A Phenome-Wide Association Study of genes associated with COVID-19 severity reveals shared genetics with complex diseases in the Million Veteran Program

BackgroundThere are certain pre-existing conditions which leads in increased risk of coronavirus disease 2019 (COVID-19) severity and mortality. The objective of this study is to determine shared genetic architecture between COVID-19 severity and other medical conditions using electronic health record (EHR) data from diverse patient populations. Methods and FindingsWe conducted Phenome-wide association study (PheWAS) of genetic variants associated with severe COVID-19 in two biobanks with EHR and genomic data: 1) Veteran Affairs (VA) Million Veteran Program (MVP), 2) United Kingdom Biobank (UKBB). Genetic variants associated with critical illness (n=48) or hospitalization (n=39) due to COVID-19 from COVID-19 Host Genetics Initiative genome-wide association studies. Phenotypes defined by International Classification of Diseases (ICD) codes mapped to clinically relevant groups using published PheWAS methods; pre-COVID-19 data used to avoid potential confounding. Among 455,683 US Veterans from MVP, variants associated with severe COVID-19 tested for association across 1,559 phenotypes; 353,365 UK Biobank participants, and 1,064 phenotypes tested. Genetic variants at ABO locus (rs550057, rs505922) associated with the largest number of phenotypes (nrs55057= 53 and nrs505922=61); strongest association with venous embolism, odds ratio (OR)rs550057 1.27 (p=5.28 x 10-116), and thrombosis ORrs505922 1.31, p=3.5 x10-183. Among 67 respiratory conditions tested, only idiopathic pulmonary fibrosis, OR rs2277732 1.17, p=1.3410-05, and asthma ORrs143334143 0.94, p=2.31 x10-04, shared variants with severe COVID-19. The RAVER1 locus (rs74956615) associated with reduced risk for autoimmune conditions, e.g. psoriasis OR 0.71, p= 1.53 x10-22, rheumatoid arthritis, OR 0.78, p=1.04 x 10-09; findings replicated in UKBB. A known functional missense variant (rs34536443, TYK2) in the region had the highest linkage disequilibrium with rs74956615, suggesting signal was likely from TYK2. In MVP, PheWAS results stratified by genetic ancestry did not demonstrate significant difference in associations across ancestry. ConclusionsShared genetic architecture between COVID-19 severity and conditions related to underlying risk factors for severe and poor COVID-19 outcomes; associations similar across genetic ancestries. Divergent association between inflammatory conditions and severe COVID may be explained by known pathway impairing signaling of inflammatory cytokines, reducing risk for autoimmunity; same pathway reduces type 1 interferon signaling, critical for viral host defense. Caution needed when targeting pathways that may balance immune tolerance and immunodeficiency to treat COVID-19.