Abstract

Background: Statins may very rarely cause drug induced liver injury (DILI) but its pathogenesis is not well understood. Genetic factors may play a role in the pathogenesis of liver injury due to selected agents. In this study, we investigated genetic factors associated with statin DILI through a genome wide association study (GWAS).

Methods: We conducted a GWAS of statin DILI. Statin DILI cases with definite, highly likely or probable causality scores were identified from Drug Induced Liver Injury Network (DILIN) studies between 2004 and 2022. Controls were statin treated patients with no DILI (statin controls) identified from Indiana Biobank (IB). Statin controls were defined as having ALT < 45 U/L and Alk P < 125 U/L during the 12 months preceding and 6 months following statin prescription and absence of ICD code 573.8 and K71. Imputed, QC passing variants from the cases and controls were combined then re-cleaned to adjust for any potential batch effects. We compared statin DILI cases to statin controls as primary GWAS then compared top three suggestive hits in 1) non-statin DILI from DILIN studies (DILI controls) and 2) not statin exposed individuals with no evidence for liver injury from Indiana Biobank (IB Controls). Association tests were conducted in cross ancestry group and EUR ancestry specific fashion while adjusting for age, sex, and several principal components. Logistic regression model with Firth correction was used.

Results:

There were 71 statin-DILI cases with definite, highly likely or probable causality and 552 statin controls. After QC, there were 5,483,580 variants available for GWAS. In both cross-ancestry and EUR ancestry specific analysis we observed 3 suggestive (p < 5e^-06) associations (Table 1). These associations were more prominent in EUR ancestry specific analyses. In the EUR comparison, rs35197737, variant in RGS1 (regulator of G Protein signaling 1), showed a strong association with statin-DILI (OR=5.03, (95% C.I.=2.77- 9.13), p=1.14x10^-07). On chromosome 10, rs75629598 in FRMD4A (FERM domain containing 4A) showed significant association with statin DILI (OR= 4.35 (95% CI: 2.33-8.12), p=3.97 x10^-06). Lastly, rs7658630, intergenic variant, was associated with an increased risk (p-value=2.68 x10^-07, OR=4.85, 95% CI: 2.66-8.85) for statin-DILI.

The comparisons between statin DILI and DILI controls and between statin DILI and IB controls are also shown in Table 1. RGS rs35197737 showed genome-wide significant association with statin DILI in comparison to IB controls (OR: 5.74, 95% C.I. 3.29-10.00, p=7.22 x10^-10) and strong association in comparison to DILI controls (OR: 4.11, 95%: 2.42-7.00, p=1.88x10^-07).

Conclusion:

rs35197737 in RGS1 (Chr 1) appears to be a significant risk factor for statin DILI. Further studies to confirm this observation and to understand the mechanisms by which it contributes to statin DILI are warranted.