INTRODUCTION: Cardiorenal syndrome (CRS) is a common and critical complication of sepsis, with high morbidity and mortality rates. Studies on biomarkers for the early prediction of septic CRS are sporadic. Classic and novel potential biomarkers were identified to explore their diagnostic performance of in patients with septic CRS. METHODS: A total of 138 patients with sepsis from Peking University People's Hospital were enrolled in this prospective observational study, which was conducted between May 2019 and June 2022. The patients were divided into non-CRS(n=106) and CRS (n=32) groups. Serum levels of cystatin C, KIM-1, NGAL and α-Klotho were detected at admission using enzyme-linked immunosorbent assay (ELISA). The relationship between the biomarker levels and risk factors of CRS were analyzed, as well as discrimination accuracy comparisons were performed. RESULTS: The incidence of CRS in patients with sepsis was 23.2% (32/138) during hospitalization, with an obvious mortality. Compared with the non-CRS group, serum cystatin C, BNP, TNI, KIM-1, and NGAL levels were both significantly elevated at admission in patients with sepsis complicated with CRS. Logistic regression analysis revealed that BNP, TNI cystatin C albumin, Lac, D-dimer were risk factors for CRS in sepsis patients. Compared with other biomarkers, serum cystatin C had moderate discriminative power for predicting septic CRS (AUROC 0.746, sensitivity 0.719, specificity 0.783). BNP combined with cystatin C and D-dimer demonstrated an excellent discrimination performance, for its AUROC was up to 0.878 (sensitivity, 0.844
specificity, 0.759). CONCLUSION: Serum cystatin C, BNP, TNI, KIM-1, and NGAL levels are elevated in patients with septic CRS. Our study provides reliable evidence that cystatin C in combination with BNP and D-dimer might better predict septic CRS upon admission. Further research on sensitive biomarkers is needed.