A Risk Prediction Model for Contrast-Associated Acute Kidney Injury (CA-AKI)
Robert H. Seitter Pérez, Yi Mu, Bernard Rosner, Donald F. Chute, Shveta S. Motwani, Gary C. Curhan, Shruti Gupta
Abstract
Background: Cancer patients undergo frequent CT scans with contrast and may be uniquely predisposed to CA-AKI due to decreased effective circulating volume or concomitant treatment with nephrotoxic chemotherapy. Nevertheless, large-scale data regarding specific risk factors for CA-AKI in this population are lacking. Methods: We collected data on all CT scans with contrast obtained in adult cancer patients without ESKD from 2016 through 2020 at 2 large cancer centers. With each scan serving as an individual unit, we collected data on demographics, comorbidities, labs, and medications related to each scan. CA-AKI was defined either as a ≥0.3 mg/dl rise in serum creatinine (SCr) from baseline within 48 hours of the CT scan or a 1.5-fold rise in SCr to the peak measurement in the 14 days following the scan. Regression models accounting for correlated data were used to identify risk factors for CA-AKI. Results: CA-AKI occurred in 2435 of 46,593 scans (5.2%). Non-white race, contrast volume, diabetes mellitus, congestive heart failure, hypoalbuminemia, thrombocytopenia, baseline proteinuria, lower baseline eGFR, and use of diuretics and ACEI/ARBs were each associated with a higher risk of CA-AKI (Table), and the risk of CA-AKI progressively increased with a higher risk score (Figure). Table: - Variable and Score Assigned for CA-AKI Variable Odds Ratio Points Non-white race 1.28 2 Contrast volume 100 to 200 ml 1.13 1 Contrast volume >200 ml 1.86 6 Diabetes melhitus 1.28 2 Congestive heart failure 1.37 3 Serum albumin <3.0 g/dL 2.72 10 Serum albumin 3 to <3.5 g/dL 1.73 5 Platelets <150,000 per μL 1.30 3 Baseline proteinuria on urinalysis 1.31 3 eGFR <30 ml/min/1.73 m2 3.82 13 eGFR 30 to <60 ml/min/1.73 m2 1.45 4 Diuretic use 1.59 5 ACEI/ARB use 1.24 2 Conclusions: A clinically relevant scoring system is predictive of CA-AKI and can be used to help risk-stratify cancer patients undergoing CT scans with contrast. Funding: Commercial Support - GE Healthcare