Original articles
Prevention of Vascular Graft Infection by Rifampin Bonding to a Gelatin-Sealed Dacron Graft Olivier Go~au-Brissonni6re, MD, PhD, Catherine Leport, MD, PhD, Francois Bacourt, MD, Claude Lebrault, MD, Raymonde Comte, Jean-Claude Pech~re, MD, Paris, France, and Geneva, Switzerland
This study examines the efficacy of rifampin bonding to a gelatin-sealed knitted Dacron graft to prevent perioperative bacteremic vascular graft infection. Antibiotic bonding was obtained by soaking grafts for 15 minutes in a 1 mg/ml saline solution of rifampin at 37~ Nineteen dogs had thoracoabdominal aortic bypass: seven (group I) received a rifampin treated graft; six (group II) received an untreated gelatin-coated graft; and six (group III) received an uncoated Dacron graft. Two days later bacteremic challenge was produced by rapid intravenous injection of 5 x 105 colony forming units of methicillin resistant Staphylococcus aureus. Grafts were harvested five days after this challenge and cut into 10 fragments, each submitted to bacterial counts. Results were expressed as CFU/cm 2 of graft material. In group I, no graft was infected, whereas all grafts in groups II and Ill were infected (p < 0.05). Median bacterial counts from the infected fragments (median _ SD) were similar in groups II (2.5 x 10 ~ CFU/cm 2) and III (4 x 104 CFU/cm2). Blood cultures at time of sacrifice were negative in all dogs in group I and positive in five of six dogs in groups II and III. Cultures of liver, spleen, kidney, and lung specimens were always negative in group I and positive in 22 of 24 specimens in group II and 23 of 24 specimens in group III. Soaking a gelatin-sealed Dacron graft in rifampin solution evidently prevents early bacteremic graft infection and secondary foci of infection in this model. (Ann Vasc Surg 1991 ;5:408-412). KEY WORDS:
Graft infection; gelatin-sealed grafts; rifampin bonding.
Vascular prosthetic graft infection remains the most serious complication of reconstructive arterial surgery, especially when the aorta is involved.
Although prophylactic systemic antibiotics have decreased the incidence of this problem significantly [1], infections still occur in approximately
From the Groupe d'Etude des Biomat(riaux, Departments of Surgery and Anesthesiology, Ambroise Par~ Hospital and Ren~ Descartes University; the Department of Infectious and Tropical Diseases, Claude Bernard Hospital, INSERM U 13, Paris, France; and the Department of Microbiology, Faculty of Medicine, Geneva, Switzerland.
Presented at the Annual Meeting of the French Vascular Surgery Society, Nancy, France, May 18-19, 1990. Reprint requests: Olivier Go~au-BrissonniOre, MD, Service de Chirurgie Visc(rale et Vasculaire, HOpital Ambroise ParO, 9 Avenue Charles de Gaulle, 92104 Boulogne Cedex, France.
408
VOLUME 5 N o . 5 - 1991
PREVENTION OF GRAFT INFECTION B Y R I F A M P I N BONDING
2% of operations in which prostheses are implanted [2-7], and other forms of prophylaxis continue to be sought. Most infections are due to a contamination at the time of implantation [5,7]. Bonding of antimicrobial agents to prosthetic materials has been suggested as a way of decreasing graft infectivity. Techniques for the fixation of antibiotics to vascular grafts have been developed using surfactant agents [8-14], collagen release [15], silver containing antibiotics [16-18], and a passive method of adding rifampin to the blood used to preclot Dacron prostheses [19]. The advent of protein-sealed textile grafts has raised the possibility of using the sealant as a vehicle for antibiotic delivery. The purpose of this experimental study was to evaluate the ability of a simple system involving soaking a gelatin-sealed Dacron graft in a solution of rifampin to prevent bacteremic Staphylococcus aureus (S. aureus) vascular graft infection. Since the sealant could have modified the infectivity of grafts, we used the same animal model to test the same Dacron graft in its unsealed form.
MATERIALS AND METHODS Vascular grafts
The prosthetic grafts chosen for this study were a commercial gelatin-sealed knitted Dacron graft (Gelsoft| and its unsealed form (VP 1200). Grafts were 8 mm internal diameter and 25 to 30 cm in length. Preparation of grafts
Experimental sealed grafts were soaked just before implantation in a 1 mg/ml normal saline solution of rifampin (Rifadine| t at 37~ for 15 minutes. Control sealed grafts were soaked in a normal saline solution at 37~ for 15 minutes, without adding rifampin. Unsealed grafts were preclotted according to the method of Sauvage (described in Yates and associates [20]). Implantation of grafts
Nineteen mongrel dogs weighing 15-20 kg were anesthetized with intravenous (IV) pentobarbital sodium and given respiratory support. Each dog underwent thoracoabdominal aortic bypass using aseptic technique. Seven dogs received an experimental, rifampin-treated graft, six received a control sealed graft and six received an unsealed graft. Grafts were anastomosed end-to-end proximally *Vascutek Limited, Inchinnan, Scotland *Merrell-Dow, Levallois-Perret, France
409
and end-to-side distally with 5-0 polypropylene sutures [21]. The distal stump of the thoracic aorta was oversewn. No systemic antibiotics were given. During the operation, hydration was maintained by IV perfusion with Ringer's solution. Dogs were anticoagulated intravenously with 100 units/kg of sodium heparinate before implantation of the grafts. Animal care complied with the "Principles of Laboratory Animal Care" and the "Guide for the Care and Use of Laboratory Animals" (National Institutes of Health Publication no. 80-23, revised 1978). Bacterial strain and bacteremic challenge
Staphylococcus aureus CMU 147, used for bacteremic challenge and isolated from an infected human wound was coagulase positive, DNAase positive, lipase positive, alphahemolysin positive, and penicillinase positive. The strain was highly resistant to penicillin G, methicillin (heterogenous type), cephalosporins, imipenem, and erythromycin, but susceptible to aminoglycosides, quinolones, vancomycin and rifampin (MIC for rifampin: 0.03 rag/L). The bacteria were stored on trypticase soy agar (TSA). When required, the strain was grown overnight at 37~ in trypticase soy broth (TSB), washed twice in 10 ml normal saline solution, and resuspended in normal saline solution to obtain approximately 5 x 106 colony forming units (CFU)/ml. The actual inoculum size was determined by formal dilution-plate colony counts on TSA. Forty-eight hours after graft implantation, transient bacteremia was produced in dogs by intravenous injection of 1 ml S. aureus suspension over a period of one minute. Graft retrieval and preparation
Five days after the bacteremic challenge, the dogs received sodium heparinate (100 units/kg) to prevent thrombus formation during isolation and excision of the graft. Venous blood was sampled just before sacrifice for quantitative blood culture. The dogs were sacrificed, and the grafts were removed surgically under aseptic conditions. Specimens of lung, liver, spleen and kidney were also taken for bacteriological culture. Remaining blood was gently expressed from the grafts. Each graft was cut into 10 fragments of roughly equal size. Each fragment was submitted to bacterial counts. Measuring the length and weight of the whole prosthesis and weight of each fragment allowed estimation of the internal surface area of each fragment. Bacterial counts
One part of each fragment was hand-crushed for two minutes after addition of 2 ml of normal saline
ANNALS OF VASCULAR SURGERY
P R E V E N T I O N OF GRAFT INFECTION B Y R I F A M P I N B O N D I N G
410
solution. The crushing effluent was submitted to appropriate 10-fold dilutions and 0.1 ml aliquots were subcultured onto TSA and incubated at 37~ for 48 hours. Colony counts were expressed as the number of CFU per cm 2 of graft material. Another part of each fragment was frozen at -20~ immediately after removal for rifampin assays. Blood and organ sampling
Quantitative bacterial counts were obtained by immediately plating 1 ml of the venous blood samples taken just before graft explantation on TSA. The plates were incubated at 37~ for 48 hours at which time colony counts were performed. Tissue specimens were cultured on TSA plates. Cultures were incubated at 37~ for 48 hours. Cultures were considered positive if the study strain of S. aureus was recovered.
TABLE I.--Results of bacteriological cultures from arterial grafts Positive culture/total number examined Grafts Fragments Rifampin grafts Control grafts Unsealed grafts
0/7*
0/70 t
Viable counts, median (range) CFU/cm 2 0t
6/6
57/60
2.5 x 105 (0 --~ 107)
6/6
60/60
4 x 104 (70 --~ 10z)
*p < 0.05 versus control and unsealed grafts tp < 0,01 versus control and unsealed grafts
Results were analyzed by Fisher's exact test or chi-square test where appropriate.
logical cultures from grafts are shown in Table I. None of the grafts from the seven dogs receiving rifampin-treated prostheses was infected, whereas all the grafts from control dogs or dogs receiving an unsealed graft grew bacteria. This difference was significant as shown by Fisher's exact test (p < 0.05). The comparison of the number of fragments yielding bacterial growth in each group of dogs showed that none of the 70 fragments of grafts explanted from dogs receiving a rifampin-treated graft grew bacteria, while 57 of the 60 fragments of grafts explanted from control dogs and all the fragments explanted from dogs receiving an unsealed graft grew bacteria (p < 0.01). There was no significant difference in the density of bacteria measured in the infected fragments of control and unsealed grafts. Bacteriological cultures of tissue specimens were constantly negative in dogs receiving a rifampintreated graft, whereas 22 of the 24 specimens from control dogs and 23 of 24 specimens from dogs receiving an unsealed graft grew bacteria (Table II). Rifampin assays were performed on all of the 40 fragments of four prostheses seven days after graft implantation. All exhibited antibiotic activities ranging from 0.1 to 0.9/~g/g of prosthesis.
RESULTS
DISCUSSION
All dogs had recovered from operation at the time of bacterial challenge. However, control dogs and dogs receiving an unsealed graft exhibited signs of infection after the challenge. All grafts were patent when removed. The actual inoculum size (mean + SD) was 9.2 -+ 8.9 x 106 CFU without significant differences between the three groups of dogs. Blood cultures at the time of graft removal were negative in all the dogs bearing a rifampin-treated graft and positive in five dogs of each other group. Results of bacterio-
Direct bonding of antibiotics to prosthetic materials is an appealing concept, especially in patients exposed to a special risk of postoperative infection such as those undergoing grafting for an abdominal aortic aneurysm [3,22,23]. Providing high levels of
~Merck-ABS, Basel, Switzerland. **Schleicher & Schuell, Zurich, Switzerland. *Difco, Chemie Brunshwig, Basel, Switzerland.
Rifampin grafts Control grafts Unsealed grafts
Rifampin assays
At the time of final analysis, frozen parts of graft fragments were thawed and vigorously washed three times in methanol ~. Methanol washouts were filtered through filters n~ ** and dried under vacuum. Residues were then taken into phosphate buffer (1 ml, 0.1 M, pH 4.5), and assayed according to a microbiological assay. The assay employed Sarcina lutea ATCC 9341 as test organism incorporated into Antibiotic Medium I* and was able to detect 0.125 p~g/ml of rifampin when a prediffsion time of four hours at +5~ was allowed. Results were expressed as ~g/g of material. Statistical analysis
TABLE II.--Results of bacteriological culture from
tissue specimens Number of specimens with positive culture/total number of specimens examined Kidney Liver Lung Spleen 0/7 5/6 5/6
0/7 6/6 6/6
0/7 6/6 6/6
0/7 5/6 6/6
VOLUME 5 N o . 5 - 1991
P R E V E N T I O N OF GRAFT I N F E C T I O N B Y R I F A M P I N B O N D I N G
an antimicrobial agent specifically in the area at risk should increase the graft resistance to infection. Several methods have been described to develop infection-resistant prostheses. Clark and Margraf treated polyester grafts with silver allantoin-heparin, which allowed bacterial inhibition for about 24 hours [24]. Harvey and Greco used benzalkoniurn chloride or tridodecylmethylammonium to bind penicillin, oxacillin or cefazolin to polytetrafluoroethylene (PTFE) or Dacron grafts [8-14]. Moore described a Dacron graft in which an amikacin collagen mixture was used as the graft sealant [15]. In an animal model, these grafts resisted an immediate perioperative bacteremic challenge with S. aureus. Benvenisty recently demonstrated the effectiveness of silver-oxacillin and silver-amikacin PTFE grafts to reduce graft infection in a stringent, direct contamination model [16]. The design of an infection-resistant graft needs to comply with several prerequisites. First, the antibiotic obviously needs to be effective against the organisms that are involved in vascular graft infections. Review of the literature indicates that S. a u r e u s remains the leading pathogen, followed by other staphylococci, non-staphylococcal Gram-positive organisms, and increasing numbers of Gramnegative organisms, including P s e u d o m o n a s spp. [2,3]. Rifampin demonstrates a wide antibacterial activity in vitro against most aerobic Gram-positive bacilli (such as M y c o b a c t e r i u m tuberculosis) and cocci (notably with remarkable antistaphylococcal potency, MICs ranging from 0.004 to 0.015 mg/ml [25]), and against aerobic Gram-negative organisms including those causing arterial graft infections. This semisynthetic N-methyl piperazine derivative of rifamycine SV has a molecular weight of 822 daltons. It inhibits DNA-dependent ribonucleic acid polymerase activity in susceptible bacterial cells [26]. However, the use of rifampin has been associated with rapid emergence of resistance [27], especially when the site of infection contains a numerous bacterial population (over a million per ml [28]) and when rifampin is used as a single drug therapy for prolonged periods [29]. One could theorize that there is a significant diminution of the risk of resistance when rifampin is used as a prophylaxis, i.e., when the inoculum is probably much smaller. Supporting this statement, studies have shown that rifampin was very efficient in eradicating relatively small inocula of staphylococci, at a size thought to be comparable to those usually involved in the clinical setting of vascular surgery [30]. In addition, the combination of rifampin with another antimicrobial agent seems effective for limiting the problem of bacterial resistance [29], suggesting that rifampin-impregnated prostheses should be used in combination with the conventional systemic antibioprophylaxis (i.e., generally a cephalosporin).
411
A second prerequisite for an effective deviceassociated antibiotic is to be devoid of toxicity and to be nonallergenic. Rifampin is, overall, a well tolerated drug. Gastrointestinal symptoms, which are the most common side effect, are unlikely to appear after nonoral administration of a single dose slowly released in the blood circulation. A "flulike" syndrome thought to be of immunological origin has been described, sometimes associated with renal impairment, but this syndrome has been related to intermittent dosing [31]. Thrombocytopenia has also been associated with intermittent dosing and with high dose regimens as well [32]. As to hepatic dysfunction, hepatitis has been observed in patients receiving both isoniazid and rifampin [33]. To place dosage levels of rifampin in context, the usual daily dose for the treatment of tuberculosis is 600 mg, while we estimate that the maximum dose loaded in a graft is 20 mg. A third prerequisite when selecting an antibiotic for bonding to an arterial prosthesis is the ability to maintain antibiotic concentrations locally for a period of time long enough to protect the graft from perioperative contamination. In a passive system described by Powell, rifampin was added to the blood that was used to preclot Dacron prostheses. After implantation of these rifampin-treated grafts for 24 hours in dogs, the in vitro inhibition of S. aureus was 94% of that measured at time 0 [19]. These results suggested the possibility of binding rifampin to the sealant of protein-coated Dacron grafts, especially since in vitro studies by Ashton and colleagues demonstrated that a simple soaking of the gelatin-sealed graft used in our study in a rifampin solution conferred antibacterial properties to the graft for four days in vitro [34]. Our in vivo study confirmed that, 48 hours after graft implantation, such a binding provided complete protection against bacteremia. That potency of rifampin was emphasized by its ability to prevent secondary foci of infection without administration of any systemic antibiotic, which suggests that these foci arise from the infected graft. We have indications that the graft retains its ability to resist infection for a period of four days after implantation and further studies are underway. In addition, significant amounts of rifampin have been found in explanted grafts after seven days of implantation. The specificity and mechanism of interaction between antibiotic and modified gelatin sealant were studied in vitro by Ashton and coworkers, and they believe that ionic bonding between positively charged rifampin molecules and the large number of carboxyl groups present in the gelatin sealant is probably involved [34,35]. Finally, the process to create chemical bonds between the antimicrobial agent and the prosthetic matefial should be easy to accomplish in the usual clinical
412
PREVENTION OF GRAFT INFECTION B Y RIFAMP1N BONDING
setting, without chemical pretreatment of the graft material. Such pretreatment of vascular prostheses could modify their characteristics and interfere with graft healing. Our study clearly indicates that a simple soaking in a rifampin solution can protect gelatin-sealed Dacron vascular grafts against bacteremic infection for at least two days. The major advantage of the described system is the simplicity of its accomplishment, thus allowing an extemporaneous choice by the surgeon, without any modification of the graft. ACKNOWLEDGMENTS We thank Pr. J.L. Vilde, MD, for helpful discussions, Mr. B. Ait, Ms. G. Daneron for technical assistance, and Ms. E. Le Levier for preparation of the manuscript. We also thank the Centre M6dicoChirurgical Foch for providing laboratory facilities. Grafts were kindly provided by Vascutek Ltd., a company of Sulzer Medica, Inchinnan, Scotland. REFERENCES 1. KAISER AB, CLAYSON KR, MULHERIN JL, et al. Antibiotic prophylaxis in vascular surgery. Ann Surg 1978; 188:283-289. 2. BUNT TJ. Synthetic vascular graft infections. I. Graft infectious. Surgery 1983:93:733-746. 3. GOEAU-BR1SSONNIERE O, PECHERE JC, LEPORT C. Comment pr6venir les infections de proth6se. In: KIEFFER E (Ed). Les andvrismes de l'aorte abdominale sous-r~nale. Paris: AERCV, 1990, pp 143-153. 4. GOLDSTONE J, MOORE WS. Infection in vascular prostheses. Clinical manifestations and surgical management. Am J Surg 1974;128:225-233. 5. LIEKWEG WJ, GREENFIELD LJ. Vascular prosthetic infections: collected experience and results of treatment. Surgery 1977;81:335-342. 6. LORENTZEN JE, NIELSEN OM, ARENDRUP H, et al. Vascular graft infection an analysis of 62 graft infections in 2411 consecutively implanted synthetic vascular graft. Surgery 1975;78:211-216. 7, SZYLAGYI DE, SMITH RF, ELLIOTT JP, et al. Infection in arterial reconstruction with synthetic grafts. Ann Surg 1972;176:321-333. 8. HARVEY RA, GRECO RS. The noncovalent bonding of antibiotics to a polytetrafluoroethylene-benzalkonium graft. Ann Surg 1981 ;194:642-647. 9. HARVEY RA, ALCID DV, GRECO RS. Antibiotic bonding to polytetrafluoroethylene with tridodecylmethylammonium chloride. Surgery 1982;92:504-512. 10. HARVEY RA, TESORIERO JV, GRECO RS. Noncovalent bonding of penicillin and cefazolin to Dacron. Am J Surg 1984;147:205-209. 11. HENRY R, HARVEY RA, GRECO RS. Antibiotic bonding to vascular prostheses. J Thorac Cardiovasc Surg 1981;82:272-277. 12. GRECO RS, HARVEY RA, SMILOW PC. Prevention of vascular prosthetic infection by a benzalkonium-oxacillin bonded polytetrafluoroethylene graft. Surg Gynecol Obstet 1982;155:28-32. 13. GRECO RS, TROOSKIN SZ, DOUETZ AP, et al. The application of antibiotic bonding to the treatment of established vascular prosthetic infection. Arch Surg 1985;120:71-75. nmm
ANNALS OF VASCULAR StJgGERY
14. PRAHLAD A, HARVEY RA, GRECO RS. Diffusion of antibiotics from a polytetrafluoroethylene-benzalkonium surface. Ann Surg 1981;194:515-518. 15. MOORE WS, CHVAPIL M, SEIFFERT G, et al. Development of an infection-resistant vascular prosthesis. Arch Surg 1981 ;116:1403-1407. 16. BENVENISTY AI, TANNENBAUM G, AHLBORN TN, et al. Control of prosthetic bacterial infection: evaluation of an easily incorporated tightly bonded silver antibiotic PTFE graft. J Surg Res 1988;44:1-7. 17. MODAK SM, SAMPATH L, FOX CL, et al. A new method for the direct incorporation of antibiotic in prosthetic vascular grafts. Surg Gynecol Obstet 1987;164:143-147. 18. WHITE JV, BENVENISTY AI, REEMTSMA K, et al. Simple methods for direct antibiotic protection of synthetic vascular grafts. J Vase Surg 1984;1:372-380. 19. POWELL TW, BURNHAM SJ. A passive system using rifampin to create an infection-resistant vascular prosthesis. Surgery 1983;94:765-769. 20. YATES SG, BARROS D'SA AAB, BERGER K, et al. The preclotting of porous arterial prostheses. Ann Surg 1978;188: 611-622. 21. GOEAU-BRISSONNIERE O, PECHERE JC, GUIDOIN R, et al. Experimental colonization of vascular grafts with Staphylococcus aureus. Can J Surg 1983;26:540-545. 22. BUCKELS JAC, FIELDING JWL, BLACK J, et al. Significance of positive bacterial cultures from aortic aneurysm contents. Br J Surg 1985;72:440~42. 23. MACBETH GA, RUBIN JR, MC INTYRE KE. The relevance of arterial wall microbiology to the treatment of prosthetic graft infections: graft infection vs arterial infection. J Vasc Surg 1984;1:750-756 24. CLARK RE, MARGRAF HW. Antibacterial vascular grafts with improved thrombo-resistance. Arch Surg 1974;109:15%162. 25. THORNSBERRY C, HILL BC, SWENSON JM, et al. Rifampin: spectrum of antibacterial activity. Rev Infect Dis 1983;5 (Suppl 3):$412-$419. 26. HARTMANN G, HONIKEL KO, KUNSEL F, et al. The specific inhibition of the DNA-directed RNA synthesis by erythromycin. Biochem Biophys Acta 1967;145:843-844. 27. ACAR F, GOLDSTEIN FW, DUVAL J. Use of rifampin for the treatment of serious staphylococcal and Gram-negative infections. Rev lnfec Dis 1983;5 (Suppl 3):$502-$507. 28. KUNIN CM, BRANDT O, WOOD HG. Bacteriologic studies of rifampin: a new semi-synthetic antibiotic. J Infect Dis 1969;119:132-138. 29. SANDE SA. The use of rifampin in the treatment of nontuberculous infections: an overview. Rev Infect Dis 1983;5 (Suppl 3);$399-$411. 30. MANDELL GL, VENT KT. Killing of intraleucocytic Staphylococcus aureus by rifampin: in vitro and in vivo studies. J Infect Dis 1972;125:486-489. 31. DICKINSON JP, MITCHISON DA, LEE SK. Serum rifampicin concentration related to dose size and to the incidence of the "flu syndrome" during intermittent rifampin administration. J Antimicrob Chemother 1977;3:445-452. 32. POOLE G, STRADLING P, WORLLEDGE S. Potentially serious side effects of high dose twice-weekly rifampin. Bio Med J 1974;3:343-347. 33, GROSSET J, LEVENTIS S. Adverse effects of rifampin. Rev Infect Dis 1983;5 (Suppl 3),$440-$450. 34, ASHTON TR, CUNNINGHAM JD, PATON D, et al. Antibiotic loading of vascular grafts. Proceedings of the 16th Annual Meeting of the Society for Biomaterials, Charleston, South Carolina, pp 235, May 20-23, 1990. 35. JONAS RA, ZIENER G, SCHOEN FJ, et al. A new sealant for knitted Dacron prostheses minimally cross-linked gelatin. J Vase Surg 1988;7:414-419.
Prevention of Vascular Graft Infection by Rifampin Bonding to a Gelatin-Sealed Dacron Graft Olivier Go~au-Brissonni6re, MD, PhD, Catherine Leport, MD, PhD, Francois Bacourt, MD, Claude Lebrault, MD, Raymonde Comte, Jean-Claude Pech~re, MD, Paris, France, and Geneva, Switzerland
This study examines the efficacy of rifampin bonding to a gelatin-sealed knitted Dacron graft to prevent perioperative bacteremic vascular graft infection. Antibiotic bonding was obtained by soaking grafts for 15 minutes in a 1 mg/ml saline solution of rifampin at 37~ Nineteen dogs had thoracoabdominal aortic bypass: seven (group I) received a rifampin treated graft; six (group II) received an untreated gelatin-coated graft; and six (group III) received an uncoated Dacron graft. Two days later bacteremic challenge was produced by rapid intravenous injection of 5 x 105 colony forming units of methicillin resistant Staphylococcus aureus. Grafts were harvested five days after this challenge and cut into 10 fragments, each submitted to bacterial counts. Results were expressed as CFU/cm 2 of graft material. In group I, no graft was infected, whereas all grafts in groups II and Ill were infected (p < 0.05). Median bacterial counts from the infected fragments (median _ SD) were similar in groups II (2.5 x 10 ~ CFU/cm 2) and III (4 x 104 CFU/cm2). Blood cultures at time of sacrifice were negative in all dogs in group I and positive in five of six dogs in groups II and III. Cultures of liver, spleen, kidney, and lung specimens were always negative in group I and positive in 22 of 24 specimens in group II and 23 of 24 specimens in group III. Soaking a gelatin-sealed Dacron graft in rifampin solution evidently prevents early bacteremic graft infection and secondary foci of infection in this model. (Ann Vasc Surg 1991 ;5:408-412). KEY WORDS:
Graft infection; gelatin-sealed grafts; rifampin bonding.
Vascular prosthetic graft infection remains the most serious complication of reconstructive arterial surgery, especially when the aorta is involved.
Although prophylactic systemic antibiotics have decreased the incidence of this problem significantly [1], infections still occur in approximately
From the Groupe d'Etude des Biomat(riaux, Departments of Surgery and Anesthesiology, Ambroise Par~ Hospital and Ren~ Descartes University; the Department of Infectious and Tropical Diseases, Claude Bernard Hospital, INSERM U 13, Paris, France; and the Department of Microbiology, Faculty of Medicine, Geneva, Switzerland.
Presented at the Annual Meeting of the French Vascular Surgery Society, Nancy, France, May 18-19, 1990. Reprint requests: Olivier Go~au-BrissonniOre, MD, Service de Chirurgie Visc(rale et Vasculaire, HOpital Ambroise ParO, 9 Avenue Charles de Gaulle, 92104 Boulogne Cedex, France.
408
VOLUME 5 N o . 5 - 1991
PREVENTION OF GRAFT INFECTION B Y R I F A M P I N BONDING
2% of operations in which prostheses are implanted [2-7], and other forms of prophylaxis continue to be sought. Most infections are due to a contamination at the time of implantation [5,7]. Bonding of antimicrobial agents to prosthetic materials has been suggested as a way of decreasing graft infectivity. Techniques for the fixation of antibiotics to vascular grafts have been developed using surfactant agents [8-14], collagen release [15], silver containing antibiotics [16-18], and a passive method of adding rifampin to the blood used to preclot Dacron prostheses [19]. The advent of protein-sealed textile grafts has raised the possibility of using the sealant as a vehicle for antibiotic delivery. The purpose of this experimental study was to evaluate the ability of a simple system involving soaking a gelatin-sealed Dacron graft in a solution of rifampin to prevent bacteremic Staphylococcus aureus (S. aureus) vascular graft infection. Since the sealant could have modified the infectivity of grafts, we used the same animal model to test the same Dacron graft in its unsealed form.
MATERIALS AND METHODS Vascular grafts
The prosthetic grafts chosen for this study were a commercial gelatin-sealed knitted Dacron graft (Gelsoft| and its unsealed form (VP 1200). Grafts were 8 mm internal diameter and 25 to 30 cm in length. Preparation of grafts
Experimental sealed grafts were soaked just before implantation in a 1 mg/ml normal saline solution of rifampin (Rifadine| t at 37~ for 15 minutes. Control sealed grafts were soaked in a normal saline solution at 37~ for 15 minutes, without adding rifampin. Unsealed grafts were preclotted according to the method of Sauvage (described in Yates and associates [20]). Implantation of grafts
Nineteen mongrel dogs weighing 15-20 kg were anesthetized with intravenous (IV) pentobarbital sodium and given respiratory support. Each dog underwent thoracoabdominal aortic bypass using aseptic technique. Seven dogs received an experimental, rifampin-treated graft, six received a control sealed graft and six received an unsealed graft. Grafts were anastomosed end-to-end proximally *Vascutek Limited, Inchinnan, Scotland *Merrell-Dow, Levallois-Perret, France
409
and end-to-side distally with 5-0 polypropylene sutures [21]. The distal stump of the thoracic aorta was oversewn. No systemic antibiotics were given. During the operation, hydration was maintained by IV perfusion with Ringer's solution. Dogs were anticoagulated intravenously with 100 units/kg of sodium heparinate before implantation of the grafts. Animal care complied with the "Principles of Laboratory Animal Care" and the "Guide for the Care and Use of Laboratory Animals" (National Institutes of Health Publication no. 80-23, revised 1978). Bacterial strain and bacteremic challenge
Staphylococcus aureus CMU 147, used for bacteremic challenge and isolated from an infected human wound was coagulase positive, DNAase positive, lipase positive, alphahemolysin positive, and penicillinase positive. The strain was highly resistant to penicillin G, methicillin (heterogenous type), cephalosporins, imipenem, and erythromycin, but susceptible to aminoglycosides, quinolones, vancomycin and rifampin (MIC for rifampin: 0.03 rag/L). The bacteria were stored on trypticase soy agar (TSA). When required, the strain was grown overnight at 37~ in trypticase soy broth (TSB), washed twice in 10 ml normal saline solution, and resuspended in normal saline solution to obtain approximately 5 x 106 colony forming units (CFU)/ml. The actual inoculum size was determined by formal dilution-plate colony counts on TSA. Forty-eight hours after graft implantation, transient bacteremia was produced in dogs by intravenous injection of 1 ml S. aureus suspension over a period of one minute. Graft retrieval and preparation
Five days after the bacteremic challenge, the dogs received sodium heparinate (100 units/kg) to prevent thrombus formation during isolation and excision of the graft. Venous blood was sampled just before sacrifice for quantitative blood culture. The dogs were sacrificed, and the grafts were removed surgically under aseptic conditions. Specimens of lung, liver, spleen and kidney were also taken for bacteriological culture. Remaining blood was gently expressed from the grafts. Each graft was cut into 10 fragments of roughly equal size. Each fragment was submitted to bacterial counts. Measuring the length and weight of the whole prosthesis and weight of each fragment allowed estimation of the internal surface area of each fragment. Bacterial counts
One part of each fragment was hand-crushed for two minutes after addition of 2 ml of normal saline
ANNALS OF VASCULAR SURGERY
P R E V E N T I O N OF GRAFT INFECTION B Y R I F A M P I N B O N D I N G
410
solution. The crushing effluent was submitted to appropriate 10-fold dilutions and 0.1 ml aliquots were subcultured onto TSA and incubated at 37~ for 48 hours. Colony counts were expressed as the number of CFU per cm 2 of graft material. Another part of each fragment was frozen at -20~ immediately after removal for rifampin assays. Blood and organ sampling
Quantitative bacterial counts were obtained by immediately plating 1 ml of the venous blood samples taken just before graft explantation on TSA. The plates were incubated at 37~ for 48 hours at which time colony counts were performed. Tissue specimens were cultured on TSA plates. Cultures were incubated at 37~ for 48 hours. Cultures were considered positive if the study strain of S. aureus was recovered.
TABLE I.--Results of bacteriological cultures from arterial grafts Positive culture/total number examined Grafts Fragments Rifampin grafts Control grafts Unsealed grafts
0/7*
0/70 t
Viable counts, median (range) CFU/cm 2 0t
6/6
57/60
2.5 x 105 (0 --~ 107)
6/6
60/60
4 x 104 (70 --~ 10z)
*p < 0.05 versus control and unsealed grafts tp < 0,01 versus control and unsealed grafts
Results were analyzed by Fisher's exact test or chi-square test where appropriate.
logical cultures from grafts are shown in Table I. None of the grafts from the seven dogs receiving rifampin-treated prostheses was infected, whereas all the grafts from control dogs or dogs receiving an unsealed graft grew bacteria. This difference was significant as shown by Fisher's exact test (p < 0.05). The comparison of the number of fragments yielding bacterial growth in each group of dogs showed that none of the 70 fragments of grafts explanted from dogs receiving a rifampin-treated graft grew bacteria, while 57 of the 60 fragments of grafts explanted from control dogs and all the fragments explanted from dogs receiving an unsealed graft grew bacteria (p < 0.01). There was no significant difference in the density of bacteria measured in the infected fragments of control and unsealed grafts. Bacteriological cultures of tissue specimens were constantly negative in dogs receiving a rifampintreated graft, whereas 22 of the 24 specimens from control dogs and 23 of 24 specimens from dogs receiving an unsealed graft grew bacteria (Table II). Rifampin assays were performed on all of the 40 fragments of four prostheses seven days after graft implantation. All exhibited antibiotic activities ranging from 0.1 to 0.9/~g/g of prosthesis.
RESULTS
DISCUSSION
All dogs had recovered from operation at the time of bacterial challenge. However, control dogs and dogs receiving an unsealed graft exhibited signs of infection after the challenge. All grafts were patent when removed. The actual inoculum size (mean + SD) was 9.2 -+ 8.9 x 106 CFU without significant differences between the three groups of dogs. Blood cultures at the time of graft removal were negative in all the dogs bearing a rifampin-treated graft and positive in five dogs of each other group. Results of bacterio-
Direct bonding of antibiotics to prosthetic materials is an appealing concept, especially in patients exposed to a special risk of postoperative infection such as those undergoing grafting for an abdominal aortic aneurysm [3,22,23]. Providing high levels of
~Merck-ABS, Basel, Switzerland. **Schleicher & Schuell, Zurich, Switzerland. *Difco, Chemie Brunshwig, Basel, Switzerland.
Rifampin grafts Control grafts Unsealed grafts
Rifampin assays
At the time of final analysis, frozen parts of graft fragments were thawed and vigorously washed three times in methanol ~. Methanol washouts were filtered through filters n~ ** and dried under vacuum. Residues were then taken into phosphate buffer (1 ml, 0.1 M, pH 4.5), and assayed according to a microbiological assay. The assay employed Sarcina lutea ATCC 9341 as test organism incorporated into Antibiotic Medium I* and was able to detect 0.125 p~g/ml of rifampin when a prediffsion time of four hours at +5~ was allowed. Results were expressed as ~g/g of material. Statistical analysis
TABLE II.--Results of bacteriological culture from
tissue specimens Number of specimens with positive culture/total number of specimens examined Kidney Liver Lung Spleen 0/7 5/6 5/6
0/7 6/6 6/6
0/7 6/6 6/6
0/7 5/6 6/6
VOLUME 5 N o . 5 - 1991
P R E V E N T I O N OF GRAFT I N F E C T I O N B Y R I F A M P I N B O N D I N G
an antimicrobial agent specifically in the area at risk should increase the graft resistance to infection. Several methods have been described to develop infection-resistant prostheses. Clark and Margraf treated polyester grafts with silver allantoin-heparin, which allowed bacterial inhibition for about 24 hours [24]. Harvey and Greco used benzalkoniurn chloride or tridodecylmethylammonium to bind penicillin, oxacillin or cefazolin to polytetrafluoroethylene (PTFE) or Dacron grafts [8-14]. Moore described a Dacron graft in which an amikacin collagen mixture was used as the graft sealant [15]. In an animal model, these grafts resisted an immediate perioperative bacteremic challenge with S. aureus. Benvenisty recently demonstrated the effectiveness of silver-oxacillin and silver-amikacin PTFE grafts to reduce graft infection in a stringent, direct contamination model [16]. The design of an infection-resistant graft needs to comply with several prerequisites. First, the antibiotic obviously needs to be effective against the organisms that are involved in vascular graft infections. Review of the literature indicates that S. a u r e u s remains the leading pathogen, followed by other staphylococci, non-staphylococcal Gram-positive organisms, and increasing numbers of Gramnegative organisms, including P s e u d o m o n a s spp. [2,3]. Rifampin demonstrates a wide antibacterial activity in vitro against most aerobic Gram-positive bacilli (such as M y c o b a c t e r i u m tuberculosis) and cocci (notably with remarkable antistaphylococcal potency, MICs ranging from 0.004 to 0.015 mg/ml [25]), and against aerobic Gram-negative organisms including those causing arterial graft infections. This semisynthetic N-methyl piperazine derivative of rifamycine SV has a molecular weight of 822 daltons. It inhibits DNA-dependent ribonucleic acid polymerase activity in susceptible bacterial cells [26]. However, the use of rifampin has been associated with rapid emergence of resistance [27], especially when the site of infection contains a numerous bacterial population (over a million per ml [28]) and when rifampin is used as a single drug therapy for prolonged periods [29]. One could theorize that there is a significant diminution of the risk of resistance when rifampin is used as a prophylaxis, i.e., when the inoculum is probably much smaller. Supporting this statement, studies have shown that rifampin was very efficient in eradicating relatively small inocula of staphylococci, at a size thought to be comparable to those usually involved in the clinical setting of vascular surgery [30]. In addition, the combination of rifampin with another antimicrobial agent seems effective for limiting the problem of bacterial resistance [29], suggesting that rifampin-impregnated prostheses should be used in combination with the conventional systemic antibioprophylaxis (i.e., generally a cephalosporin).
411
A second prerequisite for an effective deviceassociated antibiotic is to be devoid of toxicity and to be nonallergenic. Rifampin is, overall, a well tolerated drug. Gastrointestinal symptoms, which are the most common side effect, are unlikely to appear after nonoral administration of a single dose slowly released in the blood circulation. A "flulike" syndrome thought to be of immunological origin has been described, sometimes associated with renal impairment, but this syndrome has been related to intermittent dosing [31]. Thrombocytopenia has also been associated with intermittent dosing and with high dose regimens as well [32]. As to hepatic dysfunction, hepatitis has been observed in patients receiving both isoniazid and rifampin [33]. To place dosage levels of rifampin in context, the usual daily dose for the treatment of tuberculosis is 600 mg, while we estimate that the maximum dose loaded in a graft is 20 mg. A third prerequisite when selecting an antibiotic for bonding to an arterial prosthesis is the ability to maintain antibiotic concentrations locally for a period of time long enough to protect the graft from perioperative contamination. In a passive system described by Powell, rifampin was added to the blood that was used to preclot Dacron prostheses. After implantation of these rifampin-treated grafts for 24 hours in dogs, the in vitro inhibition of S. aureus was 94% of that measured at time 0 [19]. These results suggested the possibility of binding rifampin to the sealant of protein-coated Dacron grafts, especially since in vitro studies by Ashton and colleagues demonstrated that a simple soaking of the gelatin-sealed graft used in our study in a rifampin solution conferred antibacterial properties to the graft for four days in vitro [34]. Our in vivo study confirmed that, 48 hours after graft implantation, such a binding provided complete protection against bacteremia. That potency of rifampin was emphasized by its ability to prevent secondary foci of infection without administration of any systemic antibiotic, which suggests that these foci arise from the infected graft. We have indications that the graft retains its ability to resist infection for a period of four days after implantation and further studies are underway. In addition, significant amounts of rifampin have been found in explanted grafts after seven days of implantation. The specificity and mechanism of interaction between antibiotic and modified gelatin sealant were studied in vitro by Ashton and coworkers, and they believe that ionic bonding between positively charged rifampin molecules and the large number of carboxyl groups present in the gelatin sealant is probably involved [34,35]. Finally, the process to create chemical bonds between the antimicrobial agent and the prosthetic matefial should be easy to accomplish in the usual clinical
412
PREVENTION OF GRAFT INFECTION B Y RIFAMP1N BONDING
setting, without chemical pretreatment of the graft material. Such pretreatment of vascular prostheses could modify their characteristics and interfere with graft healing. Our study clearly indicates that a simple soaking in a rifampin solution can protect gelatin-sealed Dacron vascular grafts against bacteremic infection for at least two days. The major advantage of the described system is the simplicity of its accomplishment, thus allowing an extemporaneous choice by the surgeon, without any modification of the graft. ACKNOWLEDGMENTS We thank Pr. J.L. Vilde, MD, for helpful discussions, Mr. B. Ait, Ms. G. Daneron for technical assistance, and Ms. E. Le Levier for preparation of the manuscript. We also thank the Centre M6dicoChirurgical Foch for providing laboratory facilities. Grafts were kindly provided by Vascutek Ltd., a company of Sulzer Medica, Inchinnan, Scotland. REFERENCES 1. KAISER AB, CLAYSON KR, MULHERIN JL, et al. Antibiotic prophylaxis in vascular surgery. Ann Surg 1978; 188:283-289. 2. BUNT TJ. Synthetic vascular graft infections. I. Graft infectious. Surgery 1983:93:733-746. 3. GOEAU-BR1SSONNIERE O, PECHERE JC, LEPORT C. Comment pr6venir les infections de proth6se. In: KIEFFER E (Ed). Les andvrismes de l'aorte abdominale sous-r~nale. Paris: AERCV, 1990, pp 143-153. 4. GOLDSTONE J, MOORE WS. Infection in vascular prostheses. Clinical manifestations and surgical management. Am J Surg 1974;128:225-233. 5. LIEKWEG WJ, GREENFIELD LJ. Vascular prosthetic infections: collected experience and results of treatment. Surgery 1977;81:335-342. 6. LORENTZEN JE, NIELSEN OM, ARENDRUP H, et al. Vascular graft infection an analysis of 62 graft infections in 2411 consecutively implanted synthetic vascular graft. Surgery 1975;78:211-216. 7, SZYLAGYI DE, SMITH RF, ELLIOTT JP, et al. Infection in arterial reconstruction with synthetic grafts. Ann Surg 1972;176:321-333. 8. HARVEY RA, GRECO RS. The noncovalent bonding of antibiotics to a polytetrafluoroethylene-benzalkonium graft. Ann Surg 1981 ;194:642-647. 9. HARVEY RA, ALCID DV, GRECO RS. Antibiotic bonding to polytetrafluoroethylene with tridodecylmethylammonium chloride. Surgery 1982;92:504-512. 10. HARVEY RA, TESORIERO JV, GRECO RS. Noncovalent bonding of penicillin and cefazolin to Dacron. Am J Surg 1984;147:205-209. 11. HENRY R, HARVEY RA, GRECO RS. Antibiotic bonding to vascular prostheses. J Thorac Cardiovasc Surg 1981;82:272-277. 12. GRECO RS, HARVEY RA, SMILOW PC. Prevention of vascular prosthetic infection by a benzalkonium-oxacillin bonded polytetrafluoroethylene graft. Surg Gynecol Obstet 1982;155:28-32. 13. GRECO RS, TROOSKIN SZ, DOUETZ AP, et al. The application of antibiotic bonding to the treatment of established vascular prosthetic infection. Arch Surg 1985;120:71-75. nmm
ANNALS OF VASCULAR StJgGERY
14. PRAHLAD A, HARVEY RA, GRECO RS. Diffusion of antibiotics from a polytetrafluoroethylene-benzalkonium surface. Ann Surg 1981;194:515-518. 15. MOORE WS, CHVAPIL M, SEIFFERT G, et al. Development of an infection-resistant vascular prosthesis. Arch Surg 1981 ;116:1403-1407. 16. BENVENISTY AI, TANNENBAUM G, AHLBORN TN, et al. Control of prosthetic bacterial infection: evaluation of an easily incorporated tightly bonded silver antibiotic PTFE graft. J Surg Res 1988;44:1-7. 17. MODAK SM, SAMPATH L, FOX CL, et al. A new method for the direct incorporation of antibiotic in prosthetic vascular grafts. Surg Gynecol Obstet 1987;164:143-147. 18. WHITE JV, BENVENISTY AI, REEMTSMA K, et al. Simple methods for direct antibiotic protection of synthetic vascular grafts. J Vase Surg 1984;1:372-380. 19. POWELL TW, BURNHAM SJ. A passive system using rifampin to create an infection-resistant vascular prosthesis. Surgery 1983;94:765-769. 20. YATES SG, BARROS D'SA AAB, BERGER K, et al. The preclotting of porous arterial prostheses. Ann Surg 1978;188: 611-622. 21. GOEAU-BRISSONNIERE O, PECHERE JC, GUIDOIN R, et al. Experimental colonization of vascular grafts with Staphylococcus aureus. Can J Surg 1983;26:540-545. 22. BUCKELS JAC, FIELDING JWL, BLACK J, et al. Significance of positive bacterial cultures from aortic aneurysm contents. Br J Surg 1985;72:440~42. 23. MACBETH GA, RUBIN JR, MC INTYRE KE. The relevance of arterial wall microbiology to the treatment of prosthetic graft infections: graft infection vs arterial infection. J Vasc Surg 1984;1:750-756 24. CLARK RE, MARGRAF HW. Antibacterial vascular grafts with improved thrombo-resistance. Arch Surg 1974;109:15%162. 25. THORNSBERRY C, HILL BC, SWENSON JM, et al. Rifampin: spectrum of antibacterial activity. Rev Infect Dis 1983;5 (Suppl 3):$412-$419. 26. HARTMANN G, HONIKEL KO, KUNSEL F, et al. The specific inhibition of the DNA-directed RNA synthesis by erythromycin. Biochem Biophys Acta 1967;145:843-844. 27. ACAR F, GOLDSTEIN FW, DUVAL J. Use of rifampin for the treatment of serious staphylococcal and Gram-negative infections. Rev lnfec Dis 1983;5 (Suppl 3):$502-$507. 28. KUNIN CM, BRANDT O, WOOD HG. Bacteriologic studies of rifampin: a new semi-synthetic antibiotic. J Infect Dis 1969;119:132-138. 29. SANDE SA. The use of rifampin in the treatment of nontuberculous infections: an overview. Rev Infect Dis 1983;5 (Suppl 3);$399-$411. 30. MANDELL GL, VENT KT. Killing of intraleucocytic Staphylococcus aureus by rifampin: in vitro and in vivo studies. J Infect Dis 1972;125:486-489. 31. DICKINSON JP, MITCHISON DA, LEE SK. Serum rifampicin concentration related to dose size and to the incidence of the "flu syndrome" during intermittent rifampin administration. J Antimicrob Chemother 1977;3:445-452. 32. POOLE G, STRADLING P, WORLLEDGE S. Potentially serious side effects of high dose twice-weekly rifampin. Bio Med J 1974;3:343-347. 33, GROSSET J, LEVENTIS S. Adverse effects of rifampin. Rev Infect Dis 1983;5 (Suppl 3),$440-$450. 34, ASHTON TR, CUNNINGHAM JD, PATON D, et al. Antibiotic loading of vascular grafts. Proceedings of the 16th Annual Meeting of the Society for Biomaterials, Charleston, South Carolina, pp 235, May 20-23, 1990. 35. JONAS RA, ZIENER G, SCHOEN FJ, et al. A new sealant for knitted Dacron prostheses minimally cross-linked gelatin. J Vase Surg 1988;7:414-419.
