Aortic endograft infection (AEI) is a rare but life-threatening complication of endovascular aneurysm repair (EVAR). Studies have usually reported the frequencies of AEI after EVAR and aortic graft infection (AGI) after open aortic surgery together. Over time, AEIs and AGIs have been reported more frequently due to cumulative effects. AEI has been reported to occur in 0.6% of cases after EVAR, with a range of 0.2% to 8% [1-4]; however, the reported incidence may vary depending on the diagnostic criteria used and the duration of the follow-up period after EVAR.

The pathogenesis of AEI is multifactorial. It includes patient factors and any factors resulting in potential endograft exposure to the infecting organism, such as breach of aseptic technique, nearby infection, and bacteremia [5,6]. The risk factors for AEI are summarized in Table 1.

Table 1 . Risk factors for vascular graft/endograft infection.

Perioperative risk factors

Preoperative risk factors

Prolonged preoperative hospitalization

Infection in a remote or adjacent site

Recent percutaneous arterial access at the implant site

Emergency/urgent procedure

Re-intervention

Lower limb infection (ulcer, gangrene, cellulitis)

Groin incision

Intraoperative risk factors

Breach in aseptic technique

Prolonged operation time

Concomitant gastrointestinal or genitourinary procedure

Postoperative risk factors

Postoperative wound complications (infection, skin necrosis, lymphocele, seroma, hematoma)

Graft thrombosis

Patient-related risk factors/altered host defense

Malignancy

Lymphoproliferative disorder

Immune disorders

Corticosteroid administration

Chemotherapy

Malnutrition

Diabetes mellitus/perioperative hyperglycemia

Chronic renal insufficiency/end-stage renal disease

Liver disease/cirrhosis

Immunosuppression by non-suspended anti-tumor necrosis factor α

Data from the article of Back (Rutherford’s vascular surgery; 2014. p. 654-662) [5] and the article of Tatterton and Homer-Vanniasinkam (Injury 2011;42 Suppl 5:S35-S41) [6]..

An adjacent or remote infection at any site can be a causative or contributory factor for AEI. According to the time of occurrence, early AEI is usually caused by contamination during the EVAR procedure or a preexisting bacterial infection at the time of EVAR. In contrast, late AEI is mainly caused by hematogenous seeding from bacteremia (e.g., urinary or respiratory tract infection), bacterial translocation, or iatrogenic contamination during endovascular procedures at follow-up [7]. An AEI can also contribute to the development of an aortoenteric fistula (AEF). An AEF can occur due to enteric wall erosion by direct contact with a pulsatile, infected aorta, or by penetration of the metal fixing device of the endograft.

Some authors believe that AEIs might occur more frequently than AGIs, as the endograft fabric, which is surrounded only by thrombotic material, does not have tissue ingrowth on its walls, in contrast to prosthetic aortic grafts in open aortic surgery [8]. Another difference between the two techniques is that late endovascular re-interventions are more frequently performed after EVAR than after open aortic surgery. Approximately one-third of patients presenting with AEI have undergone one or more adjunctive endovascular procedures after EVAR [9]. Although its frequency is very low, endovascular reintervention appears to be related to the development of AEI.

According to the Vascular Low-Frequency Disease Consortium database, Smeds et al. [10] reported that urinary tract infection was the most common infection, complicating the initial aortic operation. There were various documented infections in 38% of patients with AEI between the time of endograft implantation and the diagnosis of AEI, including sinusitis, infected hematoma, liver abscess, spine abscess, tooth abscess, vertebral infection, Q-fever, and endocarditis. Although the exact portal of entry of the infecting organisms cannot be confirmed in some patients, various antecedent or late-occurring interval infections can be assumed to be the source of the AEI. Table 2 shows the possible sources of AEIs.

Table 2 . Possible sources of aortic endograft infection in 180 EVARs and 26 TEVARs.

Possible source of aortic endograft infection (n=205)	No. (%)
Infection at index operation	70 (34)
Groin infection	14 (7)
Urinary tract infection	16 (8)
Other infection	40 (19)
Contaminated index operation	29 (14)
Endoleak at index operation	52 (25)
No intervention	23 (11)
With intervention	29 (14)
Interval procedure after EVAR	69 (34)
Interval known infection after EVAR	78 (38)

Adapted from the article of Smeds et al. (J Vasc Surg 2016;63:332-340) [10] with original copyright holder’s permission..

EVAR, endovascular aneurysm repair; TEVAR, thoracic endovascular aneurysm repair..

Infecting organisms vary according to the source of infection (exogenous or endogenous), time of onset (early or late), and risk factors (environmental or derived from patients). AEI is most often detected within one year after EVAR, but gradually increases thereafter. Current changes in the infecting microorganisms include more variable microorganisms and an increase in polymicrobial, culture-negative, and fungal infections, in addition to gram-positive and gram-negative infections [10]. Endogenous (from patients) sources of AEI, including preexisting infections close or remote to the abdominal aorta, are more common than exogenous sources. Preexisting infection is one of the major causes of AEI, and AEIs have been reported in patients who underwent EVAR with an overlooked retroperitoneal abscess or an infected aortic aneurysm [10]. Fig. 1 shows an example of AEI development in a patient with a preexisting infection who underwent EVAR at another hospital seven months prior.

Figure 1. Serial computed tomography (CT) images showed the development of aortic endograft infection and pseudoaneurysm at the site of the suprarenal fixing barb in a patient who underwent endovascular aneurysm repair (EVAR) at another hospital. Upon history taking, the patient received antibiotic treatment for 3 weeks due to fever before EVAR. (A) CT at 1 month after EVAR showed no evidence of infection. (B) Seven months after EVAR, fever developed and CT showed a retroperitoneal fluid collection (arrow) around the suprarenal fixing device. (C) After 1 month of antibiotic therapy, follow-up CT showed increased fluid collection (arrow). (D) Follow-up CT a week later showed a saccular aneurysm (arrow) with fluid collection at the level of the left renal vein.

Meta-analysis showed that gram-positive cocci were identified in 47% (95% confidence interval [CI], 42%-51%) of AEIs, gram-negative cocci in 30% (95% CI, 27%-32%), fungi in 10% (95% CI, 8%-12%), and in the remaining 13% (95% CI, 8%-12%), the infection was polymicrobial [11]. Among the various microorganisms, Staphylococcus aureus, Enterobacteriaceae, Pseudomonas aeruginosa, and β-hemolytic streptococci were classified as virulent organisms, while skin-colonizing bacterial flora such as Staphylococcus epidermidis, Corynebacterium, and Cutibacterium acnes were classified as non-virulent agents. Prosthetic graft or endograft infections caused by virulent organisms were associated with a more severe clinical course and significantly higher reinfection rates.

In all organisms, bacterial adherence to a prosthetic graft is the initial step in the development of AEI. After initial adherence of bacteria to the prosthesis, the production of an extracellular glycocalyx (e.g., mucin and slime) by the staphylococci promotes adherence. It protects the bacteria from host defenses. This biofilm makes bacterial culture difficult, deters antibiotic penetration into the infection focus, and impairs phagocytic and antibody functions [12]. Gram-negative organisms produce proteases that contribute to vessel wall necrosis, anastomotic disruption, and pseudoaneurysm formation [13]. Salmonella are gram-negative anaerobic bacteria that can cause primary aortitis with an aortic aneurysm (or pseudoaneurysm) or infection of a previously existing aortic aneurysm [14-16]. Therefore, Salmonella-associated abdominal aortic aneurysms (AAAs) are relatively common in tropical or subtropical Salmonella-endemic areas [17]. Salmonella-associated AEI have been reported years after EVAR [18]. However, Salmonella rarely causes prosthetic graft infections, in contrast to primary aortic infections.

The reported mean time period from endograft implantation to AEI diagnosis was 25 months, ranging from 1 to 128 months. AEI is usually classified as early or late, based on this time period. When it develops within four months after EVAR, it is classified as early AEI, whereas it is classified as late AEI when it occurs thereafter. The clinical symptoms range from generalized weakness and mild fever to hemorrhagic shock due to aortic rupture or sepsis. Most patients (70%) with AEI present with pain, fever, and leukocytosis, while approximately one-third present with weight loss, fatigue, or generalized weakness [19]. When an AEI is associated with an AEF, disastrous clinical features such as massive gastrointestinal bleeding can occur. Infected aortic grafts can also cause spondylitis.

In current practice, AEI is usually diagnosed based on clinical manifestations and laboratory tests, including bacterial cultures and imaging studies. A multidisciplinary team approach is emphasized from the time of diagnosis, consisting of vascular surgeons, infectious disease specialists, medical microbiologists, radiologists, nuclear medicine specialists, and pharmacists.

A fever of unknown origin or unexplained leukocytosis with a concomitant increase in C-reactive protein (CRP) levels may be the only clinical or laboratory sign of AEI. The clinical features of AEI may differ according to the time of onset of the infection. In the early postoperative period, it should be distinguished from post-implantation syndrome (PIS), which usually develops within 48 hours after endograft implantation and is characterized by fever associated with elevated leukocyte and CRP levels. Endograft infections usually develop later than this period. Measuring blood procalcitonin levels may be helpful in distinguishing PIS from an actual bacterial infection.

In 2016, the Management of Aortic Graft Infection Collaboration (MAGIC) introduced criteria to establish the diagnosis of an aortic graft/endograft infection. There are three categories, namely clinical/surgical, imaging, and laboratory findings, which are subdivided into either “major” or “minor” depending upon the weight of evidence that they contributed towards making a definitive diagnosis of aortic graft/endograft infection (Table 3) [20]. According to the MAGIC criteria, vascular graft/endograft infection is suspected in the presence of one major criterion or two minor criteria from different categories, and is diagnosed when there is at least one major criterion plus any other criterion from another category.

Table 3 . Management of Aortic Graft Infection Collaboration (MAGIC) criteria for the diagnosis of aortic graft infection.

	Clinical/surgical	Radiology/imaging	Laboratory
Major	-. Pus (confirmed by microscopy) around graft or in aneurysm sac at surgery. -. Open wound with exposed graft or communicating sinus. -. Fistula development (e.g. AEF or ABF). -. Graft insertion in an infected site (e.g. fistula, mycotic AAA or infected pseudoaneurysm).	-. Peri-graft fluid on CT scan ≥3 months after EVAR. -. Peri-graft gas on CT scan ≥7 weeks after EVAR. -. Increase in peri-graft gas volume on serial images.	-. Organisms recovered from an explanted graft. -. Organisms recovered from an intraoperative specimen. -. Organisms recovered from a percutaneous, radiologically-guided aspirate of peri-graft fluid.
Minor	-. Localized clinical features of AGI (e.g. erythema, warmth, swelling, purulent discharge, pain). -. Fever ≥38°C with AGI as most likely cause.	-. Other e.g. suspicious peri-graft gas/fluid/soft tissue inflammation;. -. Aneurysm expansion;. -. Pseudoaneurysm formation;. -. Focal bowel wall thickening;. -. Discitis/osteomyelitis;. -. Suspicious metabolic activity on FDG-PET CT;. -. Radiolabelled leukocyte uptake.	-. Blood culture (+) and no apparent source except AGI. -. Abnormally elevated inflammatory markers with AGI as most likely cause e.g. ESR, CRP, WBC count.

Adapted from the article of Lyons et al. (Eur J Vasc Endovasc Surg 2016;52:758-763) [20] with original copyright holder’s permission..

AEF, aortoenteric fistula; ABF, aorto-bronchial fistula; AAA, abdominal aortic aneurysm; CT, computed tomography; EVAR, endovascular aortic aneurysm repair; AGI, aortic graft infection; FDG, fluorodeoxyglucose; PET, positron emission tomography; ESR, erythrocyte sedimentation rate; CRP, C-reactive protein; WBC, white blood cell..

Major clinical and surgical criteria include intraoperative identification of pus around a graft, deployment of an endovascular stent-graft into an infected field (e.g., mycotic aneurysm), and situations where direct communication between the prosthesis and a non-sterile site exists, including fistulae and exposed grafts in open wounds. Minor clinical/surgical criteria include localized features of AGI/AEI and fever ≥38°C, where AGI/AEI is the most likely cause.

Major imaging criteria include prolonged presence of peri-graft fluid (≥3 months after EVAR) or peri-graft gas (≥7 weeks after EVAR), and increasing volume of perigraft gas on serial computed tomography (CT) images. Minor imaging criteria include suspicious peri-graft gas/fluid/soft tissue inflammation, aneurysm expansion, pseudoaneurysm formation, focal bowel wall thickening, discitis/osteomyelitis, and evidence from alternative imaging techniques such as fluorodeoxyglucose (FDG) positron emission tomography (PET) CT (Fig. 2) or radiolabeled leukocyte scintigraphy. However, it may be challenging to interpret these images in patients who have undergone repeated endovascular interventions using mechanical and/or liquid devices to treat endoleaks.

Figure 2. ¹⁸F fluorodeoxyglucose (FDG) positron emission tomography/computed tomography (PET CT) in a patient with aortic endograft infection after endovascular aneurysm repair showed FDG uptake around the aortic endograft and urinary tracts.

Various imaging techniques are used to diagnose AEI. Table 4 shows the reported sensitivities and specificities of different imaging modalities for diagnosing AGI/AEI [19]. When AEI is suspected, CT is the primary diagnostic modality, whereas radioisotope leukocyte scintigraphy and FDG PET CT can be used as adjunctive diagnostic tools. Radioisotope-tagged leukocyte scintigraphy is a highly specific and sensitive imaging tool; however, it has some limitations. This procedure is time-consuming because it needs to be performed at two or more different time points (preferably 2-4 and 20-24 hours after injection) and is available in a laboratory specifically equipped to perform leukocyte labeling. The diagnostic accuracy of leukocyte scintigraphy depends on the location of the target region in the body. The tracer is eliminated via the intestinal tract and physiologically taken up by the bone marrow, leading to a problematic interpretation of AEI. When leukocyte scintigraphy was combined with single photon emission computed tomography (SPECT)/CT, the sensitivity increased to 0.99 (95% CI, 0.92-1.00), with a specificity of 0.82 (95% CI, 0.57-0.96) [21]. According to a systematic review and meta-analysis of ¹⁸F-FDG PET CT interpretation methods for the diagnosis of vascular graft and endograft infections, the pattern of FDG uptake was the most optimal interpretation point among various interpretation tools, such as FDG uptake intensity and maximum standardized uptake value [21]. ¹⁸F-FDG PET CT is usually performed when a low-grade infection is suspected but not confirmed by CT. Although the diagnostic sensitivity and specificity of ¹⁸F-FDG PET CT and leukocyte scintigraphy are high, they are not recommended as first-line imaging modalities for diagnosing AEI because of the aforementioned limitations. When an AEF is present, the prosthetic material may be observed through the duodenal wall defect on gastroduodenoscopy.

Table 4 . Sensitivities and specificities for each imaging modality in the diagnosis of vascular graft/endograft infection.

Imaging tool	Reported ranges
	Sensitivity	Specificity
CT angiography	0.64-1.00	0.00-0.86
FDG-PET	0.86-0.98	0.63-0.76
FDG-PET CT	0.80-1.00	0.60-0.92
WBC scintigraphy	0.73-1.00	0.50-1.00
WBC SPECT/CT	0.94-1.00	0.50-1.00

Data from the article of Chakfé et al. (Eur J Vasc Endovasc Surg 2020;59:339-384) [19]..

CT, computed tomography; FDG, fluorodeoxyglucose; PET, positron emission tomography; WBC, white blood cell; SPECT, single photon emission computed tomography..

Major laboratory criteria include the isolation of microorganisms from percutaneous aspirates of peri-graft fluid, explanted grafts, and other intraoperative specimens, whereas minor criteria include positive blood cultures and elevated inflammatory indices with no alternative source of infection. Infection is defined as the invasion and proliferation of microorganisms not ordinarily present in the body. Therefore, positive cultures from percutaneously or intraoperatively obtained specimens are considered the reference standard for the diagnosis of infection. However, retrieving specimens for bacterial culture can be challenging for patients with AEI. Furthermore, bacterial cultures with directly obtained specimens or blood may be negative, even in patients with active AEI, due to prior use of antimicrobial agents.

Collection of the specimen without contamination and of an adequate quantity is essential for determining the accuracy and relevance of microbiological tests. A direct specimen is a surgically explanted prosthetic material or intraoperatively obtained periaortic specimen in patients with AEI. In contrast, indirect specimens include blood and specimens obtained from superficial wounds, draining sinuses, or close anatomical structures. To identify the bacteria, tissue specimens or a portion of the graft material are superior to swab specimens from infected sites. For bacterial cultures, three or more direct specimens should be collected in sterile containers without contamination [22]. A biofilm is a thick layer of prokaryotic organisms that aggregate to form colonies. The bacterial colony attaches to a surface with a slime layer that aids in protecting the microorganisms and promoting their growth and survival. This biofilm makes collecting an infected specimen using a swab or the usual collection method difficult. Therefore, swab cultures should be avoided. Furthermore, they do not allow for the differentiation of colonizing microorganisms from true pathogens.

Enhanced sample processing techniques, such as vortex-mixing of specimens or sonication, improve the detection rate of microorganisms attached to the graft material [23-25]. Another innovative specimen collection technology (microDTTect) allows for contamination-free sampling and can also dislodge bacteria embedded in a biofilm from prosthetic surfaces [26]. Adding broad-range polymerase chain reaction (PCR) detection to sonicated fluid cultures may increase the detection rate of bacteria attached to graft materials [27].

The treatment goals for patients with AEI are to prevent aortic rupture, eradicate septic foci, and restore arterial blood flow to the pelvis and lower extremities. Antimicrobial therapy, removal of the infected endograft and periaortic tissue, and arterial flow restoration with either in situ aortic reconstruction or extra-anatomic bypass are the three main components of treatment. A multidisciplinary team approach is emphasized to accomplish these tasks.

Peri- and postoperative antibiotic therapies for patients with AEI are not standardized regarding the selection of antibiotic agents and their optimal use duration. The antibiotic choice depends on the culture and sensitivity results. In cases in which there are no identified organisms, empirical broad-spectrum antibiotics should be used. The advent of multidrug-resistant microorganisms has made antibiotic therapy more difficult. In some patients, the addition of antifungal agents should be considered, particularly those with enteric fistulas. Due to the constant evolution of antimicrobial agents, the complexity of interpreting microbiological test results, and the emergence of resistant bacterial strains, antimicrobial therapy management should be performed by an infectious disease specialist. When an infected endograft is completely removed and thorough debridement of all infected tissues is performed, a minimum of six weeks of intravenous antibiotic therapy is recommended [28]. In cases of partial removal of the endograft or aortic reconstruction using a new prosthetic graft, an extended period of antimicrobial therapy is usually proposed to prevent recurrence of the infection.

1) Surgical treatmemt

Total endograft removal with debridement of the regional infected tissue is recommended whenever possible to prevent aortic rupture or infection recurrence. Surgical explantation of an infected aortic endograft is technically challenging, particularly in cases of infection of endograft with suprarenal fixation, owing to the risk of suprarenal aortic wall injury caused by barbed metal struts. The procedure is associated with high mortality (18%-30%) and morbidity [29]. Various surgical techniques using a syringe, proctoscope, or Rummel tourniquet have been introduced to remove the suprarenal fixing device [30]. All of these techniques were devised to remove the aortic endograft by converging the suprarenal metal struts inward to avoid aortic wall injury by the metal struts. In the syringe technique, it is essential to have the syringe ready for use in the operating room. A 20 mL syringe barrel is cut to a length of approximately 4 cm, and the sharp cut edge is smoothed with a bone file (Fig. 3). During the syringe technique, it is essential to push the syringe into the aortic lumen between the stent graft and the aortic wall to avoid pulling out the endograft. Aortic cross-clamping should be performed simultaneously with removal of the endograft from the aortic lumen. Alternatively, a supraceliac aortic clamp or balloon can be used. An aortic balloon is inserted via the transfemoral route under fluoroscopic guidance, followed by endograft explantation and suprarenal clamping. According to a French multicenter study, neither supraceliac clamping nor the use of an aortic balloon was associated with increased postoperative mortality [30]. This procedure should be performed by an experienced surgeon.

Figure 3. The syringe technique is used for removal of the aortic endograft. Pushing the syringe into the aorta allowed the suprarenal fixing struts to converge into the syringe. Note that the cut edge of the syringe was smoothened to avoid aortic wall injury during the procedure.

When complete removal of the suprarenal component of the endograft is risky, the fabric component of the endograft may be excised, leaving the metal struts behind (partial removal of the endograft). In our experience, cutting the metal struts requires a strong metal cutter. Partial removal of the endograft is usually performed for the treatment of persistent endoleaks after failure of multiple endovascular treatments. Partial endograft removal carries a lower surgical risk but a relatively higher reinfection rate than total endograft removal in patients with AEI. Therefore, it is necessary to balance surgical and reinfection risks for individual patients.

Before the advent of in situ aortic reconstruction, an extra-anatomic bypass (e.g., axillobifemoral bypass) was performed together with the removal of the infected aortic graft. However, extra-anatomic reconstruction outside the infected field is associated with lower long-term graft patency compared to anatomic reconstruction, and also carries the risk of graft infection [10]. In patients with AEI with extensive contamination and gross purulence, the 2018 Society for Vascular Surgery (SVS) practice guidelines recommend performing an extra-anatomic bypass rather than in situ aortic reconstruction [31]. The American Heart Association (AHA) guidelines also state that infected graft excision and extra-anatomic bypass might be considered in patients with infection due to multidrug-resistant microorganisms, or in patients with extensive intra-abdominal abscesses or peri-graft purulence (Class IIb; Level of Evidence C) [32]. An extra-anatomic bypass can be performed in a single- or two-stage operation; however, a two-stage operation is more frequently performed (initial axillo-bifemoral bypass followed by excision of all infected grafts and aortic closure). The main drawback of this treatment is the risk of blowout of the closed aortic stump, which has been reported in 10% to 20% [32]. To avoid this disastrous complication, biological materials such as bovine pericardium, fascia lata, and omentum have been used to cover or strengthen the aortic stump [33].

Currently, in situ aortic reconstruction is performed more frequently and shows more favorable outcomes than extra-anatomic bypass after removal of the infected aortic endograft [31,34,35]. For in situ aortic reconstruction, autologous femoral vein grafts (“neo-aortoiliac system”, NAIS), cryopreserved arterial allografts, antimicrobial-treated synthetic grafts (e.g., rifampicin-bonded or silver-coated synthetic grafts) or biosynthetic grafts are used. Regarding patency and reinfection rates, autogenous vein grafts or cryopreserved allografts are superior to prosthetic grafts. However, biological grafts have limitations, particularly in urgent situations, since antimicrobial-treated prosthetic grafts are more accessible to obtain than an autogenous vein or cryopreserved allograft.

NAIS was first introduced by Clagett et al. [36] using the lower extremity deep and superficial veins to create an aortoiliac conduit. They reported an in-hospital mortality rate of 10% and an amputation rate of 10% after treating patients with infected aortobifemoral prostheses (n=17) and other complex aortic problems (n=3). In a subgroup analysis, the results were better when using only deep vein grafts than composite deep and superficial vein grafts [35].

For the preparation of a cryopreserved allograft, arterial tissue is usually procured from brain dead donors and cryopreserved in an institutional tissue bank (Fig. 4), or obtained as a pre-made product (CryoArtery, CryoLife Co.). The reported results of using cryopreserved allografts are heterogeneous [37-40]. In a systematic review and meta-analysis of cryopreserved allograft use for patients with aortoiliac infection (n=1,377) [40], the authors reported a 30-day mortality rate of 15% (95% CI, 11.78%-18.31%), peri-anastomotic rupture/allograft disruption rate of 6% (95% CI, 2.77%-9.88%), pooled aneurysmal degeneration/dilatation of 5% (95% CI, 1.60%-9.68%), pseudoaneurysm formation of 3% (95% CI, 1.60%-4.98%), allograft thrombotic or stenotic complication rate of 12% (95% CI, 7.90%-17.15%), and peri-anastomotic infection rate of 3% (95% CI, 1.90%-5.03%). Furthermore, mortality during follow-up was 19% (95% CI, 11.97%-27.58%); among them, allograft-related mortality was 4% (95% CI, 1.56%-6.15%). The pooled allograft-related reoperation rate was estimated to be 25% (95% CI, 17.89%-32.51%) [41].

Figure 4. The images showed a cryopreserved allograft. (A) After thawing the frozen grafts, a composite bifurcated graft was made at the back table with 2 segments of iliac and one aortic allograft. (B) In situ aorto-iliac reconstruction with the cryopreserved allograft was performed for a patient with aortic endograft infection. (C) In situ aortic reconstruction with cryopreserved allograft was again performed in another patient. (D) The allograft was covered with mobilized omentum after completion of the aorto-iliac reconstruction.

We reported 25 in situ abdominal aortic or aortoiliac reconstructions to treat abdominal aortic infections using cryopreserved allografts from our institutional tissue bank [42]. During the median follow-up of 19.1 months (range, 1-73 months), there were 7 (28%) deaths, including 2 (8%) surgical mortalities (<30 days) and 5 late deaths. There were 3 (12%) graft-related complications, including thrombotic occlusion, aneurysmal dilatation, and AEF. Three years after graft implantation, the patient survival rate was 74% and the event-free survival rate was 58% [42]. In our experience, careful handling of frozen allografts is crucial during thawing and back table work. Minor fractures of the crystallized graft tissue may result in late graft rupture after restoration of aortic blood flow. Moreover, covering the graft with vascularized tissue is important (Fig. 4D). The omentum is often used because of its phagocytic effects and its ability to absorb peri-graft fluid and deliver intravenous antibiotics to the target region.

Results of in situ aortic reconstruction depend on the virulence of the pathogen. Voit et al. [43] recently reported the results of in situ aortic reconstruction using autologous venous grafts and cryopreserved arterial allografts. According to the study, the 30-day mortality rate was low (12.5%) despite the high incidence (45.8%) of major postoperative complications requiring reintervention. Among the postoperative complications, recurrent intracavitary infections were the most common; notably, gram-negative and drug-resistant pathogens were the most commonly implicated organisms in recurrent intra-abdominal infections.

Antimicrobial-treated prosthetic grafts have been used to reduce the reinfection rate of prosthetic grafts in infected fields [44,45]. Antibiotic (e.g., rifampicin, vancomycin, and daptomycin)-soaked prosthetic grafts and silver-coated polyester grafts (InterGard Silver Prosthesis; Getinge) are commercially available.

Contrary to our expectations, both rifampicin-bonded and silver-impregnated grafts have been reported to have higher reinfection rates (11.5% and 11%, respectively) than autogenous or cryopreserved grafts [19]. Furthermore, the use of rifampicin-bonded grafts can lead to the emergence of rifampicin-resistant bacterial strains. Considering the cytotoxicity of highly concentrated rifampicin to the vessel wall and the risk of emergence of rifampicin-resistant bacterial organisms, bacteriophage-based enzymes, so-called endolysins, have also been suggested as potential non-antibiotic impregnation agents [46].

According to a recent systematic review of in situ aortic reconstructions for abdominal AGI/AEIs, the cumulative mortality rates after the use of biologic and prosthetic grafts were 15.6% and 27%, respectively, while the graft reinfection rates were 6.3% and 9%, respectively [47]. The cumulative mortality rate reported in studies using autologous veins was 14.8%, and 30-day reinfection rate was 5.7% [47]. Additionally, the Vascular Low-Frequency Disease Consortium database demonstrated a lower overall mortality rate in AGI/AEI patients treated with biological grafts than those treated with prosthetic grafts (Fig. 5) [10]. In a meta-analysis reporting the reinfection rate after in situ reconstruction of infected aortic grafts/endografts, the reinfection rate of polytetrafluoroethylene grafts was 20%, which was significantly higher than those observed with autogenous vein grafts (6%), cryopreserved allografts (9%), rifampicin-soaked prosthetic grafts (11%), and silver-coated prostheses (11%) [11]. Taken together, biologic grafts are superior to prosthetic grafts in terms of mortality and reinfection rates.

Figure 5. Image showed patient survival rates according to the graft material after in situ aortic reconstruction for patients with aortic endograft infection. Adapted from the article of Smeds et al. (J Vasc Surg 2016;63:332-340) [10] with original copyright holder’s permission.

Biosynthetic vascular grafts are a new type of conduit consisting of a bovine collagen matrix on a polyester mesh. It has been used for in situ aortic reconstruction in patients with an infected aorta [48,49]. According to a recently published systematic review regarding the biosynthetic graft (Omniflow II, LeMaitre) used for aortic reconstructions (n=60), early mortality was 9% (95% CI, 1.12%-20.53%) and late mortality was 18% (95% CI, 5.51%-35.34%) [49]. No graft rupture or degeneration occurred during the follow-up period of 9 to 24 months. However, 6% (95% CI, 0.39%-15.81%) developed early graft occlusion and 4% (95% CI, 0%-16.34%) developed early graft stenosis. Two cases (3%) of postoperative reinfection were reported. The freedom from reinfection was 97.71% (95% CI, 87.94%-100%). They also concluded that using a biosynthetic graft for in situ aortic reconstruction in infected fields is a feasible option, with acceptable mortality and low reinfection rates, but with a risk of limb occlusion (6.2%; 95% CI, 0.39%-15.81%) [49].

Despite improvements in outcomes, graft reinfection remains a serious clinical concern. Owing to variations in the infecting organisms and patient conditions, comparing the results of using various conduits for in situ aortic reconstruction is difficult. Table 5 summarizes the advantages and disadvantages of various graft materials used in in situ aortic reconstruction for patients with AEI.

Table 5 . Advantages and disadvantages of various aortic conduits used in in situ aortic reconstruction for patients with aortic endograft infection.

Conduit	Advantages	Disadvantages
Autogenous vein graft	Lower reinfection rate	Optimal vein is not always available
	Desirable patency	Extended surgery time
		Not readily available in emergency setting
		Uncommon but possible postoperative leg edema
Cryopreserved allograft	Lower reinfection rate than prosthetic graft	Not easily available
		Late degeneration (aneurysmal change or graft rupture)
Prosthetic graft	Readily available	Higher reinfection rates compared to biologic grafts
Antimicrobial-treated prosthetic graft	Readily available	Cytotoxicity to the vessel wall
		Emergence of resistant bacterial strain
		Antibacterial effect does not last long
Biosynthetic graft	Readily available	Outcomes need to be evaluated in the future
		Graft occlusion is common

2) Conservative treatment

Conservative treatment involves nonsurgical prolonged antimicrobial therapy with or without percutaneous drainage and antimicrobial solution irrigation of localized abscesses. Conservative treatment is sometimes the only reasonable option, especially for patients with a short life expectancy, unacceptably high surgical risk, or localized infections by organisms with low virulence and susceptibility to antibiotic therapy [50-52]. However, surgery remains the first-line treatment for patients suspected of having an AEF, and conservative treatment is not indicated for such cases.

According to a systematic review and meta-analysis of AEI, 2.5% of patients received conservative treatment [1]. Conservative treatment can be adopted as a bridge before definitive elective surgery or as palliative treatment in patients unfit for open surgery because of comorbidities or concurrent severe sepsis. Percutaneous drainage can reduce the local bacterial burden and provide microbiological specimens for antimicrobial sensitivity testing. The percutaneous catheter can be connected to a closed drainage bag, and repeated irrigation can be performed using an antimicrobial solution.

A retrospective, observational, single-center cohort study from Sweden reported the results of conservative management of aortic graft (n=8) and endograft (n=42) infections [53]. The study included patients without an aortic fistula. With targeted antibiotic therapy, they performed adjunctive procedures whenever possible, focusing on infection source control to reduce the bacterial or fungal burden. Using Kaplan-Meier estimation, conservative treatment without graft removal resulted in 98%, 88%, and 79% patient survival at 30 days, 1 year, and 3 years, respectively. They concluded that antimicrobial treatment would not be needed indefinitely, and that conservative therapy was effective in a selected group of patients with AEI [53]. Another large series of patients treated conservatively was reported by Caradu et al. [54]. Of the 74 patients with aortic graft/endograft infection treated with a conservative strategy without graft explantation, they reported an in-hospital mortality rate of 20%. Freedom from aortic-related death and overall survival rates were 77.1% and 70.4% at 1 year, and 61.7% and 43.1% at 5 years, respectively. The sepsis recurrence rate was 50.0%.

Most studies have reported poor results in patients who did not undergo explantation of the infected aortic graft and were treated conservatively. Lyons et al. [29] reported that the conservative management of AEI inevitably results in death from disease progression, usually within two years of presentation.

Spondylitis is a clinical feature associated with AEI and AGI. In such cases, conservative treatment of the vertebral bone is the standard and is expected to be generally effective [55]. Surgical treatment is indicated only for patients with spinal cord or cauda equina compression with progressive neurological deficits, spinal instability due to extensive bone destruction, significant deformity, or those in whom conservative treatment fails.

Although complete removal of the infected endograft and in situ aortic reconstruction are the best options, partial endograft explantation or conservative treatment can be performed in frail patients presenting with confined endograft infections. After weighing the expected surgical risks related to total endograft removal and the persistent threat of aortic rupture or sepsis due to an unremoved infected endograft, a multidisciplinary team should determine the optimal treatment plan. Fig. 6 shows the management scheme for patients with AEI.

Figure 6. Flow chart showed the suggested treatment algorithm for patients with aortic endograft infection.

Considering the seriousness of AEI, prevention is the best policy for patients undergoing EVAR. EVAR should not be performed in patients with a potential source of infection. Therefore, the prevention of AEI should be initiated by selecting appropriate patients for EVAR. Moreover, it is strongly recommended that any potential sources of dental sepsis be eliminated at least two weeks before the implantation of a prosthetic valve or other intracardiac or intravascular foreign material, unless the procedure is deemed urgent.

In a meta-analysis [56], antimicrobial prophylaxis with broad-spectrum systemic antibiotics significantly reduced the risk of wound infection (relative risk [RR], 0.25; 95% CI, 0.17-0.38) and early graft infection (RR, 0.31; 95% CI, 0.11-0.85) in arterial reconstructions. In all patients undergoing open aortic repair or EVAR, perioperative systemic antimicrobial prophylaxis is usually recommended before skin incision (ideally within 30 min). First or second-generation cephalosporins are the most widely used agents, owing to their tolerance profile and antibacterial spectrum, which include methicillin-susceptible staphylococci (i.e., S. aureus and coagulase-negative staphylococci), streptococci, and some gram-negative bacilli. However, antimicrobial prophylaxis for more than 24 hours did not provide any additional benefits [56].

For patients who have undergone EVAR, antibiotic prophylaxis before any surgical procedure is analogous to prosthetic cardiac valve surgery and may follow the recommendations of the European Society for Cardiology (ESC) [57] and the American College of Cardiology (ACC)/AHA [58]. The most recent guidelines recommend antimicrobial prophylaxis for all patients with aortic endografts before high-risk procedures, such as gingival or periapical tooth procedures or perforation of the oral mucosa, including scaling and root canal procedures [19,31]. For prevention, patient education regarding AEI at the time of hospital discharge is important.

Improvements in endovascular technology have allowed EVAR to be attempted in patients with challenging aortoiliac anatomies. Accordingly, adjunctive endovascular procedures or late re-interventions have been performed more frequently. There has been an anecdotal report of occurrence of an AEI after percutaneous embolization for a type II endoleak [59]. Therefore, avoiding endovascular re-interventions after EVAR is another way to prevent AEI.

An AEI is an uncommon but life-threatening complication of EVAR. Considering the serious consequences, prevention is considered the most effective policy. The MAGIC diagnostic criteria, based on clinical/surgical features, laboratory examinations, and imaging studies, are used to diagnose the condition. For patients with AEI, surgical treatment with complete endograft explantation and in situ aortic reconstruction with autogenous or other biologic grafts is the optimal management strategy. However, this is not always feasible; therefore, modified procedures or alternative conduits may be used to save a patient’s life. Conservative treatment without endograft removal can be an alternative treatment option for selected patients with unacceptably high surgical risk. A multidisciplinary team approach that weighs the risks and benefits of the intended treatment for individual patients is important.

None.

The author has nothing to disclose.