Prediction of future hepatocellular carcinoma incidence in moderate to heavy alcohol drinkers with the FIB-4 liver fibrosis index.

Abstract

Although heavy alcoholics are at heightened risk for hepatocellular carcinoma (HCC), there are no guidelines that recommend HCC screening for heavy alcoholics. This study investigated FIB-4, a noninvasive and easily applicable liver fibrosis index, as a risk factor for HCC incidence among alcohol drinkers without viral hepatitis. This retrospective cohort study included 6661 generally healthy adults who were 30 years old or older, did not have chronic viral hepatitis, and visited Seoul National University Hospital for a general, routine health evaluation. The future HCC incidence was determined from National Health Insurance medical service claims data (median follow-up, 6.2 years). With adjustments for age, sex, body mass index, smoking, and alcohol, compared with subjects with FIB-4 values less 1.00, subjects with FIB-4 values greater than or equal to 1.75 and less than 2.10 and subjects with FIB-4 values greater than or equal to 2.10 had adjusted hazard ratios (aHRs) of 5.18 (95% confidence interval [CI], 1.12-24.00) and 13.63 (95% CI, 3.77-49.33), respectively, for HCC incidence. This was heightened in subjects who drank more 30 g of alcohol per day: the aHRs were 8.39 (95% CI, 1.28-54.87) and 16.58 (95% CI, 3.87-71.04), respectively. FIB-4 was shown to have a higher predictive value for HCC incidence than ultrasonographically detected liver cirrhosis (C-index, 0.665 vs 0.527; P = .044). High FIB-4 is a risk factor with a high predictive value for HCC incidence, especially among moderate to heavy alcoholics (>30 g/d). FIB-4 is a readily available and probably cost-effective clinical tool with potential value for identifying subpopulations of alcoholics at particularly high risk who would benefit from regular HCC screening. Further investigations are warranted to validate our results; nonetheless, our study suggests that FIB-4 may be useful in HCC screening among alcoholics.