Rerouting in vascular access infections using a biosynthetic vascular graft (Omniflow® II)

Vascular access (VA) complications in patients with end-stage renal disease (ESRD) are common and often require interventions. Although loss of patency is by far the most common reason for a surgical intervention, vascular access infections (VAI) are a relevant complication, associated with morbidity and mortality due to sepsis [12]. The VAI are more frequent in arteriovenous grafts (AVG) using polytetrafluoroethylene (PTFE) than in native arteriovenous fistulas (AVF). Infection rates as high as 9% per year have been reported for AVG in comparison to AVF with 4.1% [5, 17]. The VAI may occur as an early complication of VA creation or VA revision for another reason or, in the longer term due to repeated puncture for hemodialysis (puncture site infection) or secondarily due to a systemic infection.

Treatment of VAI usually includes systemic antibiotic therapy and surgical intervention. In early AVG infections, total graft explantation is usually necessary. In cases of late AVF and AVG infections, mostly related to puncture site infections, the surgical options range from local debridement of infected and necrotic tissue to partial graft excision with interposition grafting (PGE), or total graft excision (TGE) with in situ reconstruction or rerouting with an interposition graft (IG). Figure 1 shows the technique how PGE and IG can be performed with rerouting in a safe manner [20].

The chosen surgical treatment largely depends on the extent of VAI, the patient’s condition and the surgeon’s clinical evaluation and preference. Clinical practice guidelines, such as those of the European Society of Vascular Surgery [18] and The National Kidney Foundation’s Kidney Disease Outcomes Quality Initiative [13] may help in decision making but are mostly based on expert opinions.

Older studies have shown acceptable infection control while preserving vascular access using PGE and IG, although the rate of local reinfection requiring further intervention was up to 35% [15]. PTFE was usually used for IG [11, 16], although alternative materials, such as autologous veins, xenografts or biosynthetic grafts have been considered [2].

In this study, PGE and IG is the favored surgical approach for the treatment of infected AVG. To further reduce the risk of recurrent infection, Omniflow® II was routinely used for IG. Omniflow® II is a biosynthetic vascular prosthesis, which is a composite of cross-linked ovine collagen with a polyester mesh endoskeleton (LeMaitre Vascular Inc., Burlington, MA, USA; Fig. 2). This product is mainly available in Europe, Canada and Australia but it is not approved by the U.S. Food and Drug Administration (FDA). Whenever possible, rerouting the graft should be done in healthy tissue. The aim of this study was to investigate all VAI cases at a tertiary center over a 10-year period, then identify and evaluate the outcomes of cases where Omniflow® II was used for IG after PGE in AVG infections.

This is a retrospective analysis of all patients who underwent surgical revision of VAI between January 2009 and December 2018 at a single tertiary referral center. Follow-up data were gathered until the end of May 2021. The local ethics committee approved the study and waived the requirement for written informed consent for patients who were missing documentation (decision number 2019-01677). Patients who had refused informed consent for the use of their health-related data in the context of research, were excluded. A full text search for the terms fistula or shunt and infection was performed on all operative reports of our vascular unit registered between January 2009 and December 2018. The resulting reports were manually screened for eligibility. Preoperative, procedural and postoperative data were collected from surgical and anethesia reports and electronic medical records. Data were stored in a secured REDcap database (Research Electronic Data Capture, Vanderbilt University, Nashville, TN, USA). Follow-up data were collected from nephrological records at our institution. The VA was classified as infected if the attending vascular surgeon considered an infection as highly likely due to clinical presentation (e.g., erythema, swelling, purulent liquid near site), local microbial cultures proved bacterial infection or the patient had positive blood cultures in the absence of another more likely focus. The attending surgeon indicated the time of operation and technique of surgery. Intraoperative swabs or tissue samples were cultivated for bacterial growth. Further PCR testing for specific pathogens and sonification of graft parts were not routinely performed. The PGE and IG was performed if deemed feasible. New grafts were rerouted through noninfected tissue. The choice of graft material was at the surgeon’s discretion. Antibiotic protocols were not standardized. Empirical treatment was administered in all cases and later adjusted according to the microbiological results. The duration of antibiotic treatment was determined on an individual basis, in collaboration with attending nephrologists and infectiologists.

For the statistical analysis the R statistics program version 4.1.3 was used (R Foundation for Statistical Computing, Vienna, Austria). Categorical variables are presented as absolute numbers and percentages. Continuous data are presented as median values and ranges.

Completeness of follow-up was examined using a follow-up index [23].

The patient identification process is illustrated in Fig. 3. The surgical approach to treat VAI varied greatly, emphasizing the need for tailored approaches to different extents of infection and different patient situations. Patients in need of multiple PGE and IG procedures over the years were only included with the first event. Demographic data, patient characteristics and clinical aspects at presentation are shown in Table 1.

A total of 15 rerouting PGE and IG procedures were found, of which 12 cases were reconstructed with Omniflow® II (3 cases were reconstructed with PTFE). In all these cases microbial cultures were obtained. Staphylococcus aureus was by far the most common pathogen (nine cases, including all cases that underwent reconstruction with PTFE); coagulase-negative Staphylococci and Streptococci were identified once each. No microbial growth occurred in four cases despite a clinical suspicion of infection. All of these had been preoperatively treated with antibiotics. The infection necessitating the surgical procedure occurred between 2 and 43 months after the AVG creation. Early postoperative outcomes were favorable, with freedom from local infection at discharge in all cases, without the need of secondary interventions. There was no in-hospital mortality. The duration of systemic antibiotic therapy varied greatly between 1 week and 4 months. During long-term follow-up, a surgical intervention for reinfection was necessary in 4 cases, between 298 and 515 days after the PGE and IG procedure. Of these, three reinfections occurred in the Omniflow® II group and one in the PTFE group. At the time of reinfection all patients presented with local signs of infection and three also had systemic symptoms. Staphylococcus aureus was found in two cases, the others showed no microbial growth. The PGE and IG was performed again in three of these cases and in all of them Omniflow® II was used. One Omniflow® II with a new infection was treated with local debridement and reconstruction using a xenopericardial graft (as the resulting defect was small). The time to reoperation for reinfection in the Omniflow® II group is depicted in Fig. 4 and 2 tertiary infections occurred in the Omniflow® II group after 4 and 15 months. Staphylococcus aureus was again found in the first mentioned patient, who was on home hemodialysis and PGE/IG was performed again in both cases. Loss of patency was more common and a surgical intervention for this reason was necessary eight times. Only five patients were alive at the end of data collection (including one PTFE rerouting case). No death appeared to be related to an AVG infection (Table 2). The median follow-up was 32 months (range 2–101 months) with a median follow-up index of 0.92 (range 0.18–1).

Infections related to vascular access in patients with ESRD are not infrequent and different strategies are established to resolve this problem. In the case of an early AVG infection, total graft excision is usually unavoidable. In situations with late infections related to puncture sites, where the graft has areas of complete soft tissue integration, or in cases of a potentially infected false aneurysm, a partial graft replacement, as described by Raju in 1987 [15] is a good treatment option. In this study two thirds of all identified patients with AVG infections were treated by this method. The median times of 12 months (for Omniflow® II) and 9 months (for PTFE) until the first manifestation of reinfection suggest that the VAI was related to a complication of cannulation.

In the guidelines for the management of vascular graft and endograft infections different classifications are presented [7]. The MAGIC classification was developed for aortic graft infections but with some exceptions this concept is also suitable for AVG and AVF infections for vascular access patients because, as in this case series, not all criteria (clinical or surgical signs, radiology and laboratory with the subclasses major versus minor) were present all the time [14]. During data analysis a certain lack of correct identification and description of an infected situation was encountered, especially if local microbial cultures or blood cultures were missing or inconclusive. This might lead to faulty decision-making regarding PGE and IG, as in the clinical practice it should be offered to selected patients with localized infections, where such a procedure can be safely performed with low risk of reinfection. As in the analysis of the US Renal Data System, with more than 870,000 patients with AVF, AVG or catheters by Locham [12], Staphylococcus aureus was the most common microorganism found in this population which also has the same typical profile of risk factors [12]. These findings underline the fact that a careful and aseptic cannulation technique is the key in ESRD patients who are per se susceptible to septic complications.

The first clinical results of Omniflow® II used as graft material in primary AVG placement in patients without veins were reported in the 1990s showing a primary patency of 43–64% after 3 years with a low risk of infection and a minimal risk of aneurysmatic dilatation [10, 24]. Shakarchi et al. published three cases of patients with a high risk for infection using an Omniflow® II. All three patients did not show any signs of AVG infection during follow-up [1]. In their review of four case series published between 2009–2011, with a total of 236 procedures using Omniflow® II for AVG creation, the pooled primary patency rate was 60.1% (95% confidence interval, CI 53.6–66.5%) and the secondary patency rate was 82.1% (95% CI 76.7–86.9%). The largest study in the review with 158 procedures, reported an infection rate of 5.7% using Omniflow® II as AVG [3]. The low risk for infection of Omniflow® II prostheses was a trigger to use these ovine collagen grafts in other fields of vascular surgery, especially in the lower limbs. Van de Laar et al. reported a primary and secondary patency at 1 year for Omniflow® II as arterial bypass material in an above-knee position (in comparison to PTFE) of 60% vs. 46.9% and 80.8% vs. 82.5%, respectively, with significantly more wounds and ischemia as well as a higher foot infection (WiFi) score in the Omniflow® II group [22]. Currently, Omniflow® II is not only used as an alternative to expanded PTFE (ePTFE) but also to replace infected aortic and peripheral prosthetic grafts in situ [4, 6, 9, 19]. In all these publications no graft ruptures were reported.

The technique of rerouting to avoid in situ reconstruction is a convincing concept even if three patients in the Omniflow® II group suffered from a reinfection after approximately 10–17 months. Interestingly, rerouting procedures were successfully attempted again due to a localized reinfections. It was even possible to intervene in the same manner again in two patients 4 and 15 months after the second rerouting. Successful preservation of vascular access appears to be worthwhile even when the cost for a biological prosthesis is higher than for a regular synthetic AVG.

All reinfections seemed to be related to cannulation, with its inherent risk of bacterial contamination, in a population with a general risk for infections due to renal insufficiency [21]. This case series shows that more than half of the patients could profit in the long term and use the Omniflow® II graft as a vascular access conduit for more than 3 years, with a maximum of 8 years. The patient long-term mortality rate was similar to other reported data [8].

This study also revealed vast differences in the use of adjuvant antibiotic therapy among the cases, especially considering the length of treatment. Although individualized treatment is certainly warranted for infected dialysis access sites, standardized treatment schemes, also with respect to the choice of vascular graft materials, should be developed and implemented on an institutional level or within vascular societies, helping surgeons to choose adequate therapeutic measures and reduce confounding factors for further studies investigating the outcome of treatment for VAI.

In terms of limitations, this retrospective study included procedures with considerable heterogeneity as demonstrated in Fig. 3. The data were also gathered over a long period with potentially changing treatment standards over time. As the cause of death in patients who died was unknown in the majority of cases, more infections, even severe ones, could have been missed. Nevertheless, the use of Omniflow® II for this indication is durable in the long term and even further reinfections can be repetitively treated.

In patients with a VAI and partially well-integrated AVG, the technique of rerouting using a biosynthetic graft, such as Omniflow® II as an interposition graft shows favorable results, even over time but with the necessity of secondary and even tertiary rerouting procedures in a minority of patients. The possibility to do so can help maintain a vascular access site for several years. Meticulous and standardized data collection in registries of patients with VAI is mandatory to further improve the concept of the best treatment for patients with VAI.