Results from the First in Man German Pilot Study of the Silver Graft,a Vascular Graft Impregnated with Metallic Silver

Materials

Silver Graft (SG; B.Braun Melsungen AG, Tuttlingen, Germany) is a polyester knitted double-velour vascular prosthesis whose external textile surface is coated with metallic silver instead of silver acetate for a sustained silver release over years. After the silver deposition process, the SG is gelatin sealed, as are other commonly used polyester grafts. This prosthesis was originally designed for prophylactic use to prevent graft infections. A long-lasting bactericidal effect has been confirmed in in vitro ¹ and animal studies, ² including long-term studies, with a residual silver content on the graft surface of 98% after 12 months. ¹

It has been reported in numerous studies that silver ions are effective in inhibiting microorganisms, as recently confirmed by confocal laser microscopy. ³ The ion beam–assisted silver coating of polyester fabric was confirmed to be chemically very stable. ¹ Compared with uncoated polyester prostheses, their metallic silver-coated analogues display equivalent mechanical properties, such as textile strength and radial compliance.

A standard dimension of SG, 8 mm in diameter and 60 cm in length, was provided to all investigators. According to anatomic characteristics, the SG prostheses were then length adjusted and implanted according to each institution's guidelines and experiences.

Study Design and Population

Silver Graft First in Man is a prospective single-arm cohort trial of 50 patients with peripheral artery disease or aneurysms.

Indication

Owing to the vast diversity of potential vascular reconstructions, the implantation field for the SG was limited to inguinal and infrainguinal reconstructions, with the possible use of crossover bypasses.

Inclusion Criteria

Patients suitable for inguinal and infrainguinal reconstructions with an overall health status permitting vascular surgery had to fulfill the following criteria:

Exclusion Criteria

Exclusion criteria for implantation of the SG are as follows:

Primary End Point

The primary end point was the partial or complete lack of graft-tissue integration at 2 months quantified with duplex ultrasonography (perigraft fluid detection). Clinical indications for inflammation or liver damage were recorded as well based on leukocyte counts, C-reactive protein (CRP), liver enzymes (glutamic oxaloacetic transferase, glutamic pyruvate transferase, γ-glutamyltransferase, cholinesterase), and creatinine levels.

Secondary End Points

The secondary end points were graft occlusion and the need to explant the graft.

Follow-up

The patients were postoperatively followed up at 2 and 18 months (clinical status, laboratory parameters, Doppler and duplex ultrasonography). For the follow-up, patients were requested to return for control duplex ultrasonography and further blood work. Table 1 provides an overview of the examination schedule for all relevant ultrasound imaging and blood work.

Inflammatory Markers

Postoperative laboratory findings were recorded at days 1, 3, 6, and 60 and compared with the corresponding presurgical baseline values. CRP, body temperature, and leukocyte values were routinely measured to assess the inflammatory process and any signs of postoperative infection.

Liver Enzymes

Given that an effect of silver ion release through the vascular graft wall and into the bloodstream with subsequent adverse effects on hepatocellular integrity could not be ruled out, standard liver enzymes were determined to assess the hepatocellular integrity. Glutamic oxaloacetic transferase, glutamic pyruvate transferase, γ-glutamyltransferase, and cholinesterase were determined at the scheduled intervals as displayed in Table 1.

Patient Demographics, Risk Factors, and Vascular Reconstructions

The average patient age was 69.1 ± 9.0 years. Patients were predominantly male (74%). The overall diabetes rate was 26% (13 of 50). Fontaine classifications were IIa, 4% (2 of 50); stage IIb, 66% (33 of 50); stage III, 14% (7 of 50); and stage IV, 16% (8 of 50). Other indications besides PAOD were aneurysms (2 of 50). Emergency surgery was performed in one patient, whereas 16% of all patients (8 of 50) had SG reconstructions in previously operated tissue, that is, a previously surgical incision site (eg, vein stripping). American Society of Anesthesiologists' (ASA) scores were 28% (ASA 2), 64% (ASA 3), and 8% (ASA 4).

As listed in Table 2, 54% (27 patients) of all vascular reconstructions were femoropopliteal bypasses to the first segment of the popliteal artery, whereas 30% (15 patients) had femorofemoral crossovers. Eight percent (four patients) underwent femoropopliteal II/III bypasses to the second and third segments of the popliteal artery.

Other bypasses, such as iliacofemoral, ilaco externa/profundal, iliacopopliteal, and femoroprofundal bypasses, contributed 8% (4 patients) to all vascular reconstructions.

Statistical Methods

Paired t-tests were applied for laboratory values, that is, CRP and leukocyte counts, before surgery and at the 2-month follow-up.

Evidence of Perigraft Fluid

The in-hospital ultrasound examinations prior to discharge revealed that 14% (7 of 50) of patients had a minimum 1 mm perigraft fluid layer at least partially around the prosthesis or its anastomoses. Among these seven patients, only one patient showed a minimum perigraft fluid layer at the 2-month follow-up that could not be detected at the later follow-up. This particular patient received a crossover bypass where perigraft fluid accumulated in both groins (right: 1.4 × 1.4 cm; left: 1.8 × 1.0 cm). Furthermore, the lower abdomen sonograms revealed fluid dimensions of 3 × 1 × 1.4 cm (right) and 3.4 × 1 × 0.5 cm (left). During routine ultrasonography at 13 months, these fluid pockets were completely dissolved. At postsurgical month 20, the patient died of cardiac complications. The autopsy revealed a well-incorporated bypass without signs of perigraft fluid and/or infection.

One documented case of perigraft fluid at 18 months developed between 2 months and routine computed tomography at 12 months (Table 3). At 18 months, Doppler ultrasonography detected a minimal fluid layer without progression compared with the 12-month examination in an otherwise asymptomatic patient. Besides this case, there were no indications of poor graft-tissue healing or integration.

Patency Rates

There were two bypass occlusions prior to discharge (4.0%). One occlusion occurred proximal to the graft owing to a suspected inflow mismatch. The second occlusion required lysis, percutaneous transluminal angioplasty, and subsequent stenting. Furthermore, an additional patient needed revascularization at 2 months, yielding a primary patency rate of 94.0% (Table 4). At 18 months, six grafts were not patent.

The corresponding Kaplan-Meier curve of primary (unassisted) patency is shown in Figure 1. Accounting for patient deaths, graft explantation, and previous occlusions, the primary patency at 18 months was 75.3%.

OPEN IN VIEWER

Figure 1.

Kaplan-Meier curve for primary unassisted patency.

Graft Infection and Other Serious Adverse Events

At the 2-month follow-up, a graft explantation was reported owing to a persistent methicillin-resistant Staphylococcus aureus (MRSA) infection at postsurgical day 9 in the surgical reconstruction field. This patient underwent vascular surgery (vein stripping and endovascular arterial procedure) 4 weeks prior to study inclusion.

Within the first postsurgical 18 months, an additional case of MRSA graft infection with subsequent graft explantation was reported. This patient did not show any signs of infection or perigraft fluid with a patent graft at the 2-month follow-up. The perisurgical antibiotic regimen consisted of ampicillin and a β-lactamase inhibitor.

The two observed deaths (multiorgan failure, cerebral infarction) at the 2-month follow-up were unrelated to the vascular reconstructions. At 18 months, the accumulated death rate increased to 10% (5 of 50).

There was one patient with a documented wound infection at 2 months that was localized to the superficial access site. Four patients were lost to follow-up.

Multiple Events per Patient

A number of patients had several events. For instance, there is overlap of patients with superficial infections or wound healing complications and other adverse events. In the four cases of prolonged wound healing within the first two postsurgical months, there was one of the two graft infections, one case of transient perigraft fluid, and one non–graft-related death. In addition, the patients who had any of the six bypass occlusions during the 18-month follow-up period had other events as well. Two of the occluded graft patients had a temporary perigraft reaction at some point.

Body Temperature and Leukocytes

Averaged over the entire patient population, there was a slightly elevated body temperature at days 1 and 3. At postsurgical day 6, body temperatures were not elevated. The leukocyte count was determined as part of the general postsurgical documentation. As seen in Table 5, the leukocyte count returned to presurgical values at day 60. There was no significant difference between the leukocyte counts before surgery and at postsurgical day 60 (p = .24).

C-Reactive Protein

CRP levels were elevated on day 3 (Table 6) and returned to baseline values. The CRP values were not significantly different before surgery and at postsurgical day 60 (p = .14).

Evidence of Perigraft Fluid

It is commonly accepted that perigraft seroma is indicative of poor graft incorporation. At later postsurgical time points beyond months, a fibrous pseudomembrane may form to confine perigraft seroma. ⁴ The relevant literature on perigraft seroma, however, has two limiting factors.

First, most larger retrospective studies were done in the 1970s and 1980s, when image quality and resolution were not to today's standard. ^4,5 To detect a 1 mm layer using Doppler ultrasonography was more challenging in those days.

The second limitation is the fact that the perigraft fluid layer thickness is rarely defined to yield a binary incidence rate. Ahn and colleagues defined perigraft reactions as an enlarging sterile fluid collection at the graft implantation site. ⁶ It can be hypothesized that these fluid collections had layers significantly larger than 1 mm. Ahn and colleagues reported higher incidences in extra-anatomic bypasses (4.2%) followed by aortic reconstructions (1.2%) and femoropopliteal bypasses (0.3%). Although these findings most likely describe late-stage perigraft reactions, it is noteworthy that even for femoropopliteal reconstructions, the incidences are around the 0.5% mark. In this study, seven patients had documented perigraft seromas at discharge. Among these patients, in only one could perigraft fluid be detected at the 2-month follow-up that was not detected at 18 months. At the 18-month follow-up, however, another asymptomatic patient had documented perigraft fluid. The overall incidence rate at 18 months is therefore 2% of all initially implanted grafts. Dauria and colleagues found that in arteriovenous graft placement, there is no significant difference between lower and upper extremity placement. ⁷ Their overall seroma incidence rate was 1.7%, which compares well with the 2.0% rate in this study.

Therefore, it can be concluded that SG polyester-based prostheses do not behave differently in the studied patient population. These clinical findings are in agreement with preclinical data in a porcine model in which the incorporation and patency rates of the SG were similar to those of the control group receiving non–silver-coated polyester prostheses. ²

Primary Patency Rates

Considering the studied patient population, the primary patency rate at 2 months (94%) and the assisted patency rate at 18 months of 88% are comparable to the rates observed in other studies with gelatin-coated polyester prostheses. In particular, data on the non–silver-coated analogue of the SG were published by Jensen and colleagues. ⁸ In this study, the UniGraft was randomized against a commonly available polytetrafluoroethylene graft for above-knee femoropopliteal bypasses. Jensen and colleagues reported in a patient population with risk factors such as diabetes (16%) and previous vascular surgery (40%) a 30-day mortality of 0.5%, surgical wound infection of 3%, and wound complications of 11% in similar vascular reconstructions. Two-year-patency rates were 70% (95% confidence interval 62–76%). In the Silver Graft First in Man study, six bypasses occluded at the 18-month follow-up period, which is in agreement with the findings of Jensen and colleagues in this typical patient cohort undergoing vascular reconstructions. In fact, the Kaplan-Meier primary patency rates in this study (75%) are in very good agreement with the 18-month findings of Jensen and colleagues for the non–silver-coated analogue of the SG.

Liver Enzymes

Relative to liver enzymes, CRP and leukocyte concentration measured values were in agreement with published data when available. Regarding liver enzymes, however, an effect of the anesthesia can be observed. Typically, altered liver enzymes are indicative of disturbances in hepatocellular integrity. These deviations from well-established baseline values are especially observed after the administration of volatile anesthetics. ⁹ All of these enzymes correlated with hepatocellular integrity and change as a response of the metabolized anesthesia, especially in the hours following surgery. There is a plethora of clinical trials that study the hepatocellular integrity during and after various volatile anesthesia. ^10,11 These studies focused on the early responses, with several time intervals in the range of hours up to 1 day. Longer periods beyond 24 hours could not be found in the published domain. Most of these studies, however, focused on α-glutathionine-S-transferase, which apparently has a well-accepted sensitivity in this field of clinical research. Nevertheless, after a peak in alanine transminase and aspartate transaminase within the first postsurgical hours, a slight decrease is observed when isoflurane and deflurane inhalants are used. Nevertheless, the measured liver enzymes are in agreement with the reported literature data and therefore do not give reason to suspect any adverse effect on hepatocellular integrity provoked by silver ions that may have been released into the bloodstream through the silver-coated graft wall.

Graft Infections

In our series, we observed one early MRSA graft infection (2%) at postsurgical day 9. This patient had multiple previous interventions at the operation site, including varicose vein stripping 1 month prior to SG implantation. The observed 2% early graft infection rate compares well with the 6% incidence of infrainguinal or inguinal vascular graft infections. ¹² A later MRSA graft infection after the 2-month follow-up occurred and required a partial explantation followed by a complete graft explantation. Considering the accumulated graft infection rate of 4%, it must be accepted that this novel silver-releasing vascular graft cannot guarantee the prevention of graft infections per se. Until now, there have been no randomized studies with silver- or antibiotic-coated grafts that can prove any reduction in de novo graft infections. For more evidence, a registry for use of the SG in patients with risk factors (Fontaine stage IV, diabetes mellitus, reoperation) and graft infections is currently being conducted using Internet-based questionnaires (http://www.silvergraft.com). If the use of a prosthesis in case of a MRSA infection or contamination cannot be avoided, there is a clear rationale for the use of protected grafts (silver plus rifampin) together with intravenous antibiotics. ^13–15

On the other hand, the existing data about the successful use of silver-protected grafts in case of vascular graft infections, even with MRSA, are documented. ¹⁵ Meanwhile, the use of these grafts has been accepted as one possible treatment in the new German guidelines of 2008. ¹⁶ In our experience, the long-term protection of the SG seems to enable the broader, safer, and prolonged use of vacuum therapy in complicated cases with in situ replacement. ¹⁵