1Medical Faculty, University of Zurich, Pestalozzistrasse 3, CH-8032 Zurich, Switzerland.
2Division of Chronic Disease Epidemiology, Epidemiology, Biostatistics and Prevention Institute (EBPI), University of Zurich, Hirschengraben 84, CH-8001 Zurich, Switzerland.
3Division of Cancer Epidemiology, German Cancer Research Center (DKFZ), Im Neuenheimer Feld 280, DE-69120 Heidelberg, Germany.
Associations between liver enzymes or De Ritis ratio (DRR; aspartate aminotransferase (AST)/alanine aminotransferase (ALT)) and mortality stratified by non-alcoholic fatty liver disease (NAFLD), which have rarely been analyzed in previous studies, were investigated using the National Health and Nutrition Examination Survey (NHANES) III (1988-1994). Participants without risk factors for liver diseases other than NAFLD were linked with National Death Index records through 2019 (n = 11,385) and divided into two cohorts with or without NAFLD, based on ultrasound examination. Liver enzyme concentrations were categorized into sex-specific deciles and subsequently grouped (AST and ALT: 1-3, 4-9, 10; gamma glutamyltransferase (GGT): 1-8, 9-10). DRR was categorized into tertiles. Cox proportional hazards regression models adjusted for confounders were fitted to estimate associations with mortality. Compared with low levels, high GGT and DRR in participants with and without NAFLD had significantly higher hazard ratios for all-cause mortality. Compared with intermediate concentrations, low ALT showed higher all-cause mortality in participants with and without NAFLD, whereas low AST had higher HR in participants without NAFLD and high AST in those with NAFLD. Mortality was associated with liver enzymes or DRR in participants both with and without NAFLD, indicating that the relationship is not mediated solely by hepatocellular damage.