Serum Fibrosis Tests Guide Prognosis in Metabolic Dysfunction-Associated Fatty Liver Disease Patients Referred From...

Metabolic dysfunction-associated fatty liver disease (MAFLD) is managed predominately in primary care, however, there is uncertainty regarding how to best identify patients for specialist referral. We examined the accuracy of noninvasive tests as screening tools for the prediction of outcomes in MAFLD patients referred from primary care.

Patients with MAFLD referred by primary care for specialist review to Sir Charles Gairdner Hospital (cohort 1, n = 626) or tertiary centers within Western Australia (cohort 2, n = 246) were examined. Hepascore, aspartate aminotransferase to platelet ratio, Fibrosis-4 (FIB-4), and nonalcoholic fatty liver disease fibrosis score performed at baseline were examined for their accuracy in predicting liver-related death (LRD), decompensation, and hepatocellular carcinoma. Outcomes were collected from hospital records and data linkage.

The median follow-up period was 5.0 years (range, 0.1-13.0 y) and 3.8 years (range, 0.1-10.0 y) in cohorts 1 and 2, respectively. In both cohorts, Hepascore and FIB-4 had the highest area under the curve for the prediction of LRD (0.90-0.95 and 0.83-0.94, respectively), decompensation (0.86-0.91 and 0.86-0.87, respectively), and hepatocellular carcinoma (0.75-0.90 and 0.67-0.85, respectively). The sensitivity and negative predictive values were high (>90%) for Hepascore (cut-off value, 0.60), FIB-4 (cut-off value, 1.30), and nonalcoholic fatty liver disease fibrosis score (cut-off value, -1.455) for all outcomes in cohort 1, and for predicting LRD in cohort 2. Hepascore had the highest specificity, classified the greatest proportion of patients as low risk, and was favored by decision curve analysis as providing the greatest net benefit.

Serum noninvasive tests accurately stratify risk of liver-related outcomes in MAFLD patients and can be used as a screening tool for patients referred for specialist review by primary care.