Prediction model for safe contrast volume thresholds to prevent postcontrast acute kidney injury after endovascular abdominal aortic aneurysm repair

Abstract

Objective: Post-contrast acute kidney injury (PC-AKI) is a serious complication of endovascular abdominal aortic aneurysm repair (EVAR) associated with development of chronic kidney disease, prolonged hospital stay, and perioperative mortality. Iodinated contrast is a known risk factor for PC-AKI but is a technical necessity for EVAR. The optimal volume of contrast necessary to minimize risk of PC-AKI in a patient undergoing elective EVAR is unknown. This study examines the incidence and significance of PC-AKI after EVAR and derives a patient-specific model to determine the optimal volume of contrast the surgeon should administer to mitigate the risk of PC-AKI. Methods: The Vascular Quality Initiative database was queried for patients who underwent elective EVAR. Patients with history of dialysis, kidney transplant, intraoperative coverage of a renal artery or renal artery stenting, or open conversion were excluded. Patients were stratified by development of PC-AKI. Patient characteristics were compared using bivariate and then multivariable logistic regression analysis to select key significant variables. A prediction model was developed for PC-AKI using a split-sample approach with a model training dataset and a validation dataset. The 30-day postoperative mortality for this same patient cohort was evaluated using bivariate and multivariable logistic regression analysis. Kaplan-Meier curve analysis compared survival between groups. Results: Among 49,417 patients undergoing elective EVAR, 2.6% (n = 1300) developed PC-AKI. Patients with PC-AKI were older with higher incidence of preoperative comorbidities and developed significantly more postoperative complications, including 30-day mortality (12% vs 0.5%; P < .001), compared with patients without PC-AKI. Patients who developed PC-AKI also had lower survival on Kaplan-Meier analysis (2-year survival: 74.5% vs 82.8%; P < .001). Regression analysis accounted for all other factors and showed that PC-AKI was independently associated with perioperative mortality (odds ratio [OR], 8.79; 95% confidence interval [CI], 6.60-11.64). On multivariable logistic regression, PC-AKI was independently associated with volume of contrast given (OR per mL, 1.005; 95% CI, 1.004-1.006) translating to a 5% increased PC-AKI risk for every 10 mL of contrast administered. The prediction model (area under the curve = 0.732; 95% CI, 0.706-0.759) provides surgeons with a recommended patient-specific safe volume of contrast that minimizes the risk of PC-AKI based on preoperative patient characteristics. Conclusions: PC-AKI has significant impact on patient morbidity and mortality after elective EVAR. The model provided utilizes 13 simple, patient-specific variables to generate a recommended contrast volume to minimize PC-AKI risk. In the absence of established guidelines for optimal volume of contrast to administer during elective EVAR, this model can serve as an important guide for surgeons prior to performing EVAR.