Antibiotic irrigation and conservative for major aortic graft infection

Transcription

1 Antibiotic irrigation and conservative for major aortic graft infection surgery G. E. Morris, DM, MCh, FRCS, P. J. Friend, MA, FRCS, D. J. Vassallo, FRCS, RAMC, M. Farrington, MA, MRCPath, S. Leapman, MD, FACS, and C. R. G. Quick, MS, FRCS, Cambridge, London, and Huntingdon, United Kingdom and Indianapolis, Ind. Purpose: Traditional surgical treatment for panprosthetic aortic graft infection entails radical excision of the graft, aortic stump closure, and extraanatomic revascularization of the lower limbs. This carries an early mortality rate of 24% to 45%. Amputation rates range from 11% to 37%. Multiple operations and prolonged hospital stay are usual. We have developed a more conservative management technique with the aim of improving outcome. Methods: We describe an innovative method of treating the condition with prolonged, high-dose, local antibiotic irrigation therapy, systemic antibiotic treatment, surgical debridement, and graft conservation in a prospectively studied series of 10 patients. Results: The actual 30-day patient survival rate is 90%, the 1-year survival rate is 80%, and the 4-year survival rate is 67%. Two patients died because of graft infection, and the third died, uninfected, of an unrelated cause. No limbs have been amputated. Only two patients required a second operation. Mean postoperative hospital stay was 32 days. The seven survivors have been closely followed up with regular computed tomography or indium scanning and clinical examination and appear to be free from infection at a mean of 61 months after cessation of irrigation therapy. Conclusion: The technique appears to represent a significant improvement in the management of this major complication of vascular surgery. (J VASC SURG 1994;20:88-95.) The most feared complication of vascular surgery is prosthetic graft infection. Overall rates for major aortic graft infection average approximately 2%, ranging from less than 1% to 6% in published series. 16 If an abdominal aortic prosthesis is involved, traditional teaching dictates removal of the infected graft, oversewing of the aortic stump, and, since the introduction of the technique in 1963, insertion of an extraanatomic bypass. 5,712 This management is technically complex and carries a reported early mortality rate of 24% to 45%. 132 Despite this, it From the Department of Surgery (Messrs. Morris, Friend, and Quick), Clinical Microbiology and Public Health Laboratory (Dr. Farrington), Cambridge, the Department of Surgery (Mr. Vassallo), Royal Army Medical College, Millbank, London, Transplant Unit (Dr. Leapman), University Hospital, Indianapolis, and Department of Surgery (Mr. Quick), Hinchingbrooke Hospital, Huntingdon. Reprint requests: C. R. G. Quick, MS, FRCS, Department of Surgery, Hinchingbrooke Hospital, Hinchingbrooke Park, Huntingdon, Cambs, PE18 8NT, United Kingdom. Copyright 1994 by The Society for Vascular Surgery and International Society for Cardiovascular Surgery, North American Chapter /94/$ /1/ remains the conventional surgical solution for major graft infections. 21 In an effort to reduce the morbidity and mortality rates of the condition, we have devised a simpler, more conservative method of treatment. 22 Patients with infected aortic grafts are treated by surgical drainage, antibiotic irrigation of the graft and systemic antibiotic therapy without graft removal. Encouraged by the short-term (2- to 3-year follow-up) results of this technique in four cases reported in 1990, we now present a larger series of patients with medium- to long-term follow-up. PATIENTS AND METHODS This study reports the treatment of 10 patients who have been treated for major aortic graft infections from 1985 to Table i gives the details of these cases. Anastomoses were within the infected fields in nine cases, all anastomoses were involved in four grafts, and a single anastomosis was involved in five grafts. It is important to note that suture-line integrity was preserved on all occasions. All grafts were found at surgery to be lying in gastrointestinal contents or frank pus.

2 Volume 20, Number 1 J]/lorris et al. 89 Table I. Patient and graft details Patient Age Graft type Complicating factors Time to infection Extent of infection 1 59 Aortoiliac (aneurysm) Replacement graft for 3 anastomotic 81 days Whole graft false aneurysms after aortoiliac aneurysm repair, duodenal damage 2 72 Aortoiliac (aneurysm) Colon erosion by aneurysm, synchro- 70 days Limb graft (L) nous colectomy 3 52 Aorta (aneurysm) Past UTI and septicemia, traumatic 31 days Whole graft catheterization 4 67 Aortobifemoral (aneurysm) Ruptured AAA, adhesions from pre- 18 days Limb graft (L) vious peritonitis, cecal perforation 5 40 Redo (R) limb aortobifemoral Redo for graft occlusion, adhesions 28 days Limb graft (R) 6 81 Aorta (aneurysm) Ruptured AAA, diverticulitis, large 30 days Whole graft hematoma 7 45 Aortoifiac (aneurysm) Established end-stage renal failure 26 days Limb graft (L) 8 63 Aortobifemoral (aneurysm) Ligation IMA, ischemic colon, colos- 94 days Proximal graft tomy, Hartmann's 9 74 Aorto(R)femoral (aneurysm) Ruptured AAA, absent (R) femoral 926 days Distal aortic graft pulse, A-F graft Aortoiliac (aneurysm) Redo distal false aneurysm after 15 days Whole graft ruptured AAA repair, ischemic colon, fecal peritonitis, Hartmann's operation UT/, Urinary tract infection; AAA, abdominal aortic aneurysm; A-F, aortofemoral; L, left; R, right. These patients had previously undergone major aortic reconstructions, all of which were complicated. Three were repeat reconstructions, three were emergency operations for ruptured abdominal aortic aneurysms, and the others were at increased risk of graft infection for various reasons detailed in Table I. In five cases the graft was contaminated directly by gastrointestinal contents. Infection in nine of the 10 grafts became apparent within a few months of operation (mean 4.4 months, range 0.5 to 31 months). The diagnosis of graft infection was made by computed tomography (CT) scanning alone in four cases, ultrasonography alone in one case, indiumlabeled white blood cell scanning combined with CT scanning or ultrasonography in three cases, and by clinical evidence alone in two cases. The typical findings were as reported in the literature: on CT scanning, perigraft fluid and gas; on ultrasonography, perigraft fluid; and on indium-labeled white blood cell scanning, markedly increased uptake in the region of the grafts. 21'2s26 After our early experience of the irrigation technique reported in 1990, 22 we have devised a standardized protocol for treating these cases: (1) The infected graft is exposed by use of an extraperitoneal approach, if feasible. The sac is opened, and all pus and nonviable tissue are removed. (2) Samples of pus and tissue are obtained for microbiologic culture, and the graft is thoroughly irrigated with normal saline solution. (3) Two to four silicone tube drains (5 mm suggested), with multiple side holes cut in them, are placed inside the sac along the limb(s) of the graft, an end directed toward each anastomosis. The tubes are then brought out via the extraperitoneal plane to exit the skin in the loin. The native sac is closed over the graft if possible (See Fig. 1). (4) The tubes are irrigated in rotation as follows: 20 mg gentamicin in 20 ml saline solution every 2 hours for the first 12 hours; Irrigation with the same dose and volume every 4 hours for the next 12 hours; A precautionary serum gentamicin assay is performed after 24 hours; If the measured serum concentration is less than 5 mg/l, irrigation continues with 20 mg gentamicin in 20 ml saline solution every 6 hours. If it is greater than 5 mg/l, the gentamicin dose is reduced (although we have never found this to be necessary). (5) Concurrently, intravenous gentamicin (80 to 120 mg once daily), penicillin (1 megaunit four times daily), and metronidazole (500 mg three times daily) are administered. (6) Gentamicin levels are assayed three times a week at times related to systemic administration. Predose concentrations are maintained below 2 mg/l and peak concentrations (60 minutes after the dose) between 4 to 6 mg/l, by adjusting the single intravenous dose. (7) Regular cultures of the drain effluent are made, and the antibiotic regimen is changed according to sensitivities of the pathogens isolated. (8) Irrigation is continued until three consecutive effluent cultures are sterile or as prompted by other events, such as isolation of Candida albicam from the effluent. (7)

3 90 3/lorris et al. JOURNAL OF VASCULAR SURGERY ]tdy 1994 Fig. 1. Technique for irrigation of infected graft. Three 5 mm silicone drainage tubes are placed to lie between native aneurysm sac and prosthesis, each directed toward anastomosis. For nonbifurcated prostheses, two tubes are used. Tubes enter via arteriotomy in sac, which is sutured snugly around tubes with absorbable sutures. Each tube has several side holes cut in part that lies inside sac. Drains are removed sequentially. (8) Most patients have been maintained on oral antibiotics for several weeks or months. The duration of this therapy has not been formalized. Table II shows the details of the treatment schedules for all patients in this series. Deviations from the developing protocol have occurred and will be discussed. RESULTS A summary of the results of this series of irrigation therapy for major graft infections is shown in Table III. Actual survival rates for the series are 90% at 30 days, 80% at 1 year, and 67% at 4 years. Two patients died as a result of their graft infections, and the third died at 10 months of a myocardial infarct, with the graft proven to be uninfected at postmortem examination. There have been no amputations performed in the series. Only two patients have required a second operation (one for an ischemic colostomy and one for a displaced irrigation catheter). The mean postoperative stay was 32 days (range 21 to 66 days). The mean follow-up for the seven survivors is 61 months after cessation of irrigation (range 17 to 84 months, as of September 1993). All survivors have been closely reviewed as outpatients and have tradergone regular CT or indium scanning to confirm absence of infection. Gentamicin was used as the first-fine irrigation antibiotic in six cases in this series, clavulanate potassium-potentiated amoxicillin in three and clindamycin in one. Irrigation therapy with multiple antibiotics was used in four cases, the other agents being penicillin, metronidazole, or cefotaxime. The duration of irrigation therapy was a mean of 23 days, with a range of 12 to 56 days. Concurrent multiple intravenous antibiotic therapy was administered in nine cases for a mean of 20 days, range 9 to 35 days. Longer term oral administration was continued in six patients (one indefinitely). In case 10, Enterococcusfaecalis highly resistant to gentamicin (no zone to a 100 ~g gentamicin disc)

4 Volume 20, Number 1 x~ioryis et al. 91 was isolated from the effluent fluid throughout the first 10 days of irrigation with gentamicin. This patient was also receiving intravenous metronidazole, penicillin, and gentamicin, and had previously been treated with parenteral gentamicin. At day 14, penicillin (1 megaunit three times daily) was substituted as the irrigation antibiotic and the intravenous antibiotics were replaced by oral ciprofloxacin, amoxicillin, and metronidazole. Drainage cultures became sterile after 9 days of this regimen, which was continued for a total of 13 days. Cultures taken at the time of placement of the irrigation tube drains in case 1 grew Escherichia coli, but C. albicans was isolated in heavy growth from the drainage effluent 1 day later. Amphotericin B was therefore added to both the irrigation (20 mg per day) and intravenous (20 mg per day) regimens. Effluent cultures continued to grow C. albicans for 10 days, but irrigation and intravenous therapy was stopped 3 days later, and the patient made a good recovery, with no evidence of persistent fungal infection. Irrigation therapy was unsuccessful in two patients, but both had complicating factors: Case 2 was admitted 70 days after undergoing aortoiliac bypass grafting and sigmoid colectomy with a fistula be tween the colonic anastomosis and the aortoiliac graft-the procedures were undertaken to repair a false aneurysm eroding the sigmoid colon after an iliac endarterectomy and Dacron inlay patch carried out 7 years before. After irrigation therapy the patient recovered well and remained free from graft infection for 27 months. He was readmitted with the symptoms and signs of recurrent graft infection, but he had recently had a major myocardial infarct, so was not fit for surgical intervention. Without this complicating factor, further lavage and irrigation therapy would have been possible. He continued to deteriorate and died a few days later. Case 6 had undergone repair of a ruptured abdominal aortic aneurysm 1 month previously, at which sigmoid diverticular disease was noted. He was admitted with an established graft infection, diagnosed by plain abdominal radiography (air outlining the graft) and CT scanning. 27 The graft was explored, and drains were placed. However, the drains became displaced, rendering treatment ineffective, and a further drain was placed on the nineteenth day of treatment. Two days later the patient died of a secondary hemorrhage. This patient was believed to have ongoing sepsis caused by a persistent large bowel fistula, and his treatment had failed to irrigate the graft adequately. One patient (case 7) with chronic kidney failure caused by polycystic disease underwent successful kidney transplantation 1 year after treatment. He remains well with good renal function and is free from recurrent infection more than 4 years later, despite immunosuppressive therapy. Another patient (case 3) underwent emergency coronary artery bypass grafting and aortic root replacement 5 years after treatment and remains well and free from infection 2 years later. Another patient (case 8) had an uncomplicated vein bypass of a popliteal aneurysm 1 year after treatment and is also well and free from infection 4 years later. DISCUSSION This is the first reported series of successful conservative management of aortic grafts extensively infected by enteric organisms with medium- to long-term follow-up. Its success has confirmed the encouraging short-term results published in our earlier report. 22 Conventional treatment of major aortic graft infection is by graft removal and extraanatomic bypass. Many attempts have been made to reduce the considerable morbidity and mortality rates of this procedure. Successful graft excision, debridement, and in situ replacement with a mortality rate as low as 17% was reported by Walker et al.28 However, a later series from the same unit presented less optimistic results, with a mortality rate of 83% in severe graft infections. 29 The major causes of death were recurrent sepsis and secondary hemorrhage. Prolonged hospital stay (Lorentzen et ai.4 reported a mean of 90 days), multiple operations (1.4 to 2.6), and high amputation rates (11% to 37%) are all sequelae of conventional treatment. 2-5,2,s One alternative technique to improve outcome involves placing gentamicin beads in the infected field or adjacent to the graft for i to 2 weeks. This method has been described in small series and isolated case reports in combination with graft excision to lower the risk of aortic stump blowout, and as a conservative method of treating localized groin infections. 31,32 The technique is substantially less successful as a conservative treatment of more extensive aortic graft infections. 33 Irrigation with antibiotics and antiseptics has been well described in the successful treatment of deep-seated infections in orthopedic surgery. 34 The technique has been adapted for the treatment of vascular graft infections. Kesults of a number of small series have been generally encouraging, but only for very limited infections of vascular grafts, particu-

5 92 Morris et al. July 1994 Table II. Microbiology and treatment Patient Drain/irrigation tubes Culture result at tube placement (& during irrigation) Antibiotics Irrigation Duration 1 3 x 5 mm silicone mm silicone, gentamicin beads mm silicone 4 3 x 5 mm silicone mm silicone mm silicone 7 1 x 5 mm silicone 8 3 x 5 mm silicone 9 2 x 5 mm silicone 1 x 15 mm silicone 10 3 x 5 mm silicone E. coli (C. albicans) (Citroba~ter sp.) I~bsiella sp. Proteus mirabilis S. faecalis E. coli E. coli S. faecalis, E. coli, Bacteroides sp. Clostridium sp. Proteus sp. Bacteroides sp. S. faecalis, E. coli (Peptog.OCCUS) Bacteroides fragilis (Candida albicam) S. faecalis, Enterobacter cloacae, B. fragilis (HLGR Enterococcus faecalis) Gentamicin, penicillin Amphotericin Augmentin Augmentin, cefotaxime Penicillin, gentamicin Augmentin Gentamicin, metronidazole, ampicillin Clindamycin Gentamicin Gentamicin Gentamicin, penicillin 14 days 13 days 20 days 39 days 16 days 56 days 12 days 18 days 18 days 12 days 14 days 13 days HLGR, High level gentamicin resistant (resistant to a 100 ~g gentamicin disc). larly those involving staphylococcal infections of groin incisions and the underlying prosthetic graft limbs, sss9 Experience of irrigation therapy with antibiotics/antiseptics in panprosthetic infections is extremely limited, with single case reports forming the basis of the literature. *,41 The results of our series are better than those quoted for traditional graft excision, and the method is considerably less demanding technically. No limbs have been lost, and multiple operations have been avoided. Hospital stay has been kept to a minimum. Although detailed costings are not available for patients from the British National Health Service, the technique appears to offer substantial economic benefits. Six of the seven survivors remain clinically (and on CT evidence) free of infection at 4 to 7 years after treatment. These patients fulfil the criteria of an assumed cure attributed to Szilagyi et al.2_ freedom from infection at 2 years. Szilagyi et al., however, also noted that in rare instances an apparently cured infection may flare up at a much later stage. Most of our patients had received parenteral antibiotic therapy before the operation to place the irrigation tube drains was performed. Because of the frequent involvement of colonic flora in major aortic graft infection, and concerns over the renal function of these patients, this prior therapy frequently included a broad-spectrum cephalosporin plus metronidazole. In four cases operative cultures grew enterococci, which are naturally resistant to cephalosporins. The combination of penicillin, gentamicin, and metronidazole for treatment of our latter patients was chosen primarily to treat enterococci, and also to cover the possible involvement of cephalosporinresistant coliforms. Intravenous penicillin can bc given in large doses and has good tissue penetration, but if applied locally is likely to cause sensitization. 42 Gentamicin is potentially toxic if prolonged, highlevel systemic treatment is required. Furthermore, [3-1actams and gentamicin may mutually inactivate if mixed in solution. 43 For these reasons, we have now standardized a regimen of penicillin given parenterally in high dose, and gentamicin by irrigation with a single daily intravenous dose to achieve therapeutic systemic, and high local concentrations. Serum gentamicin assay was performed three times weekly immediately before and 1 hour after these daily doses. Predose serum concentrations were maintained below 2 mg/l, and peak concentrations between 4 and 6 mg/l by adjusting the intravenous dose. No patient treated with this regimen had apparent deterioration in renal function as measured by serum urea or creatinine concentrations. The decision to cease irrigation therapy, and remove the tube drains, has necessarily been in part arbitrary, but we now recommend a period of irrigation of at least 14 days. Monitoring leucocyte counts and the erythrocyte sedimentation rate is an unreliable guide, because we have found these often only return to normal after several weeks of treatment, significantly after the drains are removed.

6 Volume 20, Number 1 Morris et al. 93 Parenteral Oral Antibiotics Duration Antibiotics Duration Amphotericin 13 days Cefuroxime, metronidazole i dose Cefbtaxime, gentamicin 35 days Augmentin Indefinite Penicillin, gentamicin 18 days - - Piperacillin, cefuroxime 21 days Augmentin 20 months Gentamicin, metronidazole 9 days - Vancomycin, imipenem, amikacin Ampicillin, gentarnicin 18 days 28 days Metronidazole Amoxicillin 3 months 1 month Gentamicin, penicillin, metronidazole 14 days Metronidazole 6 days Gentamicin, penicillin, metronidazole 14 days Metronidazole, amoxicillin, ciprofloxacin 4 months Table IlL Outcome Latest figllow-up Patient Status Cause of death graft status Follow-up investigations Duration Comment I Dead Myocardial infarction Uninfected at PM CF scanning 10 months 2 Dead Sepsis Graft reinfected CT scanning 27 months Recent myocardial infarction, unfit for surgery 3 Alive Patent, uninfected Annual CT scanning, 84 months indium scanning 4 Alive Patent, uninfected Annual CT scanning 70 months 5 Alive R limb occluded and Indium scanning 77 months excised, uninfected 6 Dead Secondary hemorrhage Graft infected CT scanning 21 days 7 Alive Patent, uninfected Annual C1 ~ scanning, 64 months Indium scanning 8 Alive Patent, uninfected Annual C'F scanning 59 months 9 Alive Patent, uninfected Annual CT scanning 53 months 10 Alive Patent, uninfected CT scanning 17 months PM, Postmortem examination; R, right. Drains displaced, inadequate therapy Kidney transplant after 11 months Indications to switch to oral continuation therapy include three sterile consecutive daily effluent cultures, and isolation of Candida sp. We suspect that late isolation of yeasts represents distal colonization of the drains, which should therefore be removed to prevent subsequent deep infection of the graft. C. albicans was grown from the drainage effluent of case i in heavy growth from the first postoperative day. Fungal infections of arterial prostheses are rare, but have previously been reported to carry a grave mortality rate unless the graft is excised with extraanatomic bypass, and 6 to 8 weeks of intravenous amphotericin B is given. 44 Combined irrigation and intravenous therapy with amphotericin B for 2 weeks was successful in case 1. Thus this technique appears to be effective in fungal and bacterial infections. Enterococci are recognized as important causes of aortic graft infection. 4 High-level (resistant to a 100 ~g gentamicin disc) aminoglycoside resistant fecal streptococci are an increasing nosocomial infection problem, and are resistant to synergistic killing by penicillin plus gentamicin. 4s Patient 10 remains free of infection 16 months after irrigation with penicillin (given, despite the theoretic risk of sensitization, to achieve exceptionally high local concentrations) and

7 lour2q~ OF VASCULAR SURGERY 94 IVlorris et al. July 1994 prolonged oral therapy with amoxicillin and ciprofloxacin. Our experience of this technique is mainly limited to early infections, but we suggest it has a place in the treatment of both early and late panprosthetic graft infections with anastomotic involvement, providing suture line integrity is preserved. We have no experience of the technique if disruption has occurred. Some form of more radical combination treatment is obviously required in these cases. Minor peripheral graft infections may be treated by this method or other previously described conservative techniques. Additionally, our experience suggests that a number of factors are of vital importance: correct drain placement is essential, and positions should be checked by contrast studies if in doubt. Effluent fluid should be cultured regularly, with antibiotic therapy adjusted according to the results without delay, and possible overgrowth of yeasts monitored. In conclusion, this method of treatment, in appropriate cases, appears to offer the dual advantages of a superior outcome in terms of infection-free survival and limb loss, while requiring a single, less hazardous surgical procedure for a critically ill patient. We believe that the technique may represent a significant improvement in the management of this major complication of vascular surgery. Clearly the number of patients presented is small, but we believe that the results reported here warrant further investigation in a larger scale multicenter study. The following surgeons (in addition to the authors) and centers have participated in this series: Professor Sir Roy Calne, Addenbrooke's Hospital, Cambridge, U.K.; Mr. J. F. Colin, Norfolk and Norwich Hospital, Norwich, U.K.; Mr. J. A. Dormandy, St. George's Hospital, London, U.K.; Mr. R. M. Heddle, Kent and Canterbury Hospital, Canterbury, U.K. REFERENCES 1. Hoffert PW, Gensler S, Haimovici H. Infection complicating arterial grafts. Arch Surg 1965;90: Szilagyi DE, Smith RF, Elliott JP, Vrandecic MP. Infection in arterial reconstruction with synthetic grafts. Ann Surg 1972; 176: Goldstone J, Moore WS. Infection in vascular prostheses: clinical manifestations and surgical management. Am J Surg 1974;128: Lorentzen JE, Nielsen OM, Arendrup H, et al. Vascular graft infection: an analysis of sixty-two graft infections in 2411 consecutively implanted synthetic vascular grafts. Surgery 1985;98: O'Hara PJ, Hertzer NR, Beven EG, Krajewski LP. Surgical management of infected abdominal aortic grafts: review of a 25-year experience. J VAsc SURG 1986;3: Earnshaw JJ. Infection after vascular reconstruction-hard graft for surgeons. Surg Infect 1991;3: BlaisdeU FW, Hall AD. Axillary-femoral artery bypass for lower extremity ischemia. Surgery 1963 ;54: Louw JH. Splenic-to-femoral and axillary-to-femoral bypass grafts in diffuse atherosclerotic occlusive disease. Lancet 1963;I: Conn JH, Hardy JD, Chavez CM, Fain WR. Infected arterial grafts: experience in 22 cases with emphasis on unusual bacteria and technics. Ann Surg 1970;171: Trout HH, Kozloff L, Giordano JM. Priority of revascularization in patients with graft enteric fistulas, infected arteries, or infected arterial prostheses. Ann Surg 1984;199: Yeager RA, McConneU DB, Sasaki TM, Vetto RM. Aortic and peripheral prosthetic graft infection: differential management and causes of mortality. Am J Surg 1985;150: Goldstone J. The infected infra-renal aortic graft. Acta Clair Scand Suppl 1987;538: Jamieson GG, DeWeese JA, Rob CG. Infected arterial grafts. Ann Surg 1975;181: Spanos PK, Gilsdorf RB, Sako Y, Najarian JS. The management of infected abdominal aortic grafts and graft-enteric fistulas. Ann Surg 1976;183: Liekweg WJ, Greenfield LJ. Vascular prosthetic infections: collected experience and results of treatment. Surgery 1977; 81: Comfier JM, Ward AS, Lagneau P, Janneau D. Infection complicating aortoiliac surgery. J Cardiovasc Surg (Torino) 1980;21: Casali RE, Tucker WE, Thompson BW, Read RC. Infected prosthetic grafts. Arch Surg 1980;115: Downs AR, Lye CR, MacKean G. Graft infections in aortoiliac arterial reconstructions. Can J Surg 1983;26: , Ricotta J}', Faggioli GL, Stella A, et al. Total excision and extra-anatomic bypass for aortic graft infection. Am J Surg 1991;162: Quifiones BW, Hernandez JJ, Moore WS. Long-term results following surgical management of aortic graft infection. Arch Surg 1991;126: O'Brien T, Coffin J. Prosthetic vascular graft infection. Br J Surg 1992;79: Quick CR, Vassallo DJ, Colin JF, Heddle RM. Conservative treatment of major aortic graft infection. Eur J Vasc Surg 1990;4: Haaga JR, Baldwin GN, Reich NE, et al. CT detection of infected synthetic grafts: preliminary report of a new sign. AJR Am J Roentgenol 1978;131: Stevick CA, Fawcett HD. Aortoiliac-graft infection: detection by leukocyte scan. Arch Surg 1981;116: Mark A, Moss AA, Lusby R, Kaiser IA. CT evaluation of complications of abdominal aortic surgery. Radiology 1982; 145: Lawrence PF, Dries DJ, Alazraki N, Albo DJ. Indium 1 l l-labeled leukocyte scanning for detection of prosthetic vascular graft infection. J VASC SURG 1985;2: Wright IF, VassaUo DJ, Dundas DD. The aortic pole. Br J Radiol 1988;61: Walker WE, Cooley DA, Duncan JM, Hallman GJ, Ott DA, Reul GJ. The management of aortoduodenal fistula by in situ replacement of the infected abdominal aortic graft. Ann Surg 1987;205: Jacobs MJ, Reul GJ, Gregoric I, Cooley DA. In-situ replace-

8 Volume 20, Number 1 Morris et al. 95 ment and extra-anatomic bypass for the treatment of infected abdominal aortic grafts. Eur J Vasc Surg 1991;5: Reilly LM, Stoney RJ, Goldstone J, Ehrenfeld WK. Improved management of aortic graft infection: the influence of operation sequence and staging. J VASC SURG 1987;5: Bailey IS, Bundred NJ, Pearson HJ, Bell PR. Successful treatment of an infected vascular graft with gentamicin beads. Eur J Vasc Surg 1987;1: Reilly DT, Grigg MJ, Mansfield AO. Intraperitoneal placement of gentamicin beads in the management of prosthetic graft sepsis. J R Coil Surg Edinb 1989;34: Nielsen OM, Noer HH, Jorgensen LG, Lorentzen JE. Gentamycin beads in the treatment of localised vascular graft infection--long term results in 17 cases. Eur J Vasc Surg 1991;5: Kawashima M, Torisu T, Kamo Y, Iwabuchi A. The treatment of pyogenic bone and joint infections by closed irrigationsuction. Clin Orthop 1980;148: Popovsky J, Singer S. Infected prosthetic grafts. Local therapy with graft preservation. Arch Surg 1980;115: Almgren B, Eriksson I. Local antibiotic irrigation in the treatment of arterial graft infections. Acta Chir Scand 1981;147: Kwaan JH, Connolly JE. Successful management of prosthetic graft infection with continuous povidone-iodine irrigation. Arch Surg 1981;116: Edwards MJ, Richardson JD, Klamer TW. Management of aortic prosthetic infections. Am J Surg 1988;155: Moran KT, Jewell ER. Local antiseptic treatment of infected prosthetic vascular grafts in the groin. Br J Surg 1988;75: Knight CJ, Famell MB, Hollier LH. Treatment of aortic graft infection with povidone-iodine irrigation. Mayo Clin Proc 1983;58: Francois F, Thevenet A. Conservative treatment of prosthetic aortic graft infection with irrigation. Ann Vasc Surg 1991; 5: Lambert HI', O'Grady FW. Peulcillins. In: Antibiotic and Chemotherapy. 6th ed. Edinburgh: Churchill Livingstone, 1992: Wallace SM, Chan LY. In vitro interaction ofaminoglycosides with beta-lactarn penicillins. Antimicrob Agents Chemother 1985;28: Doscher W, Krishnasastry KV, Deckoff SL. Fungal graft infections: case report and review of the literature. J VAsc SURG 1987;6: Murray BE. The life and times of the enterococcus. Clin Microbiol Rev 1990;3: Submitted Sept. 28, 1993; accepted Dec, 28, 1993.