Biosynthetic Grafts To Replace Infected Prosthetic Vascular Bypasses: A Single-Center Experience

Patients and Methods

Between November 2009 and March 2015, 29 patients with infected infrainguinal arterial bypass grafts (Szilagyi III) presented to our institution. Epidemiologic data and details on earlier revascularization procedures are shown in Table 1. All grafts were patent. Pre-operatively, patients were screened by sonography for superficial veins suitable for autologous replacement. In none of the cases could suitable autologous veins be detected with respect to diameter, length, and sclerotic or inflammatory changes. After pre-operative correction of fluid and electrolyte status, patients underwent prompt complete graft removal, aggressive debridement of all necrotic tissue, meticulous rinsing with Ringer's solution, and in situ replacement with a 6-mm biosynthetic graft (Omniflow^®II, LeMaitre Vascular, Burlington MA, USA). In 12 of the 15 infragenicular reconstructions, a composite graft with a short distal venous segment (whenever available) was placed to avoid kinking of the biosynthetic prosthesis in the popliteal fossa, as experienced in the early phase of this series [11].

Intra-operatively, wound swabs and specimens of the infected prosthetic grafts were obtained for microbiologic work-up. Then, a calculated broad-spectrum antibiotic was started with piperacillin/tazobactam or vancomycin according to the expected hospital-specific microbial spectrum [14]. In 21 cases, a proximal sartorius muscle transposition flap and in two cases a rectus femoris muscle flap were created for biologic protection. In these cases, groin incisions were closed incompletely to allow flap vitality assessment and were managed with negative wound pressure therapy. After five to seven days, incisions were closed secondarily or a split skin mesh graft transplantation was done.

Post-operatively, anticoagulation and antiplatelet therapy was conducted according to our protocol for patients with biosynthetic grafts (Table 2). Deviating from the protocol, one patient from the above-knee bypass group was kept on permanent antiplatelet therapy because of two bypass occlusions within the first 12 months.

Antibiotic therapy was adjusted according to the results of cultures and antibiograms. After two weeks, antibiotics were given orally and continued for 6–12 weeks guided by the clinical course of the infection and laboratory inflammation parameters (leukocyte count, C-reactive protein). In patients with negative cultures (7/29; 24%) intravenous therapy was shifted to a calculated oral antibiotic regimen (co-trimoxazole 960 mg bid; rifampicin 600 mg bid) by the end of the second post-operative week.

Data were collected prospectively in an Excel spreadsheet. Patients were followed on an out-patient basis at least every six months. During their visits, the latest history was taken, and a physical examination and a color duplex ultrasound scan were done. Blood was drawn for C-reactive protein assay and leukocyte count. Survival, limb salvage, and patency were analyzed using Kaplan-Meier curves (SPSS for Windows, V 15.0; SPSS Inc., Chicago, IL, USA).

Results

The replacement procedure could be completed in all 29 patients. In 17 cases, we used the Omniflow II graft alone. In 12, we performed a protheto-venous composite graft (Table 1). There were no peri-operative deaths or early major amputations.

The most common microbial isolate was Staphylococcus aureus (nine cases) followed by coagulase-negative staphylococci (seven cases). Enterococcus (three cases), streptococci (two cases), Enterobacter cloacae (two cases), Serratia marcescens (one case), and Corynebacterium afermentans (one case) were found in cultures. In three patients, multiple isolates could be detected. There were seven culture-negative cases.

In this series, no reinfection could be detected by clinical assessment, ultrasound scanning, or laboratory tests. Therefore, there was no reinfection-related morbidity or death, neither had we to perform a major amputation for reasons of infection.

During the mean followup of 24 months (range 5–66 mos), six patients died., but there were no deaths related to primary or secondary bypass procedures. Four patients died from cardiovascular events, one patient from pneumonia (unrelated to the graft infection), and one patient from a malignant disease. The survival rate was 92% at one year and 87% at two years by Kaplan-Meier analysis.

The limb salvage rate was 89% at one year and 83% at two years [5]. Patients with infragenicular bypasses required amputation: One below the knee for progressive ischemic tissue loss and distal occlusion of the arterial run-off, three transfemoral for recurrent bypass occlusions and loss of the single run-off vessel, and one by hip disarticulation after a secondary surgical site infection with Clostridium perfringens during a bypass revision 34 months after the Omniflow procedure.

The primary patency rate in this series was 67.6% at one year and 61.5% at two years. We achieved a secondary patency rate of 85.3% after one year and 69.0% after two years.

During followup, graft occlusions were observed in 10 patients. There were four early (within the first post-operative mo) and six late occlusions.

Early occlusions appeared only in cases with a below-knee reconstruction. As mentioned earlier, we detected kinking of the bypass in three patients with a complete infragenicular biosynthetic reconstruction by ultrasound scanning after recanalization by thrombectomy. One patient showed recurrent arteriosclerotic outflow vessel stenosis one month and eight months after the operation that led to occlusion. There were three late occlusions of both above- and below-knee bypasses. Over a mean followup of two years, we observed three occlusions in the above-knee group (3/14; 21%) and seven in the below-knee group (7/15; 47%).

Discussion

Treatment of infrainguinal graft infections has to achieve proper and durable restoration of the blood supply to the extremity and permanent extinction of the infection. If these goals are not achieved rapidly, the morbidity and mortality rates are high.

In the recent literature, the one-year mortality rate in cases with peripheral prosthetic bypass infections ranges from 8%–55% [2]. The highest mortality rate was observed in a cohort of patients treated with graft preservation and negative wound pressure therapy [5]. In our series, we observed a mortality rate of 8% at one y and 13% at two years. The fatal events were unrelated to the procedure and not caused by reinfection. These facts indicate that in experienced medical centers, even radical operative procedures with graft removal, in-situ reconstruction, and biologic protection procedures can have an acceptable mortality rate.

Infection of vascular grafts leads to major amputation in as many as 33% of the patients [5] and to relevant changes in health-related quality of life [15]. Therefore, limb salvage is a central therapeutic goal and clinical endpoint. Unfortunately, it is not always reported in a comparable fashion. In our study, there were no early major amputations in response to uncontrolled infection [11]. During followup, five limbs had to be amputated, but there was no recurrent or persistent infection detectable (limb salvage rate 89% at one y and 83% at two years). Armstrong et al. reported similar numbers with autologous reconstructions [2].

By removal of the infected graft and its replacement with a biosynthetic prosthesis, we achieved freedom from reinfection in all cases. These results are confirmed by a small series of nine patients published by Fellmer et al. [12]. Although there are limitations in interpretation because of the small sample in these two studies, similar results have yet not been reported with replacement materials other than autologous vein. We presume that rapid tissue integration and neovascularization of the graft wall secondary to the unique porosity and “natural” environment for immigrating host cells play the key role in the superior resistance to reinfection. In vitro studies by Walter et al. support this presumption [16]. Nevertheless, Bozoglan et al. found contradictory results in an animal model of graft infection that showed a higher susceptibility to infection in a biosynthetic graft compared with polytetrafluoroethylene implanted in a contaminated surgical field [17]. Larger studies and long-term results have to be obtained to judge this graft property reliably.

In this series, the primary patency rate was 67.6% at one year and 61.5% at two years, and the secondary patency rate was 85.3% after one year and 69.0% after two years. There were fewer occlusions in the above-knee than in the below-knee group (3/14 vs. 7/15; 21% vs. 47%) during followup. Although the numbers of patients in these groups are too small to be compared by Kaplan-Meier analysis, there is a trend to patency rates that are known from elective prosthetic bypasses with similar target vessels [18].

The patency-specific properties of the biosynthetic graft have to be taken into consideration during planning of the operation. In our experience, composite grafts with short venous distal segments help to avoid kinking and early occlusions in below-knee reconstructions.

The concomitant antibiotic therapy should be aligned with the specific microbial spectrum regularly found in similar patients at the treating facility. At our hospital, the microbial spectrum of vascular graft infections is monitored constantly, and the calculated initial and long-term antibiotic regimens are audited once a year by clinical microbiologists.

To date, there are neither consistent clinical data nor manufacturer's recommendations on anticoagulation or antiaggregation management of biosynthetic grafts. The protocol used for this study is a practical approach based on our experience (Table 2). There is a basic necessity to generate data on this particular topic, especially from elective bypass surgery series using biosynthetic material.

Conclusion

This study is the first to report a two-year followup of a series of patients after replacement of infected infrainguinal prosthetic grafts with biosynthetic material. It demonstrates the excellent resistance to reinfection of the Omniflow^®II prosthesis. Limb salvage, morbidity, and mortality rates are similar to those observed with autologous vein grafts. Being aware of the small number of patients and the lack of long-term data, we nevertheless argue that biosynthetic material may be a useful alternative to replace infected prosthetic grafts in patients without suitable autologous veins.