2. Materials and Methods
2.1. Study Population and Diagnostic Criteria of AGIs
From January 2008 to December 2019, 43 patients with AGI were admitted to our department (median age, 70 years; interquartile range, 57–78 years); their medical records were retrospectively collected from the hospital database for statistical analysis (see
Figure 1). AGI was diagnosed based on clinical symptoms (fever, abdominal or back pain, gastrointestinal bleeding, dyspnea or hypoxemia, and so forth) and auxiliary examination results (radiography, ultrasonography, and CT angiography), which indicated the presence of peri-graft gas or fluid, retroperitoneal abscess, graft duodenal fistula, pseudoaneurysms, and graft thrombosis (see
Figure 2A,B).
2.2. Peri-Operative Management and Operation Methods
The treatment regimen for patients with AGI was devised according to the patients’ general status, clinical symptoms, and auxiliary examination results. For those AGI patients who had insufficient etiological evidence, empirical antibiotics (usually piperacillin-tazobactam) were prescribed once they were admitted. The blood samples and abscess specimens were collected for microbiological culture whenever possible after admission to identify pathogenic bacteria and administer appropriate anti-infection therapy before the surgery. The anti-infection therapy of AGI patients would change according to their microbiological culture and drug susceptibility test results. The most common antibiotics for AGI patients include vancomycin, imipenem, and amikacin. Furthermore, the clinical course and variety of antibiotics would adjust according to the opinion of the infectious disease specialist. Conservative treatment with percutaneous drainage and irrigation was indicated for AGI patients with poor general status (severe anemia, hypoproteinemia, malnutrition, e.g.,) or serious pre-operative complications (such as advanced malignant diseases). Furthermore, conservative treatment may also be considered for certain AGI patients, such as thoracic/abdominal vascular graft/endograft infection patients with a high risk of surgery or thoracic vascular graft/endograft infection patients with no positive microbiological results, and an absence of an esophagus, the presence of an airway fistula, or severe sepsis. ISR was indicated as the first-line treatment for thoracic and abdominal AGI patients, especially when the infection is limited or the infected pathogen is a low-virulence strain, while EAR would be considered as an alternative option for patients with a large abscess or multi-resistant microorganisms. The infected stent grafts or vascular grafts would be completely removed, while an autologous vein, rifampicin-soaked ePTFE vascular grafts (600 mg rifampicin in 10–15 mL normal saline for 15–20 min at room temperature; Gore-Tex, W.L. Gore & Associates, Flagstaff, AZ, USA) or normal ePTFE vascular grafts were used as bypass conduits in ISR and ESR procedures (see
Figure 2C,D). As for the indication of bypass graft conduits, an autologous vein was the first choice as the bypass conduit for AGI patients, especially for those with limited infection or low-virulence pathogens. Rifampicin-soaked ePTFE vascular graft was indicated for patients with high-virulence pathogens, large or multifocal abscesses, and generalized sepsis, while normal ePTFE vascular graft was indicated for patients without complex infection or multi-resistant pathogens, but autologous vein was not available. Autologous veins were mostly used in abdominal AGI patients because of their limited diameter; however, they can also be used as bypass grafts for TEVAR infection patients under certain conditions. In our cohort, one patient received TEVAR because of a previous thoracic aorta-arteritis pseudoaneurysm. The infected stent graft was removed during the surgery; the diameter of the infected aorta was smaller than normal thoracic aorta due to the affected section being in the distal end of the thoracic aorta, and the patient being diagnosed with aorta-arteritis, hence autologous vein was used as a bypass graft conduit in this patient. In addition, bilateral superficial femoral veins would usually be harvested and sutured together to increase the diameter of the bypass conduit in case of diameter mismatch; furthermore, we would split the femoral vein off and use spiral type anastomosis to re-suture it in order to increase the diameter of the autologous vein bypass graft further.
Due to the possibility that an autologous vein may be used as the conduit for bypass, all AGI patients received lower limb vein ultrasound examinations to evaluate the patency and reflux of veins in their lower extremities. The superficial femoral vein was the most common autologous vein used in the ISR or EAR procedures. In addition, for certain thoracic stent graft infection patients whose CTA or other auxiliary examinations indicated the possibility of a thoracic aorto-bronchial fistula (ABF) or an aorto-esophageal fistula(AEsF), broncho- or gastro-endoscope was used to confirm whether the fistula existed. Similarly, AGI patients who previously received EVAR or a AAA open repair and who have a suspicious aorto-enteric fistula (AEnF) would undergo gastrointestinal endoscopy to detect the existence of a fistula.
For thoracic aortic stent graft infection patients who received ISR surgery, autologous tissues (muscles or pleura) were used to cover the newly implanted graft. This is in contrast to patients who had undergone EAR surgery and suffered a thoracic aortic stent graft infection, where autologous tissues were used to cover the distal end of the thoracic aorta. A specialist in cardiothoracic surgery or gastrointestinal surgery would be consulted about surgical options if there was evidence of an aorto-bronchial fistula (ABF), an aorto-esophageal fistula (AEsF), or an aorto-enteric fistula (AEnF). Direct closure of the fistula would usually be considered for AGI patients with an ABF or an AEsF, while intestinal resection would be conducted for AGI patients with an AEnF. Furthermore, the ends ABF, AEsF, and AEnF were repaired by autologous tissues. In addition, emergency stent graft implantation would be considered a temporary measure in the event of active bleeding complicating an AGI with or without an ABF, AEsF, or AEnF.
2.3. Post-Operative Treatment and Follow-Up Management
Additional specimens and infected vascular stents or stent grafts were collected for bacterial culture and other investigations. The infected region was irrigated with antibiotics and diluted in povidone-iodine and normal saline during the surgery. All patients continued to receive treatment against infection after the surgery, and oral antibiotics were administered for 6 months or more, depending on the results of the microbiological culture, clinical symptoms, and the infectious disease physician’s assessment at the hospital.
All patients were advised to return to the clinic at 3 and 6 months, and annually thereafter for further follow-up examinations. Patients underwent physical examinations for arterial pulse, routine blood tests, C-reactive protein (CRP), procalcitonin (PCT), interleukin-6 (IL-6) levels, and CTA to evaluate the outcome of treatment and determine further treatment plans during the follow-up clinical examinations (see
Figure 3).
2.4. Primary and Secondary Endpoints
The primary endpoints of our study were 30-day mortality and perioperative complication-related morbidity. The secondary endpoints were re-infection (RI) rates, primary and secondary graft patency, overall mortality, duration of antibiotic therapy, and the total number of antibiotic types used for the treatment.
2.5. Statistical Analysis and Ethics Approval
To analyze differences between the means, a student t-test, one-way analysis of variance, and Dunnett-T3 test were performed. The Pearson chi-square test or Fisher exact test was used for the analysis of categorical variables. Overall survival and RI-free survival were assessed using Kaplan–Meier curves, and differences were analyzed using the log-rank test. A p < 0.05 indicated statistical significance. All statistical analyses were performed using the SPSS 21.0 software (IBM SPSS, New York, NY, USA).
This study was approved by the Ethics Committee of West China Hospital of Sichuan University (No.2021-150). Informed consent was obtained from the patients and their family.
4. Discussion
AGI is a challenging complication with high perioperative and overall mortality [
2,
3,
6,
9,
10]. To treat the vascular prosthetic infection, a complete removal of the infected graft combined with EAR (mostly axillofemoral bypass) was considered the standard treatment. However, because of the high perioperative complication rates and overall mortality associated with EAR, ISR was introduced by the University of Texas-Houston group to treat AGI in the late 1980s, and it has since become the treatment of choice owing to its safety and durability [
11]. In our study, ISR had an acceptable early mortality, complication-related morbidity, primary graft patency, a significantly lower RI rate, and lower overall mortality rate than EAR and conservative treatments. Our study confirms that ISR is a safe treatment choice in cases of AGI, which is in line with the conclusion drawn by earlier studies [
5,
6,
7,
9,
12,
13,
14].
PCT is a precursor of calcitonin, which is produced by the thyroid gland. The PCT level is known to increase under several circumstances, such as bacterial infection, major trauma or surgery, and malignant tumors. A systematic review conducted in 2017 concluded that the PCT level is not associated with mortality, RI, mechanical ventilation, or duration of antibiotic therapy and that it has no prognostic value in cases of septic or severe septic shock [
15]. However, PCT evaluation yields high sensitivity and specificity for infection and is hence valuable for early diagnosis and informing treatment decisions [
16,
17]. In addition, a study showed that a CRP level >50 mg/L is associated with mortality in patients with AGI, and CRP has also been reported as a reliable factor associated with an increased RI rate in patients with stent infection [
18,
19]. Thus, we believe that increased levels of inflammatory chemical indicators in AGI patients, such as CRP, IL-6, and PCT, may indicate a poor prognosis and prompt surgeons toward proactive management. No previous study reported the potential association between peak PCT level and 30-day mortality. Although univariate regression results revealed the potential association between the peak PCT level and 30-day mortality, univariate and multivariate regression analysis results investigating this relationship have been inconsistent. We assume that the contradictory results may be attributed to the existence of confounding factors and the limited number of enrolled patients, therefore, more large-scale and prospective studies may be necessary.
Our study also demonstrated a higher 30-day mortality rate in patients requiring emergency intervention than in patients who underwent elective surgery, which is in line with previous studies [
6,
7,
9,
19]. A few studies have also reported age, chronic kidney dysfunction, higher CRP levels, and coronary artery disease as risk factors associated with a poorer prognosis and higher mortality [
19,
20,
21]. However, the aforementioned risk factors were not associated with mortality in our cohort.
While current practice includes the immediate administration of antibiotics to patients with AGI diagnosed with an established vascular prosthetic infection, a consensus on the best duration for oral antibiotic administration after surgery or conservative treatment and the time for hospital discharge remains lacking. In our study, although RI and the duration of antibiotic therapy were not associated, patients receiving ISR and EAR received antibiotic therapy longer than those who received conservative treatment. This could be attributed to the longer survival of patients with ISR and EAR. Furthermore, patients without RI received more types of antibiotics during treatment, which indicates the need for regular blood or abscess cultures to establish evidence of bacteria and adjust antibiotic therapy depending on the results. We recommend that the duration of antibiotic therapy be decided considering the patient’s clinical symptoms and auxiliary examination results, that blood culture tests be conducted periodically during the treatment and follow-up, and that the duration of antibiotic therapy be at least 3–6 months. For patients with AGI experiencing RI, lifetime antibiotic therapy might be required.
Various vascular conduits are used in ISR and EAR surgery, including autologous vein, cryopreserved allografts, silver-coated grafts, rifampicin bonded polyester grafts, and bovine pericardium. In our cohort, autologous vein, rifampicin-soaked, and normal ePTFE vascular grafts were used as conduits in ISR or EAR surgery. Statistical analysis demonstrated no significant differences in the patency of primary and secondary graft conduits. Furthermore, there was no significant difference in the RI rates between these conduits. A few studies have demonstrated the reputation of autologous veins for resisting infection and possessing acceptable graft patency rates in AGI patients [
22,
23,
24]. However, nearly all of these studies, like the present one, were retrospective and observational, which posed problems concerning interstudy heterogeneity and selection bias. In addition, a recent meta-analysis conducted by Batt et al. found that, despite autologous vein having the lowest RI rate of 6%, no significant differences were observed between autologous vein and other common bypass conduits [
25]. It is also revealed that the RI rate of autologous veins shows a negative correlation with age and prosthetic-duodenal fistula. In our cohort, the median age of patients was 70, and there were many patients suffering from an AEnF; thus, autologous veins may not be the best bypass conduit choice for these patients. The RI rate of autologous vein was similar to PTFE vascular grafts. As for the graft patency rate, our research demonstrated that there was no significant difference in the primary and secondary graft patency of autologous vein when compared to PTFE vascular grafts. However, a previous study showed that only 2% of autologous veins had graft occlusion issues and had significant advantages over other common bypass conduits, except for silver-coated polyester grafts [
22]. The discrepancy in our results may be attributed to the fact that the total and sub-group numbers of enrolled patients in our cohort were limited. Furthermore, our research was a retrospective and non-randomized study with only one participating institution, which may cause selection bias and affect the statistical analysis.
Our study showed that patients who underwent conservative treatment or total graft removal plus EAR had a higher risk of RI and overall mortality than patients who underwent total graft removal plus ISR. Furthermore, overall mortality and RI rates did not differ between patients receiving conservative treatment or EAR only. This could be attributed to the poor general status and low life expectancy of these patients, and our data suggest that RI is more likely to occur in these patients. Our study failed to establish whether the initial procedures of AGI patients have an influence on a patient’s prognosis. However, performing ISR or EAR surgery for patients who underwent TEVAR previously may be more challenging and take more time. This may affect the prognosis of AGI patients, and thus, more high-quality studies are necessary. A few studies indicated that partial graft removal is acceptable when the infected region is small and localized, with the remaining graft being stable [
26,
27,
28]. However, partial graft removal should only be considered when patients are not eligible for total graft removal plus ISR. An earlier study also reported that partial graft removal is associated with a risk of RI [
6]. Although ISR is now the first-line treatment for patients with AGI, EAR remains an important treatment choice for AGI. Oderich et al. and Heinola et al. indicated that EAR is suitable for patients with large peri-graft abscesses or MRSA infections [
7]. Although EAR is associated with higher mortality and RI rates, it remains the preferred choice for AGI patients.
The clinical symptoms, prognosis, and hence the treatment strategy also vary depending on the infecting pathogenic bacteria.
Bandyk et al. used ISR with rifampin-soaked gelatin-sealed polyester grafts in patients with vascular prosthetic infection caused by low-virulence pathogens, such as
Staphylococcus epidermidis or
Salmonella, and reported good outcomes with low mortality and RI rates [
29]. However, in the case of infections with high-virulence pathogens, such as MRSA or other multiple resistant bacteria, the prognosis is poor, and ISR may be unsafe [
30]. Our data suggests that patients with high-virulence pathogen infections had severe clinical presentations, higher RI rates, and a poorer prognosis than patients with low-virulence pathogen infections, although no statistically significant difference was observed between these groups. Despite no positive findings, it is important to note that the negative results may be due to the small number of enrolled patients in this study as well as the virulence of pathogens in AGI patients, which influence the treatment choice. For patients infected with high virulence pathogens, EAR may be considered first-line treatment, and the duration of antibiotics would be extended because ISR may no longer be the best option for AGI patient’s and our data suggests that high-virulence pathogens are more likely to cause RI.
At our institution, the patients’ general status, clinical presentations, and life expectancy are considered in making the decision to administer a surgical intervention. ISR is our first choice for surgical treatment of AGI. EAR is considered only if the patient has a large abscess that precludes satisfactory results with ISR or infection with unmanageable multiple resistant bacteria. In addition, we have set several criteria that patients with AGI must meet before hospital discharge, including negative CT findings, normal body temperature for at least 14 days, and 3 consecutive negative blood culture results. Lifelong follow-up is preferred if possible, given that the risk of RI is reported to increase with a prolonged follow-up period [
6].
Our study had several limitations. First, it was a retrospective and non-randomized study in which only one institution participated, which may cause selection bias and affect the statistical analysis. Furthermore, the number of overall and sub-group enrolled patients was relatively small. However, its strengths lie in the relatively long follow-up period and several positive findings regarding the factors that affect the prognosis of AGI patients.