Accuracy of Noninvasive Scoring Systems in Assessing Risk of Death and Liver-Related Endpoints in Patients With Nonalcoholic Fatty Liver Disease

BACKGROUND AND AIMS: Several non-invasive scoring systems have been developed to determine risk of advanced fibrosis in non-alcoholic fatty liver disease (NAFLD). We examined the association between 4 scoring systems and incident severe liver disease and overall mortality in a large cohort of patients with biopsy-proven NAFLD. METHODS: We performed a retrospective analysis of data from 646 patients with biopsy-proven NAFLD, recruited from 2 hospitals in Sweden, from 1971 through 2009. The NAFLD fibrosis score (NFS), FIB-4, APRI, and BARD scores were calculated at the time of the liver biopsy. Based on each score, patients were assigned to categories of low, intermediate, or high risk for advanced fibrosis. Overall mortality and severe liver disease (cirrhosis, decompensated liver disease, liver failure, or hepatocellular carcinoma) were ascertained through linkage with national registers until the end of 2014. Cox regression, area under the receiver operating characteristic (AUROC) curve, and C-statistic analyses were used to study the predictive capacity of each scoring system. RESULTS: During a mean follow-up time of 19.9 +/- 8.7 years, there were 214 deaths and 76 cases of severe liver disease. For overall mortality, AUROC curve values were: NFS, 0.72 (95% CI, 0.68-0.76); FIB-4, 0.72 (95% CI, 0.68-0.76); BARD, 0.62 (95% CI, 0.58-0.66); and APRI, 0.52 (95% CI, 0.47-0.57). For severe liver disease, AUROC curve values were: NFS, 0.72 (95% CI, 0.66-0.78); FIB-4, 0.72 (95% CI, 0.66-0.79); BARD, 0.62 (95% CI, 0.55-0.69); APRI, 0.69 (95% CI, 0.63-0.76). C-statistics for all scores were of moderate capacity to predict outcomes. CONCLUSIONS: In a retrospective analysis of data from 646 patients with biopsy-proven NAFLD, we found the NFS and the FIB-4 scores to most accurately determine risk of overall death or severe liver disease. However, the AUROC values for these scoring systems are not high enough for use in the clinic; new systems are needed to determine prognoses of patients with NAFLD.