BACKGROUND This study aimed to examine whether repeated measurements on noninvasive fibrosis scores during follow-up improve long-term nonalcoholic fatty liver disease (NAFLD) outcome prediction.

METHODS A cohort study of 2,280 NAFLD patients diagnosed at the Seoul National University Hospital from 2001 to 2015 was conducted. Multivariable Cox regression models with baseline and designated time-point measurements of the fibrosis-4 index (FIB-4) and NAFLD fibrosis score (NFS) were used to assess the association between these scores and overall mortality, liver-related outcomes, and cardiovascular events.

RESULTS Higher baseline NFS (high versus low probability for advanced fibrosis groups) was associated with higher risk of mortality (adjusted hazard ratio (aHR), (95% confidence interval (CI)), 2.80, [1.39-5.63]) and liver-related outcomes (3.70, [1.27-10.78]). Similar findings were observed for the association of baseline FIB-4 with mortality (2.49, [1.46-4.24]) and liver-related outcomes (11.50, [6.17-21.44]). In models considering designated time-point measurements of the scores, stronger associations were noted. For NFS, a higher time-point measurement was associated with a significantly higher risk of mortality (3.01, [1.65-5.49]) and liver-related outcomes (6.69, [2.62-17.06]). For FIB-4, higher time-point measurements were associated with significantly higher mortality (3.01, [1.88-4.82]) and liver-related outcomes (13.26, [6.89-25.53]). An annual increase in FIB-4 (2.70, [1.79-4.05]) or NFS (4.68, [1.52-14.44]) was associated with an increased risk of liver-related outcomes. No association between NFS/FIB-4 and risk of cardiovascular events was observed in both models.

CONCLUSIONS Higher aHRs describing the associations of FIB-4/NFS with overall mortality and liver-related outcomes were observed in the models that included designated time-point measurements of the scores. In addition to the baseline measurement, a routine monitoring on these scores may be important in predicting prognosis of NAFLD patients.