Total Graft Excision and Extraanatomic Repair for Aortic Graft Infection

Historical Background

Many early series recognized that aortic graft infection is difficult to diagnose, is associated with substantial amputation and mortality risks, and has a tendency to recur if inadequately treated. It was recognized early on that inadequately treated aortic graft infection could lead to systemic sepsis or, in propagating to the anastomosis, could lead to pseudoaneurysm and hemorrhage from anastomotic disruption. The 0.2% to 2% incidence of aortic graft infection has remained remarkably constant over the last 4 decades, with an interval between initial graft implantation and symptomatic infection ranging between 41 and 62 months and amputation and mortality rates of 11% and 40%, respectively.

Graft infection may occur at the time of initial surgery because of contamination from skin flora or exposure to bowel organisms through an unrecognized enterotomy during graft tunneling ( Fig. 33-1 ). Other factors associated with the development of aortic graft infection include emergent repairs, lengthy procedures, development of wound complications, or perioperative bacteremia from a urinary tract infection, line sepsis, or pneumonia. Aortofemoral reconstructions are more prone to infection than aortoiliac bypasses, probably because of risk of groin wound necrosis or infection. Although many aortic graft infections can be indolent, others can manifest with protean clinical symptoms of hemorrhage or systemic sepsis.

Graft infections have been categorized as early or late based upon time of presentation after implantation. Early aortic graft infections are often associated with more virulent organisms than those presenting late. Most common causative organisms are Staphylococcus, with Staphylococcus aureus often implicated in early infection and coagulase negative Staphylococcus, such as Staphylococcus epidermis, more prevalent in late infections. Enterococcus organisms, gram-negative organisms, fungi, and anaerobes can also be cultured, especially in the presence of graft-enteric erosion. In many aortic graft infections, however, an organism is not identified because of lack of suitable specimens, because of inadequate culture methods, or because antimicrobial coverage was instituted before the collection of tissue for culture.

Early reports demonstrated that simultaneous graft excision and revascularization were associated with very high mortality, so delaying or omitting revascularization was initially advocated to decrease the magnitude of the procedure, even though it was practical for only a minority of patients. Staged extraanatomic arterial reconstruction followed by graft excision was first reported by Elliott and colleagues in 1974. Objections were initially raised to this approach because of concerns related to the presumed risk of bacterial seeding of the new prosthesis, thrombosis of the extraanatomic graft from competitive flow, and risk of interval hemorrhage from the infected graft in the period prior to graft excision. Reilly and associates subsequently demonstrated in a large series that these concerns were unfounded with use of adequate antibiotic coverage and graft removal within 1 week of the initial extraanatomic bypass.

Preoperative Preparation

• Clinical manifestations. The presentation of aortic graft infection can be variable, and early diagnosis is challenging. Patients with a history of prior aortic graft placement presenting with recurrent fever and chills, pain or pulsatile mass in the abdomen or groin, or evidence of wound infection should be suspected to have aortic graft infection, and those presenting with gastrointestinal hemorrhage, systemic sepsis and shock, or vertebral osteomyelitis usually have advanced infection. The spectrum of gastrointestinal hemorrhage is wide. Suture line fistula can lead to abrupt massive bleeding, whereas graft body erosion into the gastrointestinal tract typically results in low-grade bleeding from the intestinal mucosa. Early presentation of S. aureus or gram-negative infection in the setting of aortoenteric fistula is typically more virulent and can spread rapidly, leading to anastomotic breakdown and exsanguinating hemorrhage. Methicillin-resistant S. aureus has become recognized as a particularly dangerous pathogen because of its destructive characteristics that result in high mortality and risk of limb loss. Indolent, late prosthetic infections may present with limb thrombosis, pseudoaneurysm formation, or fluid collections around poorly incorporated grafts without systemic symptoms and typically involve S. epidermis. This organism secretes a glycoprotein biofilm, which makes its detection difficult unless broth culture and graft sonification are used to isolate the organism. Physical examination may reveal exposed graft material in a contaminated wound or abscess cavity. Graft thrombosis, anastomotic pseudoaneurysm, and septic emboli to the lower extremities are consistent with but not diagnostic of chronic graft infection.

• Laboratory findings. Leukocytosis, elevated inflammatory markers, positive blood cultures, and anemia may be present but are nonspecific findings.

• Upper gastrointestinal endoscopy. Gastroduodenoscopy to the fourth portion of the duodenum should be performed on patients with a history of aortic graft placement presenting with gastrointestinal bleeding to exclude an aortoenteric fistula. This is demonstrated by exposed graft material if a graft-enteric communication is present ( Fig. 33-2 ).

  

• Computed tomography (CT) imaging. CT imaging is the initial diagnostic modality of choice, because it can assess of the extent of graft involvement, as well as provide percutaneous guidance if aspiration of a fluid collection for microbial analysis is required. CT scanning is reported to identify aortic graft infection with a sensitivity of 94% and a specificity of 85%. Characteristic CT findings suggesting the presence of late aortic prosthetic graft infection include periprosthetic gas or fluid, soft tissue thickening, hydronephrosis, vertebral body degeneration, and bowel wall thickening. Although the presence of fluid or loculated air around the graft is common in the early postoperative period after graft implantation, CT demonstration of perigraft gas more than 2 months after aortic reconstruction is suggestive of graft infection ( Fig. 33-3 ).

  

• Additional imaging studies. Although radionuclide tagged white cell scanning has been used to identify aortic graft infection, its utility is limited by false-positive results because of uptake in other inflammatory beds. Duplex ultrasonography can identify fluid collections and the presence of a pseudoaneurysm, but its usefulness is limited by bowel gas and body habitus. Magnetic resonance imaging can evaluate inflammatory changes and soft tissue attenuation and can distinguish hematoma from other fluid collections in determining the extent of graft involvement, but it does not add information to that available through a CT scan. Recent studies have suggested that fused fluoro-2-deoxy-d-glucose positron emission tomography (FDG-PET)-CT imaging may improve the diagnostic accuracy of detecting graft infection.

• Extent of graft infection. Determination of the extent of graft infection is important to successful treatment planning. If the entire graft is involved, it will likely require removal in its entirety. However, if only one limb of an aortobifemoral graft is involved, it may be possible to remove only the infected portion and salvage the remaining uninfected graft. Complete graft excision is usually required in early infections, because the graft is not incorporated. However, late infections with solitary limb involvement can be treated with limited graft excision if the rest of the graft is well incorporated. Operative inspection of tissue incorporation into the graft and microbial identification can assist in determining the amount of graft that might be salvaged.

• Assessment of lower extremity perfusion. Preoperative assessment of lower extremity perfusion is necessary to plan treatment and may be based on physical examination and lower extremity noninvasive studies. If the infected prosthesis is occluded and the lower extremities are viable, excision of the infected graft material alone should be adequate. However, if lower extremity perfusion depends on the involved graft, extraanatomic bypass is required to avoid amputation once the infected graft is removed. CT or conventional arteriography can identify patent vessels available for revascularization of the lower extremities, if required once the infected graft is removed.

• Nutritional assessment. Patients with aortic graft infection are often elderly and debilitated, especially if the process is chronic. Adequate nutritional support is critical to meet the metabolic demands of sepsis and a prolonged operative and postoperative course. In the presence of aortoenteric fistula, enteral nutrition may be delayed for a lengthy period. The transition to enteral nutrition may be facilitated by the use of feeding gastrostomy or jejunostomy.

Pitfalls and Danger Points

• Failure to control hemorrhage. Careful preoperative planning should determine options for proximal and distal arterial control and assess the extent of possible involvement of neighboring venous structures.

• Failure to control sepsis. Removal of all infected grossly foreign material is mandatory to minimize the risk of persistent or recurrent infection. Careful management of the aortic stump with an omental flap and oversewn femoral vessels with a sartorius muscle flap after graft removal is a critical component of effective treatment.

• Failure of revascularization. Assessment of options for revascularization that bypass the infected field, including the need for an axillofemoral, axilloprofunda, or axillopopliteal bypass, is essential to ensure adequate lower extremity perfusion.

Operative Strategy

Critical concepts central to the treatment of aortic synthetic graft infection include excision of the infected graft material, debridement of infected tissue, provision of adequate drainage to control local sepsis, revascularization of critical vascular beds, and appropriate antibiotic therapy for primary infection and to prevent secondary infection in the aortic bed, as well as the extraanatomic bypass.

Selection of an Extraanatomic Bypass

The selection of an extraanatomic bypass depends upon the extent of the underlying graft infection, the extent of associated occlusive disease, and the type of aortic reconstruction requiring removal. If the patient has had a prior aortobiiliac graft and the entire graft is involved, then an axillobifemoral graft is the optimal selection to provide lower extremity perfusion ( Fig. 33-4 , A ). The arm with the highest systolic blood pressure should be used for inflow for the extraanatomic bypass, but if both arm blood pressure measurements are equal, the right axillary artery is preferred, because right subclavian artery occlusive disease is less common than left subclavian artery involvement. This configuration also has the advantage of preserving the option to approach the aorta through the left retroperitoneal approach, if needed in the future. However, if the patient has an infected aortobifemoral graft requiring prosthetic removal from both femoral anastomotic sites, then bilateral axillofemoral bypass grafts tunneled through clean tissue planes to the superficial femoral or above-knee popliteal arteries may be necessary ( Fig. 33-4 , B ). Extension to the popliteal artery can usually be accomplished through a lateral subcutaneous tunnel extending across the distal thigh that traverses medially to the popliteal space at a level above the knee. After graft removal, the resulting femoral arteriotomy sites in the groin incisions are closed with an autogenous patch, such as saphenous vein or endarterectomized superficial femoral artery, and if needed are covered with a sartorius muscle flap.

If only a single distal aortobifemoral graft limb is involved with a groin infection but the remaining graft is uninfected, initial division and ligation of the uninvolved proximal limb is performed through a clean retroperitoneal incision. The distal portion of the limb is left in the wound, unless an obturator bypass is considered, and the incision is closed. Limb perfusion is then be restored by constructing an axillosuperficial femoral artery synthetic bypass graft ( Fig. 33-5 , A ). Alternatively, an obturator bypass graft is tunneled medially through the obturator foramen around the infected groin incision through uninvolved tissue planes ( Fig. 33-5 , B ). After closure and placement of occlusive dressings over the extraanatomic bypass incisions, the infected graft material is then removed from the groin and the anastomotic site in the femoral artery is reconstructed using a patch angioplasty with autogenous material.