Obesity- and Lipid-Related Traits May Causally Contribute to Sepsis-Associated Acute Kidney Injury

OBJECTIVES: Acute kidney injury (AKI) is a major complication of sepsis resulting in substantial morbidity and mortality. We used genome-wide association study (GWAS) data together with Mendelian randomization (MR) analysis in multiple cohorts of different ancestries to identify traits potentially contributing to sepsis-associated AKI (Septic-AKI). DESIGN: Natural experiment and case-control study. SETTING: ICU patients, FinnGen (finngen.fi), and U.K. Biobank participants. PATIENTS: Adults with sepsis or septic shock. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: We conducted a discovery GWAS in the U.K. Biobank by selecting, after quality control, 4584 European Septic-AKI patients as cases, and 7090 European sepsis patients without AKI as controls. Causal inference analyses using two sample MR combined these GWAS results with Integrative Epidemiology Unit Open Genome-Wide Association Studies results for 118 clinical risk factors and 386 metabolites and, separately, for 13 lipid classes from FinnGen GWAS results. We tested for replication of positive findings in two independent genotyped septic shock cohorts (Vasopressin and Septic Shock Trial [VASST] cohort n = 632 and Japanese Septic Shock Cohort [Chiba] cohort n = 536). Variants in the GALNTL6 gene were associated with Septic-AKI (rs149773593; odds ratio = 2.18; p = 3.0 × 10 –8 ) in U.K. Biobank patients. GALNTL6 is associated with cardiometabolic traits, which we then focused on. Increased body mass index was associated with increased serum creatinine in both septic shock cohorts ( p < 0.0001). Obesity and metabolic traits, most frequently related to very low-density lipoprotein, were identified as potentially causal contributors to Septic-AKI. Combining these GWAS results with FinnGen GWAS results identified associations with Septic-AKI, which replicated in both independent septic shock cohorts, for cholesterol ester ( p = 4.8 × 10 –44 ), lysophosphatidylcholine ( p = 8.5 × 10 –54 ), phosphatidylcholine ( p = 2.7 × 10 –39 ), and phosphatidylethanolamine ( p = 2.1 × 10 –28 ). CONCLUSIONS: For Septic-AKI, GWAS identified a novel genetic susceptibility locus ( GALNTL6 ), which is also associated with cardiometabolic traits. We then further found that obesity- and lipid-related traits are possible contributors to the pathogenesis of Septic-AKI.