Biomarkers and patient characteristics to predict changes in alcohol-related liver disease over time

Background and aim: Steatotic liver disease (SLD), including alcohol-related liver disease (ALD), covers a disease continuum ranging from simple steatosis to silent fibrosis and decompensated cirrhosis and poses a significant health care burden. SLD management is challenged by delayed diagnosis and a lack of insights on disease trajectories. The aim of this thesis was to evaluate the diagnostic and prognostic performance of selected non-invasive tests (NITs) and explore the dynamics of SLD and ALD over time.

Methods: In Paper 1, we included a prospective biopsy-controlled cohort study of individuals at risk of ALD and evaluated the diagnostic accuracy of M30, M65, and ActiTest for histopathological findings including hepatic inflammatory activity. We also evaluated the prognostic performance for future liver-related events (LRE) and all-cause mortality and compared findings to routine liver function tests (AST, ALT, GGT). In Paper 2, we merged two cohorts of individuals at risk of ALD who underwent liver stiffness measurements (LSM) and follow-up assessments for future LSM and outcomes. We then evaluated the prognostic performance of baseline LSM and LSM changes for predicting decompensation and all-cause mortality. Finally, in Paper 3, we included a prospective screening cohort of the general and an at-risk population undergoing SLD evaluation using LSM at two time points to evaluate the robustness of the SLD diagnosis and subclassification.

Results: Based on 265 biopsied individuals at risk of ALD, we found an excellent diagnostic accuracy for biopsy-verified severe hepatic inflammatory activity (grade 4-5) for M30 (AUROC 0.90), M65 (AUROC 0.86) and AST (AUROC 0.86), and moderate for ActiTest (AUROC 0.78). Elevated M30 (> 240 U/L) exhibited a sensitivity and specificity above 80%. Both elevated M30 and AST were associated with an increased risk of LRE and all-cause mortality, also when stratifying for abstinence. Next, we included 536 individuals at risk of ALD and reported that compensated advanced chronic liver disease (cACLD) ruledin by LSM using transient elastography (TE ≥ 10 kPa) led to a significantly decreased decompensation-free survival and 19-fold increased risk of decompensation. In 371 individuals (69%) with a follow-up LSM after a median of 25 months (IQR 17-38 months), we found that a LSM increase led to a significantly higher risk of outcomes compared to a LSM decrease in participants with cACLD (n = 125). A similar trend was observed in participants without cACLD at baseline (n = 246). Finally, based on 1,042 individuals undergoing SLD evaluation at baseline and after a median of 25 months (IQR 25-31), we found that 38% were reclassified according to the SLD diagnosis and subclassification driven by changes in steatosis and alcohol intake.

Conclusion: In individuals at risk of ALD, M30 accurately reflects hepatic inflammatory activity and is associated with outcomes, indicating progressive ALD. We also report that both baseline LSM and LSM changes hold prognostic value for decompensation and death, especially in cACLD in line with recommendations by Baveno VII. Moreover, we demonstrate that the robustness of the SLD diagnosis and subclassification is challenged, as 38% of cases are reclassified upon reassessment. Thus, longitudinal data on NITs and disease trajectories can significantly improve disease management in SLD, and more data is needed.