Noninvasive diagnosis and monitoring of nonalcoholic steatohepatitis: Present and future


  • Anna Wieckowska,

    1. Departments of Gastroenterology and Hepatology, Cleveland Clinic Foundation, Cleveland, Ohio
    Search for more papers by this author
  • Arthur J. McCullough,

    1. Departments of Gastroenterology and Hepatology, Cleveland Clinic Foundation, Cleveland, Ohio
    Search for more papers by this author
  • Ariel E. Feldstein

    Corresponding author
    1. Pediatric Gastroenterology, Cleveland Clinic Foundation, Cleveland, Ohio
    2. Cell Biology, Cleveland Clinic Foundation, Cleveland, Ohio
    • Departments of Pediatric Gastroenterology and Cell Biology, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195
    Search for more papers by this author
    • Fax: 216-444-2974

  • Potential conflict of interest: Nothing to report.


Nonalcoholic fatty liver disease (NAFLD) is the most common chronic liver disease in the United States, and its prevalence is increasing worldwide. It currently affects approximately 30% of adults and 10% of children in the United States. NAFLD represents a wide spectrum of conditions ranging from simple fatty liver which in general follows a benign nonprogressive clinical course, to nonalcoholic steatohepatitis (NASH), which is a more serious form of NAFLD that may progress to cirrhosis and end-stage liver disease. At present, a liver biopsy remains the only reliable way to diagnose NASH and establish the presence of fibrosis. Current noninvasive clinically available tests lack accuracy and reliability. In light of the dramatic increase in the prevalence of NAFLD in conjunction with the significant research effort in developing novel therapies for patients with NASH, noninvasive, simple, reproducible, and reliable biomarkers are greatly needed. They will not only help in the diagnosis of NASH, but also be useful for assessment of treatment response and prognosis and remain a research priority in the NAFLD field. (HEPATOLOGY 2007;46:582–589.)

Several factors have converged to make the development and validation of noninvasive biomarkers for nonalcoholic fatty liver disease (NAFLD) diagnosis an area of increasing interest in hepatology. First, the prevalence of NAFLD has grown to epidemic proportions over the last 2 decades, with current estimates of 80 million to 100 million Americans affected.1, 2 Second, long-term longitudinal studies now show 2 clearly demarcated natural history patterns of the disease.3, 4 Simple fatty liver or NAFL, the most common form of NAFLD representing about 80%-90% of cases, follows a remarkably benign clinical course, whereas nonalcoholic steatohepatitis (NASH) is a potentially serious condition associated with a significant increase in overall and liver-related morbidity and mortality.3, 4 Third, the currently available noninvasive tests lack sensitivity and specificity and have limited utility in general. Thus, liver biopsy remains the only reliable way of diagnosing and staging NASH. A liver biopsy provides important information regarding the degree of liver damage, changes in the overall liver architecture, as well as severity of inflammatory activity and fibrosis.5 However, it is obvious that this invasive procedure is not suitable as a screening test for a condition that affects a third of the American population. Finally, although no medication has yet been proven to be effective in stopping and/or causing disease regression in NASH patients,6 new promising therapies are currently being tested in large clinical trials. This makes us predict that clinicians will soon face an increasing array of treatment options directed to those patients with NASH. For all of these reasons, there is an urgent need to develop and validate simple, reproducible, noninvasive tests that accurately distinguish NASH from NAFL and determine the stage and grade of the disease. Such a test would not only aid clinicians in the identification of patients with NASH, but also allow for noninvasive frequent monitoring of disease progression and response to therapy. This review gives an overview of potential novel biomarkers for NASH diagnosis and disease staging.

Currently Available Noninvasive Tools

Most patients with NAFLD, including adults and children with either NAFL or NASH, are asymptomatic at presentation. When present, clinical symptoms and physical findings are nonspecific and unreliable in assessing disease severity in patients with compensated liver disease. The most common signs and symptoms are fatigue, right upper quadrant pain, hepatomegaly, and acanthosis nigricans, which tends to be more frequently seen in the pediatric population. Several clinical features as well as historical data have been shown to be independent risk factors for NASH and presence of fibrosis. The features more consistently found to be associated with disease severity include obesity, older age, diabetes, and hypertension. However, the accuracy and utility of these variables to diagnose NASH and determine the presence of fibrosis have only been assessed in a relatively small number of studies, in which specific clinical models or algorithms have been created combining these variables with several biological markers (see later).

Laboratory tests that are routinely included in the evaluation of patients with suspected NAFLD include a serum panel of liver tests (alanine aminotransferase [ALT], aspartate aminotransferase [AST], alkaline phosphatase [ALP], gamma-glutamyl-transpeptidase [GGT], albumin), prothrombin time, and complete blood count. Elevated serum ALT and AST levels are the primary abnormality seen in patients with NAFLD and tend to be higher in patients with NASH as compared to NAFL. However, liver aminotransferase levels are seldom higher than 5 times the upper limit of normal, and typically fluctuate with normal levels seen in more than two-thirds of NASH patients at any given time.7, 8 Moreover, a study by Mofrad et al. demonstrated that the entire histological spectrum of NAFLD can be seen in patients with normal ALT values.9 A subsequent study by Kunde et al. formally evaluated the diagnostic accuracy of serum ALT for NASH diagnosis in a large cohort of women undergoing gastric bypass surgery.10 The sensitivity for NASH diagnosis was found to be quite poor, at about 40%. Decreasing the upper limit of normal, as has been suggested by many, improved the sensitivity but resulted in very high false positive rates.

Serum ALP, GGT, or both are usually mildly elevated in many patients with NAFLD. However, their utility for the diagnosis of NASH is poor.1 Among other routine laboratory tests, a reversal of the AST/ALT ratio to more than 1 had been consistently reported to predict the presence of more advanced fibrosis.11 This phenomenon is also true in a variety of other chronic liver diseases including chronic hepatitis C. However, in a similar fashion as the isolated aminotransferase levels, it has a relatively poor sensitivity and negative predictive value. Hypoalbuminemia, prolonged prothrombin time, and hyperbilirubinemia may be seen with cirrhotic NASH but are not present until decompensated disease arises.

The most commonly used imaging technique in the diagnostic work-up of patients with suspected NAFLD is ultrasonography (US), in which fatty infiltration of the liver produces a diffuse increase in echogenicity of the liver parenchyma. Several studies demonstrated that the sensitivity, specificity, and positive predictive value of this technique to detect steatosis is as high as 80%-100%. However, accurate quantification of the degree of steatosis is not feasible with the current technology. Although US is useful in the detection of severe steatosis, its sensitivity decreases sharply if the degree of steatosis is 30% or less.12 Moreover, in patients with morbid obesity, a sensitivity lower than 40% has been reported,13 likely due to technical difficulties in performing US in such patients. Alternate US parameters including hepatic vein Doppler waveform and the presence of focal hypoechoic areas within the liver hilus have been evaluated as predictors of liver steatosis. However, similar to regular US, the lack of signs of steatosis has a low accuracy in excluding mild degree of fat infiltration.

Both computerized tomographic scanning and, in particular, magnetic resonance imaging seem to be more sensitive techniques for quantification of steatosis. However, none of these imaging techniques have sufficient sensitivity and specificity for staging the disease and cannot distinguish between simple steatosis and NASH with or without fibrosis.


ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; AUC, area under the curve; BMI, body mass index; CCL2, CC-chemokine ligand-2; CI, confidence interval; CK, cytokeratin; CRP, C-reactive protein; ECM, extracellular matrix; FT, FibroTest; GGT, gamma-glutamyl-transpeptidase; HA, hyaluronic acid; IL-6, interleukin-6; LDL, low-density lipoprotein; NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis; OS, oxidative stress; ROS, reactive oxygen species; TAR, total antioxidant response; TBARS, thiobarbituric acid-reacting substance; TNF, tumor necrosis factor; US, ultrasonography.

Biomarkers Under Investigation

Due to the important limitations of currently available noninvasive tests, several investigators have tried to identify potential novel biomarkers based on the current knowledge of the pathophysiologic mechanisms involved in disease progression in NAFLD (see below and Fig. 1). An ideal biomarker should be simple, reproducible, inexpensive, readily available, and accurate for a particular disease process (Table 1). In the case of NAFLD, such a biomarker should also be particularly able to distinguish NASH from NAFL, determine the extent of liver fibrosis present, predict risk of disease progression (mainly tight to fibrosis progression), and monitor response to therapeutic interventions. Identification of such a biomarker may provide as a new tool in the development of effective novel therapeutics and the identification of specific populations that would most benefit from a particular treatment. Unfortunately, to date, none of the available markers fulfill all these criteria. Most have been studied in small single-center “proof of concept” studies with the independence of its association with either NASH or extent of fibrosis as the main endpoint, with only a limited number of studies performing C-statistics and documenting the reliability and accuracy of a particular test. Moreover, a significant confounding factor in all these studies is that a liver biopsy has been used as the gold standard for assessing the utility of particular biomarkers. However, as it is the case in other chronic liver diseases such as hepatitis C, liver biopsy in NAFLD is also subject to sampling error.14, 15 Finally, the published data on these biomarkers do not allow for comparisons due to the different scoring systems used to diagnose NASH and stage of fibrosis, the different populations examined, and the different methods of analysis used.

Figure 1.

Potential rational targets for biomarkers development in NASH based on the pathophysiology of the disease. Hepatocyte steatosis is a prerequisite for the subsequent events that lead to liver injury and fibrogenesis. Mitochondrial dysfunction is thought to play a central role in the progression from steatosis to NASH to cirrhosis. Different mechanisms have been proposed including (1) an increased production of reactive oxygen species, and mitochondrial outer membrane permeabilization, resulting in a cascade of events leading to (2) inflammation, (3) hepatocellular apoptosis, and (4) fibrogenesis and fibrosis. ROS, reactive oxygene species; TNF-α, tumor necrosis factor-α; CRP, C-reactive protein; CCL2, CC-chemokine ligand-2; FasL, Fas ligand; RBP-4, retinol binding protein-4; HSC, activated stellate cell; TGF-β, tumor growth factor-β; HA, hyaluronic acid; ECM, extracellular matrix; CK-18, cytokeratin-18. Question marks are placed on potential biomarkers in which no data is currently available.

Table 1. Characteristics of an Ideal Novel Biomarker for NAFLD
Provides incremental information to currently available diagnostic tools
Easy to measure and handle
Accurate for diagnosis of NASH or stage of fibrosis
Accurate for risk stratification and monitoring response to therapy
Validated in multiple large and prospective trials

Biomarkers of Oxidative Stress

Oxidative stress (OS), characterized by an imbalance between pro-oxidant and antioxidant mechanisms in favor of the former, has long been recognized as a key mechanism responsible for liver damage and disease progression in NAFLD. Enhanced OS occurs in the liver of patients with NASH, as well as in animal models of NASH. Several oxidation pathways may play a role in the overproduction of reactive oxygen species (ROS) in NASH including mitochondrial, peroxisomal, cytochrome P-450, myeloperoxidase, and nitric oxide synthase. Each of these pathways may generate different oxidation products which could be potentially quantified. Because of the importance of OS in the pathogenesis of NASH, several groups have attempted to elucidate whether measurement of systemic markers of OS may reflect the levels of OS present in the liver. Most groups used either method to measure systemic levels of stable lipid by-products of ROS activity such as lipid peroxides and thiobarbituric acid-reacting substance (TBARS) or “total antioxidant status”. Many questions remain unanswered such as the relative importance of each of these oxidation pathways in the liver damage seen in NASH. As ROS react rapidly and in situ in the environment where they are produced, does measuring these markers in blood or breath reflect what is happening in the liver?

Chalasani et al.16 tested the blood levels of oxidized low-density lipoprotein (ox-LDL) and TBARS in a small group of patients with biopsy-proven NASH and age-matched, gender-matched, and body mass index (BMI)-matched controls. Both ox-LDL and TBARS were significantly higher in NASH patients on the univariate analysis. However, these associations disappeared on the stepwise regression analysis. In another small study, total antioxidant response (TAR) and total lipid peroxide levels were measured in biopsy-proven NASH patients and healthy controls by using a novel automated colorimetric method.17 In this study, the tests were performed within 1 year of the liver biopsy used for comparison. The TAR was found to be significantly lower, whereas total plasma peroxide levels were significantly higher in patients with NASH versus control patients. Moreover, fibrosis scores correlated with both plasma peroxides and TAR in a positive and negative fashion, respectively. Bonnefont-Rousselot and colleagues18 assessed blood OS status by testing several parameters: TBARS, plasma vitamin E levels, glutathione peroxidase (GSH-Px) activity, erythrocyte GSH-Px activity, and Cu-to-Zn superoxide dismutase (SOD) activities. A blood test was performed the day of the liver biopsy on 64 patients with suspected NAFLD or viral hepatitis. No correlation was found between these OS markers and any histological feature in patients with NAFLD. More recently, Solga et al.19 measured breath ethane, an investigational marker of OS, in patients presenting for gastric bypass surgery. Subjects with NASH did not have higher breath ethane than those without NASH, and there was no correlation between the breath test and hepatic steatosis, fibrosis, or elevated AST or ALT levels. Thus, although the existence of OS in the liver of patients with NASH has been clearly demonstrated, the results of the current studies attempting to noninvasively assess OS are mixed, and future studies will need to take a multimodal approach to better characterize which OS biomarkers are better suited for NASH diagnosis.

Biomarkers of Inflammation

A growing body of evidence supports a central role for inflammation and inflammatory cytokines in the development of the metabolic syndrome and NAFLD. A cytokine imbalance, and in particular an increase in the tumor necrosis factor α (TNF-α)/adiponectin ratio may play an important role in the development of NASH. Treatment of leptin-deficient ob/ob mice with antibodies against TNF-α improves NASH and hepatic insulin resistance.20 Mice genetically deficient in TNF type 1 receptor (TNFR1) are resistant to diet induced NAFLD.21 In addition, injection of recombinant adiponectin to ob/ob mice, decreases TNF-α levels and reverses steatohepatitis.22 Several groups have investigated the circulating levels of these cytokines in blood from patients with NAFLD and their correlation with disease severity. However, there is currently no data available on the accuracy and clinical usefulness of these markers for noninvasive NASH diagnosis. Hui et al.23 quantified serum levels of both TNF-α and adiponectin in patients with biopsy proven NAFLD. Adiponectin levels were significantly lower and TNF-α levels were significantly higher in patients with NASH as compared to controls. However, when compared to simple steatosis, only adiponectin levels remained significantly lower in NASH patients whereas no difference was noted in TNF-α levels between the 2 groups of patients. Similarly, in a study from Italy, Musso et al.24 showed that serum adiponectin levels were significantly lower in patients with NASH as compared to age-matched, sex-matched, and BMI-matched controls, whereas there were no differences in serum TNF-α values. In contrast to these studies, others reported that serum TNF-α levels are higher in patients with NASH as compared to both patients with simple steatosis and controls with no known liver disease.25, 26 Several reasons might explain these discrepancies in the results. These include the fact that in humans TNF-α has a relative short half-life and low circulating levels, which may not reflect the changes occurring in the liver tissue, as well as the sensitivities of the assays used, the differences in the population tested, and the lack of adjustment for several factors that may influence the TNF-α circulating levels.

Two other inflammatory markers, interleukin-6 (IL-6) and C-reactive protein (CRP), which have been extensively studied as markers for insulin resistance, type 2 diabetes, and cardiovascular disease, have been tested in patients with NAFLD. CRP is an acute phase-reactant synthesized by the liver that is also elevated in chronic inflammatory states. CRP levels have been shown to be closely related to obesity, in particular central or visceral fat deposition. Traditional assays were not sufficiently sensitive to monitor low levels of CRP elevation and therefore high-sensitivity assays have been developed (hsCRP). In a study from Australia, Hui et al.27 found no difference in the mean hsCRP levels between patients with NASH and those with simple steatosis, and no association between hsCRP levels and any histological feature. In another study, hsCRP, IL-6, and several chemokines, including CC-chemokine ligand-2 (CCL2), were measured in 47 patients with biopsy-proven NAFLD as well as 30 age-matched, sex-matched, and ethnicity-matched controls.28 Both IL-6 and CCL2 but not hsCRP were increased in patients with NAFLD as compared to controls. Moreover, only CCL2 levels were different between patients with NASH and those with simple steatosis. Larger future studies are needed to better characterize the role of these as well as other cytokines/chemokines such as IL-1β, macrophage inflammatory proteins, adipokines such as resistin, and the recently identified visfatin29 and retinol binding protein-4 (increasingly recognized to play an important role in insulin resistance and the associated metabolic syndrome),30 as potential biomarkers for NASH diagnosis in isolation or in combination as diagnostic panels.

Biomarkers of Hepatocyte Apoptosis

Emerging data suggest hepatocyte apoptosis, a highly organized and genetically controlled form of cell death, may play an important role in liver injury and disease progression in NAFLD.31, 32 An increase in hepatocyte cell death by apoptosis is typically present in humans with NASH as well as animal models of NASH but absent in those with simple steatosis.33 Increasing evidence suggests a role of both the so-called extrinsic (death receptor–mediated) pathway and the intrinsic (organelle-initiated) pathway of apoptosis. Expression of Fas, a death receptor member of the TNFR family, increases in experimental models of NASH and results in increased sensitivity to Fas-mediated apoptosis.34 Accumulation of free fatty acids in the liver cells results in lysosomal and mitochondrial permeabilization. Mice genetically deficient in cathepsin B, a major lysosomal cysteine protease, are protected against diet-induced steatosis and liver injury.21, 35 Although the relative importance of each of these pathways in human NAFLD remains to be elucidated, in hepatocytes, both pathways tend to converge at the level of the mitochondria resulting in permeabilization of the mitochondrial outer membrane and release of multiple proteins from the mitochondrial intermembrane space into the cytosol. A central consequence of this process is the activation of the effector caspases (mainly caspase 3) which cleave a number of different substrates inside the cell, including cytokeratin 18 (CK-18), the major intermediate filament protein in the liver, resulting in the characteristic morphologic changes of apoptosis. Recently, caspase-generated CK-18 fragments were tested in the livers as well as in plasma from patients undergoing a liver biopsy for suspected NAFLD and in healthy age-matched controls.36 CK-18 fragments were significantly elevated in the NAFLD patients as compared to controls, and the plasma levels correlated with the expression levels in the liver. Moreover, plasma CK-18 fragments independently predicted NASH on multivariate analysis. The area under the curve (AUC) was estimated to be 0.93 (95% confidence interval [CI]: 0.83-1.00). A cutoff value of 380 U/l gave a specificity for NASH diagnosis of 94% and a sensitivity of 90.5%. A large multicenter validation study is currently underway and future longitudinal studies are warranted to determine the utility of this marker to predict fibrosis progression and response to therapy over time.

Biomarkers of Fibrosis

Liver fibrosis represents the most worrisome histopathological feature in patients with NAFLD, because it suggests a more severe and progressive liver damage. Thus, as in other liver diseases, accurate assessment of fibrosis extent is essential for patients with NAFLD. Several groups created panel markers using combinations of clinical and biochemical tests to generate clinical models of fibrosis, whereas others focused on specific markers of fibrosis either in isolation or in multicomponent tests. Most of these studies involve patients with chronic hepatitis C. We will only review the markers that have been specifically tested in the context of NAFLD (Table 2). For a more extended discussion, please refer to a recently published thorough review.37

Table 2. Combined Markers of Fibrosis/Fibrogenesis in NAFLD
TestStaging systemnAUC for advanced fibrosisComponents
  • *

    Training set.

  • Validation set.

  • Components tested include HA, hyaluronic acid, collagen IV, collagen VI, TIMP-1, tissue 1 inhibitor of metalloproteinase; PIIINP, aminoterminal peptide of procollagen III.

  • Abbreviations: AST, aspartate aminotransferase; ALT, alanine aminotransferase; BMI, body mass index; AUC, area under the curve; CI, confidence interval; ECM, extracellular matrix; N/S, not stated.

BAAT scoreMetavir930.84 (CI: N/S)Age, BMI, ALT, Serum Triglycerides
FibroTestModified Brunt2670.86 (CI: 0.77, 0.91)* 0.75 (CI: 0.61, 0.83)α2-macroglobulin, apolipoprotein A1, haptoglobulin, total bilirubin, GGT
NAFLD fibrosis scoreModified Brunt7330.88 (CI: 0.85, 0.92)* 0.82 (CI: 0.76, 0.88)Age, BMI, platelet count, albumin, AST/ALT ratio, IFG/diabetes
ELFScheuer610.87 (CI: 0.66, 1.0)Combination of multiple ECM proteins and proteinases

In general, all these noninvasive markers perform well and show similar accuracy to detect advanced, severe fibrosis, but they have low negative predictive values for the presence of mild to moderate fibrosis. This represents an important limitation of these biomarkers as a screening test for patients with NAFLD. Ideally, detection of the earliest stages of fibrosis would allow initiation of therapeutic interventions before the development of advanced precirrhotic or cirrhotic stages of the disease. Moreover, most panels use a high and low cutoff value which results in significant negative and positive predictive values, respectively. However, approximately one-third of patients typically have values that fall between these 2 cutoffs, and thus the presence or absence of advanced fibrosis cannot be predicted.

Ratziu et al.38 combined 4 clinical variables to generate the BAAT score: BMI (≥28 kg/m2), age (≥50 years), ALT (≥2× normal), and serum triglycerides (≥1.7 mmol/l). The presence of each variable gives 1 point in the combined score. Whereas a total score of 0 or 1 had a negative predictive value of 100% for fibrosis, a high total score of 4 gave a sensitivity of 14% and a specificity of 100% for detection of septal fibrosis. More recently, the same group tested the utility of the FibroTest (FT) for prediction of advanced liver fibrosis in patients with NAFLD.39 This proprietary panel has been extensively studied in chronic hepatitis C and combines 5 biochemical markers including α2-macroglobulin, apolipoprotein A1, haptoglobulin, total bilirubin, and GGT. The score is computed via an undisclosed formula by entering a patient's age and sex along with the 5 components into a proprietary program. The test yields an AUC of 0.86 (95% CI:0.77-0.91) for the diagnosis of advanced fibrosis. The most frequent causes of FT failure include Gilbert's syndrome, cholestasis and acute inflammation, which result in increases in bilirubin and haptoglobulin, respectively. Another cause of FT failure in patients with NAFLD which was not recognized previously in other populations is an abnormal apolipoprotein A1 concentration which might be related to the frequent lipid abnormalities seen in these patients. Angulo and colleagues40 recently created a NAFLD fibrosis score to separate patients with or without advanced fibrosis in a large cohort of biopsy-proven NAFLD patients. An algorithm was constructed using 6 readily available laboratory and clinical variables including age, hyperglycemia, BMI, platelet count, albumin, and AST/ALT ratio. The AUC for this model was 0.88 and 0.82 in the estimation and validation groups, respectively. Using this curve, they generated 2 (high and low) cutoff values that allow the diagnosis of advanced fibrosis with high accuracy (negative predictive value between 88% and 93%, PPV between 82% and 90%). However, similarly to the FT study in which 33% of cases the presence or absence of advanced fibrosis could not be predicted, in this series there were 25% indeterminate cases.

Fibrosis is a dynamic process which may result in increased circulating levels of extracellular matrix (ECM) components. Several groups have used this reasoning to develop different blood tests using individual or a composite of ECM components. Suzuki et al.41 determined the reliability of serum hyaluronic acid (HA) to predict the severity of hepatic fibrosis in 79 patients with histologically confirmed NAFLD. Serum HA was obtained at the time of liver biopsy. The logarithm of serum HA showed a significant positive correlation with the stage of fibrosis, and this association persisted after adjusting for age and serum albumin. The test was found to be useful for predicting severe fibrosis (stages 3-4) with AUC of 0.9 (95% CI: 0.83, 0.97). The study could not evaluate the efficacy for moderate fibrosis (stage 2) due to the limited number of patients with this stage of fibrosis and showed low accuracy for mild fibrosis. Lydatakis et al.42 subsequently studied the utility of serum HA as well as laminin in 50 patients with NASH. Twenty-three of these patients had some degree of fibrosis and 27 had no evidence of fibrosis on liver biopsy. Both serum HA and laminin were significantly higher in patients with NASH-fibrosis versus those without fibrosis. However, only the mean concentration of serum HA was found to be significantly different among patients with various stages of fibrosis. AUC values were calculated to distinguish fibrotic-NASH versus NASH without fibrosis showing high accuracy. In contrast to the study by Suzuki,41 no analysis was performed to assess the accuracy of the test to separate different stages of fibrosis. Moreover, the proposed cutoff value to predict fibrosis significantly differed between the 2 abovementioned studies. This lack of reproducibility may be due to many factors including differences in the population tested and differences in assay systems. Further larger studies with sufficient number of patients with different stages of fibrosis are still required to better address both the accuracy and reproducibility of serum HA measurement as a marker of fibrosis in NAFLD. In a recent study, the European Liver Fibrosis study group examined a combination of multiple ECM-related components including HA, collagen IV, collagen VI, amino-terminal propeptide of type III collagen, tissue inhibitor of metalloproteinase 1, and others in 1021 subjects including 61 patients with NAFLD.43 An algorithm was developed to detect severe fibrosis in patients with NAFLD with a sensitivity of 89% and a negative predictive value for absence of fibrosis of 98%. The AUC for severe fibrosis was 0.87 (95% CI: 0.66, 1.0). However, similar to the other biomarkers as well as the panel markers for diagnosis of liver fibrosis, the test showed poor performance in predicting early stages of fibrosis (stages 1-2).

These studies suggest that either a combination of clinical and biochemical markers or specific markers of fibrosis may be used for noninvasive staging of NAFLD patients. However, to date, none of these markers have been independently validated in different populations in a prospective fashion. Moreover, all of these studies have been tested in a cross-sectional fashion, and the role of these biomarkers for monitoring disease progression, response to therapy, and prognosis remains completely unknown.

Recently, transient elastography, a technique based on US technology for measuring tissue elasticity, has been applied for noninvasive assessment of hepatic fibrosis. Promising results have been shown on initial studies in patients with chronic hepatitis C with AUC, specificities, and sensitivities to detect advanced fibrosis similar to that of the FT and other markers of fibrosis.44 However, a recent study from France that evaluated transient elastography in more than 1000 males with mainly viral hepatitis (less than 4% of cases had NAFLD), showed that the only independent risk factor for failure of the procedure was a BMI greater than 28 (odds ratio:10; 95% CI, 5.7-17.9, P = 0.001),45 a common characteristic in the majority of patients with NAFLD. Furthermore, it remains to be determined whether this technique is able to differentiate fibrosis from significant steatosis. Thus, the utility of transient elastography for assessing fibrosis in patients with NAFLD remains uncertain and merits further investigation.

Future Directions and Novel Approaches for Biomarker Discovery

NAFLD has become a significant and challenging epidemic. Liver biopsy remains the current gold standard for the diagnosis of NASH and assessment of fibrosis extent. New accurate noninvasive methods that are reliable and readily available are greatly needed in order to determine treatment strategies, response to therapies, and prognosis.

As important progress is made in the elucidation of the pathogenesis of NAFLD, new rational noninvasive serum biomarkers that reflect the pathobiology of the disease, including markers of OS, inflammation, apoptosis, and fibrosis, are being tested. All these markers are in the initial phases of investigation, and the clinical utility of these tests remains to be determined. Prospective, independent validation studies in various populations and laboratories are still needed.

Promising new approaches that use proteomics, metabolomics, and genomics may help identify novel biomarkers that may drive clinical decision-making, supplementing or replacing currently available techniques. To date, only a few small, single-center pilot studies have used these technologies in NAFLD.46 These types of studies are critically dependent on the collection of a large number of well-characterized cases and controls which will most likely require multicenter collaborations. A good example comes from the recent study on genetic markers in more than 900 patients with chronic hepatitis C that showed 2 single-nucleotide polymorphisms significantly associated with advanced fibrosis.47 Finally, other exciting new technologies are those related to molecular imaging with the ability to noninvasively demonstrate both the level of a specific molecular target and the functional state of the target in vivo. These techniques may potentially allow direct determination of the extent of apoptosis, inflammation, and fibrosis in the liver of patients with NAFLD.