- Email: [email protected]
Prosthetic Graft Patency in the Setting of a Polymicrobial Infection in Swine (Sus scrofa)
Prosthetic Graft Patency in the Setting of a Polymicrobial Infection in Swine (Sus scrofa) Mamie C. Stull,1,2 Michael S. Clemens,1,2 Thomas A. Heafner,1 John Devin B. Watson,1 Zachary M. Arthurs,1 and Brandon W. Propper,1 Sam Houston, Texas
Background: In the setting of vascular injury, vein interposition graft is the preferred conduit, but may have limited availability. This study seeks to develop a large animal model assessing the graft performance of polytetrafluoroethylene (PTFE) and Dacron in the setting of a polymicrobial infection. Methods: Thirty-seven animals were placed into 4 groups for a 21-day survival period. Sixmillimeter PTFE or Dacron interposition grafts were placed in the right iliac artery with a standardized bacterial inoculation. Native vessel with and without contamination served as control groups. The inoculant was 1 107 of genetically labeled Pseudomonas aeruginosa and Staphylococcus aureus. The primary end points were graft patency (determined by duplex ultrasound and necropsy) and graft infection (culture with molecular analysis). Secondary end points included physiological measurements, blood cultures, laboratory data, and histopathology. Results: PTFE and Dacron had similar infection rates of 85.7% and 75%, respectively. There was no significant difference in infectious organisms between graft materials. PTFE and Dacron exhibited bacterial ingrowth and transmigration to the intraluminal portion of the conduit. Fortyfive percent of the Dacron group and 40% of the PTFE group remained patent at postoperative day 21 (P ¼ 0.98). Clinical data, including white blood cell count, percent neutrophils, and lactate, did not vary significantly between groups. Conclusions: PTFE and Dacron perform similarly in terms of infection rates and graft failure as both have a propensity toward bacterial ingrowth and occlusion when compared with controls. This is a valid animal model to assess graft performance in the setting of polymicrobial infection and provides an avenue for studying novel prosthetic conduits.
INTRODUCTION
Presented at the Poster Presentation Competition and won first place in its category at the 2015 Vascular Annual Meeting of the Society for Vascular Surgery, Chicago, IL, June 17e20, 2015. Conflicts of Interest: none. 1
San Antonio Military Medical Center, Sam Houston, TX.
2
Clinical Research Division, 59th Medical Wing, Joint Base San Antonio e Lackland, Sam Houston, TX. Correspondence to: Mamie C. Stull, MD, ATTN: General Surgery, 3551 Roger Brooke Drive, Fort Sam Houston, TX 78234, USA; E-mail: [email protected] Ann Vasc Surg 2016; 36: 265–272 http://dx.doi.org/10.1016/j.avsg.2016.05.089 Published by Elsevier Inc. Manuscript received: February 28, 2016; manuscript accepted: May 22, 2016; published online: 15 July 2016
Autologous vein grafts remain the preferred conduit for vascular reconstruction in the setting of a contaminated field.1e3 However, circumstances remain where autologous vein is either unavailable or is a less desirable option. One common scenario is the trauma patient where prosthetic materials offer an expedient reconstructive option in the physiologically tenuous patient or when concomitant venous injury eliminates vein as a possible conduit. In these situations, surgeons must generally resort to ‘‘off-the-shelf’’ alternatives such as polytetrafluoroethylene (PTFE) or Dacron grafts. While these grafts have generally shown adequate patency in proximal vasculature, prosthetic grafts have a lower primary and 265
266 Stull et al.
primary-assisted patency and carry a 4.3e4.7% risk of infectious complications.4e6 Treatment options for patients with infected grafts are limited to explantation with extra-anatomic bypass, in-line repair with autologous vein or prosthetic material, or conservative management with graft preservation. All options are plagued with the risk of recurrent infection, limb loss, and mortality.7 The combat environment further adds unique challenges to vascular reconstruction including a high frequency of explosive injuries, lack of available equipment or expertise in far-forward situations, need for rapid aeromedical evacuation, and ubiquitous revascularization in a contaminated field.8 Combat surgeons have been attempting to address traumatic vascular reconstruction since the Korean War when Hughes’s report described the attempted repair of 269 of 304 major vessel injuries.9 However, White et al.8 identified a 12% rate of vascular injury during the wars in Afghanistan and Iraq from 2002 to 2009, nearly 5 times higher than previous conflicts. Due to a 73% rate of explosive injury, many of these patients presented with extensive soft tissue loss, simultaneous traumatic amputations, and limited options for venous reconstruction. Between 2% and 15% of combat-related vascular injuries required reconstruction with prosthetic grafts.10e12 These studies noted that infections caused 40% of graft-related complications, prompting the authors to suggest that prosthetic grafts be used as a bridge therapy to definitive reconstruction in a combat environment.11,12 Thus, the need to identify a safe, efficacious, off-the-shelf alternative to autologous vein for vascular reconstruction has been illuminated by the vascular injuries in the recent conflicts. The purpose of this study was to create a large animal model in swine (Sus scrofa) to test various prosthetic vascular conduits, specifically PTFE and Dacron, in the setting of a polymicrobial infection. The primary end points include 21-day graft patency and rate of graft infection. The secondary end points include laboratory and histological data.
METHODS The United States Air Force 59th Airwing Clinical Research Division’s Institutional Animal Care and Use Committee approval was obtained before study initiation. All protocols were performed at the Clinical Research Division, 59th Medical Wing at Joint Base San Antonio-Lackland, Texas, an Association for Assessment and Accreditation of Laboratory Animal Care-accredited facility. Animals were treated
Annals of Vascular Surgery
in accordance with The Guide for the Care and Use of Laboratory Animals (National Research Council 199). Animal Model Female Yorkshire-Landrace swine (S. scrofa) weighing between 70 and 90 kg were acclimated in the facility for 7 days before surgery. A total of 37 animals were placed into 4 groups: 10 animals received a 6mm heparin-bonded, expanded PTFE interposition graft, 11 animals received a 6-mm Dacron interposition graft, 8 animals received a sham surgery with inoculation (Sham[+]), and 8 animals received a sham surgery without inoculation (Sham[]). Sham procedures included full retroperitoneal exposure of the iliac system and anesthetic times comparable with those of the surgical groups. Preparation Before the procedure, 325 mg of oral aspirin was administered daily, for a median of 6 days (range 5e15). On the day of surgery, the animals were premedicated with ketamine 10e20 mg/kg, acepromazine 0.1 mg/kg, and atropine 0.04e0.4 mg/kg. The animals were then given buprenorphine 0.01 mg/ kg intramuscularly and intubated. Anesthesia was maintained with isoflurane 2e3%. A urinary catheter was placed for the duration of the operation. Surgical Protocol Following preoperative preparation, the animals were brought to the operating room where strict aseptic conditions and techniques were maintained during the entirety of the procedure. A right external jugular venous line was placed and baseline labs obtained. The animals were given 100 mL/hr of Lactated Ringers throughout the procedure. Bilateral femoral velocities were assessed with ultrasound. A midline incision was made and a retroperitoneal dissection was utilized to expose the right external iliac artery. All animals received 75 U/kg of systemic heparin before arterial clamping with atraumatic vascular clamps. In the animals that received a prosthetic graft, approximately 3 cm of the artery was excised. The interposition graft was then anastomosed using 60 prolene suture with a running stitch. Before completion, forward and back-bleeding was confirmed and the distal vasculature flushed with heparinized saline. Flow through the graft was confirmed using intraoperative Doppler. Heparinization was maintained with 1000 U every 30 min. Figure 1 shows the interposition graft once placed.
Volume 36, October 2016
Graft Patency in the Setting of Infection in Swine
267
excised. If a graft thrombosed before POD 3, it was excised for gross inspection and labeled as a technical failure. Graft/Native Artery Excision
Fig. 1. Representative image of the retroperitoneal exposure and insertion of the 6-mm PTFE interposition graft.
After the animals were anesthetized, sterile technique was utilized to reopen the surgical site and expose the implanted prosthetic graft. The surrounding tissue was extracted from the proximal native artery to the distal anastomosis. After explantation, the animal was euthanized with Beuthanasia-D Special (Schering-Plough Animal Health Corp., Union, NJ). The proximal anastomosis with surrounding tissue was sent for histology and the distal anastomosis was sent for microbiological studies. Bacterial Challenge Preparation
Following exposure of the artery in the control animals, 75 U/kg of heparin was given 5 min before occlusion. Proximal and distal vascular clamps were placed on the artery to occlude a 3 cm portion of the right external iliac artery. Heparinized saline was placed in the field during the 30 min of occlusion. For all animals, once the occlusion period concluded or the graft was successfully interposed, the surgical site was irrigated with 1 L of warmed saline. For the sham surgery group with inoculation (Sham[+]), PTFE group, and Dacron group, the surgical site was contaminated with 1 mL each of 1 107 of genetically labeled Pseudomonas aeruginosa and Staphylococcus aureus. For all groups, the fascia was then closed with a running polydioxanone suture. The subcutaneous tissue was approximated with a running 2-0 vicryl suture and staples were used to close the skin. Bacitracin was spread over the incision. The pigs recovered from anesthesia and returned to their cages. Pain medication was administered as needed by the veterinary staff (buprenorphine 0.01e0.05 mg/kg intramuscular). Postoperative Care and Evaluation The animals received 325 mg oral aspirin daily following the operation. The pigs were evaluated clinically using the Tarlov gait scale, laboratory data, and with duplex ultrasound evaluation of the bilateral femoral arteries and interposition grafts on postoperative days (POD) 1, 3, 7, 10, 14, 17, and 21. Doppler was used to assess flow and velocity within the graft. On POD 21 or at the earliest demonstration of graft occlusion by ultrasound, the animals were sacrificed and the grafts were
S. aureus (ATCC BAA1025) and P. aeruginosa (PAOIUW) were cultured overnight in 30 mL Thioglycollate medium at 35 C, 5% CO2. One milliliter of the bacterial cultures was then centrifuged for 10 min. The supernatant was discarded and the bacterial concentrate was resuspended into 1 mL Microbial Freeze Drying Buffer (OPS Diagnostics 500-06, City) and pooled. The bacterial suspension was aliquoted into glass vials, freeze dried, and stored at 4 C. Lyophilized samples were periodically reconstituted to confirm bacterial concentration and viability. On the day of the protocol, lyophilized bacteria were reconstituted to a concentration of 1 107. One milliliter of Pseudomonas and Staphylococcus each was utilized to inoculate the animals. Microbiological Assays At the time of excision, the distal anastomosis was placed in 5 mL of Thioglycollate medium and incubated at 35 C, 5% CO2 for 24 hr. The medium was plated and streaked on blood agar plates (trypticase soy agar (TSA) with 5% sheep blood) and MacConkey agar plates. Gram stain, catalase test, coagulase test, and a VITEK 2 compact (BioMerieux, Inc., Durham, NC) were used to identify S. aureus. P. aeruginosa was identified by gram stain, the indole test, and oxidase test. All confirmed samples of S. aureus and P. aeruginosa underwent arbitrarily primed polymerase chain reaction analysis with molecular gel staining to confirm that the strains were identical to the inoculum (Fig. 2). Blood cultures were collected into in BacT/ ALERTÒ PF bottles and monitored on the BacT/ ALERT Microbial Detection System (BioM erieux,
268 Stull et al.
Annals of Vascular Surgery
Fig. 2. Representative gels from the arbitrarily primed polymerase chain reactions ran on the cultured organisms to ensure that the recovered organism was
consistent with the standardized inoculum. ‘‘I’’ indicates that the organisms are similar genetically and ‘‘D’’ indicates that the organisms differ.
Inc.). Positive blood cultures were identified using the same steps as used to determine positive graft cultures.
appropriate. Categorical variables were assessed with Fisher’s exact test and reported as proportions. Graft patency was plotted as a survival curve using a log-rank test. Clinical and laboratory variables measured at multiple time points throughout the experiment were analyzed with repeated measures 2-way ANOVA. Statistical software utilized was GraphPad Prism version 6.0f (La Jolla, CA).
Histopathological Studies The proximal anastomosis or a proximal portion of the native artery with the surrounding tissue was collected, fixed in 10% neutral buffered formalin, sectioned at 5 microns, and processed by conventional methods. The samples were stained with hematoxylin, eosin, and gram stains. A blinded, veterinary pathologist then assessed the samples for evidence of infection and inflammation. A graded scale of inflammation from 0 (none) to 5 (severe) was utilized to differentiate degrees of inflammatory changes. Statistical Analysis All continuous variables were reported as mean (standard deviation) or median (interquartile range [IQR]), and analyzed with one-way analysis of variance (ANOVA) or a KruskaleWallis test, as
RESULTS Clinical Assessments Baseline demographic variables were similar between groups (Table I). Throughout the experiment, no significant differences were observed between groups regarding heart rate (P ¼ 0.82), temperature (P ¼ 0.92), white blood cell (WBC) (P ¼ 0.32), or lactate (P ¼ 0.93). There was a significant difference between groups regarding the rise of neutrophil percentage (P < 0.01) and post hoc comparisons revealed a higher percentage of neutrophils in each arm that received inoculum (Dacron, PTFE,
Volume 36, October 2016
Graft Patency in the Setting of Infection in Swine
269
Table I. Baseline characteristics demonstrate homogeneity Parameter
PTFE
Dacron
Sham (+)
Sham ()
P valuea
Heart rate Temperature ( C) WBC % Neutrophils Lactate (mmol/L) Velocity (cm/sec)c
79 ± 17 36.5 ± 0.7 14 ± 3.5 35 ± 7.8 1.3 (1.0e1.5) 90 (78e113)
80 ± 17 36.6 ± 0.5 16 ± 2.5 27 ± 6.8 1.2 (0.7e1.6) 101 (84e118)
83 ± 12 36.7 ± 0.7 15 ± 5.1 27 ± 11.2 1.1 (0.8e1.9) 103 (95e127)
82 ± 17 36.8 ± 0.7 18 ± 4.5 26 ± 9.0 1.2 (0.7e1.4) 122 (94e152)
0.96 0.84 0.23 0.13 0.91b 0.40b
a
One-way ANOVA unless otherwise specified. KruskaleWallis test. c Right femoral artery velocity. b
Sham[+]) compared with the Sham[] group on POD 1. While the Dacron and Sham[+] groups returned to equivalent neutrophil percentage to the Sham[] group following POD 1, PTFE had a significantly elevated neutrophil count until POD 3 and subsequently all groups remained similar. Within each group, a significant rise and fall of heart rate, temperature, WBC, and percent neutrophils (P < 0.01, each) was noted over time until excision. In general, the animals remained mobile and failed to show clinical evidence of infection such as fever (>40 C), malaise, or reduced oral intake, with the exception of 2 animals with infected Dacron grafts. Intraoperatively, there was a median occlusive time of 52 min (IQR 30e82 min).
Fig. 3. Log-rank test of graft patency.
Overall, 45% of the Dacron group and 40% of the PTFE group remained patent until POD 21 (P ¼ 0.98; Fig. 3), demonstrating no significant difference between the 2 groups. Given that the 2 sham groups remained 100% patent at 21 days, there was a significant difference between the graft and sham groups (P < 0.01). In the Dacron group, 3 of 11 animals (27%) were sacrificed within the first 72 hr and labeled as technical failures, which were excluded from the survival analysis. In the PTFE group, 2 of 10 animals (20%) were lost to technical failure with no significant difference noted between graft groups. The median graft patency was 14 days for both the Dacron and PTFE animals. All grafts lost before 21 days were found to be grossly infected and thrombosed at the time of excision without evidence of a technical source of failure.
72 hr, Dacron demonstrated a 75% infection rate compared with PTFE, which demonstrated an 87.5% infection rate (P ¼ 0.67) All grafts that thrombosed within 72 hr also demonstrated bacterial growth at the time of excision. P. aeruginosa was cultured from 7 of 8 (87.5%) PTFE grafts and 6 of 8 (75%) Dacron grafts (P ¼ 1.0). S. aureus grew from 6 (75%) PTFE and 3 (37.5%) Dacron grafts (P ¼ 0.31). Neither the Sham[+] nor the Sham[] groups demonstrated infection with the standardized inoculum upon excision on POD 21. Contaminants universally grew from all 16 sham animals. Three of 8 Dacron grafts and 4 of 8 PTFE grafts also grew at least one or more contaminants consistent with the natural flora of S. scrofa. Positive blood cultures were only seen in 2 animals, both of which were implanted with a Dacron interposition graft. In each instance, P. aeruginosa was cultured the day before graft thrombosis.
Graft Infection
Histology
There was no significant difference in the overall percent of infection or the type of organism with which Dacron or PTFE was infected (Fig. 4). Of the prosthetic grafts that remained patent beyond
Multiple prosthetic grafts demonstrated bacterial ingrowth through the PTFE and Dacron (Fig 5). In particular, the 2 animals with positive blood cultures demonstrated large multifocal colonies of
Graft Patency
270 Stull et al.
Annals of Vascular Surgery
Fig. 4. (A) Overall comparison of percent infected in Dacron and PTFE groups. (B) Specific comparison of the type of infection between all groups.
bacteria on the luminal side of the graft. However, bacterial growth into the graft material was common in both the PTFE and Dacron groups. None of the animals in either the Sham[+] or Sham[] groups had bacteria present in the native artery or the surrounding tissue submitted for gram stain. There was a significantly greater degree of inflammation in the surrounding tissue for animals with prosthetic grafts compared with the control groups (Dacron 4.2 ± 0.6; PTFE ¼ 3.8 ± 0.7; Sham [+] ¼ 1.1 ± 1.2; Sham[] ¼ 0.8 ± 1.0, P < 0.01). The majority of inflammation was granulomatous in nature.
DISCUSSION The purpose of this study was to create a model of prosthetic infection mimicking that seen in a clinical setting using 2 of the standard prosthetic vascular conduits. We demonstrated a similar, <50% patency of PTFE and Dacron in the setting of a mixed S. aureus and P. aeruginosa infection in swine. Furthermore, there was an equivalent rate of infection in each group. This study establishes a new, large animal model of polymicrobial infection allowing for continued evaluation of novel
Fig. 5. (A) Gram stain of PTFE with bacteria visible within the graft. (B) Gram stain of Dacron with bacteria visible within the graft.
approaches to vascular reconstruction in a contaminated field. This model furthers the work of previous authors who have documented in laboratory and clinical settings the failure of prosthetic material in an infected field. In 1978, Rich and Spencer’s13 Vascular Trauma recognized the risk of infection associated with prosthetic graft material and definitively stated the superiority of autologous vein reconstruction. In 1983, Shah et al.14 reported an updated model in dogs, using 1 107 S. aureus and Escherichia coli in a 21-day survival model. In the setting of infection, the vein grafts were subject to transmural necrosis and perforation with devastating consequences compared with anastomotic dehiscence with prosthetic materials. However, these studies generally noted that venous reconstruction offered a greater overall patency rate and was more amenable to treatment with systemic antibiotics. Staphylococcus infections have been shown to grow into the synthetic fibers of PTFE within 7 days and to decrease the tensile strength of
Volume 36, October 2016
reconstructions by 30 days.15 Antibiotics often cannot penetrate the biofilms created around prosthetic material, which requires explantation for adequate sterilization. Our study utilized S. aureus and P. aeruginosa because of their ubiquitous prevalence and ability to form biofilms. In addition, utilizing genetically labeled bacteria allowed us to clearly delineate those infections created by inoculation versus infection inherent to the animals. As such, 75% of the Dacron grafts and 87.5% of PTFE grafts remained infected at graft excision. Neither material demonstrated resistance to overall infection or either organism. We demonstrated a 40% patency rate of PTFE and 45% patency of Dacron in the setting of polymicrobial infection. These rates are comparable with previous models, but offer the advantage of an ultrasound graft surveillance regimen to better define the time to occlusion.1,16 More importantly, these results are similar to the retrospective analyses of modern combat complications from prosthetic materials, improving the applicability of our model to that population.11,12 Watson et al.11 noted a 69% rate of graft-related complications at 8 years compared with 23% for autologous vein after peripheral reconstructions in combat. Based on this clinical evidence, we recognize the superiority of venous reconstruction and designed our model for 2e15% of those patients with limited options who require an off-the-shelf conduit.10e12 Although the survival time was limited to 3 weeks, we believe that this model creates a foundation for future studies. Even within the 21-day study period, a large proportion of our grafts demonstrated nontechnical failure at necropsy. Although none of these animals manifested a systemic infection, it is possible that a longer survival may have elicited this response. Most importantly, the Dacron and PTFE groups both proved to create a nidus from which a profound infection could begin. Furthermore, histopathological examination reveals migration of bacteria into the inner lumen. While very few animals became bacteremic, we believe that longer survival may have also yielded more positive blood cultures, systemic sepsis, and conduit thrombosis/dehiscence. This study was limited by the small numbers and short survival inherent to large animal models. Prolonged survival models in large animals are challenging and very resource intensive, while studies with less survival time may miss important clinical findings. We selected 3 weeks as it encompasses the 9- to 18-day time frame when graft infection is usually clinically significant, and it builds on historic precedent. In addition, this model does not
Graft Patency in the Setting of Infection in Swine
271
investigate a venous conduit arm as venous options are limited in swine due to lack of a saphenous vein. The internal jugular vein is the most comparable based on size, but would require an additional surgical site differing this from the other arms of the model. Future studies will likely investigate options for venous comparison. Despite these limitations, this model offers a unique in vivo comparison of synthetic materials with natural tissue reactions with infection and foreign bodies. Furthermore, there are several strengths to this model over prior animal studies, including a uniform bacterial inoculum, a graft surveillance regimen, and molecular confirmation of cultured organisms.
CONCLUSIONS With increasing rates of vascular infection, in both the military and civilian sector, the need is evident for a commercially available conduit that is inherently resistant to infection. This model provides a backbone from which future studies can be built to test novel vascular conduits. REFERENCES 1. Bricker DL, Beall AC, DeBakey ME. The differential response to infection of autogenous vein versus Dacron arterial prosthesis. Chest 1970;58:566e70. 2. Fox CJ, Starnes BW. Vascular surgery on the modern battlefield. Surg Clin North Am 2007;87:1193e211. 3. Feliciano DV, Mattox KL, Graham JM, Bitondo CG. Fiveyear experience with PTFE grafts in vascular wounds. J Trauma 1985;25:71e82. 4. AbuRahma AF, Robinson PA, Holt SM. Prospective controlled study of polytetrafluoroethylene versus saphenous vein in claudicant patients with bilateral above knee femoropopliteal bypasses. Surgery 1999;126:594e601. discussion 601e602. 5. Chang JK, Calligaro KD, Ryan S, et al. Risk factors associated with infection of lower extremity revascularization: analysis of 365 procedures performed at a teaching hospital. Ann Vasc Surg 2003;17:91e6. 6. Siracuse JJ, Nandivada P, Giles KA, et al. Ten year experience with prosthetic graft infections involving the femoral artery. J Vasc Surg 2013;57:700e5. 7. Vardanian AJ, Chau A, Wuinones-Baldrich W, Lawrence PE. Arterial allograft allows in-line reconstruction of prosthetic graft infection with low recurrence rate and morality. Am Surg 2009;75:1000e3. 8. White JM, Stannard A, Burkhardt GE, et al. The epidemiology of vascular injury in the wars in Iraq and Afghanistan. Ann Surg 2011;253:1184e9. 9. Hughes CW, Cohen A. The repair of injured blood vessels. Surg Clin North Am 1958;38:1529e43. 10. Sohn VY, Arthurs ZM, Herbert GS, et al. Demographics, treatment and early outcomes in penetrating vascular combat trauma. Arch Surg 2008;143:783e7.
272 Stull et al.
11. Watson JD, Houston R 4th, Morrison JJ, et al. A retrospective cohort comparison of expanded polytetrafluorethylene to autologous vein for vascular reconstruction in modern combat casualty care. Ann Vasc Surg 2015;29:822e9. 12. Vertrees A, Fox CJ, Quan RW, et al. The use of prosthetic grafts in complex military vascular trauma: a limb salvage strategy for patients with severely limited autologous conduit. J Trauma 2009;66:980e3. 13. Rich NM, Spencer FC. Vascular Trauma. Philadelphia, PA: WB Saunders, 1978. p 91.
Annals of Vascular Surgery
14. Shah PM, Ito K, Clauss RH, et al. Expanded microporous polytetrafluoroethylene (PTFE) grafts in contaminated wounds: experimental and clinical study. J Trauma 1983;23:1030e3. 15. Bell on JM, Garcı´a-Carranza A, Garcı´a-Honduvilla N, et al. Tissue integration and biomechanical behavior of contaminated experimental polypropylene and expanded polytetrafluoroethylene implants. Br J Surg 2004;91:489e94. 16. Harrison JH. Influence of infection on homografts and synthetic (Teflon) grafts. AMA Arch Surg 1958;76:67e73.
Background: In the setting of vascular injury, vein interposition graft is the preferred conduit, but may have limited availability. This study seeks to develop a large animal model assessing the graft performance of polytetrafluoroethylene (PTFE) and Dacron in the setting of a polymicrobial infection. Methods: Thirty-seven animals were placed into 4 groups for a 21-day survival period. Sixmillimeter PTFE or Dacron interposition grafts were placed in the right iliac artery with a standardized bacterial inoculation. Native vessel with and without contamination served as control groups. The inoculant was 1 107 of genetically labeled Pseudomonas aeruginosa and Staphylococcus aureus. The primary end points were graft patency (determined by duplex ultrasound and necropsy) and graft infection (culture with molecular analysis). Secondary end points included physiological measurements, blood cultures, laboratory data, and histopathology. Results: PTFE and Dacron had similar infection rates of 85.7% and 75%, respectively. There was no significant difference in infectious organisms between graft materials. PTFE and Dacron exhibited bacterial ingrowth and transmigration to the intraluminal portion of the conduit. Fortyfive percent of the Dacron group and 40% of the PTFE group remained patent at postoperative day 21 (P ¼ 0.98). Clinical data, including white blood cell count, percent neutrophils, and lactate, did not vary significantly between groups. Conclusions: PTFE and Dacron perform similarly in terms of infection rates and graft failure as both have a propensity toward bacterial ingrowth and occlusion when compared with controls. This is a valid animal model to assess graft performance in the setting of polymicrobial infection and provides an avenue for studying novel prosthetic conduits.
INTRODUCTION
Presented at the Poster Presentation Competition and won first place in its category at the 2015 Vascular Annual Meeting of the Society for Vascular Surgery, Chicago, IL, June 17e20, 2015. Conflicts of Interest: none. 1
San Antonio Military Medical Center, Sam Houston, TX.
2
Clinical Research Division, 59th Medical Wing, Joint Base San Antonio e Lackland, Sam Houston, TX. Correspondence to: Mamie C. Stull, MD, ATTN: General Surgery, 3551 Roger Brooke Drive, Fort Sam Houston, TX 78234, USA; E-mail: [email protected] Ann Vasc Surg 2016; 36: 265–272 http://dx.doi.org/10.1016/j.avsg.2016.05.089 Published by Elsevier Inc. Manuscript received: February 28, 2016; manuscript accepted: May 22, 2016; published online: 15 July 2016
Autologous vein grafts remain the preferred conduit for vascular reconstruction in the setting of a contaminated field.1e3 However, circumstances remain where autologous vein is either unavailable or is a less desirable option. One common scenario is the trauma patient where prosthetic materials offer an expedient reconstructive option in the physiologically tenuous patient or when concomitant venous injury eliminates vein as a possible conduit. In these situations, surgeons must generally resort to ‘‘off-the-shelf’’ alternatives such as polytetrafluoroethylene (PTFE) or Dacron grafts. While these grafts have generally shown adequate patency in proximal vasculature, prosthetic grafts have a lower primary and 265
266 Stull et al.
primary-assisted patency and carry a 4.3e4.7% risk of infectious complications.4e6 Treatment options for patients with infected grafts are limited to explantation with extra-anatomic bypass, in-line repair with autologous vein or prosthetic material, or conservative management with graft preservation. All options are plagued with the risk of recurrent infection, limb loss, and mortality.7 The combat environment further adds unique challenges to vascular reconstruction including a high frequency of explosive injuries, lack of available equipment or expertise in far-forward situations, need for rapid aeromedical evacuation, and ubiquitous revascularization in a contaminated field.8 Combat surgeons have been attempting to address traumatic vascular reconstruction since the Korean War when Hughes’s report described the attempted repair of 269 of 304 major vessel injuries.9 However, White et al.8 identified a 12% rate of vascular injury during the wars in Afghanistan and Iraq from 2002 to 2009, nearly 5 times higher than previous conflicts. Due to a 73% rate of explosive injury, many of these patients presented with extensive soft tissue loss, simultaneous traumatic amputations, and limited options for venous reconstruction. Between 2% and 15% of combat-related vascular injuries required reconstruction with prosthetic grafts.10e12 These studies noted that infections caused 40% of graft-related complications, prompting the authors to suggest that prosthetic grafts be used as a bridge therapy to definitive reconstruction in a combat environment.11,12 Thus, the need to identify a safe, efficacious, off-the-shelf alternative to autologous vein for vascular reconstruction has been illuminated by the vascular injuries in the recent conflicts. The purpose of this study was to create a large animal model in swine (Sus scrofa) to test various prosthetic vascular conduits, specifically PTFE and Dacron, in the setting of a polymicrobial infection. The primary end points include 21-day graft patency and rate of graft infection. The secondary end points include laboratory and histological data.
METHODS The United States Air Force 59th Airwing Clinical Research Division’s Institutional Animal Care and Use Committee approval was obtained before study initiation. All protocols were performed at the Clinical Research Division, 59th Medical Wing at Joint Base San Antonio-Lackland, Texas, an Association for Assessment and Accreditation of Laboratory Animal Care-accredited facility. Animals were treated
Annals of Vascular Surgery
in accordance with The Guide for the Care and Use of Laboratory Animals (National Research Council 199). Animal Model Female Yorkshire-Landrace swine (S. scrofa) weighing between 70 and 90 kg were acclimated in the facility for 7 days before surgery. A total of 37 animals were placed into 4 groups: 10 animals received a 6mm heparin-bonded, expanded PTFE interposition graft, 11 animals received a 6-mm Dacron interposition graft, 8 animals received a sham surgery with inoculation (Sham[+]), and 8 animals received a sham surgery without inoculation (Sham[]). Sham procedures included full retroperitoneal exposure of the iliac system and anesthetic times comparable with those of the surgical groups. Preparation Before the procedure, 325 mg of oral aspirin was administered daily, for a median of 6 days (range 5e15). On the day of surgery, the animals were premedicated with ketamine 10e20 mg/kg, acepromazine 0.1 mg/kg, and atropine 0.04e0.4 mg/kg. The animals were then given buprenorphine 0.01 mg/ kg intramuscularly and intubated. Anesthesia was maintained with isoflurane 2e3%. A urinary catheter was placed for the duration of the operation. Surgical Protocol Following preoperative preparation, the animals were brought to the operating room where strict aseptic conditions and techniques were maintained during the entirety of the procedure. A right external jugular venous line was placed and baseline labs obtained. The animals were given 100 mL/hr of Lactated Ringers throughout the procedure. Bilateral femoral velocities were assessed with ultrasound. A midline incision was made and a retroperitoneal dissection was utilized to expose the right external iliac artery. All animals received 75 U/kg of systemic heparin before arterial clamping with atraumatic vascular clamps. In the animals that received a prosthetic graft, approximately 3 cm of the artery was excised. The interposition graft was then anastomosed using 60 prolene suture with a running stitch. Before completion, forward and back-bleeding was confirmed and the distal vasculature flushed with heparinized saline. Flow through the graft was confirmed using intraoperative Doppler. Heparinization was maintained with 1000 U every 30 min. Figure 1 shows the interposition graft once placed.
Volume 36, October 2016
Graft Patency in the Setting of Infection in Swine
267
excised. If a graft thrombosed before POD 3, it was excised for gross inspection and labeled as a technical failure. Graft/Native Artery Excision
Fig. 1. Representative image of the retroperitoneal exposure and insertion of the 6-mm PTFE interposition graft.
After the animals were anesthetized, sterile technique was utilized to reopen the surgical site and expose the implanted prosthetic graft. The surrounding tissue was extracted from the proximal native artery to the distal anastomosis. After explantation, the animal was euthanized with Beuthanasia-D Special (Schering-Plough Animal Health Corp., Union, NJ). The proximal anastomosis with surrounding tissue was sent for histology and the distal anastomosis was sent for microbiological studies. Bacterial Challenge Preparation
Following exposure of the artery in the control animals, 75 U/kg of heparin was given 5 min before occlusion. Proximal and distal vascular clamps were placed on the artery to occlude a 3 cm portion of the right external iliac artery. Heparinized saline was placed in the field during the 30 min of occlusion. For all animals, once the occlusion period concluded or the graft was successfully interposed, the surgical site was irrigated with 1 L of warmed saline. For the sham surgery group with inoculation (Sham[+]), PTFE group, and Dacron group, the surgical site was contaminated with 1 mL each of 1 107 of genetically labeled Pseudomonas aeruginosa and Staphylococcus aureus. For all groups, the fascia was then closed with a running polydioxanone suture. The subcutaneous tissue was approximated with a running 2-0 vicryl suture and staples were used to close the skin. Bacitracin was spread over the incision. The pigs recovered from anesthesia and returned to their cages. Pain medication was administered as needed by the veterinary staff (buprenorphine 0.01e0.05 mg/kg intramuscular). Postoperative Care and Evaluation The animals received 325 mg oral aspirin daily following the operation. The pigs were evaluated clinically using the Tarlov gait scale, laboratory data, and with duplex ultrasound evaluation of the bilateral femoral arteries and interposition grafts on postoperative days (POD) 1, 3, 7, 10, 14, 17, and 21. Doppler was used to assess flow and velocity within the graft. On POD 21 or at the earliest demonstration of graft occlusion by ultrasound, the animals were sacrificed and the grafts were
S. aureus (ATCC BAA1025) and P. aeruginosa (PAOIUW) were cultured overnight in 30 mL Thioglycollate medium at 35 C, 5% CO2. One milliliter of the bacterial cultures was then centrifuged for 10 min. The supernatant was discarded and the bacterial concentrate was resuspended into 1 mL Microbial Freeze Drying Buffer (OPS Diagnostics 500-06, City) and pooled. The bacterial suspension was aliquoted into glass vials, freeze dried, and stored at 4 C. Lyophilized samples were periodically reconstituted to confirm bacterial concentration and viability. On the day of the protocol, lyophilized bacteria were reconstituted to a concentration of 1 107. One milliliter of Pseudomonas and Staphylococcus each was utilized to inoculate the animals. Microbiological Assays At the time of excision, the distal anastomosis was placed in 5 mL of Thioglycollate medium and incubated at 35 C, 5% CO2 for 24 hr. The medium was plated and streaked on blood agar plates (trypticase soy agar (TSA) with 5% sheep blood) and MacConkey agar plates. Gram stain, catalase test, coagulase test, and a VITEK 2 compact (BioMerieux, Inc., Durham, NC) were used to identify S. aureus. P. aeruginosa was identified by gram stain, the indole test, and oxidase test. All confirmed samples of S. aureus and P. aeruginosa underwent arbitrarily primed polymerase chain reaction analysis with molecular gel staining to confirm that the strains were identical to the inoculum (Fig. 2). Blood cultures were collected into in BacT/ ALERTÒ PF bottles and monitored on the BacT/ ALERT Microbial Detection System (BioM erieux,
268 Stull et al.
Annals of Vascular Surgery
Fig. 2. Representative gels from the arbitrarily primed polymerase chain reactions ran on the cultured organisms to ensure that the recovered organism was
consistent with the standardized inoculum. ‘‘I’’ indicates that the organisms are similar genetically and ‘‘D’’ indicates that the organisms differ.
Inc.). Positive blood cultures were identified using the same steps as used to determine positive graft cultures.
appropriate. Categorical variables were assessed with Fisher’s exact test and reported as proportions. Graft patency was plotted as a survival curve using a log-rank test. Clinical and laboratory variables measured at multiple time points throughout the experiment were analyzed with repeated measures 2-way ANOVA. Statistical software utilized was GraphPad Prism version 6.0f (La Jolla, CA).
Histopathological Studies The proximal anastomosis or a proximal portion of the native artery with the surrounding tissue was collected, fixed in 10% neutral buffered formalin, sectioned at 5 microns, and processed by conventional methods. The samples were stained with hematoxylin, eosin, and gram stains. A blinded, veterinary pathologist then assessed the samples for evidence of infection and inflammation. A graded scale of inflammation from 0 (none) to 5 (severe) was utilized to differentiate degrees of inflammatory changes. Statistical Analysis All continuous variables were reported as mean (standard deviation) or median (interquartile range [IQR]), and analyzed with one-way analysis of variance (ANOVA) or a KruskaleWallis test, as
RESULTS Clinical Assessments Baseline demographic variables were similar between groups (Table I). Throughout the experiment, no significant differences were observed between groups regarding heart rate (P ¼ 0.82), temperature (P ¼ 0.92), white blood cell (WBC) (P ¼ 0.32), or lactate (P ¼ 0.93). There was a significant difference between groups regarding the rise of neutrophil percentage (P < 0.01) and post hoc comparisons revealed a higher percentage of neutrophils in each arm that received inoculum (Dacron, PTFE,
Volume 36, October 2016
Graft Patency in the Setting of Infection in Swine
269
Table I. Baseline characteristics demonstrate homogeneity Parameter
PTFE
Dacron
Sham (+)
Sham ()
P valuea
Heart rate Temperature ( C) WBC % Neutrophils Lactate (mmol/L) Velocity (cm/sec)c
79 ± 17 36.5 ± 0.7 14 ± 3.5 35 ± 7.8 1.3 (1.0e1.5) 90 (78e113)
80 ± 17 36.6 ± 0.5 16 ± 2.5 27 ± 6.8 1.2 (0.7e1.6) 101 (84e118)
83 ± 12 36.7 ± 0.7 15 ± 5.1 27 ± 11.2 1.1 (0.8e1.9) 103 (95e127)
82 ± 17 36.8 ± 0.7 18 ± 4.5 26 ± 9.0 1.2 (0.7e1.4) 122 (94e152)
0.96 0.84 0.23 0.13 0.91b 0.40b
a
One-way ANOVA unless otherwise specified. KruskaleWallis test. c Right femoral artery velocity. b
Sham[+]) compared with the Sham[] group on POD 1. While the Dacron and Sham[+] groups returned to equivalent neutrophil percentage to the Sham[] group following POD 1, PTFE had a significantly elevated neutrophil count until POD 3 and subsequently all groups remained similar. Within each group, a significant rise and fall of heart rate, temperature, WBC, and percent neutrophils (P < 0.01, each) was noted over time until excision. In general, the animals remained mobile and failed to show clinical evidence of infection such as fever (>40 C), malaise, or reduced oral intake, with the exception of 2 animals with infected Dacron grafts. Intraoperatively, there was a median occlusive time of 52 min (IQR 30e82 min).
Fig. 3. Log-rank test of graft patency.
Overall, 45% of the Dacron group and 40% of the PTFE group remained patent until POD 21 (P ¼ 0.98; Fig. 3), demonstrating no significant difference between the 2 groups. Given that the 2 sham groups remained 100% patent at 21 days, there was a significant difference between the graft and sham groups (P < 0.01). In the Dacron group, 3 of 11 animals (27%) were sacrificed within the first 72 hr and labeled as technical failures, which were excluded from the survival analysis. In the PTFE group, 2 of 10 animals (20%) were lost to technical failure with no significant difference noted between graft groups. The median graft patency was 14 days for both the Dacron and PTFE animals. All grafts lost before 21 days were found to be grossly infected and thrombosed at the time of excision without evidence of a technical source of failure.
72 hr, Dacron demonstrated a 75% infection rate compared with PTFE, which demonstrated an 87.5% infection rate (P ¼ 0.67) All grafts that thrombosed within 72 hr also demonstrated bacterial growth at the time of excision. P. aeruginosa was cultured from 7 of 8 (87.5%) PTFE grafts and 6 of 8 (75%) Dacron grafts (P ¼ 1.0). S. aureus grew from 6 (75%) PTFE and 3 (37.5%) Dacron grafts (P ¼ 0.31). Neither the Sham[+] nor the Sham[] groups demonstrated infection with the standardized inoculum upon excision on POD 21. Contaminants universally grew from all 16 sham animals. Three of 8 Dacron grafts and 4 of 8 PTFE grafts also grew at least one or more contaminants consistent with the natural flora of S. scrofa. Positive blood cultures were only seen in 2 animals, both of which were implanted with a Dacron interposition graft. In each instance, P. aeruginosa was cultured the day before graft thrombosis.
Graft Infection
Histology
There was no significant difference in the overall percent of infection or the type of organism with which Dacron or PTFE was infected (Fig. 4). Of the prosthetic grafts that remained patent beyond
Multiple prosthetic grafts demonstrated bacterial ingrowth through the PTFE and Dacron (Fig 5). In particular, the 2 animals with positive blood cultures demonstrated large multifocal colonies of
Graft Patency
270 Stull et al.
Annals of Vascular Surgery
Fig. 4. (A) Overall comparison of percent infected in Dacron and PTFE groups. (B) Specific comparison of the type of infection between all groups.
bacteria on the luminal side of the graft. However, bacterial growth into the graft material was common in both the PTFE and Dacron groups. None of the animals in either the Sham[+] or Sham[] groups had bacteria present in the native artery or the surrounding tissue submitted for gram stain. There was a significantly greater degree of inflammation in the surrounding tissue for animals with prosthetic grafts compared with the control groups (Dacron 4.2 ± 0.6; PTFE ¼ 3.8 ± 0.7; Sham [+] ¼ 1.1 ± 1.2; Sham[] ¼ 0.8 ± 1.0, P < 0.01). The majority of inflammation was granulomatous in nature.
DISCUSSION The purpose of this study was to create a model of prosthetic infection mimicking that seen in a clinical setting using 2 of the standard prosthetic vascular conduits. We demonstrated a similar, <50% patency of PTFE and Dacron in the setting of a mixed S. aureus and P. aeruginosa infection in swine. Furthermore, there was an equivalent rate of infection in each group. This study establishes a new, large animal model of polymicrobial infection allowing for continued evaluation of novel
Fig. 5. (A) Gram stain of PTFE with bacteria visible within the graft. (B) Gram stain of Dacron with bacteria visible within the graft.
approaches to vascular reconstruction in a contaminated field. This model furthers the work of previous authors who have documented in laboratory and clinical settings the failure of prosthetic material in an infected field. In 1978, Rich and Spencer’s13 Vascular Trauma recognized the risk of infection associated with prosthetic graft material and definitively stated the superiority of autologous vein reconstruction. In 1983, Shah et al.14 reported an updated model in dogs, using 1 107 S. aureus and Escherichia coli in a 21-day survival model. In the setting of infection, the vein grafts were subject to transmural necrosis and perforation with devastating consequences compared with anastomotic dehiscence with prosthetic materials. However, these studies generally noted that venous reconstruction offered a greater overall patency rate and was more amenable to treatment with systemic antibiotics. Staphylococcus infections have been shown to grow into the synthetic fibers of PTFE within 7 days and to decrease the tensile strength of
Volume 36, October 2016
reconstructions by 30 days.15 Antibiotics often cannot penetrate the biofilms created around prosthetic material, which requires explantation for adequate sterilization. Our study utilized S. aureus and P. aeruginosa because of their ubiquitous prevalence and ability to form biofilms. In addition, utilizing genetically labeled bacteria allowed us to clearly delineate those infections created by inoculation versus infection inherent to the animals. As such, 75% of the Dacron grafts and 87.5% of PTFE grafts remained infected at graft excision. Neither material demonstrated resistance to overall infection or either organism. We demonstrated a 40% patency rate of PTFE and 45% patency of Dacron in the setting of polymicrobial infection. These rates are comparable with previous models, but offer the advantage of an ultrasound graft surveillance regimen to better define the time to occlusion.1,16 More importantly, these results are similar to the retrospective analyses of modern combat complications from prosthetic materials, improving the applicability of our model to that population.11,12 Watson et al.11 noted a 69% rate of graft-related complications at 8 years compared with 23% for autologous vein after peripheral reconstructions in combat. Based on this clinical evidence, we recognize the superiority of venous reconstruction and designed our model for 2e15% of those patients with limited options who require an off-the-shelf conduit.10e12 Although the survival time was limited to 3 weeks, we believe that this model creates a foundation for future studies. Even within the 21-day study period, a large proportion of our grafts demonstrated nontechnical failure at necropsy. Although none of these animals manifested a systemic infection, it is possible that a longer survival may have elicited this response. Most importantly, the Dacron and PTFE groups both proved to create a nidus from which a profound infection could begin. Furthermore, histopathological examination reveals migration of bacteria into the inner lumen. While very few animals became bacteremic, we believe that longer survival may have also yielded more positive blood cultures, systemic sepsis, and conduit thrombosis/dehiscence. This study was limited by the small numbers and short survival inherent to large animal models. Prolonged survival models in large animals are challenging and very resource intensive, while studies with less survival time may miss important clinical findings. We selected 3 weeks as it encompasses the 9- to 18-day time frame when graft infection is usually clinically significant, and it builds on historic precedent. In addition, this model does not
Graft Patency in the Setting of Infection in Swine
271
investigate a venous conduit arm as venous options are limited in swine due to lack of a saphenous vein. The internal jugular vein is the most comparable based on size, but would require an additional surgical site differing this from the other arms of the model. Future studies will likely investigate options for venous comparison. Despite these limitations, this model offers a unique in vivo comparison of synthetic materials with natural tissue reactions with infection and foreign bodies. Furthermore, there are several strengths to this model over prior animal studies, including a uniform bacterial inoculum, a graft surveillance regimen, and molecular confirmation of cultured organisms.
CONCLUSIONS With increasing rates of vascular infection, in both the military and civilian sector, the need is evident for a commercially available conduit that is inherently resistant to infection. This model provides a backbone from which future studies can be built to test novel vascular conduits. REFERENCES 1. Bricker DL, Beall AC, DeBakey ME. The differential response to infection of autogenous vein versus Dacron arterial prosthesis. Chest 1970;58:566e70. 2. Fox CJ, Starnes BW. Vascular surgery on the modern battlefield. Surg Clin North Am 2007;87:1193e211. 3. Feliciano DV, Mattox KL, Graham JM, Bitondo CG. Fiveyear experience with PTFE grafts in vascular wounds. J Trauma 1985;25:71e82. 4. AbuRahma AF, Robinson PA, Holt SM. Prospective controlled study of polytetrafluoroethylene versus saphenous vein in claudicant patients with bilateral above knee femoropopliteal bypasses. Surgery 1999;126:594e601. discussion 601e602. 5. Chang JK, Calligaro KD, Ryan S, et al. Risk factors associated with infection of lower extremity revascularization: analysis of 365 procedures performed at a teaching hospital. Ann Vasc Surg 2003;17:91e6. 6. Siracuse JJ, Nandivada P, Giles KA, et al. Ten year experience with prosthetic graft infections involving the femoral artery. J Vasc Surg 2013;57:700e5. 7. Vardanian AJ, Chau A, Wuinones-Baldrich W, Lawrence PE. Arterial allograft allows in-line reconstruction of prosthetic graft infection with low recurrence rate and morality. Am Surg 2009;75:1000e3. 8. White JM, Stannard A, Burkhardt GE, et al. The epidemiology of vascular injury in the wars in Iraq and Afghanistan. Ann Surg 2011;253:1184e9. 9. Hughes CW, Cohen A. The repair of injured blood vessels. Surg Clin North Am 1958;38:1529e43. 10. Sohn VY, Arthurs ZM, Herbert GS, et al. Demographics, treatment and early outcomes in penetrating vascular combat trauma. Arch Surg 2008;143:783e7.
272 Stull et al.
11. Watson JD, Houston R 4th, Morrison JJ, et al. A retrospective cohort comparison of expanded polytetrafluorethylene to autologous vein for vascular reconstruction in modern combat casualty care. Ann Vasc Surg 2015;29:822e9. 12. Vertrees A, Fox CJ, Quan RW, et al. The use of prosthetic grafts in complex military vascular trauma: a limb salvage strategy for patients with severely limited autologous conduit. J Trauma 2009;66:980e3. 13. Rich NM, Spencer FC. Vascular Trauma. Philadelphia, PA: WB Saunders, 1978. p 91.
Annals of Vascular Surgery
14. Shah PM, Ito K, Clauss RH, et al. Expanded microporous polytetrafluoroethylene (PTFE) grafts in contaminated wounds: experimental and clinical study. J Trauma 1983;23:1030e3. 15. Bell on JM, Garcı´a-Carranza A, Garcı´a-Honduvilla N, et al. Tissue integration and biomechanical behavior of contaminated experimental polypropylene and expanded polytetrafluoroethylene implants. Br J Surg 2004;91:489e94. 16. Harrison JH. Influence of infection on homografts and synthetic (Teflon) grafts. AMA Arch Surg 1958;76:67e73.