Long-term Prognostic Value of the FibroTest in Patients With Non-alcoholic Fatty Liver Disease, Compared to Chronic Hepatitis C, B, and Alcoholic Liver Disease

Mona Munteanu; Raluca Pais; Valentina Peta; Olivier Deckmyn; Joseph Moussalli; Yen Ngo; Marika Rudler; Pascal Lebray; Frederic Charlotte; Vincent Thibault; Olivier Lucidarme; An Ngo; Françoise Imbert-Bismut; Chantal Housset; Dominique Thabut; Vlad Ratziu; Thierry Poynard


Aliment Pharmacol Ther. 2018;48(10):1117-1127. 

In This Article


In this study, we validated the performance of FibroTest for the prognostic of patients with NAFLD, both for the survival without liver-related deaths the primary aim, and for the overall survival. We also confirm that FibroTest has prognostic value for cardiovascular-related deaths in patients with NAFLD, mainly due to the prognostic value of ApoA1.

This long-term follow-up of a prospective cohort has several strengths and limitations.


The first strength was the sample size and the long-term follow-up, which permitted to analyse a total of 240 deaths in the NAFLD group including 38 liver-related deaths, much more than in our previous study in two different cohorts which analysed 172 deaths with only seven liver-related deaths.[9] The only other long-term prognostic study published included 360 patients with NAFLD and analysed 83 deaths with 17 liver-related deaths.[18] Without a direct comparison of tests in the same patients, in intention to diagnose analysis, it is hazardous to compare indirectly the non-invasive biomarkers.[5] However, the AUROC of FibroTest 0.941 (0.905–0.978) for liver-related deaths, observed in our cohort seems in the upper range of the performances observed in the other study, 0.853 (0.738–0.938) for the best patented blood test (Hepascore), 0.778 (0.663–0.880) for the best non-patented test (FIB4), and 0.885 (0.818–0.947) for transient elastography.[18] Here in the post hoc analyses, despite the low number of events, we retrieved same upper range of FibroTest vs FIB4 and transient elastography, in the same patients (AUROC = 0.943; 0.890–0.971) vs FIB–4 (0.906; 0.797–0.957; P = 0.33), and 0.954 (0.865–0.985) vs 0.903 (0.552–0.982; P = 0.38) respectively.

For the first time, as presumed by quantitative blood tests, we were able to assess if the severity of NASH and steatosis at baseline were predictive of survivals, independently of baseline fibrosis severity. Indeed, in the subset population with all these tests, despite the limited number of events (n = 17) NashTest–2 and SteatoTest–2 had significant predictive value. As expected, only the significance of NashTest–2 persisted when adjusted on FibroTest.

We confirmed our previous results in type 2–diabetes, that FibroTest predicted cardiovascular events, and improved the Framingham-risk score.[9] For the first time we observed that FibroTest was also predictive of cardiovascular-related deaths, both in NAFLD patients with and without diabetes and also among CHC patients.

There is evidence that low ApoA1 and low haptoglobin are two blood biomarkers of mortality risk, besides their performance for the diagnosis of cirrhosis.[10,11] They both mediate hepatoprotection. In healthy volunteers testing acetaminophen, these proteins were differentially expressed before acetaminophen intake in subjects with an increase in transaminases vs those without.[19] Similarly, in patients with drug-induced liver injury, higher serum levels of these proteins were predictive of transaminase recovery.[12] We also recently observed that individuals with lower ApoA1 or haptoglobin could be at a higher risk of developing primary liver cancer, irrespective of the presence of cirrhosis (T. Poynard et al. unpublished data.) ApoA1 was associated with an increased risk of overall cancers in a meta-analysis based on 28 epidemiologic studies as well as in a recent prospective cohort.[10,11] Here we also found a significant association between low ApoA1 and non-liver-related cancer deaths in CHC, but not in NAFLD patients. In NAFLD low ApoA1 was associated with all causes of death, and as expected, with cardiovascular mortality as extensively validated in patients with cardiovascular diseases.

For haptoglobin we found in NAFLD patients an unexpected association between high haptoglobin and cardiovascular deaths and all causes of deaths. We have no clear explanation of this association. Recent progress in the comprehension of the interactions of haptoglobin and ApoA1 in diabetics for the risk of cardiovascular mortality suggests first to analyse the prevalence of haptoglobin 2–2 haplotype individuals among our patients with NAFLD.[20,21] In these individuals, the oxidative modification of ApoA1 appears to be responsible for inducing inflammation in diabetic individuals with haptoglobin 2–2 genotype. In this scenario the "good cholesterol" (High Density Lipoprotein) can go "bad" and increase Low Density Lipoprotein-induced inflammation (File S2).


First, we have not performed an external validation. However, the prognostic value of FibroTest in NAFLD was already validated for overall survival and the risk of cardiovascular events in diabetic patients of the FibroFrance project,[9] and its diagnostic performance has been extensively validated in NAFLD patients.[16]

Even if this study had the longest follow-up and the higher number of events in comparison with the other published prognostic study in NAFLD, the number of events was still low. We found the same prevalence of primary liver cancer associated deaths than in the other prognostic studies, but more cases are needed to validate the prognostic value of biomarkers for liver cancer occurring in non-cirrhotic NAFLD or for the prediction of cholangiocarcinoma.

We did not analyse the predictive value of evolving risk factors during follow-up, such as alcohol intake or diabetes. We also did not analyse the predictive value of steatosis, being overweight, tobacco, coffee, chocolate or cannabis consumption, physical exercise or long-term drug use, all factors that could be associated not only with fibrosis but also with cardiovascular deaths.

Finally, we did not compare the prognostic value of FibroTest with other patented blood tests, and the post hoc comparisons vs FIB4 and transient elastography, validated 8 years later than Fibrotest, had a limited statistical power. For diagnostic performances, no biomarkers or elastography methods have been demonstrated superior to FibroTest in face to face study, and using intention to diagnose analysis, for chronic viral hepatitis and alcoholic liver disease.[8,16,5] To assess liver fibrosis in NAFLD, EASL guidelines recommends the use of either NAFLD-Fibrosis-score, FIB–4 score, Fibrometer, ELF as well as Fibrotest. To assess the liver-related prognosis in NAFLD there is so far no guidelines.

In Conclusion, the FibroTest has a high predictive value for survival without liver disease in patients with NAFLD, as already observed in chronic viral hepatitis and alcoholic liver disease.