Key words embolization - renal arteries - angiography - interventional procedures
Introduction
Partial nephrectomy is a well-established operative procedure for the treatment of
renal tumors [1 ]
[2 ]. This nephron-sparing surgery shows similar oncologic outcomes when compared to
total nephrectomy [3 ]
[4 ]
[5 ]. In addition, it is associated with a lower risk of post-operative chronic renal
failure [3 ]
[6 ]
[7 ]. However, partial nephrectomy increases the risk of iatrogenic vascular lesions,
such as pseudoaneurysms and arteriovenous fistulas, due to the highly vascularized
renal parenchyma [8 ]
[9 ]. They are rare but serious complications which can lead to severe hematuria, blood
loss, and hemorrhagic shock [8 ]
[10 ].
It has been reported that these arterial lesions can be treated successfully with
minimally invasive selective transcatheter embolization [11 ]
[12 ]
[13 ]
[14 ]. However, the published studies on this topic are relatively small, with the largest
study to date, to our knowledge, reporting about a cohort of 28 patients [11 ]. In addition, the effects of embolization on quantitative renal parenchymal volume
and procedure-related radiation dose have not been previously reported.
Therefore, this study was performed to further define the role of embolization in
the management of renal vascular lesions after partial nephrectomy. The purpose of
this study was to evaluate all consecutive patients who received transcatheter embolization
for the treatment of renal vascular complications after partial nephrectomy at our
institution. Technical and clinical successes, radiation doses, and the influence
of the embolization on renal function and volume are reported.
Materials and Methods
Patients and technical and functional outcomes
All consecutive patients who underwent angioembolization of iatrogenic vascular lesions
due to partial nephrectomy were identified through a search of our database, which
includes all radiological interventions performed at our department. Medical charts,
laboratory investigations, and radiological reports of all patients were carefully
reviewed to determine technical and functional outcomes of the interventions. Glomerular
filtration rate (GFR) was determined one day before and one day after partial nephrectomy
as well as one day before and one day after angioembolization. GFR was estimated using
the Cockcroft-Gault equation with the parameters age, sex, and serum creatinine levels
[15 ]. Technical success was defined as successful primary angiography-guided occlusion
of the arterial lesion. Clinical success was defined as no need for nephrectomy or
further operation on the vascular lesion during follow-up.
Pre-interventional imaging
In all patients, pre-interventional imaging was performed to localize and confirm
the clinically suspected vascular lesion. Patients underwent a biphasic contrast-enhanced
CT examination consisting of an arterial and venous phase of the abdomen as the primary
imaging modality. In patients with severely reduced GFR, contrast-enhanced ultrasound
was performed as an alternative imaging modality. CT examinations were performed on
a 64- or 128-row scanner (Siemens Healthcare® , Erlangen, Germany) with the use of 1.5 ml of contrast agent per kg of body weight
(Imeron 400® , Bracco Imaging, Konstanz, Germany) followed by a bolus of 100 ml of sodium chloride.
Sonographic evaluation consisted of an examination using B-mode, color duplex ultrasound,
and contrast-enhanced sonography. Contrast-enhanced sonography was performed after
injection of 0.8 – 1 ml of contrast agent consisting of microbubbles (SonoVue® , Bracco, Milan, Italy) followed by a bolus of 10 ml of NaCl. Ultrasound examinations
were carried out either on a LOGIQ E9® (GE Healthcare® , Little Chalfont, United Kingdom) or an S2000® (Siemens Healthcare® , Erlangen, Germany).
Interventions
Through a femoral access, a renal angiogram was first performed to localize the vascular
lesion using a 4F cobra catheter (Terumo® , Tokyo, Japan). Afterwards, superselective catheterization of the bleeding vessel
was performed using a coaxial microcatheter system (Progreat® , Terumo, Tokyo, Japan). When a stable catheter position was obtained, coil embolization
was performed until complete occlusion of the vascular lesion was achieved. The coil
type used depended on the preference of the interventionalist. The types and sizes
of the coils used will be presented in the results section. All interventions were
performed by board-certified interventional radiologists with at least five years
of experience in embolization procedures.
Volumetric analysis
Axially reconstructed images with 5-mm slice thickness from the pre- and post-interventional
CT scans in the portal venous phase were used for volumetric analysis. This was performed
using semi-automated volumetric software (Siemens Syngo.via® Oncology, Siemens Healthcare® ). A digital dash was drawn across the whole kidney, after which a volumetric segmentation
of the kidney was obtained. Manual correction of the contours was then performed.
The two volumes obtained (pre- and post-intervention) were recorded and the volume
shrinkage between pre- and post-intervention was calculated. Volumetric measurements
were performed in consensus by two diagnostic radiologists, one with three years and
the other with six years of experience in this procedure.
Ethics statement
All investigations have been conducted according to the principles expressed in the
Declaration of Helsinki. All patient records and information was anonymized and de-identified
prior to analysis. Therefore, and due to the strictly retrospective design of the
study, no written informed consent was obtained by the participants. The institutional
review board of the ethics committee of Ludwig-Maximilians-University confirmed the
study with a certificate of compliance (No. 526 – 15 UE).
Statistical analysis
Categorical variables are presented as absolute and relative frequencies. For descriptive
data with normal distribution, mean ± standard deviation (SD) was given, and comparisons
were performed by a 2-sided t test. A p-value < 0.05 indicated statistical significance. All analyses were performed
using SPSS® software (version 20, IBM SPSS Inc.).
Results
Between January, 2003 and September, 2013, a total of 1425 patients underwent partial
nephrectomy at our institution.
Of these, 39 (2.7 %) were referred to our department for transcatheter embolization
of pseudoaneurysms or arteriovenous fistulas (mean age 65.7 years; 30 males, 9 females).
Baseline demographics of the cohort are given in [Table 1 ]. In total, we had 26 patients with at least one renal pseudoaneurysm, 12 patients
with AV-fistulas, and 1 patient with a pseudoaneurysm and an additional AV-fistula.
The patients of this study underwent open, laparoscopic, or robot-assisted surgery;
the greatest proportion (82.1 %) was treated with open partial nephrectomy.
Table 1
Baseline demographics.Tab. 1 Demografische Daten zu Studienbeginn.
characteristics
number (percentage) or mean ± SD (range)
total number of patients
39 (100)
sex
30 (76.9)
9 (23.1)
age, mean ± SD (range) in years
65.7 ± 11.8 (32 – 81)
tumor size, mean ± SD (range) in cm
3.8 ± 1.6 (1.2 – 8.5)
operative approach
32 (82.1)
4 (10.3)
3 (7.6)
Patients presented clinical symptoms of the vascular lesion 15.3 days (mean) after
partial nephrectomy. Diagnosis was established by biphasic CT in 92.3 % of cases and
by CEUS in 7.7 % of cases. Occlusion of the vascular lesion was accomplished with
a mean of 4 coils (range 1 – 26). Tornado® coils (Cook, Bjaeverskov, Denmark) with a size range of 3 – 10 mm were used in 36
patients. An Interlock occlusion system® (Boston Scientific® , Marlborough, MA, United States) was employed in two patients (Interlock coils with
a size of 6x100 mm in one patient and one Interlock coil with a size of 12x200 mm
in the other patient). A detachable HydroCoil (Azur® , Terumo, Tokyo, Japan) was used for embolization in one patient. The mean intra-interventional
dose area product was 8563 cGy × cm2 (range 1287 – 36 701 cGy × cm2 ). Characteristics of the interventions are shown in [Table 2 ].
Table 2
Intervention characteristics.Tab. 2 Interventionscharakteristika.
characteristics
number (percentage) or mean ± SD/median (range)
days from operation to intervention
15.3 ± 9.7 (3 – 43)
imaging modality to confirm diagnosis
36 (92.3)
3 (7.7)
number of coils used
4 (1 – 26)
dose area product in cGy × cm2
8563 (1287 – 36 701)
fluoroscopy time in minutes
13 (2.5 – 31.4)
Overall, primary coil embolization was technically successful in 39 patients (100 %).
However, a second intervention with additional embolization was necessary for two
patients in this group (5.1 %). Clinical success, with no need for further operation
or nephrectomy, was achieved in 35 patients (89.7 %). Four patients (10.3 %) had to
undergo surgery after angioembolization due to persistent blood loss and hematuria.
In three of these four patients, persisting clinical symptoms were severe enough that
total nephrectomy had to be performed, and in one patient an operative revision with
suturing of the arterial lesion was performed. Examples for successful interventions
and preinterventional imaging are presented in [Fig. 1 ], [2 ] .
Fig. 1 A Contrast-enhanced CT demonstrating a renal artery pseudoaneurysm (white arrow) of
4.4 × 2.3 cm after partial nephrectomy on the right side. B Angiography confirms the pseudoaneurysm (white arrow) of the upper renal artery.
C Successful embolization of the feeding vessel of the pseudoaneurysm with six microcoils,
which were deployed via a co-axial system. D Postinterventional CT demonstrating complete occlusion of the feeding vessel of the
vascular lesion with microcoils.Abb. 1 A CT nach KM zeigt ein renales Pseudoaneurysma (weißer Pfeil) mit Größe von 4,4 × 2,3 cm
nach partieller Nephrektomie rechts. B Angiografische Bestätigung des Pseudoaneurysma (weißer Pfeil) des Nierenoberpolarterie.
C Erfolgreiche Embolisation des zuführenden Gefäßes des Pseudoaneurysmas mit 6 Microcoils
welche durch ein coaxiales System eingebracht wurden. D Postinterventionelles CT zeigt einen kompletten Verschluss des zuführenden Gefäßes
der vaskulären Läsion mit Microcoils.
Fig. 2 A Contrast-enhanced CT showing an arterial lesion (arrow) of 10 × 8 mm and a perirenal
hematoma at the upper pole of the right kidney. B Coronal MPR of the CT-scan. C Doppler sonography shows the blood flow of an AV-fistula with the feeding artery
and the draining vein. D CEUS confirms the AV-fistula. E Angiographic catheter position before embolization, the AV-fistula is visible (white
arrow). F Successful microcoil embolization of the artery, which fed the AV-fistula.Abb. 2 A CT nach KM zeigt eine arterielle Läsion (Pfeil) mit Größe von 10 × 8 mm und ein perirenales
Hämatom um den Oberpol der rechten Niere. B Koronare MPR der CT-Aufnahmen. C Dopplersonografie zeigt den Blutfluss der AV-Fistel mit zuführender Arterie und abführender
Vene. D Bestätigung der AV-Fistel durch CEUS. E Angiografische Katheterlage vor Embolisation, die AV-Fistel ist sichtbar (Pfeil).
F Erfolgreiche Microcoil-Embolisation des zuführenden arteriellen Gefäßes der AV-Fistel.
No typical complications of invasive angiography, such as dissections, groin hematoma,
or bleeding, were observed in any intervention. Although there was a significant reduction
of GFR between the pre-operative and post-operative days, there was no significant
difference of GFR between the pre-embolization and post-embolization days. Details
about the analysis of renal function are given in [Table 3 ].
Table 3
Changes in GFR.Tab. 3 Änderungen der GFR.
mean GFR before partial nephrectomy
64.75 ml/min
mean GFR one day after partial nephrectomy
48.82 ml/min
p < 0.01
mean GFR before embolization
mean GFR one day after embolization
p = 0.95
50.88 ml/min
50.96 ml/min
Determination of renal loss of volume could be determined in a subgroup of n = 10.
In the remaining 29 patients, post-interventional renal volume could not be determined
because patients were either lost to follow-up or underwent follow-up in a different
hospital or private practice or underwent follow-up by MRI. All pre-interventional
volume measurements were performed in the CT scan before embolization. Post-interventional
volume was measured in the first post-interventional CT scan, which was performed
100 days (median) after embolization. Mean renal volume loss was 25.2 ± 14.3 % between
the pre- and post-interventional CT scans, which is a significant loss (p < 0.05).
Examples of the volumetric measurements are presented in [Fig. 3 ]. Details about volume measurements are given in [Table 4 ].
Fig. 3 A Volumetric measurement of the left renal parenchyma before embolization. The freehand
ROI was carefully drawn around the parenchyma; perirenal hematoma and renal pelvis
was not included in the ROI. The pre-interventional volume of the left kidney was
176 ml. B Volumetric measurement 2 days after the embolization. Post-interventional renal volume
was 139 ml, so that the volume loss was 21 %. C Volumetric measurement of the left renal parenchyma in another patient. We demonstrate
the placement of the ROI in the sagittal MPR. The renal pelvis was spared out of the
measurement. The pre-interventional volume was 160 ml. D Post-interventional volumetric assessment in the same patient as in C. The post-interventional
volume of the left kidney is 135 ml, consistent with 15.6 % parenchymal loss.Abb. 3 A Volumetrie des linken Nierenparenchyms vor Embolisation. Die freihand-ROI wurde sorgfältig
um das Parenchym gezeichnet; das perirenale Hämatom und das Nierenbecken wurden nicht
in die ROI einbezogen. Das präinterventionelle Parenchymvolumen der linken Niere betrug
176 ml. B Volumetrische Messung 2 Tage nach Embolisation. Das postinterventionelle Parenchymvolumen
betrug 139 ml, der Volumenverlust lag daher bei 21 %. C Volumetrische Messung des linken Nierenparenchyms in einem weiteren Patienten. Gezeigt
wird die Lage der ROI in der sagittalen MPR. Das Nierenbecken wurde von der Messung
ausgespart. Das präinterventionelle Volumen lag bei 160 ml. D Postinterventionelle volumetrische Messung im selben Patienten wie in C. Das postinterventionelle
Volumen lag bei 135 ml, vereinbar mit 15,6 % Parenchymverlust.
Table 4
Changes in renal volume in each patient of n = 10.Tab. 4 Änderungen des Nierenvolumens in jedem Patienten von n = 10.
patient number
renal volume (ml) before embolization
renal volume (ml) after embolization
(Median 100 d)
loss of volume
(%)
1
137
124
9.5
2
192
169
12
3
182
154
15.4
4
160
135
15.6
5
143
114
20.3
6
176
139
21
7
184
127
31
8
117
80
31.6
9
165
97
41.2
10
186
85
54.3
mean loss of volume: 25.2 ± 14.3 % (p < 0.05)
Discussion
Pseudoaneurysms and arteriovenous fistulas are rare post-operative complications of
nephron-sparing partial nephrectomy. At our center, we identified 39 of 1425 patients
treated with nephron-sparing surgery (2.7 %) as being affected by one of these iatrogenic
lesions. This complication rate is consistent with other reports published in the
literature [11 ]
[16 ]
[17 ]
[18 ]
[19 ]. Since these complications are potentially life-threatening, it is mandatory to
achieve a reliable diagnosis when patients present signs and symptoms potentially
indicative of pseudoaneurysms and/or arteriovenous fistulas. Unfortunately, these
are often unspecific (as in the case of hematuria, flank pain, hypotension, and fever)
or delayed in their occurrence. It has been reported that some patients were even
asymptomatic and that the pseudoaneurysms or arteriovenous fistulas in these patients
were detected incidentally during examinations for follow-up or another condition
[17 ].
Most of our patients (82.1 %) diagnosed with either pseudoaneurysms or arteriovenous
fistulas had undergone open partial nephrectomy, while 10.3 % and 7.6 % were treated
with laparoscopic and robot-assisted procedures, respectively. This is in contrast
to other reports where a slightly higher occurrence is observed when patients received
laparoscopic surgery [12 ]
[17 ]. This discordance could be explained by the lack of large-scale randomized trials
and, possibly, the failure to take confounding factors into account. In addition,
a selection bias might be present, given the fact that patients with larger tumors
with difficult anatomy, such as a central location, and previously operated patients
are more likely to undergo open nephrectomy, which might explain the higher rate of
patients being treated with open nephrectomy in our cohort. As yet, no clear reason
is known why one or the other surgical procedure would predispose a patient to these
vascular complications [20 ].
Several diagnostic approaches were demonstrated as suitable for the assessment of
pseudoaneurysms or arteriovenous fistulas following partial nephrectomy. In 92.3 %
of our patient group, CT was used for pre-interventional imaging. This technique has
the advantage of not only the detection of pseudoaneurysms and arteriovenous fistulas
but also the diagnosis of other intra-abdominal pathologies such as hematoma or urinary
retention. Moreover, MDCT is a fast and reliable method [16 ]. However, pseudoaneurysms may be difficult to differentiate from adjacent arteries.
Thus, negative CT-examinations cannot definitely exclude pseudoaneurysms and additional
investigations may have to be performed [12 ]. In patients with severely impaired renal function, the use of nephrotoxic contrast
media can pose an unacceptably high risk. In these cases, contrast-enhanced ultrasound
or MRI without contrast agents are more suitable. For this reason, diagnostic imaging
of our patients with a severely lowered eGFR (7.7 %) was performed utilizing contrast-enhanced
ultrasound. However, contrast-enhanced ultrasound is highly dependent on the experience
and knowledge of the performing physician. Therefore, it may not be readily available
in many institutions.
Even if CECT and CEUS are preferred because of their non-invasive nature, percutaneous
angiography may be required although it presents a higher morbidity and mortality
risk and cannot image the entire urinary tract [18 ]. When there is a strong suspicion of a pseudoaneurysm or an arteriovenous fistula,
angiography with a subsequent embolization procedure may be a straightforward approach
[13 ]. Despite these alternative options, however, CECT is the first choice both for diagnosis
and for follow-up unless substantial contraindications exist [18 ].
In our cohort, symptoms indicating the existence of a vascular lesion and the consecutive
positive diagnoses thereof appeared 15.3 days (mean; range 3 – 43 days) after partial
nephrectomy. This time interval corresponds to findings in the literature which show
that pseudoaneurysms or arteriovenous fistulas were detected from one day up to several
months after surgery, with an approximate mean of between 8 and 15 days [11 ]
[12 ]
[17 ]
[18 ]
[21 ]
[22 ]. In one small study, the longest interval between operation and clinical symptoms
was even 3 months [23 ]. The largest study on this topic, which was recently published, examined the timing
and indications of transcatheter angiographic embolization for delayed hemorrhage
after percutaneous nephrolithotomy in 144 patients [24 ]. Their reported mean time to the onset of post-nephrolitomy hemorrhage was 10.5
days (range 2 – 30 days), which is in line with our results after partial nephrectomy.
These observations underline the need for alertness when symptoms begin to manifest
within the follow-up period.
Selective embolization of the leaking arteries was accomplished by complete occlusion
of the feeding arteries using microcoils. According to the interventionalists’ preferences,
several types of coils were utilized. In most cases (36 of 39 patients) Tornado® coils were employed (similar to a. o. [13 ]
[18 ]). Coil embolization proved to be highly effective and allowed for complete closure
of the target vessels. However, autologous blood clots, detachable balloons, gelatin
sponges, and several liquid embolic agents (such as glue and the non-adhesive Onyx® ) have also been shown to perform adequately. The latter impose a higher risk of reopening
or are associated with higher costs. In specific anatomical situations as well as
in patients with poor coagulation parameters, however, they may be the better choice
[13 ]. Microcoil embolization enables isolation of the vascular lesion while retaining
maximal renal vascularization, thus ensuring as much preservation of the renal parenchyma
as possible [16 ].
The intervention was technically successful in 100 % and clinically successful in
89.7 % of our patients. In two patients of our collective, a second transcatheter
embolization was required, which resulted in good clinical success, as demonstrated
by absence of bleeding in DSA and normalization of blood pressure and renal function.
These high success rates are in agreement with those reported in literature [12 ]
[13 ]
[16 ]
[17 ]
[25 ]. In four patients, the vascular lesions could not be managed with transarterial
embolization, so radical nephrectomy was required for three of them and operative
suture of vessels was necessary for the other one. No typical complications of invasive
angiography, such as dissections, groin hematoma, or bleeding, were observed; this
demonstrates that angiographic microembolization performed by experienced interventional
radiologists is a safe procedure.
A review of the eGFR values before and immediately after surgery and transcatheter
embolization shows that renal function significantly decreased after partial nephrectomy.
Embolization of the vascular lesions, however, did not impair renal function any further
as measured by eGFR. These observations are similar to those reported by Gahan et
al. [11 ]. Interestingly, although they found that patients with diabetes mellitus did show
a greater decrease in eGFR values before and after intervention, statistical significance
was absent, likely because of the small patient sample.
To our knowledge, we are the first to quantitatively evaluate renal parenchymal volume
before and after coil embolization procedures ([Fig. 3 ]). We observed a mean loss of volume of 25.2 % of renal parenchyma after embolization,
which suggests a significant reduction of functional renal parenchyma. However, this
volume loss of renal parenchyma was not associated with a significant reduction of
renal function, as proven by eGFR measurements. Only one recently published study
has semiquantitatively assessed renal parenchymal volume loss after embolization [26 ]. They reported a mean volume loss of 5 % with a range of 1 – 50 %, which is in line
with the range of our study.
Radiation exposure in patients associated with embolization of pseudoaneurysms and
arteriovenous fistulas following partial nephrectomy has not been reported previously.
The mean dose area product was 8563 cGy × cm2 , with a wide range of 1287 – 36 701 cGy × cm2 . The wide range may be due to the different generations of angiographic equipment
used during the ten-year period of the study as well as the different extent and severity
of the vascular lesions treated, which is also reflected in the wide range of coils
used (1 – 26). More effort should be made to reduce the radiation doses of these interventions
in the future. The mean fluoroscopy time was 13 minutes, which is an acceptably low
exposure time. Nevertheless, a shortening of the latter could reduce the radiation
effect since a correlation exists between the two [27 ]. Several limitations of this study inherent to its retrospective nature and small
sample size have to be taken into account. However, since pseudoaneurysms and arteriovenous
fistulas following partial nephrectomy are rare, our study group is the largest reported
to date. The analysis of renal function was performed by determining the eGFR according
to the Cockgroft-Gault formula, which is only an approximate indicator of renal function.
Volumetric analysis of the renal parenchyma could only be performed in a small subgroup
(n = 10). This was due to the fact that in the remaining 29 patients, post-interventional
renal volume could not be determined because patients were either lost to follow-up
or underwent follow-up in a different hospital or private practice or underwent follow-up
by MRI. Moreover, the time interval between embolization and the follow-up CT was
variable. Finally, a long-term follow-up of the kidney function would provide more
information.
Conclusion
Transcatheter angioembolization is a promising method for treating pseudoaneurysms
and arteriovenous fistulas, which are rare but dangerous vascular complications which
may occur after partial nephrectomy. Our results show a high technical success rate
and no significant decrease in renal function in the early post-embolization period.
We assessed the changes in parenchymal volume and radiation doses to which patients
are exposed during intervention. In order to provide a clear answer as to which method
of diagnosis and angioembolization treatment is superior with regard to maintaining
as much renal function as possible, large-scale prospective studies, including a long-term
follow-up of renal function, should be performed.
Clinical Relevance of the Study
Renal pseudoaneurysms and arteriovenous fistulae can be treated successfully and safely
with endovascular embolization.
Although embolization reduces the renal parenchymal volume, it preserves the organ
and its function.
Arterial lesions, which occur after partial nephrectomy should therefore be treated
first with embolization by interventional radiologists.
