Endovascular Treatment of Complex Aortic Aneurysms Using the Sandwich Technique

Transcription

1 J ENDOVASC THER 2012;19: CLINICAL INVESTIGATION Endovascular Treatment of Complex Aortic Aneurysms Using the Sandwich Technique Armando C. Lobato, MD, PhD, and Luciana Camacho-Lobato, MD, PhD São Paulo Vascular & Endovascular Institute (ICVE-SP), São Paulo, Brazil. Purpose: To assess the safety and efficacy of the sandwich technique for complex aortic aneurysms. Methods: Between October 2008 and March 2012, 78 patients (70 men; mean age 73.3 years) undergoing complex thoracic or abdominal endovascular aneurysm repairs were treated with the sandwich technique when the aneurysm features did not fulfill the requirements for standard endovascular techniques or total aortic arch replacement. Two thirds of the population was treated for aortoiliac aneurysm (52, 66.7%). There were 15 (19.2%) thoracoabdominal aneurysms, 6 (7.7%) isolated iliac artery aneurysms, 3 (3.8%) arch aneurysms, and 2 (2.6%) abdominal aortic aneurysm with non-diseased, short common iliac arteries bilaterally. Five (6.4%) symptomatic patients were submitted to endovascular repair in the urgent setting. Results: Technical success was 98.7%; one procedure was aborted when both renal arteries could not be cannulated. Over a mean 17-month follow-up (range 1 42), primary patency was high (96.7%) and mortality low (early: 5.1%, late: 1.3%). Overall, only 4 (5.1%) type II endoleaks persisted: 3 early with no sac increase and 1 late with sac increase that was managed conservatively on patient demand (stable at 9 months). In all other aneurysms except the 3 in the arch, there was a 5-mm reduction in size achieved by the end of the second year of follow-up, though this was significant only in the aortoiliac aneurysm group (p,0.005). Conclusion: The sandwich technique facilitates safe and effective aneurysm exclusion and target vessel revascularization in adverse anatomical scenarios, with sustained durability in midterm follow-up. J Endovasc Ther. 2012;19: Key words: endovascular aneurysm repair, thoracic endovascular aortic repair, stent-graft, aortoiliac aneurysm, arch aneurysm, thoracoabdominal aortic aneurysm, isolated iliac artery aneurysm, endoleak, mortality, target vessel, aneurysm diameter Aortic aneurysm disease is a world health problem of great magnitude and considerable economic burden, with rising incidence and prevalence rates. Endovascular aneurysm repair (EVAR) is currently recognized as the first-line treatment for abdominal aortic aneurysms (AAA) due to decreased operative mortality/morbidity and shorter patient recovery times. In the thoracic aorta, thoracic endovascular aortic repair (TEVAR) is also gaining acceptance as a valid treatment option for thoracic aortic aneurysm. Despite great technical and device improvement in the last 2 decades, complex aortic aneurysm disease, as represented by aortoiliac (AIA) and isolated iliac artery aneurysms The authors have no commercial, proprietary, or financial interest in any products or companies described in this article. Corresponding author: Armando C. Lobato, MD, PhD, R. Maestro Cardim 560/cjtos , São Paulo, Brazil. aclobato@icve.com.br Ó 2012 by the INTERNATIONAL SOCIETY OF ENDOVASCULAR SPECIALISTS Available at

2 692 SANDWICH TECHNIQUE FOR COMPLEX AORTIC ANEURYSM J ENDOVASC THER Lobato and Camacho-Lobato 2012;19: (IAA), as well as aortic arch (AcA) and thoracoabdominal aortic (TAAA) aneurysms, still presents technical challenges not completely overcome by either open surgery or currently available endovascular techniques. AIA and isolated IAA extending to the internal iliac artery (IIA) require a more demanding procedure owing to the difficulties in obtaining an adequate distal landing zone for the stent-graft limb(s), a potential site for endoleak. Another difficult scenario is AAA associated with shorter common iliac arteries (CIAs) because secure distal fixation of a bifurcated stent-graft requires a long and healthy segment of CIA. Extending the iliac limbs of the stent-graft past the IIA origin to provide a secure seal can be done in these cases, but bilateral IIA occlusion is problematic in 12% to 45% of cases. 1 Patients deemed unfit for surgical or standard endovascular management are subject to unfavorable outcomes according to their aneurysm type and size. The sandwich technique, first introduced in 2008, 2 is an endovascular approach developed to overcome current anatomical and device constraints, expanding the limits of endovascular aneurysm repair. We review the technical steps for 4 types of complex aortic aneurysms and analyze the midterm results of our single-center experience with this technique. Study Design METHODS A prospective study was conducted of all consecutive patients undergoing endovascular repair of complex aortic aneurysms with the sandwich technique at our institution between October 2008 and March Patients were eligible for the sandwich technique if the aortic aneurysm did not meet the requirements for standard EVAR or TEVAR or total aortic arch replacement. Inclusion criteria were mainly anatomical. A stepwise, fully illustrated description of the sandwich technique as applied in the treatment of these types of aneurysms is provided in the designated figures. AcA (Fig. 1): (1) ascending aorta proximal neck at least 30 mm long with a maximal diameter not to exceed 40 mm; (2) supraaortic trunks at least 30 mm long with a maximal diameter no larger than 12 mm. TAAA (Fig. 2): (1) descending thoracic aorta proximal neck at least 20 mm long with a maximal diameter of 40 mm; (2) internal diameter of the visceral arteries at least 5 mm; (3) proximal end of the superior mesenteric artery (SMA) with at least 20 mm free of side branches. AIA (Fig. 3): (1) no distal landing zone bilaterally (CIAs); (2) no unilateral distal landing zone (CIA) plus contralateral IIA with aneurysm or previous occlusion/severe stenosis; (3) AAA with short, healthy CIAs (no distal landing zone bilaterally). Isolated IAA (Fig. 4): (1) no distal landing zone bilaterally (CIAs); (2) unilateral isolated IAA with no unilateral distal landing zone (CIA) plus a contralateral IIA with aneurysm or previous occlusion/severe stenosis. Patient Population In the observation period, 78 consecutive patients (70 men; mean age 73.3 years, range 61 83) undergoing endovascular repair of complex aortic aneurysms at our institution were treated using the sandwich technique under a protocol approved by the Institutional Review Board. All patients gave written informed consent to the procedure. Patient demographics and comorbidities are reported in Table 1. Anatomical features of the aneurysms, as well as aspects of patient history that could influence endovascular procedure indications, strategies, and planning are presented in Table 2. Two thirds of the population was treated for AIA (52, 66.7%). There were 15 (19.2%) TAAAs, 6 (7.7%) isolated IAAs, 3 (3.8%) AcAs, and 2 (2.6%) AAAs with nondiseased, short CIAs bilaterally. Five (6.4%) symptomatic patients (3 AcAs and 2 TAAAs) were submitted to endovascular repair in the urgent setting. Five (6.4%) patients had previous aortic open surgery, 32 (41.0%) were unfit for open repair, and 44 (56.4%) were not candidates for branched devices. Follow-up assessment included routine office visits and imaging [computed tomographic angiography (CTA) or duplex ultrasound for

3 J ENDOVASC THER SANDWICH TECHNIQUE FOR COMPLEX AORTIC ANEURYSM ;19; Lobato and Camacho-Lobato Figure 1 Sandwich technique for an arch aneurysm. (A) Insert a 100-mm-long thoracic stent-graft through a femoral artery access. Place and deploy the proximal end just above the coronary artery ostia and the distal end 10 mm below the brachiocephalic trunk. (B) Introduce a second thoracic stent-graft (200-mm-long) through a femoral artery approach and place the proximal end in the ascending aorta just above the first thoracic stent-graft. The distal end is left in the proximal third of the thoracic aorta, at least 30 mm below the LSA. Position the 100-mm-long self-expanding covered stents, one after the other starting with the brachiocephalic trunk (right brachial artery access), followed by the left CCA (open left CCA access) and, finally, the LSA (using a left brachial access and the periscope technique). Leave the proximal end of the first 2 covered stents just above the coronary artery ostia and the distal end inside the brachiocephalic trunk and left CCA, respectively. The proximal end of the third covered stent is left inside the LSA and the distal end 10 mm below the distal end of the second thoracic stent-graft. (C) Deploy and accommodate the second thoracic stent-graft using a latex balloon. (D) Deploy the 3 covered stents. (E) The 1-month CTA in a representative patient. patients with non-dialysis-dependent chronic renal insufficiency] performed at 1, 3, 6, and 12 months and every 6 months thereafter. Endpoints and Statistical Analysis Primary endpoints included patient survival and target vessel patency, as well as aneurysm sac size evolution. Patients with AIA, AAA with short CIAs, and isolated IAAs were combined in one group (hereafter referred to as the composite aortoiliac group) for analysis. Statistical analysis comprised descriptive statistics for all collected variables (procedure and fluoroscopy times; contrast and blood transfusion volumes; hospital and intensive care lengths of stay; technical success; 30-day and late (procedure-related/unrelated) mortality; endoleak; secondary intervention; device migration; end-organ complications; and aneurysm sac size change), as well as intention-to-treat analysis of survival and patency using the Kaplan-Meier method. Patients lost to followup were contacted, and their last CTA results were used in the analysis. Non-retrievable

4 694 SANDWICH TECHNIQUE FOR COMPLEX AORTIC ANEURYSM J ENDOVASC THER Lobato and Camacho-Lobato 2012;19: Figure 2 Sandwich technique for a TAAA. (A) Insert the thoracic stent-graft through a femoral approach and deploy it across the full thoracic extent of the aneurysm, leaving its distal end 1 cm above the celiac axis. (B) Cannulate the visceral arteries preferentially through subclavian and axillary artery accesses using a long sheath and an extra-stiff guidewire. (C) Position the self-expanding covered stents in the visceral arteries. Place and deploy a bifurcated stent-graft just below the renal arteries, followed by a second thoracic stentgraft. (D) The thoracic stent-graft is deployed 10 mm below the proximal end of the covered stents, followed by molding with a latex balloon. (E) Deploy the distal end of the covered stents 20 mm inside the visceral arteries and the proximal end no more than 10 mm above the second thoracic stent-graft. (F) The 6-month CTA in a representative patient.

5 J ENDOVASC THER SANDWICH TECHNIQUE FOR COMPLEX AORTIC ANEURYSM ;19; Lobato and Camacho-Lobato Figure 3 Sandwich technique for an aortoiliac aneurysm extending to the internal iliac artery (IIA). (A) Insert the main body of any commercially available bifurcated stent-graft through a femoral approach, leaving the distal end of the ipsilateral iliac limb 10 mm above the IIA origin. (B) Cannulate the ipsilateral IIA preferentially through a left brachial access using a long sheath and an extra-stiff guidewire. Place the distal end of a self-expanding covered stent inside the IIA and position an iliac limb extension 10 mm below the proximal end of the covered stent. (C) Deploy the iliac limb extension and model it using a latex balloon. (D) Deploy the covered stent. (E) For bilateral CIA aneurysms extending to both IIAs, deploy the contralateral iliac limb and repeat steps in B to D. (F) The 12-month CTA in a representative patient. data were approached by the last-observation-carried-forward method. Student t test of the means (repeated measures) was employed to evaluate aneurysm sac size evolution over time. Continuous data are presented as the means 6 standard deviation; categorical data are given as the counts (percentage). The analyses were performed using Prism statistical software (version 6; GraphPad Software Inc., San Diego, CA, USA). RESULTS Technical success was 98.7%; the procedure was aborted in 1 patient when both renal arteries could not be cannulated by guidewire and guide catheter. Of the 127 intended target vessels, 5 (3.9%) could not be cannulated: the 2 renal arteries in the aborted procedure and 1 additional left and 2 right renal arteries. In the last 3 cases, the sandwich procedure was carried out normally, and these vessels were occluded by the endograft. The type and

6 696 SANDWICH TECHNIQUE FOR COMPLEX AORTIC ANEURYSM J ENDOVASC THER Lobato and Camacho-Lobato 2012;19: Figure 4 Sandwich technique for an isolated iliac artery aneurysm extending to the internal iliac artery (IIA). (A) Insert the iliac limb of the stent-graft through a femoral approach with its distal end placed 10 mm above the IIA origin. (B) Cannulate the ipsilateral IIA preferentially through a left brachial access using a long sheath and an extra-stiff guidewire. Place the distal end of a self-expanding covered stent inside the IIA and position a second iliac limb extension or a second covered stent 10 mm below the proximal end of the first covered stent. (C) Deploy the second iliac limb extension or the second covered stent and model it using a latex balloon. (D) Finally, deploy the first covered stent. (E) The 12-month CTA in a representative patient. number of stent-grafts, as well as the number of self-expanding covered stents used to treat target vessels, are listed in Table 3. Performance data for the sandwich procedure for the entire study group featured a mean operating time of minutes, a mean fluoroscopy time of minutes, a mean contrast volume of ml, and a mean blood transfusion volume of ml. The mean intensive care and hospital lengths of stay were and days, respectively. Complications No paraplegia or paraparesis was detected (Table 4). There were 3 (3.8%) intraoperative acute myocardial infarctions (1 death, 1 coronary stenting, and 1 conservative management); 1 (1.3%) fatal pulmonary infection; 2 (2.6%) fatal cases of colonic ischemia; 1 (1.3%) case of erectile dysfunction; 2 (2.6%) patients with temporary buttock claudication; and 8 (10.2%) patients with worsening renal function (6 had existing chronic renal insuffi-

7 J ENDOVASC THER SANDWICH TECHNIQUE FOR COMPLEX AORTIC ANEURYSM ;19; Lobato and Camacho-Lobato TABLE 1 Patient Demographics and Comorbidities Men 70 (89.7%) Age, y Hypertension 45 (57.7%) Smoking 48 (61.5%) COPD 19 (24.3%) Coronary artery disease 28 (35.9%) Coronary stent 13 (16.7%) Coronary artery bypass 6 (7.7%) Congestive heart failure 8 (10.0%) PAOD 26 (33.3%) Hypercholesterolemia 15 (19.2%) Diabetes 23 (29.5%) Chronic renal insufficiency (no dialysis) 12 (15.4%) Obesity 7 (9.0%) TIA/stroke 7 (9.0%) Hostile abdomen 4 (5.1%) Oral anticoagulation 7 (9.0%) Continuous data are presented as the means 6 standard deviation; categorical data are given as the counts (percentage). COPD: chronic obstructive pulmonary disease, PAOD: peripheral artery occlusive disease. ciency). Renal function returned to the preoperative status at the 1 month in 7 patients. Graft-related complications included 3 limb graft thromboses, 2 axillary access occlusions, 2 early brachial artery occlusions (treated medically), an inguinal hematoma, and 2 device migrations causing late type III endoleaks owing to disconnection of 2 thoracic stent-grafts in an AcA and a disconnected AIA limb. Endoleak Twenty-seven (34.6%) patients had endoleaks detected intraoperatively on the completion angiogram (Table 4): 4 type I, 13 type II, 3 type III, and 7 type IV. Type I endoleaks were successfully managed endovascularly during the main procedure, either with the deployment of a proximal thoracic endograft (TAAA and AcA groups) or an abdominal aortic cuff (AIA group). All early type II endoleaks were managed conservatively except one with backflow from the SMA. In the TABLE 2 Preoperative Aneurysm Anatomical Features AcA (n¼3) TAAA (n¼15) AIA (n¼52) AAA w/scia (n¼2) Isolated IAA (n¼6) Symptomatic Aneurysm diameter, mm Aneurysm type 2 fusiform, 1 saccular 15 fusiform (I: 4, II: 2, III: 3, IV: 6 52 fusiform 2 fusiform 5 fusiform, 1 anastomotic CIA,30 mm long 4* Bi-/unilateral aneurysm 1/2 40/12 3/3 Acute aneurysm 2 expanding 2 contained ruptures AEF 1 Previous open aortic surgery 2 JRAIA, 1 AAD 1 AAA 1 AAD Concomitant TAAA 1 1 Urgent setting 3 2 Unfit for open repair Unfit for branched devices Continuous data are presented as the means 6 standard deviation; categorical data are given as the counts (percentage). AcA: aortic arch aneurysm, TAAA: thoracoabdominal aortic aneurysm, AIA: aortoiliac aneurysm, scia: short common iliac artery, AAA: abdominal aortic aneurysm, IAA: iliac artery aneurysm, JRAIA: juxtarenal aortoiliac aneurysm, AAD: acute type A aortic dissection. * Mean mm.

8 698 SANDWICH TECHNIQUE FOR COMPLEX AORTIC ANEURYSM J ENDOVASC THER Lobato and Camacho-Lobato 2012;19: TABLE 3 Aortic Stent-Graft Types and Self-Expanding Covered Stents in the Target Vessels Stent-Graft Type (n¼118) Patients (n¼78)* Covered Stents in the Target Vessels Arch aneurysm IA LCCA LSA TAG (n¼5) Braile (n¼3) TAAA* CA SMA RRA LRA TAG (n¼29) Excluder (n¼9) Valiant (n¼7) 3 2 # 3 2 ** 1 Bifurcated Talent (n¼1) Endurant (n¼1) Relay (n¼2) Powerlink (n¼1) Aortoiliac aneurysm RHA LHA Excluder (n¼31) Talent/Endurant (n¼17) Abdominal E-Vita (n¼1) Braile (n¼1) Powerlink (n¼1) Zenith (n¼1) Isolated iliac artery aneurysm RHA LHA Excluder (n¼3) Talent/Endurant (n¼1) Braile (n¼1) Viabahn (n¼1) Aortic aneurysm with short CIA RHA LHA Excluder (n¼2) IA: innominate artery, LCCA: left common carotid artery, LSA: left subclavian artery, TAAA: thoracoabdominal aortic aneurysm, CA: celiac axis, SMA: superior mesenteric artery, RRA: right renal artery, LRA: left renal artery, RHA: right hypogastric artery, LHA: left hypogastric artery, CIA: common iliac artery. * One patient was excluded as the procedure was aborted after failure to cannulate both renal arteries; the CA and SMA were cannulated without difficulty. There were 123 successfully treated and 5 failures to cannulate in all, including 2 renal arteries in the aborted case. One spontaneous previous celiac axis occlusion. Two right renal arteries could not be cannulated by the guidewire. One left renal artery could not be cannulated with the guidewire. # One intentional celiac trunk occlusion. ** One spontaneous previous right renal artery occlusion. Two previous spontaneous left renal artery occlusions. latter, a self-expanding stent-graft was successfully deployed in the SMA at the 3-month follow-up. Eight of the 13 type II endoleaks sealed spontaneously, and 1 patient died of colonic ischemia with a type II endoleak (TAAA emergency group). No aneurysm sac increase (.5 mm in diameter) was detected in the 3 persistent early type II endoleaks so far. The 3 low-flow early type III endoleaks were managed conservatively and sealed spontaneously by the time of the first CTA assessment at 1 month. All type IV endoleaks also sealed spontaneously by the time of the first CTA. Over a mean follow-up of 17 months (range 1 42), 3 (3.8%) late endoleaks were encountered. Two type III leaks (1 AcA, 1 AIA) owing to device migration noted previously were

9 J ENDOVASC THER SANDWICH TECHNIQUE FOR COMPLEX AORTIC ANEURYSM ;19; Lobato and Camacho-Lobato TABLE 4 Outcomes and Complications Composite Aortoiliac* (n¼60) Elective TAAA (n¼13) Emergent TAAA (n¼2) AcA (n¼3) Follow-up Mean, mo 18.5 (1 42) 16.2 (1 36) (1 10) Patients lost 2 (3%) 1 (8%) Excluded (aborted procedure) 1 (8%) Technical success (per patient) 60 (100%) 12 (92%) 2 (100%) 3 (100%) Technical success (per vessel) 68/68 (100%) 37/41 (90%) 6/7 (86%) 7/7 (100%) Mortality At 30 days (procedure-related) 0 1 (8%) 2 (100%) 1 (34%) Late procedure-related (34%) Late non-procedure-related 2 (3%) 2 (15%) 0 1 (34%) Intraoperative endoleak type I 2 (3%) 1 (8%) 1 (34%) II 10 (17%) 2 (15%) 1 (50%) III 2 (3%) 1 (8%) IV 4 (7%) 2 (15%) 1 (34%) Late endoleak type II 1 (2%) III 1 (2%) 1 (34%) Persistent type II endoleak 3 (5%) 1 (8%) Target vessel primary patency 65/68 (96%) 36/37 (97%) 5/6 (83%) 7/7 (100%) Secondary interventions for Endoleaks 1 (2%) 1 (8%) 1 (34%) Limb graft thrombosis 3 (5%) Colonic ischemia 2 (100%) Axillary approach occlusion 2 (15%) Incisional hematoma 1 (2%) Device migration 1 (2%) (34%) End-organ complications Cardiac 2 (3%) 1 (34%) Renal 5 (8%) 3 (23%) Pulmonary 1 (34%) Colonic ischemia 2 (100%) Paraplegia/paraparesia/limb weakness Continuous data are presented as the means 6 standard deviation; categorical data are given as the percentage (counts) TAAA: thoracoabdominal aortic aneurysm, AcA: aortic arch aneurysm. *Includes aortoiliac aneurysm, isolated iliac artery aneurysm, and abdominal aortic aneurysms with short CIAs. diagnosed at 3 and 6 months, respectively; they were effectively treated with a thoracic stent-graft deployed inside the aortic arch and a stent-graft inside the right iliac limb endograft, respectively, within the first month of diagnosis. One late type II endoleak (AIA group) associated with a 5-mm increase in the aneurysm sac was diagnosed at the 12-month assessment. This patient refused a secondary procedure and is currently being managed conservatively at 3-month intervals; no further aneurysm sac increase was observed in the last 9 months of follow-up. No late type I and IV endoleaks were detected during follow-up. Mortality The 30-day mortality was 5.1% (4/78). One patient died of hepatic failure after intentional

10 700 SANDWICH TECHNIQUE FOR COMPLEX AORTIC ANEURYSM J ENDOVASC THER Lobato and Camacho-Lobato 2012;19: Figure 5 Kaplan-Meier survival curves based on intention-to-treat. (A) Overall cohort based on deaths related to the sandwich technique (93%) and all-cause deaths (83%). (B) Survival in the composite aortoiliac group (100% related and 94% all-cause). (C) Survival in the TAAA group of (92% related and 70% all-cause). (D) Overall target vessel patency (96%), reflecting 123 successfully revascularized target vessels using the sandwich technique; the 4 target vessel occlusions occurred in the first month of the study. celiac trunk occlusion by the endograft despite performing a recognized viability test. One arch aneurysm patient suffered a fatal myocardial infarction. Two emergency patients presenting rapid aneurysm expansion and contained TAAA rupture died of colonic infarction. All 4 deaths occurred before hospital discharge. Further investigation of factors responsible for the intestinal infarctions demonstrated that the SMA was occluded in only one of the patients. Throughout follow-up, 6 (7.7%) patients died. Only one (1.3%) was considered procedure-related: an arch aneurysm patient with aortoesophageal fistula died of massive pulmonary infection at 6 months. The other 5 deaths, starting at 10 months of follow-up, were considered unrelated to the procedure. Their causes included a TAAA open repair in another center (AcA group), pancreatic cancer (AIA group), a hemorrhagic stroke (TAAA group), a lymphoma (TAAA group), and a myocardial infarction (isolated IAA group). The overall survival rate (Table 4, Fig. 5A) was 93.1%; the rates were 100% for the composite aortoiliac group (Fig. 5B) and 92.3% for the elective TAAA (Fig. 5C) group. Target Vessel Patency Primary patency for the 123 successfully treated target vessels was 96.4% (Fig. 5D). All 4 side branches occlusions (an SMA in the TAAA group and 3 IIAs in the AIA group) occurred in 3 patients within the first month of the study; no late occlusions were detected. Two IIA occlusions resulted from placing the covered stents too distally into the gluteal artery (,3.5 mm); the third was attributed to dissection of the distal part of the IIA by an extra-stiff guidewire. Symptoms were observed in 2 of the 3 patients, who developed buttock claudication; one of them had also erectile dysfunction, which persisted despite resolution of the buttock claudication after 6 and 10 months, respectively. The SMA occlu-

11 J ENDOVASC THER SANDWICH TECHNIQUE FOR COMPLEX AORTIC ANEURYSM ;19; Lobato and Camacho-Lobato sion also resulted from dissection of the distal part of the vessel by an extra-stiff guidewire (Table 4). Secondary Interventions Eleven (14.1%) secondary interventions were undertaken (Table 4): 3 (27.3%) for endoleaks as noted above, 3 (27.3%) for limb graft thrombosis, 2 (18.2%) for colonic ischemia (laparotomy and resection), 2 (18.2%) for axillary access occlusions, and 1 (9%) surgical evacuation for an incisional inguinal hematoma. Among the graft limb complications, one ipsilateral iliac limb stent-graft with patent revascularized IIA thrombosed 12 hours after the main procedure. Patency was restored with thromboembolectomy, followed by placement of a self-expanding bare stent inside the iliac limb extension. A contralateral iliac limb stent-graft occlusion occurred on the 29 th postoperative day in a patient who had a previous unilateral CIA aneurysm and isolated IAA open repair 7 years before. A left external iliac artery dissection associated with distal clamping occurred during that procedure and was considered the cause of this event. Patency was restored through thromboembolectomy and profundaplasty, followed by self-expanding bare stent deployment to treat the previous left external iliac artery dissection. One ipsilateral external iliac artery thrombosis with patent revascularized IIA occurred 5 months after the sandwich procedure. This patient underwent a femorofemoral crossover bypass with success. Fate of the Aneurysm Sac A progressive reduction in the aneurysm (AIA, isolated IAA, and TAAA) sac size (Fig. 6) was observed over the follow-up period (Table 5), reaching the mean target 5-mm reduction at the end of the second year of follow-up. Individual 5-mm reduction, though, started earlier: at 12 months for the aortic segment and at 6 months for the iliac segment of the composite aortoiliac group and at 18 months for the TAAA group. Statistical significance was reached only for the aortic segment of the AIA þ AAA with short CIAs group Figure 6 (A) Aneurysm sac size evolution over time. Consistent progressive sac size reduction was observed for all aneurysms, reaching the target 5- mm reduction at the second year of follow-up. Statistical significance was attained for AIAs þ AAAs with short CIAs at 30 months (p¼0.044), 36 months (p¼0.001), and 42 months (p¼0.004). at 30 months (p¼0.044), 36 months (p¼0.001), and 42 months (p¼0.004). DISCUSSION The sandwich technique evolved to circumvent the anatomical constraints of commer-

12 702 SANDWICH TECHNIQUE FOR COMPLEX AORTIC ANEURYSM J ENDOVASC THER Lobato and Camacho-Lobato 2012;19: TABLE 5 Aneurysm Sac Size Change Reduction in Aneurysm Size 5 mm Follow-up, mo AIA þ AAA with scia (aortic segment) AIA þ Isolated IAA (iliac segment) TAAA 6 0/45 (0%) 3/49 (6%) 0/11 (0%) 12 1/34 (3%) 4/36 (11%) 0/9 (0%) 18 7/26 (27%) 4/26 (15%) 2/4 (50%) 24 10/23 (43%) 10/23 (43%) 2/4 (50%) 30 14/16 (87%) 12/17 (70%) 3/3 (100%) 36 7/8 (87%) 6/9 (67%) 1/1 (100%) 42 4/4 (100%) 3/4 (75%) AIA: aortoiliac aneurysm, AAA: abdominal aortic aneurysm, scia: short common iliac artery, TAAA: thoracoabdominal aortic aneurysm, IAA: iliac artery aneurysm. cially available stent-grafts for the treatment of complex aortic aneurysms, and the midterm results are encouraging. Overall high technical success and primary patency rates, combined with progressive sac size reduction and low mortality, constitute clear-cut evidence that the sandwich technique provides safe and effective aneurysm exclusion, as well as target vessel revascularization, with sustained midterm durability in adverse anatomical scenarios not routinely addressed by other techniques. 3 7 To our understanding, this derives from the use of commercially available stent-grafts familiar to the majority of endovascular surgeons, obviating the need to wait for custom-made branched devices. The outcomes reflect the technique s versatility in addressing challenging anatomical scenarios either in the emergency or elective setting (Table 6). In addition, there is no need for precise location of target vessel origins with complex image-analyzing software because the technique accommodates target vessels despite their location. As regards branched devices, fewer than half of the study patients (43.6%: 19 AIAs, 4 isolated IAAs, and 11 TAAAs) would have been eligible for TEVAR or EVAR using a branched device. This finding is consistent with existing evidence demonstrating the low anatomical applicability (38%, 8 52% 9 ) of the commercially available iliac branched devices (IBD) for IIA preservation and/or endovascular management of aortoiliac and solitary iliac aneurysms. In the present series, patients had at least one of the following morphological TABLE 6 Challenging Anatomical Scenarios Appropriate for the Sandwich Technique Arch aneurysm involving all supra-aortic trunks (no need for hybrid procedure) Chronic aortic dissection with true lumen 8 mm AIA or isolated IAA with CIA,40 mm long CIA and/or CIA aneurysm lumen 8 mm in diameter AIA or isolated IAA with very tortuous CIA anatomy Distal landing zone,10 mm long in the main hypogastric artery trunk Isolated IAA with a proximal neck 18 mm in diameter and 15 mm long Contralateral external iliac artery occlusion Long and large IIA aneurysm CIA dissection complicated by aneurysm and true lumen 8 mm in diameter Previous open AAA repair with Dacron graft complicated by CIA anastomotic false aneurysm AIA: aortoiliac aneurysm, IAA: iliac artery aneurysm, CIA: common iliac artery, AAA: abdominal aortic aneurysm.

13 J ENDOVASC THER SANDWICH TECHNIQUE FOR COMPLEX AORTIC ANEURYSM ;19; Lobato and Camacho-Lobato factors contraindicating IBD use according to the manufacturer s Instruction for Use, 10,11 as well as in other IBD series 9,12 15 : (1) patent CIA lumen,18 mm (free of intraluminal thrombus) 9,12 in 15 (25%) patients; (2) aneurysmal IIA 9 in 13 (21.7%) patients; (3) IIA trunk length,10 mm 12,14,15 in 10 (16.7%) patients; (4) CIA length,40 mm 12,14,15 in 8 (13.3%) patients; (5) severe CIA tortuosity 9 in 7 (11.7%) patients; and (6) aortic bifurcation angle, ,13 in 5 (8.3%) patients. Additionally, no commercially available or custom-made branched devices can address arch aneurysm (n¼3), emergency TAAA (n¼2), or TAAA with previous open repair and temporary spinal cord ischemia (n¼2). Although less than half of our patients would have qualified for an IBD, it may be useful to compare the IBD results to those of the sandwich technique. Ziegler et al. 13 reported a 58% technical success rate for the first-generation device and 85% for the second-generation device when attempting to treat unilateral or bilateral CIA aneurysms combined or not with EVAR. Greenberg et al. 16 had an 86% technical success rate for CIA aneurysms and 89% for TAAAs. Other investigators have reported technical success rates varying from 91% to 100% and primary patency rates of 74% to 100% over a 6- to 20-month follow-up for aortoiliac or isolated iliac aneurysms. 14,15,17 24 We currently report 100% technical success in both arch and abdominal aneurysms and 93% for TAAAs. Adverse outcomes, such as renal function impairment, pelvic ischemia, endoleaks, early deaths, and/or target vessel occlusions, were observed mainly in the first months of the study, suggesting the effect of a learning curve 25 or perhaps the consequence of advanced age and high comorbidities among the patients. The learning curve effect is especially evident in the TAAA group because all deaths were clustered in the first half of the study and were due to acute severe visceral ischemia caused either by intentional celiac axis occlusion, SMA occlusion, or coverage of SMA small branches. No graft-related deaths were observed in the second half of the study. The high mortality rate observed in the arch group, on the other hand, seems to reflect the patients adverse clinical conditions. The absence of major causes of intraoperative death, such as cardiovascular and/or pulmonary complications as well as stroke in this series, may indicate that repair with the sandwich technique is well tolerated and may impose minimum operative burden to the patients. The involvement of the visceral arteries in the aneurysm has always been of great concern during TAAA repair and has probably hindered the use of the endovascular approach in these cases. Our few instances of visceral ischemia reinforce our recommendation to revascularize all branches whenever possible to avoid end-organ ischemia. Buttock claudication and erectile dysfunction have been reported in up to 12% of patients submitted to IBD, but these were rare (2.6% and 1.3%, respectively) for the sandwich technique. The two cases of buttock claudication in our series resolved. Endoleaks, on the other hand, were fairly common at completion angiography; however, no type I endoleak was detected thereafter. Overall, only a handful of type II endoleaks persisted, and of those, only one was associated with late sac increase. Spinal cord ischemia (SCI) is a devastating complication that has great negative impact on health-related quality of life, lifestyle, and late survival rates. Overall, SCI (paraplegia, paraparesis, or transient lower extremity weakness) has been reported in 4% to 32% of TAAA patients undergoing surgical repair 26 and in up to 18% in TEVAR with branched devices. 27 Reilly and Chuter 26 recently proposed creating a temporary type Ib endoleak to acutely reverse neurological complications after TEVAR with branched devices by providing controlled aneurysm sac perfusion with consequent restoration of spinal cord perfusion. This idea was also adopted by Harrison et al. 27 with success in 7 of 10 cases. No SCI was observed in our patients treated with the sandwich technique, which might be owing to the use of protective measures (cerebrospinal fluid drainage) or suggest that disruption of intercostals was minimized. Indeed, the sandwich technique might provide some sort of protection against neurological adverse events by maintaining a degree of spinal cord perfusion from the

14 704 SANDWICH TECHNIQUE FOR COMPLEX AORTIC ANEURYSM J ENDOVASC THER Lobato and Camacho-Lobato 2012;19: Figure 7 (A) CTA of a type I TAAA endovascular repair, demonstrating there is no need for the bifurcated stent-graft. (B) CTA of an isolated iliac artery aneurysm with no adequate proximal neck. (C) Use any commercially available bifurcated stent-graft instead of the iliac extension and proceed with the sandwich technique steps described for an aortoiliac aneurysm. unintentional low-flow type III endoleaks emanating from minimal gaps in between the two thoracic stent-grafts and the covered stents, although these endoleaks sealed within a month. Aneurysm sac shrinkage was somewhat slower but consistent and progressive with the sandwich technique compared to IBD. Only one sac expanded and no secondary ruptures were encountered during the up to 42-month follow-up. These factors, in association with the absence of persistent types I and III endoleaks, support the efficacy and durability of this repair to completely exclude the aneurysm sac in complex aortic aneurysms. On a technical note, we have amassed some tips to attaining better results over our 42-month-long experience with the sandwich technique. For all groups, perform targets vessel cannulation using a inch stiff hydrophilic guidewire and 5-F catheters at least 110 cm long with VERT, MP, SIM, or IM tip curves. After cannulation, exchange the stiff hydrophilic guidewire for an extra-stiff inch floppy tip wire no longer than 30 mm. For the arch aneurysm group: (1) do not place more than one stent-graft and two covered stents in the aortic arch to avoid type III endoleak; instead, the third vessel should be revascularized using the periscope technique (Fig. 1). (2) Provide at least an 80-mm overlap between the first and second covered stent and the second thoracic stent-graft at the ascending thoracic aorta level to avoid type Ia endoleak. (3) Provide at least a 30-mm overlap between the third covered stent and the second thoracic stent-graft at the descending thoracic aorta level to avoid type Ib endoleak. For the TAAA group: (1) the covered stents should be positioned not to exceed 20 mm inside the visceral arteries (distal end) and with at least a 50-mm overlap (proximal end) with the first thoracic stent-graft to avoid endoleak type III. Special attention should be given so as not to cover SMA small branches to avoid ischemic insults. (2) Place a bare selfexpanding stent inside the covered stent to avoid kinking and future occlusion. (3) For type I TAAA, there is no need for the bifurcated stent-graft (Fig. 7A). For the AIA group: (1) covered stents should be positioned 20 mm inside the IIA (distal end) to avoid type II endoleak and with a 50-mm overlap within the ipsilateral iliac limb of the bifurcated stent-graft (the proximal end) to avoid type III endoleak. (2) The IIA must have a minimal lumen of 3.5 mm to avoid occlusion. (3) The IIA could also be catheterized through

15 J ENDOVASC THER SANDWICH TECHNIQUE FOR COMPLEX AORTIC ANEURYSM ;19; Lobato and Camacho-Lobato a contralateral femoral approach when using unibody bifurcated stent-graft. For the isolated IAA group: if the CIA aneurysm is,40 mm in length and/or has an inadequate proximal neck (Fig. 7B,C), opt for any commercially available bifurcated stent-graft instead of the iliac extension or covered stent and follow the technical steps described for the aortoiliac aneurysms. Conclusion The sandwich technique facilitates safe and effective aneurysm exclusion and target vessel revascularization in adverse anatomical scenarios, with sustained durability in midterm follow-up. Thus, the sandwich technique appears to be a good addition to the endovascular armamentarium for treatment of complex aortic aneurysms. More experience with the method is advisable, especially for arch aneurysms and TAAAs. REFERENCES 1. Razavi MK, DeGroot M, Olcott C, et al. Internal iliac artery embolization in stent-graft treatment of aortoiliac aneurysms: analysis of outcomes and complications. J Vasc Interv Radiol. 2000; 11: Lobato AC. Sandwich technique for aortoiliac aneurysms extending to the internal iliac artery or isolated common/internal iliac artery aneurysms: a new endovascular approach to preserve pelvic circulation. J Endovasc Ther. 2011; 18: Reilly LM, Chuter TA. Endovascular repair of thoracoabdominal aneurysms: design options, device construct, patient selection and complications. J Cardiovasc Surg (Torino). 2009;50: Donas KP, Torsello G, Pitoulias GA, et al. Surgical versus endovascular repair by iliac branch device of aneurysms involving the iliac bifurcation. J Vasc Surg. 2011;53: Pua U, Tan K, Rubin BB, et al. Iliac branch graft in the treatment of complex aortoiliac aneurysms: early results from a North American institution. J Vasc Interv Radiol. 2011;22: Mastracci TM, Eagleton MJ. Endovascular repair of type II and type III thoracoabdominal aneurysms. Perspect Vasc Surg Endovasc Ther. 2011;23: Lioupis C, Abraham CZ. Results and challenges for the endovascular repair of aortic arch aneurysms. Perspect Vasc Surg Endovasc Ther. 2011;23: Karthikesalingam A, Hinchliffe RJ, Malkawi AH, et al. Morphological suitability of patients with aortoiliac aneurysms for endovascular preservation of the internal iliac artery using commercially available iliac branch graft devices. J Endovasc Ther. 2010;17: Tielliu IF, Bos WT, Zeebregts CJ, et al. The role of branched endografts in preserving internal iliac arteries. J Cardiovasc Surg (Torino). 2009; 50: Cook Medical. Zenith Branch Endovascular Graft-Iliac Bifurcation with the H & L-B One- Shot Introduction System IFU-IBE/4. Queensland, Australia: William A Cook Australia Pty. Ltd., Cook Medical. Zenith Helical Iliac Sidebranch and Zenith Crimped Leg Extension Endovascular Grafts with the H&L-B One-Shot Introduction System Instructions for Use. IFU-HISD/6. Queensland, Australia: William A Cook Australia Pty. Ltd., Verzini F, Parlani G, Romano L, et al. Endovascular treatment of iliac aneurysm: concurrent comparison of side branch endograft versus hypogastric exclusion. J Vasc Surg. 2009;49: Ziegler P, Avgerinos ED, Umscheid T, et al. Branched iliac bifurcation: 6 years experience with endovascular preservation of internal iliac artery flow. J Vasc Surg. 2007;46: Serracino-Inglott F, Bray AE, Myers P, et al. Endovascular abdominal aortic aneurysm repair in patients with common iliac artery aneurysms - initial experience with the Zenith bifurcated iliac side branch device. J Vasc Surg. 2007;46: Dias NV, Resch TA, Sonesson B, et al. EVAR of aortoiliac aneurysms with branched stentgrafts. Eur J Vasc Endovasc Surg. 2008;35: Greenberg RK, West K, Pfaff K, et al. Beyond the aortic bifurcation: branched endovascular grafts for thoracoabdominal and aortoiliac aneurysms. J Vasc Surg. 2006;43: Malina M, Dirven M, Sonesson B, et al. Feasibility of a branched stent-graft in common iliac artery aneurysms. J Endovasc Ther. 2006; 13: Naik J, Hayes PD, Sadat U, et al. Internal iliac artery branch graft for common iliac artery aneurysm following previous open abdominal

16 706 SANDWICH TECHNIQUE FOR COMPLEX AORTIC ANEURYSM J ENDOVASC THER Lobato and Camacho-Lobato 2012;19: aortic aneurysm repair. Eur J Vasc Endovasc Surg. 2008;35: Tielliu IF, Zeebregts CJ, van den Dungen JJ, et al. A modified technique for iliac artery branched endografting using a tromboned sheath. J Vasc Surg. 2008;48: Oderich GS, Ricotta JJ. Novel surgeon-modified hypogastric branch stent-graft to preserve pelvic perfusion. Ann Vasc Surg. 2010;24: Karthikesalingam A, Hinchliffe RJ, Holt PJ, et al. Endovascular aneurysm repair with preservation of the internal iliac artery using the iliac branch graft device. Eur J Vasc Endovasc Surg. 2010;39: Lee WA. Branched endograft for aortoiliac artery aneurysms. Vascular. 2009;17 Suppl 3: S Huilgol RL, Denton MJ, Cohen T. The iliac bifurcation device for endovascular iliac aneurysm repair: indications, deployment options and results at 1-year follow-up of 25 cases. ANZ J Surg. 2009;79: Haulon S, Greenberg RK, Pfaff K, et al. Branched grafting for aortoiliac aneurysms. Eur J Vasc Endovasc Surg. 2007;33: Lobato AC, Rodriguez-Lopez J, Diethrich EB. Learning curve for endovascular abdominal aortic aneurysm repair: evaluation of a 277- patient single-center experience. J Endovasc Ther. 2002;9: Reilly LM, Chuter TA. Reversal of fortune: induced endoleak to resolve neurological deficit after endovascular repair of thoracoabdominal aortic aneurysm. J Endovasc Ther. 2010; 17: Harrison SC, Agu O, Harris PL, et al. Elective sac perfusion to reduce the risk of neurologic events following endovascular repair of thoracoabdominal aneurysms. J Vasc Surg. 2012;55: