Genetic variants and the risk of renal impairment in decompensated cirrhosis: a multi-ancestry genome-wide association study.

BACKGROUND: Renal impairment is a major complication of decompensated cirrhosis and a key driver of mortality. Despite its clinical significance, the genetic architecture of susceptibility to renal impairment in this context remains largely unexplored. We conducted a multi-ancestry genome-wide association study (GWAS) to identify genetic loci associated with renal impairment in decompensated cirrhosis. METHODS: We performed a GWAS for serum creatinine levels in 3220 patients with decompensated cirrhosis across five cohorts (CANONIC [n=707], PREDICT [n=825], ACLARA [n=671], Bologna [n=272], and Brussels [n=745]); with 83% of patients having European and 17% having admixed American genetic ancestry. Findings were replicated in 1728 European patients with decompensated cirrhosis across four independent cohorts (Czech-Romanian [n=988], CirVir-CIRRAL [n=139], French-Belgian [n=294], and German [n=307]). Associations were evaluated in 3139 patients with compensated cirrhosis (EuroLiver [n=851], CirVir-CIRRAL [n=1142], and the UK Biobank [n=1146]) and 402 683 individuals from the UK Biobank with no history of cirrhosis, who served as a general population comparator. Single-cell and spatial transcriptomics mapped GWAS associations to cell types in the liver, blood, and kidney. FINDINGS: We identified a genome-wide significant association at the SUSD1 locus (lead variant rs2099161[C→T]; trans-ancestry-discovery β=0·16 [95% CI 0·11 to 0·21], p=4·32 × 10-10; independent replication β=0·09 [0·03 to 0·16], p=5·34 × 10-3; combined meta-analysis β=0·13 [0·09 to 0·17], p=8·31 × 10-11; I2=0%). The effect was specific to decompensated cirrhosis, with no association in compensated cirrhosis (β= -0·0003 [-0·0460 to 0·0454], p=0·99) or the general population (β= -0·005 [-0·009 to -0·002], p=3·78 × 10-3). Each additional copy of the rs2099161(C) risk allele was associated with a higher risk of kidney failure at inclusion in the ACLARA, CANONIC, PREDICT, Bologna, Brussels, Czech-Romanian, and French-Belgian cohorts (n=4390, p=6·34 × 10-4). The risk was highest in CC homozygotes compared with patients with the TT genotype (odds ratio 1·60 [95% CI 1·21 to 2·11]). A similar pattern was observed for incident kidney failure at 1 week (CANONIC cohort n=638, p=0·035, CC vs TT genotypes 2·00 [1·10 to 3·65]), and terlipressin requirement (PREDICT, ACLARA, and Brussels cohorts, n=2161, p=0·027, CC vs TT genotypes 2·06 [1·08 to 3·93]). Transcriptomic analyses identified plasmacytoid dendritic cells as the primary SUSD1-expressing cell-type with elevated expression in patients with renal impairment across compartments. Patients with renal impairment showed increased abundance of plasmacytoid dendritic cells and type-I-interferon signalling in liver and kidney tissues. INTERPRETATION: SUSD1 is a risk locus for renal impairment in decompensated cirrhosis, with genetic effects that appear context-dependent across cirrhosis stages and the general population. Our findings implicate the plasmacytoid dendritic cell-interferon axis as a candidate mechanism linking liver dysfunction to kidney injury. FUNDING: European Foundation for the Study of Chronic Liver Failure and Horizon Europe.