Background: As a life-threatening complication in patients undergoing surgery, acute kidney injury (AKI) is strongly associated with a worse prognosis. Urinary dickkopf-related protein 3 (DKK3) has been identified as a biomarker for predicting postoperative AKI in patients undergoing cardiac surgery.
Objective: To investigate the predictive value of urinary DKK3 on postoperative AKI and develop a clinical model based on the predictor for predicting the development of AKI within seven days for patients undergoing noncardiac surgery.
Methods: All patients who were admitted to the Intensive Care Unit (ICU) after noncardiac surgery from March 2023 to June 2023 were included in this study. The patients' baseline data on demographic characteristics, lifestyle risk factors, medical history, clinical features, and laboratory tests before surgery were collected at the time of admission. Besides, the blood samples for cystatin C and routine laboratory tests and the urine samples for DKK3 tests were simultaneously collected at the time of admission to the ICU. In addition, the independent predictors of postoperative AKI were identified by univariate, multivariate, and LASSO analyses. Moreover, a nomogram for predicting postoperative AKI was developed based on these independent predictors. Finally, the nomogram was evaluated through calibration and decision curve analyses.
Results: A total of 166 patients with a median age of 67 years old were included in this study, including 94 (56.63%) males. Among these patients, 47 patients (28.3%) developed postoperative AKI. Additionally, 7 independent risk factors, including preoperative serum creatinine, preoperative eGFR, preoperative serum albumin, preoperative serum potassium ion, cystatin C, uDKK3/uCr, and SOFA score, were selected by univariate and multivariate regression analyses. Eventually, 4 independent risk factors (including preoperative eGFR, cystatin C, uDKK3/uCr, and SOFA score) identified in this study by LASSO analyses were used to establish the nomogram. The area under the receiver operating characteristic (ROC) curve (AUC) for the prediction model was 0.868. The calibration curve and decision curve analysis results demonstrated that the nomogram had good prediction performance.
Conclusions: Urinary DKK3/creatinine was independently associated with postoperative AKI for patients in the ICU after noncardiac surgery. The nomogram constructed based on uDKK3/uCr, preoperative eGFR, cystatin C, and SOFA score showed a higher accuracy in predicting postoperative AKI.