Establishing and validating a risk prediction model for acute kidney injury in patients with acute heart failure

Objective To explore the risk factors of acute kidney injury(AKI)in patients with acute heart failure(AHF)and establish a risk prediction model of facilitating an early identification of high risks for lowering the incidence of AKI.Methods From January 2016 to December 2017,a total of 386 AHF in patients at First Affiliated Hospital of Soochow University were selected as the study subjects.They were randomized into two groups of model(n=257)and verification(n=129).The risk factors of AKI were examined in AHF patients according to β scores of various factors.Through a risk prediction model,the area under curve(AUC)of receiver operating characteristic(ROC)and Hosmer-Lemeshow goodness-of-fit test were employed for evaluating resolutions and calibrations.Results The incidence of AKI was 38.9% in AHF patients.Logistic regression analysis revealed that estimated glomerular filtration rate(eGFR)<60 ml·min^-1·(1.73 m²)^-1 age ≥72 years,diastolic blood pressure(DBP)≤78 mmHg,fasting plasma glucose(FPG)≥6.0 mmol/L and uric acid(UA)≥430 μmol/L were independent risk factors for AKI in AHF patients.The above risk factors were β-scored on the basis of risk prediction model.The AUC of model and verification groups were 0.773 and 0.758 respectively,and the Hosmer-Lemeshow goodness-of-fit test P values were 0.806 and 0.785 respectively,indicating that the model had decent resolution and calibration.Conclusion The above risk prediction model based upon eGFR,age,DBP,FPG and UA has decent resolution and calibration.It should be further popularized in clinical practices.