- Email: [email protected]
A multicenter experience with the surgical treatment of infected abdominal aortic endografts
From the Southern Association for Vascular Surgery
A multicenter experience with the surgical treatment of infected abdominal aortic endografts Victor J. Davila, MD,a William Stone, MD,a Audra A. Duncan, MD,b Emily Wood, MD,b William D. Jordan Jr, MD,c Nicholas Zea, MD,c W. Charles Sternbergh III, MD,d and Samuel R. Money, MD, MBA,a Phoenix, Ariz; Rochester, Minn; Birmingham, Ala; and New Orleans, La Objective: Single-center experiences with the treatment of infected endografts after endovascular aortic repair (I-EVAR) have been reported. We performed a multicenter review of the surgical care of these patients to elucidate short-term and long-term outcomes. Methods: A retrospective analysis of all EVAR explants from 1997 to 2014 at four institutions was performed. Patients with I-EVAR undergoing surgical treatment were reviewed. Data were obtained detailing preoperative demographics, and postoperative morbidity and mortality. Results: Thirty-six patients (30 male) were treated with endovascular graft excision and revascularization for I-EVAR with a median age of 69 years (range, 54-80 years). Average time from the initial EVAR to presentation was 589 days (range, 432466 days). Preoperative comorbidities included hypertension, 32 (89%); tobacco use, 31(86%); coronary artery disease, 26 (72%); hyperlipidemia, 25 (69%), peripheral artery disease, 13 (36%); cerebrovascular disease, 10 (28%); diabetes, 10 (28%); chronic obstructive pulmonary disease, 9 (25%); and chronic kidney disease, 9 (25%). The most common presenting patient characteristics were leukocytosis, 23 (63%); pain, 21 (58%); and fever, 20 (56%), which were present an average of 65 days (range, 0-514 days) before explantation. Nine different types of endograft were removed. Three patients (8%) underwent emergency explantation. Thirty-four patients (89%) underwent total graft excision, and two patients (6%) underwent partial excision. Methods of reconstruction were in situ in 27 (75%) and extra-anatomic in nine (28%). Conduits used were Dacron (DuPont, Wilmington, Del), with or without rifampin, polytetrafluoroethylene, cryopreserved allograft, and femoral vein. Forty-nine organisms grew from operative cultures. Gram-positive organisms were the most common, found in 24 (67%), including Staphylococcus in 13 (36%) and Streptococcus in six (17%). Anaerobes were cultured in 6 patients (17%), gramnegative organisms in 6 (17%), and fungus in 5 (14%). Thirty-one patients (86%) received long-term antibiotics. Early complications included acute renal failure requiring dialysis, 12 (33%); respiratory failure, 3 (8%); bleeding, 4 (11%); and sepsis, 2 (6%). Six patients required re-exploration due to hematoma, infected hematoma, lymphatic leak, bowel perforation, open abdomen at initial operation, and anastomotic bleeding. Perioperative mortality was 8% (3 of 36), and long-term mortality was 25% (9 of 36) at a mean follow-up of 569 days (range, 0-3079 days). Type of reconstruction (in situ vs extra-anatomic) or conduit type did not affect perioperative or overall mortality. Conclusions: I-EVAR is a rare but potentially devastating clinical problem. Although perioperative mortality is acceptable, long-term mortality is high. The most common postoperative complication was acute renal failure requiring dialysis. Although this is the largest series of I-EVAR, further studies are needed to understand the risk factors and preventive measures. (J Vasc Surg 2015;62:877-83.)
From the Division of Vascular Surgery, Department of Surgery, Mayo Clinic Arizona, Phoenixa; the Division of Vascular Surgery, Department of Surgery, Mayo Clinic, Rochesterb; the Division of Vascular Surgery, Department of Surgery, University of Alabama at Birmingham, Birminghamc; and the Division of Vascular Surgery, Department of Surgery, Ochsner Clinic Foundation, New Orleans.d Author conflict of interest: W.D.J.: clinical investigator (paid to University of Alabama at Birmingham) for Medtronic, Gore, Cook, Endologix, Aptus, and Lombard; Cordis: consultant (paid to University of Alabama at Birmingham) for Medtronic, Gore, Endologix, Aptus, Colvano, and Lombard. S.R.M: consultant for Cook Medical and Gore Medical. Presented at the Thirty-ninth Annual Meeting of the Southern Association for Vascular Surgery, Scottsdale, Ariz, January 14-17, 2015. Correspondence: Samuel R. Money, MD, MBA, Mayo Clinic, Division of Vascular Surgery, 5777 E Mayo Blvd, Phoenix, AZ 85054 (e-mail: [email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest. 0741-5214 Copyright Ó 2015 by the Society for Vascular Surgery. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jvs.2015.04.440
More than 1 million Americans have abdominal aortic aneurysms (AAAs), many of which may require intervention. Open surgical intervention has slowly been replaced by endovascular AAA repair (EVAR) using aortic endografts.1 EVAR was initially introduced for high-risk individuals whose open operative risk was deemed excessive. Advancements in endograft development have brought a shift in the way that most AAAs are managed; currently, most AAA repairs performed in the United States are by an endovascular approach.2 Aortic endovascular graft infection (I-EVAR) is an unusual complication of endograft repair and has a reported incidence between 0.05% and 5%.3 The higher rates of infection reported may be the result of endograft deployment in infected fields, such as those in patients with mycotic aneurysms or aortoenteric fistulae.4 Mortality from endograft explantation secondary to infection has been reported as high as 30% and is similar to infection of open aortic grafts.5 To date, the treatment of I-EVAR 877
JOURNAL OF VASCULAR SURGERY October 2015
878 Davila et al
Table I. Survival summary Interval
Survival, % Time, months No. at risk 95% CI
1-2 months 2-10 months 10-21 months 21-28 months 28-101 months 101 months
95 90 84 74 55 0
1 2 10 21 28 101
21 20 14 8 4 1
87-100 79-100 69-100 53-100 29-100 NA
CI, Confidence interval.
has been reported mainly by single centers. This report details the results of a multicenter experience from four geographically distinct medical centers, which will assist delineating procedural techniques, clinical challenges, and results in this patient population. METHODS After approval from the Institutional Review Boards of the four medical centers involved in this study (Mayo Clinic, Rochester, Minn and Phoenix, Ariz; University of Alabama at Birmingham, Birmingham, Ala; and Ochsner Medical Institutions, New Orleans, LA), databases were queried to identify patients who underwent excision of infected aortic endografts from AAAs between 1997 and 2014. The Investigational Review Board reviewer approved waiver of the requirement to obtain informed consent. Patients who underwent explantation were included for review regardless of the institution where the endograft was initially placed. A retrospective review included pertinent history, presentation, physical findings, results of microbiologic testing, location of infection, management, and outcomes. Demographic data for each patient were collected and evaluated, including comorbidities. Time from the original EVAR implantation until the emergence of presenting signs or symptoms of infection and the type of endograft placed at original operation were also included for review. Patients were deemed to have I-EVAR based on presence of positive blood cultures, radiologic evidence of abscess or infection, intraoperative evidence of infection as described by the operating surgeon, or positive intraoperative cultures from explanted graft material, aortic wall, or aneurysm sac contents. Management was based on the surgeon’s judgment, including method of reconstruction, timing of the operation, antibiotic duration and type, method of explantation, including proximal control, and postoperative management. Operative data collected included indications for intervention, urgency of the procedure, and the method of reconstruction. Early complications were defined as those #30 days of removal of the infected endograft, and late complications were those that occurred afterward. Perioperative mortality was defined as death #30 days. Long-term mortality was defined as death >30 days. Demographic data describing the frequency of postoperative complications, explanted device type, and symptoms
Fig 1. Survival curve for infected endograft after endovascular aortic aneurysm (AAA) repair (I-EVAR).
at presentation are presented. Because of the small cohort size, most data are reported as mean and range. KaplanMeier survival curves were used to provide 10-month and 21-month survival estimates (Table I and Fig 1). These time frames are somewhat shifted from normal yearly intervals based on patient dropout and length of follow-up in this retrospective study. Survival was summarized after each death and reported for 1 and 2 years. All analysis was performed in R 3.1.2 software (The R Foundation for Statistical Computing, http://www.r-project.org/foundation/), and the package survival version 2.37-7 was used for survival analysis. RESULTS Thirty-six patients, including 30 men (83%), with a mean age of 69 years (range, 54-80 years) underwent explantation of an aortic endograft for infection and were included for review. The preoperative comorbidities, including hypertension, previous or current tobacco use, coronary artery disease, and others, appear to be those commonly seen in patients with aortic aneurysmal disease (Table I). Presenting symptoms varied, however; a majority (21 patients [58%]) presented with abdominal pain or fever (20 patients [56%]), or both (Table II). Blood cultures were positive in 10 of 34 patients (29%). Mean duration of these symptoms in these patients was 65 days before explantation (range, 0-514 days). Mean duration from initial endograft placement until development of symptoms was 589 days (range, 43-2466 days). Multiple diagnostic imaging modalities were used to establish the diagnosis. Computed tomography (CT) was used in all patients and confirmed an infected aortic endograft in 33 of 36 patients (91.7%; Fig 2). The CT in three patients did not demonstrate obvious perigraft infection. One
JOURNAL OF VASCULAR SURGERY Volume 62, Number 4
Davila et al 879
Table II. Presenting characteristics
Table III. Explanted endografts
Characteristic
No. (%)a
Type of graft
No. (%)
Pain Pain (alone) Pain and fever Pain and leukocytosis Pain and fever and leukocytosis Fever Fever (alone) Fever and pain Fever and leukocytosis Fever and pain and leukocytosis Fever and GI bleeding Leukocytosis Leukocytosis (alone) Leukocytosis and pain Leukocytosis and fever Leukocytosis and pain and fever Leukocytosis and GI bleeding
21 6 2 6 7 20 3 2 7 7 1 23 1 6 7 7 1
Excluder (Gore, Flagstaff, Ariz) AneuRx (Medtronic, Minneapolis, Minn) Zenith (Cook, Bloomington, Ind) Endologix (Endologix, Irvine, Calif) Endurant (Medtronic) Viabahn (Gore) Ancure (Guidant, Indianapolis, Ind) Talent (Medtronic) TAG (Gore)
13 9 6 2 2 1 1 1 1
(58) (17) (6) (17) (19) (56) (8) (6) (19) (19) (3) (63) (3) (17) (19) (19) (3)
GI, Gastrointestinal. a One patient was asymptomatic but had evidence of infection on computed tomography (CT) imaging.
Fig 2. Representative computed tomography (CT) image of an infected endograft after endovascular aortic aneurysm (AAA) repair (I-EVAR).
patient had leukocytosis with postprandial pain, with CT showing perigraft fluid but no secondary signs of infection (air or fat stranding), and was found to have an abscess pocket along the lateral side of the aneurysm sac. One patient presented with a psoas abscess after a spinal procedure, which was found intraoperatively to be in continuity with the endovascular prosthesis. The remaining patient was undergoing a workup as an outpatient for possible endograft infection when he presented with a signs and symptoms of a ruptured aneurysm, and subsequent intraoperative cultures were positive. Magnetic resonance imaging was performed in 12 patients (33%) and identified an infected endograft in nine of these patients (75%). Tagged white blood cell scans were positive in 21 of 28 patients (75%).
(36) (25) (16) (6) (6) (3) (3) (3) (3)
Nine different types of endograft were explanted for treatment of infection (Table III). Emergency explantation was performed in three patients (8%). One patient presented with a ruptured aneurysm, and two patients presented with acute onset of back pain and large aneurysms. Thirty-four patients (94%) underwent total graft excision, and two patients (6%) underwent partial excision. Partial endograft excision was used when only a portion of the endograft was determined to be infected. This judgment, which was made by the operating surgeon, primarily involved good incorporation of the remaining endograft into the surrounding aortic or iliac artery wall. Of the two patients who underwent partial excision, one patient was alive at the last follow-up of 811 days, and one patient died on postoperative day 44 from an anastomotic dehiscence after in situ femoral vein reconstruction. Suprarenal fixation was present in 24 of the explanted endografts (67%). Twenty-five percent (6 of 24) of these patients developed acute renal failure requiring dialysis in the postoperative period. Acute renal failure requiring dialysis developed in 56% of patients (6 of 12) with only infrarenal fixation. After endograft explantation, methods of reconstruction included in situ graft placement in 27 patients (75%) and/or extra-anatomic axillobifemoral bypass in nine (28%). In situ reconstructions consisted of rifampinsoaked Dacron (DuPont, Wilmington, Del) graft in 14 patients (39%), standard nonantibiotic-treated Dacron grafts in 2 (6%), polytetrafluoroethylene in 2 (6%), femoral vein in 4 (11%), cryopreserved graft (CryoLife Inc, Kennesaw, Ga) in 4 (11%), and a combination of femoral vein and cryopreserved graft in 1 patient (3%). Axillobifemoral bypass (n ¼ 9) was performed during the EVAR explantation (n ¼ 6) or in a staged fashion (n ¼ 3). The conduits used for extra-anatomic bypass included polytetrafluoroethylene in 6 patients (17%), rifampin-soaked Dacron in 2 (6%), and a combination of rifampin-soaked Dacron and femoral vein in 1 (3%). Intraoperative cultures were obtained from the explanted graft, aneurysm wall, or aneurysm sac contents in all patients and were positive in 28 patients (81%). Gram-positive cocci were the most common isolates and were identified in 24 patients (67%; Table IV). Patients were treated with broad-spectrum intravenous antibiotics and then transitioned to oral antibiotic therapy determined
JOURNAL OF VASCULAR SURGERY October 2015
880 Davila et al
Table IV. Microbiology of infected endografts
Variable Organisms Gram-positive Staphylococcus Anaerobes Gram-negative Streptococcus Fungus Patients with positive intraoperative cultures Organisms per patient, average (range), No.
Table V. Early complications
Organisms isolated Patients (n ¼ 49), No. (%) (n ¼ 36), No. (%) 32 13 11 8 6 5
(65) (27) (22) (16) (12) (10)
24 13 6 6 6 5 28/36
(67) (36) (17) (17) (17) (14) (80)
1.4 (0-5)
by culture sensitivities at the discretion of the treatment team. Lifelong antibiotic regimens were continued in 19 patients (53%), 10 (28%) were maintained on antibiotics for periods >5 weeks and up to 2 years, 2 (6%) received <6 weeks of antibiotic therapy, and 5 (14%) did not have adequate documentation to determine the duration of antibiotic therapy. Three patients (8%) died #30 days after explantation in this series. Early deaths occurred in two patients in the in situ reconstruction group and in one in the extraanatomic reconstruction group. Acute renal failure requiring dialysis was the most common early complication. Other complications included respiratory failure requiring tracheostomy, bleeding, and sepsis (Table V). Six patients (17%) required re-exploration after explantation for indications that included hematoma, infected hematoma, lymph leak, and small-bowel perforation in one patient each. An additional patient developed an aortic anastomotic dehiscence after in situ reconstruction with femoral vein resulting in hemorrhage and was classified as an early death. One other patient had a planned reoperation for secondary closure of the abdominal wall. Mortality for patients undergoing reoperation was significant (n ¼ 3 [50%]). Two patients needing reoperation were early deaths: one had sepsis after aortoiliac reconstruction with a Dacron graft and omental flap, and one had anastomotic bleeding. The third early death was secondary to aspiration and subsequent respiratory failure. Overall mortality was 25% (9 patients) at a mean follow-up of 570 days (range, 0-3079 days). Late deaths included one death from proximal aneurysm rupture, one death from a chronic inflammatory illness, and the cause of death was not determined in four patients. There were four late deaths in the in situ reconstruction group and two late deaths in the extra-anatomic reconstruction group. Although one form of reconstruction did not appear to be superior in this series, the sample size was too small for any meaningful statistical comparison. Of the 36 patients studied, there were three early deaths. These deaths were not included in the survival analysis for those who successfully endured the revision. Of the
Early complications
No. (%)
Acute renal failure Bleedinga Respiratory failure requiring tracheostomy Sepsis Need for re-exploration in operating room Hematoma Infected hematoma Lymphatic leak Small-bowel perforation Open abdomen Anastomotic bleeding
12 4 3 2 6 1 1 1 1 1 1
(33) (11) (8) (6) (17) (3) (3) (3) (3) (3) (3)
a One patient each developed hemothorax, a retroperitoneal hematoma, gastrointestinal bleeding managed endoscopically, and an anastomotic dehiscence of the aortic reconstruction requiring reexploration.
remaining 33 patients, follow-up data existed for 26 patients. Six of those patients died during the follow-up period, for a total death rate of 25% (9 of 36). KaplanMeier analysis was performed only on those patients for whom follow-up data were available, making the death rate during the observation period 23% (6 of 26). Survival was 84% (95% confidence interval, 69%-100%; 14 of 26 at risk) at 1 year and 74% (95% confidence interval, 53%100%; 8 of 26 at risk) at 2 years. An accurate estimate of median survival time cannot be provided from these data due to censoring, but appears to be in excess of 30 months. Survival is summarized in Table I. DISCUSSION Patients who develop I-EVARs carry a significant risk, but our multi-institutional series demonstrates that they can be treated by excision and revascularization with a reasonably low mortality (8%). Long-term survival in these patients was 75% with a mean follow-up of >18 months. Only two patients died of complications attributable to their aortic disease. One early death was caused by an anastomotic dehiscence, and one late death from proximal aortic aneurysm rupture. The results of this multiinstitutional review compare favorably with the results other studies.6,7 However, despite what may be considered an acceptable 30-day mortality, the risk of postoperative morbidity is significant (n ¼ 21 [58%]), with acute renal failure requiring dialysis occurring as the most common event in one-third of the patients. In addition, six patients (17%) required re-exploration due to a variety of causes, including small-bowel perforation, anastomotic bleeding, lymphatic leak, and infected hematoma. Mortality for patients undergoing re-exploration was significant (50%). Median survival cannot be estimated in this study due to the large amount of dropout. In general, this is a heterogeneous subgroup, and interpretation of Kaplan-Meier results from a small, heterogeneous population should be done with care. Infection remains a rare complication after endovascular aortic graft placement (0.5%-5%).5 The causes of this complication can be clearly delineated in some cases. The
JOURNAL OF VASCULAR SURGERY Volume 62, Number 4
etiology of early endograft infection could be related to endograft deployment through an infected field or in an area known to have aortic infection. Endografts have been used as a temporizing measure to stabilize patients with aortoenteric fistula.4 In this clinical scenario, endografts without suprarenal fixation should be used to make future explantation technically easier. Other reasons for infection to develop in aortic endografts may be multifactorial. Patients with systemic bacteremia from ongoing infections, such as urinary tract infections, pneumonia, or even from dental surgery, may cause secondary infection of an aortic endograft. In a survey by Lockhart et al8 in 2002, 35% of infectious disease physicians would “always” or “usually” recommend antibiotic prophylaxis for dental procedures in patients who have a vascular graft prosthesis, and 30% would “never” recommend prophylaxis before a dental procedure. The potential for secondary infection of an aortic endograft could follow additional procedures other than dental. Percutaneous coiling or placement of sealant or glue into an aneurysm sac could potentially introduce bacteria and lead to secondary infection. As with the placement of any prosthetic device, the endovascular graft has the potential to become infected during initial placement. Theoretically, EVAR minimizes contamination of the prosthetic device compared with standard open surgery because the device is housed in a sterile delivery system. However, Ducasse et al9 noted that endografts placed in interventional radiology suites have a higher rate of infection compared with traditional operating rooms, possibly because of different levels of sterility when equipment is handled. Mycotic aneurysms represent another situation where an endograft may be deployed in an infected field. In our retrospective review, a few patients may have had mycotic aneurysms at the original placement of the aortic endograft; for example, one patient developed vertebral body osteomyelitis likely secondary to a mycotic aneurysm. Although the initial imaging showed no evidence of this, imaging showed erosion of the vertebral body and subsequent spread of infection into the aneurysm sac. Multiple potential causes of I-EVAR are possible; however, the actual cause is not clearly discernible in many cases. It is important to closely evaluate the surrounding aortic wall and structures for abnormalities before the initial endograft is placed to avoid deployment in a potentially infected field because of the implications of resultant endograft infection. The presentation of I-EVAR is often insidious. Many of the patients in our series presented with pain (58%) and fever (56%), which are not specific to I-EVAR. Leukocytosis was present only in approximately two-thirds of the patients. The best diagnostic study was a thin-slice CT scan, which was performed in all patients and diagnostic in 91.7%. Tagged white blood cell scanning was positive in 73% of patients when ordered and may be of additional benefit when clear-cut evidence is not present on CT scanning. Most of the patients in this series had a very subtle
Davila et al 881
presentation, which can be beneficial because it allows time to establish a diagnosis, optimize comorbidities, and formulate a defined treatment plan tailored to the patient. However, the nebulous presentation can also be detrimental, because a delay in diagnosis can result in aneurysm rupture and exsanguination. This subtle presentation may explain why the patients in this study presented with symptoms for an average of 65 days (range 0-514 days) before graft explantation. Various options for reconstruction exist after explantation. Classically, extra-anatomic reconstruction in the form of an axillobifemoral bypass, followed by endograft explantation and aortic ligation, was the standard approach for most infected aortic prostheses. In situ reconstructions allow for prevention of aortic stump blowout and the complications associated with possibly tenuous extra-anatomic bypass, including a reduced patency rate.10,11 Cryopreserved grafts or rifampin-soaked Dacron can be used for in situ reconstruction. Fatima et al3 demonstrated superior results with full coverage with omentum in those patients undergoing in situ replacement with rifampin-soaked Dacron grafts. They describe the technique of mobilizing the omentum, raising a tongue of omentum in the retrocolic configuration, and performing a 360 omental wrap around the Dacron graft. Microorganisms were isolated from operative cultures in 28 of 36 patients (77%) and were predominantly gram-positive organisms. Anaerobes and gram-negative organisms were also present but not as commonly identified. Cultures were positive for fungus in five patients (14%). Most of these patients demonstrated a polymicrobial process. Cultures of eight patients (22%) were negative. These patients may have had preoperative antibiotic therapy precluding growth from samples obtained in the operating room. All patients with negative cultures were diagnosed with I-EVAR through extensive preoperative workup. A spectrum of findings leads to the diagnosis of I-EVAR in these cases, including air in the aneurysm sac, perigraft fluid, and retroperitoneal abscess or fat stranding. One patient, who was not diagnosed by CT scan and had no evidence of microbial growth, had a large psoas abscess. Exploration of the abscess showed it was in continuum with the aorta and permeated through the native aortic wall in contact with the endograft. An in situ repair with femoral vein was performed in this patient. This study is in accordance with others that place emphasis on CT scanning for definitive diagnosis.3 Numerous small studies have examined the role of nonoperative treatment of I-EVAR. Moulakakis et al12 published a meta-analysis of 17 reviews including 29 patients with I-EVAR who were treated with preservation of the stent graft. The in-hospital mortality was 21%, and seven more patients died during a mean follow-up of 11.4 months, for an overall mortality of 45%.12 This reported mortality appears greater than our 30-day mortality of 8% and long-term mortality of 25%. We believe that nonexcisional treatment of I-EVAR should be limited to patients with a prohibitive operative risk.
JOURNAL OF VASCULAR SURGERY October 2015
882 Davila et al
There are no set standards for the duration of antibiotic treatment in infected aortic grafts or in infected endovascular aortic grafts. Duration of antibiotic treatment in these patients postoperatively was left to the treating surgeon and his or her team. Most of these patients received intravenous antibiotics for 4 to 6 weeks, with most transitioned to a lifetime of oral antibiotic therapy. This regimen is based purely on physician preference, without appropriate evidence-based studies in the literature. The lack of scientific rigor to guide our management after resection of the infected endograft requires attention. Treating surgeons should understand that this is truly a polymicrobial process, with a wide array of infecting organisms. Antibiotic treatment must start with broad-spectrum coverage and be tailored to meet the individual microbial susceptibilities. In addition, fungal infections were present in 10%, and this must not be overlooked. With the evolution of newer endovascular devices, the configuration of the devices themselves is bound to change. Several pearls for successful explantation are elucidated here. Proximal control can be obtained in several ways. Traditional teaching states that supraceliac control of the aorta is safest if suprarenal fixation is present; otherwise, suprarenal control is adequate. Alternatively, an aortic occlusion balloon can be inflated proximally under fluoroscopic guidance and the device explanted with care taken not to rupture the balloon. There are technical challenges and techniques that require consideration when these devices must be explanted. The barbs of the suprarenal fixation can be technically challenging to remove because they serve as a robust anchoring system for the endograft. Several strategies are outlined in a report by Usatii et al.13 These strategies include graft infolding, transection of the graft without removal of the suprarenal fixation struts, and use of a Rumel tourniquet or Javid clamp to constrain the proximal end of the endograft to dislodge the struts from the aortic wall. In addition, we have had success with transecting the fabric and metal supports with a heavy scissors and removing the suprarenal fixation from the aortic wall with a heavy needle driver and a metal Andrews suction. Although it may be tempting to conclude that explantation of devices with suprarenal fixation may lead to a higher rate of postoperative acute renal failure, our study did not demonstrate this statistically. I-EVAR remains a rare problem, and although this is the largest series to date, a few unanswered questions remain. What is the optimal method of reconstruction? The gold standard treatment is extra-anatomic bypass with excision of all infected material and aortic ligation. This study has shown that in situ reconstruction with a variety of conduits is possible. Although it is tempting to draw definitive conclusions, none of the methods of reconstruction were significantly better secondary to the rare nature of I-EVAR. A general approach to I-EVAR may be to consider the intraoperative findings. If there is a frank abscess cavity with purulent and necrotic debris, extraanatomic bypass may be a safer choice. In the setting of perigraft fluid collection without frank purulence, in situ
reconstruction with omental wrapping may be an acceptable alternative. Unfortunately, statistical analysis of our data does not bear out a clear answer about the optimal reconstructive approach. Although this represents the largest series of I-EVARs undergoing explantation, it is a retrospective review that has significant limitations. There remains no standard protocol to follow when initiating treatment, and we cannot make any definitive statements regarding the “ideal” management. The patients were treated at four geographically disperse medical centers with different surgeons and support personnel performing the procedures. Decisions regarding diagnosis, timing of treatment, method of reconstruction, and postoperative care, were left to the discretion of the operating surgeons. This design precluded a control group, and there was no way to accrue patients in equal numbers to all of the methods of reconstruction. Despite the lack of standardized protocol and despite the variability in the treatment based on the institution, results were still acceptable compared with those reported in the literature.3,6,14 CONCLUSIONS The growing number of endovascular stent graft procedures performed each year will inevitably lead to a larger number of infectious complications. The management of IEVAR, especially those grafts with suprarenal fixation, poses significant challenges to vascular surgeons. Reviewing a shared experience such as we have documented may assist the practicing surgeon with these challenges. Prompt diagnosis of this pathologic process is key with a thorough understanding of the anatomical challenges. Morbidity and mortality can be minimized with proper planning and diagnosis by the surgeon. In short, we believe that this potentially complex complication of infected endovascular grafts can be treated with acceptable morbidity and mortality. AUTHOR CONTRIBUTIONS Conception and design: VD, AD, EW, WJ, NZ, WCS, SM Analysis and interpretation: VD, WS, AD, WJ, WCS, SM Data collection: VD, AD, EW, SM Writing the article: VD, WS, AD, WJ, WCS, SM Critical revision of the article: VD, WS, AD, WJ, WCS, SM Final approval of the article: VD, WS, AD, WJ, WCS, SM Statistical analysis: VD, AD, SM Obtained funding: Not applicable Overall responsibility: VD REFERENCES 1. Giles KA, Pomposelli F, Hamdan A, Wyers M, Jhaveri A, Schermerhorn ML. Decrease in total aneurysm-related deaths in the era of endovascular aneurysm repair. J Vasc Surg 2009;3:543-9. 2. Dua A, Kuy S, Lee CJ, Upchurch GR Jr, Desai SS. Epidemiology of aortic aneurysm repair in the United States from 2000 to 2010. J Vasc Surg 2014;59:1512-7. 3. Fatima J, Duncan AA, de Grandis E, Oderich GS, Kalra M, Gloviczki P, et al. Treatment strategies and outcomes in patients with infected aortic endografts. J Vasc Surg 2013;58:371-9.
JOURNAL OF VASCULAR SURGERY Volume 62, Number 4
4. Setacci C, de Donato G, Setacci F. Endografts for the treatment of aortic infection. Semin Vasc Surg 2011;4:242-9. 5. Cernohorsky P, Reijnen MM, Tielliu IF, van Sterkenburg SM, van den Dungen JJ, Zeebregts CJ. The relevance of aortic endograft prosthetic infection. J Vasc Surg 2011;54:327-33. 6. Lyons OT, Patel AS, Saha P, Clough RE, Price N, Taylor PR. A 14year experience with aortic endograft infection: management and results. Eur J Vasc Endovasc Surg 2013;46:306-13. 7. Laser A, Baker N, Rectenwald J, Eliason JL, Criado-Pallares E, Upchurch GR Jr. Graft infection after endovascular abdominal aortic aneurysm repair. J Vasc Surg 2011;54:58-63. 8. Lockhart PB, Brennan MT, Fox PC, Norton HJ, Jernigan DB, Strausbaugh LJ. Decision-making on the use of antimicrobial prophylaxis for dental procedures: a survey of infectious disease consultants and review. Clin Infect Dis 2002;34:1621-6. 9. Ducasse E, Calisti A, Speziale F, Rizzo L, Misuraca M, Fiorani P. Aortoiliac stent graft infection current problems and management. Ann Vasc Surg 2003;18:521-6. 10. Berger P, Moll FL. Aortic graft infections: is there still a role for axillobifemoral reconstruction? Semin Vasc Surg 2011;24:205-10.
Davila et al 883
11. Liedenbaum MH, Verdam FJ, Spelt D, de Groot HG, van der Waal J, van der Laan L. The outcome of the axillofemoral bypass: a retrospective analysis of 45 patients. World J Surg 2009;33: 2490-6. 12. Moulakakis KG, Sfyroeras GS, Mylonas SN, Mantas G, Papapetrou A, Antonopoulos CN, et al. Outcome after preservation of infected abdominal aortic endografts. J Endovasc Ther 2014;21:448-55. 13. Usatii A, Payne W, Santilli S. Removal of an infected aortic endograft and open aortic reconstruction: technical remarks. Ann Vasc Surg 2013;27:679-83. 14. Murphy EH, Szeto WY, Herdrich BJ, Jackson BM, Wang GJ, Bavaria JE, et al. The management of endograft infections following endovascular thoracic and abdominal aneurysm repair. J Vasc Surg 2013;58:1179-85.
Submitted Jan 22, 2015; accepted Apr 23, 2015.
The CME exam for this article can be accessed at http:// www.jvascsurg.org/cme/home.
DISCUSSION Dr Martin R. Back (Tampa, Fla). Dr Davila and colleagues have analyzed operative experiences from four tertiary centers managing infected abdominal aortic endografts over the last 15 years. In the 36-patient cohort, presentations with abdominal pain, leukocytosis, fever or bacteremia were common. Computed tomography imaging diagnosed 92% of cases, and intraoperative culture microbiology revealed prevalent gram-positive organisms, especially Staphylococcus species. All these features being similar to open aortic prosthetic graft infection series; however, in contradistinction to open aortic experiences, infected endografts can present more insidiously (>2 months’ symptom evolution before diagnosis) and occur earlier after implantation (<2 years vs 3 to 4 years for biofilm infection of open grafts). Complete endograft explantation was performed in all but two cases, with lower torso revascularization done by various in situ conduits in three-quarters of patients or axillobifemoral bypass in one quarter. Thirty-day morbidity and mortality, need for reoperation, and cumulative mortality to a mean of 2 years were appreciable in the series and certainly no better than reported outcomes for open aortic graft infections. These results just elucidate how serious this uncommon complication is after minimally invasive aortic endografting. I have 3 queries for the authors. Were any enteric erosions or developing bowel fistula observed/mentioned at operation during endograft explantation? Postoperative renal dysfunction was frequent. To facilitate complete endograft explantation, were suprarenal stents removed
and/or pararenal intimal disruption occur that could help explain renal dysfunction in addition to aortic clamp location and renal ischemic duration? Lastly, exactly what in situ conduit revascularization configurations were used (ie, aortofemoral, aortoiliac, or aortic tube)? Dr Victor J. Davila. Thank you very much for the questions. I think you bring up some good points. With respect to the first question, we do not always know the cause of endograft infections. I know there was at least one case of aortoenteric fistula that did seed the aneurysm sac with bacteria and led to the endograft infection. I think it is an important etiology to be aware of and sought out during the explantation procedure. With respect to the second question, suprarenal fixation was examined, and we did not find any statistically significant difference between the patients who developed renal failure and the presence of suprarenal fixation. However, acute renal failure is multifactorial, and we need to make sure to minimize suprarenal aortic clamp time as well as minimizing the aortic intimal disruption around the renal arteries. I can think of one case where a patient did develop acute renal failure in the postoperative period secondary to an intimal flap and was corrected with renal artery stent deployment on postoperative day 2. With regards to the third question, a variety of conduits were used in the in situ reconstructions, including aortofemoral, aortoiliac, and aortic tube grafts. The most commonly utilized reconstruction method was aortoiliac reconstruction.
A multicenter experience with the surgical treatment of infected abdominal aortic endografts Victor J. Davila, MD,a William Stone, MD,a Audra A. Duncan, MD,b Emily Wood, MD,b William D. Jordan Jr, MD,c Nicholas Zea, MD,c W. Charles Sternbergh III, MD,d and Samuel R. Money, MD, MBA,a Phoenix, Ariz; Rochester, Minn; Birmingham, Ala; and New Orleans, La Objective: Single-center experiences with the treatment of infected endografts after endovascular aortic repair (I-EVAR) have been reported. We performed a multicenter review of the surgical care of these patients to elucidate short-term and long-term outcomes. Methods: A retrospective analysis of all EVAR explants from 1997 to 2014 at four institutions was performed. Patients with I-EVAR undergoing surgical treatment were reviewed. Data were obtained detailing preoperative demographics, and postoperative morbidity and mortality. Results: Thirty-six patients (30 male) were treated with endovascular graft excision and revascularization for I-EVAR with a median age of 69 years (range, 54-80 years). Average time from the initial EVAR to presentation was 589 days (range, 432466 days). Preoperative comorbidities included hypertension, 32 (89%); tobacco use, 31(86%); coronary artery disease, 26 (72%); hyperlipidemia, 25 (69%), peripheral artery disease, 13 (36%); cerebrovascular disease, 10 (28%); diabetes, 10 (28%); chronic obstructive pulmonary disease, 9 (25%); and chronic kidney disease, 9 (25%). The most common presenting patient characteristics were leukocytosis, 23 (63%); pain, 21 (58%); and fever, 20 (56%), which were present an average of 65 days (range, 0-514 days) before explantation. Nine different types of endograft were removed. Three patients (8%) underwent emergency explantation. Thirty-four patients (89%) underwent total graft excision, and two patients (6%) underwent partial excision. Methods of reconstruction were in situ in 27 (75%) and extra-anatomic in nine (28%). Conduits used were Dacron (DuPont, Wilmington, Del), with or without rifampin, polytetrafluoroethylene, cryopreserved allograft, and femoral vein. Forty-nine organisms grew from operative cultures. Gram-positive organisms were the most common, found in 24 (67%), including Staphylococcus in 13 (36%) and Streptococcus in six (17%). Anaerobes were cultured in 6 patients (17%), gramnegative organisms in 6 (17%), and fungus in 5 (14%). Thirty-one patients (86%) received long-term antibiotics. Early complications included acute renal failure requiring dialysis, 12 (33%); respiratory failure, 3 (8%); bleeding, 4 (11%); and sepsis, 2 (6%). Six patients required re-exploration due to hematoma, infected hematoma, lymphatic leak, bowel perforation, open abdomen at initial operation, and anastomotic bleeding. Perioperative mortality was 8% (3 of 36), and long-term mortality was 25% (9 of 36) at a mean follow-up of 569 days (range, 0-3079 days). Type of reconstruction (in situ vs extra-anatomic) or conduit type did not affect perioperative or overall mortality. Conclusions: I-EVAR is a rare but potentially devastating clinical problem. Although perioperative mortality is acceptable, long-term mortality is high. The most common postoperative complication was acute renal failure requiring dialysis. Although this is the largest series of I-EVAR, further studies are needed to understand the risk factors and preventive measures. (J Vasc Surg 2015;62:877-83.)
From the Division of Vascular Surgery, Department of Surgery, Mayo Clinic Arizona, Phoenixa; the Division of Vascular Surgery, Department of Surgery, Mayo Clinic, Rochesterb; the Division of Vascular Surgery, Department of Surgery, University of Alabama at Birmingham, Birminghamc; and the Division of Vascular Surgery, Department of Surgery, Ochsner Clinic Foundation, New Orleans.d Author conflict of interest: W.D.J.: clinical investigator (paid to University of Alabama at Birmingham) for Medtronic, Gore, Cook, Endologix, Aptus, and Lombard; Cordis: consultant (paid to University of Alabama at Birmingham) for Medtronic, Gore, Endologix, Aptus, Colvano, and Lombard. S.R.M: consultant for Cook Medical and Gore Medical. Presented at the Thirty-ninth Annual Meeting of the Southern Association for Vascular Surgery, Scottsdale, Ariz, January 14-17, 2015. Correspondence: Samuel R. Money, MD, MBA, Mayo Clinic, Division of Vascular Surgery, 5777 E Mayo Blvd, Phoenix, AZ 85054 (e-mail: [email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest. 0741-5214 Copyright Ó 2015 by the Society for Vascular Surgery. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jvs.2015.04.440
More than 1 million Americans have abdominal aortic aneurysms (AAAs), many of which may require intervention. Open surgical intervention has slowly been replaced by endovascular AAA repair (EVAR) using aortic endografts.1 EVAR was initially introduced for high-risk individuals whose open operative risk was deemed excessive. Advancements in endograft development have brought a shift in the way that most AAAs are managed; currently, most AAA repairs performed in the United States are by an endovascular approach.2 Aortic endovascular graft infection (I-EVAR) is an unusual complication of endograft repair and has a reported incidence between 0.05% and 5%.3 The higher rates of infection reported may be the result of endograft deployment in infected fields, such as those in patients with mycotic aneurysms or aortoenteric fistulae.4 Mortality from endograft explantation secondary to infection has been reported as high as 30% and is similar to infection of open aortic grafts.5 To date, the treatment of I-EVAR 877
JOURNAL OF VASCULAR SURGERY October 2015
878 Davila et al
Table I. Survival summary Interval
Survival, % Time, months No. at risk 95% CI
1-2 months 2-10 months 10-21 months 21-28 months 28-101 months 101 months
95 90 84 74 55 0
1 2 10 21 28 101
21 20 14 8 4 1
87-100 79-100 69-100 53-100 29-100 NA
CI, Confidence interval.
has been reported mainly by single centers. This report details the results of a multicenter experience from four geographically distinct medical centers, which will assist delineating procedural techniques, clinical challenges, and results in this patient population. METHODS After approval from the Institutional Review Boards of the four medical centers involved in this study (Mayo Clinic, Rochester, Minn and Phoenix, Ariz; University of Alabama at Birmingham, Birmingham, Ala; and Ochsner Medical Institutions, New Orleans, LA), databases were queried to identify patients who underwent excision of infected aortic endografts from AAAs between 1997 and 2014. The Investigational Review Board reviewer approved waiver of the requirement to obtain informed consent. Patients who underwent explantation were included for review regardless of the institution where the endograft was initially placed. A retrospective review included pertinent history, presentation, physical findings, results of microbiologic testing, location of infection, management, and outcomes. Demographic data for each patient were collected and evaluated, including comorbidities. Time from the original EVAR implantation until the emergence of presenting signs or symptoms of infection and the type of endograft placed at original operation were also included for review. Patients were deemed to have I-EVAR based on presence of positive blood cultures, radiologic evidence of abscess or infection, intraoperative evidence of infection as described by the operating surgeon, or positive intraoperative cultures from explanted graft material, aortic wall, or aneurysm sac contents. Management was based on the surgeon’s judgment, including method of reconstruction, timing of the operation, antibiotic duration and type, method of explantation, including proximal control, and postoperative management. Operative data collected included indications for intervention, urgency of the procedure, and the method of reconstruction. Early complications were defined as those #30 days of removal of the infected endograft, and late complications were those that occurred afterward. Perioperative mortality was defined as death #30 days. Long-term mortality was defined as death >30 days. Demographic data describing the frequency of postoperative complications, explanted device type, and symptoms
Fig 1. Survival curve for infected endograft after endovascular aortic aneurysm (AAA) repair (I-EVAR).
at presentation are presented. Because of the small cohort size, most data are reported as mean and range. KaplanMeier survival curves were used to provide 10-month and 21-month survival estimates (Table I and Fig 1). These time frames are somewhat shifted from normal yearly intervals based on patient dropout and length of follow-up in this retrospective study. Survival was summarized after each death and reported for 1 and 2 years. All analysis was performed in R 3.1.2 software (The R Foundation for Statistical Computing, http://www.r-project.org/foundation/), and the package survival version 2.37-7 was used for survival analysis. RESULTS Thirty-six patients, including 30 men (83%), with a mean age of 69 years (range, 54-80 years) underwent explantation of an aortic endograft for infection and were included for review. The preoperative comorbidities, including hypertension, previous or current tobacco use, coronary artery disease, and others, appear to be those commonly seen in patients with aortic aneurysmal disease (Table I). Presenting symptoms varied, however; a majority (21 patients [58%]) presented with abdominal pain or fever (20 patients [56%]), or both (Table II). Blood cultures were positive in 10 of 34 patients (29%). Mean duration of these symptoms in these patients was 65 days before explantation (range, 0-514 days). Mean duration from initial endograft placement until development of symptoms was 589 days (range, 43-2466 days). Multiple diagnostic imaging modalities were used to establish the diagnosis. Computed tomography (CT) was used in all patients and confirmed an infected aortic endograft in 33 of 36 patients (91.7%; Fig 2). The CT in three patients did not demonstrate obvious perigraft infection. One
JOURNAL OF VASCULAR SURGERY Volume 62, Number 4
Davila et al 879
Table II. Presenting characteristics
Table III. Explanted endografts
Characteristic
No. (%)a
Type of graft
No. (%)
Pain Pain (alone) Pain and fever Pain and leukocytosis Pain and fever and leukocytosis Fever Fever (alone) Fever and pain Fever and leukocytosis Fever and pain and leukocytosis Fever and GI bleeding Leukocytosis Leukocytosis (alone) Leukocytosis and pain Leukocytosis and fever Leukocytosis and pain and fever Leukocytosis and GI bleeding
21 6 2 6 7 20 3 2 7 7 1 23 1 6 7 7 1
Excluder (Gore, Flagstaff, Ariz) AneuRx (Medtronic, Minneapolis, Minn) Zenith (Cook, Bloomington, Ind) Endologix (Endologix, Irvine, Calif) Endurant (Medtronic) Viabahn (Gore) Ancure (Guidant, Indianapolis, Ind) Talent (Medtronic) TAG (Gore)
13 9 6 2 2 1 1 1 1
(58) (17) (6) (17) (19) (56) (8) (6) (19) (19) (3) (63) (3) (17) (19) (19) (3)
GI, Gastrointestinal. a One patient was asymptomatic but had evidence of infection on computed tomography (CT) imaging.
Fig 2. Representative computed tomography (CT) image of an infected endograft after endovascular aortic aneurysm (AAA) repair (I-EVAR).
patient had leukocytosis with postprandial pain, with CT showing perigraft fluid but no secondary signs of infection (air or fat stranding), and was found to have an abscess pocket along the lateral side of the aneurysm sac. One patient presented with a psoas abscess after a spinal procedure, which was found intraoperatively to be in continuity with the endovascular prosthesis. The remaining patient was undergoing a workup as an outpatient for possible endograft infection when he presented with a signs and symptoms of a ruptured aneurysm, and subsequent intraoperative cultures were positive. Magnetic resonance imaging was performed in 12 patients (33%) and identified an infected endograft in nine of these patients (75%). Tagged white blood cell scans were positive in 21 of 28 patients (75%).
(36) (25) (16) (6) (6) (3) (3) (3) (3)
Nine different types of endograft were explanted for treatment of infection (Table III). Emergency explantation was performed in three patients (8%). One patient presented with a ruptured aneurysm, and two patients presented with acute onset of back pain and large aneurysms. Thirty-four patients (94%) underwent total graft excision, and two patients (6%) underwent partial excision. Partial endograft excision was used when only a portion of the endograft was determined to be infected. This judgment, which was made by the operating surgeon, primarily involved good incorporation of the remaining endograft into the surrounding aortic or iliac artery wall. Of the two patients who underwent partial excision, one patient was alive at the last follow-up of 811 days, and one patient died on postoperative day 44 from an anastomotic dehiscence after in situ femoral vein reconstruction. Suprarenal fixation was present in 24 of the explanted endografts (67%). Twenty-five percent (6 of 24) of these patients developed acute renal failure requiring dialysis in the postoperative period. Acute renal failure requiring dialysis developed in 56% of patients (6 of 12) with only infrarenal fixation. After endograft explantation, methods of reconstruction included in situ graft placement in 27 patients (75%) and/or extra-anatomic axillobifemoral bypass in nine (28%). In situ reconstructions consisted of rifampinsoaked Dacron (DuPont, Wilmington, Del) graft in 14 patients (39%), standard nonantibiotic-treated Dacron grafts in 2 (6%), polytetrafluoroethylene in 2 (6%), femoral vein in 4 (11%), cryopreserved graft (CryoLife Inc, Kennesaw, Ga) in 4 (11%), and a combination of femoral vein and cryopreserved graft in 1 patient (3%). Axillobifemoral bypass (n ¼ 9) was performed during the EVAR explantation (n ¼ 6) or in a staged fashion (n ¼ 3). The conduits used for extra-anatomic bypass included polytetrafluoroethylene in 6 patients (17%), rifampin-soaked Dacron in 2 (6%), and a combination of rifampin-soaked Dacron and femoral vein in 1 (3%). Intraoperative cultures were obtained from the explanted graft, aneurysm wall, or aneurysm sac contents in all patients and were positive in 28 patients (81%). Gram-positive cocci were the most common isolates and were identified in 24 patients (67%; Table IV). Patients were treated with broad-spectrum intravenous antibiotics and then transitioned to oral antibiotic therapy determined
JOURNAL OF VASCULAR SURGERY October 2015
880 Davila et al
Table IV. Microbiology of infected endografts
Variable Organisms Gram-positive Staphylococcus Anaerobes Gram-negative Streptococcus Fungus Patients with positive intraoperative cultures Organisms per patient, average (range), No.
Table V. Early complications
Organisms isolated Patients (n ¼ 49), No. (%) (n ¼ 36), No. (%) 32 13 11 8 6 5
(65) (27) (22) (16) (12) (10)
24 13 6 6 6 5 28/36
(67) (36) (17) (17) (17) (14) (80)
1.4 (0-5)
by culture sensitivities at the discretion of the treatment team. Lifelong antibiotic regimens were continued in 19 patients (53%), 10 (28%) were maintained on antibiotics for periods >5 weeks and up to 2 years, 2 (6%) received <6 weeks of antibiotic therapy, and 5 (14%) did not have adequate documentation to determine the duration of antibiotic therapy. Three patients (8%) died #30 days after explantation in this series. Early deaths occurred in two patients in the in situ reconstruction group and in one in the extraanatomic reconstruction group. Acute renal failure requiring dialysis was the most common early complication. Other complications included respiratory failure requiring tracheostomy, bleeding, and sepsis (Table V). Six patients (17%) required re-exploration after explantation for indications that included hematoma, infected hematoma, lymph leak, and small-bowel perforation in one patient each. An additional patient developed an aortic anastomotic dehiscence after in situ reconstruction with femoral vein resulting in hemorrhage and was classified as an early death. One other patient had a planned reoperation for secondary closure of the abdominal wall. Mortality for patients undergoing reoperation was significant (n ¼ 3 [50%]). Two patients needing reoperation were early deaths: one had sepsis after aortoiliac reconstruction with a Dacron graft and omental flap, and one had anastomotic bleeding. The third early death was secondary to aspiration and subsequent respiratory failure. Overall mortality was 25% (9 patients) at a mean follow-up of 570 days (range, 0-3079 days). Late deaths included one death from proximal aneurysm rupture, one death from a chronic inflammatory illness, and the cause of death was not determined in four patients. There were four late deaths in the in situ reconstruction group and two late deaths in the extra-anatomic reconstruction group. Although one form of reconstruction did not appear to be superior in this series, the sample size was too small for any meaningful statistical comparison. Of the 36 patients studied, there were three early deaths. These deaths were not included in the survival analysis for those who successfully endured the revision. Of the
Early complications
No. (%)
Acute renal failure Bleedinga Respiratory failure requiring tracheostomy Sepsis Need for re-exploration in operating room Hematoma Infected hematoma Lymphatic leak Small-bowel perforation Open abdomen Anastomotic bleeding
12 4 3 2 6 1 1 1 1 1 1
(33) (11) (8) (6) (17) (3) (3) (3) (3) (3) (3)
a One patient each developed hemothorax, a retroperitoneal hematoma, gastrointestinal bleeding managed endoscopically, and an anastomotic dehiscence of the aortic reconstruction requiring reexploration.
remaining 33 patients, follow-up data existed for 26 patients. Six of those patients died during the follow-up period, for a total death rate of 25% (9 of 36). KaplanMeier analysis was performed only on those patients for whom follow-up data were available, making the death rate during the observation period 23% (6 of 26). Survival was 84% (95% confidence interval, 69%-100%; 14 of 26 at risk) at 1 year and 74% (95% confidence interval, 53%100%; 8 of 26 at risk) at 2 years. An accurate estimate of median survival time cannot be provided from these data due to censoring, but appears to be in excess of 30 months. Survival is summarized in Table I. DISCUSSION Patients who develop I-EVARs carry a significant risk, but our multi-institutional series demonstrates that they can be treated by excision and revascularization with a reasonably low mortality (8%). Long-term survival in these patients was 75% with a mean follow-up of >18 months. Only two patients died of complications attributable to their aortic disease. One early death was caused by an anastomotic dehiscence, and one late death from proximal aortic aneurysm rupture. The results of this multiinstitutional review compare favorably with the results other studies.6,7 However, despite what may be considered an acceptable 30-day mortality, the risk of postoperative morbidity is significant (n ¼ 21 [58%]), with acute renal failure requiring dialysis occurring as the most common event in one-third of the patients. In addition, six patients (17%) required re-exploration due to a variety of causes, including small-bowel perforation, anastomotic bleeding, lymphatic leak, and infected hematoma. Mortality for patients undergoing re-exploration was significant (50%). Median survival cannot be estimated in this study due to the large amount of dropout. In general, this is a heterogeneous subgroup, and interpretation of Kaplan-Meier results from a small, heterogeneous population should be done with care. Infection remains a rare complication after endovascular aortic graft placement (0.5%-5%).5 The causes of this complication can be clearly delineated in some cases. The
JOURNAL OF VASCULAR SURGERY Volume 62, Number 4
etiology of early endograft infection could be related to endograft deployment through an infected field or in an area known to have aortic infection. Endografts have been used as a temporizing measure to stabilize patients with aortoenteric fistula.4 In this clinical scenario, endografts without suprarenal fixation should be used to make future explantation technically easier. Other reasons for infection to develop in aortic endografts may be multifactorial. Patients with systemic bacteremia from ongoing infections, such as urinary tract infections, pneumonia, or even from dental surgery, may cause secondary infection of an aortic endograft. In a survey by Lockhart et al8 in 2002, 35% of infectious disease physicians would “always” or “usually” recommend antibiotic prophylaxis for dental procedures in patients who have a vascular graft prosthesis, and 30% would “never” recommend prophylaxis before a dental procedure. The potential for secondary infection of an aortic endograft could follow additional procedures other than dental. Percutaneous coiling or placement of sealant or glue into an aneurysm sac could potentially introduce bacteria and lead to secondary infection. As with the placement of any prosthetic device, the endovascular graft has the potential to become infected during initial placement. Theoretically, EVAR minimizes contamination of the prosthetic device compared with standard open surgery because the device is housed in a sterile delivery system. However, Ducasse et al9 noted that endografts placed in interventional radiology suites have a higher rate of infection compared with traditional operating rooms, possibly because of different levels of sterility when equipment is handled. Mycotic aneurysms represent another situation where an endograft may be deployed in an infected field. In our retrospective review, a few patients may have had mycotic aneurysms at the original placement of the aortic endograft; for example, one patient developed vertebral body osteomyelitis likely secondary to a mycotic aneurysm. Although the initial imaging showed no evidence of this, imaging showed erosion of the vertebral body and subsequent spread of infection into the aneurysm sac. Multiple potential causes of I-EVAR are possible; however, the actual cause is not clearly discernible in many cases. It is important to closely evaluate the surrounding aortic wall and structures for abnormalities before the initial endograft is placed to avoid deployment in a potentially infected field because of the implications of resultant endograft infection. The presentation of I-EVAR is often insidious. Many of the patients in our series presented with pain (58%) and fever (56%), which are not specific to I-EVAR. Leukocytosis was present only in approximately two-thirds of the patients. The best diagnostic study was a thin-slice CT scan, which was performed in all patients and diagnostic in 91.7%. Tagged white blood cell scanning was positive in 73% of patients when ordered and may be of additional benefit when clear-cut evidence is not present on CT scanning. Most of the patients in this series had a very subtle
Davila et al 881
presentation, which can be beneficial because it allows time to establish a diagnosis, optimize comorbidities, and formulate a defined treatment plan tailored to the patient. However, the nebulous presentation can also be detrimental, because a delay in diagnosis can result in aneurysm rupture and exsanguination. This subtle presentation may explain why the patients in this study presented with symptoms for an average of 65 days (range 0-514 days) before graft explantation. Various options for reconstruction exist after explantation. Classically, extra-anatomic reconstruction in the form of an axillobifemoral bypass, followed by endograft explantation and aortic ligation, was the standard approach for most infected aortic prostheses. In situ reconstructions allow for prevention of aortic stump blowout and the complications associated with possibly tenuous extra-anatomic bypass, including a reduced patency rate.10,11 Cryopreserved grafts or rifampin-soaked Dacron can be used for in situ reconstruction. Fatima et al3 demonstrated superior results with full coverage with omentum in those patients undergoing in situ replacement with rifampin-soaked Dacron grafts. They describe the technique of mobilizing the omentum, raising a tongue of omentum in the retrocolic configuration, and performing a 360 omental wrap around the Dacron graft. Microorganisms were isolated from operative cultures in 28 of 36 patients (77%) and were predominantly gram-positive organisms. Anaerobes and gram-negative organisms were also present but not as commonly identified. Cultures were positive for fungus in five patients (14%). Most of these patients demonstrated a polymicrobial process. Cultures of eight patients (22%) were negative. These patients may have had preoperative antibiotic therapy precluding growth from samples obtained in the operating room. All patients with negative cultures were diagnosed with I-EVAR through extensive preoperative workup. A spectrum of findings leads to the diagnosis of I-EVAR in these cases, including air in the aneurysm sac, perigraft fluid, and retroperitoneal abscess or fat stranding. One patient, who was not diagnosed by CT scan and had no evidence of microbial growth, had a large psoas abscess. Exploration of the abscess showed it was in continuum with the aorta and permeated through the native aortic wall in contact with the endograft. An in situ repair with femoral vein was performed in this patient. This study is in accordance with others that place emphasis on CT scanning for definitive diagnosis.3 Numerous small studies have examined the role of nonoperative treatment of I-EVAR. Moulakakis et al12 published a meta-analysis of 17 reviews including 29 patients with I-EVAR who were treated with preservation of the stent graft. The in-hospital mortality was 21%, and seven more patients died during a mean follow-up of 11.4 months, for an overall mortality of 45%.12 This reported mortality appears greater than our 30-day mortality of 8% and long-term mortality of 25%. We believe that nonexcisional treatment of I-EVAR should be limited to patients with a prohibitive operative risk.
JOURNAL OF VASCULAR SURGERY October 2015
882 Davila et al
There are no set standards for the duration of antibiotic treatment in infected aortic grafts or in infected endovascular aortic grafts. Duration of antibiotic treatment in these patients postoperatively was left to the treating surgeon and his or her team. Most of these patients received intravenous antibiotics for 4 to 6 weeks, with most transitioned to a lifetime of oral antibiotic therapy. This regimen is based purely on physician preference, without appropriate evidence-based studies in the literature. The lack of scientific rigor to guide our management after resection of the infected endograft requires attention. Treating surgeons should understand that this is truly a polymicrobial process, with a wide array of infecting organisms. Antibiotic treatment must start with broad-spectrum coverage and be tailored to meet the individual microbial susceptibilities. In addition, fungal infections were present in 10%, and this must not be overlooked. With the evolution of newer endovascular devices, the configuration of the devices themselves is bound to change. Several pearls for successful explantation are elucidated here. Proximal control can be obtained in several ways. Traditional teaching states that supraceliac control of the aorta is safest if suprarenal fixation is present; otherwise, suprarenal control is adequate. Alternatively, an aortic occlusion balloon can be inflated proximally under fluoroscopic guidance and the device explanted with care taken not to rupture the balloon. There are technical challenges and techniques that require consideration when these devices must be explanted. The barbs of the suprarenal fixation can be technically challenging to remove because they serve as a robust anchoring system for the endograft. Several strategies are outlined in a report by Usatii et al.13 These strategies include graft infolding, transection of the graft without removal of the suprarenal fixation struts, and use of a Rumel tourniquet or Javid clamp to constrain the proximal end of the endograft to dislodge the struts from the aortic wall. In addition, we have had success with transecting the fabric and metal supports with a heavy scissors and removing the suprarenal fixation from the aortic wall with a heavy needle driver and a metal Andrews suction. Although it may be tempting to conclude that explantation of devices with suprarenal fixation may lead to a higher rate of postoperative acute renal failure, our study did not demonstrate this statistically. I-EVAR remains a rare problem, and although this is the largest series to date, a few unanswered questions remain. What is the optimal method of reconstruction? The gold standard treatment is extra-anatomic bypass with excision of all infected material and aortic ligation. This study has shown that in situ reconstruction with a variety of conduits is possible. Although it is tempting to draw definitive conclusions, none of the methods of reconstruction were significantly better secondary to the rare nature of I-EVAR. A general approach to I-EVAR may be to consider the intraoperative findings. If there is a frank abscess cavity with purulent and necrotic debris, extraanatomic bypass may be a safer choice. In the setting of perigraft fluid collection without frank purulence, in situ
reconstruction with omental wrapping may be an acceptable alternative. Unfortunately, statistical analysis of our data does not bear out a clear answer about the optimal reconstructive approach. Although this represents the largest series of I-EVARs undergoing explantation, it is a retrospective review that has significant limitations. There remains no standard protocol to follow when initiating treatment, and we cannot make any definitive statements regarding the “ideal” management. The patients were treated at four geographically disperse medical centers with different surgeons and support personnel performing the procedures. Decisions regarding diagnosis, timing of treatment, method of reconstruction, and postoperative care, were left to the discretion of the operating surgeons. This design precluded a control group, and there was no way to accrue patients in equal numbers to all of the methods of reconstruction. Despite the lack of standardized protocol and despite the variability in the treatment based on the institution, results were still acceptable compared with those reported in the literature.3,6,14 CONCLUSIONS The growing number of endovascular stent graft procedures performed each year will inevitably lead to a larger number of infectious complications. The management of IEVAR, especially those grafts with suprarenal fixation, poses significant challenges to vascular surgeons. Reviewing a shared experience such as we have documented may assist the practicing surgeon with these challenges. Prompt diagnosis of this pathologic process is key with a thorough understanding of the anatomical challenges. Morbidity and mortality can be minimized with proper planning and diagnosis by the surgeon. In short, we believe that this potentially complex complication of infected endovascular grafts can be treated with acceptable morbidity and mortality. AUTHOR CONTRIBUTIONS Conception and design: VD, AD, EW, WJ, NZ, WCS, SM Analysis and interpretation: VD, WS, AD, WJ, WCS, SM Data collection: VD, AD, EW, SM Writing the article: VD, WS, AD, WJ, WCS, SM Critical revision of the article: VD, WS, AD, WJ, WCS, SM Final approval of the article: VD, WS, AD, WJ, WCS, SM Statistical analysis: VD, AD, SM Obtained funding: Not applicable Overall responsibility: VD REFERENCES 1. Giles KA, Pomposelli F, Hamdan A, Wyers M, Jhaveri A, Schermerhorn ML. Decrease in total aneurysm-related deaths in the era of endovascular aneurysm repair. J Vasc Surg 2009;3:543-9. 2. Dua A, Kuy S, Lee CJ, Upchurch GR Jr, Desai SS. Epidemiology of aortic aneurysm repair in the United States from 2000 to 2010. J Vasc Surg 2014;59:1512-7. 3. Fatima J, Duncan AA, de Grandis E, Oderich GS, Kalra M, Gloviczki P, et al. Treatment strategies and outcomes in patients with infected aortic endografts. J Vasc Surg 2013;58:371-9.
JOURNAL OF VASCULAR SURGERY Volume 62, Number 4
4. Setacci C, de Donato G, Setacci F. Endografts for the treatment of aortic infection. Semin Vasc Surg 2011;4:242-9. 5. Cernohorsky P, Reijnen MM, Tielliu IF, van Sterkenburg SM, van den Dungen JJ, Zeebregts CJ. The relevance of aortic endograft prosthetic infection. J Vasc Surg 2011;54:327-33. 6. Lyons OT, Patel AS, Saha P, Clough RE, Price N, Taylor PR. A 14year experience with aortic endograft infection: management and results. Eur J Vasc Endovasc Surg 2013;46:306-13. 7. Laser A, Baker N, Rectenwald J, Eliason JL, Criado-Pallares E, Upchurch GR Jr. Graft infection after endovascular abdominal aortic aneurysm repair. J Vasc Surg 2011;54:58-63. 8. Lockhart PB, Brennan MT, Fox PC, Norton HJ, Jernigan DB, Strausbaugh LJ. Decision-making on the use of antimicrobial prophylaxis for dental procedures: a survey of infectious disease consultants and review. Clin Infect Dis 2002;34:1621-6. 9. Ducasse E, Calisti A, Speziale F, Rizzo L, Misuraca M, Fiorani P. Aortoiliac stent graft infection current problems and management. Ann Vasc Surg 2003;18:521-6. 10. Berger P, Moll FL. Aortic graft infections: is there still a role for axillobifemoral reconstruction? Semin Vasc Surg 2011;24:205-10.
Davila et al 883
11. Liedenbaum MH, Verdam FJ, Spelt D, de Groot HG, van der Waal J, van der Laan L. The outcome of the axillofemoral bypass: a retrospective analysis of 45 patients. World J Surg 2009;33: 2490-6. 12. Moulakakis KG, Sfyroeras GS, Mylonas SN, Mantas G, Papapetrou A, Antonopoulos CN, et al. Outcome after preservation of infected abdominal aortic endografts. J Endovasc Ther 2014;21:448-55. 13. Usatii A, Payne W, Santilli S. Removal of an infected aortic endograft and open aortic reconstruction: technical remarks. Ann Vasc Surg 2013;27:679-83. 14. Murphy EH, Szeto WY, Herdrich BJ, Jackson BM, Wang GJ, Bavaria JE, et al. The management of endograft infections following endovascular thoracic and abdominal aneurysm repair. J Vasc Surg 2013;58:1179-85.
Submitted Jan 22, 2015; accepted Apr 23, 2015.
The CME exam for this article can be accessed at http:// www.jvascsurg.org/cme/home.
DISCUSSION Dr Martin R. Back (Tampa, Fla). Dr Davila and colleagues have analyzed operative experiences from four tertiary centers managing infected abdominal aortic endografts over the last 15 years. In the 36-patient cohort, presentations with abdominal pain, leukocytosis, fever or bacteremia were common. Computed tomography imaging diagnosed 92% of cases, and intraoperative culture microbiology revealed prevalent gram-positive organisms, especially Staphylococcus species. All these features being similar to open aortic prosthetic graft infection series; however, in contradistinction to open aortic experiences, infected endografts can present more insidiously (>2 months’ symptom evolution before diagnosis) and occur earlier after implantation (<2 years vs 3 to 4 years for biofilm infection of open grafts). Complete endograft explantation was performed in all but two cases, with lower torso revascularization done by various in situ conduits in three-quarters of patients or axillobifemoral bypass in one quarter. Thirty-day morbidity and mortality, need for reoperation, and cumulative mortality to a mean of 2 years were appreciable in the series and certainly no better than reported outcomes for open aortic graft infections. These results just elucidate how serious this uncommon complication is after minimally invasive aortic endografting. I have 3 queries for the authors. Were any enteric erosions or developing bowel fistula observed/mentioned at operation during endograft explantation? Postoperative renal dysfunction was frequent. To facilitate complete endograft explantation, were suprarenal stents removed
and/or pararenal intimal disruption occur that could help explain renal dysfunction in addition to aortic clamp location and renal ischemic duration? Lastly, exactly what in situ conduit revascularization configurations were used (ie, aortofemoral, aortoiliac, or aortic tube)? Dr Victor J. Davila. Thank you very much for the questions. I think you bring up some good points. With respect to the first question, we do not always know the cause of endograft infections. I know there was at least one case of aortoenteric fistula that did seed the aneurysm sac with bacteria and led to the endograft infection. I think it is an important etiology to be aware of and sought out during the explantation procedure. With respect to the second question, suprarenal fixation was examined, and we did not find any statistically significant difference between the patients who developed renal failure and the presence of suprarenal fixation. However, acute renal failure is multifactorial, and we need to make sure to minimize suprarenal aortic clamp time as well as minimizing the aortic intimal disruption around the renal arteries. I can think of one case where a patient did develop acute renal failure in the postoperative period secondary to an intimal flap and was corrected with renal artery stent deployment on postoperative day 2. With regards to the third question, a variety of conduits were used in the in situ reconstructions, including aortofemoral, aortoiliac, and aortic tube grafts. The most commonly utilized reconstruction method was aortoiliac reconstruction.