Feasibility and Technical Aspects of Proximal Nellix-in-Nellix Extension for Late Caudal Endograft Migration

677037JETXXX10.1177/1526602816677037Journal of Endovascular TherapyDonselaar et al research-article2016 A SAGE Publication Clinical Investigation Feasibility and Technical Aspects of Proximal Nellix-in-Nellix Extension for Late Caudal Endograft Migration Journal of Endovascular Therapy 2017, Vol. 24(2) 210 217 The Author(s) 2016 Reprints and permissions: sagepub.com/journalspermissions.nav DOI: 10.1177/1526602816677037 www.jevt.org Esmé J. Donselaar, MD 1, Andrew Holden, MBChB, FRANZCR, EBIR 2, Aleksandra C. Zoethout, BSc 3, Clark J. Zeebregts, MD, PhD 3, and Michel M. P. J. Reijnen, MD, PhD 1 Abstract Purpose: To describe the feasibility and technical aspects of a proximal Nellix-in-Nellix extension to treat caudal stentgraft migration after endovascular aneurysm sealing (EVAS) in the in vitro and in vivo settings. Methods: In vitro studies were designed (1) to assess inner diameters of Nellix-in-Nellix extensions after postdilation with 12-mm balloons and (2) to test wall apposition in tubes with different diameters using a Nellix-in-Nellix stent-graft that extended out of the original Nellix stent-graft lumen by 10, 20, 30, and 40 mm. Simulated-use experiments were performed using silicone models in conjunction with a pulsatile flow pump. In the clinical setting, 5 patients (median age 74 years, range 73 83) presented at 2 centers with type Ia endoleak secondary to caudal Nellix stent-graft migration measuring a median 9 mm (range 7 15) on the left and 7 mm (range 0-11) on the right. Median polymer fill volume at the initial EVAS procedure was 42.5 ml (range 25 71). The median time to reintervention with a proximal Nellix extension was 15 months (range 13 32). Results: In vitro, the inner diameters of the Nellix-in-Nellix extensions were consistent after postdilation. Cases with 10 and 20 mm of exposed endobag resulted in a poor seal with endoleak, while cases with 30 and 40 mm of exposed endobag length exhibited angiographic seal. Fill line pressures of the second Nellix were higher than expected. In the 5 clinical cases, chimney grafts were required in each case to create an adequate proximal landing zone. The Nellix-in-Nellix procedure was successful in all patients. There were no procedure-related complications, and no endoleaks were observed during a median 12-month follow-up. Reinterventions were performed in 2 patients because of in-stent stenosis and chimney graft compression, respectively. Conclusion: Proximal Nellix-in-Nellix extension can be used to treat caudally migrated Nellix stent-grafts and to treat the consequent type Ia endoleak, but the technique differs from primary EVAS. The development of dedicated proximal extensions is desirable. Keywords aortic aneurysm, endograft, endoleak, endovascular aneurysm sealing, Nellix-in-Nellix technique, proximal extension, stent-graft migration Introduction Endovascular aneurysm sealing (EVAS) of abdominal aortic aneurysms (AAA) was commercially introduced in early 2013 as an alternative endovascular treatment to reduce the reintervention rate after endovascular aneurysm repair (EVAR). 1 Since the introduction of the Nellix EndoVascular Aneurysm Sealing System (Endologix, Irvine, CA, USA), over 5000 patients have been treated, and the early results are promising. In the EVAS Investigational Device Exemption trial, 2 all 150 patients were treated within the instructions for use (IFU); there were no aneurysm ruptures, conversions, or limb thromboses within 30 days. Only 1 (0.7%) secondary intervention was performed to treat inadvertent coverage of a renal artery. Nine (6%) endoleaks were detected by the core laboratory at 30-day follow-up, including 1 type I and 8 type II endoleaks. 2 Similar good results were observed in the EVAS FORWARD GLOBAL registry recently reported by 1 Department of Surgery, Rijnstate Hospital, Arnhem, the Netherlands 2 Department of Radiology, Auckland City Hospital, Auckland, New Zealand 3 Department of Surgery, Division of Vascular Surgery, University Medical Center Groningen, University of Groningen, the Netherlands Corresponding Author: Michel M. P. J. Reijnen, Department of Surgery, Rijnstate Hospital, Wagnerlaan 55, 6815 AD Arnhem, the Netherlands. Email: mmpj.reijnen@gmail.com

Donselaar et al 211 Thompson et al. 3 The majority (72%) of the 277 patients were treated within the IFU. The 1-year estimates for freedom from secondary intervention and from any endoleak were 92% and 95%, respectively. Silingardi et al 4 concurrently reported the first midterm outcomes of 65 EVAS patients treated at 2 centers; their estimated 18-month freedom from secondary interventions was 95%. Although these early to midterm results are most encouraging, late caudal migration of the stents can occur. The incidence of this complication is unknown at present, but England et al 5 documented caudal migration 4 mm (mean 6.6 mm at 1 year) in 6 (17%) of 35 Nellix stentgrafts deployed in 18 patients; there were no associated endoleaks or distal migration. Both downward force on the polymer-filled endobags and lateral acceleration force within curvatures in the stent-grafts could contribute to loss of proximal stent-graft attachment, which could cause a flow channel (type Ia endoleak) to open adjacent to the endobag. In our experience, caudal migration is often related to a lateral shift of one or both Nellix stents within the aneurysm. Because dedicated proximal Nellix extensions do not exist today, the only solution currently is the use of a commercially available Nellix device, although it is outside the IFU. 6 The therapeutic application of Nellix-in-Nellix requires the placement of 1 Nellix system within the lumen of an already implanted Nellix system. This approach inherently limits the degree to which the second endobag may open as it is partially constrained within the lumen of the first implant. To determine any effect this may have on endobag filling and wall apposition, a benchtop deployment test was performed; the results are described along with a small case series. Methods In Vitro Postdilation and Endobag Geometry Tests To determine the effect of balloon postdilation on the inner diameter (ID) of the Nellix stent-graft, 3 Nellix devices were dilated at the proximal (n=3) and distal ends (n=2) using a 12-mm balloon-expandable stent delivery balloon (LifeStream; Bard Medical, Covington, GA, USA) and a custom-made Nellix balloon of the same design and materials as the standard 10-mm Nellix balloon but with a 13-mm diameter at nominal inflation. The implants were first deployed using the 10-mm Nellix balloons dilated to their nominal pressure and removed from the Nellix delivery system. The ID was measured using a pin gauge with 0.001-inch resolution (Meyer Gage Company, South Windsor, CT, USA), which inherently measures the minimal ID. The pin was inserted into the deployed Nellix device until the pin met slight resistance due to contact with the opposite wall of the stent. The pin gauge was inserted a minimum 30 mm into each sample, which was considered the 0 atmosphere (atm) state. From there, 1 of the 2 larger balloons was placed into the implant lumen and inflated to 3, 5, 7, 10, and 15 atm; the balloon was deflated between each pressure step to allow stent recoil and ID measurement. The outer diameter of the balloons at the given pressure increments was measured using a laser micrometer (AcuScan 5040; Beta Lasermike, Dayton, OH, USA) to characterize unconstrained balloon geometry. To analyze the geometry of the endobag in the extension Nellix device, a single Nellix stent-graft was deployed and the outer endobag was removed for ease of visualization. A second single Nellix was tracked within the lumen of the first Nellix stent-graft and extended out of the lumen a distance of 10, 20, 30, or 40 mm. This distance was measured from the top of the proximal stent-graft of the first system to the top of the inner proximal stent-graft of the extension system. The second Nellix was then inflated to a pressure of 60 mm Hg to simulate the pressure differential of an endobag fill pressure of 180 mm Hg in a patient with a systolic pressure of 120 mm Hg. The implant was then photographed in 2 orthogonal orientations, sagittal and coronal, with a ruler for scale. Given that the endobag is made from a flat piece of polyurethane and is not symmetrical, it exhibits different dimensions in different planes (Figure 1). Measurements were then performed on the photographs using ImageJ software (National Institutes of Health, Bethesda, MD, USA; http://rsb.info.nih.gov/ij/). This was repeated for 2 Nellix systems in 24.5-mm and 30-mm ID rigid acrylic tubes with similar device spacing (Figure 1A). The apposition to the wall of the tube was qualitatively examined and photographed. Simulated Use Deployments To confirm the results of the endobag geometrical analysis, simulated use experiments were performed in a similar manner, one deployment for each distance of Nellix-in- Nellix overlap (10, 20, 30, and 40 mm). Silicone anatomical models were used in conjunction with a pulsatile flow pump (SuperPump; Vivitro Labs Inc, Victoria, British Columbia, Canada). The system was tuned to a pressure of ~120/80 mm Hg, with 4.5- to 5-L/min flow at a rate of 70 beats/min. A first Nellix system was implanted at varying distances below the most caudal renal artery. Similar to the endobag geometrical analysis, the targeted distance from the renal artery to the first Nellix was 10, 20, 30, or 40 mm. A second 100-mm-long Nellix system was then tracked and deployed within the first system to a target location of just below the most caudal renal artery. The Nellix stent-graft balloons were inflated to their nominal pressure (7 atm) during deployment.

212 Journal of Endovascular Therapy 24(2) Figure 1. (A) Results of the endobag protrusion experiment with color coding corresponding to the distance from the original proximal stent-graft edge to the extension proximal stent-graft edge as shown in the schematic (B) of a 10-cm-long endobag design. All measurements are in millimeters. After implantation, multiplanar angiography was performed using a C-arm fluoroscopic system [OEC 9800 Plus (General Electric, Fairfield, CT, USA) or Mini C-arm (Hologic, Marlborough, MA, USA)] and 50% contrast (MD-76R; Mallinckrodt, Inc., Dublin, Ireland) delivered by a Medrad Mark V ProVis injector (Bayer HealthCare LLC, Whippany, NJ, USA) to confirm the presence or absence of seal. The nested Nellix system was unfurled by performing a prefill with water in an attempt to maximize endobag opening and obtain an accurate prefill volume estimate because fill line pinching may occur (Figure 2), causing artificially high pressures. The lumen volume per length of stent-graft was subtracted from the prefill volume and used as the target polymer volume. The prefill volume was calculated at 180 mm Hg fill pressure. Secondary fills were performed to a pressure of 250 mm Hg on implants demonstrating lack of seal after primary fill. Nellix-in-Nellix Case Series Five patients (median age 74 years; 3 men) with a type Ia endoleak due to caudally migrated EVAS stent-grafts were treated with a proximal Nellix extension at 2 institutions in the Netherlands and New Zealand, which together had Figure 2. (A) Nellix stent-graft internal diameter (ID) after recoil and as a function of balloon pressure for the 2 balloons used in the experiments. ATM, atmosphere; OD, outer diameter. (B) Nellix lumen design features. eptfe, expanded polytetrafluoroethylene; PET, polyethylene terephthalate. experience of 346 EVAS procedures. The baseline characteristics are given in Table 1. The AAAs in these patients had a median diameter of 57.6 mm (range 52 65) and a median neck length of 16 mm (range 1 24). Four of the

Donselaar et al 213 Table 1. Patient Characteristics. Median age, y 74 (range 73 83) Men 3 Hypertension 4 Diabetes mellitus 1 Dyslipidemia 2 History of stroke/tia 2 Aneurysm morphology per 1 2 3 4 5 patient a Neck length, mm 19 1 24 16 15 Diameter at lowest RA, mm 16 26 19 21 20 Diameter 1 cm below 18 43 22 21 22 lowest RA, mm Infrarenal angulation, deg 40 41 22 62 15 Aneurysm diameter, mm 52 65 58 55 60 Abbreviations: RA, renal artery; TIA, transient ischemic attack. a At initial treatment. Table 2. Preoperative Aneurysm Morphology Outside the Nellix Instructions for Use. Patient Variable Criteria per IFU 1 Neck diameter 16.3 mm 18 32 mm 2 Neck length 1 mm a 10 mm 4 β-angulation >60 60 5 Neck length 7 mm Right CIA occlusion b 10 mm CIA diameters 9 35 mm Abbreviations: IFU, instructions for use; CIA, common iliac artery. a Endovascular aneurysm sealing with one chimney was performed. b A single Nellix stent-graft was placed. initial EVAS procedures were performed outside the IFU because of challenging anatomy (Table 2). At the initial EVAS procedure, the median fill volume of the endobags was 42.5 ml (range 25 71), with a median fill pressure of 183 mm Hg (range 180 204). The full length of the infrarenal neck was not utilized due to low positioning of the stents, with an infrarenal neck seal length after implantation of 6.0 mm (range 4 8). There were no early postprocedural events or endoleaks. The median time from EVAS to reintervention was 15 months (range 13 32). Results In Vitro Studies In the postdilation experiment, the 2 balloons used for postdilation experiments expanded to different outer diameters at the same pressure. The resultant Nellix implant lumen, after recoil, at a given pressure was found to be very consistent (Figure 2A). The difference between the balloon outer diameter and resultant lumen ID was likely due to the low compliance polyethylene terephthalate (PET) sleeve within the endobag, which limits the degree of expansion (Figure 2B). Figure 3. A subtraction angiogram of the 20-mm extension simulated use deployment showing endoleak (arrow) between the 2 Nellix systems. In the endobag analysis, a schematic displaying endobag exposure with varying degrees of Nellix-in-Nellix extension is shown in Figure 1B. The 10-mm-wide colored bands correspond to the 10-, 20-, 30-, and 40-mm distance from the proximal edge of the stent-graft in Figure 1A. In the latter figure, the degree of bag exposure increases significantly from 20 to 30 mm of extension. At 30 mm of extension the bag is no longer tapered, while in the 10- and 20-mm exposure lengths only the tapered portion is exposed. In the lower portion of Figure 1A, 2 Nellix systems can be seen within a tube at 20, 30, and 40 mm of exposed length. The 10-mm exposure length is clearly lacking wall apposition. The 20-mm exposure length does have a slight gap or gutter in wall apposition at the 12 o clock position, not seen at 30- or 40-mm exposure lengths. In all simulated use experiments, some space between the proximal and distal endobag was observed. Similar results were found in simulated use models. The cases with 10 and 20 mm of exposed length resulted in poor seal with notable endoleak (Figure 3). The cases with 30 and 40 mm of exposed length exhibited angiographic seal. During simulated use testing, it was noted that fill line pressures may be higher than expected. This is possibly artificially high due to fill line pinching because of the fill line being sandwiched between the 2 stent frames (Figure 4A).

214 Journal of Endovascular Therapy 24(2) In all patients, one or more parallel grafts to the visceral arteries were required to create an appropriate seal length >20 mm (Figure 6). Two patients had a 3-vessel chimney EVAS (chevas) for the superior mesenteric artery (SMA) and bilateral renal arteries. In 2 patients, bilateral renal chimneys were required, and in 1 patient a single chevas was performed (right renal artery). The median procedure time was 185 minutes (range 135 247). Technical success was achieved in all patients, and no intervention-related complications occurred. The median hospitalization time after the procedure was 6 days (range 3 8). Through the last follow-up (median 12 months, range 4 15), there was 1 recurrent migration at 9 months that requires retreatment. Two patients underwent reinterventions for bilateral in-stent stenosis, likely caused by excessive endobag material, and chimney graft compression in the SMA, respectively. The in-stent stenoses were successfully treated using balloon-expandable covered stents. The other patient underwent redilation of the parallel graft. One patient with preexisting chronic renal insufficiency had deterioration of renal insufficiency, but dialysis was not required. Figure 4. (A) Fill line pinching is due to the line being sandwiched (arrow) between the 2 Nellix stent frames. (B) Fill line pressures during filling with the extension Nellix balloons inflated ( up ) or deflated ( down ). During prefill, the pressure may be higher when the Nellix balloons are inflated (Figure 4B) due to fill line compression. Pressure readings are ~200 (balloons deflated) and ~300 (balloons inflated) mm Hg even though the endobags were visually confirmed as not being full at this point and the flow loop system was generating only 120 mm Hg systolic pressure. Nellix-in-Nellix Case Series All 5 patients presented with a type Ia endoleak due to caudal stent-graft migration (Figure 5). The median increase in aneurysm sac diameter was 5 mm (range 0 8) at the time of diagnosis. The median migration distances of the left and right stents were 9 mm (range 7 15) and 7 mm (range 0 11), respectively. There were no cases of inadequate seal in the iliac arteries, and the distal seal positions had been stable throughout follow-up. Four patients were treated with a dual proximal Nellix extension and the fifth patient was treated only on the left side. In all cases, the proximal bare part of the Nellix stentgrafts was flared to 12 mm. Unfurling of the endobag and regular prefill were performed in all cases. The median fill volume of the endobags during reintervention was 14 ml (range 8 30) with a median fill pressure of 230 mm Hg (range 220 320). Discussion This study demonstrated that the application of a proximal Nellix-in-Nellix extension can effectively treat late caudal Nellix stent-graft migration. There are several important factors to be taken into account, as the technique differs from a primary EVAS procedure. Careful preoperative procedure planning is paramount, as the proper distance between the primary Nellix stentgraft and the intended sealing zone is vital for a good seal. Based on the in vitro study, the endobags of the Nellix extension should protrude at least 2 to 3 cm above the primary stent-graft to provide good wall apposition. Given this minimum sealing length requirement, parallel grafts to the renal arteries are often required, as was the case in all reported patients. The aortic diameter at the intended sealing zone must be a sufficient diameter to accommodate the Nellix and parallel chimney stents. An important step that differs from the primary Nellix EVAS procedure is the unfurling of the endobags prior to expanding the extension stent-grafts. This step is needed to get as much of the endobag as possible outside the primary Nellix stent-graft to optimize wall apposition of the endobags. When using the device for primary EVAS, the shape of the endobags will create a flat surface at the proximal seal zone. However, with a Nellix-in-Nellix extension, part of the endobag will be caught in between the stents, and the shape of the shoulder of the endobag will likely to be pointed downward, thus reducing the sealing zone. Unfurling may minimize this phenomenon, increasing the chances of long-term success. Unfurling may also reduce

Donselaar et al 215 Figure 5. Images from an 85-year-old woman originally treated with endovascular aneurysm sealing (EVAS) outside the instructions for use due to a conical neck with some thrombus. Axial computed tomography angiography (CTA) at (A) 1 month, (B) 1 year, and (C, D) 2 years after EVAS; A C are taken at the same level and show aneurysm growth. C and D show separation of the endobags. A type Ia endoleak (arrow) between the endobags is evident in D. Coronal CTA reconstructions at (E) 1 month and (F) 2 years after EVAS, which clearly shows caudal stent-graft migration. the amount of excess endobag material in between the stents that could cause a stenosis in the overlap zone, as was the case in one of our patients. The minimum required overlap between Nellix devices is unknown but is unlikely to be problematic as the shortest available stent-graft is 100 mm. An advantage of a longer overlap zone is that the entire stent-graft configuration gets stiffer, which may prevent further lateral movement of the stent-grafts.

216 Journal of Endovascular Therapy 24(2) Figure 6. (A) Angiography after polymer fill of the endobags with Nellix balloons and renal chimney balloons inflated. Note the complete seal with the superior mesenteric artery still filling. (B) Completion angiography confirming complete aneurysm seal with patent Nellix and renal artery stent-grafts. (C) Volume rendered reconstruction from the 1-month computed tomography angiography scan. In our patients, the proximal segments of the original Nellix stent-grafts were flared with a 12-mm balloon to avoid compression of the fill lines. This flaring can be performed only at the proximal bare segment as the PET sleeve around the stent-graft will limit overdilation. Whether this procedural step is needed is unknown. During primary EVAS, it is advocated to leave the Nellix balloons inflated during the entire procedure, especially in complex anatomy. When using the Nellix-in-Nellix application, however, this may interfere with the pressure measurements and give immediate high pressures. For this reason, the angioplasty balloons are deflated during polymer fill. A crucial step during primary EVAS is the prefill phase using saline solution with or without contrast. When filling a proximal Nellix-in-Nellix extension, the required volume to reach the intended 180 mm Hg is very low. In this situation, saline prefill may be counterproductive as much of the prefill may not be able to be aspirated, potentially interfering with the polymerization process. In addition, according to the manufacturer, a ratio of >10% contrast to polymer in the endobag might delay the polymerization process. In these low-volume applications, a contrast/polymer ratio >10% is quickly reached, so contrast should not be added to the saline solution used for unfurling. A small fill volume is associated with a steep volumepressure curve where small added volumes will lead to a steep increase in endobag pressure. For this reason, the polymer should be injected with maximum care to avoid endobag prolapse or inadvertent aortic wall injury. In the primary procedures of the 5 patients, the median fill volume of polymer was only 43 ml to get a median fill pressure of 183 mmhg, which is a low volume in comparison with the median used polymer fill of 75 ml in the Nellix pivotal trial. 2 Whether a low polymer volume is a risk factor for migration remains to be investigated. Limitations The in vitro experiments may not translate directly to clinical application and should be interpreted with caution. The benchtop geometrical analysis of endobag filling did not deploy the second Nellix stent, only filling of the endobag with colored water. Although the impact that this may have is suspected to be minimal, it may have altered the measured dimensions. The simulated use experiments were performed with water and silicone models, which have different properties than blood and tissue, respectively. Compared with blood, water is of lower viscosity and without clotting properties, which may reduce the appearances of gutters or endoleaks. The in vitro characterization experiments performed are small sample sizes. Given the consistent and repeatable nature of the closely controlled implant design, it is considered here to be adequate for basic characterization. In the current case series, 4 of 5 caudal migrations occurred in patients treated outside the IFU. Due to challenging anatomy, EVAS was the only endovascular solution in these patients with multiple comorbidities. In all cases, open surgery was considered undesirable due to the general condition of the patients. Anatomical characteristics, such as the quality of the infrarenal neck and the patent aneurysm flow volume, as well as procedural aspects could potentially have attributed to late caudal migration. However, in our case series, there is no clear relationship between the shape of the aortic neck and the occurrence of caudal migration, although the sample size is very small. In addition, progression of disease could have been involved. According to the current standard, the sealing length in the neck should be 10 mm, and the endobags should be positioned immediately below the lowest renal artery to achieve the longest seal length possible given the patient anatomy. This strategy is likely to reduce the risk of caudal migration.

Donselaar et al 217 Conclusion A proximal Nellix-in-Nellix extension can be used to treat caudal endograft migration after EVAS. The development of dedicated proximal extensions remains desirable to reduce the required sealing length and potential for procedure-related complications related to excessive endobag material or insufficient wall apposition. 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: Andrew Holden and Michel M. P. J. Reijnen are consultants for Endologix. Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Funding for the in vitro study was obtained from Endologix Inc. References 1. van den Ham LH, Zeebregts CJ, de Vries JP, et al. Abdominal aortic aneurysm repair using Nellix EndoVascular Aneurysm Sealing. Surg Technol Int. 2015;26:226 231. 2. Carpenter JP, Cuff R, Buckley C, et al; Nellix Investigators. Results of the Nellix system investigational device exemption pivotal trial for endovascular aneurysm sealing. J Vasc Surg. 2016;63:23 31.e1. 3. Thompson MM, Heyligers JM, Hayes PD, et al, for the EVAS FORWARD Global Registry Investigators. Endovascular aneurysm sealing: early and midterm results from the EVAS FORWARD Global Registry. J Endovasc Ther. 2016;23: 685 692. 4. Silingardi R, Coppi G, Ferrero E, et al. Midterm outcomes of the Nellix Endovascular Aneurysm Sealing System: a dualcenter experience. J Endovasc Ther. 2016;23:695 700. 5. England A, Torella F, Fisher RK, et al. Migration of the Nellix endoprosthesis. J Vasc Surg. 2016;64:306 312. 6. Zerwes S. Commentary: seduction and its impact on instructions for use. J Endovasc Ther. 2016;23:693 694.