Abstract
BACKGROUND AND AIMS: Individuals with genetic polymorphisms in UGT1A1 exhibit bilirubin levels that belie their risk of liver disease (Gilbert's syndrome) but it is not known if this phenomenon extends to other common liver blood tests (LBTs). METHODS: A genome-wide association analysis of 10 LBTs was conducted using the UK biobank. Polygenic scores (PGS) were created from discordant loci (e.g. loci associated with the LBT but not associated with cirrhosis morbidity risk). Participants were assigned to a low, intermediate or high PGS for each LBT. A high PGS approximates Gilbert's syndrome (i.e. elevated LBT without an analogous increase in disease risk). The prognostic significance of an 'elevated' LBT-and how this differs by PGS-was assessed through competing risk survival analysis. RESULTS: This study included 157 005 and 166 871 participants for the discovery and validation phases, respectively. Elevated LBTs were more prevalent in the high versus low PGS group, yet the 10-year risk of cirrhosis morbidity was comparable. For example, in the low PGS group, 4.3% had an elevated gamma-glutamyltransferase (GGT) and the 10-year risk of cirrhosis morbidity was .45%. Conversely, in the high PGS group, 21.2% had an elevated GGT and the 10-year risk was .38%. Accordingly, the 10-year risk of cirrhosis morbidity for individuals with an elevated GGT was markedly different in the low vs. high group (4.2% vs. 1.2%; p < .001). Similar results were apparent for Fibrosis-4 index, total bilirubin, and platelet count. CONCLUSION: Variability in LBTs is influenced by genetic polymorphisms that have a neutral effect on disease risk. These findings have implications for interpreting elevated LBTs in clinical practice.