Endoleaks, or persistent flow into the aneurysm sac, are the most common complication of EVAR, occurring in up to a third of cases. They occur at the time of graft implantation, or arise during followup, with varying implications. The Society for Vascular Surgery classifies endoleaks into five categories ( Fig. 17.1 ) :
Type I. Attachment site endoleak
IA. Proximal attachment site endoleak
IB. Distal attachment site endoleak
Type II (the most common). Back-filling of the aneurysm sac through branch vessels of the aorta
IIA. Single causative vessel (often referred to as a “to-and-fro” endoleak)
IIB. Multiple involved vessels
Type III. Graft defect or component misalignment
IIIA. Leaks from junctions of the endograft components
IIIB. Leak from fabric tear/disruption
Type IV. Leakage through the graft itself caused by the porosity of some endografts
Type V. Also called endotension, Type V endoleaks are poorly understood, but thought to result from aortic pressure transmitted through the graft or thrombus to the aneurysm sac.
Most endoleaks seen on completion angiogram mandate attempts at correction prior to leaving the operating room. Notable exceptions are most Type II endoleaks, which often resolve as the vessels thrombose over time, and Type IV endoleaks, which resolve with cessation of heparinization. However, surgeons must pay particular attention to completion angiograms, as Type IV endoleaks appear very similar to Type I. Type II endoleaks discovered at follow-up can often be observed as long as they are not associated with significant sac expansion (defined as ≥5 mm). Late Type I or Type III endoleaks, however, almost universally require intervention. While the aneurysm sac remains excluded from aortic pressure, it weakens and atrophies. This atrophic sac is at increased risk of rupture from a high-pressure leak such as a Type I or Type III.
Type Ia Endoleaks
Type Ia endoleaks result from inadequate seal at the proximal attachment site (the neck). Type Ia endoleaks that occur at the time of the operation usually result from unsuitable anatomy. Consequently, careful preoperative planning may help avoid many of these complications ( Box 17.1 ). Surgeons should ensure an adequate neck length of at least 15 mm, preferably without significant calcification (>50% of the diameter of the vessel), neck angulation (>60 degrees), or conical shape (increase in size from proximal to distal). Short or angulated necks, or segments of aorta without parallel walls, benefit from fenestrated, custom-made, surgeon-modified endografts, or open repair. Although the instructions for use (IFU) of many devices include neck lengths as short as 10 mm, the advent of fenestrated devices obviated the need to push the tolerances of prior devices. As such, we rarely deploy standard endografts in these marginal cases, even within the IFU. In addition, although the IFU for many devices include neck diameters of up to 32 mm, such wide necks represent early aneurysmal degeneration ( Box 17.2 ). Even if no Type Ia occurs at the time of implant, one often develops over time as the neck continues to expand, creating a potentially life-threatening leak. Fortunately, surgeons can rely on a range of alternative strategies to deal with unsuitable necks, from fenestrated, snorkel, custom-made, or surgeon-modified devices to traditional open repair. Short or wide necks are situations in which other options such as open surgery or fenestrated, scalloped or snorkel devices should be considered. In our practice, we rarely place grafts wider than 32 mm below the renal arteries.
The most important factor in the prevention of endoleaks is careful preoperative planning.
Although the instructions for use for many devices allow up to a 32 mm neck, we do not advise sealing into a neck this size because it probably represents early aneurysmal degeneration and carries significant risk of late Type IA endoleak.
If a Type Ia endoleak does occur at the time of the initial procedure, many will resolve with ballooning of the seal zone using a compliant, molding balloon such as a Coda (Cook Medical, Bloomington, Indiana) or Reliant (Medtronic, Santa Rosa, California). If ballooning is unsuccessful, the next step is to place an aortic cuff if there is an adequate landing zone (>5 mm). Some advocate placing a giant Palmaz stent (Cordis Corp., Bridgewater, New Jersey) if a cuff fails to resolve the leak, but many centers (including ours) now avoid placing Palmaz stents except in special circumstances such as an incompletely deployed stent. A Palmaz rarely succeeds long-term. In addition, should the Palmaz fail to resolve the leak, its tight mesh complicates further endovascular maneuvers such as deploying a fenestrated cuff. Placement of an aortic cuff or a giant Palmaz stent should be followed by additional ballooning of the landing zone. In our practice, if ballooning and/or an aortic cuff are unsuccessful, we reinforce the landing zone using the Heli-FX EndoAnchor (Aptus Endosystems, Sunnyvale, California). EndoAnchors (also known as screws) should be avoided in heavily calcified necks, because the screws will not penetrate the calcium and will actually push the graft away from the aortic wall.
At this point, should a leak persist, the surgeon must assess the quality of the leak and consider whether the leak requires immediate action. After the above maneuvers, the only option available during the index procedure (although almost never used) is open conversion, as any further endovascular therapies require planning time. Fortunately, most leaks can be observed for a limited time while further action is planned and we avoid open conversion during the index case except for the rarest of circumstances. One particularly challenging scenario is an undersized graft. If the initial graft is undersized, then unfortunately screws (and likely most endovascular maneuvers) rarely succeed. To deal with a proximal endoleak with an undersized graft, some surgeons place a new cuff and parallel grafts in chimney fashion, although we prefer custom-made or surgeon-modified fenestration. Some may require conversion to open. Large or even moderate leaks need further intervention, but many small gutter leaks will resolve after reversal of heparinization. Some surgeons will follow these smaller leaks with serial computed tomography (CT) scans at frequent intervals (1 month, 6 months, and 1 year if the endoleak persists). Moderate to large leaks, or those that fail to resolve in serial followup, require further intervention.
In recent years, the armamentarium of the modern vascular surgeon increased dramatically. Subsequent efforts to correct a proximal endoleak should take into account the surgeon’s familiarity with these novel techniques, as well as the anatomy of the surgical neck. If there is an adequate seal zone above the newly-placed endograft or cuff, the surgeon can place another cuff or place a second EVAR device. If there is no seal zone below the renal arteries, the Cook Z-Fen (Cook Medical, Bloomington, Indiana) can extend the seal zone above the renal arteries. However, commercially available Z-Fen devices taper to 24 mm and extend at least 5 cm distally toward the bifurcation (and longer lengths for grafts greater than 24 mm to allow for a flaring cuff), which limit its ability to deploy within most pre-existing EVAR. An alternative strategy is placement of a Z-Fen cuff and then an aorto-uni-iliac (AUI) device. This avoids the space issue, but requires a subsequent femorofemoral bypass. Consequently, the most commonly applied therapy in our hands is the physician-modified endograft (PMEG). A custom-modified body of a Cook Alpha device (Cook Medical, Bloomington, Indiana) avoids the aforementioned problems with the Z-Fen, because it is shorter and does not taper down. However, these devices are still experimental and should be only be used by surgeons experienced in their use and with approval from their institution’s Institutional Review Board. Currently, PMEGs are performed under an Investigational Device Exemption and/or after a thorough discussion with the patient concerning the off-label modification of an endograft. In limited series and anecdotal reports, perioperative mortality following PMEG is lower than open conversion, but long-term durability remains unknown.
If these rescue maneuvers fail to resolve the leak, the surgeon can convert to open repair ( Fig. 17.2 ). Open conversion should not be undertaken lightly, as it carries substantial morbidity and mortality (10% rate of death within 30 days). The two primary approaches are to explant the device and perform a traditional open repair or to band the proximal neck with felt pledgets. Often the iliac limbs are left in place at the time of open conversion.