RESUMO
INTRODUCTION: Cardiac surgery-associated acute kidney injury (CSA-AKI) is associated with increased morbidity and mortality. OBJECTIVES: We aimed to identify potentially modifiable risk factors for CSA-AKI. METHODS: This was asingle-center retrospective cohort study of 495 adult patients undergoing cardiac surgery. AKI was diagnosed and staged using full KDIGO criteria incorporating baseline serum creatinine (SC) levels and correction of postoperative SC levels for fluid balance. We examined the association of routinely available clinical and laboratory data with AKI using multivariate logistic regression modeling. RESULTS: A total of 103 (20.8%) patients developed AKI: 16 (15.5%) patients were diagnosed with AKI upon hospital admission, and 87 (84.5%) patients were diagnosed with CSA-AKI. Correction of SC levels for fluid balance increased the number of AKI cases to 104 (21.0%), with 6 patients categorized to different AKI stages. Univariate logistic regression analysis identified five preoperative (age, sex, diabetes mellitus, preoperative systolic pulmonary arterial pressure [PSPAP], acute decompensated heart failure) and five intraoperative predictors of AKI (age, sex, red blood cell [RBC] volume transfused, use of minimally invasive surgery, duration of cardiopulmonary bypass). When all preoperative and intraoperative variables were incorporated into one model, six predictors remained significant (age, sex, use of minimally invasive surgery, RBC volume transfused, PSPAP, duration of cardiopulmonary bypass). Model discrimination performance showed an area under the curve of 0.69 for the model including only preoperative variables, 0.76 for the model including only intraoperative variables, and 0.77 for the model including all preoperative and intraoperative variables. CONCLUSIONS: Use of minimally invasive surgery and therapies mitigating PSPAP and intraoperative blood loss may offer protection against CSA-AKI.