CO2 Automated Angiography in Endovascular Aortic Repair Preserves Renal Function to a Greater Extent Compared with Iodinated Contrast Medium. Analysis of Technical and Anatomical Details

Background: Contrast induced nephropathy occurs in up to 7.5% of cases in endovascular aortic repair (EVAR). Carbon dioxide (CO2) has been proposed as an alternative agent to iodinated contrast medium (ICM); however, specific protocols are not universally adopted, and the visualization of the renal arteries may be suboptimal in some cases. The aim of this study was to analyze our CO2-EVAR experience with automatic injections, in order to identify the anatomical characteristics associated with the best visualization of all the aortic vessels, with particular attention to the lowest renal artery (LoRA).Methods: From 2016 to 2019, all EVAR performed with either CO2 or ICM were analyzed and compared. CO2-EVAR was performed using an automated injector (600 mm Hg pressure; 100 cc volume); a small amount of ICM was injected in case of difficulty in LoRA visualization or doubts at the completion angiogram. Clinical and CT-Scan preoperative characteristics were considered. The study endpoints were technical success, amount of ICM and radiation dose, postoperative renal function and possible CO2-related adverse events. Statistical analysis was by Fisher's exact, t-Student, Mann-Whitney tests and ROC curve.Results: In the considered period, 321 EVAR procedures, 72 (22.4%) with CO2 and 249 (77.6%) with ICM, were performed. The 2 groups were similar for clinical characteristics and preoperative renal function. ICM was injected in a significantly lower amount in the CO2-EVAR group (52.8 +/- 6.1 vs. 88.1 +/- 9.2 cc, P < 0.001), which received a significantly higher mean radiation dose (Total DAP: 500,550.8 +/- 377,394.6 mGy/cm(2) CO2-EVAR vs. 332,301.8 +/- 230,139.3 mGy/cm(2) ICMEVAR, P = 0.001). Postoperative eGFR decreased significantly less in the CO2 -EVAR (2.3 +/- 1.1 mL/min) compared with the ICM-EVAR group (10.6 +/- 5.3 mL/min), P < 0.001. LoRA was correctly visualized in 50/72 (69.4%) cases of CO2-EVAR, which had a significantly longer proximal neck (Median [IQR]: 30 [14] vs. 18 [15] mm, P = 0.001). At ROC curve, a proximal neck length >24.5 mm was predictive of LoRA visualization (72.1% sensitivity, 73.8% specificity). Three CO2-EVAR cases had intraoperative transient hypotension with no consequences. Sixteen/72 (22.2%) CO2 EVAR procedures were performed using 0 cc of ICM.Conclusions: CO2-EVAR by automated injections is safe and requires a lower amount of ICM if compared with ICM-EVAR, with a consequent significant benefit on postoperative renal function. If specific anatomical situations are present, ICM may be completely unnecessary. The radiation dose is however significantly higher, therefore procedural protocols need further refinements.