Abstract
INTRODUCTION: Limited understanding of kidney dysfunction types and acute kidney injury (AKI) recovery exists due to biopsy risks in decompensated cirrhosis patients. To inform this, we analyzed aptamer-based proteomics and metabolomics to differentiate by AKI diagnosis and recovery. METHODS: A case-control study of 97 patients hospitalized at our institution. We performed aptamer-based proteomics and metabolomics on serum biospecimens obtained within 72 h of admission. We compared the proteome and metabolome by the AKI phenotype (i.e., HRS-AKI, ATN) and by AKI recovery (decrease in sCr within 0.3 mg/dL of baseline) using ANCOVA analyses adjusting for demographics and clinical characteristics. We completed Random Forest (RF) analyses to identify metabolites and proteins associated with AKI phenotype and recovery. Lasso regression models were developed to highlight metabolites and proteins that could improve diagnostic accuracy. RESULTS: ANCOVA analyses showed no metabolomic or proteomic differences by AKI phenotype while identifying differences by AKI recovery status. Our RF and Lasso analyses showed that metabolomics can improve the diagnostic accuracy of both AKI diagnosis and recovery, and aptamer-based proteomics can enhance the diagnostic accuracy of AKI recovery. DISCUSSION: Our analyses provide novel insight into pathophysiologic pathways, highlighting the metabolomic and proteomic similarities between patients with cirrhosis with HRS-AKI and ATN while also identifying differences between those with and without AKI recovery.