Keywords Type II endoleak - Digital subtraction angiography - Transarterial embolization -
Procedural success - Outcome analysis
Introduction
With an incidence of 10–25%, type II endoleaks are more common than any other type
of endoleak and represent the most frequent complication after endovascular aortic
repair (EVAR) [1 ]. Type II endoleaks facilitate retrograde perfusion of an aneurysm sac through lateral
feeder vessels, most often from lumbar segmental arteries or the inferior mesenteric
artery but also from intercostal arteries in the case of thoracic EVAR [2 ]. In order to detect endoleaks promptly, patients are advised to undergo lifelong,
periodic imaging after EVAR [3 ]. CT angiography and contrast-enhanced ultrasound are the examination methods of
choice, whereas diagnostic digital subtraction angiography (DSA) has become less important
for follow-up imaging due to its invasiveness [4 ]. MRI may also play a role in long-term follow-up [5 ].
If a type II endoleak is documented and the aneurysm’s size increases substantially,
endovascular feeder embolization represents a potential method to cut off the blood
flow into the aneurysm sac. While other routes exist, the most common techniques are
retrograde catheterization via the common femoral artery and percutaneous translumbar
or transabdominal access [6 ]. Regardless of the access route, coils or liquid embolic agents or a combination
of both can be used to disrupt the perfusion of feeder vessels [7 ]. While the former come in the form of microcoils and macrocoils, the group of liquid
embolic agents includes cyanoacrylate derivates or ethylene vinyl alcohol copolymer-based
adhesives [8 ]. Due to their material properties, liquid embolic agents fill a target volume more
completely than coils, which may support embolization of the endoleak nidus, i.e.,
the intraaortic cavity or connecting zone of the afferent and efferent supply arteries
[9 ]. However, a disadvantage is the higher risk of vascular injury if the catheter becomes
stuck in the target area. In addition, non-target embolization occurs more frequently
with liquid embolization than with the use of coils [10 ].
Due to the heterogeneous nature of the data and the fact that embolization of type
II endoleaks is a relatively rare and highly specialized procedure, the 2024 clinical
practice guidelines of the European Society for Vascular Surgery (ESVS) do not contain
a clear recommendation regarding the optimal access route and embolization material
for interventional closure of type II endoleaks [4 ]. Therefore, this study was designed to address the need for more data as well as
to investigate the safety and efficacy of transarterial embolization of type II endoleaks
via the common femoral artery following EVAR.
Materials and Methods
Study design
In this monocentric study, the data of patients who underwent endovascular treatment
of a type II endoleak between January 2008 and September 2023 at a tertiary-care university
hospital were retrieved from the institution’s electronic database and analyzed retrospectively.
By definition, previous EVAR of an abdominal or thoracic aortic aneurysm was necessary
for study inclusion. Patients with recurrent endoleaks and repeated endovascular treatment
after EVAR were also included. The study design received the approval of the local
ethics committee (2023072601), who waived the need to obtain separate written informed
consent due to the retrospective nature of the investigation.
Patient analysis
Sociodemographic parameters and clinical risk profiles were registered for each patient.
The analyzed sample comprised 36 individuals (75.9 ± 6.6 years), including 33 men
(91.7%). Nine patients (22.2%) received at least two endovascular interventions, bringing
the total number of procedures to 50. All but three patients were regularly taking
anticoagulant medication (91.7%). [Table 1 ] provides an overview of the risk profile of the study sample including common underlying
diseases. Notably, 341 abdominal and 130 thoracic EVAR procedures were performed in
our hospital during the study enrollment period.
Table 1 Patient sample.
Patient age
75.9 ± 6.6 years (mean ± standard deviation)
Patient sex
33 men (91.7%) and 3 women (8.3%)
Risk factor
Frequency
Risk factor
Frequency
Note: Relative frequencies were calculated for the entire patient sample (i.e., 36
individuals). COPD: chronic obstructive pulmonary disease; TIA: transitory ischemic
attack.
Arterial hypertension
29 (80.6%)
Atrial fibrillation
7 (19.4%)
Smoking (current or past)
17 (47.2%)
Arteriosclerosis
6 (16.7%)
Dyslipidemia
14 (38.9%)
Heart failure
4 (11.1%)
Coronary heart disease
12 (33.3%)
COPD
4 (11.1%)
Obesity
10 (27.8%)
Post-stroke
3 (8.3%)
Diabetes mellitus
9 (25.0%)
Post-TIA
1 (2.8%)
Aneurysm analysis
Both the aneurysm morphology and the position of the inserted stent graft were assessed.
The maximum diameter of the aneurysm (centerline measurements in CT angiography or
determined via contrast-enhanced ultrasound) was recorded for each patient at different
time points: The first size measurement was performed before EVAR, and further control
points were defined before the primary and each subsequent endovascular procedure.
Based on these measurements, the time intervals between EVAR and type II endoleak
detection, between EVAR and the angiographic intervention, and between the first endoleak
detection and angiography were evaluated.
Endoleak analysis
Due to the study-specific inclusion criteria, all recorded and treated endoleaks were
type II leaks according to the classification of White and Stavropoulos [7 ]. The number of feeder vessels perfusing the aneurysm sac was recorded before each
intervention. If a single feeder vessel could be identified, the endoleak was determined
as type IIa (i.e., simple), whereas endoleaks with two or more feeders were classified
as type IIb (i.e., complex). For each feeder, the originating vessel was determined.
Procedure analysis
Transarterial endoleak embolization was performed in dedicated angiography suites.
Either the Axiom Artis Zee (Siemens Healthineers, Forchheim, Germany) or Azurion Clarity
IQ (Philips Medical Systems, Best, The Netherlands) angiography system was used to
perform DSA during the study period. All procedures were performed by a board-certified
interventional radiologist with 27 years of clinical experience. During interventions,
both intermittent fluoroscopy and DSA (with both selective/super-selective manual
injections as well as non-selective, high-pressure injections) were employed to identify
and catheterize feeder vessels before embolization. Once the desired catheter position
was verified by imaging, either coils or liquid embolic agents, or a combination thereof
was utilized to occlude the feeder vessel. (Push) coil embolization was the preferred
method if super-selective catheterization of an endoleak’s nidus was achieved, whereas
liquid embolization was performed in cases with distal rarefication of feeding vessels
preventing microcatheter placement within the nidus. A combination of coils and liquids
was used in certain cases to avoid non-target embolization. Frequently used coil types
included the manufacturer-specific pushable models Nester (Cook Medical, Bjaeverskov,
Denmark), VortX (Boston Scientific, Marlborough, MA, USA), and Tornado (Cook Medical),
or the detachable Concerto microcoils (Medtronic, Dublin, Ireland) in various lengths
and sizes; the number of coils was documented for coil-based interventions. The two
most commonly used liquid embolic agents were n-butyl-2-cyanoacrylate (Histoacryl;
B.Braun, Melsungen, Germany) and synthetic surgical glue (Glubran 2; GEM, Viareggio,
Italy).
Outcome analysis
The success of interventions was measured according to two criteria. Technically successful
embolization required the complete lack of blood flow in the previously perfused aneurysm
sac, i.e., no accumulation of contrast medium in the final DSA series. In addition,
all procedural complications were recorded and the complication rate in the patient
population was calculated. A complication was defined as any unforeseen event with
potentially negative consequences for a patient's state of health that originated
from the intervention and would not have occurred otherwise. Complications were classified
according to the reporting standards of the Society of Interventional Radiology as
major or minor [11 ].
In order to evaluate the long-term clinical success, the image-based control examinations
following the respective angiography examination were analyzed. The parameters assessed
in follow-up imaging included the perfusion of the aneurysm sac via a persistent or
new endoleak, the maximum extent of the aneurysm sac, and the change in the sac’s
maximum diameter compared to pre-angiographic imaging. In accordance with Iwakoshi
et al. and a meta-analysis by Akmal et al., embolization was considered successful
if the maximum diameter of the aneurysm sac either decreased, remained constant, or
increased ≤ 0.5 cm over all follow-up examinations [12 ]
[13 ]. Leaks that continued to cause perfusion of the aneurysm sac in the first follow-up
examination after endovascular therapy were defined as persistent endoleaks. In contrast,
a feeder which could not be delineated on the DSA scan used to guide the embolization
procedure was considered a new endoleak. Based on this consideration, the freedom
rates for recurrence and persistent endoleaks were calculated. In addition, the re-intervention
rate was determined as the ratio of patients who underwent multiple transarterial
embolization treatments for type II endoleaks to the total size of the patient sample.
Statistical analysis
Data analyses were performed using specific software (SPSS Statistics version 28.0;
IBM, Armonk, New York, USA). In a first step, metric variables were tested for the
presence of normal distribution using Kolmogorov-Smirnov tests. For metric items without
normal distribution and ordinal or nominal items, the absolute and relative frequencies
are given with median values and interquartile ranges. For normally distributed metric
variables, mean values and standard deviations are provided instead. In order to examine
the connection between two items with a metric scale level, correlation analyses were
carried out using the Pearson coefficient (r). If one variable was nominally scaled,
the Eta coefficient was determined instead. In the case of two nominally or dichotomously
scaled items, the Phi coefficient was used. Null hypotheses were rejected if the corresponding
alpha error was below a value of 0.05.
Results
Aneurysm morphology
Based on preoperative imaging, 17 aneurysms were classified as fusiform, 14 as sacciform,
and 5 as a fusisacciform mixed type. Preexisting occlusion of the inferior mesenteric
artery was ascertained in 19.4% of patients. Meanwhile the median number of perfused
lumbar arteries was 4 (interquartile range 2–6). Most patients displayed minor intra-aneurysmal
thrombosis (91.7%). During EVAR, the vast majority of patients had received an aorto-bi-iliac
prosthesis (80.6%). Before stent graft insertion, the average maximum aneurysm diameter
was 5.8 cm (5.3–6.3 cm). Before endoleak embolization, the median aneurysm diameter
was 6.7 cm (6.1–7.9 cm), which corresponds to an average size increase of 0.8 cm (0.6–1.2
cm) between insertion of the stent graft and the interventional endoleak therapy.
The average time between implantation of the endograft and first detection of a type
II endoleak was 8 days (0–196 days). In eleven patients (30.6%), the first documentation
of a type II endoleak already occurred intraoperatively during EVAR. The average time
between EVAR and the first transarterial embolization was 820 days (364–1436 days),
while the interval between endoleak detection and first endovascular therapy was 511
days (87–1246 days).
Endoleak characterization
Of the 50 type II endoleaks treated, a single feeder vessel (type IIa) was identified
in 32 cases. More than one feeder vessel (type IIb) was detected in the remaining
18 cases. The majority of endoleaks analyzed in this study originated in one of the
iliac vessels. Less common were endoleaks stemming from the mesenteric arteries and
the lumbar or gluteal arteries. In the two patients with thoracic stent grafts, type
II endoleaks originated from the intercostal arteries. [Table 2 ] summarizes the morphological characteristics of aneurysms, stent grafts, and endoleaks
in the study sample, while Supplemental Table S1 provides a detailed overview of the endoleak origins in all 50 procedures.
Table 2 Aneurysm and endoleak characterization.
Parameters
Frequency
Note : Relative frequencies were calculated either for the entire patient sample (i.e.,
36 individuals) or the entirety of interventions (i.e., 50 procedures).
Aneurysm morphology
Reference: 36 patients
Fusiform
17 (47.2%)
Sacciform
14 (38.9%)
Fusisacciform
5 (13.9%)
Type of endograft
Reference: 36 patients
Aorto-bi-iliac graft
29 (80.6%)
Aorto-mono-iliac graft
4 (11.1%)
Thoracic tube graft
2 (5.6%)
Abdominal tube graft
1 (2.8%)
Number of feeder vessels
Reference: 50 interventions
One feeder (type IIa)
32 (64.0%)
Two feeders (type IIb)
11 (22.0%)
Three feeders (type IIb)
6 (12.0%)
Four feeders (type IIb)
1 (2.0%)
Technical success
The exclusive use of coils was recorded for 29 embolizations (58.0%), with the number
of coils varying from a minimum of one to a maximum of 39 coils during one procedure
([Table 3 ]). A successful coil embolization is shown in [Fig. 1 ]. In seven interventions (14.0%), a liquid embolic agent was used in addition to
coil embolization. In seven other procedures (14.0%), endoleaks were treated exclusively
with liquid embolic agents ([Fig. 2 ]).
Table 3 Number of coils used per embolization procedure.
Number of coils
Frequency
Note: Relative frequencies were calculated for coil-based interventions (i.e., 36
of 50).
< 5
9 (25.0%)
5–10
10 (27.8%)
10–20
11 (30.6%)
> 20
6 (16.7%)
Fig. 1 Successful transarterial embolization for the treatment of a type II endoleak originating
from the inferior mesenteric artery after EVAR (A ). Two 5 mm coils were used in this 83-year-old patient (B ). Aneurysm perfusion was permanently absent after the interventional procedure (C ).
Fig. 2 In a 64-year-old man, a lumbar artery branch was seen feeding a type II endoleak after
EVAR had been performed for an abdominal aortic aneurysm with insertion of a bi-iliac
stentgraft. Due to a continuously enlarging right iliac artery aneurysm associated
with the endoleak, endovascular therapy was requested (A ). Technically successful super-selective embolization was performed using a combination
of cyanoacrylate glue and lipiodol because of the feeder vessel’s corkscrew-like morphology
(B ). No recurring endoleak was observed during follow-up.
Successful embolization of target vessels with complete elimination of aneurysm perfusion
was achieved in 42 procedures. Therefore, the technical success rate was 84%. Reasons
for unsuccessful embolization were pronounced kinking of the iliac arteries (n = 2),
a corkscrew-like tortuous course of the lumbar arteries (n = 1), or the presence of
a particularly narrow caliber vascular plexus (n = 5). Follow-up imaging via CTA was
available after six unsuccessful embolization attempts. In all of these cases, the
initially targeted type II endoleak was classified as persistent with four aneurysm
sacs showing further enlargement. However, only one patient was referred to open surgery
(i.e., ligation). The other patients received periodic follow-up scans. Nine patients
required more than one endovascular embolization procedure during the observation
period, resulting in a re-intervention rate of 25%. Five patients (13.9%) underwent
a single re-intervention, three patients (8.3%) required two further interventions,
and one patient (2.8%) underwent three re-interventions for recurrent type II endoleaks
([Fig. 3 ]).
Fig. 3 Multiple transarterial embolization procedures for recurrent type II endoleaks in
an 83-year-old man finally resulting in permanent cessation of aneurysm growth. A/B Before and after coiling of a feeder from the main trunk of the left internal iliac
artery. C/D Before and after coiling of a total of three feeder vessels from the right internal
iliac artery. E/F Before and after coiling of a feeder from the left superior gluteal artery. G/H Before and after coiling of a feeder vessel from the inferior mesenteric artery.
Clinical success
Since no follow-up was available for nine of the 50 endovascular interventions, the
following key figures could only be calculated for the 41 interventions with subsequent
CT angiography and/or ultrasound controls. The median follow-up interval was 221 days
(101–615 days). The average time interval between endoleak embolization and the first
follow-up examination was 88 days (31–117 days) ([Fig. 4 ]). No persistent endoleak was detectable in 31 of the earliest follow-up examinations.
Therefore, the clinical success rate in this respect was 75.6%. Over the entire control
period, no endoleak recurrence was recorded after 19 procedures, hence the recurrence-free
rate was 46.3%. Size progression of the aneurysm sac during follow-up was detected
after 21 interventions (51.2%). However, a progression of > 0.5 cm was only ascertained
in eleven instances (26.8%) ([Table 4 ]). No significant correlation could be determined between any of the recorded risk
factors and the technical or clinical success parameters (all p > 0.05). In contrast,
the maximum aneurysm sac size before endovascular therapy correlated strongly with
the largest diameter during follow-up (r = 0.87; 95% confidence interval 0.76–0.93).
Fig. 4 Median time intervals with interquartile ranges in parentheses.
Table 4 Maximum aneurysm diameter during follow-up.
Note: Relative frequencies were calculated for the number of interventions with available
imaging follow-up with either contrast-enhanced ultrasound or CT angiography (i.e.,
41 of 50).
Size difference
Frequency
Regression
All
12 (29.3%)
–0.1 to -0.5 cm
8 (19.5%)
–0.6 to -1.0 cm
4 (9.8%)
Consistency
All
8 (19.5%)
Progression
All
21 (51.2%)
+0.2 to +0.5 cm
10 (24.4%)
+0.6 to +2.0 cm
8 (19.5%)
more than +2.0 cm
3 (7.3%)
Rate of complications
No major complications were recorded in the investigated patient sample. After one
endovascular embolization, a self-limiting pseudoaneurysm of the right common femoral
artery was registered, accounting for a minor complication. Accordingly, the overall
procedural complication rate was 2%.
Discussion
In this large single-center study, a high technical success rate (84%) and low complication
rate (2%) were established for transfemoral embolization of type II endoleaks after
previous endovascular aortic repair. Clinical success, defined by the absence of a
persistent endoleak in the earliest imaging follow-up, was ascertained in 75.6% of
cases, demonstrating the general efficacy of endovascular therapy. While the endoleak
recurrence rate was comparatively high at 53.7%, transarterial embolization was able
to prevent a relevant growth of the aneurysm sac in 73.2% of the interventions. As
a major finding, individual risk profiles were not substantially related to the angiographic
or clinical success.
Regarding the effect of patient risk profiles, a study by Vandelbulcke et al. postulated
that active smoking and hyperlipidemia would have a positive influence on the “radiological
success”, which they defined as follows: No persistent endoleak and no increase in
maximal axial diameter in the first postinterventional control [14 ]. Interestingly, the results of Sarac et al., who found that smokers have a higher
risk of postinterventional expansion of the aneurysm sac, are in exact contrast to
this. In this study, hyperlipidemia was the only variable associated with a higher
risk of a second interventional procedure [15 ]. The contrasting findings of Vandelbulcke et al. and Sarac et al. underline that
a scientific consensus on the effects of the individual risk profile does not exist
so far. Therefore, the absence of a substantial correlation between risk profile and
outcome within our data appears plausible. Another central aim of many studies is
to determine which endoleak- or aneurysm-related characteristics influence the success
of treatment. In our study, we observed that the maximum pre-interventional diameter
is a predictive variable for the maximum post-interventional diameter of the aneurysm
sac. This is consistent with the results of Iwakoshi et al., who postulated that the
maximum diameter before endovascular treatment represents a risk factor for further
aneurysm growth after the intervention [12 ].
The technical success rates in the present investigation are comparable with the results
of other studies. While Vance et al. recorded an overall success rate of 98%, the
authors compared the success of different access routes [16 ]. While direct translumbar or transabdominal access led to success in 100% of cases,
perigraft access was successful in 86%. It should be noted that the transarterial
route, which was used exclusively in the present patient sample, resulted in a success
rate of merely 76% for Vance et al. [16 ]. A meta-analysis by Ultee et al. found different success rates depending on the
chosen access route: 98.7% for translumbar access, 93.3% for transcaval procedures,
and 98.1% for open ligation. Interestingly, the technical success rate for transarterial
access was again significantly lower at 84% [8 ]. Since transarterial embolization is obviously a demanding procedure, our high success
rates may be attributed to the profound expertise of the specialized interventional
radiologist who performed all 50 procedures in the present study.
Another important indicator for evaluating an interventional procedure is the number
of complications. In this study, the complication rate was 2%, as only one self-limiting
inguinal pseudoaneurysm occurred, classified as a minor complication. In contrast,
Ultee et al. described minor and major complications in 3.8% of 1073 interventions
in their meta-analysis. These included, among others, ischemic colitis, coil misplacement,
and cardiac decompensation [8 ]. It can be assumed that the low complication rate in the present study is partly
due to the high number of transarterial embolization procedures in a single center.
For comparison, Iwakoshi et al. were able to include a total of 315 patients from
19 Japanese centers, which corresponds to approximately 17 patients per center [12 ]. More than twice as many patients were included in the current study, supporting
the hypothesis that highly specialized procedures, such as transarterial embolization
of type II endoleaks, should be conducted in high-performance centers whenever possible.
The re-intervention rate in the current study was 25%, which is in line with the results
of Iwakoshi et al. (24.7%) and Sarac et al. (24%) [12 ]
[15 ]. In the study sample of Arenas Azofra et al., the frequency of additional procedures
was even higher at 35.7% [17 ]. It should be noted that the overall re-intervention rate described by Ultee et
al. was by far the lowest with just 14.7% [8 ]. However, this may be attributed to the meta-analysis also including several studies
in which an open surgical procedure was used, since some of these studies recorded
a re-intervention rate of 0% [18 ]
[19 ]
[20 ]. In the earliest follow-up, no persistent endoleak was determined after 75.6% of
interventions in our study, while the goal of postinterventional endoleak elimination
was only achieved in 41.7% of cases by Vandenbulcke et al. [14 ]. A potential explanation lies in the fact that endoleaks were primarily closed with
coils in the current study (83.7%). In contrast, Vandenbulcke et al. used liquid embolization
exclusively in 48.3% of interventions, additional coils in 43.3% and coils exclusively
in only 5% of procedures [14 ]. Whereas 73.2% of our patients did not experience aneurysm growth > 0.5 cm during
follow-up, other groups reported weaker results in this respect. In the investigations
of Iwakoshi et al., Sarac et al., and Arenas Azofra et al., only 55.4%, 44%, and 62.5%
of interventions were deemed clinically successful under the same criteria (no diameter
increase > 0.5 cm) [12 ]
[15 ]
[17 ]. In the light of these findings, the use of coils may be considered most effective
if the nidus of an endoleak can be probed directly. In this scenario, higher procedure
safety constitutes a particular advantage of coil usage due to a lower rate of non-target
embolization. In other situations, however, we believe that liquid embolic agents
should be selected. Notably, the 2024 ESVS guidelines do not recommend endovascular
embolization for aneurysm sac growth of less than 1 cm [4 ], hence the clinical impact of enlargement between 0.5 and 1 cm may be lower than
expected in the past.
Several methodological limitations must be taken into account when interpreting the
results of the present investigation. First, it should be emphasized that this study
is based on a monocentric data analysis with a retrospective design over a period
of 15 years. Compared to multicenter studies, there was no external validation of
results. Second, the follow-up varied individually without a standardized scheme,
which makes a structured comparison of patient aftercare difficult. Furthermore, a
median follow-up interval of 221 days is rather short when considering the frequently
slow progress of type II endoleaks. Third, no CT-guided embolization was performed
in this patient series. The main reason for this lies in our hospital’s policy of
subjecting patients to a surgical approach, if type II endoleaks have a blurred or
unsharp character on pre-interventional CTA. Therefore, this alternative option for
type II endoleak treatment was not explored in the present study. Fourth, the techniques
and materials used for transarterial access of the endoleak nidus and the embolic
agents employed for its occlusion were at the discretion of the interventional radiologist
who performed all 50 procedures. Fifth, since neither CT angiography nor contrast-enhanced
ultrasound are performed directly after an embolization procedure, the DSA from the
embolization itself was used as the baseline to differentiate persistent from newly
arising endoleaks. Sixth, no prophylactic mesenteric or lumbar artery embolization
took place prior to EVAR in the investigated patient sample. Therefore, the effect
of this procedure could not be analyzed. Finally, there is a lack of scientific consensus
as to whether coils or liquid embolic agents provide better long-term results for
the treatment of type II endoleaks. The best way to evaluate this would be in a prospective,
randomized controlled study design. However, as the treatment of type II endoleaks
is a rare intervention for most interventional radiologists, a multicenter study with
standardized conditions would be desirable.
Conclusion
Despite being a challenging interventional procedure, transarterial embolization of
type II endoleaks after EVAR constitutes a safe form of treatment with a low complication
rate. While endoleak recurrences limit the effectiveness in the long-term prevention
of aneurysm enlargement, most interventions succeeded in permanently eliminating the
targeted feeder vessels.
