Hepatitis B: are non-invasive markers of liver fibrosis reliable?

Authors

  • Laurent Castera

    Corresponding author
    1. Service d'Hépatologie, Hôpital Beaujon, Assistance Publique-Hôpitaux de Paris, INSERM U773 CRB 3, Université Denis Diderot, Paris-7, Clichy, France
    • Correspondence

      Laurent Castera, Service d'Hépatologie, Hôpital Beaujon, Assistance Publique-Hôpitaux de Paris, 100 boulevard du General Leclerc, 92110 Clichy, France

      Tel: +33 1 40 87 57 64

      Fax: +33 1 40 87 51 09

      e-mail: laurent.castera@bjn.aphp.fr

    Search for more papers by this author

Abstract

Liver biopsy, which was traditionally considered to be the gold standard for the staging of fibrosis, has been challenged in the past decade by non-invasive techniques. These techniques rely on two distinct but complementary approaches: a ‘biological’ approach, based on the quantification of biomarkers of fibrosis in serum, and a ‘physical’ approach, based on the measurement of liver stiffness using elastography-based technologies. Advantages of serum biomarkers include their high applicability (>95%) and good reproducibility. However, as none are liver specific their results can be influenced by comorbid conditions (risk of false positive results with FibroTest® in patients with Gilbert's syndrome or with APRI in case of acute hepatitis). Transient elastograpy has the advantages of being a user's friendly procedure that can be performed at the bedside or in an outpatient clinic with high performance for detecting cirrhosis. However, its applicability is lower (80%) than that of serum biomarker (particularly in case of ascites, obesity and limited operator experience) with the risk of false positive results in case of ALT flares. Although these non-invasive methods were initially developed and validated in patients with chronic hepatitis C, they are now increasingly used in patients with hepatitis B, reducing the need for liver biopsy.

Abbreviations
HBeAg

hepatitis Be antigen

HBV

hepatitis B virus

HCC

hepatocellular carcinoma

NPV

negative predictive value

PPV

positive predictive value

ULN

upper limit of normal

The prognosis and management of patients with chronic hepatitis B virus (HBV) depend on the amount and progression of liver fibrosis and the risk of cirrhosis. The assessment of liver fibrosis is not only important to determine the prognosis of disease but also to identify patients who require antiviral therapy. For decades, liver biopsy was considered to be the gold standard for staging liver fibrosis using semiquantitative scoring [1]. There were two clinically relevant endpoints: (i) the detection of significant fibrosis (METAVIR F ≥ 2), which indicates that patients should receive antiviral treatment and (ii) the detection of cirrhosis (METAVIR F4), which indicates that patients should be monitored for complications related to portal hypertension and hepatocellular carcinoma (HCC) [2]. However, because of the limitations of liver biopsy as well as the development of powerful virological tools and new antiviral drugs, the use of this method in the management of patients with hepatitis B has rapidly decreased. In the past decade, non-invasive techniques have been developed to assess liver fibrosis [3]. Besides assessing liver fibrosis, these methods can be used to decide whether to treat a patient or defer antiviral treatment as well as to monitor response to treatment.

This article reviews existing non-invasive methods to assess liver fibrosis and discusses their advantages and disadvantages in the management of patients with HBV.

Currently available non-invasive methods

Non-invasive methods are based on two distinct approaches: a ‘biological’ approach (quantifying biomarkers in serum samples) or a ‘physical’ approach (measuring liver stiffness) (Table 1). Although these approaches are complementary, they rationale is different for each. Liver stiffness corresponds to an intrinsic physical property of the liver parenchyma, while serum biomarkers indicate several, non-liver-specific features in the blood that have been associated with the stage of fibrosis.

Table 1. Respective advantages and disadvantages of currently available non-invasive methods in patients with hepatitis B
 Serum biomarkersTransient elastography
Advantages

Good reproducibility High applicability (95%)

No cost and wide availability (non patented)

Well validated

Can be performed in the outpatient clinic

Liver stiffness is a genuine physical property of liver tissue

Good reproducibility

Well-validated

High performance for cirrhosis (AUROC >0.9)

Most widely used non-invasive technique

Can be performed at bedside or in the outpatient clinic

User-friendly (<5 min, immediate results, short learning curve)

Prognostic value in cirrhosis

Disadvantages

Non specific of the liver

Unable to discriminate between intermediate stages of fibrosis

Performance not as good as TE for cirrhosis

Results not immediately available

Cost and limited availability (proprietary)

Limitations (haemolysis, Gilbert syndrome, inflammation…)

Applicability (80%) lower than that of serum biomarkers: (ascites, obesity, limited operator experience)

False positive in case of acute hepatitis, extra-hepatic cholestasis and liver congestion

Unable to discriminate between intermediate stages of fibrosis

Requires a dedicated device

Region of interest cannot be chosen

Serum markers of liver fibrosis

Many serum biomarkers have been evaluated to determine the stage of liver fibrosis, mainly in patients with chronic hepatitis C [for review see [4]]. The so−called direct markers reflect the deposition or removal of the extracellular matrix in the liver. These include glycoproteins such as serum hyaluronate, laminin and YKL-40, collagens such as procollagen III N-peptide and type IV collagen, collagenases, and their inhibitors such as matrix metalloproteases and tissue inhibitory metalloprotease-1. The so−called indirect markers include factors that can be measured in routine blood tests, such as the prothrombin index, platelet count and ratio of aspartate to alanine aminotransferase (AST/ALT), which indicate alterations in hepatic function. The results from measurements of direct and indirect markers can be combined and used as a diagnostic tool. The FibroTest® (patented formula) (Biopredictive, Paris, France) was the first algorithm to combine these data in patients with hepatitis C [5] or in hepatitis B [6]. Non-patented methods use published models based on routinely available laboratory tests such as the Aspartate Platelet Ratio Index (APRI) [7]. Although many scores have been proposed for hepatitis C (summarized in [4], only two scores have been proposed specifically for hepatitis B [8, 9].

The practical advantages of analysing serum biomarkers to measure fibrosis include their high applicability (>95%) and inter-laboratory reproducibility [10], as well as their potential widespread availability (Table 1). However, because none are liver specific, their results can be influenced by comorbid conditions and they require critical interpretation of results. For instance the FibroTest® produces false positive results in patients with Gilbert's syndrome or haemolysis, because these patients have hyperbilirubinaemia, while acute hepatitis can produce false positive results in the APRI, which measures levels of aminotransferases.

Transient Elastography

Liver fibrosis can be staged using 1-dimensional ultrasound transient elastography (TE) (FibroScan®, Echosens, France) [11], which measures the velocity of a low-frequency (50 Hz) elastic shear wave propagating through the liver. This velocity is directly related to tissue stiffness, called the elastic modulus (expressed as E = 3ρv2, where v is the shear velocity and ρ is the density of tissue, assumed to be constant). The stiffer the tissue, the faster the shear wave propagates. TE measures liver stiffness in a volume that approximates a cylinder that is 1 cm wide and 4 cm long, 25–65 mm below skin surface. The results are expressed in kilopascals (kPa), and range from 2.5 to 75 kPa with normal value around 5 kPa [12].

Advantages to TE include a short procedure time (<5 min), immediate results, and a test that can be performed at the bedside or in an outpatient clinic that is not a difficult procedure to learn (Table 1). However, accurate results require careful interpretation of data, based on at least 10 validated measurements, a success rate (the ratio of valid measurements to the total number of measurement) above 60%, and an interquartile range (IQR, reflects variations among measurements) of less than 30% of the median value (IQR/LSM ≤30%) [12]. Although TE analysis has excellent inter- and intra-observer agreement [13], its applicability (80%) is not as good as that of serum biomarkers. Failure to obtain a measurement has been reported in 3% of cases and unreliable results (not meeting manufacturer's recommendations) have been reported in 16% [14], mostly because of patient obesity or limited operator experience. A new probe (XL) has been proposed to overcome the limitations in overweight and obese patients [15]. Failure (no valid shot) was significantly less frequent with the XL probe than with the standard M probe (1.1% vs 16%; < 0.00005). However, unreliable results were still observed with the XL probe in 25% of the cases compared with 50% with the M probe (P < 0.00005). Also, liver stiffness values measured by the XL probe are around 1.5 kPa lower than that by the M probe, thus appropriate cut-off values must still be defined and validated for the former.

Except for obese patients, TE results can also be difficult to obtain in patients with a narrow intercostal space and are impossible to obtain in patients with ascites [11]. The liver is an organ that is wrapped in a distensible but non-elastic envelope (Glisson's capsula), so additional space-occupying tissue abnormalities such as oedema, inflammation, extra-hepatic cholestasis or congestion can interfere with measurements of liver stiffness, independent from fibrosis [4].

Diagnostic performance

Serum markers of liver fibrosis

The Zeng score including alpha-2-macroglobulin, age, gamma glutamyl transpeptidase and hyaluronic acid detected significant fibrosis with an AUROC of 0.77 in hepatitis Be antigen (HBeAg)-positive Asian patients [9]. With a cut-off of <3.0, the presence of significant fibrosis could be excluded with a 90.9% negative predictive value (NPV) and 98.0% sensitivity. With a cut-off of >8.7 the presence of significant fibrosis could be correctly identified with a 84.8% positive predictive value (PPV) and 90.4% specificity. Overall, 35.5% of patients could be correctly classified and avoid liver biopsy. The Hui score including BMI, platelet count, albumin and total bilirubin detected significant fibrosis with an AUROC of 0.76 [8]. With a cut-off of <0.15, the presence of significant fibrosis could be excluded with a 92% NPV. Overall, 49% of patients could be correctly classified. Unfortunately these scores have not been externally validated. So far the most widely used and validated biomarkers are the APRI and the FibroTest®. Both are less accurate in detecting intermediate stages of fibrosis than cirrhosis. A meta-analysis by the developer [16] that analysed data from 6378 subjects (individual data from 3282 subjects) with the FibroTest® and biopsies [3501 with hepatitis C virus (HCV) infection and 1457 with hepatitis B virus (HBV)] found that the mean standardized AUROC for the diagnosis of significant fibrosis was 0.84, without significant differences between patients with HCV (0.85) and HBV (0.80). Another recent meta-analysis of APRI in 1798 HBV patients found mean AUROC values of 0.79 and 0.75 for the diagnosis of significant fibrosis and cirrhosis respectively [17]. In the largest study, which prospectively compared the FibroTest® with APRI in patients with viral hepatitis (913 HCV and 284 HBV patients), the AUROC values for significant fibrosis (0.78 vs 0.72, respectively) and cirrhosis (0.82 vs 0.77, respectively) did not differ significantly [18].

Transient elastography

Although the ability of TE to quantify liver fibrosis was first studied in patients with chronic hepatitis C, its results have been largely confirmed in patients with hepatitis B [18-27]. When compared between patients with hepatitis B and C, the diagnostic accuracy of TE did not differ [25, 26]. TE more accurately detects cirrhosis (AUROC values, 0.85–0.97; correct classification ranging from 64% to 94%) than significant fibrosis (AUROC values, 0.65–0.97; correct classification from 59% to 90%) (Table 2). Several meta-analyses [28, 29] have confirmed the better diagnostic performance of TE for cirrhosis than for fibrosis, with mean AUROC values of 0.94 and 0.84 respectively [28]. In a recent meta-analysis of 18 studies including 2772 HBV patients [30], mean AUROC values for diagnosing cirrhosis and significant fibrosis were 0.93 and 0.86 respectively. However, we are still lacking a meta-analysis of data from individual patients.

Table 2. Diagnostic performance of TE for significant fibrosis (F ≥ 2) and cirrhosis (F4) in patients with viral hepatitis B
AuthorsCountryYearPatient (n)F ≥ 2 (%)F4 (%)Cut-offs (Kpa)AUROCSe (%)Sp (%)CC (%)
  1. AUROC, area under ROC curve; Se, sensitivity; Sp, specificity; CC, correctly classified: true positive and negative; NA, not available.

  2. a

    HBV and HCV patients were analysed together.

  3. b

    More than half of patients with «clinical» cirrhosis.

  4. c

    Adapted to ALT levels.

Coco et al. [19]Italy2007228a62 8.30.93859187
     50b14.00.96789888
Oliveri et al. [20]Italy2008188   267.50.97948890
     20b11.80.97869694
Marcellin et al. [21]France200917350 7.20.81708376
     811.00.93938794
Chan et al. [22]Hong Kong2009161 2512–13.4c0.93987585
Kim Do et al. [23]South Korea200991 4310.3080597864
Wang et al. [24]Taiwan200988NA 8.00.868077NA
     NA10.00.898588NA
Degos et al. [18]France201028442 5.20.78893859
     1012.90.85529389
Sporea et al. [25]Romania201014076 7.00.655970NA
     513.60.978699NA
Cardoso et al. [26]France201220242 7.20.87748882
     811.00.93759089
Goyal et al. [27]India201335725 6.00.848267NA
     69.00.938190NA

Use in clinical practice

Assessing the stage of fibrosis

In clinical practice, determining the stage of fibrosis does not need to be as exact as the pathological scoring system. The absolute stage is less important than determining whether patients have mild or advanced liver disease. To identify patients with significant fibrosis, sensitivities and specificities above 85% can be considered to be sufficient, because there are no relevant clinical consequences of false positives or false negatives. Because the performances of TE and serum biomarkers have been shown to be equivalent [18], the use of either method depends on what is available. Identifying patients with cirrhosis requires more sensitive tests, because patients might need specific therapies and because patients must be screened for complications. TE appears to be the most well-suited test to screen for cirrhosis because it has better results than biomarker assays [18].

Although strategies to combine non-invasive methods have been shown to increase the diagnostic accuracy in patients with HCV [31-34], they have not yet been validated in patients with HBV [35-38]. Serum levels of aminotransferases should also be taken into account when interpreting results from TE in patients with hepatitis B [39]. To avoid the risk of false positives, certain authors have suggested that TE cut-offs should be adapted according to ALT levels [22]—a strategy that might not apply to patients with fluctuating levels of ALT or hepatitis flares. Conversely, in HBeAg-negative patients with normal ALT levels, non-invasive methods, particularly TE, could be used as adjunct tools to HBV DNA measurements, to follow inactive carriers or better identify patients who require liver biopsy (those with ongoing disease activity or significant fibrosis, despite normal ALT levels) [20, 40-42].

Deciding to provide or defer antiviral therapy

Treatment of hepatitis B is not curative and is usually prolonged. Besides staging of fibrosis, HBeAg, ALT and HBV DNA measurements play important roles in the decision to treat patients with hepatitis B. Non-invasive tests for fibrosis are used less systematically in the management of HBV than HCV because there have been fewer studies on the topic and because liver inflammation and HBV replication confound the interpretation of test results. In treatment-naïve patients, non-invasive tests could used for patients with ALT levels <2-fold the upper limit of normal (ULN) and levels of HBV DNA >20 000 IU/ml (for HBeAg-positive patients) or >2000 IU/ml (for HBeAg-negative patients) [2]. Results from studies combining TE and biomarker assays are still too preliminary to make recommendations. Some researchers have proposed diagnostic algorithms that use dual TE cut-offs, for the positive and negative prediction of significant fibrosis [43]. Obviously, the applicability of this dual cut-off strategy is determined by the setting and the probability that patients have significant fibrosis. For instance, a cut-off value <6.2 kPa could exclude a diagnosis of significant fibrosis in nearly all patients (87%) with a low probability of significant fibrosis such as inactive carriers. Alternatively, a cut-off value of >9.4 kPa accurately predicted fibrosis (92%) in patients with a higher probability of significant fibrosis, such as middle-aged, HBeAg-negative patients with persistently abnormal ALT levels. The remaining patients should still undergo liver biopsy [2].

Monitoring treatment response

A major advantage of non-invasive methods compared with liver biopsy is that the former can easily be repeated over time to monitor patients. For instance, in patients already receiving antiviral therapy, TE and biomarker assays could be used to monitor response to treatment and to evaluate the regression of fibrosis. Significant histological improvement has been documented in studies of paired liver biopsies from patients with chronic hepatitis B who received long-term antiviral therapy [44, 45].

Several recent studies have reported a significant decrease in liver stiffness and biomarker values in HBV-infected patients treated with analogues [46-50]. Despite these encouraging results, follow-up of treated patients using TE or biomarkers can be confounded by changing ALT levels and inflammation. Certain tests for serum biomarkers include parameters that could be affected by the resolution of inflammation (including ALT and AST levels). Similarly, a decrease in liver stiffness could be a result of a reduction in inflammatory activity, rather than fibrosis. Thus, serial liver stiffness measurements should be performed after ALT levels have normalized, during the course of treatment. In patients with cirrhosis, post-treatment assessment of liver stiffness should not replace the recommended, periodic monitoring for HCC based on ultrasound and α-fetoprotein levels.

Conclusions and Perspectives

Significant progress has been made over the past decade in the non-invasive assessment of liver disease in patients with hepatitis B. Nevertheless, non-invasive methods will reduce, but not completely end the need for liver biopsy. Liver biopsies and non-invasive methods should be used as an integrated system for more efficient evaluation of patients with hepatitis B [51]. It is also important to investigate the prognostic value of non-invasive methods, particularly TE in patients with cirrhosis [52]. These tests could be used to better classify patients with cirrhosis and assign them to different categories of risk for clinical outcomes. Finally, TE has limitations and is challenged by other technologies that measure liver stiffness, such as ARFI [53] and MR elastography [54], whose place in clinical practice remains to be defined.

Ancillary