Steatosis non-invasive tests accurately predict metabolic dysfunction-associated steatotic liver disease, while fibrosis non-invasive tests fall short: Validation in U.S. adult population

Introduction

Metabolic dysfunction associated steatotic liver disease (MASLD) has replaced NAFLD as the diagnostic standard. This study aimed to validate non-invasive tests (NITs) for hepatic steatosis and fibrosis, originally used for NAFLD, in predicting MASLD and advanced fibrosis, respectively, compared to transient elastography (TE).

Methods

This cross-sectional study used the NHANES database (2017–2020). Adults aged ≥18 years with valid TE status and without other steatotic liver diseases were included. Hepatic steatosis NITs were compared against controlled attenuation parameter (CAP)- diagnosed MASLD and fibrosis NITS were compared against liver stiffness measurement (LSM)-diagnosed advanced fibrosis. The diagnostic accuracy was assessed using weighted ROC analysis.

Results

Among 5305 participants (49.1 % males), MASLD prevalence was 42.8 % (CAP-diagnosed) and advanced fibrosis was identified in 10.4 % (LSM-diagnosed) in those with MASLD. Steatosis NITs showed good diagnostic accuracy for predicting MASLD (AUROC 0.836 to 0.862), with fatty liver index having the maximum Youden index (0.55). Fibrosis NITs indicated poor to fair diagnostic accuracy for predicting advanced fibrosis (AUROC 0.564 to 0.691) but indicated high negative predicted value (89 %-94 %). In age categorized subgroup analysis, fibrosis NITs performed poorly in those aged ≤ 35 years and had low specificity to exclude fibrosis in those aged ≥65 years but performed comparably to the overall MASLD cohort in those aged 36–64 years.

Conclusion

In this population-based cohort of U.S. adults, all steatosis NITs demonstrated good diagnostic accuracy for MASLD. However, the fibrosis NITs showed limited diagnostic ability and were influenced by age, suggesting they should be used cautiously in general population.