Air Embolism During TEVAR: Carbon Dioxide Flushing Decreases the Amount of Gas Released From Thoracic Stent-Grafts During Deployment

675621JETXXX10.1177/1526602816675621Journal of Endovascular TherapyRohlffs et al research-article2016 Experimental Investigation Air Embolism During TEVAR: Carbon Dioxide Flushing Decreases the Amount of Gas Released From Thoracic Stent-Grafts During Deployment Journal of Endovascular Therapy 2017, Vol. 24(1) 84 88 The Author(s) 2016 Reprints and permissions: sagepub.com/journalspermissions.nav DOI: 10.1177/1526602816675621 www.jevt.org Fiona Rohlffs, MD 1, Nikolaos Tsilimparis, MD, PhD 1, Vasilis Saleptsis, MD 1, Holger Diener, MD 1, E. Sebastian Debus, MD, PhD 1, and Tilo Kölbel, MD, PhD 1 Abstract Purpose: To investigate the amount of gas released from Zenith thoracic stent-grafts using standard saline flushing vs the carbon dioxide flushing technique. Methods: In an experimental bench setting, 20 thoracic stent-grafts were separated into 2 groups of 10 endografts. One group of grafts was flushed with 60 ml saline and the other group was flushed with carbon dioxide for 5 minutes followed by 60 ml saline. All grafts were deployed into a water-filled container with a curved plastic pipe; the deployment was recorded and released gas was measured using a calibrated setup. Results: Gas was released from all grafts in both study groups during endograft deployment. The average amount of released gas per graft was significantly lower in the study group with carbon dioxide flushing (0.79 vs 0.51 ml, p=0.005). Conclusion: Thoracic endografts release significant amounts of air during deployment if flushed according to the instructions for use. Application of carbon dioxide for the flushing of thoracic stent-grafts prior to standard saline flush significantly reduces the amount of gas released during deployment. The additional use of carbon dioxide should be considered as a standard flush technique for aortic stent-grafts, especially in those implanted in proximal aortic segments, to reduce the risk of air embolism and stroke. Keywords air embolism, aorta, aortic aneurysm, carbon dioxide, endovascular procedures, endovascular technique, stroke, thoracic aorta, thoracic endovascular aortic repair Introduction Endovascular strategies for the treatment of proximal thoracic aortic pathologies are increasingly applied in regular clinical practice. In the descending aorta, thoracic endovascular aortic repair (TEVAR) is considered the gold standard and has replaced open surgical therapy due to reduced mortality and morbidity. 1,2 Clinical experience is also increasing in endovascular treatment of pathologies involving the aortic arch and the ascending aorta. These new minimally invasive techniques broaden the spectrum of treatment options, especially for patients not suitable for open surgery. 3 5 Despite an overall decrease in morbidity and mortality due to the use of endovascular techniques, stroke remains a major drawback of TEVAR, with reported frequencies of 2.3% to 8.2% for standard tubular TEVAR and as high as 16% for more complex procedures. 6,7 The general hypothesis for the origin of stroke is release of atherosclerotic particles during wire manipulation and graft deployment, but this theory is not verified in studies of TEVAR procedures and does not explain the incidence of stroke in younger TEVAR patients without relevant atherosclerotic disease. Another potential source of stroke is embolism of air released during graft deployment. Air has been identified as the source of neurological deficits after other procedures, such as coronary artery bypass grafting, endoscopy, and interventional procedures, 8 10 although the etiology of stroke might be different compared to TEVAR. The role of air embolism in TEVAR is not well studied, but it may play an important role, especially for the frequent clinically silent lesions present on diffusion-weighted magnetic 1 German Aortic Center, University Heart Center, University Hospital Hamburg-Eppendorf, Hamburg, Germany Corresponding Author: Fiona Rohlffs, German Aortic Center, University Heart Center, University Hospital Hamburg-Eppendorf, Martinistraße 52, Hamburg 20246, Germany. Email: f.rohlffs@uke.de

Rohlffs et al 85 resonance imaging in a high percentage of patients after TEVAR. 11 Recently, the carbon dioxide flushing technique was introduced using high-pressure pure carbon dioxide from a cylinder prior to flushing the stent-graft with normal saline. 12 Using this technique, room air filled spaces around the stent-graft can be replaced by carbon dioxide in advance of the normal saline flush. Applied for complex proximal TEVAR, the use of carbon dioxide flush has been associated with a low cerebrovascular event rate in a group of 36 patients. 12 Nevertheless, this study included a small number of patients, and the results should be verified in a larger number of procedures. Physical and chemical characteristics of carbon dioxide include a low viscosity, a much higher solubility than nitrogen-rich ambient air, and immediate dissolution in blood. 13,14 Hence, flushing stent-grafts additionally with carbon dioxide potentially may reduce the amount of air trapped in and released from the stent-graft during TEVAR, thereby decreasing the risk of stroke. The present study investigates the amount of gas released from standard thoracic endografts after flushing according to the instructions for use (IFU), and the effect of carbon dioxide flushing on the amount of released gas in an experimental setting. Methods Study Design To study the amount of released gas, 20 identical tubular thoracic stent-grafts (Zenith TX2 ProForm; Cook Medical, Bloomington, IN, USA) were separated into 2 equal groups and deployed into a translucent water-filled tray with a 50-mm-wide curved testing pipe (Figure 1). A different flushing technique was applied to each group. The volume of released gas was collected, quantified, and compared between the groups. The deployment process was filmed, and relevant observations were documented. Preparation of the Deployment System The 50-L translucent tray was filled with regular tap water on room temperature (21 C). To better visualize the air bubbles, the water was stained light blue by adding 2 ml of methylene blue. For graft deployment, a 50-mm-wide curved testing pipe was positioned and fixed to the deployment tray, sealing equally with the water level. One end was fixed to the bottom of the tray to enable introduction of the stent-grafts without contamination of room air. A conical tip at the upper end of the testing pipe was fitted with a Luerlock connection just above the water level and attached to a 3-way stopcock. This was connected to a 60-cm tube with another 3-way stopcock attached to a 1-mL syringe and Figure 1. Draft of the experimental setup. The 50-mmwide plastic testing pipe was fixed to the bottom of the tray. To measure the amount of gas released during stent-graft deployment, the conical tip of the testing pipe was connected to a 60-cm line with 20-mL and 1-mL syringes. To ensure collection of the total amount of released gas, the dilator tip of each stentgraft was advanced into the conical section of the testing pipe, and the total length of the graft was released into the vertical part of the testing pipe. another 20-mL syringe (Figure 1). The 20-mL syringe was used to fill the line and the conical tip of the plastic pipe with water so that gas was removed from the system. Bubbles on the wall of the testing pipe or the deployment tray were wiped away. The 1-mL syringe contained a scale enabling measurement of volumes in 0.01-mL increments. With this setup, the released air from the deployed stentgraft could be collected into the conical tip of the plastic pipe and aspirated through the 60-cm line into the 20-mL syringe. After collecting all gas into the 20-mL syringe, it was transferred into the 1-mL syringe for accurate measurement. To calibrate the setup, known amounts of air were applied in advance and measured using the syringe maneuver. This technique could accurately measure air amounts down to 0.02 ml. After each measurement, the system was completely cleared of any air. Stent-Grafts Twenty Zenith TX2 ProForm thoracic stent-grafts (ZDEG-PT-34-199-PF) were used in the experiment. These stent-grafts are loaded on a Z-Trak Plus introducer system with a 20-F hydrophilic sheath. The system contains a flushing chamber with a captor valve, which can be closed and opened, as well as a separate side port with a 2-way stopcock to apply the flush. The IFU recommend flushing of this stent-graft with 60 ml of saline prior to deployment.

86 Journal of Endovascular Therapy 24(1) Table 1. Overview on the Total Amount of Gas Released From Each Stent-Graft and Average Amount of Gas for All Stent-Grafts for Both Study Groups. a Graft Number Loss While Flushing Flush at Tip Loss While Resting Big Bubbles Small Bubbles Total Gas, ml Mean Total Gas, ml Group A, flush with 60 ml saline 0.79±0.2 A 1 +++ +++ +++ 0.91 A 2 +++ +++ +++ 0.65 A 3 +++ +++ +++ 0.85 A 4 +++ +++ +++ 0.62 A 5 + +++ +++ +++ 0.89 A 6 +++ +++ +++ 0.92 A 7 +++ + +++ +++ 0.87 A 8 +++ +++ +++ 1.02 A 9 +++ +++ +++ 0.89 A 10 +++ ++ ++ 0.34 Group B, flush with 100% CO 2 (5 minutes) and 60 ml saline 0.51±0.19 B 1 +++ ++ ++ 0.51 B 2 +++ ++ ++ 0.28 B 3 +++ ++ ++ 0.67 B 4 +++ ++ ++ 0.60 B 5 + +++ +++ ++ 0.86 B 6 +++ + ++ 0.49 B 7 +++ + ++ 0.25 B 8 +++ + ++ 0.44 B 9 +++ ++ +++ 0.65 B 10 +++ + ++ 0.31 a Key: denotes none, + indicates few, ++ signifies intermediate, and +++ means many/much. Flushing and Deployment of the Stent-Grafts To prepare the grafts for flushing, the central cannula wire and peel-away sheath were removed from every stent-graft. The sidearm of the flushing chamber and the 2-way stopcock were firmly tightened. The captor valve was closed firmly and taped for flushing; the tip of the graft was elevated by 40. The flush was applied through the side arm of the flushing chamber. The 10 grafts in group A (A1 10) were flushed with 60 ml 0.9% saline. The 10 grafts of group B (B1 10) were flushed with carbon dioxide from a cylinder for 5 minutes then 60 ml of 0.9% saline. The 2-way stopcock was closed during change of the 20-mL syringes. The stent-grafts were allowed to rest for 10 minutes on the bench after flushing. They were then deployed by introducing each stent-graft system into the plastic pipe; the captor valve was opened. The stent-graft was released by retraction of the covering sheath according to the IFU. As described above, the released gas bubbles from each graft were collected at the conical tip of the plastic pipe, and the released gas was quantified with the 1-mL syringe. Additional documented observations included the release of large bubbles at the beginning of sheath retraction (large bubbles), the release of small bubbles from the total length of the graft (small bubbles), the loss of flush from the closed captor valve while flushing (loss while flushing), the drainage of flush from the tip of the graft while flushing (flush at tip), and the loss of flush from the closed captor valve while resting (loss while resting). The volumes of gas released by the stent-grafts were averaged and compared between groups using SPSS for Macintosh (version 22.0; IBM Corporation, Somers, NY, USA). Results Amounts of released gas and additional observations are summarized in Table 1. The average amount of released gas per graft was significantly (p=0.005) lower in group B compared with group A (0.51 vs 0.79 ml, respectively; Figure 2). In group B, the highest amount of released gas appeared in stent-graft B5, which showed a leak from the captor valve with loss of saline during flushing. In group B, large bubbles and small bubbles were found, but fewer than in group A. Large bubbles exclusively appeared at the tip of the graft when sheath retraction was started (Figure 3). Small bubbles were found along the length of the graft, having been released from the crimped fabric (Figure 4). Discussion This study demonstrates that measurable quantities of gas are released during deployment of tubular sheath constrained

Rohlffs et al 87 Figure 2. Box plot comparing the released gas amounts of both groups. The mean amount of gas released for group A (flush with 0.9% saline) was 0.79 ml. The mean amount for group B (carbon dioxide flushing technique) was 0.51 ml. The total amount of released gas was significantly lower for group B (p=0.005). Figure 4. Image of a collection of small bubbles (black arrow). Small bubbles were found at the stent-graft shaft, having been released from small spaces between crimps in the fabric. Figure 3. (A-C) Sequence of a big bubble (black arrows) being released from the tip of the graft immediately after sheath retraction was begun. Big bubbles exclusively appeared at the tip of the graft. thoracic stent-grafts in an experimental setting, which supports the hypothesis that air embolism may play a role as a source of perioperative stroke in TEVAR. This finding is in accord with a study by Bismuth et al 15 from 2011, in which transcranial Doppler during TEVAR detected microembolic events and showed the highest rate during stent-graft opening and not during wire manipulation in landing zones 0 2. 15 Furthermore, our findings demonstrate that carbon dioxide flushing significantly reduces the amount of gas released from thoracic stent-grafts during deployment. One potential explanation of this finding is the ability of carbon dioxide to dissolve in the saline solution during the resting period. Another potential explanation is the influence of the different viscosities of the gases on the ability of a saline flushing maneuver to drive out the gas from the stent-graft assembly. In addition to the reduced volume of gas, the less harmful properties of carbon dioxide may account for a potential reduction in the size and number of strokes when used in proximal TEVAR cases. Adding carbon dioxide as a standard flush technique does not add any obvious additional risk for the patient as there is no known allergic reaction, no recirculation, or renal toxicity. Applied in controlled amounts, carbon dioxide, especially in large vessels, is known to cause less harm compared to room air and rapidly dissolves in blood. 13,14 Carbon dioxide is eliminated through the lungs and larger amounts do not cause acid-base disturbances. 12 Since carbon dioxide is heavier than room air, we suggested elevating the tip of the graft during carbon dioxide flushing so that a column of carbon dioxide gas can be built from the bottom to the tip of the graft, avoiding formation of a potential room air level in the sheath of the graft. The risk of release of trapped air from endografts is increasingly important for more complex devices, such as fenestrated or branched stent-grafts. Factors such as degree of compression of the endografts may influence the amount of this so-called trapped air within the graft. As endovascular therapy is increasingly applied in the aortic arch and ascending aorta using fenestrated and branched techniques, this consideration necessitates revisiting potential side effects of air embolism. 16 To use carbon dioxide in the thoracic aorta as a contrast medium in large amounts (40 100 ml) is contraindicated. However, it is even more hazardous to release room air in

88 Journal of Endovascular Therapy 24(1) this location, even though it can serve as a good contrast agent in cadavers. 17 To clarify: we do not discuss the use of carbon dioxide as a contrast agent but consider carbon dioxide to be the preferred gas compared with room air when released in small amounts from endografts in the proximal aorta. Limitations Our findings were produced in an experimental setting using stent-grafts from only one manufacturer. Moreover, the released gas was not analyzed, so no conclusion can be drawn about whether carbon dioxide flushing succeeded in replacing the room air. However, the recently published clinical finding of a surprisingly low stroke rate after complex aortic arch stent-grafts flushed with carbon dioxide enforces this suggestion. 12 Conclusion Thoracic endografts release significant amounts of air during deployment if flushed according to the IFU with saline alone. In our study, additional use of carbon dioxide prior to flushing with saline significantly reduced the amount of gas released during stent-graft deployment. Its use should be considered as an adjunct to the standard flushing technique for aortic stent-grafts, especially in proximal aortic segments to reduce the risk of air embolism and stroke. Declaration of Conflicting Interests The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Tilo Kölbel acts as a proctor for and has intellectual property with Cook Medical. He also receives travel and research grants from Cook Medical. Equipment and stent-grafts were provided by Cook Medical. Funding The author(s) received no financial support for the research, authorship, and/or publication of this article. References 1. Cheng D, Martin J, Shennib H, et al. Endovascular aortic repair versus open surgical repair for descending thoracic aortic disease a systematic review and meta-analysis of comparative studies. J Am Coll Cardiol. 2010;55:986 1001. 2. Matsumura JS, Melissano G, Cambria RP, et al. Five-year results of thoracic endovascular aortic repair with the Zenith TX2. J Vasc Surg. 2014;60:1 10. 3. Murphy EH, Stanley GA, Ilves M, et al. Thoracic endovascular repair (TEVAR) in the management of aortic arch pathology. Ann Vasc Surg. 2012;26:55 66. 4. Rohlffs F, Tsilimparis N, Detter C, et al. New advances in endovascular therapy: endovascular repair of a chronic DeBakey type II aortic dissection with a scalloped stent-graft designed for the ascending aorta. J Endovasc Ther. 2016;23:182 185. 5. Tsilimparis N, Debus ES, Oderich GS, et al. International experience with endovascular therapy of the ascending aorta with a dedicated endograft. J Vasc Surg. 2016;63:1476 1482. 6. Gutsche JT, Cheung AT, McGarvey ML, et al. Risk factors for perioperative stroke after thoracic endovascular aortic repair. Ann Thorac Surg. 2007;84:1195 1200. 7. Haulon S, Greenberg RK, Spear R, et al. Global experience with an inner branched arch endograft. J Thorac Cardiovasc Surg. 2014;148:1709 1716. 8. Lynch JE, Riley JB. Microemboli detection on extracorporeal bypass circuits. Perfusion. 2008;23:23 32. 9. Eoh EJ, Derrick B, Moon R. Cerebral arterial gas embolism during upper endoscopy. A A Case Rep. 2015;5:93 94. 10. Zakhari N, Castillo M, Torres C. Unusual cerebral emboli. Neuroimaging Clin N Am. 2016;26:147 163. 11. Kahlert P, Eggebrecht H, Janosi RA, et al. Silent cerebral ischemia after thoracic endovascular aortic repair: a neuroimaging study. Ann Thorac Surg. 2014;98:53 58. 12. Kölbel T, Rohlffs F, Wipper S, et al. Flushing technique to prevent cerebral arterial air embolism and stroke during TEVAR. J Endovasc Ther. 2016;23:393 395. 13. Kerns SR, Hawkins IF. Carbon dioxide digital subtraction angiography: expanding applications and technical evolution. AJR Am J Roentgenol. 1995;164:735 741. 14. Cho KJ. Carbon dioxide angiography: scientific principles and practice. Vasc Specialist Int. 2015;31:67 80. 15. Bismuth J, Garami Z, Anaya-Ayala JE, et al. Transcranial Doppler findings during thoracic endovascular aortic repair. J Vasc Surg. 2011;54:364 369. 16. Vallabhajosyula P, Szeto WY. Current paradigms in aortic arch repair: Striking the balance between open surgery and endovascular repair. J Thorac Cardiovasc Surg. 2015;150:1399 1400. 17. Wipper S, Lohrenz C, Ahlbrecht O, et al. Antegrade side branch access in branched aortic arch endografts: a porcine feasibility study. J Endovasc Ther. 2013;20:233 241.