FibroTest/Fibrosure for Significant Liver Fibrosis and Cirrhosis in Chronic Hepatitis B

A Meta-analysis

Nermin N Salkic MD; PhD; Predrag Jovanovic MD; PhD; Goran Hauser MD, PhD; Majda Brcic MD

Disclosures

Am J Gastroenterol. 2014;109(6):796-809. 

In This Article

Abstract and Introduction

Abstract

Objectives: Extent of liver fibrosis is one of the most important factors in determining prognosis and the need for active treatment in chronic hepatitis B (CHB). Noninvasive alternatives such as FibroTest/Fibrosure (FT) have been developed in order to overcome the shortcomings of liver biopsy (LB). We aimed to systematically review studies describing the diagnostic accuracy of FT for predicting CHB-related fibrosis.

Methods: MEDLINE and EMBASE searches and hand searching methods were performed to identify studies that assessed the diagnostic accuracy of FibroTest in HB patients using LB as a reference standard. We used a hierarchical summary receiver operating curves model and the bivariate model to produce summary receiver operating characteristic curves and pooled estimates of sensitivity and specificity.

Results: We included 16 studies (N=2494) and 13 studies (N=1754) in the heterogenous meta-analysis for liver fibrosis and cirrhosis, respectively. The area under the hierarchical summary receiver operating curve for significant liver fibrosis and for all included studies was 0.84 (95% confidence interval (CI): 0.78–0.88). At the FT threshold of 0.48, the sensitivity, specificity, and diagnostic odds ratio (DOR) of FT for significant fibrosis were 61 (48–72%), 80 (72–86%), and 6.2% (3.3–11.9), respectively. The area under the hierarchical summary receiver operating curve for liver cirrhosis and for all included studies was 0.87 (95% CI: 0.85–0.90). At the FT threshold of 0.74, the sensitivity, specificity, and DOR of FT for cirrhosis were 62 (47–75%), 91 (88–93%), and 15.7% (8.6–28.8), respectively.

Conclusions: FibroTest is of value in exclusion of patients with CHB-related cirrhosis, but has suboptimal accuracy in the detection of significant fibrosis and cirrhosis. It is necessary to further improve the test or combine it with other noninvasive modalities in order to improve accuracy.

Introduction

Chronic hepatitis B (CHB) is a major public health problem worldwide, with more than two billion people showing evidence of exposure and more than 240 million people showing evidence of chronic infection.[1]

Substantial progress has been made in the treatment of chronically infected patients in the last couple of years. However, precise definition of the extent of the liver fibrosis in CHB remains one of the most important factors determining both the risk of further progression of the disease and the need for active treatment.[2] The most recent guidelines by the European Association for Study of Liver recommend liver biopsy (LB) to determine the degree of necroinflammation and fibrosis, as hepatic histology can assist the decision to start treatment.[3] In addition, there is a need to monitor treatment effects, with recovery of liver histology being one of the most important signs of success.

LB with subsequent histological analysis is considered as a gold standard that allows evaluation of presence and extent of the fibrotic process in liver tissue.[4] However, LB is an invasive procedure associated with significant patient discomfort and a small but important risk of complications, with reported risk of hospitalization ranging from 1 to 5%, with a risk of severe complications of 0.57% and mortality rates varying from 0.009 to 0.12%.[5–7] It is also prone to sample variability and accuracy seems to depend highly on the size of the sample.[8,9] Another important limitation derives from the fact that the biopsy sample is evaluated by histological scores, which have a certain amount of interobserver variability and are also dependent on the experience of the pathologist.[10]

As a result of these issues, numerous investigators have attempted to devise noninvasive methods of assessing hepatic fibrosis, resulting in more than 20 different clinical scores or imaging modalities with variable diagnostic accuracy.[11] Most attention has been focused on whether noninvasive methodologies can detect the presence or absence of minimal (i.e., F0–F1), significant (i.e., ≥F2), or advanced (i.e., ≥F3–F4) fibrosis according to the METAVIR histological score.[12,13]

One of the most investigated and most frequently used tools is the FibroTest/Fibrosure (FT) (proprietary formula; Biopredictive, Paris, France)—a patented calculation of the combination of five serum biochemical parameters (α-2-macroglobulin, apolipoprotein A1, haptoglobin, L-glutamyltranspeptidase, and bilirubin)—which was developed by Poynard and colleagues.[14,15] Clear advantages of FT include high applicability (>95%), widespread availability, and inter-laboratory reproducibility;[16] however, there are also numerous drawbacks such as cost, failed external validation, lack of specificity for liver disease (results can be severely impaired by comorbidities, i.e., Gilbert's syndrome or hemolysis),[17] and difficulty in differentiating intermediate stages of fibrosis.[18] Assessment of liver fibrosis without biopsy is very tempting, and despite the fact that recommendations suggest that noninvasive tests are still not ready to replace LB,[2,18] FT has become widely present in clinical practice. Its accuracy for detection of fibrosis or even disease prognosis has been evaluated extensively in a variety of liver diseases and in several systematic reviews.[19–24]

Despite its omnipresence, FT has not been as extensively studied in the CHB population as in chronic hepatitis C. There are several possible reasons for this: apart from the stage of fibrosis, other factors such as HBe antigen positivity, levels of alanine aminotransferase, and HB virus (HBV) DNA have important roles in deciding how and when to treat patients with CHB, and liver inflammation and HBV replication may confound interpretation of FT results.[11] The most recent guidelines have indeed recognized the unresolved issue of the true place of FT and all noninvasive markers in the evaluation and follow-up of CHB patients, and recommend further development.[3]

Currently, we are aware of only one meta-analysis focused on accuracy of FT and other noninvasive markers of hepatic fibrosis in CHB patients.[22] However, it is somewhat limited by the relatively small number of included studies, along with the inclusion of predominantly single-center data on FT, which may impair its reproducibility and interpretability. In additon, there is an apparent diversity in the statistical methodology of meta-analysis in most of the reviews; most of them use various methods to analyze areas under the receiver operating characteristics curve (AUCs), yet none are prepared according to the methodology recommended by the Cochrane Collaboration Diagnostic Test Accuracy (DTA) Working Group.[25] This may be an important issue; from a clinical standpoint, reported AUCs do give insights into the overall accuracy of a diagnostic test, but it is much more useful for a clinician to know how well tests perform at a certain threshold, measured by sensitivity, specificity, positive and negative predictive values, and likelihood ratio (LR).

Our primary objective was therefore to perform an independent meta-analysis of the diagnostic accuracy of FT for the prediction of significant liver fibrosis (F2–F4 vs. F0–F1) in CHB patients. Our secondary goal was to evaluate the diagnostic accuracy of FT for the prediction of liver cirrhosis (F4 vs. F0–F3) in CHB patients.

processing....