Abstract Background Bile acids are produced in the liver and are important for lipid digestion. Higher circulating bile acid levels, however, have been linked to metabolic disorders, inflammation, and gut microbiota dysbiosis, which have been implicated in liver carcinogenesis. To date, few epidemiological studies have explored the association between circulating bile acids and liver cancer risk. Methods We conducted a nested case-control study among 12 prospective cohort studies located in the United States. Fifteen pre-diagnostic circulating bile acids were measured from blood samples among 872 individuals who developed liver cancer and 872 matched control participants. Odds ratios (OR) and 95% confidence intervals (CI) were estimated using multivariable-adjusted conditional logistic regression analysis of circulating bile acid levels and liver cancer risk. Results Primary conjugated bile acid concentrations were positively associated with higher risk of liver cancer (OR per doubling in concentrations log2 of glycocholic acid, 95% CI: 1.32, 1.24-1.40; glycochenodeoxycholic acid: 1.33, 1.24-1.43; taurocholic acid: 1.28, 1.22-1.35; and taurchenodeoxycholic acid: 1.32, 1.24-1.39). Secondary conjugated bile acids were also positively associated with liver cancer risk (doubling of concentrations OR ranged from 1.11 to 1.22). Unconjugated bile acid concentrations were generally not associated with liver cancer risk, except lithocholic acid (OR per doubling: 1.27, 1.16-1.39). When analyses were separated into the two main subtypes of liver cancer, hepatocellular carcinoma (HCC; 438 cases/438 controls) and intrahepatic cholangiocarcinoma (ICC; 111 cases/111 controls), significant heterogeneity was observed for primary conjugated bile acid concentrations (all p-values 0.001) that showed positive significant associations with HCC but not ICC. Conclusions These results suggest that bile acids may be important markers of HCC risk and contribute to hepatocarcinogenesis; however, further research using serial measurements is needed.
Watling et al. (Tue,) studied this question.