GENOME-WIDE AND PATHWAY-SPECIFIC POLYGENIC EFFECTS ON CORTICOSTEROID RESPONSE IN PATIENTS WITH SEVERE INFECTIONS

Introduction: Corticosteroids (CS) improve survival in severe infections, but variable responses have been reported. We aimed to identify genetic variants that could be associated with the differential CS response. Methods: We performed a genome-wide association study (GWAS) of 90day mortality in a cohort of hospitalized COVID-19 patients treated with CS from the SCOURGE consortium (Ncases=375, Ncontrols=1,850). Logistic regression models were conducted in ?9.5M TOPMed-imputed variants (significance set at p=5x10-8). Then, we tested the polygenic overlap of the CS response between COVID-19 and all-cause sepsis using PRSice-2 to perform polygenic risk scores (PRS). The PRS were tested on the independent CS-treated all-cause sepsis patients (Ncases=21, Ncontrols=69) from the GEN-SEP cohort. The best PRS model was subject to pathway-specific PRS studies using PRSet to evaluate which pathways were related to the differential CS response. Results: Three PLCG2 variants, two intronic and one in the 3?-UTR region, were associated with CS response in COVID-19 (plowest=2.79x10?8). A PRS model including 31,374 variants reaching a pthreshold=0.0298 in the GWAS was nominally associated with CS response among all-cause sepsis patients (p=0.044). Pathway-specific PRS analyses revealed the regulation of the cholesterol biosynthesis pathway as the most significantly associated with CS response in all-cause sepsis (p=0.001). Conclusions: Overlapping polygenic effects of the CS response between COVID-19 and all-cause sepsis were observed. Genetic variants affecting the regulation of cholesterol biosynthesis pathway could underlie the CS response differences during critical illness.