Introduction The use of non-invasive liver tests (NILTs) for assessing hepatic fibrosis carries fewer risks than liver biopsy. However, their clinical utility and cost-effectiveness, in patients with suspected alcohol-related liver disease, are unknown. This study explored whether the Enhanced Liver Fibrosis (ELF) test, FibroTest, and FibroScan are likely to be cost-effective in a UK setting.
Methods Systematic reviews of the clinical effectiveness of the three NILTs under study and of the morbidity/mortality associated with liver biopsy were undertaken and used, where possible, to populate a mathematical model constructed to simulate the experience of patients suspected of having alcohol-related liver injury. The model used the following strategies: (1) biopsy all; (2) use a NILT for triage and biopsy more selectively; and (3) use a NILT alone to determine diagnosis and subsequent management. Thirty-six separate scenarios were evaluated varying the type and sensitivity of the liver biopsy, the accuracy of the non-invasive tests, and the disutility associated with the biopsy procedure.
Results No clear conclusions could be drawn, following the systematic reviews, about the accuracy of these three NILTs in the diagnosis of alcohol-related liver disease because there was significant heterogeneity in the published studies, particularly in relation to drinking behaviour and diagnostic cut-offs. The assessment was also confounded by the paucity/absence of data on: (1) the sensitivity of the liver biopsy; and (2) the difference in costs and health between abstinent patients and those who continued to drink. There was wide variation in the threshold level for a decrease in abstinence rates using NILTs in both triage and replacement strategies. The maximum threshold value was 6.1 percentage points with the minimum value being zero percentage points. There was also wide variation in the threshold levels for the incidental gains associated with biopsy using NILTs in both triage and replacement strategies. The maximum value was 0.188 QALYs with the minimum value being zero QALYs.
Conclusion The clinical utility and cost-effectiveness of these three NILTs cannot be assessed, at the present time, primarily because information on the key variables needed to populate the model are lacking. Thus, caution should be exercised in their use, in patients with suspected alcohol-related liver disease, at the present time.
Competing interests None declared.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.