Abstract
Keywords
Introduction
The therapy of complex abdominal aortic aneurysm (AAA) and thoracoabdominal aortic aneurysm (TAAA) is challenging.1–5 Despite improvements in perioperative management over the past decades that have led to reduced morbidity and mortality,6,7 open repair of these complex aneurysms is still associated with relevant death and complication rates even in low- to medium-risk patients. These rates significantly increase in emergent settings 4 ; as a consequence, patients with a high perioperative risk are often denied open aortic repair.
Endovascular aneurysm repair (EVAR) has become an effective treatment option for AAA. In complex aneurysms, demanding endovascular techniques such as fenestrated and branched endografts have demonstrated their benefit. 8 However, these devices are custom-made to fit individual patient anatomy, and planning, production, and delivery may take up to 12 weeks. In symptomatic or ruptured complex aneurysms in high-risk patients, customized endografts are not available. Currently, there is only 1 off-the-shelf thoracoabdominal endograft, the t-Branch (Cook Medical, Bloomington, IN, USA), which is applicable to 60% of TAAA patients. For patients with symptomatic aneurysms or contained rupture who are in a hemodynamically stable condition, “surgeon-modified fenestrated and branched stent-grafts” (sm-FBSGs), also called physician-modified endografts (PMEG), can be used by modifying a commercially available aortic stent-graft to create the necessary fenestrations and branches for the renovisceral vessels directly prior to surgery. The purpose of this report is to review our experience with sm-FBSGs in the treatment of complex abdominal and thoracoabdominal pathologies.
Methods
Study Design and Patient Selection
A retrospective analysis was performed of 21 consecutive patients (mean age 70 years, range 58–78; 16 men) treated in emergency with a sm-FBSG between April 2014 and September 2016 (sm-FBSGs have not been used in elective cases). The indications for surgical treatment were complex abdominal and thoracoabdominal aortic pathologies involving the renal, superior mesenteric (SMA), or celiac arteries. Indications for urgent treatment were symptomatic lesions (n=8) or contained rupture (n=13). Patients were classified as symptomatic if there was tenderness on palpation or new onset of abdominal or back pain without other explanation on clinical examination or computed tomography (CT). All patients were hemodynamically stable so that modification of the endograft could be performed. Fourteen patients were classified as American Society of Anesthesiologists (ASA) risk class IV, 6 patients as ASA III, and 1 patient as ASA II. The mean Society for Vascular Surgery cardiac, respiratory, renal, hypertension, and age scores 11 were 5, 1, 1, 2, and 2, with a mean total score of 10 (range 4–20). The mean maximal aneurysm size was 7.4±2.3 cm (range 4.0–11.4 cm). Further patient characteristics are summarized in Table 1.
Baseline Characteristics of the 21-Patient Cohort. a
Abbreviations: CABG, coronary artery bypass grafting; COPD, chronic obstructive pulmonary disease; MI, myocardial infarction; PAU, penetrating aortic ulcer; PTCA, percutaneous transluminal coronary angioplasty.
Continuous data are presented as the means (range); categorical data are given as the counts.
Creatinine >1.5 mg/dL.
The decision between open and endovascular repair was made based on patient comorbidities. Only patients with high perioperative risk were candidates for sm-FBSG repair. Comorbidities included coronary artery disease and reduced cardiac ejection fraction, as well as pulmonary dysfunction and hostile abdomen due to prior abdominal or aortic surgery, among others.
All patients were informed that the use of the stent-graft was not according to the instructions for use and that major modifications potentially affecting the durability of the graft were undertaken. The specific graft and procedural risks were explicitly discussed, and patients completed a written consent whenever possible.
Technical success was described as successful stent-graft deployment with revascularization of all intended vessels and aneurysm exclusion. Primary outcomes were technical success, 30-day mortality, 30-day reintervention, and complications. Secondary outcomes were long-term survival, graft patency, late reintervention, and endoleak incidence. All patients had a CT scan before discharge and at regular intervals thereafter.
Technique
Three-dimensional CT angiography reconstructions and measurements of renovisceral vessels, clock position, and distances were performed with the Intuition workstation (TeraRecon Inc, San Mateo, CA, USA). The commercially available aortic stent-grafts used for surgical modification were the Zenith Flex AAA (bifurcated), Zenith TX2 TAA with Pro-Form (tubular), Zenith Renu AAA main body extensions, Zenith Alpha Thoracic, and the Zenith t-Branch (Cook Medical), as well as a Valiant Captivia (Medtronic, Minneapolis, MN, USA) according to patient anatomy and lesion characteristics (Table 1).
The technique of surgeon-modified stent-grafts has been previously described (Figure 1).9,10 Modification of the aortic stent-grafts was performed prior to operation under sterile condition. The number of surgeon-modified scallops, fenestrations, or directional branches depended on aneurysm extent. Fenestrations were usually sized 6 to 8 mm in diameter, positioned between stent struts, and reinforced with a radiopaque wire for intraoperative orientation under fluoroscopy.
Modification of a stent-graft. (A) Reinforcement of a fenestration with a radiopaque wire. (B) Introduction of a constraining wire and (C) diameter-reducing ties. (D) The 4-vessel surgeon-modified stent-graft (E) before resheathing.
For implantation, open surgical access was obtained from both femoral arteries. Typically, the main graft was introduced from the right femoral artery, and visceral artery catheterization was achieved from the left femoral artery. Depending on visceral vessel anatomy an additional access from the right brachial artery was established in selected cases to catheterize target vessels from above. Bridging stents were implanted through the fenestrations into the target vessels using a balloon-expandable stent-graft; if necessary, relining was performed with a self-expanding nitinol stent. Stent-grafts were proximally flared with a 10- or 20-mm balloon to seal within the main graft. To address the risk of spinal cord ischemia (SCI), cerebrospinal fluid (CSF) drainage was strongly considered when 2 of the following applied: (1) planned coverage of >20 cm of the thoracic aorta, (2) prior infrarenal aortic repair, (3) occlusion of a hypogastric or vertebral artery, and (4) planned coverage of the left subclavian artery. Postoperative spinal cord ischemia surveillance and protocol involved permissive hypertension (goal mean arterial pressure of 80–90 mm Hg), continuous neurological examination, and hemoglobin maintenance >9 g/dL. In case of suspected SCI, a postoperative cerebrospinal fluid drain was applied. All patients received a CT scan before discharge.
Results
Four patients were treated with a single fenestration or scallop, 16 patients had a modification with 3 or 4 fenestrations (Figure 2), and 1 patient received a modified t-Branch with 2 directional branches for the SMA and celiac trunk. A total of 69 vessels were targeted with 58 fenestrations, 9 scallops, and 2 directional branches. Further technical characteristics are reported in Table 2.
(A) Intraprocedural fluoroscopy with catheterization of the renal, superior mesenteric, and celiac arteries. (B) Final angiography of a surgeon-modified stent-graft with 4 fenestrations.
Types of Repair in the 21-Patient Cohort.
Abbreviations: SMA, superior mesenteric artery.
Prophylactic spinal drainage was applied in 1 patient who received a 4-vessel sm-FBSG for a symptomatic false lumen aneurysm due to prior thoracic endovascular aortic repair (TEVAR). Adjuvant procedures included an iliac conduit in 2 cases due to external iliac artery rupture and 1 right subclavian artery conduit. This specific patient had left-sided iliac occlusion, and all sheaths for the fenestrations were delivered though a trans-subclavian approach. A type I TAAA patient received a scallop for the SMA and additional coil embolization of the aneurysm sac in a case of contained rupture. Simultaneous transcatheter aortic valve replacement was performed with a 3-vessel sm-FBSG in a patient with a contained rupture of a pararenal penetrating ulcer and high-grade valvular aortic stenosis.
Technical Results
Successful aneurysm exclusion as well as target vessel revascularization was achieved for all aneurysms and target vessels. The mean procedure time without modification was 360 minutes (range 135–635). Among the 14 4-vessel fenestrated stent-grafts, the mean procedure time of the first 7 cases was 453 minutes (range 200–635) while the mean time of the second half was 342 minutes (range 265–435). The mean volume of contrast agent was 151 mL (range 44–220), and the mean fluoroscopy time was 88 minutes (range 31–258).
The mean length of stay in the intensive care unit was 9 days (range 1–32) and the mean total in-hospital stay was 19 days (range 1–78, median 11). Twelve patients were discharged home; 6 patients were discharged to a nursing facility or rehabilitation center.
Mortality and Morbidity
Thirty-day survival was 95%. One patient who received a 4-vessel fenestrated stent-graft for a symptomatic pararenal aneurysm died in the early postoperative period due to massive reperfusion syndrome. The procedure was strikingly prolonged due to relevant intraoperative difficulties that delayed reperfusion of the extremities up to 6 hours. After restoring limb flow, the condition of the patient worsened, and upper and lower legs on both sides presented with compartment syndrome. The patient immediately received a fasciotomy, partially under cardiopulmonary resuscitation due to hyperkalemia. Dialysis and maximal conservative treatment was promptly started, but the patient nevertheless died because of the therapy-refractory hyperkalemia and consequent cardiac arrest.
Two further patients died in-hospital. One suffered from sepsis and multiorgan failure due to an aortogastric fistula after sm-FBSG, as well as gastrectomy and necrotizing pancreatitis requiring necrosectomy. 12 The second patient initially presented with a type Ia endoleak after endovascular aneurysm repair (EVAR) for contained rupture of a type IV aneurysm. This patient received a sm-FBSG but died due to a septic course refractory to antibiotic therapy. The cause of sepsis was unclear, but a mycotic aneurysm was suspected due to the morphology and rapid growth of the aneurysm. Major adverse events are presented in Table 3.
Major Adverse Events in the 21-Patient Cohort.
Four patients suffered neurological symptoms due to SCI; 2 had permanent paralysis. The first of the 4 patients had a type I TAAA and received a long-segment TEVAR with coverage of the celiac trunk and a scallop for the SMA. After onset of mild SCI, the patient underwent revascularization of the stenotic right hypogastric artery and had complete remission of neurological symptoms (SCI grade 1). The second patient had a 4-vessel fenestrated endograft for a ruptured pararenal penetrating ulcer. The patient suffered an intraoperative iliac rupture associated with hypotension and a prolonged procedure. Postoperatively, permanent paraplegia occurred despite immediately applied CSF drainage (SCI grade 3c). The third patient had a ruptured type IV aneurysm with a 4-vessel sm-FBSG. Risk factors for SCI were moderate hypotension intraoperatively due to an ejection fraction of 25% as well as 800-mL blood loss. The patient had CSF drainage postoperatively but only had slight remission of symptoms (SCI grade 3b). The fourth patient presented with mild SCI in terms of hypoesthesia of both feet after 4-vessel repair of a symptomatic pararenal penetrating ulcer with intraoperative revascularization of the right hypogastric artery. After CSF drainage was applied, neurological symptoms completely remitted (SCI grade 1).
Two patients required prolonged intubation with tracheostomy. Acute renal failure occurred in 3 patients, 2 of whom had chronic renal insufficiency prior to the index procedure. One patient received temporary dialysis and 1 patient required permanent dialysis.
Two early reinterventions and 2 reoperations were conducted, including the aforementioned revascularization of a hypogastric artery. The other endovascular reintervention involved implantation of an Optimus XXL stent (AndraTec, Koblenz, Germany) to prevent late endoleak due to proximal infolding of the prosthesis. As mentioned above, one patient with a secondary aortogastric fistula and pancreatitis required gastrectomy and pancreatic necrosectomy. Exploratory laparotomy was conducted in a patient with suspected acute mesenteric ischemia, which was not intraoperatively confirmed. There was no perioperative stroke, major cardiac event, extremity ischemia, or bowel ischemia requiring resection. Three patients suffered from minor bleeding (1 access site and 2 non–aneurysm-related).
All patients had postprocedural CT angiography within 1 week after the procedure (Figure 3). Six early endoleaks were detected (3 type II and 3 minor type III), but no reinterventions were undertaken. All target renovisceral vessels were patent.
Postoperative reconstruction computed tomography angiography scan of a 4-vessel surgeon-modified stent-graft.
Follow-up
The mean follow-up was 11.2 months (range 2–33) in 17 patients. Thirteen of 18 discharged patients received a CT scan as follow-up imaging. In all these scans, all target vessels were patent; 2 type II and 1 type I endoleaks were seen. Four patients died during follow-up of an aneurysm-unrelated cause, 3 patients because of decompensated heart failure and 1 patient due to gastric cancer.
Additionally, 1 late aneurysm-related death was observed. This patient has been treated with a 4-vessel fenestrated endograft due to a type Ia endoleak after prior EVAR; he presented with a contained rupture of a type IV aneurysm, which was suspected to be mycotic. Four months after the sm-FBSG, he presented with progression of the aortic aneurysm in the thoracic aorta and a new type Ia endoleak, which was attributed to continued infection due to an aortoenteric fistula. The patient subsequently received a TEVAR and chimney graft for the celiac trunk, as well as an open surgical treatment with laparotomy, resection of the fistula, and Roux-en-Y gastrojejunostomy. Although the patient was discharged, he had a recurrent septic course with suspected graft and aneurysm sac infection refractory to drainage and antibiotics. The patient was considered unsuitable for open aneurysm repair and died after recurrent aortic bleeding 6.5 months after the sm-FBSG.
Discussion
In emergent settings, the available options for TAAA and AAA patients at high perioperative risk are limited. Mortality of open repair increases to 40% to 80% in complex aortic aneurysm,13,14 therefore endovascular techniques have found their way into emergent treatment strategies. Several reports of chimney graft repairs of complex abdominal or thoracoabdominal have been published, mainly concerning feasibility of this procedure,15–21 but long-term results are still missing. Although chimney grafting is a useful technique in urgent cases, 22 it does not establish an anatomically correct repair.
Fenestrated and branched stent-grafts for complex abdominal and thoracoabdominal aortic repair in high-risk patients and elective cases have become a safe approach, with low perioperative morbidity and mortality.8,22–24 Since customized endografts take weeks for delivery and off-the-shelf devices for complex aortic anatomy are limited, the surgeon-modified stent-graft appears to be an effective endovascular option because it facilitates an anatomically correct reconstruction; modification of the endograft takes only about 1 to 2 hours. Limitation of this technique is that it can be applied only to hemodynamically stable patients due to the modification time.
This study presents a large series of high-risk patients treated with a sm-FBSG for urgent complex abdominal and thoracoabdominal aortic pathologies. Despite the significant comorbidities of the treated patients, 30-day survival was 95%, which demonstrates the short-term benefit of this approach for these high-risk patients. During the same 2.5-year period, 106 patients were treated at our center with a 3 or 4-vessel fenestrated or branched EVAR using custom-made and off-the-shelf devices, underlining our center’s experience with complex aortic aneurysm repair. Among these patients, 20 were symptomatic and 6 had contained ruptures treated with the t-Branch; 30-day mortality was 5% for elective cases and 19.2% for urgent cases.
Georgiadis et al 25 recently published a systematic review comparing off-the-shelf fenestrated/branched grafts (OSFG) with sm-FBSG in 308 patients with complex aortic aneurysm. One-third of all patients had emergent indications for operation. Major adverse events occurred in 12.8% of sm-FBSG and 7.4% of OSFG patients; death occurred in 3.2% of the sm-FBSG group. As regards clinical success, graft patency, and mortality, the authors concluded that both techniques were effective and safe, with comparable results.
The 30-day mortality of our series was relatively low considering the emergent indications for operation and comorbidities of the patients. The 1 patient who died during the 30-day period was our fourth patient treated with a 4-vessel sm-FBSG and the first and only case of fulminant reperfusion and compartment syndrome among all fenestrated and branched EVAR patients. This case sharpened our awareness for this lethal complication, and we thereafter adjusted our procedural standards toward earlier removal of the sheaths if possible, smaller sheaths, distal perfusion over sheaths, and improved communication with our anesthesiologists before and during reperfusion of the legs.
Of the 2 further in-hospital deaths, 1 patient died after treatment with a 4-vessel fenestrated endograft for a type Ia endoleak after EVAR for contained rupture of a type IV aneurysm, which was suspected to be infected. A fenestrated EVAR procedure was selected due to severe comorbidities and a hostile abdomen after pancreatectomy. The patient died due to aneurysm progression, bleeding, and septic course. Our limited experience with a few patients presenting with infected aneurysm demonstrates that a repair with sm-FBSG is technically possible in the acute situation, but survival remains poor due to recurrent sepsis despite the use of rifampicin-impregnated stent-grafts, drainage, and long-term antibiotics. Hence, endovascular therapy in patients with suspected infected aneurysm should be selected very restrictively and seen as a bailout option in the emergent setting.
The 4 late aneurysm-unrelated deaths underline the fact that the treated cohort was truly at high risk due to comorbidities, even while treated endovascularly. As a consequence, the indication for operation of these patients should be evaluated very strictly. Also, patients at high risk should be very well informed about morbidity and mortality with and without treatment.
The mean 360-minute procedure time is explained by the complexity of a few cases, including the 2 external iliac artery ruptures with conduit implantation, 1 right subclavian artery conduit, and the deceased patient with compartment syndrome after a prolonged procedure. In all 14 4-vessel sm-FBSG patients, the mean procedure time was reduced by 111 minutes for the second half of the patients, which reflects the learning curve with this sophisticated technique.
In our study, there was relevant postoperative morbidity, which needs to be discussed in further detail. Overall, the incidence of SCI was relatively high, although 2 of the 4 patients had complete regression of their symptoms. Onset of SCI in general is multifactorial. Perioperative blood loss, lack of CSF drainage, need for iliac conduit,26,27 hypotension, as well as procedure time and patient comorbidities probably contributed to the occurrence of SCI. Nevertheless, this devastating complication of endovascular therapy heightened our awareness during intra- and postoperative monitoring and SCI therapy, as well as during procedure planning. Regarding the option of prophylactic CSF drainage, we are cautious in the emergent setting as a failed puncture might result in intraspinal bleeding or hematoma owing to intraoperative heparinization. However, we have currently become more liberal with the use of a CSF drain in emergencies.
There was no perioperative stroke, major cardiac event, or extremity ischemia in our cohort. Three patients had postoperative renal complication but two had chronic renal insufficiency prior to the operation. Thus, patients need to be informed about the risk of postoperative dialysis dependency, and operators should restrict the amount of nephrotoxic contrast agent and, if possible, use a nontoxic contrast agent such as CO2. Graft patency and endoleak rates appeared satisfactory; only 1 reintervention was conducted to treat a type Ia endoleak with a chimney graft and TEVAR.
The study is limited by its retrospective single-center design, as well as the limited number of patients and missing midterm follow-up. Also, there was no open surgical repair cohort for comparison as this is a highly selected patient population. Despite these limitations, the results support the surgeon-modified endovascular technique as a valuable alternative in selected high-risk patients requiring emergent treatment for complex aortic pathology.
Conclusion
Sm-FBSG can be utilized for urgent treatment of symptomatic complex abdominal and thoracoabdominal aortic pathology or contained aortic rupture in hemodynamically stable, high-risk patients with anatomy unsuitable for commercial stent-grafts. Despite relevant perioperative morbidity and mortality, patients may benefit by avoiding the higher risk of open surgical treatment. Short-term graft patency was satisfactory. Until more off-the-shelf fenestrated-branched devices are manufactured that fit different anatomies and are available without delay, sm-FBSG may be a good treatment alternative to open repair or chimney EVAR.
Footnotes
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 has intellectual property with Cook Inc. Nikolaos Tsilimparis is a proctor for Cook Inc.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.