ABSTRACT
Abstract Objective: To discover potentially modifiable perioperative predictors for renal replacement therapy (RRT) in patients with cardiac surgery-associated acute kidney injury (CSA-AKI). Methods: A cohort of 1773 consecutive cardiac surgery patients with postoperative acute kidney injury (AKI) from January 2013 to December 2015 were included retrospectively. AKI was defined according to the Kidney Disease: Improving Global Outcomes (KDIGO) criteria. The primary outcome was CSA-AKI requiring renal replacement therapy (AKI-RRT). The initiation of RRT was based on clinical judgment regarding severe volume overload, metabolic abnormality (e.g., acidosis, hyperkalemia), and oliguria. Patients with AKI-RRT were matched 1:1 with patients without AKI-RRT by a propensity score, to exclude the influence of patients' demographics, comorbidities, and baseline renal function. Multivariable regression was performed to identify the predictors in the matched sample. Results: AKI-RRT occurred in 4.4% of the entire cohort (n=78/1773), with 28.2% of in-hospital mortality (n=22/78). With the propensity score, 78 pairs of patients were matched 1:1 and the variables found to be predictors of AKI-RRT included the contrast exposure within 3 days before surgery (odds ratio [OR]=2.932), central venous pressure (CVP) >10 mmHg on intensive care unit (ICU) admission (OR=1.646 per mmHg increase), and erythrocyte transfusions on the 1st day of surgery (OR=1.742 per unit increase). Conclusion: AKI-RRT is associated with high mortality. The potentially modifiable predictors found in this study require concern and interventions to prevent CSA-AKI patients from worsening prognosis.