Introduction

Open aneurysm repair is a major operation, which reached its zenith some years ago. Population-based studies suggest the morbidity and mortality rates are significant. Mortality may be as high as 8%, and 10% may suffer cardiac complications [1]. One in five patients would not undergo the operation again knowing the recovery process involved [2]. There is room for improvement.

It has been over a decade since the pioneering work of Parodi and Volodos culminated in the first endovascular aneurysm repair (EVAR) [3, 4]. Improving technology and experience have addressed many problems associated with the early devices. Conversion to open repair, as high as 20% in some early reports, is now rare [5, 6]. Long-term results with early-generation devices have been disappointing. There have been numerous secondary interventions and conversions. Ruptures have occurred at a rate of 1% per year, which is the same as for untreated small aneurysms <5.5 cm [7]. However, in randomized controlled trials latest-generation devices have performed equally as well as open repair in short- to midterm follow-up.

Morphological Constraints

Early experiences with EVAR demonstrated that outcome was dependent on the morphology of the aneurysms being treated [8]. These observations were confirmed by later studies with second-generation stent-grafts. In particular, breaching the proximal neck guidelines resulted in a fourfold increase in proximal endoleak. Combined morphological deviations multiplied the risk [9].

Unfortunately only a limited number of patients have a normal artery proximal and distal to the aneurysm. Investigations of aneurysm morphology revealed that about 60% of patients are suitable for EVAR [10, 11]. Larger aneurysms may have more adverse morphology [12]. Adverse proximal neck morphology accounts for the majority of unsuitable aneurysms. Almost a third (29%) of aneurysms have an inadequate neck length (<15 mm) [10]. These limitations of standard EVAR have been a driving force for fenestrated and branch stent-graft technology.

However, it is possible to successfully exclude aneurysms whose morphology is not ”ideal”using standard EVAR. In a series of 13 patients with aortic necks <10 mm long, Greenberg was able to demonstrate that such aneurysms could be excluded using a stent-graft with a suprarenal component [13]. Others have found more complex anatomy possible. Wide necks are a surmountable problem. In a study of 16 patients with wide necks (>30 mm) from Leicester, no endoleaks were observed during a 12-month follow-up period. In addition, the aortic necks appeared to significantly decrease in diameter [14]. More recent controlled studies have also confirmed the ability of standard infrarenal aortic stent-grafts to exclude aneurysms with one or more adverse proximal aortic neck features [15]. All these results are relatively short-term and should be viewed with some caution prior to widespread adoption of adverse morphology.

Improvements in graft technology including a reduction in sheath diameters have increased the applicability of EVAR, allowing more aneurysms to be treated. Delivery systems have increased trackability and conformability, allowing graft insertion through more tortuous iliac arteries. The use of a suprarenal anchor stents has reduced the length of the aortic neck required to provide an adequate seal.

Experimental evidence from Nottingham using an in vitro model (Gianturco-Dacron stent-graft) suggested that an angulation of >30 degrees significantly increases endoleak flow by lifting the graft away from the neck wall [16].

Graft Configuration

Aorto-aortic tube grafts have been largely consigned to history, few patients having the prerequisite distal aortic neck [17]. The uni-iliac configuration was originally developed to facilitate graft manufacture (often by the surgeon) and the use of smaller sheath sizes. It was later demonstrated that more aneurysms were suitable for uni-iliac EVAR than for bifurcated (on account of accommodating unilateral CIA aneurysms) [10]. However, improving technology and the use of coil embolization have allowed more patients to be treated with bifurcated stent-grafts. The bifurcated approach is more favorable because of its ”physiological” configuration.

Fenestrated EVAR

Fenestrated stent-graft technology was developed in order to increase the morphological applicability of EVAR and offer an alternative to open repair. The technique was developed in 1996 and has subsequently been refined into a system (based on the Zenith stent-graft; William A. Cook Ltd., Brisbane, Australia) which is now commercially available [18, 19].

Fenestrated stent-grafts are individually customized and require some weeks to manufacture. Consequently they are not suitable for patients presenting emergently with acute symptomatic or ruptured abdominal aortic aneurysm (AAA). Major refinements in the fenestrated EVAR technique have included the use of a modular system to assist alignment of the stent-graft in the aortic neck and bifurcation. Further improvements, including the use of diameter-reducing restraining ties which limit initial stent-graft expansion to permit alignment of the stent-graft prior to catheterization of target vessels and reinforced fenestrations to facilitate catheterization, likely have made an impact on results.

Reports on fenestrated stent-grafts were published between 2005 and 2006. Sun and co-workers performed a systematic review of the literature from 1999 to 2006 [20]. Short-term results are encouraging. Perioperative target vessel patency rate was 97% (95% CI, 92%–100%) and 90% (85%–95%) during follow-up. No conversions to open surgery were required. Perioperative mortality was 1.1% (0.4% – 2.7%) and the endoleak rate after 30 days was 9.4% (2.6%–16.3%). Long-term data are limited, with only four studies reporting more than 12 months’ mean follow-up.

Iliac Branch Grafts

The consequences of internal iliac artery (IIA) occlusion are variable. Patients may remain asymptomatic, some may suffer buttock claudication, and others may develop frank bowel and pelvic ischemia or infarction. In general, unilateral IIA occlusion is rarely associated with severe ischemic complications [21]. Although rarely associated with severe ischemic complications, when a similar system was used in Leicester, this approach resulted in buttock claudication in 40% of cases [22]. Bilateral IIA occlusion represents a higher risk of complications but does not necessarily result in clinically significant ischemia or infarction due to an extensive collateral pelvic blood supply. In a recent study from St. George’s Vascular Institute bilateral IIA embolization prior to EVAR resulted in claudication in 31% of patients. There were no limb/life-threatening complications [23]. Unsurprisingly, embolization of the main IIA trunk resulted in fewer complications than distal embolization. There was no benefit found from undertaking a sequential compared with a simultaneous approach to embolization.

Predictors of clinically significant pelvic ischemia would be desirable if either unilateral or bilateral IIA is being considered to facilitate EVAR. Endovascular repair of ruptured AAA appears to confer a higher risk of bowel ischemia and infarction [24]. The authors of that study consequently modified their practice, suggesting that no patient having EVAR of ruptured AAA should undergo bilateral IIA occlusion. Presumably the hypoperfusion resulted in the occlusion of collateral vessels.

Mehta et al. retrospectively identified 154 patients who had undergone elective IIA occlusion either unilaterally or bilaterally during either EVAR or open aneurysm repair [25]. They speculated that their low incidence (12%) of ischemic complications was due to the preservation of external iliac artery and common femoral artery collaterals, the avoidance of shock, and distal embolization.

The method of IIA occlusion also appears to play a role in the development of ischemic complications [26]. Endovascular coil embolization increases the possibility of distal ischemia because of the technical difficulties involved in occluding the IIA. In one retrospective study, coil embolization resulted in buttock claudication in 45% of patients, compared to 27% of those who received occlusion of the IIA at its origin through placement of a stent-graft across it [27].

In order to prevent the complications associated with IIA occlusion and increase the number of patients treatable by the endovascular method, iliac branch stent-grafts have been developed. Both currently available systems are based on the Zenith stent-graft and incorporate a small side arm which is cannulated with a wire into the IIA. A covered stent is then placed from the branch graft into the IIA, thus maintaining patency of the IIA. Experience with this system is limited to a small number of experienced centers but initial results are encouraging, with high rates of technical success.

Assessment of Aneurysm Morphology

Preoperative assessment of aneurysm morphology is vital to the successful outcome of EVAR. Spiral CT with intravascular contrast (spiral CTA) remains the most popular method of assessment and provides all the necessary information for successful EVAR [28]. Invasive calibration angiography, once used routinely, is no longer required. Angiography was used to aid the determination of aorto-iliac length and the detection of accessory renal arteries. However, spiral CTA with three-dimensional (3D) multiplanar reconstruction provides adequate prediction of length, although some length measurement discrepancy between the two modalities may occur within large aneurysm sacs [29]. Measurements of aorto-iliac length are much less critical with modular systems. Consequently computer software has been developed to facilitate endograft sizing. They enable the user to accurately and reproducibly predict the required size of endograft from 3D spiral CT reconstructions (e.g., TeraRecon) [30].

Magnetic resonance angiography (MRA) is a valid alternative method of preoperative assessment of aneurysm morphology, which is particularly useful in patients with renal impairment [31]. It remains more expensive than spiral CTA and requires detailed postimaging processing. Additional problems arise with the change of imaging modality which will be required if a stainless-steel endograft is deployed. Stainless-steel stents cause considerable image distortion on MRA.

The value of intraoperative intravascular ultrasound (IVUS) continues to be debated. It is currently expensive and requires adequate training before it can be used effectively. Proponents of the technique suggest that it may almost completely replace periprocedural angiography, with its attendant nephrotoxicity [32]. Others argue that IVUS is an unnecessary luxury [33]. In the latter study, selectively used IVUS was only able to detect one lesion that required treatment and had not been identified on angiography. Clearly it is not vital to the satisfactory outcome in the majority of patients. A more realistic view may be that it is an additional instrument of quality control.

Contrast nephrotoxicity is not an uncommon problem in patients undergoing EVAR, where large volumes of contrast can be used in patients who frequently have pre-existing renal impairment. Strategies to reduce the incidence of contrast nephrotoxicity include preoperative intravenous hydration in addition to the use of carbon dioxide [34]. The role of antioxidants (such as N-acetyl cysteine) in preventing renal impairment continues to be debated. Any beneficial effect is likely to be small.

Anesthesia and Percutaneous Stent-Graft Delivery

The feasibility of EVAR under local anesthesia has been demonstrated at a number of centers [35]. Percutaneous delivery of stent-grafts also has the potential to further reduce the adverse physiological consequences of EVAR. However, these techniques may not be suitable for all patients and the impact of these techniques on patient outcome has not been confirmed. In one study the physiological effects of local anesthesia were only slightly reduced compared with those of general anesthesia [36]. In fact in another study, de Virgilio et al. were unable to demonstrate any difference in the number of postoperative cardiac or pulmonary complications in a total of 229 patients undergoing EVAR (general anesthesia, n=158; local anesthesia, n=71) [37].

Percutaneous delivery of stent-grafts has allowed some centers to perform EVAR on a day-case basis [38]. Others have found that <30% of patients who have their EVAR performed under local anesthesia are even suitable for discharge the following day [39]. The benefits of a percutaneous procedure appear small. Despite perceived benefits, many of the encouraging results of these percutaneous closure devices have been obtained following interventional cardiological procedures, which require smaller sheaths, possibly in less diseased vessels. In one study of patients undergoing EVAR, 15% of patients required conversion to a conventional groin incision and one patient died from retroperitoneal hemorrhage [40]. Similar results have been published by others [41]. Both studies underscored the importance of careful patient selection to reduce complications. Obesity, scarred groins, and calcification in the former study, and large sheath size in the latter, predicted procedural failure.

Endoleak

Experimental work using a bench-model has successfully demonstrated that all endoleaks, irrespective of diameter and length, are able to transmit systemic pressure [42]. In contrast, in vitro work suggests that thrombosed endoleaks may not behave in a similar fashion [43]. Short and wide thrombosed endoleaks are capable of transmitting greater pressures than long narrow ones. This may offer the reason why thrombosed type I endoleaks do not appear to be safe, whereas long, narrow thrombosed type II endoleaks are usually benign.

In an experimental model, Parodi and Ferreira found that the presence of an outflow channel in an aneurysm sac with endoleak might bring about pressure reduction [44]. The results suggested that increasing the outflow reduced both mean and systolic intrasac pressures.

Type I endoleak is usually the result of incorrect case selection with unsuitable aneurysm morphology, incorrect choice of stent-graft, or maldeployment of the stent-graft. Late cases of type I endoleak may be attributed to migration and/or neck dilation. It is generally accepted that type I endoleaks are associated with rupture, and consequently, treatment is mandatory no matter when they are discovered.

Type II endoleak does not appear to be a graft-related complication of EVAR. The etiology is retrograde perfusion of the aneurysm sac via the inferior mesenteric artery (IMA) or lumbar arteries (especially the fourth pair) and, rarely, other branches such as an accessory renal artery. It has been estimated that one-quarter of preoperatively patent IMAs will subsequently persist to perfuse the aneurysm [45]. Thrombosed IMAs do not appear to reperfuse the aneurysm sac (although this may not be the case following attempted treatment of the endoleak). Flow of blood in patent side branches is variable. In addition to simple inflow to the aneurysm sac, a to-and-fro movement of blood within type II endoleaks can often be detected on duplex ultrasound. The clinical significance of this motion is unknown. Patent side branches can also act as outflow vessels for other types of endoleak, including inflow from either the IMA or lumbar arteries or even other types of endoleak [46].

A number of centers have been unable to identify any preoperative factors which will reliably predict the development of type II endoleak [47]. In contrast, others have found that patients with a large, patent IMA or more than two lumbar arteries on preoperative spiral CTA are at higher risk for the development of persistent type II endoleaks [48].

The natural history of type II (side-branch) endoleak is generally benign but remains incompletely understood. Roughly two-thirds of these endoleaks will spontaneously thrombose in the perioperative period. There have been a minority of cases in which type II endoleaks appear to behave in an aggressive fashion. These include isolated aneurysm ruptures, but it is not yet clear why these type II endoleaks should behave so differently than the majority [49, 50].

Intraoperative studies during EVAR and open AAA repair have suggested that the pressure within side branches would likely be insufficient to result in persistent pressurization of the aneurysm sac [51]. These findings led the authors to hypothesize that high intrasac pressure was more likely to be transmitted directly through the graft rather than patent side branches. This report is in contrast with the intrasac pressure measurements performed by Baum and colleagues [52]. They recorded systemic or near-systemic pressures with pulsatile waveforms in all patients with endoleak, irrespective of type. Pulsatile pressures have also been found in type II endoleaks at laparotomy [46]. One suggestion for these discrepancies and the general benign nature of type II endoleaks is that type II endoleaks only cause a localized rise in sac pressure. Consequently the decision about which type II endoleaks require intervention has been controversial. The prevailing opinion among experts in the field is to take a noninterventional approach unless the aneurysm is getting bigger [53].

Type III endoleak is the result of graft failure. It is associated with subsequent aneurysm rupture. There are two main modes of failure. First is disintegration of the graft fabric and second is modular limb disconnection. These failures were more common with some of the less robust first-generation endovascular stent-grafts. Graft fabric disintegration was associated with some of the thin-walled graft fabrics [54]. Modular limb disconnection with earlier devices occurred because there was insufficient frictional force to prevent distraction at the junctional zone.

Type IV endoleaks are a consequence of thin-walled grafts which remain porous in the perioperative period. Many of these endoleaks will resolve spontaneously within 1 month of EVAR [55].

Endotension is a condition associated with endoluminal vascular grafts, defined by persistent or recurrent pressurization of an aneurysm sac after endovascular repair, without evidence of endoleak [56]. There are a number of reasons why pressure may be transmitted to the aneurysm sac in the absence of detectable blood flow (endoleak). The first and most widely accepted is the transmission of pressure by thrombus. This phenomenon has been recognized for some time in vascular surgery and is the reason why some aneurysms rupture despite being thrombosed [57]. Other plausible explanations include the presence of an intermittent or low-flow endoleak not amenable to detection by current imaging modalities or, alternatively, hygroma and seroma formation. Ultrafiltration of blood through PTFE grafts has been noted in open surgery [58, 59]. There have even been suggestions that an underlying infective process may be responsible in some patients with endotension.

In vitro analysis of pressure transmission through thrombus suggested that thrombosed endoleak channels do not behave in the same fashion as channels that are patent [43]. In thrombosed endoleaks in vitro, pressure reduction is directly proportional to the length and inversely proportional to the diameter of the channel. As previously suggested, this may account for the benign nature of thrombosed type II endoleaks, in contrast to the aggressive behavior of their type I counterparts.

Endotension is usually identified indirectly by increasing sac diameter, volume, or, possibly, aneurysm pulsatility. Unfortunately, however, endotension may be present in the absence of these signs and can, if left unrecognized or untreated, present with aneurysm rupture [60].

The measurement of pressure within the aneurysm sac may help in the management of patients following EVAR. At present it is only possible to measure pressure for short periods of time by invasive means. Catheters must be placed either by endovascular (with the potential for erroneous readings created by an endoleak channel around the catheter) or translumbar routes and connected to a pressure transducer. An indirect method of detecting a pressurized aneurysm is by echo-tracking ultrasound, which measures sac compliance by detecting wall motion [61]. In vivo measurements have confirmed a pressure reduction in successfully excluded aneurysm sacs [62]. Likewise, the pressure remains high in the presence of endoleak [63].

New technology currently undergoing in vitro and animal testing holds the promise of implantable, chronic pressure telemetry. Preliminary results from animal experiments of 3 months’ duration have been encouraging, with close agreement of measurements taken from wireless versus wired pressure sensors [64]. Although these devices would appear to be extremely useful, much work must be done to define their role. They must be shown to detect endoleak as reliably as current imaging techniques. In addition, their position in the aneurysm sac will be vital (as the pressure within a nonhomogenous sac is likely to be variable), as will their long-term accuracy and repeatability. Most critically, we have yet to define what level of pressure renders any particular aneurysm safe.

Endoleak Treatment

Endovascular treatment for endoleak is nearly always possible, and open conversion should be performed only as a last resort. Large balloon-expandable stents (e.g., the giant Palmaz stent) are useful in the treatment of type I endoleak. They improve the apposition of the stent-graft to the aortic neck, which is not attainable by simple angioplasty alone [65].

Type II endoleak is usually managed by coil embolization. Coils are deployed via the endovascular or translumbar routes. Special care must be taken to ensure that the channel is completely occluded, as pressure may be transmitted through an incompletely embolized vessel [66]. Preoperative occlusion (coil embolization) of side branches has been advocated in some publications in order to reduce the incidence of type II endoleak [67]. However, others have suggested that this policy may not be beneficial. In one study, type II endoleak developed in 20% of patients who had undergone preoperative embolization, compared with 23% of patients who had not undergone preoperative treatment [68].

Migration

The ”healing,” which occurs between a polyester graft and the aorta, is insufficient and consequently requires an alternative method of fixation [69]. Incorporated in endovascular stent-grafts are one or more design features which take the place of nonabsorbable sutures during open surgery. These include radial force of the stents in the aortic neck and hooks (±barbs) to engage either the aortic neck or the suprarenal aorta. Some stent-grafts have also incorporated high columnar strength into their design.

Many first-generation stent-grafts suffered from high rates of migration due to inadequate fixation. The Chuter device had vestigial hooks, which did not penetrate the full thickness of the aortic wall. In Nottingham, 57% migrated during a 7-year follow-up period [70]. Migrations tend not to appear until a number of months following device insertion. In one study the mean delay was 18 months, and in another, 24 months [70, 71]. It is not clear why migration appears to peak at this time but it may be intimately related with neck dilation. Some authors have suggested that as the neck appears to dilate at 1 mm per year, migration might not occur for 2 to 3 years because of graft oversizing [72]. It is still unclear which precedes the other, neck dilation or migration (suffice to say some grafts migrate without evidence of neck dilation).

A study by Resch et al. revealed that there was a great difference in the forces required to pull out stent-grafts from cadaveric aorta. Balloon-expandable stents and those with robust barbs placed in the suprarenal aorta were associated with the most secure fixation in that study [73]. A number of the other designs provided insufficient strength to prevent the forces placed on the grafts in vivo, which equate to ∼10 N. However, caution must be used when considering the results of this paper, as there are many more factors that contribute to endograft fixation and migration in vivo, including aneurysm morphology and, possibly, columnar strength which some endografts possess. Because there are many variables, it is difficult to identify which patients are likely to be at greater risk of developing migration. Hence, the precise risk factors for migration remain uncertain [74].

The effect of stent-graft design on the incidence of migration has been demonstrated in a number of studies. Stent-grafts with robust barbs placed in the suprarenal aorta, engaging the full thickness of the aortic wall, appear to be associated with less migration than their first-generation counterparts with or without vestigial hooks and barbs [75]. The evidence for the efficacy of columnar strength is not yet known because some of the stent-grafts with this feature have additional forms of fixation including either hooks and barbs or suprarenal stents. In addition, longer-term follow-up data, which may reveal the true migration prevalence, are not yet available for these grafts. New techniques are being developed to prevent and treat migration. These include endovascular and laparoscopic stapling [76].

Results of EVAR

In a recent systematic review of EVAR, the 30-day mortality rate was 1.6% in randomized controlled trials and 2% in nonrandomized trials and case series [77]. The results of the EVAR 1 trial suggested that patients considered to be ”normal” risk for aortic surgery had a 3% better aneurysm-related survival compared with patients undergoing open repair.

Some patients are not offered open repair because they are ”high risk.“ Using one definition of high risk, Jordan et al. experienced a perioperative mortality of 8.3% for open AAA repair and 2.3% after EVAR [78].

These encouraging results in high-risk patients undergoing EVAR mean that current scoring systems in vascular surgery may not be valid. In a retrospective study, Ananda and colleagues found that all variants of POSSUM overpredicted mortality and suggested that aneurysm morphology needed to be incorporated in future risk-stratification tools for vascular surgery [79].

The long-term outcome following EVAR is improving. In the EVAR 1 trial, the early reduction in aneurysm-related mortality benefit persists to at least 4 years (although all-cause mortality was not different).

Emergency EVAR

Critically ill patients with ruptured AAA are just the cohort of patients who may benefit the most from the reduced physiological insult associated with EVAR.

Early reports demonstrated the feasibility of the technique and identified and surmounted a number of the challenging areas [80, 81]. Subsequent reports have continued the early work, with improved outcome in both selected patients and ”all-comers” [82]. Additional refinements have also been made, including the use of local anesthesia [83]. One of the greatest obstacles to its success is the organization and delivery of an emergency EVAR service.

Future Prospects

The large multicenter randomized trials have demonstrated that EVAR is a viable alternative to open repair for the majority of patients presenting with AAA. It is associated with lower perioperative morbidity and mortality and outcomes which are at least as good at midterm follow-up. EVAR continues to be developed to become more applicable and durable. The prospects of new graft technology, fenestrated and branched grafts, in addition to endovascular stapling, hold promise to fulfill these requirements [84, 85]. A greater understanding of endoleak, particularly type II endoleak, is required. Intrasac pressure telemetry remains unproven but may offer rationalization of follow-up from the current intensive schedules.

EVAR has come a long way in a short time. Many challenges lie ahead but the future for EVAR appears bright.