Abstract
BACKGROUND: This study aimed to comprehensively investigate the observational and Mendelian randomization (MR) associations between liver enzymes (ALP, ALT, AST and GGT) with a wide spectrum of human diseases. METHODS: UK Biobank data was used for observational analysis. Cox proportional hazard models were used to quantify the observational associations between liver enzymes and risks of 1109 diseases, adjusted for covariates. A correction was applied to allow for multiple comparisons. Two-sample MR was performed to replicate the identified observational associations. We assessed the strength, consistency and pleiotropy of these MR associations. RESULTS: In 438571 participants with a median 12.4 years of follow-up, we found evidence linking liver enzymes and 654 diseases, 254 of which were further assessed in subsequent MR. In MR, 52 associations were replicated (10 for ALP, 14 for ALT, 10 for AST and 18 for GGT). Specifically, ALT was associated with liver conditions, diabetes, abnormal bleeding from the female genital tract, and prostate cancer. AST showed associations with prostate cancer, anxiety, diabetic neuropathy, and stroke. GGT was linked to hepatic and biliary conditions, coronary atherosclerosis, and pneumonia, while ALP displayed positive associations with vitamin D-calcium-parathyroid-bone conditions. CONCLUSIONS: Elevated liver enzymes are risk factors for a wide range of hepatic and extrahepatic diseases, predominantly digestive, circulatory, metabolic/endocrine (including diabetes) diseases. Furthermore, the four liver enzymes demonstrated distinct patterns of disease associations. KEY MESSAGES: • Elevated liver enzymes are risk factors for a wide range of hepatic and extrahepatic diseases. • The four liver enzymes demonstrated distinct patterns of disease associations.