Stent Grafts For Abdominal Aortic Aneurysms

Stent grafts for abdominal aortic aneurysms

INDICATIONS FOR INTERVENTION — An abdominal aortic aneurysm is defined as an aortic diameter at least one and one-half times the diameter measured at the level of the renal arteries. The normal value at this level is approximately 2.0 cm (range 1.4 to 3.0 cm) in most individuals; a diameter greater than 3.0 cm is generally considered aneurysmal. The natural history is one of expansion and eventual rupture. Aneurysms expand on average about 0.3 to 0.4 cm per year. Elective resection is considered for abdominal aortic aneurysms that have reached 5.5 cm in diameter, for those that increase in diameter by more than 0.5 cm within a six month interval, or for those that are symptomatic (tenderness or abdominal or back pain). However, the risk of surgical intervention must be weighed against the potential benefit. The perioperative mortality rate for surgical aortic aneurysm repair was 2.7 and 5.8 percent in two trials. Mortality increases substantially when surgery is performed on an emergent basis due to rupture, and in the setting of comorbid conditions such as advanced age, renal insufficiency, cirrhosis, or cardiopulmonary disease. These issues are discussed in detail separately. (See "Natural history and management of abdominal aortic aneurysm", section on Surgery versus watchful waiting). The concept of endovascular repair of abdominal aneurysms developed from the desire to reduce morbidity and mortality and to provide an alternative to patients who cannot undergo standard surgical therapy. Since the technique is less invasive, it was also speculated that endovascular repair would be less expensive than conventional open surgery. In 2005, the American College of Cardiology/American Heart Association (ACC/AHA), published guidelines on the diagnosis and management of peripheral arterial disease in collaboration with major vascular medicine, vascular surgery, and interventional radiology societies [1]. The following recommendations were made for the choice of intervention: •

Open surgical repair was recommended for patients at low or average risk of operative complications.

•

Endovascular repair was suggested in patients at high risk of complications from open operations.

•

Endovascular repair may be considered in patients who are not at high surgical risk, but evidence of benefit is less well established in this setting. The use and potential role of endovascular repair in patients with an abdominal aortic aneurysm are discussed here. General issues regarding the management of abdominal aortic aneurysm, and the clinical features and diagnosis of this condition, are presented separately. (See "Natural history and management of abdominal aortic aneurysm" and see "Epidemiology, clinical features, and diagnosis of abdominal aortic aneurysm"). ENDOGRAFTS — Endovascular repair consists of insertion of an endograft into the lumen of the aneurysm that effectively excludes the aneurysm from flow through the aorta, thereby minimizing the risk of rupture. First generation endografts were "home made," constructed from conventional stents with fabric sewn onto them. There are now multiple commercially produced systems available. Although there are significant variations from brand to brand, each device consists of three key components:

•

A delivery system to allow placement via a femoral artery approach

•

An attachment system that forms a blood tight seal between the graft and the abdominal aorta

•

Graft fabric, which acts as a conduit for blood flow that has been diverted from the diseased segment of aorta

Endografts are described by terms that characterize their method of assembly in vivo, configuration, and support structure. Assembly in vivo — Devices that are assembled in vivo may have a modular or unibody design; these terms refer to how the bifurcated endograft is constructed. •

A unibody device is placed as one piece; access from the contralateral groin is used to pull the second iliac limb down.

•

Modular devices consist of a main body component and a contralateral limb component assembled within the aorta. Configuration — Tube endografts are used when an distal neck (the distance that separates the inferior extent of the aneurysm from the aortic bifurcation) is present in the aorta, permitting a distal attachment site in the aorta rather than the iliac arteries. However, most aneurysms do not have a suitable distal neck for a tube endograft. In these cases, an aorta-uni-iliac device, which has a distal attachment site in one iliac artery, may be used. It is placed in conjunction with a femoral artery-tofemoral artery bypass and occlusion of the contralateral iliac artery to prevent retrograde filling of the aneurysm. Bifurcated endografts with distal attachment sites in both iliac arteries may also be used. Support structure — Supported units, such as the Talent, contain a metallic framework that adds support throughout the length of the device (show figure 1). The EVT Ancure device is an example of an unsupported model. This device is primarily composed of flexible graft fabric; metallic support is present only at the sites where the device is anchored to the aorta [2]. There is at present no clear advantage of one design over the other. Proponents of the unsupported design claim that it is more plastic, being better able to adapt to changes in aneurysm configuration with time. On the other hand, some physicians feel that supported endografts are less prone to kinking and subsequent thrombosis. Future trials will resolve this issue. INITIAL RADIOGRAPHIC EVALUATION — Although the vast majority of abdominal aortic aneurysms are infrarenal, less than 50 percent are amenable to endovascular repair due to anatomic considerations [3]. Angiography and computed tomography (CT) are commonly obtained as initial radiographic studies to determine both the feasibility of an endograft and the appropriate size and configuration of the endograft. Angiography — The usefulness of angiography alone is limited because of errors resulting from parallax and magnification. Since the inner lumen is imaged, but not the wall of the aorta, angiography cannot evaluate the true lumen diameter, extent of thrombus, plaque, and calcification. Errors in length measurement may occur when catheters follow the shortest distance around curves. Computed tomography — Although CT images both the inner and outer vessel lumen, errors in measurement may occur due to volume averaging. In addition, measurements of diameters can be problematic if the aorta is angulated and the longitudinal axis is not perpendicular to the imaging plane. In these situations, vessel diameter tends to be overestimated. (See "Principles of conventional and helical CT scanning"). CT angiography — CT angiography with three dimensional reconstruction (3D CTA) allows measurements perpendicular to the true axis of the aorta and is therefore more accurate (show radiograph 1). Since angiography is more sensitive for measuring small, axially oriented vessels and vascular stenoses than CTA, combining angiography with spiral CT and 3D CTA provides the maximum amount of information; it also limits error due to intrinsic limitations of each modality [4]. ANATOMIC CONSIDERATIONS — Endografts are constructed specifically for individual patients to meet size criteria dictated by the size and configuration of the aneurysm. Manufacturers require certain measurements, determined by the above radiographic tests, to appropriately tailor the endograft. Proximal neck length — The proximal neck of an aneurysm is the length of normal aorta measured from the lowest renal artery to the most superior extent of the aneurysm. This segment of aorta is where the proximal attachment site of the endograft is placed. The minimum length needed varies from device to device but may be as long as 15 mm. Newer designs allow suprarenal fixation and require a shorter proximal neck length. Qualitative assessment of the proximal neck is also important. Ideally, it should be a normal appearing segment of aorta, without abundant thrombus or heavy circumferential calcification. A vessel that is

lined with thrombus or heavily calcified may not provide a suitable attachment site for a device that uses hooks or barbs. Distal neck length — The distal neck of an aneurysm is the distance that separates the inferior extent of the aneurysm from the aortic bifurcation. A tube endograft may be placed if a distal neck is of sufficient length and diameter as specified by the manufacturer. As an example, the Ancure device requires a distal neck at least 12 mm in length and no greater than 26 mm in diameter [5]. Aneurysm diameter — The diameters of the proximal and distal attachment sites for the endograft need to be measured since devices are available in different diameters. Undersizing the diameter of the endograft may lead to an inadequate seal and failure to exclude the aneurysm, while oversizing may lead to kinking of the device, causing a nidus for thrombus formation or a leak. Results from some clinical series suggest that endovascular repair may be less successful in patients with larger aneurysms. In a report from a European registry of patients undergoing endografting (the EUROSTAR registry), postoperative complications, postoperative mortality, late rupture, and aneurysm-related death were all more frequent in patients with a preoperative aneurysm diameter ≥6.5 cm. Similar observations were made in a series of patients from the Cleveland Clinic; patients with aneurysms ≥5.5 cm in diameter had poorer survival and greater risk of aneurysm-related death at 24 months. Angulation — Angulation is defined as the angle formed between the vertical plane and a line that transects the long axis of either the neck or the aneurysm. An angle of 60º or more leads to difficulties in implantation, kinking, leakage, and the possibility of downward migration of the device. As a result, severe angulation has been considered to be a contraindication to device placement. Iliac attachment site — The majority of patients with aneurysms have anatomy that is unsuitable for a tube endograft, largely because of an inadequate distal anastomosis site. In this setting, the distal attachment sites will be in the iliac arteries. Manufacturers specify a minimal length and maximum diameter of healthy iliac artery required to provide an adequate attachment. A wider or shorter vessel may contribute to an ineffective seal and lead to an endoleak. The common iliac artery is the preferred attachment site, although the external iliac artery may be used. When the attachment site is in the external iliac artery, the hypogastric artery will be covered by the endograft. If the hypogastric artery is patent, the potential exists to back fill the aneurysm via back flow through this vessel into the sac, resulting in ineffective exclusion of the aneurysm. This problem can be prevented by preprocedural embolization of the hypogastric artery. Femoral artery diameter — To accommodate the delivery system of most devices, a minimal femoral artery diameter of 8 mm is usually required. Focal narrowing and mild angulation can be overcome with standard guidewire and catheter techniques, while diffuse narrowing or abundant calcification is more problematic. When access is attempted, but the device fails to pass, an iliac conduit can be created through a retroperitoneal incision to allow device insertion and implantation [8]. Accessory renal arteries — Accessory renal arteries are present in up to 30 percent of the population and commonly originate from the lumbar aorta [9]. Exclusion of an accessory renal vessel by an endograft can result in partial renal infarction. Inferior mesenteric artery — The inferior mesenteric artery is not uncommonly occluded in patients with an abdominal aortic aneurysm; in these cases, excluding the inferior mesenteric artery with an endograft will be of no importance. However, when the inferior mesenteric artery is patent and there is significant stenosis of the superior mesenteric artery, the inferior mesenteric artery may supply important collateral blood flow to the bowel. In such patients, covering a patent inferior mesenteric artery with an endograft may compromise blood flow to the bowel. Thus, endograft placement is contraindicated [8]. PATIENT FOLLOW-UP — The size of the abdominal aortic aneurysm following endovascular repair and endograft placement is followed with angiography, plain film of the abdomen, and CT scan. •

Angiography is performed in the operating suite once the device has been placed to document successful exclusion of the aneurysm.

•

Follow-up protocols vary, but a typical schedule after uncomplicated endograft placement consists of abdominal plain films before discharge and at one, six, and 12 months, and then every year thereafter. Abdominal plain films are an economical and quick way to evaluate the integrity of the graft and the stability of graft appearance, alignment, and position.

•

CT scans, which are obtained on a similar schedule, are used to evaluate the diameter and volume of the aneurysm and to look for signs of endoleak or of endograft migration (show radiograph 2) [10]. (See "Endoleaks" below and see "Device migration" below).

Abdominal ultrasonography has been used with mixed success for the detection of complications [11,12] and is not recommended for routine follow-up [13]. Because it is expensive and invasive, late angiography is reserved for the evaluation of specific problems such as decreased limb flow or thrombosis or a documented endoleak, or to measure aneurysm sac pressure when an enlarging aneurysm is identified in the absence of an endoleak [14]. In the setting of a known complication such as continued aneurysm growth or an endoleak, the frequency of any of the above examinations may be increased to help determine when intervention is necessary. COMPLICATIONS — Complications that have been reported with endograft use include vascular injury during deployment (sometimes leading to aneurysm rupture), inadequate fixation or sealing of the graft to the vessel wall, stent frame fractures and separations, and breakdown of the graft material. These problems occur with varying frequency depending upon the specific device. Such complications have led to public health notifications in the United States and the United Kingdom, and several devices have been withdrawn from the market [13,15,16]. Even after an aneurysm is successfully treated with an endograft, it remains a dynamic entity. The aneurysm will eventually thrombose and, by 12 months, approximately 50 percent of aneurysm sacs have shrunk in diameter. The integrity of an endograft may be sensitive to the changing configuration of the aneurysm that applies mechanical stresses; as an example, changes in the aneurysm may lead to angulation, kinking, thrombosis, or migration of the endograft (show radiograph 3 and show radiograph 4A-4B). Such changes can be evaluated by abdominal plain films and should prompt careful evaluation by CT because an unexpected device failure may be detected [17-19]. Endoleaks — Endoleak is a term that describes the presence on angiography of persistent flow of blood into the aneurysm sac after device placement, indicating failure to completely exclude the aneurysm from the aortic circulation. If not treated, they are associated with a continued risk for aneurysm expansion or rupture. The management of some types of endoleak remains controversial, although most can be successfully occluded with surgery, further stent implantation, or embolization [20]. Four types of endoleaks have been defined, based upon their proposed etiology (show figure 2) [21]: Type I — A type I endoleak, which occurs in 0 to 10 percent of endovascular aortic aneurysm repairs, is due to an incompetent seal at either the proximal or distal attachment site [21,22]. Etiologies include undersizing of the diameter of the endograft at the attachment site and ineffective attachment to a vessel wall that is heavily calcified or surrounded by thick thrombus. Although such a leak can occur immediately after placement, a delayed type I endoleak may be seen on follow-up studies if the device is deployed into a diseased segment of aorta that dilates over time, leading to a breach in the seal at the attachment site. Type I endoleaks must be repaired as soon as they are discovered because the aneurysm sac remains exposed to systemic pressure, predisposing to aneurysmal rupture, and spontaneous closure of the leak is rare [21,22]. If discovered at the time of initial placement, repair may consist of reversal of anticoagulation and reinflation of the deployment balloon for an extended period of time. These leaks may also be repaired with small extension grafts that are placed over the affected end. These methods are usually sufficient to exclude the aneurysm. Conversion to an open surgical repair may be needed in the rare situation in which the leak is refractory to percutaneous treatment [5]. Type II — Type II endoleaks are the most prevalent type, occurring in 10 to 25 percent of endovascular aortic aneurysm repairs [22], and describe flow into and out of the aneurysm sac from patent branch vessels [21]. They are most often identified on the postprocedural CT, appearing as collections of contrast outside of the endograft, but within the aneurysm sac. The most frequent sources of type II endoleaks are collateral back flow through patent lumbar arteries and a patent inferior mesenteric artery. Because the sac fills through a collateral network, the endoleak may not be visualized on the arterial phase of CT scanning; thus, delayed imaging is required. The significance and management of type II endoleaks is controversial. Some investigators argue that since spontaneous resolution occurs in 30 to 100 percent of cases, a "wait and see" approach is preferable, while carefully following aneurysm volume and morphology on CT imaging [22,23]. However, systemic pressures have been noted within the aneurysm sac in the presence of type II endoleaks, indicating a more tenuous situation. For this reason, we and other investigators recommend occluding type II endoleaks that have not spontaneously thrombosed within one month [23]. Type III/IV — Type III and type IV endoleaks are much less common [21]. •

Type III endoleaks represent flow into the aneurysm sac from separation between components of a modular system, or tears in the endograft fabric (show radiograph 5)

•

Type IV endoleaks are due to egress of blood through the pores in the fabric Type IV leaks heal spontaneously, while type III leaks are repaired with an additional endograft to eliminate systemic flow and pressure in the aneurysm. Postimplantation syndrome — Patients undergoing endovascular stent graft placement often experience an acute inflammatory syndrome characterized by fever, leukocytosis, elevation of serum C-reactive protein (CRP) concentration, and perigraft air during the first week to 10 days after implantation [24,25]. Elevated levels of endotoxin and interleukin-6 have been detected after the procedure [26], and platelet activation has also been demonstrated [27]. The etiology of the postimplantation syndrome is unknown, but it does not appear to be due to infection. Device migration — Device migration is one of the major causes of secondary intervention after endovascular aneurysm repair [28]. If untreated, potential complications include endoleak, aneurysm expansion, and rupture. The mid-term and long-term incidence of migration with two different endografts (AneuRx and Zenith) was evaluated in a series of 130 patients with fusiform, infrarenal aneurysms and a minimum follow-up of 12 months [29]. Migration was detected by CT scan and was defined as caudal movement of ≥5 or ≥10 mm or as any migration with a related clinical event. The following observations were noted:

•

With the AneuRx device, the rate of freedom from device migration (≥10 mm or a clinical event) was 96, 90, 78, and 72 percent at one, two, three, and four years, respectively. Twelve of the fourteen patients underwent 14 secondary procedures (13 endovascular, one open conversion). Initial neck length was shorter in patients with migration (22 versus 31 mm).

•

With the Zenith device, the rate of freedom from device migration (≥10 mm or a clinical event) was 100, 98, 98, and 98 percent at one, two, three, and four years, respectively. The single patient with stent migration did not require therapy.

•

Significant (≥3 mm) aortic neck dilation occurred in 22 percent. The authors concluded that careful surveillance of stent migration was an essential component of follow-up in patients receiving stent grafts. Because of shorter follow-up with the Zenith graft, caution was observed in the apparently lower incidence of stent migration with this device. CLINICAL OUTCOME Short term — The short term technical success rate for endovascular aneurysm repair ranges from 83 to over 95 percent [24,30-32]. In one study, for example, an endoluminal stent graft was inserted in 154 patients with an abdominal aortic aneurysm [24]. Primary success, defined as complete exclusion of the aneurysm from the circulation, was achieved in 87 percent. Primary failures were mostly due to leaks, and 2 percent of patients required open surgical repair. Minor or major complications occurred in 10 percent. A 2007 systematic review identified four randomized trials of 1532 patients who were considered suitable candidates for either endovascular or open repair of nonruptured abdominal aortic aneurysms larger than 5.0 cm in diameter [33]. The 30 day all-cause mortality was significantly lower with endovascular repair (1.6 versus 4.8 percent, relative risk 0.33, 95% CI 0.17-0.64). The two principle contributing trials to this systematic review were EVAR 1:

•

The EVAR 1 trial included 1082 patients who were at least 60 years of age, with aneurysms at least 5.5 cm in diameter [34]. At 30 days, mortality was significantly lower with endovascular than with open repair (1.6 versus 4.6 percent, adjusted odds ratio 0.34, 95% CI 0.15-0.74). Endovascular repair was also associated with a significantly shorter hospital stay (7 versus 12 days), although more secondary interventions (additional surgical procedures) were required with endovascular repair (9.8 versus 5.8 percent).

•

The DREAM trial evaluated 345 patients with aneurysms of at least 5 cm in diameter [35]. There was an almost significant trend toward lower operative mortality with endografting than with surgery (1.2 versus 4.6 percent, risk ratio 0.26, 95% CI 0.03-1.10). Moderate and severe systemic complications (cardiac, pulmonary, renal) were more frequent with open repair (26 versus 12 percent), while moderate and severe local vascular or implant-related complications were more frequent with endovascular repair (16 versus 9 percent).

The short-term survival advantage of endovascular repair appears to be much greater when endovascular repair is limited to patients at highest risk from open surgery. This was illustrated in a report of 454 consecutive patients who underwent elective repair (206 endovascular and 248 open surgery) of an abdominal aortic aneurysm [36]. The overall 30-day mortality rates not significantly different for endografting and surgery (2.4 and 4.8 percent, respectively). However, among patients at highest surgical risk (American Society of Anesthesiologists class IV), the 30-day mortality rates were much lower with endovascular repair (4.7 versus 19.2 percent with open surgery). Long term — Long term outcomes of patients who have undergone endovascular repair of abdominal aneurysms have been evaluated with and without comparison to patients who have undergone open repair. Survival compared to open repair — The early survival benefit seen with endovascular repair compared to open repair described above is lost between one and four years, after which survival appears equivalent. This observation was seen in the 2007 systematic review discussed above, which included follow-up of 1473 patients from the DREAM (two years) and EVAR 1 randomized trials (four years) [33], as well as from a report using data from the United State Medicare program [37]. The DREAM trial evaluated 345 patients with aneurysms of at least 5 cm in diameter; there was an almost significant trend toward lower operative mortality with endografting than with surgery (1.2 versus 4.6 percent, risk ratio 0.26, 95% CI 0.03-1.10) [35]. The following findings were noted at two years [38]: •

There was no difference in cumulative survival (89.7 versus 89.6 percent with surgical repair), an effect that was present by one year.

•

The cumulative rate of aneurysm-related death was lower with endovascular repair (2.1 versus 5.7 percent), which was entirely due to a lower rate of perioperative mortality. However, late ruptures are likely to be missed if autopsy is not performed [39].

•

The likelihood of freedom from severe events was nonsignificantly higher with open repair (83.1 versus 80.6 percent), which was again due to a lower rate of perioperative events; there were no documented postoperative aneurysm ruptures.

•

The rate of reintervention at nine months was significantly higher with endovascular repair (11 versus 4 percent, hazard ratio 2.9). Thereafter, the reintervention rates were parallel. Thus, the perioperative survival advantage associated with endovascular repair was no longer apparent by one year. Potential caveats are that there were only 38 deaths overall and the trial was neither designed nor powered to assess long-term outcomes. Findings virtually identical to those in DREAM (lower perioperative mortality with endovascular repair that did not persist at two years) have been noted in two retrospective studies that compared endovascular and surgical repair [40,41]. The larger EVAR-1 trial, which randomly assigned 1082 patients to endovascular or open repair reported comparable overall survival in both groups four years after randomization [42]. Patients treated with endovascular rather than open repair had a lower incidence of aneurysm-related mortality (4 versus 7 percent), but experienced significantly more complications during follow-up (41 versus 9 percent). When the need for reintervention was included in the final analysis, the total cost associated with endovascular repair was greater than the cost of open surgical repair. Possible explanations for the apparently greater risk of late mortality with endovascular repair in these trials include chance, precipitation of death with open repair in high-risk patients who are more likely to die in the first year with endovascular repair, and failure to prevent aneurysm rupture with endovascular repair [39]. Using the United States Medicare database, long-term survival was evaluated in 22,830 matched pairs of patients who underwent elective repair with either open or endovascular technique between 2001 and 2004 [37]. There was a significantly lower rate of perioperative mortality with endovascular repair (1.2 versus 4.8 percent), a benefit that was higher with increasing age. However, mortality at three to four years was nearly identical in the two groups. In terms of late complications, reintervention related to the aneurysm (mostly minor) was more common after endovascular repair (9.0 versus 1.7 percent with open repair), while laparotomy-related reintervention and hospitalizations were more common after open surgery (9.7 versus 4.1 percent). Survival compared to no repair — The EVAR-2 trial included 338 patients who were evaluated for inclusion in EVAR-1 but were judged unfit for open repair; the patients were randomly assigned to endovascular repair or observation [43]. In this very ill group, there was no difference in all-cause or aneurysm-related mortality after four years.

Other outcomes — Observational studies have evaluated the rate of complications following endovascular repair [44,45]. Data from the EUROSTAR registry on 2464 patients suggest a risk of aneurysm rupture after endovascular repair of approximately 1 percent per year [45], which is similar to the rate in patients with a 4 to 5 cm aneurysm who have not undergone surgery [46-48]. Rupture was significantly more likely in patients with an endoleak or with graft migration or kinking. The overall failure rate of the procedure, including death, aneurysm rupture, or requirement for subsequent surgical repair, was 3 percent per year. Another report from the EUROSTAR registry evaluated the rate of secondary intervention in 1023 patients who were followed for at least one year [28]. A secondary intervention, mostly for migration or rupture, was performed in 18 percent at a mean of 14 months after the original endograft repair. The rate of freedom from secondary intervention at one, three, and four years was 89, 67, and 62 percent, respectively. The presence of an endoleak may predict continued aneurysm expansion. This was illustrated in a review of the literature in which a 24 percent prevalence of endoleaks was noted [49]. Among the patients with endoleaks, aneurysms grew in a majority and failed to shrink in many patients. Spontaneous thrombosis (and therefore closure of the endoleak) occurred in only 21 percent. As noted above, long-term complication rates also appear to be greater in patients who have a larger preoperative aneurysm diameter (see "Aneurysm diameter" above). CONCLUSIONS — Endovascular repair of an abdominal aortic aneurysm represents a potential alternative to open surgical repair. However, the precise role of endografts has yet to be clearly defined. The preprocedural evaluation of patients undergoing endovascular repair requires careful scrutiny of both quantitative and qualitative measures of aortic anatomy. Some patients are not suitable candidates for endografting because of the site, extent, or morphology of the aneurysm. (See "Initial radiographic evaluation" above and see "Anatomic considerations" above). The short-term morbidity and mortality rates of endovascular therapy compared favorably with those of open surgical repair in both randomized trials [34,35] and large observational studies [50-52]. The benefit is greatest for patients at highest surgical risk in whom the short-term mortality is significantly lower with endovascular repair (4.7 versus 19.2 percent at 30 days in one report) [36]. Other shortterm benefits of endovascular repair include reductions in major morbidity, intubation time, and intensive care unit and total hospital stay. Compared to patients undergoing conventional surgical repair, those undergoing an endovascular repair have less blood loss, recover more quickly, and return to normal function faster. (See "Short term" above). However, it is not clear that endovascular repair is preferable to open surgical repair in the long term, even among patients at highest risk for surgery [53]. In both the DREAM trial and two observational studies, the perioperative mortality benefit of endovascular compared to surgical repair was no longer present at one to two years [38,40,41]. Other concerns include the increased need for reintervention, the significance of endoleaks, the potential for graft migration or kinking, and the risk of aneurysm rupture [28,38,45]. (See "Complications" above and see "Long term" above). In patients who do undergo endovascular repair, diligent follow-up is required. Such patients should have imaging studies annually to evaluate the status of the endograft; these annual assessments should be performed for the remainder of the patient's life [53], since there is appreciable rate of secondary intervention [28]. (See "Patient follow-up" above). As noted above, endograft complications have raised concerns with respect to some specific devices, leading to public health notifications and the withdrawal of some devices from the market [13,15,16]. Confidence in the reliability of available data has not been enhanced by a report that the manufacturers of one device had used threats of legal action to prevent publication of a paper on mortality rates [54]. In 2005, the American College of Cardiology/American Heart Association (ACC/AHA), published guidelines on the diagnosis and management of peripheral arterial disease in collaboration with major vascular medicine, vascular surgery, and interventional radiology societies [1]. The following recommendations were made for the choice of intervention: •

Open surgical repair was recommended for patients at low or average risk of operative complications.

•

Endovascular repair was suggested in patients at high risk of complications from open operations.

•

Endovascular repair may be considered in patients who are not at high surgical risk, but evidence of benefit is less well established in this setting.