作者: Debbie A. Lawlor , Marianne Benn , Luisa Zuccolo , N. Maneka G. De Silva , Anne Tybjaerg-Hansen
DOI: 10.1371/JOURNAL.PONE.0114294
关键词:
摘要: Background: The effect of alcohol consumption on liver function is difficult to determine because reporting bias and potential residual confounding. Our aim was this using genetic variants proxy for the unbiased alcohol. Methods: We used in ADH1B ADH1C genes as instrumental variables (IV) estimate causal long-term alanine aminotransferase (ALT), c-glutamyl-transferase (c-GT), alkaline phosphatase (ALP), bilirubin prothrombin action. Analyses were undertaken 58,313 Danes (mean age 56). Results: In both confounder adjusted multivariable genetic-IV analyses greater consumption, amongst those who drank any alcohol, associated with higher ALT [mean difference per doubling consumption: 3.4% (95% CI: 3.1, 3.7) from 3.7% (24.5, 11.9) analyses] c-GT [8.2% (7.8, 8.5) 6.8% (22.8, 16.5)]. point estimates two methods very similar statistically results consistent each other effects (both pdiff.0.3). Results suggested a weak inverse association ALP [21.5% (21.7, 21.3)], which differed