This article is more than 5 years old. Information may no longer be current.

Beyond HCV: Non-invasive Assessment of Non-Alcoholic Fatty Liver Disease

Non-alcoholic fatty liver disease is one of the most common causes of liver disease in developed countries with an estimated prevalence of 30% and 33% in the United States and European populations, respectively. After excluding significant alcohol consumption and other secondary causes of steatosis (ie, medications), NAFLD can be defined by the presence of hepatic steatosis identified by either imaging or histology.

NAFLD describes a spectrum of diseases ranging from simple steatosis to non-alcoholic steatohepatitis (NASH), which most commonly develops in individuals with metabolic risk factors, including diabetes mellitus, hyperlipidemia and obesity. Underlying all of these metabolic derangements is insulin resistance, which has been recognized as a key pathogenic feature for liver disease progression in individuals with NASH. While the presence of simple steatosis has conventionally been associated with absent or delayed histological progression, individuals with NASH show marked hepatocyte damage and inflammation resulting in an increased risk for progression to advanced fibrosis, cirrhosis, and hepatocellular carcinoma. Indeed, NASH is projected to become the leading indication for liver transplantation in the U.S. by 2025.

NAFLD is significantly underestimated and underdiagnosed, particularly in the primary care setting, which is contributing tremendously to health-related expenditures and the burden of chronic liver disease. Although it has conventionally been thought that only those individuals with a diagnosis of NASH are at risk for histologic progression or development of HCC, recent data based on a meta-analysis of paired liver biopsies suggest that certain individuals with simple steatosis may still be at risk for progressive liver disease, although at a slower annual fibrosis progression rate than NASH (one stage over 14.3 years vs. 7.1 years, respectively). NAFLD is highly prevalent in patients with diabetes and obesity, however insufficient data exist at present to recommend routine screening for NAFLD in these high prevalence groups. Indeed, approximately 5% of severely obese patients undergoing bariatric surgery will have undiagnosed cirrhosis.

Given the frequency of NAFLD/NASH within the general U.S. population, the lack of awareness of the disease among primary care physicians, and the fact that many promising investigational agents are currently in clinical trials, it is imperative that all physicians — including those treating hepatitis C virus — who are likely to encounter NAFLD/NASH patients in their clinical practice, be aware of the risk factors and the methods presently employed to establish a diagnosis.

Current guidelines authored by the American Association for Study of Liver Diseases recommend calculation of the NAFLD fibrosis score (NFS; http://nafldscore.com) and the presence of the metabolic syndrome to identify those NAFLD patients at highest risk for NASH and advanced fibrosis. On the basis of noninvasive testing, patients with a high probability of a NASH diagnosis, those at the more severe end of the disease spectrum, are the ones most likely to benefit from a liver biopsy. Although the liver biopsy remains the gold standard for the diagnosis of NAFLD/NASH and the differentiation between the two entities, this technique is not without limitations including procedural complications, sampling error and cost.

In addition, we continue to identify patients who do not belong to groups traditionally thought to be at high risk for NASH. For example, NAFLD is being increasingly recognized among non-obese and non-diabetic individuals who may have a genetic predisposition. Therefore, given the high prevalence of NAFLD/NASH and the difficulty in identifying patients with a NAFLD/NASH diagnosis, accurate and reproducible non-invasive tests are needed to select those at risk for progressive liver disease who should be referred to a liver specialist for further evaluation.

NAFLD Pathogenesis and the Role for Non-invasive Biomarkers

The development and progression of NAFLD is thought to occur through ‘multiple hits’ involving a complex interaction between various inflammatory cytokines and adipokines, cytokines secreted by visceral adipose tissue, which lead to an increase in free fatty acids (Figure 1). Certain adipokines, including adiponectin and leptin, have been shown to be involved in the regulation of inflammation and in the development of insulin resistance. Adiponectin demonstrates anti-inflammatory effects and lower levels have been found in those with NASH compared with those with steatosis and with healthy controls. In the setting of obesity, the release of inflammatory cytokines and reactive oxygen species may interfere with apoptosis regulation and induce hepatocarcinogenesis, even in the non-cirrhotic liver. Although the exact incidence is not known, cases of HCC have occurred in patients with histologically confirmed NASH in the absence of underlying cirrhosis.

In addition to an increase in pro-inflammatory mediators, such as tumor necrosis factor-alpha, insulin resistance can lead to an increase in oxidative stress that further promotes hepatic steatosis and fibrogenesis via activation of hepatic stellate cells. Insulin resistance, therefore, has emerged as a key underlying feature in NAFLD pathogenesis, which itself is exacerbated by the presence of intracellular hepatic lipid accumulation resulting in hyperinsulinemia from reduced insulin clearance. Given the increased understanding of pathogenic mechanisms in NAFLD/NASH, multiple inflammatory and apoptotic mediators have been proposed as markers of disease activity as is discussed below. Many of these mediators are also being studied as potential therapeutic modalities.

Although NAFLD more commonly occurs in the setting of obesity, it is also increasingly being recognized in lean individuals, as stated previously, with an estimated prevalence of approximately 15% to 20% in patients with a BMI less than 25, primarily in the Asian population. Although the exact mechanisms leading to steatosis in the non-obese are unclear, upregulation of inflammatory mediators remains a potential causative factor. Additionally, hypertriglyceridemia and/or genetic susceptibility may play a role, particularly in Asian and Hispanic populations. Of the genetic markers, PNPLA3, which is involved in lipid metabolism, was the first to be identified as a candidate gene in susceptibility to NAFLD. Specifically, the variant G allele was shown to be associated with NASH, independent of age, BMI and metabolic syndrome. Honda et al showed that in a Japanese NAFLD cohort, the PNPLA3 GG genotype (rs738409) occurred more frequently in non-obese (n = 134) compared with obese individuals (n = 40; 47.8% vs. 36.5%; OR 4.15 vs. 2.76, respectively). The allele was found to be associated with more severe disease, specifically lobular inflammation and hepatocyte ballooning. PNPLA3 was also shown to confer increased risk for HCC in non-obese patients without obvious metabolic risk factors. Although these observations suggest that genetic testing might be beneficial, at present data are insufficient to recommend inclusion of genetic testing in the routine evaluation of NAFLD/NASH.

The predilection to develop insulin resistance at a lower BMI, as influenced by race and ethnicity, is an important consideration in the development of NAFLD/NASH. For example, several studies have illustrated that the diagnosis of NAFLD/NASH derived using BMI values mainly based on studies performed among predominantly white populations may not be appropriate for Asians. Deurenberg et al reported that for a given percent body fat, Asians have a BMI that is 3 kg/m2 to 4 kg/m2 lower than whites. Asians also have higher levels of visceral or intra-abdominal fat, a potentially stronger predictor of insulin resistance and diabetes risk compared with total body fat. Asian American women were also found to have higher visceral fat compared with European women (0.85 vs 0.51 kg; P = .0002), even after adjusting for age and total body fat. A comparable prevalence of metabolic syndrome observed in white patients with a BMI of 25 kg/m2 was seen in Asian men with a BMI of 19.9 kg/m2. Therefore, it is important to keep in mind different cutoffs for BMI when evaluating patients from different races and ethnicities for the potential presence of NAFLD/NASH.

Non-invasive Modalities in Diagnosis and Staging of NAFLD

Vast amounts of data have accrued evaluating the role of imaging techniques as well as circulating biomarkers, both individually and in combinatorial indices, for their ability to accurately diagnose NAFLD/NASH and its histological progression. Although results obtained to date are promising, no single modality alone has yet achieved the ability to supplant liver biopsy as the gold standard for the diagnosis and assessment of histological severity in NAFLD/NASH. In view of the increasing number of individuals in whom a diagnosis of NAFLD is suspected, liver biopsies cannot be feasibly performed in all patients. Consequently, an urgent need exists for an improved understanding of the factors that promote histologic progression and to identify more accurate, non-invasive, and cost-effective tools by which to classify patients with a NAFLD/NASH diagnosis who are likely to have histologic progression.

Imaging Modalities

Ultrasonography (US) is widely available and is the most commonly used radiologic technique to assess for hepatic steatosis. It has variable sensitivity (60%-94%), which is reduced especially in the setting of lesser degrees (<30%) of hepatic steatosis. Principal limitations of US are the inability to accurately differentiate NASH from simple steatosis and the high operator dependency. As an advance over US, a sophisticated MRI technique, specifically magnetic resonance spectroscopy (MRS), has emerged as a promising, highly sensitive tool that can quantify varying degrees of hepatic steatosis with good accuracy (area under the receiver operating curve (AUROC) >0.85). The technique has also been shown to correlate well with liver fat content (P < .001). MRS, however, is less precise in the setting of advanced fibrosis and it requires a high level of radiologic expertise to accurately interpret the results. Limited availability and expense also greatly restrict its use.

Transient elastography (TE), a non-invasive tool used to assess liver stiffness, has demonstrated high diagnostic accuracy in identifying more advanced fibrosis (stage 3 or more) (AUROC = 0.94) and a negative predictive value of 95% in a study comparing the performance of various non-invasive markers/indices, including the aspartate aminotransferase (AST) to platelet ratio (APRI) and NFS. However, because of the difficulty in accurately assessing liver stiffness in obese (BMI >30) individuals, its accuracy may be reduced in that setting. Recently an XL probe, designed specifically for obese patients, was compared with the standard M probe; it yielded higher diagnostic accuracy in patients with skin to liver capsule >2.5cm. Greater skin-to-liver capsule distances (>3.4cm) and more severe obesity (BMI >40), however, were risk factors for probe failure and decreased the ability of TE to accurately estimate the stage of hepatic fibrosis. Another recent improvement in TE, specifically applied to the diagnosis of NAFLD/NASH, has been the development of vibration controlled transient elastography (VCTE). VCTE is an adaptation that incorporates calculation of a controlled attenuation parameter that more accurately estimates varying degrees of steatosis, including identification of earlier steatosis, than is possible with standard TE.

Laboratory Markers

Routinely performed laboratory tests, such as AST and alanine aminotransferase (ALT) levels, are markers of liver injury; however, they demonstrate low predictive value for estimation of the stage of fibrosis and can severely underestimate the degree of underlying liver damage. An AST/ALT ratio less than one was found to be indicative of underlying NAFLD, although approximately one half of individuals with NAFLD/NASH have normal aminotransferase levels. Although aminotransferase levels do not have sufficient predictive ability as screening tests, the presence of persistently abnormal biochemistries in the setting of the metabolic syndrome or insulin resistance may help to identify those at risk for advanced fibrosis and those in whom a liver biopsy could be considered.

Based on the results obtained from genome-wide association studies and proteomics studies, several potential biomarkers of NASH including inflammatory cytokines (ie, TNF-alpha, IL-6, CCL2, ferritin), chemokines, regulators of hepatocyte apoptosis (ie, cytokeratin 18 [CK18]) and extracellular matrix components (ie, hyaluronic acid). CK18 represents one of the more extensively studied and promising circulating biomarkers used to distinguish patients with NASH from those with simple steatosis (AUROC = 0.82, sensitivity = 78%, specificity = 87%). Cusi et al demonstrated less favorable results for CK18 compared with MRS, however, yielding poorer sensitivity (58% and 54%) and negative predictive values (49% and 56%) for NASH and advanced fibrosis, respectively. Although CK18 may be a useful biomarker when combined with other parameters or clinical data, testing for CK18 alone to establish a diagnosis of NASH is not currently supported by professional society guidelines.

Diagnostic Panels

As there is currently no individual biomarker with sufficient performance to identify the presence of advanced fibrosis in NAFLD, many biomarkers, including CK18, have been used in combination with other clinical parameters to improve their diagnostic utility (Table 1). The reproducibility and generalizability of many of these scoring systems, including the enhanced liver fibrosis (ELF) panel, are unclear as study populations have been limited to certain ethnic groups or populations (ie, severely obese/bariatric patients). For example, the ELF panel, which incorporates plasma levels of matrix turnover proteins including hyaluronic acid, demonstrated good accuracy in distinguishing those with severe fibrosis (AUROC = 0.90, sensitivity = 80%, specificity = 90%). However, the study population included patients with a mean BMI of 32. Similar to the Hepatitis C Fibrosure (Laboratory Corporation of America), other biomarker panels have been studied in NAFLD patients, including NASHtest and NASH Fibrosure, yielding AUROCs of 0.79 and 0.81 for detection of NASH and at least moderate fibrosis (F>2), respectively. These panels were studied in patients with high BMI (>27) and were validated in cohorts limited to bariatric patients with morbid obesity.

As shown in a large meta-analysis by Musso et al, the NFS is one of the most extensively validated panels used to identify the presence of advanced fibrosis (AUROC = 0.85). This score incorporates six readily available, clinical variables including age, BMI, hyperglycemia, platelet count, albumin, and the AST/ALT ratio. AASLD guidelines recommend the use of the NFS and the presence of the metabolic syndrome to identify individuals at an increased risk for advanced fibrosis and in whom a liver biopsy might be indicated. The utility of NFS in certain populations, including patients without diabetes and without obesity (discussed below), may be limited and results in these populations should be interpreted cautiously. Furthermore, meta-analysis data showed that up to 58% of patients with NAFLD are classified as indeterminate for advanced fibrosis, which could result in a diagnostic dilemma for physicians.

Data suggest that combining the results of non-invasive modalities, such as TE in combination with peripheral markers, may improve their diagnostic yields compared with each modality individually. TE, when used in combination with NFS, more accurately identified those with severe liver fibrosis compared with TE in combination with FIB4, a panel that combines age, ALT, AST and platelet count, or to either modality alone. On the other hand, other studies have challenged the diagnostic accuracy of NFS. Since type 2 diabetes is a recognized metabolic risk factor for progressive NAFLD, the NASH Clinical Research Network sought to develop a predictive model in 435 biopsy-proven, NAFLD patients with diabetes to identify those with NASH and advanced fibrosis. The proposed model included several biochemical and clinical variables, including BMI, Hispanic ethnicity, ALT/AST ratio and bilirubin levels, and was shown to perform better than the NFS in the prediction of advanced fibrosis (AUROC 0.80 vs. 0.76, respectively; P < 0.05).

Non-Invasive Assessment of NAFLD in Non-obese Population

In non-obese individuals, NAFLD risk factors have been somewhat elusive making it more difficult to identify those with a higher likelihood of progressive liver disease. Since NFS incorporates BMI and the presence of diabetes, its utility may be limited in normoglycemic individuals with low BMI, leading to falsely lower scores in these individuals. Although the same BMI benchmarks may not apply uniformly to all ethnic and racial groups, insulin resistance and increased waist circumference are still thought to be potential risk factors for NAFLD. Wei et al analyzed a cohort of 911 Chinese individuals and identified 262 patients with NAFLD, of whom 135 were non-obese (prevalence 19.3%), defined by BMI <25. All NAFLD patients had similar intrahepatic triglyceride content, measured by proton-MRS, however increased plasma triglycerides, increased prevalence of insulin resistance (measured by HOMA-IR) and the PNPLA3 G allele variant were among the independent risk factors associated with non-obese NAFLD patients (P < .001). Non-obese NAFLD patients with normoglycemia were also found to be more insulin resistant than those without NAFLD (P < .001).

Recently, Rosso et al showed that the oral glucose sensitivity index (OGIS), a measurement of peripheral insulin resistance, was able to distinguish F2 from F3/4 fibrosis (P < .003) in a cohort of non-diabetic patients with NAFLD (n = 145). Interestingly, when compared with NFS, OGIS was a better predictor of advanced fibrosis (F3/4) (OR = 0.57; P < .001). The authors concluded that OGIS should be incorporated into the clinical assessment of patients with NAFLD depending on their phenotype. Although further studies are needed to determine the best diagnostic approach, data support a significant role for insulin resistance in NAFLD progression among both obese and non-obese patient populations.

Conclusion, Future Directions

NAFLD/NASH is an important cause of chronic liver disease and cirrhosis in which both environmental and genetic factors likely play important roles in disease progression. In view of the rising prevalence of NAFLD/NASH and the prediction that it will become the leading indication for liver transplantation, there is an urgent need to identify those with either advanced fibrosis or cirrhosis as well as those who are at risk for progressive fibrosis. Another diagnostic challenge is the differentiation between simple steatosis and NASH without reliance on liver biopsy. To accomplish both of these goals, and to expand the ability of primary care physicians to play an important role in the routine diagnosis of NAFLD/NASH, simple, cost-effective, accurate non-invasive modalities are required. The use of the NFS, especially in the presence of metabolic syndrome, has been endorsed by professional societies to identify those at greatest risk for fibrosis progression. Other studies, however, have documented improved performance of other noninvasive assays (ie, OGIS) compared with the NFS. Although numerous individual and combinations of biomarkers have been proposed, additional studies are needed to determine their role in the routine diagnosis of NAFLD/NASH particularly in patients without the typical phenotype. Future studies will also aim to identify the role of these modalities in monitoring disease progression, in predicting clinical outcomes and in response to therapies. Given the plethora of investigational agents presently under study, these modalities are likely to have an increasing role in patient management in the very near future.

• References:
• Angulo P et al. Hepatology. 2007;45(4);8346-8354.
• Blachier M et al. J Hepatol. 2013;58(3):593-608.
• Bazick J et al. Diabetes Care. 2015;38(7):1347-1355.
• Berres ML et al. Dig Dis. 2010;28(1):192-196.
• Bertola A et al. PLoS ONE. 2010;5(10):e13577.
• Blachier M et al. J Hepatol. 2013 Mar;58(3):593-608
• Chalasani N. Gastroenterology. 2012;142(7):1592-1609.
• Charlton MR et al. Gastroenterology. 2011 Oct;141(4):1249-1253.
• Chitturi S et al. J Gastroenterol Hepatol. 2011;26 Suppl 1:163-172.
• Cusi K et al. J Hepatol. 2014;60(1):167-174.
• Dietrich P et al. Best Prac Res Clin Gastroenterol. 2014;28:637-653.
• Duarte MS et al. FEBS Journal. 2014.
• Deurenberg P et al. Obes Rev. 2002;3(3):141-146.
• Deurenberg P et al. Int J Obes Relat Metab Disord. 1998;22(12):1164–1171.
• Festi D et al. Aliment Pharmacol Ther. 2013;37:392-400.
• Fujimoto WY et al. Obes Res. 1995;3 Suppl 2:179S–186S
• Gori M et al. Biomed Res Int. 2014;741465.
• Guha I et al. Hepatology. 2008;47:455-460.
• Harrison SA et al. Gut. 2008;57:1441-1447.
• Haukeland JW et al. J Hepatol. 2006;44:1167-1174.
• Honda Y et al. Hepatol Res. 2016 (Epub ahead of print).
• Kanda H et al. J Clin Invest. 2006;116:1494-1505.
• Karlas et al. PLoS One. 2014;9(3):e91987.
• Kawada N et al. J Gastroenterol. 2009;44:1190-1194.
• Kim NH et al. Liver Int. 2014;34(4):604-611.
• Kumar R et al. Dig Dis Sci. 2013;58:265-274.
• Kwok R et al. Aliment Pharmacol. 2014. 39(3):254-69.
• Lassailly G et al. Europ J Gastro Hepatol. 2011;23(6):499-506.
• Liu YL et al. J Hepatol. 2014; 6(1): 75-81.
• Lim JW et al. World J Gastroenterol. 2014;20(26):8325-40.
• Loomba R et al. Hepatology. 2012;56(3):943-951.
• Lovejoy JC et al. Metabolism. 1996;45(9):1119-1124.
• Macchado MV et al. Ann Gastroenterol. 2012;25(1):1-2.
• Miller M et al. J Gastroenterol Hepatol. 2014;29(10):1839-1847.
• Musso G et al. Ann Med. 2011;43(8):617-649.
• Ohlson LO et al. Diabetes. 1985;34(10):1055-1058.
• Palaniappan LP et al. Int J Obes (Lond). 2011;35:393-400.
• Park YW et al. Obes Res. 2001;9(7):381-387.
• Petta D et al. Liver Int. 2014.
• Palmer C et al. Gut. 2010;59(3):397-404.
• Pearce SG et al. Biomarker Research. 2013;1:7,1-11.
• Rajpathak SN et al. Ethn Dis. 2010; Summer;20(3):225-230.
• Ratziu V et al. BMC Gastroenterol. 2006;6:6.
• Reeder SB et al. Magn Reson Imaging Clin N Am. 2010;18(3):337-357.
• Rosso C et al. Hepatology. 2016;63:107-116.
• Shah A et al. Clin Gastroenterol Hepatol. 2009;7:1104-1112.
• Shen J et al. Aliment Pharmacol Ther. 2014;39(5):532-539.
• Singh S et al. Clin Gastroenterol Hepatol. 2015;13(4):643-654.
• Spengler EK et al. Mayo Clin Proc. 2015:90(9):1233-1246.
• Starley BQ et al. Hepatology. 2010;51(5):1820-1832.
• Stern MP et al. Arteriosclerosis. 1986;6(2):123-130.
• Sumida Y et al. World J Gastroenterol. 2014;20(2):475-485.
• Verma S et al. Liver Int. 2013;33:1398-1405.
• Vernon G et al. Aliment Pharmacol. 2011;34(3):274-285.
• Wei JL et al. Am J Gastroenterol. 2015;110(9):1306-1314.
• Wieland AC et al. Dig Dis Sci. 2013;58:2809-2816.
• Wree A et al. Nat Rev Gastroenterol Hepatol. 2013;10(11):627-636.
• Wood KL et al. BMJ Open Gastroenterol. 2015.
• Younossi ZM et al. Obes Surg. 2011;21:431-439.

• For more information:
• Alia Hasham, MD, is a third-year fellow at Jacobs School of Medicine and Biomedical Sciences at University at Buffalo, specializing in gastroenterology and hepatology. She can be reached at aliahash@buffalo.edu.
• Andrew Talal, MD, MPH, is a professor of medicine at Jacobs School of Medicine and Biomedical Sciences at University at Buffalo. He can be reached at the Clinical and Translational Research Center, 875 Ellicott Street, Suite 6089, Buffalo, NY 14203; email: ahtalal@buffalo.edu.

Disclosures: Hasham reports no relevant financial disclosures. Talal reports research and grant support from Abbott, AbbVie, Galactin, Gilead Sciences, Intercept Pharmaceuticals, Lily, Merck, Tibotec and Tobira; serving on committee for or as advisor for Abbott Diagnostics, AbbVie, the Chronic Liver Disease Foundation, Merck and Pfizer; and serving on the speakers bureau for the Chronic Liver Disease Foundation and the Empire Liver Disease Foundation.