (MetS).23 The latter was present in ∼ 70% of Chinese patients with NAFLD whereas the figures for the general population are ∼ 7% (MetS).24 Similar patterns were found in India, Korea, Taiwan, Sri Lanka, and other Asian countries.14,17,25–28 There is a bidirectional relationship between NAFLD and MetS-related disorders. Thus, individuals with MetS-related conditions have a higher future risk of NAFLD and in turn, fatty liver confers an increased future risk of developing these disorders. In a prospective study of 4401 Japanese subjects undergoing routine health evaluation, MetS increased 4-fold and 11-fold the risk of incident fatty liver in men and women, respectively.29 Further, regression in hepatic steatosis occurred in 16%, but this was less likely to occur among those with MetS at baseline.

For individuals with NAFLD, what is the future risk of developing metabolic disorders? The best estimates that are available were derived from a Chinese study and place the odds of developing type 2 diabetes, hypertriglyceridemia, obesity and hypertension at 4.6. 3.3, 3.4 and 2.9, respectively.30 For other regions, individual and other ethnic influences underpinning the MetS should be taken into account.

Glucose tolerance and fatty liver

The prevalence of fatty liver increases in parallel with progressive degrees of abnormal glucose tolerance. In a study of 541 subjects from Chennai, India, the prevalence of fatty liver in persons with normal glucose tolerance, prediabetes (impaired glucose tolerance and impaired fasting glucose) and diabetes was 23%, 33% and 55%, respectively.14 Type 2 diabetes is an important predictor of advanced hepatic fibrosis1 and therefore subjects with type 2 diabetes should be carefully evaluated for signs of advanced liver disease.

Detecting diabetes early could also pave the way for early therapeutic intervention. Overt diabetes, as measured by fasting plasma glucose assessment, is present in only about a third of cases. However, recent studies have emphasized the importance of detecting abnormalities of glucose tolerance even in patients with normal fasting glucose levels. When an oral glucose tolerance test (OGTT) is performed in non-diabetic NAFLD patients, impaired glucose tolerance and newly diagnosed type 2 diabetes are observed in one-third to over half of the patients.24,31,32 Therefore, we propose that all non-diabetic patients with NAFLD should undergo an OGTT. These would be in addition to the mandatory measurement of anthropometric indices, fasting glucose and lipid levels.11

Pediatric NAFLD

A disturbing trend has been the rise in childhood obesity levels, and, consequently, cases of pediatric NAFLD. This is to be expected as overweight and obese adolescents have a 4–5 fold risk of developing hepatic steatosis.33 The importance of central obesity has also been highlighted in children. In one Taiwanese study, with every 5 cm increase in waist circumference, there was a 1.4 fold increase in the risk of developing fatty liver.34 Viral hepatitis-endemic regions are now seeing NAFLD emerge as the main source of abnormal liver test profiles.35 Five Taiwanese cross-sectional studies conducted after universal hepatitis B vaccination was introduced, have shown that fatty liver accounts for nearly half (46%) of the cases of alanine aminotransferase (ALT) elevation in adolescents.

The influence of genetic factors has not been well studied in Asia (see Pathogenesis).36 In one study of 234 obese Taiwanese children, those carrying variant UGT1A1*6 genotypes associated with hyperbilirubinemia had a lower risk of having NAFLD (OR 0.31, 0.11–0.91).36 The authors hypothesize that this is due to the antioxidant effect of higher bilirubin levels.

Impact of NAFLD on chronic viral hepatitis B and C

Interactions between different chronic liver diseases (e.g. alcohol-related and hepatitis C; hepatitis C and overweight/obesity) are often synergistic and lead to progressive liver injury. For patients with chronic hepatitis C, in particular genotype 1 infections, host factors such as central obesity are linked to the grade of hepatic steatosis and advanced hepatic fibrosis. Hepatitis C virus infection also induces insulin resistance37 and is associated with a 2–3 fold risk of developing type 2 diabetes. Therefore, there is great interest in studying whether a similar association is also present with the dominant viral hepatitis infection (hepatitis B) in Asia. Co-existing hepatic steatosis is reported in at least a quarter of patients with chronic hepatitis B (CHB),38 but, to date, the intrahepatic fat is related only to host factors such as body mass index, lipid levels and insulin resistance and the MetS.39 Further, individuals with and without hepatitis B infection had comparable indices of insulin resistance. Inexplicably, studies from Taiwan and China have even shown an inverse relationship between hepatitis B surface antigen positivity and the metabolic syndrome!40,41 (Luo and Jan). Taken together, these reports reiterate the lack of association of hepatitis B virus infection with insulin resistance. However, this does not mean that metabolic factors are not important in patients with CHB. In fact, one Chinese study of patients with CHB showed co-existence of MetS was associated with a greater risk of cirrhosis (OR 1.7), and this has recently been confirmed in North America.42 Further, the strength of association with cirrhosis increased progressively with the number of MetS components present (OR of 1.4, 2.6, 4.1, 4 and 5.5 for patients with one, two, three, four and five components of MetS, respectively).43


The key pathogenic processes documented in studies elsewhere include insulin resistance, hypoadiponectinemia and oxidative stress.1 Their importance in contributing to NAFLD has been reiterated in Asian studies. It is beyond the scope of this article to review the pathogenesis of NAFLD44 but data on recent studies implicating genetic factors will be briefly discussed. Until recently, the genetic contribution to NAFLD has been largely ignored because alterations in lifestyle have been blamed for the fatty liver epidemic. This perception is likely to change with the publication of studies such as the one by Schwimmer et al. in the USA, who observed a high frequency of fatty liver among siblings (59%) and parents (79%) of children with NAFLD.45 This has fuelled interest in genetic studies. However, most of these studies in Asia (and elsewhere) have mainly involved evaluation of candidate genes. The latter were chosen for their known associations with insulin resistance, MetS, inflammatory and adipocytokine responses and hepatic fibrogenesis.

Most of the reported studies are small and underpowered to detect significant differences. As expected, single nucleotide polymorphisms (SNPs) related to genes coding for tumor necrosis factor-alpha (TNF-α), TNF-α related apoptosis-inducing ligand (TRAIL), leptin, adiponectin, peroxisome proliferator-activated receptors (PPAR) and angiotensin receptors have showed significant association.46–48 but the findings have not been consistent.49 To date, there has been only one adequately powered candidate gene study in Asian subjects with fatty liver. Lean Indian men with NAFLD were found to carry two gain-of-function single-nucleotide polymorphisms (SNPs) within the gene encoding apolipoprotein 3 (APOC3).50 The variant allele carriers had a 30% increase in plasma apolipoprotein C3 and a 60% increase in plasma triglycerides and marked insulin resistance. The prevalence of NAFLD among carriers of the two variant SNPs (C-482T and T-455C) was 38% (absent in wild type homozygotes). Through its effects on inhibiting hepatic lipase activity and delaying catabolism of triglyceride-rich particles, the physiological actions of APOC3 are pro-steatotic; inheritance of variants, known to be associated with increased APOC3 levels, would exacerbate this tendency. A similar but less striking association (9% prevalence of NAFLD in carriers) was also seen in non-Asian Indian men. This suggests that this genetic abnormality is neither exclusive to Asian Indians nor completely accounts for fatty liver in that ethnic group.

With the emergence of genome wide association studies (GWAS), fresh hypothesis-free approaches to examining genetic contribution to polygenic diseases have become available. Outside Asia, one of the main genes identified in GWAS studies is the patatin-like phospholipase domain-containing 3 protein (PNPLA3). SNPs within PNPLA3 have been linked to hepatic steatosis, inflammation and fibrosis.51,52 Subjects carrying APOC3 as well as PNPLA3 variants have a higher prevalence of fatty liver than those carrying APOC3 alone.50 Other GWAS have identified an association between the NAFLD activity score and farnesyl diphosphate farnesyl transferase 1 (involved in cholesterol biosynthesis) and other SNPs within or in the vicinity of genes involved in hepatic fibrogenesis (e.g. platelet derived growth factor A).53

Natural history

The natural history of this disorder is well documented in European populations and is defined largely by histologic subtype.1 Persons with simple steatosis usually have a benign non-progressive course,54 while 10% to 15% with nonalcoholic steatohepatitis (NASH) can develop progressive hepatic fibrosis and cirrhosis.1 The outcome of fatty liver-related cirrhosis is poor and the survival curves for persons with hepatic decompensation are similar to those seen in patients with end-stage viral hepatitis.55 There is a small but additional risk of the fatty liver substrates (obesity, T2D) contributing to the risk of hepatocellular carcinoma (HCC).56,57 Recently, two groups have confirmed in a larger retrospective analysis that the incidence of HCC is broadly similar between patients with NASH-related and hepatitis C-related cirrhosis (annual incidence 2.6% and 4.0%, respectively).58

Asian longitudinal studies evaluating outcome are scarce. In small retrospective series, liver decompensation and HCC are rarely seen.59,60 On the other hand, when NAFLD patients have progressed to cirrhosis, the clinical outcome is not different from that of patients with cirrhosis of other causes. In a retrospective study of 68 Japanese patients with NASH-related cirrhosis and 69 patients with hepatitis C-related cirrhosis, the 5-year survival rates were 75% and 74%, respectively.61 HCC was the cause of death in 47% and 68%, respectively. Takuma et al. reviewed 94 published cases of NASH-related HCC.62 The majority were male (64; mean age 66 years) and most had features of the metabolic syndrome; 68% were obese, 66% had diabetes and 24% had dyslipidemia. More than two-thirds (69%) had multinodular tumors (1.4–13 cm; mean 3.5 cm) but a quarter of these lesions (26%) arose in a non-cirrhotic liver.62 Surveillance programs for HCC should be instituted for patients with NAFLD. Whether this applies to all cases is unclear but patients at risk (older age, advanced hepatic fibrosis, previous alcohol over-usage) should be monitored.63 Dr Okanoue has written a detailed review of NAFLD and NASH in Japan, highlighting the importance of HCC as a complication of type 2 diabetes as well as other aspects, which accompanies the present article in this 25th anniversary supplement of JGH.64

Disease severity

Table 2 lists cross-sectional studies in the Asia-Pacific region where histologic details were provided. Overall, over half of these patients had NASH.27,59,66–72 However, it is important to note that the definition of NASH has not been uniform among studies. Histologic studies are biased towards patients seen at tertiary centres, usually presenting with abnormal liver function tests and multiple metabolic risk factors; patients are therefore likely to have more active disease. Even so, advanced liver fibrosis or cirrhosis have been relatively uncommonly reported. For example, in a study of 246 NAFLD patients from France and Hong Kong, advanced fibrosis or cirrhosis was found in 28% of Caucasian patients but in only 17% of Chinese patients.74

Table 2.  Histologic features of NAFLD patients in Asia-Pacific: cross-sectional studies
AuthorYearCountrynMean age (years)Male (%)NASH (%)Advanced fibrosis or cirrhosis n (%)Predictors of advanced hepatic fibrosis
  • Characteristics of the cohort as a whole (n = 100).

  • Portal area-predominant inflammation and fibrosis.

  • AST, aspartate aminotransferase; HOMA-IR, homeostasis assessment method-insulin resistance; ULN, upper limit of normal.

Hui652004Australia10941–56627331 (28%)HOMA-IR
Wong592004Hong Kong, China424752861 (2%)Type 2 diabetes
Tsang272006Hong Kong, China604663976 (10%)AST > 2 x ULN, hepatic necroinflammatory score, type 2 diabetes
Duseja662007India383870534 (11%) 
Malik672007Malaysia7547778418 (24%)Men, Indian ethnicity
Park682007Korea393377623 (13%) 
Yoneda692008Japan97524127 (28%)Liver stiffness score
Ko702009Korea801284787 (9%)Type 2 NASH
Wong712009Hong Kong, China17347546519 (11%)Age, fasting glucose, HOMA-IR
Sobhonslidsuk722009Thailand30541710015 (50%)HOMA > 3.5, grade of portal inflammation

The reason for the apparently lower prevalence of advanced disease in Asian NAFLD patients is not completely understood. We believe it is likely to be due to a combination of both genetic and environmental factors, as well as socio-economic history between geographic regions. Thus, it is likely that the probability of whether a patient would develop cirrhosis and its complications is linked to the duration of “metabolic overload”. Since the economic surge in many Asian countries only began in the 1980s and 1990s, it is possible that current patients who exhibit only signs of mild liver injury may yet present later with more severe complications, consistent with the clinical observation in Australia that the vast majority of patients with liver complications from NAFLD are aged older than 60 years. Further, with increasing adult and also childhood obesity, the number of Asian patients with developing advanced liver disease is expected to rise. A small case series of Sri Lankan children with advanced hepatic fibrosis secondary to NASH is a case in point; 4 of the 5 children were obese (BMI 26–31 kg/m2) and all were insulin-resistant.34 Finally, cases of NASH-related cirrhosis masquerading as “cryptogenic” cirrhosis should also be considered in the tally of patients developing advanced liver disease, as discussed next.

Cryptogenic cirrhosis

By general agreement, the majority of Western cases with cryptogenic cirrhosis (CC) represent “burnt-out” NAFLD.75 Whether these considerations also apply to patients with CC in a viral hepatitis-endemic region has not been well studied. Here too, published data would suggest that many such cases are also secondary to fatty liver. In a study of liver explants from 30 Indian patients with CC, 19 (63%) showed histologic features consistent with NAFLD.76 The bigger question is the contribution of NAFLD/NASH to cirrhosis in the general population. Here too, there are some unsettling trends. In a population-based Indian study, the prevalence of fatty liver-related cirrhosis was 0.2% in West Bengal.15 This is less than described in a North American setting (2% of NAFLD subjects in the Olmsted county study).54 Differences in the prevalence of obesity (25% vs 75%) may have contributed. However, what is remarkable is that the Indian study was conducted in a region where nearly half the population (47%) was underweight (BMI < 18.5 kg/m2) and yet, many exhibited markers of increased adiposity (e.g. percentage of body fat). Meanwhile, in more affluent Asian countries, NASH-related cirrhosis is already on the rise. NASH accounts for 2.1% of Japanese cases of cirrhosis.77 In a North American study, NAFLD accounted for 14.7% of cases with cirrhosis. Although the Japanese data seem to indicate a lesser problem, the tally of cases with NAFLD-related cirrhosis could be higher (5%–6%) if some cases of CC were also included.

Histologic progression

In a retrospective study from Hong Kong, 17 patients underwent paired liver biopsies at a median interval of 6.1 years (range 3.8–8.0 years).78 Progression of hepatic fibrosis was noted in 9 (53%) patients. Recently, the same group carried out a prospective study of 52 NAFLD patients with planned paired liver biopsies three years apart.79 Overall, 14 (27%) patients had fibrosis progression by one stage or more. In addition, over half of the patients with simple steatosis developed NASH or borderline hepatic necroinflammatory activity. On the other hand, reduction in BMI and waist circumference was associated with a non-progressive course. This suggests that simple steatosis is not a completely inert disease but may progress with unfavourable changes in metabolic profile. In another study involving 39 Japanese patients, paired liver biopsies were performed at a median interval of 2.4 years (range 1.0–8.5 years).80 Liver fibrosis progressed in 11 (28%) patients, remained static in 16 (41%), and improved in 12 (31%). The authors observed that tight glycemic control, as measured by changes in glycosylated hemoglobin, was associated with improvement in liver fibrosis. Both these studies show that improving the metabolic profile can be helpful in retarding the progression of NAFLD.


When a patient presents with features of NAFLD, the assessment should include: (i) confirmation of the diagnosis; (ii) assessing disease severity; and (iii) detecting concomitant metabolic disorders and cardiovascular diseases.

The current guidelines endorse hepatic ultrasound imaging as the first step of diagnostic evaluation.7 Characteristics of NAFLD on ultrasound scan include increased liver echogenicity, vascular blurring, and deep attenuation of the ultrasound signal. A combination of these three ultrasound criteria has good accuracy in detecting fatty liver, and correlates well with visceral obesity and MetS.81 The diagnosis of NAFLD also requires exclusion of other liver diseases, particularly hepatitis B and C infections and also alcoholic liver disease. Individuals with an alcohol intake of > 20 g/d (male) and > 10 g/d (female) (or 140 g/wk and 70 g/wk, respectively) are excluded.7 These limits may be seen as being excessively restrictive by some or permissive by others but these were set down to provide an operational definition that seeks a balance between too low or excessive alcohol consumption. In support of these imposed limits, light-to-moderate use of alcohol use was found to be protective against NAFLD in population based studies such as the Italian Dionysos study.82 Similar Asian data are now available. In one Japanese study of over 60 000 individuals undergoing routine evaluation, the prevalence of fatty liver was lower (8%–9%) in persons consuming 23 g/d of alcohol (“moderate drinkers”) than for non- and occasional drinkers (12%–28%), respectively. Still, the risks of fatty liver were significant (19%) in men consuming 2–3 drinks/day (46–69 g of alcohol).83 Similarly, in Guangzhou, China, while obesity along with diabetes, lipid levels and glucose were strongly associated with fatty liver, so too was alcohol abuse (OR 18.6).84 Therefore, the current definitions of alcohol consumption thresholds should continue to be applied. It is also now clear that the risk of cirrhosis in persons who consume excess alcohol are greatest among those with obesity, insulin resistance and T2D,42 and the link between earlier alcohol consumption and increased risk of HCC was mentioned earlier.42 Hence while a definition of NAFLD based on restrictive levels of current alcohol intake is required for disease definition, many patients fall outside this in real life practice, and their risks of liver complications may be higher than those with “pure” NAFLD, as currently defined.

Liver histology remains the gold standard for assessing disease severity in NAFLD. However, its invasive nature renders it unsuitable for community studies and in particular, for studying hepatic fibrosis progression. Further, sampling errors are substantial in histological assessment of NAFLD, and this often leads to understaging of hepatic fibrosis, particularly when biopsies are too small. Therefore, alternative methods to assess liver disease severity are being evaluated.

Noninvasive assessment of liver fibrosis

Two main methods have been evaluated—image-based tests and serum biomarkers.

Hepatic imaging

Of the various image-based tests, transient elastography using FibroScan (Echosens, Paris, France) has been extensively studied. A shear wave generated by the device is transmitted across the liver parenchyma. The velocity of the shear wave increases with liver stiffness. The latter provides an estimate of the degree of liver fibrosis. Two Asian studies have examined the performance of transient elastography in NAFLD subjects.69,73 Overall, successful acquisitions were obtained in over 97% of subjects with BMI < 30 kg/m2 but this dropped to 75% for subjects with BMI > 30 kg/m2.73 Among patients for whom successful liver stiffness measurements could be made, the accuracy of transient elastography was good. The area under the receiver operating characteristic (ROC) curves for diagnosing bridging fibrosis and cirrhosis was over 80% and 90%, respectively.

Acoustic radiation force impulse elastography is another ultrasound-based technique for measuring liver stiffness using short-duration acoustic pulses. The advantage of this test is its integration with conventional ultrasound devices. In a study of 54 Japanese patients with biopsy-confirmed NAFLD, this technique had 100% sensitivity and 91% specificity in detecting bridging fibrosis, values similar to those obtained by Fibroscan.85 More studies are required to better define the accuracy, reproducibility and limitations of this new method.

Serum biomarkers

Liver fibrosis has also been evaluated using serum biomarkers and prediction scores utilising multiple clinical and biochemical variables. Of the former, hyaluronic acid, a component of the extracellular matrix, shows promise as a predictor of severe fibrosis (bridging fibrosis and cirrhosis). In a study of 148 Japanese NAFLD patients,86 it had a negative predictive value of 100% for severe fibrosis with good specificity (89%, 95% C.I 80–94%). On the other hand, a low platelet count (< 160 000/mm3) was better at excluding cirrhosis than HA levels. The high negative predictive of hyaluronic acid in excluding severe hepatic fibrosis was also noted in a North American study.87

Prediction scores

A multi-centre study involving North American, European and Australian centres developed the NAFLD fibrosis score. The latter includes six variables—age, hyperglycemia, BMI, platelet count, albumin, and aspartate aminotransferase (AST)-to-ALT ratio—and had good accuracy in detecting advanced fibrosis.88 However, its performance was less satisfactory when used in Chinese subjects, with areas under ROC curves of only 67% and 64% for F2 and F3 disease, respectively.89 The differences in the performance of NAFLD fibrosis may due to differences in case selection. The Chinese study included fewer patients with advanced liver disease and early liver decompensation, in which platelet count, albumin and AST/ALT ratio might have better discriminating power. Furthermore, owing to the differences in fat distribution between Asian and Caucasian subjects, prediction scores including BMI might need further calibration and modification before being used in Asian studies. Among various prediction scores reported to date, the FIB-4 index, based on age, AST, ALT and platelet counts, appears to have higher accuracy than the others to detect liver fibrosis in both Caucasians and Chinese.72,90 Overall, scoring systems are good-to-excellent in identifying patients with advanced fibrosis but are less impressive in identifying cases with mild fibrosis, at which point therapeutic intervention is likely to be more effective.91

Non-invasive diagnosis of NASH

In comparison to hepatic fibrosis, there have been fewer developments in developing non-invasive tests for diagnosing NASH. Of these, measurement of plasma cytokeratin-18, a marker of hepatocyte apoptosis, could be of value. Hepatocyte apoptosis is a characteristic feature of NASH as opposed to simple steatosis.92,93 Recently, a prospective study in Chinese patients with paired liver biopsies confirmed that alterations in serum cytokeratin-18 fragment level correlated well with changes in the NAFLD activity score.79 Likewise, serum levels of adipokines have been tested in NAFLD subjects. In general, patients with NASH tend to have lower serum levels of adiponectin and higher tumor necrosis factor-alpha and interleukin-6 level.24,65 However, the overall accuracy of these markers has not been fully evaluated and is probably limited by their variability with time.

NAFLD and cardiovascular disease

As the hepatic manifestation of the MetS, it is expected that coronary artery disease (CAD) will be an important cause of morbidity and mortality in longitudinal studies. This has been borne out in both population-based as well as clinic-based studies. However, data are accruing that the CAD risk with NAFLD may be greater than that expected through its association with the MetS.94 Possible mechanisms include the contributions of NAFLD-related pathogenetic processes and epiphenomena such as oxidative stress, inflammatory cytokine alterations, changes in blood coagulation and an unfavorable atherogenic lipid profile.

Asian studies

In a study of 317 adult Iranian patients undergoing coronary angiography, the detection of fatty liver by ultrasound scan increased 8-fold the risk of significant coronary artery disease.95 In addition, there are several studies showing an association with other markers of general cardiovascular risk such as carotid intima-media thickness,96,97 and total Framingham risk score98 as well as those specific to CAD (coronary artery calcium score).99 However, since prospective data linking NAFLD and hard cardiovascular outcomes are not consistent among studies, routine workup for coronary artery disease cannot be recommended at this stage. Nevertheless, clinicians should be alert for symptoms and signs of vascular diseases.


Lifestyle modification is the cornerstone of management of NAFLD. In observational studies, even modest weight loss (2–3 kg) is associated with reduction in hepatic steatosis and other histological improvement.79,100 Lifestyle programs emphasizing calorie and fat restriction and regular exercise have been successfully implemented both in adults101–103 and also children.104 Aerobic exercise training has been shown to reduce intrahepatic triglycerides and visceral fat even in the absence of significant weight changes. In a randomized controlled trial conducted in Australia, 19 NAFLD patients were randomized to aerobic exercise training or usual treatment for 4 weeks.105 Using magnetic resonance spectroscopy, patients undergoing aerobic exercise training showed a 21% reduction in hepatic triglyceride content and a 12% reduction in visceral fat. However, a combination of diet and exercise appears to be superior to either diet or exercise alone. In a three-arm Taiwanese study, 54 patients were assigned to diet plus exercise, exercise or to a control arm.106 As compared with controls, both the intervention groups showed improvement in lipid profiles, insulin sensitivity and anthropometric indices but the improvement in metabolic profile was greater in the combined diet-exercise group than those assigned to exercise only.

At present, there is no registered drug treatment for NAFLD. Early studies suggest that insulin sensitizers and antioxidants may confer some benefit whereas ursodeoxycholic acid107 and pentoxifylline108 have not survived the scrutiny of randomized trials. In patients with morbid obesity, bariatric surgery appears safe and may improve hepatic steatosis and necroinflammation.109 While bariatric surgery has become more widely available in Asia, data on outcomes with respect to NAFLD are awaited but improvement in BMI and liver tests were reported in one Japanese study.110

Future directions

Stemming the tide of the metabolic syndrome and its consequences will be a considerable challenge in Asia, as elsewhere. With respect to NAFLD, the approach to management will have to encompass both narrow and broad perspectives. With respect to the latter, these should include efforts to prevent the development of metabolic syndrome (e.g. by lifestyle interventions in childhood), public education and facilitating and encouraging physical activity and more appropriate (healthier) dietary habits among adults. Equally important is the need to retain a narrow focus on those individuals at risk of hepatic and/or metabolic complications. These would include not only individuals with type 2 diabetes and the obese, but also the “average” individual (either slightly overweight or not) who may still be at risk of serious sequelae. Identifying host susceptibility factors through collaborative efforts and enrolment in genome wide association studies is critical. On the other hand, the influence of environmental factors such as diet needs to be explored further. Asian diets vary considerably and studying how these nutritional factors might influence fatty liver will be important. Finally, current studies addressing the relationship between this liver disorder and cardiovascular disease have been mainly cross-sectional or retrospective in design. The ultimate acceptance of NAFLD into the fold of the metabolic syndrome rests on well-conducted prospective studies to clarify this association.