Professional Documents
Culture Documents
Efficacy and Morbidity of Therapeutic Renal Embolization in The Spectrum of Urologic Disease
Efficacy and Morbidity of Therapeutic Renal Embolization in The Spectrum of Urologic Disease
ABSTRACT
Purpose: We report the largest series of renal embolizations performed for a variety of indications.
Patients and Methods: A retrospective analysis was performed on embolizations performed in our institu-
tion from 1997 to 2002 encompassing 36 patients who underwent 44 procedures.
Results: Embolization was successful on the first attempt in 87% of the patients. A second embolization was
performed in four of the five unsuccessful cases, three successfully, increasing the success rate to 95%. The
mean postoperative narcotic use was 27.2 mg of morphine equivalent, and 10 mg or less was required by 45%
of the patients. In the 14 patients who had not also undergone a surgical procedure, the mean narcotic use
was 21 mg, and 64% required 10 mg or less. Only 15% of the patients developed fever, which resolved within
2 days in all cases. Leukocytosis was seen in 47%. Follow-up creatinine and hypertension information was
available in 16 and 18 patients, respectively. After a mean follow-up of 269 days, only one patient had a clin-
ically significant rise in the creatinine concentration. After a mean follow-up of 496 days, two patients had
new-onset hypertension. There was no statistically significant difference in the success rate, narcotic use, com-
plications, creatinine concentrations, or the likelihood of fever, leukocytosis, or hypertension according to the
indication for embolization or the agent used. Use of a microcatheter was associated with less parenchymal
loss, and decreased parenchymal loss was associated with a significant reduction of narcotic use.
Conclusions: Renal embolization is a highly effective and well-tolerated procedure in a variety of urologic
conditions. The indications and material used did not have a significant effect on the outcome. Reducing
parenchymal loss can significantly reduce morbidity.
INTRODUCTION hort, and the data for this subgroup were not analyzed sepa-
rately.
Departments of 1Urology and 2 Radiology, Long Island Jewish Medical Center, New Hyde Park, New York.
385
386 JACOBSON ET AL.
dures, including four rescue procedures after failed primary at- microcatheter was utilized to reach smaller peripheral vessels.
tempts. The patient age ranged from 16 to 77 years, with a mean After embolization, the catheter was withdrawn into the main
of 47.6 years. Twenty-eight percent of the patients had elevated renal artery, and an angiogram was performed to confirm ab-
serum creatinine concentrations (.1.3 mg/dL) before the pro- sence of flow to the embolized segment and evaluate the re-
cedure, and 25% had controlled hypertension. The most com- maining renal vessels. All postembolization angiograms were
mon indication for embolization was control of hemorrhage af- reviewed by two radiologists, who estimated the percent
ter percutaneous renal surgery (N 5 23): nephrolithotomy (17), parenchymal loss. The mean of the two values was used for
endopyelotomy (4), or resection of transitional-cell carcinoma further analysis.
(2) (Fig. 1). In this group of patients, 43% presented immedi- All statistics were analyzed using commercially available
ately postoperatively, 13% had bleeding on removal of the software. Continuous variables were analyzed with the Mann-
nephrostomy tube, and 43% returned with hematuria after hav- Whitney rank-sum test or the Kruskal-Wallis test. Differences
ing been discharged home. Patients in the latter group were in proportions were analyzed using the chi-square test or the
treated a mean of 9.8 days after the initial surgery. The other Fisher exact test, as appropriate.
indications for embolization were angiomyolipoma (Fig. 2),
nonfunctioning kidneys (Fig. 3), congenital arteriovenous mal-
formation (Fig. 4), postbiopsy or postheminephrectomy bleed- RESULTS
ing, and prophylaxis of bleeding during nephrectomy for a large
renal-cell carcinoma with tumor thrombus (Table 1). The em- The procedure was successful at the first attempt in 34 of
bolization material was steel coils alone in 60%, coils and the 39 cases (87%). A second embolization was performed in
polyvinyl alcohol particles (PVA) in 23%, and PVA alone in four of the five initially failed cases. The fifth patient was suf-
12%. One patient was embolized using N-BCA glue, and in a fering from advanced AIDS and was found to be coagulopathic;
final patient, a 5F cobra catheter became occlusive, and no fur- bleeding resolved after multiple transfusions and treatment of
ther embolization material was required. the coagulopathy. The second embolization was successful in
The procedure was performed through a 6F sheath placed in three of the four cases. If these are included as successes, the
the femoral artery following puncture using the standard total success rate improves to 95%. In the patient who under-
Seldinger technique. Digital subtraction angiography was uti- went embolization prior to radical nephrectomy for a large re-
lized in all cases. Selective arteriography was accomplished af- nal cancer with vein thrombus, success was not objectively as-
ter abdominal aortography with a 5F flush catheter by ad- sessable.
vancement of a 5F cobra catheter or, when necessary, a 5F The mean narcotic use was 27.2 mg of morphine equivalent.
Simmon catheter, into the renal artery. The catheter was then Only 10 mg or less was required by 45% of patients, and, of
advanced selectively over a nontraumatic wire until it was as these, half required none. However, it is difficult to distinguish
close as possible to the desired site of embolization, and the postembolization pain from postoperative pain in patients who
embolization material was deployed. When appropriate, a 2.8F had undergone a recent surgical procedure. The subgroup of 14
A B
FIG. 1. Control of hemorrhage after percutaneous nephrolithotomy. (A) Postprocedure bleeding. (B) After embolization.
RENAL EMBOLIZATION IN UROLOGY 387
B C
FIG. 2. Embolization of angiomyolipoma. (A) CT scan in patient with tuberous sclerosis and bilateral angiomyolipomas.
(B) Preembolization angiogram. (C) Postembolization angiogram.
patients who had not undergone surgery therefore was evalu- Delayed morbidity of embolization can include renal failure
ated separately. Their mean narcotic use was 21 mg, with 64% and hypertension. Although it is admittedly an insensitive
required ,10 mg and 36% requiring none. marker, the serum creatinine concentration was used as a clin-
Fever was noted in only 15% of patients and resolved within ical measure of renal function. Follow-up creatinine values were
2 days in all. However, transient leukocytosis was common, oc- available in 16 patients at a mean of 269 days. Only one pa-
curring in 47% of the patients. Complications were seen in 10% tient was noted to have a significant rise, from 1.2 mg/dL to
of cases. None was directly related to the technique of em- 1.9 mg/dL, and it is unlikely that the rise was attributable to the
bolization, and there were no nontarget embolizations and no embolization, as the procedure was performed for a nonfunc-
abscesses. One patient suffered a myocardial infarction. Two tioning, chronically infected kidney. Postoperative blood pres-
patients continued to have profuse bleeding and developed hy- sure measurements were available in 18 patients, with a mean
povolemic shock. One patient, who failed initial and repeat em- follow-up of 496 days. Sixty-seven percent of the patients were
bolization, underwent nephrectomy and died of sepsis in the normotensive, 22% had hypertension unchanged from prior to
postoperative period. embolization, and 11% (2 patients) developed new-onset hy-
388 JACOBSON ET AL.
A B
FIG. 3. Embolization of nonfunctioning kidney. (A) Angiogram before embolization. (B) Postembolization angiogram.
pertension. Both patients’ hypertension was controlled, one ference between the observers was 2.6%, and the Spearman
with a single daily dose of a beta-blocker and one without med- rank correlation coefficient was 0.96 with a P value of ,0.05.
ication (Table 2). The mean parenchymal loss was 33.8%. However, in eight pa-
Postembolization angiograms were available in 39 patients. tients, the aim was 100% embolization, so these patients were
The interobserver correlation was very strong: the mean dif- excluded from the second analysis of attempted selective em-
A B
FIG. 4. Control of arteriovenous malformation. (A) Pretreatment angiogram shows upper-pole lesion. (B) Postembolization
angiogram.
RENAL EMBOLIZATION IN UROLOGY 389
TABLE 1. INDICATIONS FOR EMBOLIZAT ION IN 40 PATIENTS ,10 mg of morphine equivalent; of these, half did not require
No. any narcotics at all. Fever, was noted in only 15% of patients,
and all fevers resolved within 2 days. The most common find-
PCNL complication 17 ing was an elevated leukocyte count, which occurred in 47%
Angiomyolipoma 8 of patients. However, although leukocytosis may be a source
Antegrade endopyelotomy complication 4 of concern to physician and patient alike, on its own, it does
Nonfunction 4 not cause symptoms and should not be included as a source of
Percutaneous resection of TCC 2 morbidity.
Arteriovenous malformation 2
Long-term morbidity was assessed with creatinine concen-
Renal biopsy 1
Partial nephrectomy 1 trations and the development of hypertension. A significant rise
Prenephrectomy 1 in serum creatinine was noted in only 1 of 16 patients in whom
this follow-up information was available. As mentioned earlier,
this patient underwent embolization for a chronically infected
nonfunctioning kidney, so it is unlikely that the embolization
bolizations. In these patients, the mean parenchymal loss was contributed to the renal failure. New-onset hypertension was
16.7%. In 49% of patients and 61% of the patients having se- seen in 2 of 18 patients in whom follow-up blood pressure mea-
lective embolization, a 2.8F coaxial microcatheter was utilized. surements were available. Both of these patients were controlled
In cases of selective embolization, the mean parenchymal loss by minimal therapy. There is very little postrenal-embolization
in the microcatheter group was significantly reduced (12.2% v follow-up information available in the literature. Takebayashi
23.7%, P 5 0.017). Parenchymal loss was also analyzed with and associates17 reported on 30 patients treated with emboliza-
regard to its effect on success, fever, leukocyte count, narcotic tion for renal arteriovenous malformations. After a mean fol-
use, and complications. Although successful and uncomplicated low-up of 6.2 years, only one patient had developed hyperten-
cases had less parenchymal loss (25.9% v 45.8% and 29.5% v sion, and all had normal creatinine concentrations.
63.5%), only the difference in narcotic use was found to be sta- The various indications for embolization did not significantly
tistically significant. Patients who required .10 mg of narcotics affect outcome. However, it is possible that our small subgroups
had significantly greater parenchymal loss than those who re- lacked the power to demonstrate differences. In Hemingway
quired ,10 mg (38.5% v 19.9%; P 5 0.01). and Allison’s study, embolization was associated with postem-
Success, narcotic use, fever, leukocyte count, complications, bolization syndrome in 52.2% of tumor cases and only 11.1%
and parenchymal loss were also evaluated in relation to the in- of bleeding cases.13 This study was not specifically aimed at
dications for the embolization and the embolization materials renal procedures. However, in the kidney, the same seems to
used (Tables 3 and 4). The only significant difference was that hold true. Wallace and colleagues18 evaluated 100 cases of em-
parenchymal loss was significantly greater in patients with non- bolization for renal-cell carcinoma and reported that postem-
functioning kidneys than in those treated after percutaneous bolization syndrome occurred in almost every patient. Transient
stone extraction (P 5 0.046). However, in the former group, hypertension was also seen frequently but was not persistent in
complete renal embolization was attempted in three of the four any case. In two series that evaluated embolization for bleed-
patients. The remaining patient had embolization of a non- ing, fever was seen in 1 of 15 and 1 of 16 patients, and only 1
functioning upper-pole moiety of a duplex system. The mater- patient developed hypertension, which did not persist.19,20 The
ial used did not significantly affect any outcome measure. increased morbidity in the tumor group might be explainable
by a larger area of ischemia (although this information was not
available) or by tumor necrosis.
DISCUSSION
TABLE 2. MORBIDITY OF EMBOLIZATION
The efficacy of therapeutic renal embolization has been shown
in numerous series to be between 82% and 100%.6,10–12,14–17 In No. (%)
our patients, the success rate of the initial procedure was 87%. of patients
In four patients who failed initial embolization, a second em-
Early
bolization was attempted and was successful in three, improv-
Pain necessitating .10 mg 22 (55%)
ing the overall success rate to 95%. Open exploration, espe-
narcotics
cially in the face of severe bleeding, mandates a large incision In patients without other 5 (36%)
and carries a significant risk of nephrectomy. In hemodynami- surgery
cally unstable patients, this is clearly warranted. However, in Fever 6 (15%)
the stable patient, a 75% success rate justifies an attempt at res- Leukocytosis 18 (47%)
cue embolization prior to submitting the patient to the morbid- Complication 4 (10%)
ity of an open procedure. Hypovolemic shock 2
In this series, renal embolization was well tolerated. Classi- Myocardial infarction 1
cally, embolization is associated with a transient (3–5 days) Death 1
Late
postembolization syndrome, which consists of pain, fever, and
Creatinine rise 1 (6%)
leukocytosis. Although some discomfort was common, in pa- New-onset hypertension 2 (11%)
tients who had no other important source of pain, 64% required
390 JACOBSON ET AL.
Renal
Narcotic loss
Success .10 mg Fever WBC Complication Cr HTN (%)
In other studies where large-volume ablation was performed, Gelfoam is reabsorbed within days to weeks. However, in an
morbidity was also increased. Da Baere et al12 reported 20 cases end-artery system such as the kidney, one would not expect
of embolization for persistent urine leak. All except one patient nephrons to survive this prolonged period of ischemia. Absolute
complained of fever and nausea, and pain was reported by 55%. alcohol has also been promoted as the embolization agent of
In a series of embolizations for polycystic kidney disease,11 choice.22 Its effect is mediated through direct endothelial dam-
where an extremely large area was targeted, pain and fever was age and perivascular necrosis. Vascular resistance can increase
seen in all patients. In fact, pain was so severe that most pa- rapidly, causing reflux into unselected vessels.23 As well, al-
tients required epidural anesthesia. In our series, pain was sig- though it is rarely seen if the agent is used properly, systemic
nificantly associated with greater parenchymal loss, and use of toxicity is always a concern. Other inert materials such as de-
a microcatheter, which can be passed coaxially through stan- tachable balloons24,25 and cyanoacrylate26 have also been de-
dard angiographic catheters into 1- to 2-mm peripheral vascu- scribed. Despite our personal preferences, until a significant ad-
lar branches, was significantly related to less parenchymal loss. vantage is demonstrated, all commercially available agents are
The use of a microcatheter, when appropriate, therefore should acceptable. It is key that the physician performing the emboliza-
decrease patient morbidity. tion have knowledge of the physical properties of each material
Many embolization agents have been employed. The most and understand their potential advantages and limitations.
popular are absorbable materials such as Gelfoam and autolo-
gous clot and nonabsorbable materials such as metal coils, PVA
particles, and absolute alcohol. To our knowledge, there has CONCLUSION
been no study demonstrating a significant advantage of one ma-
terial over another. Our personal preference is steel coils, which Renal embolization is highly effective and can be performed
are inexpensive, easily deployed, and highly effective. Coils with minimal short- and long-term morbidity and excellent re-
alone were utilized in 60% of our cases. In patients in whom nal preservation. Embolization therefore represents an attrac-
only partial occlusion was noted, PVA particles were added to tive alternative to open surgery for a wide spectrum of urologic
complete the occlusion, and in 12% of patients, PVA particles conditions. In the present study, different indications and em-
were used alone. No significant difference was found in the re- bolization agents were not significantly associated with out-
sults according to the material used. Other authors have advo- come measures. However, the percent parenchymal loss was
cated Gelfoam as the material of choice for renal emboliza- found to be associated with the amount of narcotic use. The
tion.15 Although Gelfoam is compressible when wet, it has a goal of embolization should be to target the affected area as
tendency to fragment or plug the catheter, and when force is specifically as possible, sparing normal parenchyma. The use
used to eject the plugs, the turbulence can lead to reflux and of a microcatheter for superselective embolization significantly
nontarget embolization.21 The theoretical advantage is that decreased damage to nontarget areas.
Renal
Narcotic loss
Success .10 mg Fever WBC Complication Cr HTN (%)