Causal Analyses of Statin to Prevent Liver Disease Progression: A Nationwide Study Using Superlearning Targeted Maximum Likelihood Estimation

Abstract

Many studies have shown that statins reduce the risk of progression to liver cirrhosis (LC) and hepatocellularcarcinoma (HCC) among at-risk populations. However, causality has not been proved. This study examined whether statinscould prevent LC and HCC in patients with progressive and worsening chronic liver disease, using a robust methodologyfor causality. Between 2002 and 2013, 52,145 patients with chronic liver diseases were identified from the National HealthInsurance Service database in South Korea. The inverse probability weighting (IPW) and superlearning targeted maximumlikelihood estimation (TMLE) were used to assess the causality of statin use on the risk of LC and HCC, adjusting forsex, age, comorbidities, and co-medications. Multivariable superlearning TMLE revealed that statin use was associated withreduction in the incidence risk of LC (Marginal odds ratio (MOR) 0.59, 95% confidence interval [CI] 0.50-0.65) and HCC(MOR 0.59, 95% CI 0.50-0.67). Such a protective effect was more evident with atorvastatin and lipophilic statin. Thispopulation-based observational study indicated the benefit of statin use, particularly atorvastatin and lipophilic statin, forcausally reducing the risk of LC and HCC.