AUTHOR=Tai An-Shun , Huang Yen-Tsung , Yang Hwai-I , Lan Lauren V. , Lin Sheng-Hsuan 

TITLE=G-Computation to Causal Mediation Analysis With Sequential Multiple Mediators—Investigating the Vulnerable Time Window of HBV Activity for the Mechanism of HCV Induced Hepatocellular Carcinoma

JOURNAL=Frontiers in Public Health

VOLUME=Volume 9 - 2021

YEAR=2022

URL=https://www.frontiersin.org/journals/public-health/articles/10.3389/fpubh.2021.757942

DOI=10.3389/fpubh.2021.757942

ISSN=2296-2565

ABSTRACT=The regression-based approach has been widely applied to causal mediation analysis. However, for multiple mediators, mediation analysis becomes complicated and challenging, and in this case, the regression-based approach is inefficient to address the estimation issue. Hence, the development of a flexible approach to mediation analysis is required. We developed a method based on g-computation algorithm to conduct causal mediation analysis in multiple ordered mediators. Unlike the regression-based approach, our simulation-based approach improves upon model choice flexibility and outcome scale range. Applying our method on the Taiwanese Cohort Study dataset, we investigated the mechanism of HCV infection affecting hepatocellular carcinoma (HCC) mediated by early and late HBV viral load among HBV seropositive patients (n = 2878; HCV carrier n=123). Our results indicate that HCV decreased the incidence of HCC primarily through the decrease in HBV viral load in the early stage. In addition, the early exposure of low HBV viral load affected HCC by the lag effect on HCC incidence (OR = 0.873, 95% CI = (0.853, 0.893)), slightly more dominant than the persistent low viral load on HCC incidence (OR = 0.918, 95% CI = (0.896, 0.941)).