부산대학교 도서관

Background: The incidence of acute kidney injury (AKI) is greater than 50% among coronavirus disease-2019 (COVID-19) patients admitted to the intensive care unit (ICU). However, the literature on predictors and kinetics of renal recovery remains unclear. Patients and methods: This observational study was conducted in a 30-bedded mixed ICU of a tertiary care center from May 2020 to July 2021. A total of 200 consecutive adult COVID-19 patients who had AKI in ICU were included. Using logistic regression with the best subset selection, predictors of renal recovery were identified. Outcomes and kinetics of AKI recovery were determined. Results: Among 200 patients, 67 recovered from AKI, of which 38, 17, and 12 patients had transient AKI, persistent AKI, and acute kidney disease (AKD), respectively. A total of 25 patients had AKI relapse, primarily associated with hospital-acquired infections. Results of logistic regression showed that the combination of Acute Physiology and Chronic Health Evaluation (APACHE II) {odds ratio (OR) 1.1 [p < 0.001; 95% confidence interval (CI) 1.06--1.16]}, day onset of AKI [OR 1.6 (p = 0.001; 1.24--2.24)] and severity of AKI [OR 2.9 (p < 0.001; 2.03--4.36)] were the predictors associated with poor renal recovery. This model had sufficient discrimination with the area under the curve (AUC) of 0.86. Renal replacement therapy requirement and mortality among COVID--AKI patients were 68 and 84%, respectively. Conclusion: A higher APACHE II at admission, a longer time to onset of AKI, and the severity of AKI during ICU stay predicted poor renal recovery. Study results emphasize the need for stepping-up dialysis resources in the likely case of future waves of COVID-19. The relapse of AKI was associated with sepsis, and mortality rates were substantially high. [ABSTRACT FROM AUTHOR]