What is liver fibrosis?
In 1997 a group of European investigators convened by Professor Michael Arthur embarked on a project led by Professor William Rosenberg funded by Bayer Healthcare to identify serum markers of liver fibrosis. This program of research has been continuously active for over a decade and has resulted in the identification of a panel of direct markers that have been validated, CE marked, and are now being marketed in Europe by iQur Limited and Siemens.
The original European Liver Fibrosis project recruited over 1,000 patients having investigation with a liver biopsy at 13 centres across Europe. The subjects had a wide range of CLD reflecting clinical practice. Over 40% had CHC or CHB. All patients had a fasting serum sample taken at the time of their biopsy. This was sent to a central laboratory for analysis of a panel of candidate analytes representing direct markers of fibrosis as well as a number of indirect makers. The individual sandwich ELISA test for each of the direct markers were carefully developed to a high standard of accuracy, reproducibility and repeatability. Using logistic regression and multivariate analysis, those markers that most accurately reflected the stage of liver fibrosis assigned by a central pathologist were identified in a “training” cohort of 521 patients and then confirmed in a “validation” cohort of 400. This revealed that the combination of HA, P3NP and TIMP-1 combined in an algorithm, originally incorporating Age, could be used to determine the severity of liver fibrosis with good accuracy.14 Subsequently, the team have established that Age could be omitted from the algorithm to generate the Enhanced Liver Fibrosis test or ELF Test.
Subsequent to the original study, investigators around the world have conducted validation studies in independent populations to further assess the performance of the markers. Studies in CHC, NAFLD and PBC have all confirmed that the markers accurately reflect the severity of fibrosis as staged on liver biopsy15-19. In these studies the area under the receiver operator characteristic curve (AUROC) ranges around 0.8. This level of performance is considered to be a threshold for acceptance in clinical practice.
However, the investigators have been limited in their ability to assess the performance of the markers by the errors inherent in biopsy staging. In an attempt to overcome this “glass ceiling,” the ELF team have begun to investigate the ability of the serum markers to predict long-term clinical outcomes of CLD including the development of portal hypertension, decompensation of CLD, the development of hepatocellular cancer, liver transplantation and death from liver disease, as well as all cause mortality. Whilst not yet complete, interim analysis of the 7-year follow-up of over 500 patients has shown that the ELF markers are at least as good, if not better than liver histology at predicting clinical outcome. Similar work has been conducted in a cohort of patients with PBC.20
Advantages Over Single Markers
Systematic reviews of the literature have consistently shown that single markers of fibrosis can detect cirrhosis with some degree of accuracy. However, single-marker tests are less accurate than panels of markers in detecting lesser degrees of fibrosis.21,22 While the detection of cirrhosis is important, clinical judgement can often identify patients who are likely to have end-stage CLD. It is of greater importance to be able to detect patients with mild or moderate fibrosis, which is usually asymptomatic, in order to be able to intervene with lifestyle modification or treatment before the liver becomes irreparably damaged. It is in the screening and management of patients with CLD that the ELF markers have great potential.
Advantages Compared to Indirect Markers
The careful and extensive validation program employed by the ELF development team has taken years to come to fruition. A number of algorithms derived employing indirect markers that are available in general clinical chemistry laboratories as routine analytes. While the performance of many of these algorithms, such as Fibrotest/Fibrosure, APRI, Forns Index and Hepascore is good, the majority are unreliable in patients undergoing treatment for viral hepatitis (where aminotransferases are altered due to therapy) or when bilirubin levels may be elevated due to haemolysis (such as when ribavirin is administered) or cholestasis.23-29
Use in Clinical Practice and Impact on Patient Management
The coming year will see the introduction of the ELF test into clinical practice, and its impact in the management of patients will begin to be appreciated. Doctors in primary care are flooded with patients with obesity and hazardous drinking all of whom could be at risk of CLD. The ability to use a simple blood test to accurately identify those with significant liver disease will greatly aid triage and the appropriate targeting of interventions including weight loss, exercise and drug interventions. In secondary care, the ELF test will be of great use in the early evaluation of patients with a wide range of CLD. It will not replace liver biopsy in the detailed assessment of liver inflammation, architectural damage of pathology. However it can be used to prioritise patients for investigation, to determine the severity of fibrosis in patients unwilling or unable to undergo biopsy, and to complement biopsy given the inaccuracies that surround histological staging arising from sampling error and observer error. Once a diagnosis of CLD has been made, most patients will undergo some form of treatment and long-term follow-up. Further repeated biopsies are not acceptable in the vast majority of patients, but further knowledge of the severity of fibrosis, its progression or regression would be highly valuable to both the patient and the doctor. The ELF test will make this information accessible through a simple blood test that could be repeated at frequent intervals.