Aortoiliac Disease: Open Extra-Anatomic Bypass

Femorofemoral Bypass

Femorofemoral bypass was first described as a stand-alone procedure in 1952, but the first substantial series of patients with follow-up information was published in 1960. The procedure involves diversion of some blood from one iliac system to the contralateral leg and depends on the concept that one healthy iliac artery has the capacity to supply adequate flow to both the ipsilateral donor leg and the contralateral recipient leg. Femorofemoral bypass has been a frequently applied procedure in patients with dominant unilateral iliac artery disease and more recently in patients whose suboptimal “donor” iliac artery can be improved with endovascular techniques. Even more recently, endovascular techniques have continued to improve and many patients who would have been treated with femorofemoral bypass may now be completely treated with endovascular techniques. Femorofemoral bypass has found a new application as an accompaniment to aortouni-iliac endovascular aortic aneurysm repair and seems to perform better in that application than when used for chronic arterial occlusive disease. ^, One should also consider the more direct common iliac to femoral bypass, which appears to have the same performance as the benchmark aortofemoral bypass procedure. ^,

If there is any question about the adequacy of the inflow because of disease in the iliac arterial system one should consider a pharmacologic physiologic stress test of the inflow, ^, treat the diseased inflow with endovascular techniques as discussed above, or abandon the plan for femorofemoral bypass in favor of an alternative procedure. With respect to outflow, either a minimally diseased normal caliber deep femoral or minimally diseased superficial femoral artery is sufficient outflow in most cases to support graft patency (as discussed further below) and adequately improve perfusion of the leg. It is rarely necessary to add a simultaneous infrainguinal bypass procedure on the recipient side to enhance distal perfusion, at least to treat resting ischemia of the recipient limb.

Femorofemoral bypass is performed with the patient in the supine position and can be performed with general, regional, or even with local anesthesia in selected patients. Prophylactic antibiotics are appropriate in elective cases and targeted antibiotics are appropriate for patients with active infection who require urgent treatment. The operation may be performed with a limited area of sterile preparation to include only the lower abdomen, groins, and upper anterior thighs, but in most cases we would perform the operation with preparation and draping to expose essentially the entire anterior abdomen to allow access to the iliac arteries or even the aorta should this become necessary. The preparation and draping of the patient may also include circumferential exposure of both legs to allow intraoperative assessment of perfusion to the level of the ankles upon completion of the operation. In most cases, longitudinal groin incisions provide the most flexibility for exploring and inspecting the donor and recipient femoral arterial systems. A graft is tunneled from one groin to the other in an inverted U configuration, usually in the subcutaneous area (anterior to the abdominal wall fascia), but on rare occasion in the retrofascial space. This tunnel can be created with a specialized tunneler or simply with aggressive finger dissection and a large clamp. Care must be taken to prevent entering the peritoneal cavity and in the case of previous abdominal surgery, great care must be taken when tunneling through scarred areas to prevent visceral injury. The first ever femorofemoral bypass was performed with an endarterectomized occluded superficial femoral artery autograft. In current practice, despite the vast majority of femorofemoral bypasses being performed with prosthetic materials, there is a small fraction of procedures performed with venous or arterial autografts, or even allografts when the infection is extensive or in patients with intestinal or urinary stomas. Although multiple prosthetic graft types have been touted as superior to their alternatives, there is no compelling evidence that polyester fabric, supported polyester, ePTFE, supported ePTFE, heparin bonding or other graft lining or any other graft type performs better than others in terms of hemodynamics or patency in the femorofemoral application. ^, Notwithstanding, a recent publication demonstrated a trend toward improved patency using femoral vein compared to prosthetic conduit in a uniquely large femorofemoral bypass experience using vein conduit. The author’s conduit preference is a 6-mm or 7-mm supported ePTFE graft, depending on patient size and native arterial diameter.

Care should be taken to prevent kinking of the graft in the anteroposterior direction, especially in patients with protuberant abdomens. Technical nuances such as shortening of the anastomosis, placing the anastomosis more distally (possibly to include the deep femoral or superficial femoral arteries), or both will allow a less acute angle at the heel of the anastomosis and decrease the possibility of graft kinking. Another key point is to create the subcutaneous tunnel in a more uniform larger radius arc, generally by making the most superior part of the graft well superior to the pubis ( Fig. 110.1 ) to reduce the risk of kinking in the coronal plane.

Intravenous heparin, or in the case of patient heparin sensitivity another suitable short-acting anticoagulant is administered prior to placement of vascular clamps. Arteriotomy and anastomotic sites are selected based on review of the preoperative arteriogram and local examination of the artery at the time of exploration of the femoral arteries. These may include the common, deep, or superficial femoral artery or some combination thereof. A conventional end graft-to-side artery anastomosis as described in earlier chapters in this text is nearly always appropriate on both sides. The anastomosis may be facilitated by extending the arteriotomy and the toe of the graft onto the deep or superficial femoral artery or even placing the entire anastomosis to the deep or superficial femoral artery. The right and left anastomoses may be performed simultaneously, thus saving significant time if two surgeons are available. If the anatomy dictates, extension of the anastomosis out onto the deep femoral artery, may actually reduce the tendency of the graft to kink adjacent to the anastomosis as well. Once the anastomoses have been completed and the clamps removed, a sterile continuous wave Doppler probe should be used to confirm that flow is enhanced in the recipient side outflow artery or arteries and that there is no apparent deterioration in flow in the donor side outflow arteries. Once again, as long as there is a normal diameter, minimally diseased or adequately improved donor iliac artery, the graft should not “steal” from the donor leg at rest.

Some authors have observed that femorofemoral graft patency is better if the superficial femoral artery is patent on the recipient side, but the author and others have observed no difference. ^, We make sure that there is adequate outflow to at least the deep or superficial femoral artery, using adjunctive techniques such as using the graft anastomotic hood to perform “patch angioplasty”, and insuring that a 3.5-mm diameter dilator will pass easily beyond the toe of the anastomosis into at least one of the recipient femoral arteries. We have found that this is generally associated with adequate patency.

Patients are treated long-term with antiplatelet agents and hospital stay is determined primarily by comorbidities. Some patients may be ready for discharge on postoperative day one or two. Although the literature is sparse with respect to the value of surveillance noninvasive testing, we have adhered to a program of surveillance with ankle–brachial indices, continuous wave Doppler waveforms at the ankle, and graft duplex scans. Our protocol calls for a graft duplex every three months for the first year, twice in the second year, and yearly thereafter if graft performance is stable, similar to our protocol for infrainguinal bypass grafts.

Operative mortality for femorofemoral bypass should be much less than 5% and should be lower for femorofemoral than for aortofemoral bypass given otherwise similar patients. A femorofemoral graft does not as perform as well as a direct aortofemoral bypass based on patency or hemodynamic improvement. In particular, femorofemoral bypass may not be a hemodynamically complete treatment for claudication. However, as measured by freedom from major adverse limb events, femorofemoral bypass is nearly if not as effective as aortofemoral bypass and may be a better choice than direct aortofemoral bypass based on patient factors. Predicted primary patency at 5 years for patients whose indication is atherosclerotic occlusive disease is approximately 70% based on assessment of recent pertinent publications and is probably better in patients having femorofemoral bypass as part of aortouniiliac endovascular repair of aortic aneurysm. ^, Finally, femorofemoral bypass has become a key salvage technique for limb thrombosis after endovascular repair of aortic aneurysm. Causes of femorofemoral bypass graft failure include technical errors, progression of inflow or outflow disease, and perianastomotic stenosis due to intimal hyperplasia. There are few complications specific to femorofemoral bypass, but they would include visceral injury, particularly if the graft is tunneled in a retrofascial position or if there is a hernia. In difficult circumstances with a high-risk patient, a femorofemoral bypass is a relatively low risk way to reperfuse a leg with dominant unilateral iliac artery disease and is likely to be an essential skill for the vascular surgeon for the foreseeable future.

Axillofemoral Bypass

The first reported axillofemoral bypass procedures were performed to treat patients at unacceptable risk for conventional aortofemoral bypass and for extraanatomic bypass at the time of removal of an infected graft. These are examples of two of the three primary indications for axillofemoral bypass: (1) patients with symptomatic aortic and/or bi-iliac arterial occlusive disease thought to be at excessively high physiologic risk for direct aortic repair; (2) patients with infected native aorta or aortic graft or the closely related problem of aortoenteric fistula; and (3) patients with “hostile abdomen,” generally with multiple previous surgeries, active intra-abdominal infection, or the presence of intestinal or urinary stomas. Although there was a period of popularity of axillofemoral bypass as a primary procedure for aortoiliac occlusive disease, for example as promoted by Johnson et al. and later by the group at the Oregon Health Sciences University, it is fair to say that this approach was not widely adopted. There was also a period of interest, for example at the Albany Medical College, in axillofemoral bypass combined with ligation of the infrarenal aorta as primary treatment for abdominal aortic aneurysm, but this approach was generally abandoned when it was observed that a number of such patients later died of ruptured abdominal aortic aneurysm. Conventional thinking in the peak period of the operation was that axillofemoral bypass should nearly always be performed in a bifemoral configuration to improve outflow of the graft. However, some of the same people who advocated this later reported that axillounifemoral bypass performed just as well as axillobifemoral bypass configurations. Thus the choice of axillouni- vs. axillobifemoral bypass should be based on whether one or both legs would benefit from improved flow.

Many patients with aortic and/or bi-iliac arterial occlusive disease are now treated completely or partially with the ever-expanding endovascular armamentarium. Furthermore, it is likely that some of these patients who might have been selected for axillofemoral bypass in the past are now seen as manageably low risk for direct aortofemoral bypass. Thus, axillofemoral bypass has likely declined in volume from a likely peak in the 1970s and 1980s. Nevertheless, this procedure remains an essential tool for the vascular surgeon. Similar to the assumption with femorofemoral bypass, axillofemoral bypass assumes that one axillosubclavian artery has adequate blood volume flow capacity to supply the donor side arm and one or both legs, at least at rest.

One unique preprocedure question is which axillary artery is to be used for the donor? The graft is generally tunneled in the midaxillary line and lateral abdomen from the donor iliac artery to the ipsilateral femoral recipient artery. Certainly coexistent thoracic or abdominal problems may dictate the side of the procedure. For example, the impending need for thoracotomy or abdominal surgery for other problems may dictate use of the contralateral axillary artery. The presence of intestinal or urinary stomas may dictate placement of the axillofemoral graft contralateral to those stomas. Some surgeons ask the patient on which side they sleep and will place the graft contralateral to that, although the need for this questioning is equivocal. If none of the above seems to point to one or the other side, most surgeons would place the graft on the side of the more symptomatic leg if axillounifemoral configuration is to be used. We would recommend against placement of an axillofemoral bypass based on the side of an existing arteriovenous hemoaccess fistula, since this might provoke or worsen existing steal symptoms in that arm. We insist on a triphasic brachial artery Doppler waveform and a blood pressure no more than 10 mm Hg lower in the proposed donor arm than in the contralateral arm. If these criteria are not met then we will proceed to upper extremity angiography and if necessary inflow artery treatment prior to surgery.

The procedure is performed in the supine position. Although it may be possible to perform segments of the operation with local or regional anesthesia, creating the axillofemoral tunnel would be difficult and for practical purposes the operation is always performed with general anesthesia. We place a gel pad or a towel-wrapped 1 liter bag of fluid between the operating table and the posterolateral back on the side of the axillofemoral graft to elevate the flank and lower chest to allow slightly more posterior prepping and draping for better visualization during tunneling. Some surgeons position the ipsilateral arm at the patient’s side, but we have always positioned the ipsilateral arm at 90° abduction both to allow visualization of the ipsilateral lateral chest and abdomen during tunneling and to elevate the clavicle slightly to enhance exposure of the donor axillary artery. The sterile preparation should be wide and should include at a minimum the anterior and ipsilateral neck base and supraclavicular fossa over to the anterior shoulder area, the axilla, anterior and ipsilateral chest and abdomen and the groin and superior anterior thigh. In particular, we have always thought it prudent to prep widely to allow a thoracotomy in the case of an injury or other problem with the axillosubclavian donor artery, although we have never had to proceed to thoracotomy. In the case of axillobifemoral bypass the preparation and draping must include the contralateral lower abdomen, groin, and anterior upper thigh. Prophylactic or, in the case of active infection, targeted antibiotics are administered.

Groin exposure is consistent with that for femorofemoral bypass as discussed above. The donor axillary artery is exposed with a transverse incision a few centimeters inferior to the clavicle and by splitting of the pectoralis major muscle fibers. The axillary artery should be exposed and controlled from the clavicle medially to the pectoralis minor muscle. This is typically about 3–4 cm in length. This artery is more fragile than the more familiar common femoral artery and great care must be taken not to injure the artery during dissection and clamp placement.

A subcutaneous tunnel is created from the ipsilateral groin incision to the axillary artery exposure incision. This tunnel must be kept in the midaxillary line to minimize changes in path length during flexion and extension of the torso and to avoid graft compression at the costal margin, which is often more prominent anterior to the midaxillary line. In the early experience with axillofemoral bypass a separate flank incision was required about half the distance from the axillary artery exposure incision to the ipsilateral groin incision since there were no adequate tunnelers to traverse the entire distance. However, there are now several tunnelers that will traverse the entire distance and the intermediate incision in the flank can be abandoned. Care must be taken to remain subcutaneous and superficial to the abdominal wall deep fascia and to avoid entering the thoracic cavity during tunneling. Although some have advocated tunneling anterior to the pectoralis minor muscle or even division of the pectoralis minor, we have always tunneled posterior to the pectoralis minor.

Axillofemoral bypass is virtually always performed with a prosthetic graft and as with femorofemoral bypass there is no clear advantage of one graft type over another. Although both polyester and ePTFE grafts can be purchased with pre-constructed connections between the axillofemoral and femorofemoral components, given the current relative infrequency of axillofemoral bypass, we perform the procedure with two segments of straight graft anastomosing one graft to the other ( Fig. 110.2 ). We tend to use an 8-mm supported ePTFE graft for the axillofemoral component and if the bypass is to be axillobifemoral, we typically use a 6-mm supported ePTFE graft for the femorofemoral component. On rare occasions when performing an axillounifemoral bypass in a small patient, a 6 mm diameter graft may be used in part to try to enhance velocity, which may improve patency. ^,

Heparin or other suitable short-acting anticoagulant is administered before clamps are placed on the arteries. As with femorofemoral bypass, conventional end graft-to-side artery anastomoses are usually appropriate. An axillobifemoral bypass with two separate graft segments requires four anastomoses including one that connects the axillofemoral and femorofemoral components and pairs of these anastomoses may be performed simultaneously. Thus, axillofemoral bypass, particularly axillobifemoral bypass, is greatly facilitated if two surgeons are present.

It is critical to create the axillary artery to graft anastomosis in such a way that there is no tension on this anastomosis when the arm is abducted. We employ two technical concepts to try to avoid this problem. First, the anastomosis should be placed as medial as possible on the axillary artery since this part of the artery will move less than the more lateral part of the artery when the arm is abducted. This also tends to bring the graft into a more nearly parallel relationship to the artery, also reducing the tendency to produce traction on the anastomosis when the arm is abducted. The principle of leaving the pectoralis minor muscle intact and tunneling the graft posterior to this muscle tends to force the surgeon to keep the anastomosis medial. A second principle is to allow some redundancy of the graft and even to allow some bowing of the graft into the axilla posterior to the pectoralis minor muscle, again hoping to reduce the risk of traction on the anastomosis with ipsilateral arm movements.

In the “heyday” of axillobifemoral bypass a commonly recommended technical point was to place the connection between the femorofemoral and axillofemoral components as close to the femoral artery anastomosis as possible to try to maximize flow throughout the entire length of the axillofemoral graft. This concept is plausible and we adhere to it, but there is no observational evidence that this provides a real advantage. Nevertheless it is usually simple to either complete a conventional femorofemoral bypass and then sew the axillofemoral graft to a graftotomy in the ipsilateral anastomotic hood of the femorofemoral graft or alternatively to complete the anastomosis between the inferior end of the axillofemoral graft and the ipsilateral groin target artery and then sew the ipsilateral side of the femorofemoral graft to a graftotomy in the femoral anastomotic hood of the axillofemoral component (as is depicted in Figure 110.3 ). In either case, unlike a pliable native artery where a simple arteriotomy creates an adequate opening, a prosthetic graft tends to be rigid and not to gap adequately after a simple longitudinal graftotomy. Therefore, anastomosing a graft to the side of a prosthetic graft requires excision of an oval of the graft material at the point where anastomosis is to be created. No matter whether the graft is performed in axillounifemoral or axillobifemoral configuration, the same principles described above for femorofemoral bypass to assure adequate outflow (for example using the graft hood as a patch and assuring that a 3.5 mm diameter metal dilator will pass easily into at least one outflow artery) are equally applicable to axillofemoral bypass.

A sterile handheld Doppler probe is used to confirm that flow is augmented in the outflow artery or arteries with the graft open as compared to with the graft clamped. The surgeon must also assure that there is normal flow in the distal arm on the donor side, for example by confirming the presence of a palpable pulse in the radial artery or by confirming satisfactory pulse oximetry in a donor side digit with the graft open (with blood flowing to the lower extremities).

Just as is the case with femorofemoral bypass, an axillofemoral graft does not perform as well as a direct aortofemoral bypass based on patency or hemodynamic improvement. In fact, our previous work suggests that the predicted ankle–brachial index after axillofemoral bypass with normal (undiseased) infrainguinal outflow is less than 0.7, compared to slightly more than 1.0 after direct aortofemoral bypass. However, axillofemoral bypass performs acceptably and may be a better choice than direct aortofemoral bypass based on patient factors.

The literature of axillofemoral bypass has one of the broadest range of reported long-term patencies for any open vascular surgical procedure, probably related to the broad range of patient characteristics in that literature. However, a fair estimate for predicted primary patency at 5 years among patients whose indication is atherosclerotic occlusive disease should be about 60%–70%. In a patient with advanced age and comorbidities, bilateral iliac artery disease that cannot be improved adequately on at least one side with endovascular techniques to allow a femorofemoral bypass, a patient with an infected native aorta or aortic graft, or a “hostile” abdomen, an axillofemoral bypass provides a less invasive approach to reperfusing one or both legs. In practice, patients selected for axillofemoral bypass tend to have both shorter survival and an increased risk of major adverse limb events. ^,

With respect to surveillance of axillofemoral bypass, the literature is even less revealing than for femorofemoral bypass. Just as with femorofemoral bypass, since the predicted patency for axillofemoral bypass is clearly less than for aortofemoral bypass, we are consequently much more aggressive with noninvasive surveillance and obtain ankle pressures, ankle-level Doppler waveforms, and graft duplex scans every three months for the first year, twice in the second year, and yearly after that. Complications specific to axillofemoral bypass include disruption of the axillary artery to graft anastomosis and thromboembolic complications in the native arteries after thrombosis of the axillofemoral graft.