- Email: [email protected]
A multicenter experience with infected abdominal aortic endograft explantation
A multicenter experience with infected abdominal aortic endograft explantation Xavier Chaufour, MD, PhD,a Julien Gaudric, MD,b Yann Goueffic, MD, PhD,c Réda Hassen Khodja, MD,d Patrick Feugier, MD, PhD,e Sergei Malikov, MD,f Guillaume Beraud, MD, PhD,g and Jean-Baptiste Ricco, MD, PhD,h for the AURC (French University Surgeons Association) collaborators,* Toulouse, Paris, Nantes, Nice, Lyon, Nancy, and Poitiers, France
ABSTRACT Objective: Endovascular aneurysm repair (EVAR) is widely used with excellent results, but its infectious complications can be devastating. In this paper, we report a multicenter experience with infected EVAR, symptoms, and options for explantation and their outcome. Methods: We have reviewed all consecutive endograft explants for infection at 11 French university centers following EVAR, defined as index EVAR, from 1998 to 2015. Diagnosis of infected aortic endograft was made on the basis of clinical findings, cultures, imaging studies, and intraoperative findings. Results: Thirty-three patients with an infected aortic endograft were identified. In this group, at index EVAR, six patients (18%) presented with a groin or psoas infection and six patients (18%) presented with a general infection, including catheter-related infection (n ¼ 3), prostatitis (n ¼ 1), cholecystitis (n ¼ 1), and pneumonia (n ¼ 1). After index EVAR, eight patients underwent successful inferior mesenteric artery embolization for a type II endoleak within 6 months of index EVAR and one patient received an additional stent for a type Ib endoleak 1 week after index EVAR. Median time between the first clinical signs of infection and endograft explantation was 30 days (range, 1 day to 2.2 years). The most common presenting characteristics were pain and fever in 21 patients (64%) and fever alone in 8 patients (24%). Suprarenal fixation was present in 20 of 33 endografts (60%). All patients underwent endograft explantation, with bowel resection in 12 patients (36%) presenting with an endograft-enteric fistula. Methods of reconstruction were graft placement in situ in 30 patients and extra-anatomic bypass in 3 patients. In situ conduits were aortic cryopreserved allografts in 23, polyester silver graft in 5, and autogenous femoral vein in 2. Microbiology specimens obtained from the endograft and the aneurysm were positive in 24 patients (74%). Gram-positive organisms were the most commonly found in 18 patients (55%). Early mortality (30 days or in the hospital) was 39% (n ¼ 13) in relation to graft blowout (n ¼ 3), multiple organ failure (n ¼ 6), colon necrosis (n ¼ 3), and peripheral embolism (n ¼ 1). At 1 year, the rates of patient survival, graft-related complications, and reinfection were 44%, 10%, and 5%, respectively. Conclusions: Abdominal aortic endograft explantation for infection is high risk and associated with graft-enteric fistula in one-third of the cases. Larger multicenter studies are needed to better understand the risk factors and to improve preventive measures at index EVAR and during follow-up. (J Vasc Surg 2016;-:1-9.)
From the Department of Vascular Surgery, University of Toulouse, Toulousea; the Department of Vascular Surgery, Hôpital Pitié-Salpétrière, Parisb; the Department of Vascular Surgery, University of Nantes, Nantesc; the Department of Vascular Surgery, University of Nice, Niced; the Department of Vascular Surgery, University of Lyon, Lyone; the Department of Vascular Surgery, University of Nancy, Nancyf; and the Department of Infectious Diseasesg and Department of Vascular Surgery,h University of Poitiers, Poitiers.
*The names of the AURC collaborators participating in the study can be found in the Appendix at the end of this article. Author conflict of interest: none. Correspondence: Jean-Baptiste Ricco, MD, PhD, Division of Vascular Surgery, Department of Cardiothoracic and Vascular Surgery, University of Poitiers, Medical School, 6, rue de la Milétrie, 86073 Poitiers, France (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 2016 by the Society for Vascular Surgery. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jvs.2016.07.126
Extensive use of endovascular aneurysm repair (EVAR) following prospective multicenter trials as a primary treatment modality for abdominal aortic aneurysm1,2 is associated with short series of endograft infection, a rare complication with an incidence between 0.4% and 3%3-10 but with a postoperative mortality as high as 30%, comparable to infection of open aortic grafts.5,8 One large multi-institutional study recently published in the Journal of Vascular Surgery10 suggested that in these cases, complete removal of the infected endograft with débridement of infected tissue and in situ or extraanatomic replacement may be the option most likely to eradicate the infectious process despite a high early postoperative mortality. Furthermore, because of the low incidence of endograft infection, diagnosis may be delayed, and optimal strategy of care has yet to be defined. The aim of this study was to assess the outcomes of aortic endograft infection following EVAR 1
2
Journal of Vascular Surgery
Chaufour et al
---
and to report the technical challenges of aortic endograft explantation in a consecutive series of patients treated in 11 French tertiary vascular centers that are members of the Association Universitaire de Recherche en Chirurgie (AURC, French University Surgeons Association). The AURC is a nonprofit academic association for clinical research in vascular surgery founded 30 years ago. France has 32 university hospitals, and in each of them there is a member of the AURC heading the department of vascular surgery. In our study, 11 university hospitals participated. Results were analyzed by the leading author and a clinical research assistant and were reviewed by all members of the AURC participating in the study.
METHODS After approval by the Institutional Review Boards of the university medical center members of the AURC participating in the study, institutional databases were analyzed to identify all consecutive patients who had undergone explantation of an infected abdominal aortic endograft between January 1998 and January 2015. Patient consent was waived by all Institutional Review Boards because of the retrospective nature of the study. Records were reviewed at these institutions regardless of where EVAR, defined as index EVAR, was performed. Patients with an endograft implanted primarily in an infected field to treat an aortoenteric fistula or an infected aneurysm were excluded from the study. For each patient, demographic and index EVAR details with subsequent follow-up were examined. Clinical indicators of endograft infection were analyzed along with results of cultures, computed tomography, and white blood cell scan. Operative details including strategy of intervention and material used for reconstruction were reviewed. When endograft infection was suspected, an institutional algorithm including computed tomography angiography (CTA) with injection of contrast material and blood cultures before initiation of antibiotics was applied. In case of doubt, a tagged white blood cell scan was performed to confirm the diagnosis of endograft infection. Whereas complete removal of the endograft was recommended by the AURC group, choice of revascularization was left to the vascular surgeon, depending on the presence of gross contamination or enteric fistula, bleeding, or other emergent situations. Postoperative outcomes examining morbidity and mortality were evaluated on electronic medical records and reviewed on site for missing details. The postoperative period was defined as duration of hospitalization regardless of the number of days or within 30 days of the explantation. Data are reported as median with range using nonparametric tests. Kaplan-Meier plot was used to illustrate 1-year survival. All data were entered in a password-encrypted database, and analyses were
2016
Table I. Clinical presentation before initial endovascular aneurysm repair (EVAR) and events following EVAR (n ¼ 33) Patient characteristics
Median (range) or No. (%)
Age, years, median (range)
69 (57-87)
Male gender
33 (100)
Current smoking
20 (60)
Hypertension
19 (57)
Heart failure
18 (54)
Chronic lung disease
12 (36)
Hostile abdomen Previous aortic surgery
8 (24) 3 (9)
ASA class 3 and 4
25 (75)
Patients with early post-EVAR infection
12 (36)
Groin or psoas abscess following index EVARa
6 (18)
Early general post-EVAR infectionb
6 (18)
Post-EVAR endovascular procedures
10 (30)
Embolization for type II endoleak
8 (24)
Embolization of both hypogastric arteriesc
1 (3)
Additional stent for type Ib endoleak
1 (3)
ASA, American Society of Anesthesiologists. a Including one groin abscess, two false femoral artery aneurysms, one infected femoral crossover bypass after index EVAR in a patient with an aortomonoiliac endograft, and two patients with a psoas abscess following coil embolization for a type II endoleak (n ¼ 1) and for a hypogastric aneurysm (n ¼ 1). b Infection with bacteremia following central venous catheter infection (n ¼ 3), bacterial prostatitis (n ¼ 1), cholecystitis (n ¼ 1), and pneumonia (n ¼ 1). c Embolization of one hypogastric artery before EVAR, followed by contralateral hypogastric artery embolization during EVAR.
performed using SPSS software version 23 (IBM Corp, Armonk, NY).
RESULTS Patient demographics, index EVAR, and possible sources of infection. Between January 1998 and January 2015, 33 patients with a median age of 69 years (range, 57-87 years) were treated for an aortic abdominal endograft infection in 11 French university centers. During the study period, a total of 6057 EVARs were carried out in the 11 tertiary centers. Of the 33 infected endografts reported in our series, 18 index EVARs were initially performed in the 11 tertiary centers with a ratio of endograft infection of 0.3% (95% confidence interval, 0.2%-0.4%). Demographics and clinical presentation at index EVAR are presented in Table I, and events following index EVAR are presented in Table II. Indications for EVAR were an aortic aneurysmal disease in 32 patients and an aortic pseudoaneurysm following prior open aortic repair in 1 patient. The endograft was deployed in abdominal aortic aneurysms with a median diameter of 60 mm (range, 50-93 mm) and a median diameter of the aortic neck of 23 mm (18-31 mm). The median length of the aortic neck was 20 mm (10-38 mm). Aortic
Journal of Vascular Surgery Volume
-,
Number
Chaufour et al
3
-
Table II. Index endovascular aneurysm repair (EVAR), events following index EVAR, presenting symptoms of infection, and imaging Age, No. years
Index EVAR
Events following index EVAR
Time from EVAR to symptoms, days
Presenting symptoms
Positive imaging
Time from symptoms to explantation, days
730
Fever, abdominal pain
CTA, psoas abscess
388
1
69
Elective EVAR, ABI
2
73
Elective EVAR, AMI
Groin abscess, drainage
847
Fever, abdominal pain
CTA, psoas abscess
7
3
66
Elective EVAR, ABI
Postoperative fever
453
Fever
CTA, fistula
12
4
57
Urgent EVAR, ABI
Postoperative fever
1638
Fever, abdominal pain
CTA, fistula
820
5
67
Elective EVAR, ABI
Central catheter infected, septicemia
104
Fever, abdominal pain
CTA
209
6
70
Elective EVAR, ABI
Pneumonia, septicemia
19
Fever, abdominal pain
CTA
34
7
77
Elective EVAR, AMI
Crossover femoral bypass infectedd removal
63
Fever, abdominal pain
CTA
280
8
67
Elective EVAR, ABI
Central catheter infected, septicemia
26
Fever, abdominal pain
CTA, psoas abscess
31
9
71
Elective EVAR, ABI
Type II endoleak, IMA embolization
351
Fever
CTA
5
10
62
Elective EVAR, AMI
3468
Fever, abdominal pain
CTA, psoas abscess
67
11
70
Elective EVAR, ABI
2291
Abdominal pain
CTA, fistula
40
12
69
Urgent EVAR, ABI
97
Fever, abdominal pain, hematemesis
CTA, fistula
9
13
75
Urgent EVAR, AMI
674
Fever, abdominal pain, hematemesis
CTA, fistula
55
14
69
Elective EVAR, ABI
82
Fever, abdominal pain
CTA, psoas abscess
166
15
78
Elective EVAR, ABI
644
Abdominal pain
CTA
3
16
67
Elective EVAR, ABI
1190
Fever
CTA
55
17
59
Elective EVAR, AMI
182
Fever, abdominal pain
CTA
89
18
70
Elective EVAR, ABI
670
Fever
CTA
26
19
78
Elective EVAR AMI
1133
Fever, abdominal pain
CTA, fistula
9
20
64
Elective EVAR, AMI
Fever
CTA, psoas abscess
12
21
59
Elective EVAR, ABI
850
Fever, abdominal pain
CTA, psoas abscess
18
22
81
Elective EVAR, ABI
277
Fever, abdominal pain
CTA, fistula
35
23
63
Elective EVAR, ABI
260
Fever, abdominal pain
CTA, fistula
45
24
71
Elective EVAR, AMI
414
Fever, abdominal pain
CTA, fistula
20
25
84
Elective EVAR, AMI
1244
Fever, abdominal pain, hematemesis
CTA, fistula
1
Prostatitis
Type II endoleak, IMA embolization
Groin abscess, drainage
Cholecystitis
6
(Continued on next page)
4
Journal of Vascular Surgery
Chaufour et al
---
2016
Table II. Continued. Age, No. years
Index EVAR
Events following index EVAR Groin abscess, drainage
Time from EVAR to symptoms, days
Presenting symptoms
Positive imaging
Time from symptoms to explantation, days
145
Fever, abdominal pain
CTA
67
180
Fever
CTA, fistula
36
1746
Abdominal pain
CTA, fistula
8
22
Fever, abdominal pain, septicemia
CTA
34
8
Fever, abdominal pain
CTA
30
26
63
Elective EVAR, AMI
27
76
Elective EVAR, ABI
28
68
Elective EVAR, ABI
29
64
Elective EVAR, ABI
30
87
Elective EVAR, ABI
31
74
Elective EVAR, ABI
368
Fever
CTA
45
32
67
Elective EVAR, ABI
443
Abdominal pain
CTA, psoas abscess
45
33
74
Elective EVAR, ABI
1713
Fever
CTA
4
Type II endoleak, IMA embolization
Central catheter infected, septicemia
Type Ib endoleak, stenting
ABI, Aortobi-iliac endograft; AMI, aortomonoiliac endograft with a crossover femorofemoral graft; CTA, computed tomography angiography with contrast media showing signs of infection; IMA, inferior mesenteric artery.
neck angulation was <60 degrees in all patients of this series, and no stent migration was seen on CTA before explantation. There were 27 endografts (81.8%) implanted in the operating room and six endografts (18.2%) implanted in the radiology suite. All index EVARs were performed using a femoral cutdown. Prophylactic intravenous antibiotics were administered in all patients according to local protocols. Infected endografts were Cook Zenith (Cook Inc, Bloomington, Ind) in 14 cases (43%), Medtronic Talent (Medtronic, Minneapolis, Minn) in 7 cases (22%), Gore Excluder (W. L. Gore & Associates, Flagstaff, Ariz) in 6 cases (18%), Anaconda (Vascutek, Terumo, Inchinnan, United Kingdom) in 4 cases (12%), and home-made endografts in 2 cases (5%). During the same period, national statistics of the different brands of endografts used in France showed that the Cook Zenith endograft was the most widely used, with 42% of the national market and 42% among the AURC members participating in the study with 43% of Cook Zenith endografts infected. Therefore, the high ratio of infection of the Zenith endograft is related to its use. In this group, after index EVAR (Table I), four patients (12%) presented with a groin infection and two patients developed a psoas abscess, one after coil embolization of an internal iliac artery aneurysm and one after coil embolization for a type II endoleak. Six patients (18%) presented with a general infection with bacteremia associated with a central venous catheter (n ¼ 3), prostatitis (n ¼ 1), cholecystitis (n ¼ 1), and pneumonia (n ¼ 1). In addition, eight patients underwent inferior mesenteric artery embolization for a type II endoleak. All coil embolizations for type II endoleak were performed
through transfemoral catheterization. One patient received an additional stent for a type Ib endoleak without any complication. Symptoms and diagnosis. Median time elapsed between index EVAR and the first clinical signs of infection was 414 days (range, 6 days to 9 years). Median time between the first clinical signs of infection and endograft explantation was 30 days (range, 1 day to 2.2 years) and >2 months in 8 patients (24%). Three patients (9%) had the endograft removed on an emergency basis because of gastrointestinal bleeding in relation to a duodenal fistula. The potential influence of the delay between the first signs of infection and explantation of the endograft was analyzed using a Mann-Whitney U test. Results indicated that the median delay in patients eventually dying of endograft infection was not significantly longer than the median delay in surviving patients (P ¼ .65). All in all, median time elapsed between endograft implantation and explantation was 14.2 months (range, 20 days to 9.5 years). Before explantation, occlusion of one limb of the endograft occurred in one patient and increase of the abdominal aortic aneurysm diameter of >10 mm was observed in eight patients (24%) without any evidence of endoleak. Symptoms of infection varied (Table II); 21 patients (64%) presented with fever and abdominal pain, 8 patients (24%) had fever only, and 4 patients (12%) had abdominal pain only. Blood cultures were positive in 12 patients (27.2%), and cultures of the endograft or of the aneurysm were positive in 24 patients (74%). Imaging identified the infection in all patients, most commonly with CTA scans, in which 31 patients showed signs of infection (94%). Tagged white
Journal of Vascular Surgery Volume
-,
Number
Chaufour et al
5
-
Table III. Microbiology of infected endografts Variables
Table IV. Antimicrobial therapy after endograft removal Cultures (N ¼ 37a)
Organisms
Antimicrobial therapy
Patients, No. (%)
Main antibiotics used Vancomycin
13 (39)
Sterile
9
Gentamicin
12 (36)
Staphylococcus aureus (4), S. epidermidis (5)
9
Tazocillin
11 (33)
Streptococcus
5
Amoxicillin
9 (27)
Enterococcus
4
Metronidazole
5 (15)
Gram-negative (Escherichia coli, Serratia, Proteus, Salmonella)
6
Ofloxacin
5 (15)
Rifampicin
4 (12)
Ceftriaxone
4 (12)
Anaerobes (Bacteroides fragilis)
2
Candida (fungus)
2
Patients with polymicrobial cultures
4
a There were 28 organisms found in 24 patients with a positive culture. Four patients had two bacteria found in the endograft or in the aortic aneurysm.
blood cell scans were performed in 13 patients and were positive in 8 of them (61%), including the 2 patients with negative CTA. CTA confirmed the diagnosis in all patients, with evidence of endograft-enteric fistula (EEF) in 12 patients (36%). Microbiology and antibiotic treatment. Microbiology specimens were obtained from the explanted endograft, aneurysm wall, or thrombus within the aneurysmal sac in all patients. Cultures were positive in 24 patients (74%). Gram-positive cocci were the most common findings in 18 patients (55%; Table III). Cultures were negative in nine patients. All patients with negative cultures were diagnosed with infected aortic endograft through extensive preoperative workup, including CTA showing air in the aneurysm sac, perigraft fluid, or retroperitoneal abscess. A Pearson c2 test showed no significant association between a negative culture and postoperative survival (P ¼ .52). Polymicrobial cultures were found in 4 patients, and 29 patients (88%) received two or more antibiotics (Table IV). Following microbiologic documentation, medical management with broad-spectrum antibiotics including those capable of biofilm penetration was undertaken, including piperacillin-tazobactam plus vancomycin or daptomycin. In 10 patients with abscess or purulence noted intraoperatively, oral antibiotics were continued as lifelong therapy. Patients with negative blood and aortic wall cultures and no purulence were treated within a finite period of antibiotic therapy for a median of 145 days (range, 93-351 days). Adequate documentation was not available to determine the duration of antibiotic therapy following endograft explantation in one patient. Two patients had evidence of reinfection at the time of last follow-up. Surgical management. In 13 patients (40%) without suprarenal fixation of the endograft, suprarenal clamping was performed using a midline laparotomy, with complete retrieval of the unfolded endograft in all cases.
Oxacillin
3 (9)
Othera
7 (21)
Number of antibiotics per patient One
4 (12)
Two
22 (67)
Three or more
7 (21)
Antimicrobial therapy consisted of 16 different antibiotics. a Ciprofloxacin (n ¼ 2), levofloxacin (n ¼ 1), fluconazole (n ¼ 1), cloxacillin (n ¼ 1), ceftazidime (n ¼ 1), teicoplanin (n ¼ 1).
In 20 patients (60%) with suprarenal fixation of the endograft, celiac clamping of the aorta through a midline laparotomy (n ¼ 10) or a transperitoneal left retrorenal aortic approach (n ¼ 4) was performed. In six patients, a supraceliac aortic balloon inserted under fluoroscopic guidance through the transfemoral route was used, followed by endograft explantation and suprarenal clamping. Celiac clamping or use of an aortic balloon was not associated with an increase in postoperative mortality. In most cases, retrieval of the suprarenal hooks was achieved, separating the stents from the fabric by cutting or thermocutting the stitches holding the stents to the fabric. Once this has been done, the pressure of the stents on the aortic wall decreases and allows disconnection of the hooks one by one without damaging the aortic wall. In other cases, retrieval of the suprarenal hooks was achieved with the help of a Rumel tourniquet and a vascular clamp as shown in Fig 1 or with a barrel of a 20-mL syringe (Fig 2) as described by Usatii et al,11 completed if necessary by transection of the fabric and metal supports with scissors and removal of the suprarenal fixation with a needle driver. Complete removal of the suprarenal component of the endograft was impossible in three patients with an inflammatory aorta, leading to excision of the endograft fabric but leaving the suprarenal struts in situ, with one postoperative death due to sepsis and multiple organ failure. Reparation of the duodenum with intestinal resection was associated with endograft explantation in the 12 patients presenting with an EEF. Among those patients, eight had an endograft with hooks (60%). Among 21 patients without an EEF, 12 (57%) likewise had an endograft with hooks. During aortic graft explantation, in
6
Journal of Vascular Surgery
Chaufour et al
---
2016
Fig 1. A, After celiac aortic clamping, removal of the suprarenal hooks of the endograft was achieved with the help of a Rumel tourniquet and a vascular clamp as shown in the schematic drawing (B).
Fig 3. A, Intraoperative photograph of the completed cryopreserved aortic allograft. After aortic revascularization, resection of the aneurysm sac was completed and omentum was wrapped around the aortic allograft. B, Endograft explanted.
Fig 2. A, Total explantation of the endograft after temporary celiac aortic occlusion with a balloon. B, The suprarenal stents were compressed in the barrel of a 20-mL syringe (arrow) as described by Usatii et al.11
addition to duodenal repair for an enteric fistula in 12 patients, vertebral body débridement was achieved in 2 patients. Methods of aortic reconstruction included in situ graft placement in 30 patients (91%) with aortobi-iliac graft configuration in 25 patients and an aortic tube in five patients. Extra-anatomic axillobifemoral bypass was performed in three patients (9%). In situ reconstruction consisted of cryopreserved aortic allograft in 23 patients (Fig 3), autogenous femoral veins in 2 patients, and Intergard silver graft (Maquet, Rastatt, Germany) soaked in rifampin in 5 patients.12 After débridement, the surgical field and the tunnels were irrigated with povidoneiodine solution and hydrogen peroxide. The new conduit was covered with a pedicle omentoplasty mobilized in a retrocolic configuration and wrapped around the aortic graft.6 Outcomes. Median length of stay was 20 days (range, 1-67 days). Death rate at 30 days or during a prolonged continuous hospitalization exceeding 30 days was
39% (n ¼ 13). Causes of these early deaths are shown in Table V. Death was related to rupture of the graft in situ in 3 patients (two aortic allografts and one autogenous femoral vein), multiple organ failure following sepsis in 6 patients, colon necrosis in 3 patients, and irreversible lower limb ischemia in 1 patient. One allograft body ruptured because of reinfection, one allograft ruptured at the anastomotic site, and one femoral vein ruptured at the level of its main body. Although no single form of reconstruction in this series appeared to be superior, its size was too small for statistical comparison. At 1 year, survival was 44% 6 10% (Fig 4), with a nonsignificant trend toward better survival at 1 year in patients without an EEF (51% 6 12%) compared with patients with an EEF (32% 6 12%; P ¼ .50). Three patients were reoperated on, one for thrombectomy of an axillobifemoral bypass at 9 months, one for rupture of an autogenous femoral vein graft at 2 months, and one for a fistula between the duodenum and the aortic allograft at 3 years. All three patients survived. At 1 year, graft-related complication and reinfection rates were 10% and 5%, respectively.
DISCUSSION According to previous studies, infection of abdominal aortic endografts is reported with an incidence of 0.2%
Journal of Vascular Surgery Volume
-,
Number
Chaufour et al
7
-
Table V. Causes of death at 30 days or in the hospital after removal of the infected endograft Causes of death New graft rupture
Patients
Time, days
3
2, 20, 40
Cryopreserved aortic allograft
2
2, 20
Autogenous femoral vein
1
40
Sepsis (multiple organ failure)a
6
2, 2, 12, 15, 45, 56
Colon necrosisb
3
5, 5, 15
Lower limb ischemiac
1
30
a
Two reoperations for recurrent aortic sepsis. b Three colectomies. c Bilateral trash foot, patient refused major amputation.
to 5%.5,8 Our multicenter experience during a 17-year period following that of Smeds et al10 could not provide data on the incidence of endograft infection because many patients had been referred from other centers where the total number of aortic endografts was unknown. What our series has shown is that explantation of an infected aortic endograft results in perioperative mortality of 39% and overall survival at 1 year of 44%. In this series, infection of the endograft occurred early after implantation, and 12 patients had an infectious complication in the perioperative period of index EVAR (Table I). Lockhart et al13 and Vogel et al14 have shown the potential for secondary infection of an aortic endograft by contaminated blood from these complications. In addition, 10 patients underwent a complementary endovascular procedure, followed in 2 of them by a psoas abscess. Considering these findings, systematic prophylactic antibiotic therapy for any subsequent endovascular or even dental procedures after EVAR13 should be recommended, and any infection occurring at or after EVAR should be treated by adequate antibiotic therapy to decrease possible contamination of the endograft. Another possible source of early aortic endograft infection in this series may be the aortic aneurysm itself, which was ruptured in three patients, and an aortic pseudoaneurysm in one patient; but in these cases, there was no evidence of infection at that time. Other potential sources of infection include mechanical erosion due to stent migration or excessive aortic neck angulation, causing repetitive frictional tear, and fixation hooks, resulting in endograft erosion through the aortic wall into the duodenum. However, aortic neck angulation in all patients of this series was <60 degrees, and no stent migration was seen on CTA. On another score, EEF was present in 12 patients (36%). Whereas intramural hooks may have contributed to fistulization and infection, endografts with hooks were used in the same ratio in patients with and without EEF. Furthermore, we did not find bowel perforation to be directly related to the hooks of the endografts, even though some hooks were felt under the finger outstripping the aortic wall. In this series, all aortoenteric fistulas
were developed at the level of the aneurysm sac and seemed associated more with an inflammatory process in relation to the endograft than with a direct perforation of the bowel by the endograft hooks. Erosion of the aorta and bowel by embolization coils is another hypothesis15 potentially applicable in two patients. In this series, median time between the first clinical signs of infection and endograft explantation was 30 days (1 day-2.2 years), and it was not significantly different in the surviving patients and in those who died of endograft infection. This delay could be explained by nonspecific signs of endograft infection6,16 and also by frequent use of antibiotic therapy started elsewhere, making the diagnosis of endograft infection even more difficult,17 particularly in nine patients with negative cultures. However, there was no association between negative culture and postoperative survival. Microorganisms were isolated from intraoperative cultures in 24 of 33 patients (72%) and, as previously reported,6,8,10,15 were predominantly gram-positive organisms. Postoperative antibiotic therapy was guided by culture results in 24 patients. If abscess or purulence was noted intraoperatively, oral antibiotics were continued as lifelong therapy. There is currently no consensus or data that recommend an appropriate duration of antibiotics after explantation of an infected endograft.6 However, analogy with endocarditis suggests 6 weeks when endograft removal is complete. In this series, we followed the standard of care for EVAR infection with endograft explantation and débridement of the infected tissues. We preferably used in situ reconstruction, but despite omentum coverage, three patients died during the postoperative period of graft rupture, two with an aortic allograft and one with an autogenous femoral vein reconstruction. These three patients had presented with an EEF associated with the endograft infection. Among the five patients receiving an in situ Intergard silver graft, two of them with an EEF, no rupture or reinfection was observed. In cases in which aortic replacement with autogenous tissue is not possible, in situ antimicrobial prosthetic graft could consequently be suggested. Furthermore, use of autogenous femoral veins should be reserved for patients who are able to tolerate a long operative time, and cryopreserved aortic allografts are not without complications, including rupture, aneurysmal degeneration, and thrombosis.18 An axillobifemoral bypass, followed by endograft explantation and aortic suture, was used as an alternative in three patients with no aortic stump blowout, but occlusion of one axillobifemoral bypass occurred during the postoperative period. We found explantation of the infected endografts, particularly those with suprarenal hooks, to be technically challenging,11 but it was somewhat facilitated by positioning a supraceliac aortic balloon under
8
Journal of Vascular Surgery
Chaufour et al
---
Fig 4. Patient survival after endograft explantation. At 1 year, global survival was 44% 6 10% with a nonsignificant trend toward a better survival in patients without an endograft-enteric fistula (EEF; 51% 6 12%) compared with patients with an EEF (32% 6 12%; P ¼ .50).
fluoroscopic guidance. This lifesaving technique could explain the fact that removal of endografts with suprarenal hooks was not associated with a higher risk of postoperative deaths. Nevertheless, the theoretical benefit of complete endograft removal including suprarenal struts should be balanced against the added operative risk in these quite ill patients. Nonexcisional therapy of infected aortic endografts was not used in this series. Leaving the infected endograft in place entails mortality rates as high as 70%.8,15,19 In these situations, percutaneous drainage, instillation of antibiotics through drains, and intravenous antibiotics can improve the systemic sepsis temporarily but with progression to death in almost all cases. Despite our poor results, we advocate removal of the infected material as the only hope for cure. Likewise, endovascular extensions with further stent grafts have been used elsewhere in salvage situations to prevent immediate death from rupture or bleeding but fail in the short term without secondary explantation.18 In this series, the postoperative death rate following infected endograft explantation was 39% and comparable to that reported by others.8,15 In univariate analysis, there was a nonsignificant trend toward better survival at 1 year in patients without an EEF (51% 6 12%) compared with patients with an EEF (32% 6 12%). The retrospective nature of this study and the small sample introduce limitations, with lack of power for statistical analysis, which is inevitable in a rare disease. What remains sure is that the growing number of EVAR procedures will inevitably lead to a larger number of infectious complications in the future. Better appreciation of 17
2016
the potential risk of endograft infection and painstaking attention to detail in preparation of the patient and surgical technique for EVAR are of paramount importance, particularly timely use of appropriate antibiotic prophylaxis during not only primary but also secondary interventions,20 which are not consistent across institutions. Prevention is essential in this matter. A high index of suspicion is also required for diagnosis because the clinical presentation could be varied. In this study and in others,8,15,19 none of the methods of reconstruction were significantly better secondary to the rare nature of infection. We recommend that close attention be paid to the intraoperative findings. If there is an abscess cavity or an EEF, extra-anatomic bypass may be a safer choice, provided there is enough aortic neck for a safe aortic suture without compromising patency of the renal arteries. In cases of infected EVAR without purulence or fistula or with a short proximal aortic neck, in situ reconstruction with omental wrapping may be considered an acceptable alternative, even with use of a prosthetic graft.
CONCLUSIONS Aortic endograft infection is high risk, comparable to what is observed following explantation of an infected aortic graft inserted by laparotomy.12,18 Whereas prevention of infection is essential, explantation is the only technique that can potentially result in cure of an aortic endograft infection.
AUTHOR CONTRIBUTIONS Conception and design: XC, GB, JR Analysis and interpretation: XC, JG, YG, RK, PF, SM, GB, JR Data collection: XC, JG, YG, RK, PF, SM, GB, JR Writing the article: XC, GB, JR Critical revision of the article: XC, JG, YG, RK, PF, SM, GB, JR Final approval of the article: XC, JG, YG, RK, PF, SM, GB, JR Statistical analysis: XC, JG, JR Obtained funding: Not applicable Overall responsibility: JR
REFERENCES 1. Greenhalgh RM, Brown LC, Kwong GP, Powell JT, Thompson SG. Comparison of endovascular aneurysm repair with open repair in patients with abdominal aortic aneurysm (EVAR trial 1), 30-day operative mortality results: randomised controlled trial. Lancet 2004;364:843-8. 2. Prinssen M, Verhoeven EL, Buth J, Cuypers PW, van Sambeek MR, Balm R, et al. A randomized trial comparing conventional and endovascular repair of abdominal aortic aneurysms. N Engl J Med 2004;351:1607-18. 3. Capoccia L, Mestres G, Riambau V. Current technology for the treatment of infection following abdominal aortic aneurysm (AAA) fixation by endovascular repair (EVAR). J Cardiovasc Surg (Torino) 2014;55:381-9.
Journal of Vascular Surgery Volume
-,
Number
Chaufour et al
9
-
4. Ducasse E, Calisti A, Speziale F, Rizzo L, Misuraca M, Fiorani P. Aortoiliac stent graft infection: current problems and management. Ann Vasc Surg 2004;18:521-6. 5. 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. 6. 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. 7. 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. 8. 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. 9. Setacci C, De Donato G, Setacci F, Chisci E, Perulli A, Galzerano G, et al. Management of abdominal endograft infection. J Cardiovasc Surg (Torino) 2010;51:33-41. 10. Smeds MR, Duncan AA, Harlander-Locke MP, Lawrence PF, Lyden S, Fatima J, et al. Vascular Low-Frequency Disease Consortium. Treatment and outcomes of aortic endograft infection. J Vasc Surg 2016;63:332-40. 11. Usatii A, Payne W, Santilli S. Removal of an infected endograft and open aortic reconstruction: technical remarks. Ann Vasc Surg 2013;27:679-83. 12. Delva JC, Déglise S, Bérard X, Dubuisson V, Delva F, Stecken L, et al. In-situ revascularization for secondary aorto-enteric fistulae: the success of silver-coated Dacron is closely linked to a suitable bowel repair. Eur J Vasc Endovasc Surg 2012;44: 417-24. 13. 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. 14. Vogel TR, Symons R, Flum DR. The incidence and factors associated with graft infection after aortic aneurysm repair. J Vasc Surg 2008;47:264-9.
15. Sharif MA, Lee B, Lau LL, Ellis PK, Collins AJ, Blair PH, et al. Prosthetic stent graft infection after endovascular abdominal aortic aneurysm repair. J Vasc Surg 2007;46:442-8. 16. Davila VJ, Stone W, Duncan AA, Wood E, Jordan WD Jr, Zea N, et al. A multicenter experience with the surgical treatment of infected abdominal aortic endografts. J Vasc Surg 2015;62: 877-83. 17. Bandyk DF. Antibioticsdwhy so many and when should we use them? Semin Vasc Surg 2002;15:268-74. 18. Touma J, Cochennec F, Parisot J, Fialaire Legendre A, Becquemin JP, Desgranges P. In situ reconstruction in native and prosthetic aortic infections using cryopreserved arterial allografts. Eur J Vasc Endovasc Surg 2014;48:292-9. 19. 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. 20. Kirksey L, Brener BJ, Hertz S, Parsonnet V. Prophylactic antibiotics prior to bacteremia decrease endovascular graft infection in dogs. Vasc Endovascular Surg 2002;36:171-8.
Submitted Apr 1, 2016; accepted Jul 24, 2016.
APPENDIX. AURC collaborators participating in the study Xavier Chaufour, MD (Toulouse), Julien Gaudric, MD (Paris), Yann Goueffic, MD, PhD (Nantes), Éric Steinmetz, MD (Dijon), Réda Hassen Khodja, MD (Nice), Dominique Midy, MD, PhD (Bordeaux), Patrick Feugier, MD (Lyon), Sergei Malikov, MD (Nancy), Eugénio Rosset, MD (ClermontFerrand), Alain Cardon, MD (Rennes), Malcom Legall, MD (Toulouse), Jean-Baptiste Ricco, MD, PhD (Poitiers).
ABSTRACT Objective: Endovascular aneurysm repair (EVAR) is widely used with excellent results, but its infectious complications can be devastating. In this paper, we report a multicenter experience with infected EVAR, symptoms, and options for explantation and their outcome. Methods: We have reviewed all consecutive endograft explants for infection at 11 French university centers following EVAR, defined as index EVAR, from 1998 to 2015. Diagnosis of infected aortic endograft was made on the basis of clinical findings, cultures, imaging studies, and intraoperative findings. Results: Thirty-three patients with an infected aortic endograft were identified. In this group, at index EVAR, six patients (18%) presented with a groin or psoas infection and six patients (18%) presented with a general infection, including catheter-related infection (n ¼ 3), prostatitis (n ¼ 1), cholecystitis (n ¼ 1), and pneumonia (n ¼ 1). After index EVAR, eight patients underwent successful inferior mesenteric artery embolization for a type II endoleak within 6 months of index EVAR and one patient received an additional stent for a type Ib endoleak 1 week after index EVAR. Median time between the first clinical signs of infection and endograft explantation was 30 days (range, 1 day to 2.2 years). The most common presenting characteristics were pain and fever in 21 patients (64%) and fever alone in 8 patients (24%). Suprarenal fixation was present in 20 of 33 endografts (60%). All patients underwent endograft explantation, with bowel resection in 12 patients (36%) presenting with an endograft-enteric fistula. Methods of reconstruction were graft placement in situ in 30 patients and extra-anatomic bypass in 3 patients. In situ conduits were aortic cryopreserved allografts in 23, polyester silver graft in 5, and autogenous femoral vein in 2. Microbiology specimens obtained from the endograft and the aneurysm were positive in 24 patients (74%). Gram-positive organisms were the most commonly found in 18 patients (55%). Early mortality (30 days or in the hospital) was 39% (n ¼ 13) in relation to graft blowout (n ¼ 3), multiple organ failure (n ¼ 6), colon necrosis (n ¼ 3), and peripheral embolism (n ¼ 1). At 1 year, the rates of patient survival, graft-related complications, and reinfection were 44%, 10%, and 5%, respectively. Conclusions: Abdominal aortic endograft explantation for infection is high risk and associated with graft-enteric fistula in one-third of the cases. Larger multicenter studies are needed to better understand the risk factors and to improve preventive measures at index EVAR and during follow-up. (J Vasc Surg 2016;-:1-9.)
From the Department of Vascular Surgery, University of Toulouse, Toulousea; the Department of Vascular Surgery, Hôpital Pitié-Salpétrière, Parisb; the Department of Vascular Surgery, University of Nantes, Nantesc; the Department of Vascular Surgery, University of Nice, Niced; the Department of Vascular Surgery, University of Lyon, Lyone; the Department of Vascular Surgery, University of Nancy, Nancyf; and the Department of Infectious Diseasesg and Department of Vascular Surgery,h University of Poitiers, Poitiers.
*The names of the AURC collaborators participating in the study can be found in the Appendix at the end of this article. Author conflict of interest: none. Correspondence: Jean-Baptiste Ricco, MD, PhD, Division of Vascular Surgery, Department of Cardiothoracic and Vascular Surgery, University of Poitiers, Medical School, 6, rue de la Milétrie, 86073 Poitiers, France (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 2016 by the Society for Vascular Surgery. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jvs.2016.07.126
Extensive use of endovascular aneurysm repair (EVAR) following prospective multicenter trials as a primary treatment modality for abdominal aortic aneurysm1,2 is associated with short series of endograft infection, a rare complication with an incidence between 0.4% and 3%3-10 but with a postoperative mortality as high as 30%, comparable to infection of open aortic grafts.5,8 One large multi-institutional study recently published in the Journal of Vascular Surgery10 suggested that in these cases, complete removal of the infected endograft with débridement of infected tissue and in situ or extraanatomic replacement may be the option most likely to eradicate the infectious process despite a high early postoperative mortality. Furthermore, because of the low incidence of endograft infection, diagnosis may be delayed, and optimal strategy of care has yet to be defined. The aim of this study was to assess the outcomes of aortic endograft infection following EVAR 1
2
Journal of Vascular Surgery
Chaufour et al
---
and to report the technical challenges of aortic endograft explantation in a consecutive series of patients treated in 11 French tertiary vascular centers that are members of the Association Universitaire de Recherche en Chirurgie (AURC, French University Surgeons Association). The AURC is a nonprofit academic association for clinical research in vascular surgery founded 30 years ago. France has 32 university hospitals, and in each of them there is a member of the AURC heading the department of vascular surgery. In our study, 11 university hospitals participated. Results were analyzed by the leading author and a clinical research assistant and were reviewed by all members of the AURC participating in the study.
METHODS After approval by the Institutional Review Boards of the university medical center members of the AURC participating in the study, institutional databases were analyzed to identify all consecutive patients who had undergone explantation of an infected abdominal aortic endograft between January 1998 and January 2015. Patient consent was waived by all Institutional Review Boards because of the retrospective nature of the study. Records were reviewed at these institutions regardless of where EVAR, defined as index EVAR, was performed. Patients with an endograft implanted primarily in an infected field to treat an aortoenteric fistula or an infected aneurysm were excluded from the study. For each patient, demographic and index EVAR details with subsequent follow-up were examined. Clinical indicators of endograft infection were analyzed along with results of cultures, computed tomography, and white blood cell scan. Operative details including strategy of intervention and material used for reconstruction were reviewed. When endograft infection was suspected, an institutional algorithm including computed tomography angiography (CTA) with injection of contrast material and blood cultures before initiation of antibiotics was applied. In case of doubt, a tagged white blood cell scan was performed to confirm the diagnosis of endograft infection. Whereas complete removal of the endograft was recommended by the AURC group, choice of revascularization was left to the vascular surgeon, depending on the presence of gross contamination or enteric fistula, bleeding, or other emergent situations. Postoperative outcomes examining morbidity and mortality were evaluated on electronic medical records and reviewed on site for missing details. The postoperative period was defined as duration of hospitalization regardless of the number of days or within 30 days of the explantation. Data are reported as median with range using nonparametric tests. Kaplan-Meier plot was used to illustrate 1-year survival. All data were entered in a password-encrypted database, and analyses were
2016
Table I. Clinical presentation before initial endovascular aneurysm repair (EVAR) and events following EVAR (n ¼ 33) Patient characteristics
Median (range) or No. (%)
Age, years, median (range)
69 (57-87)
Male gender
33 (100)
Current smoking
20 (60)
Hypertension
19 (57)
Heart failure
18 (54)
Chronic lung disease
12 (36)
Hostile abdomen Previous aortic surgery
8 (24) 3 (9)
ASA class 3 and 4
25 (75)
Patients with early post-EVAR infection
12 (36)
Groin or psoas abscess following index EVARa
6 (18)
Early general post-EVAR infectionb
6 (18)
Post-EVAR endovascular procedures
10 (30)
Embolization for type II endoleak
8 (24)
Embolization of both hypogastric arteriesc
1 (3)
Additional stent for type Ib endoleak
1 (3)
ASA, American Society of Anesthesiologists. a Including one groin abscess, two false femoral artery aneurysms, one infected femoral crossover bypass after index EVAR in a patient with an aortomonoiliac endograft, and two patients with a psoas abscess following coil embolization for a type II endoleak (n ¼ 1) and for a hypogastric aneurysm (n ¼ 1). b Infection with bacteremia following central venous catheter infection (n ¼ 3), bacterial prostatitis (n ¼ 1), cholecystitis (n ¼ 1), and pneumonia (n ¼ 1). c Embolization of one hypogastric artery before EVAR, followed by contralateral hypogastric artery embolization during EVAR.
performed using SPSS software version 23 (IBM Corp, Armonk, NY).
RESULTS Patient demographics, index EVAR, and possible sources of infection. Between January 1998 and January 2015, 33 patients with a median age of 69 years (range, 57-87 years) were treated for an aortic abdominal endograft infection in 11 French university centers. During the study period, a total of 6057 EVARs were carried out in the 11 tertiary centers. Of the 33 infected endografts reported in our series, 18 index EVARs were initially performed in the 11 tertiary centers with a ratio of endograft infection of 0.3% (95% confidence interval, 0.2%-0.4%). Demographics and clinical presentation at index EVAR are presented in Table I, and events following index EVAR are presented in Table II. Indications for EVAR were an aortic aneurysmal disease in 32 patients and an aortic pseudoaneurysm following prior open aortic repair in 1 patient. The endograft was deployed in abdominal aortic aneurysms with a median diameter of 60 mm (range, 50-93 mm) and a median diameter of the aortic neck of 23 mm (18-31 mm). The median length of the aortic neck was 20 mm (10-38 mm). Aortic
Journal of Vascular Surgery Volume
-,
Number
Chaufour et al
3
-
Table II. Index endovascular aneurysm repair (EVAR), events following index EVAR, presenting symptoms of infection, and imaging Age, No. years
Index EVAR
Events following index EVAR
Time from EVAR to symptoms, days
Presenting symptoms
Positive imaging
Time from symptoms to explantation, days
730
Fever, abdominal pain
CTA, psoas abscess
388
1
69
Elective EVAR, ABI
2
73
Elective EVAR, AMI
Groin abscess, drainage
847
Fever, abdominal pain
CTA, psoas abscess
7
3
66
Elective EVAR, ABI
Postoperative fever
453
Fever
CTA, fistula
12
4
57
Urgent EVAR, ABI
Postoperative fever
1638
Fever, abdominal pain
CTA, fistula
820
5
67
Elective EVAR, ABI
Central catheter infected, septicemia
104
Fever, abdominal pain
CTA
209
6
70
Elective EVAR, ABI
Pneumonia, septicemia
19
Fever, abdominal pain
CTA
34
7
77
Elective EVAR, AMI
Crossover femoral bypass infectedd removal
63
Fever, abdominal pain
CTA
280
8
67
Elective EVAR, ABI
Central catheter infected, septicemia
26
Fever, abdominal pain
CTA, psoas abscess
31
9
71
Elective EVAR, ABI
Type II endoleak, IMA embolization
351
Fever
CTA
5
10
62
Elective EVAR, AMI
3468
Fever, abdominal pain
CTA, psoas abscess
67
11
70
Elective EVAR, ABI
2291
Abdominal pain
CTA, fistula
40
12
69
Urgent EVAR, ABI
97
Fever, abdominal pain, hematemesis
CTA, fistula
9
13
75
Urgent EVAR, AMI
674
Fever, abdominal pain, hematemesis
CTA, fistula
55
14
69
Elective EVAR, ABI
82
Fever, abdominal pain
CTA, psoas abscess
166
15
78
Elective EVAR, ABI
644
Abdominal pain
CTA
3
16
67
Elective EVAR, ABI
1190
Fever
CTA
55
17
59
Elective EVAR, AMI
182
Fever, abdominal pain
CTA
89
18
70
Elective EVAR, ABI
670
Fever
CTA
26
19
78
Elective EVAR AMI
1133
Fever, abdominal pain
CTA, fistula
9
20
64
Elective EVAR, AMI
Fever
CTA, psoas abscess
12
21
59
Elective EVAR, ABI
850
Fever, abdominal pain
CTA, psoas abscess
18
22
81
Elective EVAR, ABI
277
Fever, abdominal pain
CTA, fistula
35
23
63
Elective EVAR, ABI
260
Fever, abdominal pain
CTA, fistula
45
24
71
Elective EVAR, AMI
414
Fever, abdominal pain
CTA, fistula
20
25
84
Elective EVAR, AMI
1244
Fever, abdominal pain, hematemesis
CTA, fistula
1
Prostatitis
Type II endoleak, IMA embolization
Groin abscess, drainage
Cholecystitis
6
(Continued on next page)
4
Journal of Vascular Surgery
Chaufour et al
---
2016
Table II. Continued. Age, No. years
Index EVAR
Events following index EVAR Groin abscess, drainage
Time from EVAR to symptoms, days
Presenting symptoms
Positive imaging
Time from symptoms to explantation, days
145
Fever, abdominal pain
CTA
67
180
Fever
CTA, fistula
36
1746
Abdominal pain
CTA, fistula
8
22
Fever, abdominal pain, septicemia
CTA
34
8
Fever, abdominal pain
CTA
30
26
63
Elective EVAR, AMI
27
76
Elective EVAR, ABI
28
68
Elective EVAR, ABI
29
64
Elective EVAR, ABI
30
87
Elective EVAR, ABI
31
74
Elective EVAR, ABI
368
Fever
CTA
45
32
67
Elective EVAR, ABI
443
Abdominal pain
CTA, psoas abscess
45
33
74
Elective EVAR, ABI
1713
Fever
CTA
4
Type II endoleak, IMA embolization
Central catheter infected, septicemia
Type Ib endoleak, stenting
ABI, Aortobi-iliac endograft; AMI, aortomonoiliac endograft with a crossover femorofemoral graft; CTA, computed tomography angiography with contrast media showing signs of infection; IMA, inferior mesenteric artery.
neck angulation was <60 degrees in all patients of this series, and no stent migration was seen on CTA before explantation. There were 27 endografts (81.8%) implanted in the operating room and six endografts (18.2%) implanted in the radiology suite. All index EVARs were performed using a femoral cutdown. Prophylactic intravenous antibiotics were administered in all patients according to local protocols. Infected endografts were Cook Zenith (Cook Inc, Bloomington, Ind) in 14 cases (43%), Medtronic Talent (Medtronic, Minneapolis, Minn) in 7 cases (22%), Gore Excluder (W. L. Gore & Associates, Flagstaff, Ariz) in 6 cases (18%), Anaconda (Vascutek, Terumo, Inchinnan, United Kingdom) in 4 cases (12%), and home-made endografts in 2 cases (5%). During the same period, national statistics of the different brands of endografts used in France showed that the Cook Zenith endograft was the most widely used, with 42% of the national market and 42% among the AURC members participating in the study with 43% of Cook Zenith endografts infected. Therefore, the high ratio of infection of the Zenith endograft is related to its use. In this group, after index EVAR (Table I), four patients (12%) presented with a groin infection and two patients developed a psoas abscess, one after coil embolization of an internal iliac artery aneurysm and one after coil embolization for a type II endoleak. Six patients (18%) presented with a general infection with bacteremia associated with a central venous catheter (n ¼ 3), prostatitis (n ¼ 1), cholecystitis (n ¼ 1), and pneumonia (n ¼ 1). In addition, eight patients underwent inferior mesenteric artery embolization for a type II endoleak. All coil embolizations for type II endoleak were performed
through transfemoral catheterization. One patient received an additional stent for a type Ib endoleak without any complication. Symptoms and diagnosis. Median time elapsed between index EVAR and the first clinical signs of infection was 414 days (range, 6 days to 9 years). Median time between the first clinical signs of infection and endograft explantation was 30 days (range, 1 day to 2.2 years) and >2 months in 8 patients (24%). Three patients (9%) had the endograft removed on an emergency basis because of gastrointestinal bleeding in relation to a duodenal fistula. The potential influence of the delay between the first signs of infection and explantation of the endograft was analyzed using a Mann-Whitney U test. Results indicated that the median delay in patients eventually dying of endograft infection was not significantly longer than the median delay in surviving patients (P ¼ .65). All in all, median time elapsed between endograft implantation and explantation was 14.2 months (range, 20 days to 9.5 years). Before explantation, occlusion of one limb of the endograft occurred in one patient and increase of the abdominal aortic aneurysm diameter of >10 mm was observed in eight patients (24%) without any evidence of endoleak. Symptoms of infection varied (Table II); 21 patients (64%) presented with fever and abdominal pain, 8 patients (24%) had fever only, and 4 patients (12%) had abdominal pain only. Blood cultures were positive in 12 patients (27.2%), and cultures of the endograft or of the aneurysm were positive in 24 patients (74%). Imaging identified the infection in all patients, most commonly with CTA scans, in which 31 patients showed signs of infection (94%). Tagged white
Journal of Vascular Surgery Volume
-,
Number
Chaufour et al
5
-
Table III. Microbiology of infected endografts Variables
Table IV. Antimicrobial therapy after endograft removal Cultures (N ¼ 37a)
Organisms
Antimicrobial therapy
Patients, No. (%)
Main antibiotics used Vancomycin
13 (39)
Sterile
9
Gentamicin
12 (36)
Staphylococcus aureus (4), S. epidermidis (5)
9
Tazocillin
11 (33)
Streptococcus
5
Amoxicillin
9 (27)
Enterococcus
4
Metronidazole
5 (15)
Gram-negative (Escherichia coli, Serratia, Proteus, Salmonella)
6
Ofloxacin
5 (15)
Rifampicin
4 (12)
Ceftriaxone
4 (12)
Anaerobes (Bacteroides fragilis)
2
Candida (fungus)
2
Patients with polymicrobial cultures
4
a There were 28 organisms found in 24 patients with a positive culture. Four patients had two bacteria found in the endograft or in the aortic aneurysm.
blood cell scans were performed in 13 patients and were positive in 8 of them (61%), including the 2 patients with negative CTA. CTA confirmed the diagnosis in all patients, with evidence of endograft-enteric fistula (EEF) in 12 patients (36%). Microbiology and antibiotic treatment. Microbiology specimens were obtained from the explanted endograft, aneurysm wall, or thrombus within the aneurysmal sac in all patients. Cultures were positive in 24 patients (74%). Gram-positive cocci were the most common findings in 18 patients (55%; Table III). Cultures were negative in nine patients. All patients with negative cultures were diagnosed with infected aortic endograft through extensive preoperative workup, including CTA showing air in the aneurysm sac, perigraft fluid, or retroperitoneal abscess. A Pearson c2 test showed no significant association between a negative culture and postoperative survival (P ¼ .52). Polymicrobial cultures were found in 4 patients, and 29 patients (88%) received two or more antibiotics (Table IV). Following microbiologic documentation, medical management with broad-spectrum antibiotics including those capable of biofilm penetration was undertaken, including piperacillin-tazobactam plus vancomycin or daptomycin. In 10 patients with abscess or purulence noted intraoperatively, oral antibiotics were continued as lifelong therapy. Patients with negative blood and aortic wall cultures and no purulence were treated within a finite period of antibiotic therapy for a median of 145 days (range, 93-351 days). Adequate documentation was not available to determine the duration of antibiotic therapy following endograft explantation in one patient. Two patients had evidence of reinfection at the time of last follow-up. Surgical management. In 13 patients (40%) without suprarenal fixation of the endograft, suprarenal clamping was performed using a midline laparotomy, with complete retrieval of the unfolded endograft in all cases.
Oxacillin
3 (9)
Othera
7 (21)
Number of antibiotics per patient One
4 (12)
Two
22 (67)
Three or more
7 (21)
Antimicrobial therapy consisted of 16 different antibiotics. a Ciprofloxacin (n ¼ 2), levofloxacin (n ¼ 1), fluconazole (n ¼ 1), cloxacillin (n ¼ 1), ceftazidime (n ¼ 1), teicoplanin (n ¼ 1).
In 20 patients (60%) with suprarenal fixation of the endograft, celiac clamping of the aorta through a midline laparotomy (n ¼ 10) or a transperitoneal left retrorenal aortic approach (n ¼ 4) was performed. In six patients, a supraceliac aortic balloon inserted under fluoroscopic guidance through the transfemoral route was used, followed by endograft explantation and suprarenal clamping. Celiac clamping or use of an aortic balloon was not associated with an increase in postoperative mortality. In most cases, retrieval of the suprarenal hooks was achieved, separating the stents from the fabric by cutting or thermocutting the stitches holding the stents to the fabric. Once this has been done, the pressure of the stents on the aortic wall decreases and allows disconnection of the hooks one by one without damaging the aortic wall. In other cases, retrieval of the suprarenal hooks was achieved with the help of a Rumel tourniquet and a vascular clamp as shown in Fig 1 or with a barrel of a 20-mL syringe (Fig 2) as described by Usatii et al,11 completed if necessary by transection of the fabric and metal supports with scissors and removal of the suprarenal fixation with a needle driver. Complete removal of the suprarenal component of the endograft was impossible in three patients with an inflammatory aorta, leading to excision of the endograft fabric but leaving the suprarenal struts in situ, with one postoperative death due to sepsis and multiple organ failure. Reparation of the duodenum with intestinal resection was associated with endograft explantation in the 12 patients presenting with an EEF. Among those patients, eight had an endograft with hooks (60%). Among 21 patients without an EEF, 12 (57%) likewise had an endograft with hooks. During aortic graft explantation, in
6
Journal of Vascular Surgery
Chaufour et al
---
2016
Fig 1. A, After celiac aortic clamping, removal of the suprarenal hooks of the endograft was achieved with the help of a Rumel tourniquet and a vascular clamp as shown in the schematic drawing (B).
Fig 3. A, Intraoperative photograph of the completed cryopreserved aortic allograft. After aortic revascularization, resection of the aneurysm sac was completed and omentum was wrapped around the aortic allograft. B, Endograft explanted.
Fig 2. A, Total explantation of the endograft after temporary celiac aortic occlusion with a balloon. B, The suprarenal stents were compressed in the barrel of a 20-mL syringe (arrow) as described by Usatii et al.11
addition to duodenal repair for an enteric fistula in 12 patients, vertebral body débridement was achieved in 2 patients. Methods of aortic reconstruction included in situ graft placement in 30 patients (91%) with aortobi-iliac graft configuration in 25 patients and an aortic tube in five patients. Extra-anatomic axillobifemoral bypass was performed in three patients (9%). In situ reconstruction consisted of cryopreserved aortic allograft in 23 patients (Fig 3), autogenous femoral veins in 2 patients, and Intergard silver graft (Maquet, Rastatt, Germany) soaked in rifampin in 5 patients.12 After débridement, the surgical field and the tunnels were irrigated with povidoneiodine solution and hydrogen peroxide. The new conduit was covered with a pedicle omentoplasty mobilized in a retrocolic configuration and wrapped around the aortic graft.6 Outcomes. Median length of stay was 20 days (range, 1-67 days). Death rate at 30 days or during a prolonged continuous hospitalization exceeding 30 days was
39% (n ¼ 13). Causes of these early deaths are shown in Table V. Death was related to rupture of the graft in situ in 3 patients (two aortic allografts and one autogenous femoral vein), multiple organ failure following sepsis in 6 patients, colon necrosis in 3 patients, and irreversible lower limb ischemia in 1 patient. One allograft body ruptured because of reinfection, one allograft ruptured at the anastomotic site, and one femoral vein ruptured at the level of its main body. Although no single form of reconstruction in this series appeared to be superior, its size was too small for statistical comparison. At 1 year, survival was 44% 6 10% (Fig 4), with a nonsignificant trend toward better survival at 1 year in patients without an EEF (51% 6 12%) compared with patients with an EEF (32% 6 12%; P ¼ .50). Three patients were reoperated on, one for thrombectomy of an axillobifemoral bypass at 9 months, one for rupture of an autogenous femoral vein graft at 2 months, and one for a fistula between the duodenum and the aortic allograft at 3 years. All three patients survived. At 1 year, graft-related complication and reinfection rates were 10% and 5%, respectively.
DISCUSSION According to previous studies, infection of abdominal aortic endografts is reported with an incidence of 0.2%
Journal of Vascular Surgery Volume
-,
Number
Chaufour et al
7
-
Table V. Causes of death at 30 days or in the hospital after removal of the infected endograft Causes of death New graft rupture
Patients
Time, days
3
2, 20, 40
Cryopreserved aortic allograft
2
2, 20
Autogenous femoral vein
1
40
Sepsis (multiple organ failure)a
6
2, 2, 12, 15, 45, 56
Colon necrosisb
3
5, 5, 15
Lower limb ischemiac
1
30
a
Two reoperations for recurrent aortic sepsis. b Three colectomies. c Bilateral trash foot, patient refused major amputation.
to 5%.5,8 Our multicenter experience during a 17-year period following that of Smeds et al10 could not provide data on the incidence of endograft infection because many patients had been referred from other centers where the total number of aortic endografts was unknown. What our series has shown is that explantation of an infected aortic endograft results in perioperative mortality of 39% and overall survival at 1 year of 44%. In this series, infection of the endograft occurred early after implantation, and 12 patients had an infectious complication in the perioperative period of index EVAR (Table I). Lockhart et al13 and Vogel et al14 have shown the potential for secondary infection of an aortic endograft by contaminated blood from these complications. In addition, 10 patients underwent a complementary endovascular procedure, followed in 2 of them by a psoas abscess. Considering these findings, systematic prophylactic antibiotic therapy for any subsequent endovascular or even dental procedures after EVAR13 should be recommended, and any infection occurring at or after EVAR should be treated by adequate antibiotic therapy to decrease possible contamination of the endograft. Another possible source of early aortic endograft infection in this series may be the aortic aneurysm itself, which was ruptured in three patients, and an aortic pseudoaneurysm in one patient; but in these cases, there was no evidence of infection at that time. Other potential sources of infection include mechanical erosion due to stent migration or excessive aortic neck angulation, causing repetitive frictional tear, and fixation hooks, resulting in endograft erosion through the aortic wall into the duodenum. However, aortic neck angulation in all patients of this series was <60 degrees, and no stent migration was seen on CTA. On another score, EEF was present in 12 patients (36%). Whereas intramural hooks may have contributed to fistulization and infection, endografts with hooks were used in the same ratio in patients with and without EEF. Furthermore, we did not find bowel perforation to be directly related to the hooks of the endografts, even though some hooks were felt under the finger outstripping the aortic wall. In this series, all aortoenteric fistulas
were developed at the level of the aneurysm sac and seemed associated more with an inflammatory process in relation to the endograft than with a direct perforation of the bowel by the endograft hooks. Erosion of the aorta and bowel by embolization coils is another hypothesis15 potentially applicable in two patients. In this series, median time between the first clinical signs of infection and endograft explantation was 30 days (1 day-2.2 years), and it was not significantly different in the surviving patients and in those who died of endograft infection. This delay could be explained by nonspecific signs of endograft infection6,16 and also by frequent use of antibiotic therapy started elsewhere, making the diagnosis of endograft infection even more difficult,17 particularly in nine patients with negative cultures. However, there was no association between negative culture and postoperative survival. Microorganisms were isolated from intraoperative cultures in 24 of 33 patients (72%) and, as previously reported,6,8,10,15 were predominantly gram-positive organisms. Postoperative antibiotic therapy was guided by culture results in 24 patients. If abscess or purulence was noted intraoperatively, oral antibiotics were continued as lifelong therapy. There is currently no consensus or data that recommend an appropriate duration of antibiotics after explantation of an infected endograft.6 However, analogy with endocarditis suggests 6 weeks when endograft removal is complete. In this series, we followed the standard of care for EVAR infection with endograft explantation and débridement of the infected tissues. We preferably used in situ reconstruction, but despite omentum coverage, three patients died during the postoperative period of graft rupture, two with an aortic allograft and one with an autogenous femoral vein reconstruction. These three patients had presented with an EEF associated with the endograft infection. Among the five patients receiving an in situ Intergard silver graft, two of them with an EEF, no rupture or reinfection was observed. In cases in which aortic replacement with autogenous tissue is not possible, in situ antimicrobial prosthetic graft could consequently be suggested. Furthermore, use of autogenous femoral veins should be reserved for patients who are able to tolerate a long operative time, and cryopreserved aortic allografts are not without complications, including rupture, aneurysmal degeneration, and thrombosis.18 An axillobifemoral bypass, followed by endograft explantation and aortic suture, was used as an alternative in three patients with no aortic stump blowout, but occlusion of one axillobifemoral bypass occurred during the postoperative period. We found explantation of the infected endografts, particularly those with suprarenal hooks, to be technically challenging,11 but it was somewhat facilitated by positioning a supraceliac aortic balloon under
8
Journal of Vascular Surgery
Chaufour et al
---
Fig 4. Patient survival after endograft explantation. At 1 year, global survival was 44% 6 10% with a nonsignificant trend toward a better survival in patients without an endograft-enteric fistula (EEF; 51% 6 12%) compared with patients with an EEF (32% 6 12%; P ¼ .50).
fluoroscopic guidance. This lifesaving technique could explain the fact that removal of endografts with suprarenal hooks was not associated with a higher risk of postoperative deaths. Nevertheless, the theoretical benefit of complete endograft removal including suprarenal struts should be balanced against the added operative risk in these quite ill patients. Nonexcisional therapy of infected aortic endografts was not used in this series. Leaving the infected endograft in place entails mortality rates as high as 70%.8,15,19 In these situations, percutaneous drainage, instillation of antibiotics through drains, and intravenous antibiotics can improve the systemic sepsis temporarily but with progression to death in almost all cases. Despite our poor results, we advocate removal of the infected material as the only hope for cure. Likewise, endovascular extensions with further stent grafts have been used elsewhere in salvage situations to prevent immediate death from rupture or bleeding but fail in the short term without secondary explantation.18 In this series, the postoperative death rate following infected endograft explantation was 39% and comparable to that reported by others.8,15 In univariate analysis, there was a nonsignificant trend toward better survival at 1 year in patients without an EEF (51% 6 12%) compared with patients with an EEF (32% 6 12%). The retrospective nature of this study and the small sample introduce limitations, with lack of power for statistical analysis, which is inevitable in a rare disease. What remains sure is that the growing number of EVAR procedures will inevitably lead to a larger number of infectious complications in the future. Better appreciation of 17
2016
the potential risk of endograft infection and painstaking attention to detail in preparation of the patient and surgical technique for EVAR are of paramount importance, particularly timely use of appropriate antibiotic prophylaxis during not only primary but also secondary interventions,20 which are not consistent across institutions. Prevention is essential in this matter. A high index of suspicion is also required for diagnosis because the clinical presentation could be varied. In this study and in others,8,15,19 none of the methods of reconstruction were significantly better secondary to the rare nature of infection. We recommend that close attention be paid to the intraoperative findings. If there is an abscess cavity or an EEF, extra-anatomic bypass may be a safer choice, provided there is enough aortic neck for a safe aortic suture without compromising patency of the renal arteries. In cases of infected EVAR without purulence or fistula or with a short proximal aortic neck, in situ reconstruction with omental wrapping may be considered an acceptable alternative, even with use of a prosthetic graft.
CONCLUSIONS Aortic endograft infection is high risk, comparable to what is observed following explantation of an infected aortic graft inserted by laparotomy.12,18 Whereas prevention of infection is essential, explantation is the only technique that can potentially result in cure of an aortic endograft infection.
AUTHOR CONTRIBUTIONS Conception and design: XC, GB, JR Analysis and interpretation: XC, JG, YG, RK, PF, SM, GB, JR Data collection: XC, JG, YG, RK, PF, SM, GB, JR Writing the article: XC, GB, JR Critical revision of the article: XC, JG, YG, RK, PF, SM, GB, JR Final approval of the article: XC, JG, YG, RK, PF, SM, GB, JR Statistical analysis: XC, JG, JR Obtained funding: Not applicable Overall responsibility: JR
REFERENCES 1. Greenhalgh RM, Brown LC, Kwong GP, Powell JT, Thompson SG. Comparison of endovascular aneurysm repair with open repair in patients with abdominal aortic aneurysm (EVAR trial 1), 30-day operative mortality results: randomised controlled trial. Lancet 2004;364:843-8. 2. Prinssen M, Verhoeven EL, Buth J, Cuypers PW, van Sambeek MR, Balm R, et al. A randomized trial comparing conventional and endovascular repair of abdominal aortic aneurysms. N Engl J Med 2004;351:1607-18. 3. Capoccia L, Mestres G, Riambau V. Current technology for the treatment of infection following abdominal aortic aneurysm (AAA) fixation by endovascular repair (EVAR). J Cardiovasc Surg (Torino) 2014;55:381-9.
Journal of Vascular Surgery Volume
-,
Number
Chaufour et al
9
-
4. Ducasse E, Calisti A, Speziale F, Rizzo L, Misuraca M, Fiorani P. Aortoiliac stent graft infection: current problems and management. Ann Vasc Surg 2004;18:521-6. 5. 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. 6. 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. 7. 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. 8. 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. 9. Setacci C, De Donato G, Setacci F, Chisci E, Perulli A, Galzerano G, et al. Management of abdominal endograft infection. J Cardiovasc Surg (Torino) 2010;51:33-41. 10. Smeds MR, Duncan AA, Harlander-Locke MP, Lawrence PF, Lyden S, Fatima J, et al. Vascular Low-Frequency Disease Consortium. Treatment and outcomes of aortic endograft infection. J Vasc Surg 2016;63:332-40. 11. Usatii A, Payne W, Santilli S. Removal of an infected endograft and open aortic reconstruction: technical remarks. Ann Vasc Surg 2013;27:679-83. 12. Delva JC, Déglise S, Bérard X, Dubuisson V, Delva F, Stecken L, et al. In-situ revascularization for secondary aorto-enteric fistulae: the success of silver-coated Dacron is closely linked to a suitable bowel repair. Eur J Vasc Endovasc Surg 2012;44: 417-24. 13. 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. 14. Vogel TR, Symons R, Flum DR. The incidence and factors associated with graft infection after aortic aneurysm repair. J Vasc Surg 2008;47:264-9.
15. Sharif MA, Lee B, Lau LL, Ellis PK, Collins AJ, Blair PH, et al. Prosthetic stent graft infection after endovascular abdominal aortic aneurysm repair. J Vasc Surg 2007;46:442-8. 16. Davila VJ, Stone W, Duncan AA, Wood E, Jordan WD Jr, Zea N, et al. A multicenter experience with the surgical treatment of infected abdominal aortic endografts. J Vasc Surg 2015;62: 877-83. 17. Bandyk DF. Antibioticsdwhy so many and when should we use them? Semin Vasc Surg 2002;15:268-74. 18. Touma J, Cochennec F, Parisot J, Fialaire Legendre A, Becquemin JP, Desgranges P. In situ reconstruction in native and prosthetic aortic infections using cryopreserved arterial allografts. Eur J Vasc Endovasc Surg 2014;48:292-9. 19. 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. 20. Kirksey L, Brener BJ, Hertz S, Parsonnet V. Prophylactic antibiotics prior to bacteremia decrease endovascular graft infection in dogs. Vasc Endovascular Surg 2002;36:171-8.
Submitted Apr 1, 2016; accepted Jul 24, 2016.
APPENDIX. AURC collaborators participating in the study Xavier Chaufour, MD (Toulouse), Julien Gaudric, MD (Paris), Yann Goueffic, MD, PhD (Nantes), Éric Steinmetz, MD (Dijon), Réda Hassen Khodja, MD (Nice), Dominique Midy, MD, PhD (Bordeaux), Patrick Feugier, MD (Lyon), Sergei Malikov, MD (Nancy), Eugénio Rosset, MD (ClermontFerrand), Alain Cardon, MD (Rennes), Malcom Legall, MD (Toulouse), Jean-Baptiste Ricco, MD, PhD (Poitiers).