Journal of cardiothoracic and vascular anesthesia | 2021
Prediction Scores for Any-Stage and Stage-3 Acute Kidney Injury After Adult Cardiac Surgery in a Chinese Population.
Abstract
OBJECTIVES\nThis study was performed to internally derive and then validate risk score systems using preoperative and intraoperative variables to predict the occurrence of any-stage (stage 1, 2, 3) and stage-3 acute kidney injury (AKI) within seven days of cardiac surgery.\n\n\nDESIGN\nSingle-center, retrospective, observational study.\n\n\nSETTING\nSingle, large, tertiary care center.\n\n\nPARTICIPANTS\nAdult patients undergoing open cardiac surgery between January 1, 2012, and January 1, 2019.\n\n\nMEASUREMENTS AND MAIN RESULTS\nThe clinical data were divided into the following two groups: a derivation cohort (n\u202f=\u202f43,799) and a validation cohort (n\u202f=\u202f14,600). AKI was defined using the Kidney Disease: Improving Global Outcomes criteria. Multivariate logistic regression analysis was used to develop the prediction models. The overall prevalence of any-stage AKI and stage-3 AKI after cardiac surgery were 34.3% and 1.7%, respectively. The discriminatory ability of the any-stage AKI prediction model measured with the area under the curve (AUC) was acceptable (AUC\u202f=\u202f0.69, 95% confidence interval 0.68-0.69), and the calibration measured with the Hosmer-Lemeshow test was good (p\u202f=\u202f0.95). The AUC for the stage-3 AKI prediction model was 0.84 (95% confidence interval 0.83-0.85), and the Hosmer-Lemeshow test also indicated a good calibration (p\u202f=\u202f0.73).\n\n\nCONCLUSIONS\nThis research study, which used preoperative and intraoperative variables, derived and internally validated two predictive scoring systems for any-stage AKI and stage-3 AKI as defined by modified Kidney Disease: Improving Global Outcomes criteria using a very large cohort of Chinese cardiac surgical patients.