BACKGROUND Population-based data regarding the prognostic implications of hepatic steatosis have been inconsistent. We examined the association between the fatty liver index (FLI) with all-cause and disease-specific mortality in the general population.

METHODS We included subjects who underwent a health examination in 2009 using the Korean nationwide health screening database. Death and cause of death data were provided by Statistics Korea. The causes of death were classified using 10th Revision of the International Classification of Diseases codes.

RESULTS Among the included 10,585,844 participants, there were 418,296 deaths during a median follow-up period of 8.3 years. When adjusting for possible confounding factors, the risk of all-cause mortality linearly increased with a higher FLI score (hazard ratio [HR], 95% confidence interval [CI]: FLI 30-59, 1.19, 1.18-1.20; FLI ≥ 60, 1.67, 1.65-1.69, P for trend <0.001). The risk of disease-specific mortality including cardiovascular disease (CVD), cancer, respiratory disease and liver disease, linearly increased as the FLI score became higher (HR, 95% CI: FLI 30-59, 1.18, 1.16-1.20, FLI ≥ 60: 1.61, 1.56-1.65 for CVD; FLI 30-59, 1.13, 1.11-1.14, FLI ≥ 60, 1.41, 1.38-1.44 for cancer; FLI 30-59, 1.26, 1.22-1.29, FLI ≥ 60, 1.96, 1.88-2.05 for respiratory disease, FLI 30-59, 2.29, 2.21-2.38, FLI ≥ 60, 5.57, 5.31-5.85 for liver disease). The risk of all-cause mortality increased as the FLI score became higher across all the body mass index groups, and the greatest risk was observed in those who were underweight (HR, 95% CI = 2.43, 2.09-2.82 in FLI ≥ 60).

CONCLUSION FLI may serve as a prognostic indicator of death and a high FLI is associated with a poor prognosis particularly in the underweight group.