Post-EVAR surveillance — Type I–V endoleak mechanisms, CTA / duplex imaging signatures, and management pathways after endovascular aneurysm repair
A Type I endoleak arises when the seal between the endograft and the native artery fails at either the proximal (Type Ia) or distal (Type Ib) attachment zone. Because the leak communicates directly with the aortic lumen upstream of any flow-limiting narrowing, the sac is repressurized to near-systemic arterial pressure — the same pressure that was driving the aneurysm toward rupture before treatment. Type I endoleaks carry the strongest and most consistent association with subsequent aneurysm rupture of any endoleak category and are treated as a surgical emergency whenever identified.
The endograft depends on circumferential apposition between its sealing stent and the native aortic wall (proximal neck) or iliac artery (distal landing zone). Seal failure occurs through several converging mechanisms:
• Inadequate neck anatomy: a short (<10–15mm), angulated (>60°), conical, or heavily calcified/thrombus-lined proximal neck prevents complete circumferential apposition • Undersizing: a graft diameter chosen with insufficient oversizing (target 10–20% over native artery diameter) fails to generate adequate radial seal force • Neck dilation over time: the aortic neck itself can dilate 1–3mm per year after EVAR as it is exposed to radial stress from the stent, progressively degrading a seal that was adequate at implant • Migration: caudal graft migration (>10mm) physically separates the sealing stent from its intended landing zone, most often from inadequate proximal fixation (barbs/hooks) in a hostile neck • Distal (Type Ib) failure: iliac limb landing in a dilating or short common iliac artery, often requiring extension into the external iliac with hypogastric coverage/embolization
Type Ia (proximal) leaks are more common and more dangerous than Type Ib (distal) because the proximal neck bears the full unattenuated force of aortic pulsatile flow. Any Type I leak found intraoperatively is corrected immediately with an aortic cuff, balloon molding, or a proximal extension before the patient leaves the angiography suite.
On contrast-enhanced CT angiography (the surveillance gold standard), a Type I endoleak appears as contrast opacification of the sac that is contiguous with the attachment zone in the arterial phase — the jet of contrast can typically be traced directly from the graft edge into the perigraft space, distinguishing it from the more posteriorly located, non-contiguous pooling of a Type II leak. Sac contrast in Type I leaks is typically dense and appears early (arterial phase), reflecting high flow and high pressure.
Duplex ultrasound, used as an adjunct or primary surveillance tool in some centers, shows high-velocity, often turbulent flow entering the sac directly at the stent-artery interface, with color Doppler demonstrating a jet oriented along the axis of the graft edge. Because a Type I leak is functionally equivalent to an untreated aneurysm neck, any sac growth in the presence of a demonstrated Type I leak is considered confirmatory, not merely suggestive.
Because Type I endoleaks maintain near-systemic sac pressure, they are treated urgently regardless of sac size at diagnosis:
1. Balloon angioplasty of the attachment zone — first-line for a marginal, low-volume leak with otherwise adequate anatomy 2. Proximal aortic cuff or distal limb extension — extends the seal zone into healthier, less angulated, or larger-diameter artery 3. Endoanchors — transmural anchors deployed across the stent into the aortic wall to augment fixation in a hostile or degenerating neck 4. Chimney/snorkel or fenestrated cuff — preserves branch vessel (renal/mesenteric) perfusion while extending the seal proximally above the original neck 5. Open conversion — reserved for anatomy unsuitable for any endovascular fix, or for infected/eroding grafts
A Type II endoleak occurs when a branch vessel that was patent before repair — most often a lumbar artery, less commonly the inferior mesenteric artery (IMA), and occasionally an accessory renal or median sacral artery — continues to fill the excluded sac in a retrograde direction from collateral circulation. Because this pathway crosses at least one high-resistance collateral network before reaching the sac, the pressure delivered is a fraction of systemic pressure. Type II leaks are found in 10–25% of patients after EVAR and are the single most common endoleak encountered on surveillance imaging.
The pre-aneurysmal aortic segment normally perfuses paired lumbar arteries and the IMA, which in turn anastomose with a rich network of collaterals — the internal iliac arteries via the middle sacral and iliolumbar circulation, and the superior mesenteric artery via the marginal artery of Drummond and arc of Riolan. After EVAR excludes the sac from direct antegrade aortic flow, these same vessels can reverse their flow direction, drawing blood backward through the collateral network into the now-isolated sac.
Because this retrograde pathway must traverse a chain of smaller-caliber collateral vessels rather than the aorta directly, the pressure delivered to the sac is substantially attenuated. Not every patent lumbar or IMA produces a clinically significant leak — flow must be sufficient and the network patent enough to sustain sac perfusion against sac wall tension and thrombus resistance.
On CTA, Type II leaks classically appear as a focus of contrast within the sac that is NOT contiguous with the graft attachment zones and is often located posteriorly (near the expected origin of lumbar arteries) or, less commonly, anteriorly near the IMA origin. Contrast typically appears later than a Type I leak — often more conspicuous on delayed-phase imaging as the slower collateral flow continues to fill the sac after arterial-phase contrast has washed out of faster-flowing structures.
A feeding and a draining vessel can sometimes both be identified — an "in-and-out" pattern in which one lumbar artery feeds the sac while another lumbar artery or the IMA drains it, sustaining flow through the nidus without a fixed single jet. Duplex ultrasound shows low-velocity, often to-and-fro or continuous low-amplitude flow within the sac rather than a discrete high-velocity jet.
Not all Type II leaks are equal: a leak with a stable or shrinking sac is a benign incidental finding requiring no action beyond continued surveillance, whereas a Type II leak associated with sac growth ≥5mm over 6–12 months is treated as a hemodynamically significant leak requiring intervention.
Roughly 40–60% of Type II leaks thrombose spontaneously within the first year, particularly when only a single feeding vessel is involved. The prevailing management philosophy is therefore watchful waiting with serial imaging (typically at 1, 6, and 12 months, then annually):
• Stable or shrinking sac: continue surveillance only — no intervention, regardless of whether the leak is still visible • Sac growth ≥5mm without another identifiable cause: transarterial or translumbar embolization of the feeding lumbar artery/IMA with coils, glue, or Onyx • Persistent leak beyond 6–12 months with any growth: lower threshold for embolization given diminishing likelihood of spontaneous resolution • Recurrent leak after embolization: repeat embolization, sac laparoscopic ligation of feeding vessels, or, rarely, open conversion in refractory cases
A Type III endoleak results from a structural failure of the graft itself: either a fabric defect (Type IIIb, a hole or tear in the graft material) or a separation between two overlapping modular components (Type IIIa, a junctional leak). Both create a direct, low-resistance communication between the high-pressure graft lumen and the sac — mechanically indistinguishable from a Type I leak in terms of the pressure delivered — and both are managed with the same reintervention urgency.
Type IIIa (junctional separation): modern endografts are modular — a main body plus one or two iliac limbs that are deployed with an intended overlap zone (typically ≥2 stent rings, ~1.5–2cm). Over years of cardiac pulsation, the components can migrate relative to one another, and if overlap is lost entirely, a gap opens between segments through which blood escapes directly into the sac. This is more common with certain graft designs, undersized limb selection, or excessive component angulation at implant.
Type IIIb (fabric tear): the polyester or ePTFE graft fabric can fail from metal strut erosion (a stent wire fatiguing and puncturing the adjacent fabric after millions of cardiac cycles), suture-line dehiscence, or fabric degradation, most often manifesting years after implant. Unlike Type IV porosity leaks, a Type IIIb tear is a true structural defect that does not self-seal.
On CTA, a Type IIIa leak appears as contrast tracking through a visible gap between two graft components, often best appreciated on multiplanar reformats aligned along the graft axis, sometimes accompanied by frank limb separation visible on the scout/topogram. A Type IIIb fabric tear may show a focal jet of contrast emanating from a point along the mid-graft body, unrelated to any attachment zone or component junction — the sac fills asymmetrically from a single point source.
Both subtypes can be subtle on axial CTA alone and are sometimes only definitively identified on catheter angiography during planned reintervention, or inferred indirectly from a rapidly enlarging sac with no branch vessel or attachment-zone source identified on cross-sectional imaging.
Because Type III leaks mechanically re-expose the sac to full aortic pressure, a rapidly growing sac with no clear Type I or Type II source on CTA should prompt dedicated graft-junction review and, if needed, catheter angiography to exclude an occult Type III leak before it is mislabeled as endotension.
Type III endoleaks are treated urgently, analogous to Type I:
1. Relining stent-graft: a new bridging component is deployed across the separated junction (Type IIIa) or across the fabric defect (Type IIIb), re-establishing a continuous sealed lumen 2. Extension cuffs: additional overlap length is added at the affected junction to prevent recurrent separation 3. Covered stent placement: for a focal fabric tear, a shorter covered stent can be placed directly across the defect 4. Explantation: reserved for extensive fabric degradation, graft infection, or anatomy unsuitable for relining
A Type IV endoleak is caused by blood weeping through the microscopic porosity of the graft fabric itself, rather than through any structural defect or attachment failure. It is essentially unique to the immediate perioperative period, when patients remain systemically anticoagulated and the fresh graft fabric has not yet been sealed by an adherent fibrin layer. It is a diagnosis of exclusion, made intraoperatively or on the earliest postoperative imaging, and by definition resolves without any specific treatment.
All woven or knitted graft fabrics have some intrinsic porosity — microscopic gaps between fibers that are functionally sealed once a thin fibrin layer deposits across the fabric surface, typically within the first few days after implantation. Before that fibrin seal forms — particularly while the patient is still systemically heparinized during the implant procedure itself — a diffuse "sweating" of blood can occur uniformly across the graft surface rather than from any single, localized defect.
Older-generation graft fabrics (particularly some early ePTFE and low-density polyester weaves) were more prone to clinically apparent Type IV leaks; modern low-porosity, tightly woven polyester fabrics have made this endoleak type substantially less common than it was in the early EVAR era of the 1990s.
On intraoperative angiography or immediate postoperative CTA, a Type IV leak appears as a faint, diffuse blush of contrast uniformly across the entire graft surface rather than as a discrete jet from any single point — there is no identifiable focal source, and the pattern does not localize to an attachment zone, a branch vessel, or a component junction. Because the leak is caused by transient fabric porosity rather than a fixed anatomic defect, serial imaging over hours to days shows progressive fading and eventual complete resolution of the blush without any intervention.
Type IV is a diagnosis of exclusion: it should only be called after Type I, II, and III sources have been actively ruled out. A "diffuse blush" that persists beyond the first week, or that is later found to localize to one region, is not a Type IV leak and warrants re-evaluation for an occult structural cause.
No graft-directed intervention is indicated for a true Type IV leak:
1. Confirm the diagnosis of exclusion — no attachment-zone, branch-vessel, or junctional source on careful review 2. Reverse or reduce anticoagulation if intraoperative bleeding is clinically significant (this is a systemic, not a graft, intervention) 3. Document the finding and repeat imaging within days to confirm resolution 4. Resume standard EVAR surveillance schedule once resolution is confirmed — no additional long-term monitoring is required specifically for this episode
Endotension describes continued or recurrent aneurysm sac expansion after EVAR in the absence of any leak jet identifiable on CTA, duplex ultrasound, or even catheter angiography. It implies that the sac continues to be exposed to pressure despite apparently complete radiographic exclusion — a physiological paradox that has generated substantial debate about mechanism, and that represents the most diagnostically challenging category in the endoleak classification system.
Several non-exclusive mechanisms have been proposed to explain endotension, none universally accepted:
• Pressure transmission through organized sac thrombus: laminated thrombus may transmit systemic pulse pressure to the sac wall even without any net flow of blood or contrast, analogous to how pressure can be transmitted through a solid medium • Ultrafiltration across graft fabric: plasma (without formed blood elements or contrast-visible flow) may seep through intact but permeable graft fabric over long time periods, insufficient in volume to appear as a discrete leak on any single imaging study • Sub-radiographic leaks: a true endoleak below the spatial or temporal resolution of standard CTA protocols — a Type I, II, or III leak simply too small or too slow to be visualized, particularly if imaging timing or slice thickness is suboptimal • Measurement/technical artifact: apparent sac growth that reflects thrombus reorganization or measurement inconsistency between studies rather than true pressurization
A systematic workup is applied whenever sac enlargement is identified on surveillance imaging, working from the most to least common causes before endotension is accepted as the diagnosis:
1. Review the surveillance CTA in both arterial and delayed phases — a delayed phase is essential, since Type II leaks fill more slowly and can be missed on arterial-phase-only protocols 2. Scrutinize the attachment zones on multiplanar reformats for a subtle Type Ia/Ib jet or early neck dilation/migration 3. Trace graft component overlap zones for evidence of Type IIIa separation, and inspect the graft body for a focal Type IIIb jet 4. If CTA is negative but the sac has grown ≥5mm, proceed to contrast-enhanced ultrasound (CEUS) or time-resolved MRA, both more sensitive than single-phase CTA for slow-flow leaks 5. If non-invasive imaging remains negative, proceed to catheter angiography with selective injections of the lumbar arteries, IMA, and the graft limbs individually to unmask an occult leak 6. Only after all of the above are exhausted and no source is identified is a diagnosis of true endotension (Type V) appropriate
Classification at a glance — Type I and Type III leaks re-expose the sac to full systemic pressure and are treated urgently; Type II is a low-pressure collateral phenomenon usually managed by observation unless the sac grows; Type IV is a transient perioperative fabric-porosity phenomenon requiring no treatment; Type V (endotension) is sac growth with no demonstrable leak on any imaging modality and requires escalating surveillance and, if growth continues, exploratory reintervention.
Because no leak source is identifiable, endotension cannot be treated by simply embolizing or relining a defect — management is necessarily more conservative and individualized:
1. Intensify surveillance interval (every 3–6 months rather than annually) once endotension is suspected 2. Consider direct sac pressure measurement via translumbar or transarterial puncture in centers with this capability — persistently elevated sac pressure supports ongoing risk despite the absence of a visible leak 3. For progressive growth despite negative exhaustive imaging, consider exploratory reintervention: relining the entire graft, converting to a different graft design, or sac fenestration/ligation of all potential collateral inflow at open exploration 4. Open conversion (graft explantation with surgical aneurysm repair) remains the definitive option for relentless growth or any concern for impending rupture, and is reserved for cases where endovascular options have been exhausted
| Product | Indication | Trial Design | Key Result |
|---|---|---|---|
| Type I — Attachment | Proximal (Ia) / distal (Ib) seal failure | Direct systemic-pressure jet at graft edge; visible on arterial-phase CTA | Urgent reintervention |
| Type II — Branch retrograde | Lumbar arteries, IMA collateral filling | Low-pressure posterior/non-contiguous blush, often delayed-phase | Observe unless sac grows |
| Type III — Graft defect | Fabric tear (IIIb) or modular separation (IIIa) | Systemic-pressure jet from junction or mid-graft point source | Urgent reintervention |
| Type IV — Porosity | Fabric weeping while anticoagulated | Diffuse uniform blush, perioperative only, fades over days | Observe — self-resolving |
| Type V — Endotension | Mechanism unresolved / occult | Sac growth with no leak on any imaging modality | Intensive surveillance |