A validation study comparing existing prediction models of acute kidney injury in patients with acute heart failure

Area of Science:

• Nephrology
• Cardiology
• Clinical Prediction Models

Background:

• Acute kidney injury (AKI) frequently complicates acute heart failure (AHF).
• AKI in AHF is linked to longer hospital stays and higher mortality rates.
• Accurate prediction of AKI in AHF patients is essential for timely intervention.

Purpose of the Study:

• To externally validate existing prediction models for AKI in AHF patients.
• To identify the most accurate and reliable models for clinical use.

Main Methods:

• Utilized data from 10,364 AHF patients (2008-2018) from the Chang Gung Research Database.
• Defined AKI using the KDIGO criteria.
• Assessed model performance using Area Under the Receiver Operating Characteristic (AUC) curves for discrimination and calibration.

Main Results:

• Five prediction models were validated.
• The Forman risk score and the Wang et al. model demonstrated superior discrimination and calibration.
• AUCs for the Wang et al. model were 0.73 (AKI), 0.858 (AKI stage 3), and 0.845 (dialysis within 7 days).
• AUCs for the Forman risk score were 0.696 (AKI), 0.829 (AKI stage 3), and 0.817 (dialysis within 7 days).

Conclusions:

• The Forman risk score and Wang et al. model are effective tools for predicting AKI in AHF.
• These models offer simple yet accurate methods for risk assessment in clinical practice.