Prevention of perioperative vascular prosthetic infection with a novel triple antimicrobial-bonded arterial graft

Prevention of perioperative vascular prosthetic infection with a novel triple antimicrobial-bonded arterial graft Ibrahim Aboshady, MD,a,b Issam Raad, MD,c Deborah Vela, MD,a,b Mohamed Hassan, MD,d Yara Aboshady, BS,e Hazim J. Safi, MD,f,g,h L. Maximilian Buja, MD,a,b and Kamal G. Khalil, MD,f,g,h Houston, Tex Objective: Previously, we investigated a locally developed technique of bonding arterial grafts with three antimicrobials to protect against early (within 2 weeks) perioperative bacterial contamination encountered occasionally during aortic graft prosthetic reconstruction. Vascular graft infections are classified by their appearance time (early [<4 months] vs late [>4 months] after graft implantation), degree of incorporation into the surrounding vessel wall, connectivity to the postoperative wound, and extent of graft involvement. In the current phase of testing, we evaluated the ability of our novel triple antimicrobial-bonded graft to prevent infection in the first 8 weeks after implantation. Methods: In nine Sinclair miniature pigs, we surgically implanted a 6-mm vascular Dacron patch graft in the infrarenal abdominal aorta. Five pigs received grafts chemically bonded with a 60-mg/mL solution of rifampin, minocycline, and chlorhexidine, and four pigs received unbonded grafts. Before implantation, the five bonded grafts and three of the unbonded grafts were immersed for 15 minutes in a 2-mL solution containing 1-2 3 107 colony-forming units (CFUs)/mL of Staphylococcus aureus (ATCC 29213); the fourth unbonded graft served as a control. Results: At week 9, all of the grafts were explanted. All S aureus-inoculated bonded grafts (n [ 5) showed no bacterial growth. The unbonded, uninoculated graft (n [ 1) showed low-level bacterial growth (<1.2 3 103 CFUs); S cohnii spp urealyticus, but not S aureus, was isolated, which suggested accidental direct perioperative contamination. Two pigs that received S aureus-inoculated, unbonded grafts were euthanized because of severe S aureus infection (<6.56 3 108 CFUs per graft). Results of histopathologic analysis were concordant with the microbiologic findings. Most intergroup differences were observed in the inflammatory infiltrate in the aortic wall at the site of graft implantation. In all pigs that received bonded grafts, Gram staining showed no bacteria. Conclusions: Our triple-bonded aortic graft prevented perioperative aortic graft infection for at least 8 weeks in a porcine model. The synergistic antimicrobial activity of this graft was sufficient to prevent and/or eradicate infection during that period. Further studies are needed to assess the graft’s ability to combat early-onset vascular graft infection for up to 4 months. (J Vasc Surg 2015;-:1-10.) Clinical Relevance: This novel vascular graft was developed to combat devastating periprosthetic aortic infections associated with major arterial reconstructive surgery. When the graft’s ability to provide extended protection has been verified, the device could be recommended for in situ replacement of infected grafts and possibly for routine primary implantationdespecially in immunocompromised patients and those with heavily contaminated fields such as perforating body cavity wounds, hostile abdomen, and redo procedures. The graft’s preventive effects against Staphylococcus aureus could be extended for prevention of other, less common causes of perioperative infections, such as Pseudomonas aeruginosa, other members of the Staphylococcus family, and fungi.

Aortic graft infections pose limb- and life-threatening complications in patients who undergo arterial reconstruction. The incidence of aortic graft infection is 0.6% to 3%,1 with a mortality of #40%, reinfection rate of #18%, and amputation

rate of approximately 25%.2,3 These outcomes are associated with “open” repairs and endovascular techniques.4 The time of onset of arterial graft infection can be early or late. In early-onset infections (<4 months postoperative),

From the Department of Cardiovascular Pathology, Texas Heart Institutea; the Department of Pathology and Laboratory Medicine,b Department of Internal Medicine,d and Department of Cardiothoracic and Vascular Surgery,f The University of Texas Health Science Center; the Department of Infectious Diseases, Infection Control and Employee Health, The University of Texas MD Anderson Cancer Centerc; the Department of Biochemistry, Rice Universitye; the Department of Cardiothoracic and Vascular Surgery, Baylor College of Medicineg; and the Memorial Hermann Heart and Vascular Institute.h This study was supported by The University of Texas Health Science Center, Houston, Tex (L.M.B). Author conflict of interest: I.R. is a coinventor on two patents associated with antibiotic-coated devices. These patents are the property of the University of Texas MD Anderson Cancer Center and Baylor College of Medicine. Both patents were licensed to Cook Critical Care, American Medical

Systems, Biomet, and TyRx with royalty rights to the institutions and inventors involved. One other patent has been licensed to Akorn. I.R. is also on the speakers’ bureau for, and has received grants from, Cook Inc. Presented at the Twenty-sixth Annual Transcatheter Cardiovascular Therapeutics (TCT), Washington, D.C., September 13-17, 2014. Correspondence: Kamal G. Khalil, MD, University of Texas Health Science Center, 6400 Fannin St, Ste 2850, Houston, TX 77030 (e-mail: kamal. [email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest. 0741-5214 Copyright Ó 2015 by the Society for Vascular Surgery. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jvs.2015.09.061

1

JOURNAL OF VASCULAR SURGERY --- 2015

2 Aboshady et al

the main features might be toxicity with fever and leukocytosis. Other clinical features might include septicemia, wound infection, unexpected thrombosis, and graft dehiscence with false aneurysm or profuse anastomotic bleeding.4 In late-onset infections (>4 months postoperative), the clinical picture might be more subtle and the body temperature normal, even though complications such as false aneurysm, bleeding, fistulas, or remote infections might be present.5-7 Multiple pathogens, including Gram-positive and Gramnegative bacteria and (rarely) fungi, have been implicated in aortic graft infections. Early infection is usually monomicrobial, the most common cause being Staphylococcus aureus. Late infection is generally polymicrobial,8 S epidermidis often being the responsible pathogen.9 Methicillin-resistant S aureus (MRSA) is an emerging organism that causes aggressive infection and increases mortality.4 Regardless of the causative organism or route of infection, a common pathogenetic mechanism is involved. First, body fluids (blood, serum, lymph) surrounding the graft are inadequately perfused.5 The poorly vascularized, perigraft fluid collections help bacteria proliferate; even in low numbers, these bacteria can trigger a graft infection. However, if there is no earlier perigraft contamination, the graft can link to adjacent tissues and obliterate the perigraft space, and become more resistant to infection. The classic management strategy is total excision of the graft, extensive debridement of infected material, and extra-anatomic bypass to a distal vessel.10 More recently, surgeons have adopted in situ reconstruction, using several types of conduits, including autogenous veins, coated synthetic material, and cryopreserved allografts.11 Our group has often performed in situ replacement of the thoracic aorta with a Dacron graft wrapped with a greater omental flap, which necessitates lifelong antibiotic therapy.12 This maneuver avoids aortic stump “blow-out” if the thoracic or abdominal aorta is sewn shut. Despite elaborate planning and execution, complications of graft infection are devastating. Hence, the primary goal should be prevention of such infection. In an earlier 2-week pilot study, we investigated Dacron grafts that were coated with a combination of chlorhexidine, rifampin, and minocycline (M/RCHX), inoculated with S aureus, and implanted in the infrarenal aorta of Sinclair miniature pigs.13 In the current study, we hypothesized that bonding three antimicrobial agents to the graft would prevent or minimize perioperative graft infections in a pig model for at least 8 weeks. METHODS These studies followed the guidelines of the Animal Welfare Committee at The University of Texas Health Science Center, Houston. Animal model. Nine 8-month-old, 30 to 35 kg Sinclair miniature swine (Sinclair Bio Resources, Columbia, Mo) were divided into three groups (Fig 1):

d

d

d

Group one (n ¼ 1): This single pig (control animal) received an unbonded, uninoculated graft. Group two (n ¼ 3): These pigs received unbonded, inoculated grafts. Before implantation, the grafts were immersed for 15 minutes in a 2-mL bacterial solution containing 1-2  107 colony-forming units (CFUs)/mL of S aureus (ATCC 29213). We immersed the graft before implantation to mimic accidental graft contamination during surgery and provide a uniform, reproducible model as opposed to the local contamination method. The bacterial count in the inoculum was high enough to ensure that the soaked graft would be heavily inoculated with bacteria and that the bonded graft would be tested under conditions similar to severe aortic graft infection. Group three (n ¼ 5): These pigs received grafts bonded with triple antibiotics (M/RCHX) and inoculated with bacteria similar to those in group 2.

Surgical protocol. Each pig was given an intramuscular injection of a 1-mL mixture of zolamine (5 mg/ mL), butorphanol (1 mg/mL), ketamine (10 mg/mL), and xylazine (2 mg/mL). The animal was then intubated, and anesthesia was maintained with isoflurane inhalation (0.5%-3%) throughout the procedure. For analgesia, we used 0.25% bupivacaine (1 mL/kg) before the incision was made and buprenorphine (0.01-0.02 mg/kg, subcutaneously or intramuscularly) postoperatively. A fentanyl transdermal patch (75-100 mg/kg/h) was placed immediately after surgery. Surgical details were described in the pilot phase of this study.13 Briefly, the infrarenal abdominal aorta was exposed via a retroperitoneal incision in the left flank and was cross-clamped. A 20-  6-mm Dacron roof patch was sutured to the aortotomy incision with a continuous fine monofilament polypropylene suture. Routine sterile surgical technique was used. At the end of the procedure, the animals were extubated and followed up in our largeanimal veterinary care unit for 9 weeks. They were monitored for clinical and laboratory signs of infection. No antibiotic agents or blood transfusions were administered. Autopsy examination was done for pigs 2 and 3 at postoperative weeks 1 and 2, respectively, and during week 9 for the rest of the animals. The aorta was excised, and the grafts were processed for microbiologic and histopathologic studies. Grafts. Knitted Dacron vascular grafts (Terumo Cardiovascular Systems Corp, Ann Arbor, Mich) were impregnated with M/RCHX using a modified proprietary method that involved precoating the grafts with a chlorhexidine solution and soaking them in a solvent solution of minocycline and rifampin. The doses of rifampin and minocycline (30 and 15 mg/mL, respectively) were selected to ensure that the largest graft contained less than the normal daily dose of both drugs used clinically. Before implantation, each graft was soaked for 15 minutes in a bacterial solution of S aureus (ATTC strain 29213).14,15

JOURNAL OF VASCULAR SURGERY Volume -, Number -

Aboshady et al 3

Fig 1. Flow chart of the pilot phase of the study. AM, Antimicrobials; Bact, bacteria; Gr, graft; IV, intravenous.

Microbiologic assays. Samples of vascular tissue surrounding each graft were homogenized for 2 minutes in 5 mL of 0.9% sterile saline by using a Stomacher Lab Blender (Tekmar Corp, Cincinnati, Ohio). The resulting homogenate was quantitatively cultured to determine the efficacy of the antimicrobial grafts. Samples identified as S aureus were further typed genetically for molecular comparison. Samples from other areas with signs of infection (abscesses) were cultured and typed for comparison with the inoculated strain. Genetic strain typing from collected S aureus isolates. Isolates recovered from inoculated pigs and identified as S aureus were genetically typed via repetitive sequence-based polymerase chain reaction methods using the DiversiLab repetitive sequence-based polymerase chain reaction strain typing system (bioMérieux, Marcy-l’Étoile, France). Strains were compared with the inoculated strain to determine whether the recovered S aureus isolates were the same as the inoculated ATCC strain (Fig 2). Histopathologic studies. At the aortic anastomosis site, circumferential slices (diameter, 1.0-2.5 cm) were obtained every 3 to 4 mm, fixed in formalin, and embedded in paraffin. Sections (5-mm thick) were stained with hematoxylin and eosin, Movat pentachrome, and Gram stain. Histopathologic and microbiologic examinations were performed with

Olympus MicroSuite FIVE Imaging Software connected to an Olympus BX61 microscope (Olympus America, Melville, NY). For each 5-mm section of the paraffin-embedded specimens, sections were examined for signs of inflammation in two low-power (magnification 4) fields of the vascular wall. Statistical analysis. A single, inoculated positive control animal was used to ensure that a positive culture could be obtained in an untreated animal. The culture was positive, and this animal was not further statistically evaluated. Because all values for the CFU variable were zero in the bonded-graft group, the normal distributional assumptions required for a two-sample t-test were unmet. Therefore, we used a nonparametric Wilcoxon rank sum test to compare the CFU counts in the unbonded vs bonded groups. The data were analyzed with SAS software, version 9.4 (SAS Institute, Cary, NC). RESULTS Survival and microbiologic analysis. All pigs survived surgery. The control pig showed low-level bacterial growth (1.2  103 CFUs per graft) with S cohnii spp urealyticus and S chromogenes, but not S aureus, suggestive of accidental direct perioperative contamination, with no significant clinical, microbiologic, or histologic signs of infection. All bonded grafts had an intact suture line

4 Aboshady et al

JOURNAL OF VASCULAR SURGERY --- 2015

Fig 2. After repetitive sequence-based polymerase chain reaction tests, the dendrogram and virtual banding patterns showed that the inoculated Staphylococcus aureus (SA) strain (ATCC 29213) recovered from the specimens was the same as the strain in the initial inoculum.

without gross or microscopic signs of infection after 8 weeks (Fig 3). The bonded grafts were fully incorporated into the perigraft soft tissue. We take this as indirect evidence of residual antibacterial activity at 8 weeks, at least against accidental bacteremia. Two of the three pigs in group 2 (inoculated, unbonded) were euthanized after showing signs of severe postoperative bloodstream infection and septicemia (fever, malaise, increased total leukocyte count, reduced food intake, and altered behavior) at 1 and 2 weeks, respectively. On excision, both grafts were torn loose; portions of the aortic wall were missing across from the graft, and extensive intra-abdominal hemorrhage was present. Histologic and microbiologic examination confirmed the clinical signs and showed moderate adventitial thickening in the graft’s vicinity. At microscopic examination, both grafts, especially that from pig 2, were lined by a thick layer of fibrin strands, in which small clusters of bacterial colonies (Gram-positive cocci) were embedded at various points. Vascular engorgement and hemorrhagic infiltration were present. Mild to moderate fibrin deposition was observed in the outer layers. These findings (Figs 4 and 5) resembled those in group 2 of our pilot study.13 In the current group 2, S aureus isolated from the culture of the grafts in two pigs had a high bacterial count (6.56  108 and 7.60  106 CFUs per graft, respectively; P < .001). The third pig in group 2 showed no severe clinical, microbiologic, or histologic signs of infection; bacterial counts were significantly lower for its excised tissues (1.8  103 CFUs per graft)

Fig 3. Ex vivo aortic graft from one pig (group 3) that received a bonded graft with the triple antimicrobial agents. The graft shows no macroscopic signs of infection.

than for specimens of the S aureus-treated grafts from the pigs euthanized early (Fig 6; Table). All S aureus-treated bonded grafts (n ¼ 5) showed no bacterial growth for 8 weeks at the time of microbiologic examination. Graft specimens from groups 2 and 3 were

JOURNAL OF VASCULAR SURGERY Volume -, Number -

Aboshady et al 5

Fig 4. Grafted infrarenal aorta of pig 2 (group 2, unbonded grafts). A, Movat stain showing a cross-section of the aorta containing a Dacron graft (g). B, Higher magnification, hematoxylin and eosin stain of the inset area marked in (A), showing a region of the aortic wall (a) where the Dacron graft (g) was attached. A mixture of acute and lytic thrombus (*) covers the luminal surface of the graft. Neutrophilic infiltration is present within the thrombus and in the media of the adjacent aortic wall (white arrows). A circular space left by a removed suture can be seen at the aorta-graft junction (black arrow). C, Higher magnification Gram stain of the inset in (A), showing large clusters of bacterial colonies (stained dark blue) within the thrombus. Bar ¼ 2 mm in (A), 1 mm in (B), and 100 mm in (C) (original magnification 1, 4, and 40, respectively).

Fig 5. Histologic sample from pig 3 (group 2, unbonded grafts). A, The luminal surface of the Dacron graft (g) is covered by a thick layer of thrombus (t) seeded with abundant bacterial colonies (arrows). B, Higher magnification of the inset in (A) showing large clusters of bacterial colonies (stained blue) within the graft and in the adjacent thrombus. Stains: hematoxylin and eosin in (A), Gram stain in (B). Bar ¼ 1 mm in (A) and 100 mm in (B) (original magnification 4 and 40, respectively).

infected with the same strain. The dendrogram and virtual banding patterns showed that all S aureus isolates from inoculated pigs were the same strain used in the initial inoculum (Fig 2). Bacterial growth from the ex vivo graft material and surrounding tissues of the control pig (group 1) indicated direct accidental perioperative contamination. The organisms from the control pig were S cohnii spp urealyticus and S chromogenes, not S aureus. The CFU counts for the bonded vs unbonded groups differed significantly according to the Wilcoxon rank-sum test (P < .017). Histopathologic findings. Histopathologic findings were generally concordant with microbiologic ones. For the two group 2 animals euthanized early, histologic examination confirmed the clinical signs and microbiologic results. At the time of microscopy, the luminal surface of both grafts was covered with mixed acute and lytic thrombus, seeded with numerous small clusters of bacterial colonies. Portions of the arterial wall proximal to the graft showed moderate neutrophilic infiltration, particularly in

the media (Figs 4 and 5). The adventitia showed acute and chronic inflammatory infiltration. Vascular engorgement and hemorrhagic infiltration were also present. These findings resembled those in group 2 of our 2-week pilot study.13 Of the animals that completed the 8 weeks of the current study, all had a relatively similar healing pattern (Figs 7 and 8). All presented disruption of the arterial wall, with interruption of the endothelial, medial, and adventitial layers at the graft site and marked neointimal formation over the aortic defect. Dacron graft fragments were surrounded by a marked granulomatous foreign body response with multinucleated giant cell formation. Chronically inflamed fibrovascular tissue was common at the outer margins. The adventitia showed chronic inflammation and thickening, marked collagen deposition, and mild to moderate vascularization. Clusters of hemosiderin-laden macrophages were occasionally observed.

JOURNAL OF VASCULAR SURGERY --- 2015

6 Aboshady et al

1.0E+09 1.0E+08 1.0E+07 1.0E+06 1.0E+05 1.0E+04

CFU/tissue

1.0E+03 1.0E+02 1.0E+01 1.0E+00

Fig 6. Histogram of microbiologic results. Samples from pigs 2 and 3 (group 2, unbonded grafts) showed significantly greater bacterial growth than samples from pig 1 (group 1, control), pig 4 (group 2, unbonded grafts), and pigs 5 to 9 (group 3, bonded grafts). CFU, Colony-forming unit.

Table. Microbiologic findings from cultured aortic grafts

Group Pig

Graft

Bacterial colony count (CFUs per graft)

Isolated bacteria

1

1 Unbonded, uninoculated

1.2  103

2

2 Unbonded, inoculated 3 Unbonded, inoculated 4 Unbonded, inoculated

6.56  108

5 M/RCHX inoculated 6 M/RCHX inoculated 7 M/RCHX inoculated 8 M/RCHX inoculated 9 M/RCHX inoculated

0

Staphylococcus cohnii spp urealyticus and S chromogenes S aureus (ATCC 29213) S aureus (ATCC 29213) S cohnii spp urealyticus and S chromogenes e

0

e

0

e

0

e

0

e

2 2 3 3 3 3 3

7.60  106 1.8  103

CFU, Colony-forming unit; M/RCHX, chlorhexidine, rifampin, and minocycline.

Most intergroup differences were observed in the inflammatory infiltrate in the aortic wall near the graft site (Fig 9). The infected animal (pig 7, group 2) showed a more marked neutrophilic presence and a very small amount of bacterial colonization within the graft

fragments. No bacteria were revealed using Gram stain in groups 1 or 3. DISCUSSION Aortic graft infection is a dreaded complication after arterial surgery.16 In vitro studies have shown that antibiotic-soaked grafts might be efficacious in prevention of infection.6,17-19 Our novel triple-antimicrobial-bonded graft completely prevented perioperative aortic graft infections in the original pilot study13 (six pigs followed for 2 weeks) and the current study (nine pigs followed for 9 weeks). Russu et al20 used the classification proposed by Szilagy et al21 to characterize graft infections according to their appearance time (<4 months vs >4 months after graft implantation), relationship to the postoperative wound, and extent of involvement. The resulting grades are: I, cellulitis involving the wound; II, infection involving subcutaneous tissue; and III, infection involving the vascular prosthesis. An early infection correlates with Szilagyi grade III and is caused by virulent hospital-acquired bacteria. Patients have signs of sepsis (fever, leukocytosis, bacteremia, etc) and wound infection (inflamed surrounding tissues, pus seepage, etc). They have extensive medial and adventitial infiltration of neutrophils near the graft and have bacterial colonies within the thrombus that lines the graft. These findings were prominent in our two pigs that received unbonded grafts compared with those that received bonded grafts and the control animal. This explains why the two pigs had to be euthanized early because of septic shock.22,23 Legout et al24 reported similar observations and concluded that fever, septic shock, and nonuse of

JOURNAL OF VASCULAR SURGERY Volume -, Number -

Aboshady et al 7

Fig 7. Aortic tissue sample from the control pig in group 1. A, Movat stain of a cross-section of a portion of the aortic defect repaired with a Dacron graft (arrow). The site of graft implantation shows disruption of the normal arterial wall architecture. The defect was filled by myofibrotic tissue (*). A foreign-body granulomatous response surrounded the graft. B, Gram stain of a portion of the graft (g) and adjacent myofibrotic tissue (mf) shows no bacterial growth. Bar ¼ 2 mm in (A) and 100 mm in (B) (original magnification 1 and 40, respectively).

Fig 8. Photomicrographs of aortic tissue samples from a pig in group 3. A, Movat stain of a cross-section of a portion of infrarenal aorta repaired with a Dacron graft (arrows). The graft fragments are surrounded by marked foreign-body granulomatous inflammation (white asterisks); myofibrotic tissue growth filled the defect on the luminal side (black asterisks). The adventitial layer (a) displayed chronic inflammation and thickening, with marked collagen deposition. B, Gram stain of a portion of the graft (g) and adjacent chronic inflammation (c) revealed no bacteria. Bar ¼ 2 mm in (A) and 100 mm in (B) (original magnification 1 and 40, respectively).

antibiotic regimens containing rifampin entail poor outcomes and treatment failure in staphylococcal graft infection. We used Sinclair miniature pigs because their vascular system has many similarities to that of humans, including anatomy and electrophysiology. This can expedite the transition from research to clinical application. Also, their abdominal aorta can accommodate the type and diameter of our graft.25,26 Faress et al27 developed a swine model for studying vascular prostheses while comparing the infectability of polyester (Dacron) and glutaraldehyde-treated bovine pericardium in a topically infected environment. The researchers found no significant difference between Dacron and glutaraldehyde-treated bovine pericardium with regard to resistance of bacterial infection at 3 weeks. Rifampin has excellent bactericidal activity against MRSA.17 However, using it as a single agent could result

in rapid development of resistance. That is why we used it in combination with two other antimicrobial agents; each works with a different mechanism of action, and thus reduces the potential for development of resistance. In a recent study of central venous catheter (CVC) infections,28 we tested a novel second-generation catheter coated with M/RCHX. Central venous and peripherally inserted central catheters (PICCs) were impregnated with chlorhexidine-minocycline/rifampin (CHX-M/R) and compared with first-generation M/R catheters, chlorhexidine/ silver-sulfadiazine-treated CVCs, chlorhexidine-treated PICCs, and uncoated catheters. A biofilm catheter colonization model was used to assess their efficacy against MRSA, vancomycin-resistant Enterococcus faecium (VRE), Pseudomonas aeruginosa, Candida albicans, and C glabrata. The CHX-M/R-impregnated device was the only antimicrobial catheter that completely inhibited biofilm

8 Aboshady et al

JOURNAL OF VASCULAR SURGERY --- 2015

Fig 9. Inflammatory response. A and B, Group two, unbonded graft: neutrophilic infiltration in the media and adventitia, respectively, in a region near the graft in pig 2. C, Group two, unbonded graft: abundant collection of degenerating neutrophils (n) within the thrombus overlying the graft (g) in pig 3. Note a large cluster of bacterial colonies (b). D-F, Inflammatory infiltrates surrounding the graft at 8 weeks in pigs from groups 2, 1, and 3, respectively, are composed mostly of neutrophils (n), mononuclear chronic inflammatory cells (c), and multinucleated giant cells (arrows). F, Chronically inflamed advanced granulation tissue. Hematoxylin and eosin stain. Bar ¼ 200 mm in (B) (original magnification 20).

colonization of all resistant bacterial and fungal organisms tested at all time intervals. It significantly surpassed the uncoated catheters (all P values # .003). The study showed that the novel CHX-M/R catheter has unique properties for completely inhibiting biofilm colonization of MRSA, VRE, P aeruginosa, and fungi in a manner superior to that of the M/R- and chlorhexidine-treated catheters. In another similar study, the same group showed that the CHX-M/R treated CVC completely prevented the biofilm colonization of multiantibiotic resistant Gram-negative bacteria in a manner superior to other antimicrobial catheters.29 When an aortic graft infection is clinically suspected, it might be beneficial to use intravenous antibiotics to reinforce prevention and/or treatment. The clinical end points for assessment of the treatment’s efficacy should include leukocytosis, increased C-reactive protein levels, peripheral septic embolization, septic shock, ileus, hematemesis, hematuria, or abdominal distension. Radiologic end points should include contrast-enhanced computed tomography (CT), ultrasonography, upper gastrointestinal endoscopy, or ureteroscopy to confirm the presence of a graft-ureteral fistula. Fluorodeoxyglucose-(18F)-positron emission tomography-CT, indium 111-labeled white blood cell-scan approaches, or CT-guided aspiration of cavitary perigraft fluid collections also might be helpful. For patients already colonized with MRSA/VRE or other resistant pathogens, treatment should involve excision of the colonized prosthetic graft, in situ implantation of the

bonded graft, and coverage with greater omentum, followed by 8 weeks of intravenous antibiotics. Rifampin-treated Dacron endografting is a logical extension of the principles of open Dacron graft replacement and might improve outcomes in patients unsuited for definitive open surgery.30 Escobar et al29 recently described exposure of Dacron endografts with rifampin delivered via an injection port or via the sheath before graft deployment in selected patients with aortic infections. On the basis of rifampin use in open vascular repairs, treatment of Dacron endografts with rifampin might add similar antimicrobial resistance against selected aortic infections. Using rifampin as a single agent could result in the rapid development of resistance to this antibiotic; this is why we combined it with two other antimicrobial agents, each of which works with a different mechanism of action. For all organisms, bacterial adherence to the prosthetic graft is the initial event in the process of infection. Adherence depends on the graft’s physical characteristics, such as pore size and surface area, as well as chemical properties, such as hydrophobicity. Production of an extracellular glycocalyx (mucin, slime) by staphylococci promotes adherence to biomaterials and protects against host defenses. This biofilm decreases antibiotic penetration and impairs phagocyte and antibody functions but will stimulate a chronic inflammatory process around an infected prosthesis.31 As many as 60% of microbial infections are related to biofilm formation.4,9 A biofilm comprises an intricate

JOURNAL OF VASCULAR SURGERY Volume -, Number -

community of microorganisms that are embedded in extracellular polymeric substancesdprimarily polysaccharides, proteins, and nucleic acids. The community retains nutrients for its component cells and protects them from the host immune response and antimicrobial agents. On indwelling catheters, biofilm formation entails a complex process characterized by primary microbial adherence followed by cell to cell interaction, microcolony formation, and development of a mature biofilm. To test for antimicrobial efficacy and durability in a widely used model of biofilm colonization,22 we created an innovative gendine (gentian violet with chlorhexidine)-coated PICC, then used highly pathogenic, resistant organisms obtained from our hospital to evaluate the device. We also compared the catheter with commercial M/R and CHX PICCs.32 Moreover, to test the gendine catheter for acute biocompatibility in a rabbit intravascular model, we performed histopathologic studies and evaluated the pharmacokinetics of circulating antiseptic levels in blood by using liquid chromatography/mass spectrometry. In a retrospective, comparative study from 1996 to 2006, Larena-Avellaneda et al33 described the prophylactic use of a silver acetate-coated graft (InterGard Silver polyester graft; Intervascular, Datascope, Inc, La Ciotat, France) in arterial occlusive disease. They concluded that the silver-coated prosthesis did not differ from standard materials. Silver had no significant effect on the risk of graft infection. Also, Jeanmonod et al34 described early animal implantation of a silver acetate-coated Dacron arterial graft not yet approved by the United States Food and Drug Administration. They analyzed the biocompatibility and vascularization of uncoated (Dacron) grafts vs silver acetate-coated (Dacron Silver) grafts after implantation into the dorsal skinfold chamber of C57BL/6 mice. The researchers used repetitive intravital fluorescence microscopy to study angiogenesis and leukocytic inflammation at the implantation site over a 14-day period. Afterward, histology and immunohistochemistry methods were used to analyze collagen formation, apoptosis, and cell proliferation in the newly developed granulation tissue surrounding the implants. Compared with uncoated grafts, the silver acetate-coated grafts exhibited improved vascularization, indicated by a significantly increased functional capillary density. This finding was not associated with a stronger leukocytic inflammatory host-tissue response to the implants. Moreover, the silver-acetate coating did not affect collagen formation, apoptosis, and cell proliferation at the implantation site. Therefore, material modification might improve incorporation of implants into host tissue and decrease the risk of infection. However, past experience with silver nitrate-coated prosthetic heart valves (Silzone valves; St Jude Medical, St Paul, Minn) in treating native valve endocarditis has shown that major paravalvular leaks that required device explantation or redo surgery or that were implicated in death occurred in 18 of 403 patients with Silzone valves and 4 of 404 patients without Silzone valves (2-year

Aboshady et al 9

event-free rates, 91.1% for silver vs 98.9% for conventional valves; P ¼ .003).35 The mechanism responsible for the leaks was not culture-negative endocarditis. The silverimpregnated valves explanted from AVERT (Artificial Valve Endocarditis Reduction Trial) patients exhibited poor tissue ingrowth and loosening of sutures. Thus, the Silzone coating appeared to inhibit the normal fibroblastic response and incorporation of the sewing-ring fabric into the host tissues of some patients. Butany et al36 reported pathologic findings from 19 Silzone valves explanted from patients with 176 such valves. Many sewing cuffs showed large regions of pannus, granulation tissue, and purulent exudates, but no positive blood cultures or microorganisms were observed with special stains. Atypical tissue incorporation with partial dehiscence from the native annulus might have been due to fibroblast infiltration or silver allergy. For these reasons, we prefer our current model of triple antimicrobial graft bonding because it avoids the silver ion. CONCLUSIONS Our triple-bonded aortic graft prevented perioperative aortic graft infection for at least 8 weeks in pigs. The synergistic antimicrobial activity prevented and/or eradicated infection during that period, suggesting that it could potentially have done so for even longer. Further animal studies and a collaborative clinical trial are needed to assess the graft’s ability to combat early-onset graft infections for up to 4 months and to set the final recommendations before initiating protocol-controlled human implantations. Limitations. We focused on early graft infection, partly because it occurs in most cases (53%-70%). Future studies will assess our graft’s ability to prevent infection up to 4 months or longer. Our pigs received no perioperative antibiotics, because we intended to subject our hypothesis to the maximum challenge. Although perioperative antibiotics are widely used and might reduce the incidence and severity of infections, they do not completely protect from early graft infection. Finally, we used only one control animal and will add more in future studies. AUTHOR CONTRIBUTIONS Conception and design: IA, IR, LB, KK Analysis and interpretation: IA, IR, LB, KK Data collection: IA, DV, MH, KK, YA Writing the article: IA, YA Critical revision of the article: IR, HS, LB, KK Final approval of the article: IA, IR, DV, MH, HS, LB, KK Statistical analysis: IA Obtained funding: LB Overall responsibility: KK REFERENCES 1. Alankar S, Barth MH, Shin DD, Hong JR, Rosenberg WR. Aortoduodenal fistula and associated rupture of abdominal aortic aneurysm after endoluminal stent graft repair. J Vasc Surg 2003;37:465-8.

10 Aboshady et al

2. O’Hara PJ, Hertzer NR, Beven EG, Krajewski LP. Surgical management of infected abdominal aortic grafts: review of a 25-year experience. J Vasc Surg 1986;3:725-31. 3. Reilly L. Aortic graft infection: evolution in management. Cardiovasc Surg 2002;10:372-7. 4. Nasim A, Thompson MM, Naylor AR, Bell PR, London NJ. The impact of MRSA on vascular surgery. Eur J Vasc Endovasc Surg 2001;22:211-4. 5. Chiesa R, Astore D, Frigerio S, Garriboli L, Piccolo G, Castellano R, et al. Vascular prosthetic graft infection: epidemiology, bacteriology, pathogenesis and treatment. Acta Chir Belg 2002;102:238-47. 6. Lew W, Moore W. Antibiotic-impregnated grafts for aortic reconstruction. Semin Vasc Surg 2011;24:211-9. 7. Goeau-Brissonniere O, Javerliat I, Koskas F, Coggia M, Pechere JC. Rifampin-bonded vascular grafts and postoperative infections. Ann Vasc Surg 2011;25:134-42. 8. Perera GB, Fujitani RM, Kubaska SM. Aortic graft infection: update on management and treatment options. Vasc Endovascular Surg 2006;40: 1-10. 9. Cronenwett JL, Johnston KW, Rutherford RB. Local complications: graft infection. In: Cronenwett JL, Johnston KW, editors. Rutherford’s vascular surgery. 7th ed. Philadelphia: Saunders/Elsevier; 2010. p. 643-61. 10. McKinsey JF. Extra-anatomic reconstruction. Surg Clin North Am 1995;75:731-40. 11. O’Connor S, Andrew P, Batt M, Becquemin JP. A systematic review and meta-analysis of treatments for aortic graft infection. J Vasc Surg 2006;44:38-45. 12. Coselli JS, Crawford ES, Williams TW Jr, Bradshaw MW, Wiemer DR, Harris RL, et al. Treatment of postoperative infection of ascending aorta and transverse aortic arch, including use of viable omentum and muscle flaps. Ann Thorac Surg 1990;50:868-81. 13. Aboshady I, Raad I, Shah AS, Vela D, Dvorak T, Safi HJ, et al. A pilot study of a triple antimicrobial-bonded Dacron graft for the prevention of aortic graft infection. J Vasc Surg 2012;56:794-801. 14. Shenk NS, Ney AL, Tsukayama DT, Olson ME, Bubrick MP. Tobramycin-adhesive in preventing and treating PTFE vascular graft infections. J Surg Res 1989;47:487-92. 15. Alfonsi P, Coggia M, Leflon-Guibout V, Sessler DI, GoeauBrissonniere O, Chauvin M. Mild hypothermia does not increase bacterial proliferation on implanted vascular grafts. Am J Surg 2002;184:37-40. 16. Quinones-Baldrich WJ, Hernandez JJ, Moore WS. Long-term results following surgical management of aortic graft infection. Arch Surg 1991;126:507-11. 17. Charlton-Ouw KM, Kubrusly F, Sandhu HK, Swick MC, Leake SS, Gulbis BE, et al. In vitro efficacy of antibiotic beads in treating abdominal vascular graft infections. J Vasc Surg 2015;62:1048-53. 18. Schneider F, O’Connor S, Becquemin JP. Efficacy of collagen silvercoated polyester and rifampin-soaked vascular grafts to resist infection from MRSA and Escherichia coli in a dog model. Ann Vasc Surg 2008;22:815-21. 19. Healy AH, Reid BB, Allred BD, Doty JR. Antibiotic-impregnated beads for the treatment of aortic graft infection. Ann Thorac Surg 2012;93:984-5. 20. Russu E, MureS¸an A, Grigorescu B. Vascular graft infections management. Management in Health 2011;15:16-9.

JOURNAL OF VASCULAR SURGERY --- 2015

21. Szilagyi DE, Smith RF, Elliott JP, Vrandecic MP. Infection in arterial reconstruction with synthetic grafts. Ann Surg 1972;176:321-33. 22. Mussa FF, Hedayati N, Zhou W, El-Sayed HF, Kougias P, Darouiche RO, et al. Prevention and treatment of aortic graft infection. Expert Rev Anti Infect Ther 2007;5:305-15. 23. Kalman PG, Rappaport DC, Merchant N, Clarke K, Johnston KW. The value of late computed tomographic scanning in identification of vascular abnormalities after abdominal aortic aneurysm repair. J Vasc Surg 1999;29:442-50. 24. Legout L, Delia P, Sarraz-Bournet B, Rouyer C, Massongo M, Valette M, et al. Factors predictive of treatment failure in staphylococcal prosthetic vascular graft infections: a prospective observational cohort study: impact of rifampin. BMC Infect Dis 2014;14:228. 25. Hughes HC. Swine in cardiovascular research. Lab Anim Sci 1986;36: 348-50. 26. Nunoya T, Shibuya K, Saitoh T, Yazawa H, Nakamura K, Baba Y, et al. Use of miniature pig for biomedical research, with reference to toxicologic studies. J Toxicol Pathol 2007;20:125-32. 27. Faress JA, McKinney LA, Semaan MT, Byrd RP Jr, Mehta JB, Roy TM. Mycobacterium xenopi pneumonia in the southeastern United States. South Med J 2003;96:596-9. 28. Raad I, Mohamed JA, Reitzel RA, Jiang Y, Raad S, Al Shuaibi M, et al. Improved antibiotic impregnated catheters with extended spectrum activity against resistant bacteria and fungi. Antimicrob Agents Chemother 2011;56:935-41. 29. Escobar GA, Eliason JL, Hurie J, Arya S, Rectenwald JE, Coleman DM. Rifampin soaking dacron-based endografts for implantation in infected aortic aneurysmsenew application of a time-tested principle. Ann Vasc Surg 2014;28:744-8. 30. Golzarian J, Struyven J. Imaging of complications after endoluminal treatment of abdominal aortic aneurysms. Eur Radiol 2001;11: 2244-51. 31. Kitamura T, Morota T, Motomura N, Ono M, Shibata K, Ueno K, et al. Management of infected grafts and aneurysms of the aorta. Ann Vasc Surg 2005;19:335-42. 32. Jamal MA, Hachem RY, Rosenblatt J, McArthur MJ, Felix E, Jiang Y, et al. Development of antimicrobial gendine coated central catheters: in vivo biocompatability and in vitro efficacy. Antimicrob Agents Chemother 2015;59:5611-8. 33. Larena-Avellaneda A, Russmann S, Fein M, Debus ES. Prophylactic use of the silver-acetate-coated graft in arterial occlusive disease: a retrospective, comparative study. J Vasc Surg 2009;50:790-8. 34. Jeanmonod P, Laschke MW, Gola N, von Heesen M, Glanemann M, Dold S, et al. Silver acetate coating promotes early vascularization of Dacron vascular grafts without inducing host tissue inflammation. J Vasc Surg 2013;58:1637-43. 35. Schaff HV, Carrel TP, Jamieson WR, Jones KW, Rufilanchas JJ, Cooley DA, et al. Paravalvular leak and other events in silzone-coated mechanical heart valves: a report from AVERT. Ann Thorac Surg 2002;73:785-92. 36. Butany J, Leask RL, Desai ND, Jegatheeswaran A, Silversides C, Scully HE, et al. Pathologic analysis of 19 heart valves with silvercoated sewing rings. J Card Surg 2006;21:530-8.

Submitted May 1, 2015; accepted Sep 3, 2015.