Novel risk models to predict acute kidney disease and its outcomes in a Chinese hospitalized population with acute kidney injury
Ye-Qing Xiao,Wei Cheng,Xi Wu,Ping Yan,Li-Xin Feng,Ning-Ya Zhang,Xu-Wei Li,Xiang-Jie Duan,Hong-Shen Wang,Jin-Cheng Peng,Qian Liu,Fei Zhao,Ying-Hao Deng,Shi-Kun Yang,Song Feng,Shao-Bin Duan
DOI: https://doi.org/10.1038/s41598-020-72651-x
IF: 4.6
2020-09-24
Scientific Reports
Abstract:Abstract Acute kidney disease (AKD) is a state between acute kidney injury (AKI) and chronic kidney disease (CKD), but the prognosis of AKD is unclear and there are no risk-prediction tools to identify high-risk patients. 2,556 AKI patients were selected from 277,898 inpatients of three affiliated hospitals of Central South University from January 2015 to December 2015. The primary point was whether AKI patients developed AKD. The endpoint was death or end stage renal disease (ESRD) 90 days after AKI diagnosis. Multivariable Cox regression was used for 90-day mortality and two prediction models were established by using multivariable logistic regression. Our study found that the incidence of AKD was 53.17% (1,359/2,556), while the mortality rate and incidence of ESRD in AKD cohort was 19.13% (260/1,359) and 3.02% (41/1,359), respectively. Furthermore, adjusted hazard ratio of mortality for AKD versus no AKD was 1.980 (95% CI 1.427–2.747). In scoring model 1, age, gender, hepatorenal syndromes, organic kidney diseases, oliguria or anuria, respiratory failure, blood urea nitrogen (BUN) and acute kidney injury stage were independently associated with AKI progression into AKD. In addition, oliguria or anuria, respiratory failure, shock, central nervous system failure, malignancy, RDW-CV ≥ 13.7% were independent risk factors for death or ESRD in AKD patients in scoring model 2 (goodness-of fit, P 1 = 0.930, P 2 = 0.105; AUROC 1 = 0.879 (95% CI 0.862–0.896), AUROC 2 = 0.845 (95% CI 0.813–0.877), respectively). Thus, our study demonstrated AKD was independently associated with increased 90-day mortality in hospitalized AKI patients. A new prediction model system was able to predict AKD following AKI and 90-day prognosis of AKD patients to identify high-risk patients.
multidisciplinary sciences