Tobramycin-adhesive in preventing and treating PTFE vascular graft infections

JOURNAL OF SURGICAL RESEARCH 47,48-i-492

(1989)

Tobramycin-Adhesive in Preventing and Treating PTFE Vascular Graft Infections JOHN S. SHENK, M.D., ARTHUR MERLE E. OLSON, D.V.M., Hennepin

County Medical Center, Department Submitted

L. NEY, M.D.,’ DEAN T. TSUKAYAMA, M.D., M.Sc., AND MELVIN P. BUBRICK, M.D. of Surgery, 701 Park Aueme, Minneapolis, for publication

55415

June 13, 1988

in these situations violates this principle and exposes the patient to the high morbidity and mortality of vascular graft infection [2,3]. Infected prostheticgrafts are usually managed by extraanatomical bypass [4,5]. Placement of prosthetic material in situ as an alternative has been associated with high infection rates [6-91. This study was designed to decrease the infection rate of prosthetic grafts used in contaminated or infected fields by delivering local antibiotics to the vascular graft using a tissue adhesive as a vehicle.

The present study was designed to determine the effectiveness of N-butyl-2-cyanoacrylate as a vehicle to deliver antibiotics locally to contaminated vascular graft sites and to grafts with established infections. Phase IContaminated wound model: Sixteen dogs had a l-cm section of infrarenal aorta replaced with a PTFE graft. Prior to placement, the graft was immersed in solutions of Escherichia coli 3X10’ CFU/ml and then Staphylococcus aureus 3X 10’ CFU/ml. After anastomosis, 1 cc of each solution was placed directly over the graft. Eleven dogs served as controls and 5 as treatment dogs. Parenteral cefonecid was given preoperatively and daily until sacrifice. Treatment animals had the anastomoses and graft sealed with a suspension of N-butyl-2-cyanoacrylate and 1.2 g tobramycin powder (antibiotic glue, ANGL) after contamination. All dogs were reoperated on the third postoperative day. Results: Eleven of 11 control dogs had positive cultures for S. aureus and 9 of 11 had positive cultures for E. coli. Seven of 11 had pseudoaneurysms, 1 exsanguinated. None of the 4 treatment dogs had positive cultures (P = 0.0002), pseudoaneurysms (P = 0.017), or local signs of sepsis. Phase II-Infected graft model: The 10 surviving infected control dogs served as the established graft infection model. These dogs were randomized into two groups; Group 1 control (N = 5) had the graft replaced; Group 2 treatment (N = 5) had the graft replaced and ANGL treatment. Dogs were sacrificed after 2 weeks, Results: Graft cultures were positive in all 4 control dogs and negative in the 4 treatment dogs (P = 0.005). One dog in each group was eliminated secondary to failure to obtain graft culture. The data show that ANGL can be effective in the prevention and treatment of prosthetic graft infection. 0 1999 Academic Press, Inc.

METHODS

The first phase involved the use of a contaminated wound model with and without local antibiotic treatment. The second phase of the study evaluated local antibiotic delivery in an infected graft model. Infected controls from the first phase were used as the model for the second phase (Fig. 1). Antibiotic

Glue (ANGL)

ANGL was prepared by mixing 1.2 g of pharmaceutical tobramycin powder (Dista Products, Indianapolis, IN) with 3.3 g of N-butyl-2-cyanoacrylate (3M, Minneapolis, MN). This formed a fluid suspension which would polymerize within 10 set of placement in uiuo. The total volume of the mix was 4 cc and the entire 4 cc was used for each application of ANGL. Tobramycin elutions from ANGL in normal saline are shown in Fig. 2. Elutions were done by placing 1 g of ANGL in 10 cc of phosphatebuffered saline (PBS) with a pH of 7.4. The PBS was collected at 24-hr intervals, replaced with an equal amount of fresh buffer, and assayed for tobramycin concentration by fluorescence polarization immunoassay (TDX, Abbott Diagnostics, North Chicago, IL). Bacteria

INTRODUCTION

Penicillin-sensitive Staphylococcus aureus was obtained from a patient with endocarditis (Table 1). A standard lab Escherichia coli was used (American type culture collection, Difco Lab, Detroit, MI). Each was incubated overnight on sheep blood, taken directly off the plates

It is a basic surgical principle that foreign bodies should be avoided in the presence of wound contamination or active infection [ 11. Insertion of a prosthetic vascular graft ’ To whom reprint

Minnesota

requests should be addressed. 487

0022-4804/89 $1.60 Copyright 0 1989 by Academic Press, Inc. All rights of reproduction in any form reserved.

JOURNAL

488

OF SURGICAL

RESEARCH:

Contaminated Qralt Model (N=lB)

,

VOL. 47, NO. 6, DECEMBER

1989

_

NO TREATMENT

, Contemlnsted Wound Control Qroup and Infected Qrsft Model (N-1 1)

Contemlnated Treatment Qroup (N-6)

NO TREATMENT

I

1 Death Preop

Qraft Control

FIG.

1.

Flow diagram of study.

and standardized separately in normal saline to 1 McFarland (equivalent to 3 X 10’ CFU/ml). Contaminated The use of (36-59 pental

Wound Model

National Institute of Health’s guide for care and lab animals was followed. Sixteen mongrel dogs lb each) had anesthesia induced with sodium thioand maintained with halothane. Aseptic technique

was used. The abdomen was shaved, skin was prepped with 10% povidone-iodine solution, and a midline incision was made. The infrarenal aorta was mobilized from the iliac arteries to 2 cm below the renal arteries. All lumbar arteries and the inferior mesenteric artery were ligated and transected. After establishing proximal and distal control, a short segment of aorta was resected at the level of the inferior mesenteric artery. The cut ends of the aorta were retracted, leaving a 3- to 4-cm gap. A matching length TABLE Sensitivity

of Staph

Antibiotic

100

! 0

t 10

20

30

40

days FIG. PBS.

2.

Tobramycin

elutions

@g/ml) from 1 g ANGL

into 10 ml

Ampicillin Oxacillin Erythromycin Cephalothin Chloroamphenic Gentamicin Tetracycline Vancomycin Clindamycin Penicillin

1 Coagulase

Interpretation

01

Sensitive Sensitive Sensitive Sensitive Sensitive Sensitive Sensitive Sensitive Sensitive Sensitive Sensitive

Positive MIC (pglml) <0.25 12.0 <2.0 <2.0 2.0 <0.5 2.0 <0.5 <0.5 0.06

Not available

SHENK

TABLE Contaminated

Wound

ET AL.: LOCAL

IN VASCULAR

GRAFT

TABLE

Model-Culture

Results

Infected

Graft

ANGL (N=5)

11 11 9

0 0 0

(P = 0.0002) (P = 0.0002) (P = 0.0027)

of 6- or &mm-diameter PTFE was interposed using 5-O Prolene with running suture technique in an end to end anastomosis. Prior to placement, the graft was emersed sequentially in a solution of 3 X 10s CFU/ml E. coli and 3 X lo8 CFU/ml penicillin-sensitive S. aureus. After graft placement, 1 cc of each solution was placed over the graft and anastomoses. The retroperitoneum was closed with 4-O polyglycolic acid suture, and the abdominal wall was closed in layers with polyglycolic acid suture. All animals were given 250 mg cefonicid parenterally pre-op and daily until 3 days before sacrifice. To facilitate iv administration and blood drawing, a central line was placed by cut down to the external jugular vein and tunneled to the nape of the neck. The five treated dogs had received 4 cc ANGL to the graft and anastomosis 5 min after contamination. The posterior peritoneum was placed over the graft and held in place by the ANGL. Serum tobramycin levels were drawn immediately postoperatively and daily for 5 days in selected dogs. All surviving dogs were reoperated on Postoperative Day 3. At the time of surgery, distal pulses were assessed for graft patency. Blood clot was evacuated and the graft was excised with approximately 0.5 cm of aorta proximally and distally. The graft was transected and one anastomosis was sent for bacteriologic culture and the other frozen and later prepared for hematoxylineosin (H&E) staining. Infected

Wound Model

The surviving control dogs made up the established graft infection model. Each dog was randomized into either a treatment group or a control group. Control dogs had the graft excised and replaced only. Treatment dogs

TABLE Contaminated

Wound

Control (N = 11) Complications Patency Pseudoaneurysms Deaths

8 10 7 1

Model-Culture

4 4 2

0 0 0

had the graft excised, replaced, and then sealed, with ANGL. The debridement of the retroperitoneum and graft placement were done before randomization into treatment or control groups. No irrigation was used. The dogs were sacrificed between Postoperative Days 12 and 16. Patency was assessed. Grafts were harvested and transected. One anastamosis was sent for bacteriologic culture; the other was frozen as previously described. RESULTS

In the contaminated wound model, 11 of 11 cultures were positive for S. aureus and 9 of 11 were positive for E. coli. None of the five treated animals had positive cultures (P = 0.0002) (Table 2). Pseudoaneurysms formed in seven controls and no treatment dogs (P = 0.017). The only death was in the control group secondary to exsanguination from a pseudoaneurysm (Table 3). In the infected graft groups, four of four control dogs’ cultures were positive for S. aureus, and two of four were positive for E. coli. None of the four treated dogs had positive cultures (P = 0.005) (Table 4). There were no deaths, pseudoaneurysms, or graft thromboses in the control or treated groups (Table 5). Serum tobramycin levels were in the low therapeutic range initially and fell to below detectable levels rapidly. No toxic serum levels were obtained (Table 6). Histology was considered positive if bacteria were present. If no bacteria were seen on H&E stain, gram stains were done. In the contaminated wound model, 6 of 11 controls and 5 of 5 treated specimens were examined. All 6 controls were positive and none of the 5 in the treated

Infected

ANGL (N=5)

Graft

5

Model-Complications Control (N = 5)

P = 0.0072 P = 0.0167 NS

(P = 0.0051) (P = 0.0051)

from each group secondary to graft culture not

TABLE

0 5 0 0

Results

ANGL” (N = 5)

3

Model-Complications

4

Control’ (N = 5) Positive cultures S. aureu.5 E. coli a One dog eliminated obtained.

489

INFECTIONS

2

Control (N = 11) Positive cultures s. aureus E. coli

ANTIBIOTICS

Complications Patency Pseudoaneurysms Deaths

0 4 0 0

ANGL (N=6) 0 4 0 0

JOURNAL

490

TABLE Serum

Range k/ml)

OF SURGICAL

RESEARCH:

6

Tobramycin

Levels

Postop

Pod 1

Pod 2

Pod 3

Pod 4

3.31-1.76 (N = 6)

1.88-1.14 (N=6)

1.48-0.6 (N=6)

0.8-0.37 (N=6)

0.61-0.50 (N = 3)

Note. Serum tobramycin ANGL treatment.

levels drawn in a sampling

of dogs after

group were positive. Gram stains of the negative H&E specimens showed no bacteria. In the infected graft model, 4 of 4 controls and 3 of 4 treated specimens were examined. None of the 4 controls had positive H&E but 1 of these had bacteria seen on gram stain. The 3 treated specimens had negative H&E and gram stains (Table 7). DISCUSSION

The optimum management of vascular injury in the presence of known contamination and the management of infected vascular prostheses are both unresolved problems in vascular surgery. The use of vascular grafts in infected or contaminated wounds has been condemned because of the high risk of infectious complications [4, 51. These complications include pseudoaneurysms, anastomotic breakdown with exsanguination, or thrombosis. Mortality rates of 30% and amputation rates as high as 50% have been reported for established graft infections ]2, 39% 101. Infectious complications occur at a rate of l-6% in clean surgical wounds when prosthetic graft material is used [2,3]. Infection results from contamination from a variety of sources including skin, lymphatic drainage, hematogenous seeding, enteric spillage, and unrecognized mycotic aneurysm [ll, 121. Most previous studies have focused on the prevention of graft infection with antibiotics. Kaiser et al. [13] demonstrated a role for prophylactic systemic antibiotics with a reduction in the infection rate from 6.8 to 0.9% in clean surgical wounds. Several studies have evaluated antibiotic irrigation and local debridement, but these methods are effective only when the infection is localized and does not involve the anastomosis [ 14-171. Presoaking the graft in antibiotic solution before placement has been tried but the activity of the antibiotic in this situation is brief, and the drug has been shown to wash out in minutes [3, 181. More prolonged antibiotic activity can be obtained by bonding the antibiotic to the graft [ 19-221. Greco et al. were able to bond Oxacillin to PTFE by cationic surfactant, and this method was effective in reducing infections in both contaminated wounds and in the treatment of established graft infections in situ [23, 241. Many authors currently propose that the prosthetic graft is the nidus for infection and that breakdown occurs

VOL. 47, NO. 6, DECEMBER

1989

at the anastomosis [5,6,8,9]. Our premise was that bacteria at the suture line have not been controlled by the previous methods of local antibiotic delivery. For a treatment to be effective in infected or contaminated wounds, it must seal the anastomosis and graft and simultaneously deliver high dose antibiotics locally. In this study, the combination of tissue adhesive and tobramycin powder has been shown to accomplish this goal. Historically, local delivery of antibacterial agents to the site of infection predated the modern antibiotic era and began with various chemical disinfectants. With the subsequent proliferation of potent orally and parenterally administered antimicrobial agents, local antibiotic therapy was relegated to a relatively minor role in the treatment of serious infection. Recently, renewed interest in the concept of delivery of antibiotics to the infection site has been generated by the experience of Klemm and others [25, 261 with a novel delivery system utilizing polymethylmethacrylate (PMMA) cement. In 1973, Klemm introduced the use of gentamicin-impregnated PMMA beads for the treatment of chronic osteomyelitis. Use of these antibiotic-impregnated beads resulted in a high local concentration of gentamicin at the site of infection, while serum and urine levels remained very low [25, 271. This approach proved to be clinically effective and carried minimal risk of systemic toxicity [26]. A major drawback was the additional surgery required to remove the antibiotic-impregnanted beads after the completion of therapy. One possible solution to this problem is delivery of the antibiotic in a biodegradable carrier which can be resorbed over a period of time. The cyanoacrylate tissue adhesives were introduced in the late 1950s [28]. These monomers will rapidly polymerize when exposed to wet surfaces. The general formula is illustrated below: CN I CH2=C I COOR

TABLE

7

Histology-Contaminated and Infected Graft Contaminated wound model

Positive histology H&E Gram stain

Wound Model Infected graft model

Control

ANGL

Control

ANGL

w 616

O/5 O/5

l/4 014 l/4

O/3 O/3 O/3

O/5

Note. Number of specimens with bacteria present over the number of specimens examined in each group.

SHENK

ET AL.: LOCAL

ANTIBIOTICS

The R represents a side chain which can vary in length [29, 301. Early studies concentrated on the methyl-2-cyanoacrylate which showed severe local tissue toxicity as a result of rapid degredation to formaldehyde and cyanoacrylate 131-381. N-Butyl-2-cyanoacrylate has proven to be nontoxic to tissues because of its slower degredation, allowing the host to clear the by-products from the local site before toxic concentrations accumulate [39,40]. Since polymorization occurs within 10 set, it can still be used to seal tissue [41]. The choice of N-Butyl-2-cyanoacrylate was based on its capacity to protect and seal the graft and anastomosis, as well as its ability to carry antibiotics and its low toxicity. The effectiveness of ANGL in this study has raised several questions. It is not known if the N-butyl-2-cyanoacrylate alone would be effective and this is currently being evaluated in our laboratory. Second, the long-term effect of ANGL in vivo has not been determined nor has the duration of antibiotic activity been assessed. The presumed high concentration of tobramycin in the local tissues needs to be documented. Also, application of ANGL with other graft materials needs to be addressed. In spite of these questions, the results to date are promising. The data demonstrate that ANGL will effectively reduce the infection rate for grafts in both contaminated wounds and established graft infections. Potential clinical applications for ANGL include the management of hemodialysis shunt infections particularly when alternate access sites are limited and the treatment of trauma patients with vascular injuries occurring in contaminated wounds. In both situations, the application of high dose local antibiotics and sealing of the graft and anastomoses may provide a safe alternative to conventional techniques. The use of other antibiotics and different dosing schedules may also be possible in the future and this could allow great flexibility in the prevention and treatment of a variety of surgical infections.

IN VASCULAR

1.

Schwartz, S., and Shires, G. T. Principles York: McGraw-Hill, 1983. Pp. 289-310.

2.

Liekweg, W. G., and Greenfield, L. J. Vascular prosthetic infections-Collected experience and results of treatment. Surgery 81(3): 335,1977.

3.

Bunt, T. J. Synthetic 1983.

4.

Fry, W. J. Vascular prosthesis infections. 52: 1419,1972.

5.

Szilaqy, D., Smith, R., Elliott, J., and Vandrecic, M. Infection in arterial reconstruction with synthetic grafts. Ann. Sug. 176: 321, 1972.

6.

DeRose, G., and Provan, J. L. Infected arterial grafts: Clinical manifestations and surgical management. J. Cardiouasc. Surg. 25: 51, 1984.

7.

Shah, P., Ito, K., Clauss, R., Babu, S., Reynolds, B., and Stahl, W. Expanded microporous polytetrafluoroethylene (PTFE) grafts in contaminated wounds: Experimental and clinical study. J. Trauma 23(12): 1030, 1983.

8.

Knott, L., Crawford, F., and Grogan, J. Comparison of autogenous vein, Dacron, and Gore-Tex in infected wounds. J. Surg. Res. 24: 288,197s.

9.

Rich, N., and Hughes, C. The fate of prosthetic material used to repair vascular injuries in contaminated wounds. J. Trauma 12(6): 459,1972.

10.

Conn, J. H., Hardy, J. D., Chavez, C. M., and Fain, W. R. Infected arterial grafts-Experience in 22 cases with emphasis on unusual bacteria and technics. Ann. Surg. 171: 704, 1970.

11.

White, J., Freda, J., Kozar, R., Serfass, D., Cindy, K., Comerata, H., and Ritchie, W. Does bacteremia pose a direct threat to synthetic vascular grafts? Surgery 102: 402, 1987.

12.

Moore, W., Rosson, C., Hall, A., and Thomas, A. Transient teremia. Amer. J. Surg. 117: 342, 1969.

13.

Kaiser, A., Clayson, K., Mulherin, J., Roach, A., Allen, T., Edwards, W., and Dale, W. Antibiotic prophylaxis in vascular surgery. Ann. Surg. 188:283,1978.

14.

Algren, B., and Eriksson, I. Local treatment of infected grafts. Acta Chir. &and. (Suppl.) 529: 91, 1985.

15.

Popovsky, J., and Singer, S. Infected prosthetic grafts local therapy with graft preservation. Arch. Surg. 115: 203, 1980.

16.

Kwan, J., and Connally, J. Successful management of prosthetic graft infection with continuous providone-iodine irrigation. Arch. Surg. 116: 716, 1981.

17.

Bhat, D., Tellis, V., Kohberg, W., Driscoll, B., and Veith, F. Management of sepsis involving expanded polyte-trafluoroethylene grafts for hemodialysis access. Surgery 87: 445, 1980.

18.

Baker, W., and Bodensteiner, J. The administration of antibiotics in vascular reconstructive surgery. J. Thorac. Cardiouasc. Surg. 64: 301, 1972.

19.

Greco, R., and Harvey, R. Role of antibiotic bonding in the prevention of vascular prosthetic infection. Ann. Surg. 195: 168,1982.

20.

Modak, S., Sampath, L., Fox, C., Benvisty, A., Nowygrod, R., and Reemstmau, K. A new method for the direct incorporation of antibiotic in prosthetic vascular grafts. S’urg. Gynecol. O&et. 164: 143, 1987.

21.

Moore, W., Chvapil, M., Seiffert, G., and Keown, K. Development of an infection-resistant vascular prosthesis. Arch. Surg. 116: 1403, 1981.

22.

White, J., Benvenisty, A., Reemtsma, K., Voorhees, A., Fox, C., Modak, S., and Nowygrad, R. Simple method for direct antibiotic protection of synthetic vascular grafts. J. Vast. Surg. 1: 372,1984.

ACKNOWLEDGMENTS The authors thank the Minneapolis Medical Research Foundation, the National Kidney Foundation of the Upper Midwest, and 3M.

491

INFECTIONS

REFERENCES

SUMMARY

Animal models for a prosthetic graft contamination and infection were developed using standardized concentrations of E. coli and S. aureus and PTFE in dogs. Graft infections occurred in 11 of 11 untreated contaminated controls and none of 5 contaminated dogs treated with a suspension of N-butyl-2-cyanoacrylate and tobramycin powder (ANGL). Established graft infections were successfully treated in four of four dogs using ANGL in combination with graft replacement. The data suggest that ANGL may be of value in the prevention and treatment of prosthetic graft infections.

GRAFT

of Surgery, 4th ed. New

vascular graft infections.

Surgery 93: 733,

Surg. Clin. North Amer.

bac-

arterial

492

JOURNAL

OF SURGICAL

RESEARCH:

VOL. 47, NO. 6, DECEMBER

1989

23.

Greco, R., Trooskin, S., Donety, A., and Harvey, R. The application of antibiotic bonding to the treatment of established vascular prosthetic infection. Arch. Surg. 120: 71, 1985.

31.

Ota, K., Mori, S., Mieuno, K., and Inou, T. Experimental and clinical use of adhesive on parenchymatous organs. Arch. Surg. g&231,1968.

24.

Greco, R., Harvey, R., Snilow, P., and Tesouero, J. Prevention of vascular graft infections by a benzalkonium-oxacillin bonded polytetrafluoroethylene graft. Surg. Gynecol. Obstet. 155: 28, 1982.

32.

25.

Klemm, K. Septopal-A new way of local antibiotic therapy. In T. J. G. Van Rens and F. H. Kayser (Eds.), Local Antibiotic Treatment in Osteomyelitis and Soft Tissue Infections. Amsterdam: Excerpta Medica, 1981. Pp. 24.

Lehman, R., Hayes, G., and Leonard, F. Toxicity of alkyl2-syanoacrylates. Arch. Surg. 93: 441, 1966. Vasko, J., and Brochman, S. Clinical and experimental experience with plastic adhesives. Ann. Surg. 162: 123, 1965. Matsumoto, T. Use of tissue adhesive for arterial anastomosis. Arch. Surg. 96: 405, 1968. Healey, J., Gallager, H., Moore, E., Clark, R., Sheena, K., and McBride, C. Experience with plastic adhesive in the nonsuture repair of body tissues. Amer. J. Surg. 109: 416, 1965. Matsumoto, T. Use of tissue adhesive for arterial anastomosis. Arch. Surg. 96: 405, 1968. Woodward, S., Herrmann, J., Camaren, J., Brandes, G., Pulaski, E., and Leonard, F. Histotoxicity of cyanoacrylate tissue adhesive in the rat. Ann. Surg. 162: 113, 1965. Awe, W. C., Roberts, W., and Braunwald, N. Rapidly polymerizing adhesive as hemostatic agent: Study of tissue response and bacteriological properties. Surgery 54: 322, 1963. Lehman, R., West, R., and Leonard, F. Toxicity of alkyl-2-cyanoacrylates. Arch. Surg. 93: 447, 1966. Leonard, F. The N-alkylalphacyanoacrylate tissue adhesives. Ann. N. Y. Acad. Sci. 146: 203, 1968. Collins, J. A., Pani, K. C., Lehman, R. A., and Leonard, F. Biological substrates and cure rates of cyanoacrylate tissue adhesives. Arch. Surg. 93:428,1966.

26.

Grieben, A. Clinical results in septopal in bone and joint infections. A survey of clinical trials. In T. J. G. Van Rens and F. H. Kayser (Eds.), Local Antibiotic Treatment in Osteomyelitis and Soft Tissue Infections. Amsterdam: Excerpta Medica, 1981. Pp. 144-154.

27.

Walenkamp, G., Vree, T. B., and Van Rens, T. G. B. GentamicinPMMA beads. Pharmaco kinetic and nephrotoxicological study. Clin. Orthq. 205: 171, 1986.

28.

Coover, H. W., Jayner, F. B., Shearer, N. H., and Wicker, T. H. Chemistry and performance of cyanoacrylate adhesives. Sot. Plost. Eng. J. 15: 413,1959.

29.

Avery, B. S., and Ord, R. A. The use of butyl cyanoacrylate as a tissue adhesive in maxillo-facial and crania-facial surgery. Brit. J. Oral Surg. 20: 84, 1982.

30.

Gali, K., Schofield, I. D., and Wright, G. Z. Effect of N-butyl-2cyanoacrylate (histoacryl blue) on the healing of skin wounds. J. Canad. Dent. Assoc. 7: 565, 1984.

33. 34. 35.

36. 37.

38.

39. 40. 41.