Prospective External Validation of a New Non-invasive Test for the Diagnosis of Non-alcoholic Steatohepatitis in Patients With Type 2 Diabetes

Thierry Poynard; Valérie Paradis; Jimmy Mullaert; Olivier Deckmyn; Nathalie Gault; Estelle Marcault; Pauline Manchon; Nassima Si Mohammed; Beatrice Parfait; Mark Ibberson; Jean-Francois Gautier; Christian Boitard; Sébastien Czernichow; Etienne Larger; Fabienne Drane; Jean Marie Castille; Valentina Peta; Angélique Brzustowski; Benoit Terris; Anais Vallet-Pichard; Dominique Roulot; Cédric Laouénan; Pierre Bedossa; Laurent Castera; Stanislas Pol; Dominique Valla


Aliment Pharmacol Ther. 2021;54(7):952-966. 

In This Article

Abstract and Introduction


Background: One of the unmet needs in patients with type 2 diabetes mellitus (T2DM) is the prediction of non-alcoholic liver disease by non-invasive blood tests, for each of the three main histological features, fibrosis, non-alcoholic steatohepatitis (NASH) and steatosis.

Aims: To validate externally the performances of a recent panel, Nash-FibroTest, for the assessment of the severity of fibrosis stages, NASH grades and steatosis grades.

Methods: We prospectively analysed 272 patients with T2DM. Standard definitions of stages and grades were used, and analyses were centralised and blinded. The performances of the FibroTest, NashTest-2 and SteatoTest-2 were assessed using the Obuchowski measure (OM), the main outcome recommended as a summary measure of accuracy includeing all pairwise stages and grades comparisons, which is not provided par the extensively used binary area under the ROC curve.

Results: The diagnostic performance of each component of the panel was significant. OM (SE; significance) of the FibroTest, the NashTest-2 and the SteatoTest-2 was 0.862 (0.012; P < 0.001), 0.827 (0.015; P < 0.001) and 0.794 (0.020; P < 0.01), respectively. For ballooning and lobular inflammation, OM was 0.794 (0.021; P < 0.001) and 0.821 (0.017; P < 0.001), respectively. In a post hoc analysis the FibroTest outperformed VCTE by 4.1% (2.5–6.5; P < 0.001) for reliability, with a non-significant difference for OM for fibrosis staging, 0.859 (0.012) for FibroTest vs 0.870 (0.009) for VCTE.

Conclusions: From a single blood sample, the panel provides non-invasive diagnosis of the stages of fibrosis, and the grades of NASH and steatosis in patients with T2DM.

Trial registration number: NCT03634098.


One of the unmet needs in patients with components of the metabolic syndrome such as android obesity, type 2 diabetes mellitus (type 2 diabetes), hyperlipidaemia and hypertension is easy access to non-invasive tests (NITs) to assess the severity of non-alcoholic steato-hepatitis (NASH), including its three main histological features, steatosis, activity and fibrosis. Although the courses and clinical relevance of these features differ, they are intercorrelated.[1,2] Steatosis is the early feature of disease, the progression of fibrosis is the most accurate predictor of mortality and severe liver events, and inflammatory 'activity' is the biological driver of the progression of fibrosis.[3,4]

Thus, the availability of one NIT for each feature would provide a simple alternative to liver biopsy for the surveillance and treatment strategy in patients at risk of NASH. Among the numerous available NITS for the diagnosis of NASH, one panel, called the 'Nash FibroTest' panel, provides a specific test for each of the three features including the FibroTest (FibroSure in the USA);[5] NashTest-2 for NASH;[6] and SteatoTest-2 for steatosis.[7] This panel was constructed and internally validated in 600 patients at risk of NAFLD in a multicentre cohort, using the steatosis activity fibrosis score (SAF-score), which defines the grades of NASH and steatosis and their associated clinical outcomes independently.[8–10]

These tests were then used in a prospective phase-2 trial of selonsertib,[7] and in an ongoing phase-3 trial of obeticholic acid,[11] and validated in a retrospective analysis of 220 patients, all with type 2 diabetes.[12] The latter study is the only existing trial evaluating type 2 diabetes patients. In a review of 25 NITs of NAFLD, the prevalence of type 2 diabetes ranged between 14% and 50%.[13] Of all the components of the metabolic syndrome, type 2 diabetes is the most important risk factor for NAFLD and non-alcoholic steatohepatitis (NASH), and the most important clinical predictor of adverse outcomes such as advanced liver fibrosis and mortality.[14] Indeed, a meta-analysis of population-based observational studies found that type 2 diabetes is associated with a more than twofold increase in the risk of developing severe liver disease.[15] The performance of NITs for the diagnosis of the features of liver disease in type 2 diabetes patients is controversial,[16,17] because of the absence of prospective evaluations as well as methodological limitations such as biopsy sampling variability,[18] intra- and interobserver variability for scoring the features,[18] the impact of the spectrum effect on binary AUROCs,[19–21] and inappropriate histological references to assess the performance of NITS such as the NAS score.[8,9,22] Thus, the first aim of the prospective QUID-NASH research program focusing on type 2 diabetes (, was to externally validate the performance of the 'Nash FibroTest' panel in the specific context of diabetology outpatient clinics using a centralised biopsy review, and appropriate methods.