Derivation and validation of a prediction score for acute kidney injury in patients hospitalized with acute heart failure in a Chinese cohort

AIMS: Acute kidney injury (AKI) is a common complication among patients hospitalized for acute heart failure (AHF), and is associated with increased mortality. The goal of this study was to derive and validate a prediction score for AKI in AHF patients. METHODS: The hospital medical records of 1709 patients with AHF were reviewed. AKI was defined as an increase in serum creatinine (SCr) of ≥26.4 μmol/L or ≥50% within 48 h. A multivariate logistic regression analysis was undertaken to develop a new prediction score. The area under the receiver operating characteristic (ROC) curve and the Hosmer-Lemeshow goodness-of-fit statistic test were calculated to assess the discrimination and calibration of the prediction score, respectively. RESULTS: Acute kidney injury developed in 32.2% of patients with AHF. Factors independently associated with the risk of AKI included: ≥70 years of age, ≥3 previous hospital admissions for AHF, systolic blood pressure <90 mmHg, serum sodium <130 mmol/L, heart functional class IV, proteinuria, SCr ≥104 μmol/L and intravenous furosemide dose ≥80 mg/day. A prediction score for AKI was derived based on the β coefficients of each risk factor. Patients with ≥8 points would be considered at high risk for development of AKI (55.1% incidence vs 18% in those with <8 points, P < 0.001). Both the derived and validated datasets showed adequate discrimination (area under ROC curve was 0.76 in both datasets) and calibration (Hosmer-Lemeshow statistic test, P = 0.98 and 0.13, respectively). CONCLUSION: The newly derived and validated clinical prediction score may effectively predict AKI in the patients hospitalized with AHF.