Performance of contrast-associated acute kidney injury predictive risk models in Thai cardiac angiography or angioplasty patients

Abstract

Several risk prediction models of Contrast-associated acute kidney injury (CA-AKI) in patients undergoing cardiac angiography or angioplasty are available. However, the lack of extensive external validations limits generalizability and clinical acceptance. This study conducted the external validation of three CA-AKI predictive risk models (Chen's, Inohara's, and Tziakas' risk models) and determined the incidence of CA-AKI in Thai patients undergoing cardiac angiography or angioplasty. A total of 647 medical records of patients who underwent elective cardiac angiography (n=446) and angioplasty (n=201) were reviewed. Fifty-five percent were male, mean age 62.6±10.2 years, and mean estimated glomerular filtration rate (eGFR) 69.93±24.30 ml/min/1.73 m2). Incidents of CA-AKI, defined as an absolute increase of serum creatinine of at least 0.3 mg/dL within 48 hours or a relative increase of at least 50% within seven days after the procedure, were collected. The results showed that 78 patients (12.1%) had developed CA-AKI. Chen's, Inohara's, and Tziakas' predictive risk models exhibited low discriminative ability with c-statistic of 0.571, 0.551, and 0.530, respectively. Due to low discriminative capability, these risk models may have low sensitivity to predict CA-AKI in Thai patients undergoing elective cardiac angiography or angioplasty.