Statin and the risk of hepatocellular carcinoma and death in a hospital-based hepatitis B-infected population: A propensity score landmark analysis

Abstract

Background & aims: The use of statin in hepatocellular carcinoma (HCC) and death prevention is still uncertain among hepatitis B infected (HBV) patients. This study aimed to examine the effect of statin on HCC and death in a HBV population.

Methods: We conducted a hospital-based population study of HBV patients, using the Hospital Authority database in Hong Kong. We defined statin use by landmark analysis to abrogate "immortal time bias" and propensity score (PS) weighting to minimize baseline confounders and "indication bias". Multiple imputations for missing data were performed. The weighted Cox regression analyses was performed for the risk of HCC (adjusting for competing mortality) and death.

Results: A total of 73,499 patients with a crude HCC incidence of 1.75 per 100 patient-years were entered into the 2-year landmark analysis. After landmark analysis and PS weighting of baseline covariates, statin users had a 32% risk reduction in HCC (weighted sub-hazard ratio (SHR) 0.68; 95% CI 0.48-0.97) compared to non-users. There was no decreased risk of death in statin users (weighted HR 0.92; 0.76-1.11, p=0.386). In subgroup analysis, concurrent statin and nucleos(t)ide analogue (NA) use was associated with 59% risk reduction in HCC (weighted SHR 0.41; 0.19-0.89, p=0.023) compared to NA use alone.

Conclusion: In this HBV cohort adjusted for confounders and biases, statin use is associated with reduced HCC risk by 32%. Additive HCC chemopreventive effect was seen with the concomitant use of NA and statin. Further prospective studies are warranted to investigate the potential use of statin in NA users.

Keywords: Alanine aminotransferase; Albumin; Death; Hepatitis B; Hepatocellular carcinoma; Landmark analysis; Nucleos(t)ide analogue; Statin.