Secondary Aortoduodenal Fistulas: Value of Initial Axillofemoral Bypass Patrice Bergeron, MD, Hugo Espinoza, MD, Philippe Rudondy, MD, Michel Ferdani, MD, Jacques Martin, MD, Jean-Michel Jausseran, MD, Robert Courbier, MD, Marseille, France
Between January 1970 and April 1989, 20 patients underwent operation for secondary aortoduodenal fistulas. When the preoperative diagnosis was certain and emergency control of bleeding not required, initial axillofe,~oral bypass was performed before ablation of the infected aortic prosthetic graft during the same operation. When diagnosis was uncertain or severity of bleeding required emergency laparotomy, the therapeutic plan varied over time. Until 1980, we performed either a direct repair (three cases) or the ablation of the aortic graft followed by secondary axillofemoral bypass (four cases). After 1980, the order of procedures was 1) control of bleeding whenever necessary, 2) axillofemoral bypass, and 3) ablation of the aortic graft. Postoperative mortality was two of 13 in patients undergoing initial axillofemoral bypass, compared with six of seven patients undergoing direct surgery or initial ablation of the aortic graft. Of the 12 patients surviving the postoperative period, three died of aortic stump hemorrhage, four, 12, and 14 months after operation. Two patients had a new aortic graft inserted. Repeat replacement of the abdominal aorta graft was performed in one case and ascending thoracic aortobifemoral bypass in the other because of secondary thrombosis of the axillofemoral bypass. We conclude that initial axillofemoral bypass before dealing with the aortic graft improves the immediate prognosis in operations for secondary aortoduodenal fistulas. This procedure does not, however, preclude the possibility of aortic stump infection which can lead to recurrent aortoduodenal fistula. The risk of infection or secondary occlusion of axillofemoral bypass is minimal. Secondary prosthetic replacement is not systematically necessary. (Ann Vasc Surg 1991 ;5:4--7). KEY WORDS:
Aortoduodenal fistula; fistula; axillofemoral bypass.
Secondary aortoduodenal fistulas carry a very high mortality rate and therapeutic options remain controversial [1-4]. In 1981, we reported five cases of secondary aortoduodenal fistula and called attenFrom the Service de Chirurgie Cardio-Vasculaire, Fondation HOpital Saint Joseph, Marseille, France. Presented at the Annual Meeting of the Soci~t~ de Chirurgie Vasculaire de Langue Franfaise, May 18-19, 1990, Nancy, France. Reprint requests: P. Bergeron, MD, Service de Chirurgie Cardiovasculaire, Fondation HOpital Saint-Joseph, 26 bd de Louvain, 13008 Marseille, France.
tion to the value of initial axillofemoral bypass before approaching the aortoduodenal fistula [5]. We now report our experience with 20 patients followed for six months to nine years, 13 of them treated by this two-stage procedure. This policy has become routine since 1980 and, in our experience, has contributed to a sharp decrease in early postoperative mortality. MATERIAL AND METHODS Twenty patients, all men, ages ranging from 50 to 76 years (mean: 66 years) were operated on between
VOLUME 5 N o 1 - 1991
S E C O N D A R Y A OR TOD UODENA L H S TULA S
5
TABLE I.--Summary of twenty patients with secondary aortoduodenal fistulas Arterial disease
Symptoms
Emergency operation
AIOD* AIOD
Sepsis Hemorrhage
No Yes
69
AAA**
Hemorrhage
No
64
AAA
Hemorrhage
Yes
52
AIOD
Hemorrhage
No
60 62
AIOD AIOD
Sepsis Hemorrhage
No Yes
Group II 59
AIOD
Sepsis
No
70
AIOD
Sepsis
Yes
66
AIOD
No
63
AIOD
Hemorrhage Sepsis Hemorrhage
Yes
64
AAA
Hemorrhage
Yes
76
AAA
Hemorrhage
Yes
50
AIOD
No
58
AAA
Hemorrhage Sepsis Hemorrhage
Yes
68
AIOD
Sepsis
No
60
AAA
Hemorrhage
Yes
57
AAA
Hemorrhage
No
73
AIOD
Hemorrhage
No
73
AAA
Sepsis
No
Age Group / 62 days 34
Vascular procedure Suture 1) Ablation aortic graft 2) Axillofemoral bypass 1) Ablation aortic graft 2) Axillofemoral bypass 1) Ablation aortic graft 2) Axillofemoral bypass 1) Ablation aortic graft 2) Axillofemoral bypass Suture Replacement aortic graft 1) Axillofemoral bypass 2) Ablation aortic graft 1) Axillofemoral bypass 2) Ablation aortic graft 1) Axillofemoral bypass 2) Ablation aortic graft 1) Exploratory laparotomy 2) Axillofemoral bypass 3) Ablation aortic graft 1) Exploratory laparotomy 2) Axillofemoral bypass 3) Ablation aortic graft 1) Axillofemoral bypass 2) Ablation aortic graft 1) Axillofemoral bypass 2) Ablation aortic graft 1) Axillofemoral bypass 2) Ablation aortic graft 1) Axillofemoral bypass 2) Ablation aortic graft 1) Axillofemoral bypass 2) Ablation aortic graft 1) Exploratory laparotomy 2) Axillofemoral bypass 3) Ablation aortic graft 1) Ablation aortic graft 2) Axillofemoral bypass 1) Exploratory laparotomy 2) Axillofemoral bypass 3) Ablation aortic graft
Outcome Died recurrent ADF t 10 days Died MOF w 3 days Died MOF 2 days Died recurrent ADF 1 day Died recurrent ADF 4 months Died septicemia 10 days Died MOF 1 day Died recurrent ADF 1 year Died recurrent ADF 15 days Alive 6 years Alive 9.5 years Alive 9 years Alive 3 years Alive 2.5 years Alive 5 years tt Alive 6 months Died recurrent ADF 14 months Alive 16 months Alive 2.5 years w167 Alive 20 months Alive 7 years
*AIOD Aortoiliac occlusive disease t A D F - AorLoduodenal fistula w - Multiple organ failure **AAA = Abdominal aortic aneurysm ttOccluded axillofemoral bypass at 2.5 years w167 bypass at one year
January 1970 and April 1989 for secondary aortoduodenal fistula. All aortoduodenal fistulas were located in the fourth portion of the duodenum. They represented 0.7% of the 2877 aortic operations performed during the same period of time. The initial aortic operation had been performed on the average 4.5 years earlier (range: three months to 14 years). In eight cases, the initial lesion was an aneurysm of the abdominal aorta and in 12 cases, aortoiliac atherosclerotic occlusive disease in which the prosthetic graft was implanted laterally. All patients underwent gastrointestinal endoscopic investigation in the search for aortoduodenal fistula [6].
The mechanism of aortoduodenal fistula was a true fistula in three cases (15%) and a paraprosthetic fistula due to duodenal erosion in 17 cases (85%). Diagnosis w a s made preoperatively in 10 cases (50%): by endoscopy in five cases [6], by fistulography through the groin in two cases, by computed tomography (CT) scanning in two cases and by barium meal in one case. Clinical characteristics of the 20 patients are listed in Table I. Patients were divided into two groups according to the type of treatment performed (Table I). Before 1980 (group I, seven patients), all patients under-
6
ANNALS OF VASCULAR SURGERY
SECONDARY AORTODUODENAL FISTULAS
went an initial abdominal approach through which the aortic graft was removed and then an axillofemoral bypass was performed. Three patients had direct surgery (two sutures and one replacement of the aortic prosthesis). After 1980, (group II, 13 patients), an axillofemoral bypass was performed before treating the aortoduodenal fistula, either as soon as the diagnosis was made if there was no emergency (eight cases) or after laparotomy (five cases) whenever one was required for diagnosis or life-threatening complication. If this were the case, the operation sequence consisted of three operative steps: diagnostic laparotomy and control of bleeding, axillofemoral bypass, and last treatment of the aortoduodenal fistula by duodenal suture and removal of the aortic graft. The aortic stump was closed either directly by a 3/0 polypropylene suture or with fascial reinforcement. Pedicled omental transposition was associated in all cases. The duodenal rent was closed by simple suture in 18 cases and by segmental resection of the duodenum with end-to-end anastomosis in two cases. Axillofemoral bypass was performed with polytetrafluoroethylene (PTFE) prosthetic graft in all cases. In five cases, the bypass was axillobifemoral. In 12 cases, the bypass was done as a biaxillofemoral bypass in order to avoid reopening the groin incision and to allow implantation of the distal end of the extraanatomic graft on either the superficial or the deep femoral artery. In these cases, ablation of the graft was incomplete as the femoral limbs of the prosthesis were initially left intact and then removed eight days later in order to avoid infection of the axillofemoral bypass. Evaluation of results was based on operative and late mortality. Follow-up ranged from six months to nine years.
RESULTS In group I, six of seven patients (85%) died rapidly of recurrent aortoduodenal fistula or multiple organ failure. The only surviving patient died four months later of recurrent hemorrhage due to infection of the aortic stump. In group II, two of 13 patients (15%) died during the initial 15 days of hospitalization of recurrent aortoduodenal fistula. Of the I 1 surviving patients, two aortoduodenal fistulas recurred, 12 and 14 months after operation, leading to death. One further patient had a thrombosed axillobifemoral bypass two years later and was treated by bypass originating from the ascending thoracic aorta. A new prosthetic graft was inserted routinely one year later in one patient. No secondary infection of axillobifemoral bypass was noted. Survival in this
group was 81% (9/11) with a 48 month mean follow-up (range 6 to 98 months).
DISCUSSION Since the first case of aortoduodenal fistula treated successfully by Brock in 1953 [7], mortality has remained high and surgical treatment modalities are still controversial. Certain authors advocate local procedures (simple suture or in situ prosthetic replacement) and have reported surprisingly good results [2,8-10]. The obvious advantage of these techniques is the simplicity of repair compared to multiple procedures. The major risk is prosthetic infection [11] which can lead to repeat hemorrhage, and for certain authors, requires lifelong antibiotic protection [2]. Others like us, prefer to remove the infected graft and to perform an extraanatomic revascularization procedure during the same operation, usually with a biaxillofemoral bypass [3,4,12-15]. The advantages of this technique are the low risk of infection of the new graft, and if infection occurs, its relative ease of treatment. When the aortic graft is implanted end-to-end, the principal drawback of the classical treatment (ablation of the graft followed by axillofemoral bypass) is prolonged ischemia to the lower extremities, which is responsible for the mortality in up to 50% of cases in certain series [3]. Primary axillofemoral bypass precludes ischemia to the lower extremities and reduces the magnitude of consequent general complications. For these reasons, we have adopted the technique of primary axillofemoral bypass beginning in 1980. When the diagnosis is certain, and there are no life-threatening complications, axillofemoral bypass was performed prior to laparotomy to treat the aortoduodenal fistula (two stage operation). By comparison, primary laparotomy was performed either to confirm diagnosis or to control bleeding. The laparotomy wound was then closed with adhesive sterile drapes and followed by axillofemoral bypass and then ablation of the aortic graft (three stage operation). As shown by our results, immediate mortality was considerably reduced. However, this therapeutic policy does not solve all problems. Closure of the aortic stump remains a difficult technical problem [16,17]. Pedicled omental transposition is not sufficient to control local infection and to preclude recurrent hemorrhage (two early and two late repeat aortoduodenal fistulas in our series of 13 patients). The risk of rupture of the aortic stump has been evaluated at 80% at two years by Szi!agyi and colleagues [4]. The diagnosis of aortoduodenal fistula must be certain before performing primary axillofemoral by-
VoLumE 5
No 1 - 1991
SECONDAR Y AORTODUODENAL F I S T U L A S
pass. One patient (not included in this study) with an aortic graft who was taking oral anticoagulant therapy presented with severe upper digestive tract hemorrhage due to duodenal ulcer. He received an unnecessary aortofemoral bypass which had to be explanted at the end of the operation. Secondary infection of the axillofemoral bypass [11,18] does not seem to be frequent if the distal implantation site is made on the superficial or profunda femoris arteries, thus avoiding the femoral triangles, and if it can be performed separately with regard to the septic abdominal operation. Secondary occlusion of the axillofemoral bypass requires secondary direct revascularization either by in situ prosthetic graft or by a graft originating on the thoracic aorta. In our experience [19], long-term patency of an axillofemoral bypass has been satisfactory, making it possible to preserve this type of revascularization procedure in the aged or poor status patient. On the other hand, in the young subject, routine secondary aortic surgery should be considered. Technical solutions for in situ revascularization without preservation or utilization of prosthetic material have been proposed. Extended aortoiliac endarterectomy is possible if the aortic graft was implanted laterally [20]. In the future, aortic allografts, preserved in a sterile milieu at 4~ with antibiotics, might represent the ideal choice for secondary revascularization in aortoduodenal fistula because of their excellent tolerance in the septic setting [21]. The limited number of donors and duration of conservation are possible drawbacks of this alternative. In conclusion, initial axillofemoral bypass in the treatment of aortoduodenal fistula can substantially reduce operative mortality. Recurrence of infection at the aortic stump is not completely obviated. Routine secondary aortic surgery could be the ultimate step to optimal treatment.
REFERENCES 1. BACOURT F, BARRET A, CORMIER M, et al. Fistules aorto-digestives. In: ActualitOs Chirurgicales (84e Congr6s de l'Association Fran~aise de Chirurgie), Tome III, Paris, Masson, 1983; pp 1-22. 2_ EASTCOTT HHG. Aorto-enteric fistula: possibilities for direct repair, in: GREENHALGH RM (ed) Extra-anatomic and Secondary Arterial Reconstruction. London, Pitman, 1982; pp 58-65.
mmm
7
3. TILANUS HW, TERPSTRA OT, DE SMIT P, et al. Diagnosis and management of graft-enteric fistulae. Br J Surg 1988;75:915-916. 4. BERNHARD VM. Management of graft infections following abdominal aortic aneurysm replacement. WorldJ Surg 1980; 4:679-688. 5. COURBIER R, BERGERON P, JAUSSERAN JM, et al. Aorto-enteric fistula. Arch Chit Thorac Cardiovasc 1981 ;3: 4-10. 6. MARTIN J, CANO N, DI-COSTANZO J, et al. Quatre cas d'h6morragies digestives par fistule intestinale paraproth6tique aortique: 6tude des stades endoscopiques et approche thdrapeutique. Acta Endoscopica 1981 ;11:225-232. 7. BROCK RC. Aortic homografting: a report of six successful cases. Guy's Hosp Rep 1953;102:204-228. 8. DAUGHERTY M, SHEARER GR, ERNST CB. Primary aortoduodenal fistula extra-anatomic vascular reconstruction not required for successful management. Surgery 1979; 86:399M01. 9. THOMAS WEG, BAIRD RN. Secondary aorto enteric fistulae: towards a more conservative approach. Br J Surg 1986;73:875-878. 10. WALKER WE, COOLEY DA, DUNCAN JM, et al. The management of aortoduodenal fistula by in situ replacement of the infected abdominal aortic graft. Ann Surg 1987;205: 727-732. 11. YASHAR JJ, WEYMAN AK, BURNARD RJ, et al. Survival and limb salvage in patients with infected arterial prostheses. Am J Surg 1978;135:499-504. 12. BUNT TJ. Synthetic vascular graft infections, lI. Graft enteric erosions and graft enteric fistulas. Surgery 1983;94: 1-9. 13. TROUT HH IIl, KOZLOFF L, GIORDANO JM. Priority of revascularization in patients with graft enteric fistulas, infected arteries, or infected arterial prostheses. Ann Surg 1984;199:66%683. 14. O'DONNELL TF Jr, SCOTT G, SHEPARD A, et al. Improvements in the diagnosis and management of aortoenteric fistula. Am J Surg 1985;149:481~86. 15. GLOCK Y, BOUISSOU E, TASRINI J, et al. Fistules aorto-entdriques et paraproth6tiques: ~t propos de 5 cas. J Chir 1984;121:347-353. 16. BACOURT F, GOEAU-BRISSONN1ERE O, KOSKAS F. Aortic stump closure a new technique. Ann Chit Vasc 1986;1:271-272. 17. SCRIBNER RG, BAKER MS, TAWES RL, et al. Recurrent aortoduodenal fistula. Arch Surg 1977;112:1265 1270. 18. O'MARA C, IMBEMBO AL. Paraprosthetic-enteric fistula. Surgew 1977;81;556-566. 19. BERGERON P, MANGIALARDI N, SARRADON P, et al. Pontages axillo-fdmoraux: r6sultats 5 long terme. J Chir 1989;126:575-582. 20. EHRENFELD WK, WILBUR BG, OLCOTT C IV, et al. Autogenous tissue reconstruction in the management of infected prosthetic grafls. Surge:3' 1979;85:82-92. 21. BAHNIN1 A, KOSKAS F, MARTIN-MONDIERE C, et al. Traitement des infections aortiques post-opdratoires par allogreffe art6rielle in situ: r6sultats prdcoces chez onze malades. Ann Vasc Surg 1991; in press.