Anda di halaman 1dari 8

The British Journal of Radiology, 82 (2009), 724731

Pre-operative renal arterial embolisation does not provide


survival benefit in patients with radical nephrectomy for renal
cell carcinoma
1

M MAY, MD, 1S BROOKMAN-AMISSAH,


F KENDEL, PhD

MD,

S PFLANZ,

MD,

J ROIGAS,

MD,

B HOSCHKE,

MD

and

Department of Urology, Carl-Thiem Hospital, Cottbus, Berlin, 2Department of Urology, Vivantes-Clinic Am Urban, Berlin
and 3Institute of Medical Psychology, Charite - Universitatsmedizin Berlin, Germany

ABSTRACT. Currently, there is no widespread use of percutaneous renal artery


embolisation (PRAE) as a pre-operative treatment in the management of renal cell
carcinoma (RCC). There is also a scarcity of studies concerning the potential benefits of
this procedure. All patients with RCC who underwent pre-operative PRAE before
nephrectomy (n5227) and all patients solely undergoing surgery (n5607) at our
institution from 1992 to 2006 were included. Information on techniques used,
perioperative transfusion requirements, pathological and clinical variables, acute
toxicity and complications were obtained from a retrospective review of medical
records. Propensity modelling techniques were used to compare cancer-specific survival
(CSS) and overall survival (OS) in both groups. Propensity scores were calculated from a
logistic matching model including age, gender, clinical tumour size, grading, pN stage,
cM stage, pT stage, histology and microvascular invasion. This resulted in 189 matches.
The mean follow-up of the entire group of matched patients was 81 months. The 5year actuarial CSS and OS for the total group of matched patients was 80.8% and
73.9%, respectively. CSS and OS did not show any significant differences between the
matched treatment groups. There were no statistical differences in surgical
complications between all patients treated with pre-operative PRAE (n5227) and all
patients without PRAE (n5607), except for blood transfusion (61% vs 24%; p,0.01).
Symptoms of post-embolization syndrome, including lumbar pain, fever, nausea,
hypertension and macroscopic haematuria, were reported by 202 patients (89%), in
most cases being mild and self-limited. There is no conclusive evidence that preoperative PRAE provides survival benefits in the management of surgically resected
RCC.

The incidence of renal cell carcinoma (RCC) has been


increasing continuously since the 1970s from 2.3% to
4.3% per year, resulting in more than 35 000 new
diagnoses and almost 13 000 deaths in 2005 in the USA
alone [1]. Currently, more than 60% of RCC cases are
detected incidentally [2]. When discovered in its early
stages, the disease is often curable, whereas advanced or
metastatic RCC usually has a fatal prognosis and
outcome.
Radical nephrectomy, performed as a traditional open
surgical approach, has been the standard operation for
non-metastatic RCC. However, over the past two
decades, there was a significant shift in the surgical
management of kidney cancer to partial nephrectomy,
laparoscopic nephrectomy and laparoscopic partial
nephrectomy as alternative treatment options [3].
In the 1970s, percutaneous renal artery embolisation
(PRAE) was introduced into clinical practice [4, 5]. The
Address correspondence to: Matthias May, Department of Urology,
Carl-Thiem Hospital Cottbus, University Teaching Hospital,
Charite zu Berlin, Thiemstrasse 111, D-03048 Cottbus, Germany.
E-mail: matthias.may1@web.de

724

Received 19 May 2008


Revised 22 September
2008
Accepted 6 October 2008
DOI: 10.1259/bjr/17514226
2009 The British Institute of
Radiology

initial indications for PRAE were limited to symptomatic


haematuria and palliation for metastatic RCC [4, 5]. With
technical advances and growing experience, the indication has extended to include pre-operative embolisation
in RCC patients before nephrectomy. Although a
significant number of studies on pre-operative PRAE
have been conducted on patients with RCC, there is as
yet no consensus on the pros and cons of this procedure
[68].
However, PRAE can facilitate nephrectomy by reduction of intra-operative blood loss and operative time [9,
10]. An enhancement of prognosis could probably be
realised by prevention of tumour cell spillage after
primary ligation of the renal vein and by immunomodulation with a tumour-specific response, stimulated by
a PRAE-induced tumour necrosis factor [6, 1113].
Nevertheless, there is currently no widespread use of
PRAE as a pre-operative treatment in the management of
RCC [14]. To date, there is also a scarcity of studies
concerning the potential benefits provided by this
procedure.
We therefore evaluated our experience with PRAE in
patients with RCC undergoing consecutive nephrectomy
The British Journal of Radiology, September 2009

Value of pre-operative renal artery embolisation in RCC

in order to determine the influence of pre-operative


PRAE on survival.

acute toxicity of PRAE and all relevant post-operative


complications were assessed within 30 days after surgery.

Methods and materials


Study subjects

Pre-operative percutaneous renal artery


embolisation

The study group (SG) of the present retrospective


single centre evaluation included 834 patients (528 male,
306 female) with RCC, who underwent nephrectomy
(n5710) or nephron-sparing surgery (n5124) in the time
between 1992 to 2006. In all patients, RCC was assured in
the definitive histological report after surgery. Patients
with bilateral synchronous tumours, von HippelLindau
syndrome and histology incompatible with RCC were
excluded from the study. Additional criteria for exclusion were defined as follows: ductus Bellini carcinoma,
non-classified histology and a follow-up ,12 months
without a defined end point. In 227 patients (27.2%) of
the SG a percutaneous renal artery embolisation (PRAE)
was performed 112 days prior to surgery.

Between 1992 and 1997, a total of 227 patients were


treated with a planned pre-operative PRAE in order to
achieve a complete cessation of the arterial flow. The
time of embolisation varied. Embolisation was performed 13 days before surgery in 161 (70.9%) patients,
45 days before surgery in 63 patients (27.8%) and
exactly 12 days before surgery in the remaining three
patients (1.3%). In all patients, PRAE was performed
with local anaesthesia (Seldinger procedure using
Gelfoam; Pharmacia & Upjohn, Kalamazoo, MI).
Additionally, GianturcoWallace devices (2.5 cm wire
coil with strands of Dacron) were applied to 216 patients
(95.2%). In all patients, the success of PRAE was affirmed
by subsequent renal angiography.

Study variables

Statistical analysis, aftercare and study groups

In all patients of the SG, pre-operative CT of the


abdomen was performed, completed by MRI where
applicable. Furthermore, patients underwent X-ray or CT
of the thorax pre-operatively. In case of specific
symptoms, an additional cranial CT (CCT) and/or
radioisotope bone scan was performed. Hence, pretherapeutic variables maximum clinical tumour size
and cM stage, as well as patients age and gender were
assessed.
Apart from pT stage, which was assessed according to
the sixth edition of the World Health Organization
classification (2002), histological cell type, maximum
pathological tumour size, grading by Thoenes, pN stage
and microvascular invasion (MVI) were evaluated.
In all patients with pre-operative PRAE, the time
interval between this intervention and nephrectomy,

Within the first 2 years after surgery, post-operative


aftercare included physical examination and ultrasound
every 3 months, as well as CT of the abdomen and
thorax every 6 months. After the second year, physical
and ultrasound examination was performed every
6 months, and CT of the abdomen and thorax once a
year. Further diagnostic investigation (i.e. MRI, CCT and
radioisotope bone scanning) was carried out individually
and according to existing symptoms of the patient and in
case of tumour progress. Adjuvant treatment was not
indicated in any patient. If metastatic lesions appeared
during the follow-up, various therapeutic options, i.e.
surgical treatment of metastases, systemic immunotherapy, radiotherapy and best supportive care, were
individually applied according to the distribution of
metastasis and general condition of the patient. In all

1.00

Propensity score

0.80

0.60
No embolisation

Embolisation

0.40

0.20

0.00
100

80

60

40

20

0
20
Frequency

The British Journal of Radiology, September 2009

40

60

80

100

Figure 1. Mirrored histogram distribution of propensity scores for


patients in the embolisation (percutaneous renal artery embolisation)
and non-embolisation (nephrectomy alone) groups.
725

M May, S Brookman-Amissah, S Pflanz et al


Table 1. Patient characteristics of matched patients (n5378) (total group and stratified according to status of pre-operative
PRAE)
Variables

Mean follow-up (SD) (months)


Mean age (SD) (years)
Male gender
pT stage
pT1a
pT1b
pT2
pT3a
pT3b/pT4
pN stage
pN0/pNx
pN+
cM stage
M0
M1
Grading
G1
G2
G3
Microvascular invasion
Absent
Present
Clinical tumour size (SD) (cm)
Histology
Clear cell carcinoma
Papillary carcinoma
Chromophobe carcinoma
Spindle cell carcinoma (pleomorph)

All patients (n5378)

81.0 (54.4)
60.3 (90.4)
227 (60.1%)
51
154
65
48
60

(13.5%)
(40.7%)
(17.2%)
(12.7%)
(15.9%)

PRAE (study group 1)


(n5189)

99.7 (56.0)
60.4 (8.0)
114 (60.3%)
22
84
35
19
29

(11.6%)
(44.4%)
(18.5%)
(10.1%)
(15.3%)

Nephrectomy alone
(study group 2) (n5189)

62.3 (45.7)
60.2 (10.6)
113 (59.8%)
29
70
30
29
31

p-valuea

,0.01
0.77
1.00
0.58

(15.3%)
(37.0%)
(15.9%)
(15.3%)
(16.4%)
0.79

364 (96.3%)
13 (3.7%)

181 (95.8%)
8 (4.2%)

183 (96.8%)
6 (3.2%)

348 (92.1%)
30 (7.9%)

174 (92.1%)
15 (7.9%)

174 (92.1%)
15 (7.9%)

20 (5.3%)
309 (81.7%)
49 (13.0%)

9 (4.8%)
157 (83.1%)
23 (12.2%)

11 (5.8%)
152 (80.4%)
26 (13.8%)

324 (85.7%)
54 (14.3%)
6.83 (2.37)

161 (85.2%)
28 (14.8%)
6.85 (2.14)

163 (86.2%)
26 (13.8%)
6.81 (2.58)

297
48
25
8

144 (76.2%)
29 (15.3%)
13 (6.9%)
3 (1.6%)

153 (81.0%)
19 (10.1%)
12 (6.3%)
5 (2.6%)

1.00

0.90

0.88

(78.6%)
(12.7%)
(6.6%)
(2.1%)

0.83
0.41

PRAE, percutaneous renal artery embolisation; SD, standard deviation.


a
p-Value refers to the comparison between matched groups.

cases of solitary local relapse (LR), surgical resection of


the LR was intended.
Cancer-specific survival (CSS) and overall survival
(OS) were defined as primary end points. Survival
curves were estimated by the product limit method of
KaplanMeier and compared by log-rank statistics.
Univariate Cox proportional hazard models were fitted
to determine the clinical and pathological features that
were significantly associated with survival. A multivariate model was built using a stepwise selection
procedure, with the p-value for a feature to enter or
leave the model appointed at p,0.10.
To compare survival rates between the two treatment
groups, patients were matched with respect to oncological and baseline characteristics. A propensity score, in
logit units, was then obtained by calculating the
predicted value for each observation using a logistic
model, including age, gender, clinical tumour size,
grading, pN stage, cM stage, pT stage, histology and
MVI. Thus, the propensity score represented a summary
of the variables included in the logistic model for each
observation. Greedy matching techniques were then
used to select counterparts of the non-embolisation
group from the patients of the embolisation group by
choosing the patient with the nearest propensity score
[15]. The distribution of propensity scores (Figure 1)
resulted in 189 matches (83% of the embolisation patients
and 31% of the non-embolisation patients), reflecting a
large overlap in baseline characteristics among the two
treatment groups. The balance in covariates between
726

treated and untreated subjects was assessed using the


paired t-test and the Wilcoxon signed-rank test or
McNemars test, respectively. Univariable comparisons
were made among the unmatched groups using the t-test
for independent samples and the x2 test.
A detailed description of important patient and
tumour parameters in the matched 189 pre-operatively
embolised patients (SG 1) and 189 patients undergoing
radical nephrectomy alone (SG 2) is given in Table 1.
Mean follow-up of the entire group of matched patients
was 81 months. In those patients who were still alive at
the end of follow-up (n5241; 64%), mean follow-up was
100.3 months. Statistical analyses were performed using
SPSSH (version 15.0; SPSS Inc., Chicago, IL). In all tests, a
p-value #0.05 was considered significant.

Results
In all 189 matched patients who underwent preoperative PRAE, the result was considered satisfactory at
control angiography. However, in six cases (3.2%),
incomplete embolisation was recorded by the surgeon.
There were no cases of coil migration, adjacent organ
injury or PRAE-related death.
The 5-year actuarial CSS and OS for the entire group of
378 matched patients was 80.8% and 73.9%, respectively.
A comparison of survival between the 189 SG 1 patients
treated with pre-operative PRAE and the matched 189
SG 2 patients treated with nephrectomy alone showed a
The British Journal of Radiology, September 2009

Value of pre-operative renal artery embolisation in RCC

Cancer-specific survival

1.0
Nephrectomy
alone
PRAE

0.8

0.6

0.4

0.2
Figure 2. Cancer-specific survival

p = 0.085

stratified according to status of


pre-operative percutaneous renal
artery embolisation (PRAE) in 378
patients with renal cell carcinoma
(matched groups).

0.0
0

50
100
150
Months since nephrectomy

200

differences in surgical complications between patients


treated with PRAE (n5227) and patients treated with
nephrectomy alone (n5607) except for blood transfusion
(61% and 24%, respectively; p,0.01). Symptoms of postembolisation syndrome, including lumbar pain, fever,
nausea, hypertension and macroscopic haematuria, were
reported in 202 patients (89%). The vast majority of these
symptoms were mild and self-limited (Table 5).

1-, 5- and 8-year CSS of 95%, 79% and 70%, respectively,


for the former, and 93%, 83% and 79%, respectively, for
the latter group (p50.085; Table 2, Figure 2). OS after 1, 5
and 8 years was 93%, 73% and 62%, respectively, for SG
1 compared with 91%, 75% and 67%, respectively, for SG
2 (p50.677; Table 2, Figure 3).
Cox multiple regression models were used to determine whether age, gender, pT stage, pN stage, cM stage,
grading, tumour size, MVI or pre-operative PRAE
influenced post-operative CSS and OS. Tables 3 and 4
demonstrate that a higher pathological T stage, a positive
pN stage, clinical metastases, a higher grading and a
clinical tumour size .5.0 cm independently predicted
worse CSS and OS after renal surgery. In contrast, age
.60 years, gender, positive MVI and status of PRAE
were not independent predictors for CSS (Table 3) and
OS (Table 4) after nephrectomy.
Surgical complications within 30 days after surgery
were recorded for the entire study cohort (n5834,
Table 5): 8 (1.0%) patients died within this time period.
Re-operation because of secondary haemorrhage was
recorded in 24 (2.9%) patients. There were no statistical

Discussion
In an attempt to improve the management of RCC,
artificial necrosis of the tumour-bearing kidney by
percutaneous PRAE was introduced by Almgard et al
[4] in 1973. Pre-operative PRAE was used as a surgical
adjunct to facilitate radical nephrectomy by decreasing
intra-operative blood loss and by causing oedema in
tissue planes. The technique was also administered in
patients with non-resectable tumours receiving palliative
care and in those with significant symptoms, such as
haemorrhage [16, 17].

1.0
Nephrectomy
alone
PRAE

Overall survival

0.8

0.6

0.4

0.2
Figure 3. Overall survival stratified

p = 0.677
0.0
0

50
100
150
Months since nephrectomy

The British Journal of Radiology, September 2009

200

according to status of pre-operative


percutaneous renal artery embolisation (PRAE) in 378 patients with
renal cell carcinoma (matched
groups).

727

M May, S Brookman-Amissah, S Pflanz et al


Table 2. Cancer-specific survival and overall survival in 378 matched patients treated with surgery for renal cell carcinoma:
KaplanMeier method
Feature

Age
#60 years
.60 years
Gender
Male
Female
pT stage
pT1a
pT1b
pT2
pT3a
pT3b/pT4
pN stage
pN0/pNx
pN+
cM stage
M0
M1
Grading
G1
G2
G3
Histology
Clear cell carcinoma
Papillary carcinoma
Chromophobe carcinoma
Spindle cell carcinoma (pleomorph)
Clinical tumour size (cut-off: 5 cm)
#5 cm
.5 cm
Microvascular invasion
Absent
Present
Status of PRAE
PRAE (Group 1)
Nephrectomy alone (Group 2)

n (%)

172 (45.5%)
206 (54.5%)
227 (60.1%)
161 (39.9%)
51 (13.5%)
154 (40.7%)
65 (17.2%)
48 (12.7%)
60 (15.9%)
364 (96.3%)
13 (3.7%)
348 (92.1%)
30 (7.9%)
20 (5.3%)
309 (81.7%)
49 (13.0%)
297 (78.6%)
48 (12.7%)
25 (6.6%)
8 (2.1%)
108 (28.6%)
270 (71.4%)
324 (85.7%)
54 (14.3%)
189 (50.0%)
189 (50.0%)

CSS (1 year, 5 years, 8 years)

OS (1 year, 5 years, 8 years)

p50.854
94.1%, 82.9%, 75.1%
93.5%, 78.8%, 72.5%
p50.242
92.8%, 78.5%, 72.4%
95.3%, 84.3%, 75.8%
p,0.001
100%, 93.3%, 85.5%
98.7%, 88.2%, 84.2%
92.3%, 81.4%, 78.8%
80.2%, 67.3%, 61.0%
87.5%, 58.2%, 37.4%
p,0.001
95.0%, 83.1%, 76.6%
61.5%, 23.1%, 07.7%
p,0.001
96.2%, 84.9%, 77.4%
63.0%, 27.3%, 27.3%
p,0.001
95.0%, 90.0%, 83.1%
97.4%, 84.9%, 77.0%
70.0%, 46.3%, 46.3%
p50.740
94.9%, 82.8%, 74.3%
88.9%, 74.5%, 69.5%
91.7%, 72.6%, 72.6%
87.5%, 72.9%, 72.9%
p,0.001
99.1%, 90.8%, 89.2%
91.6%, 76.6%, 67.1%
p,0.001
94.7%, 83.5%, 78.7%
88.0%, 63.6%, 42.0%
p50.085
94.6%, 79.1%, 70.4%
93.0%, 82.8%, 79.0%

p50.084
92.4%, 78.7%, 69.9%
90.8%, 69.7%, 58.5%
p50.129
90.3%, 69.9%, 61.8%
93.4%, 80.0%, 67.1%
p,0.001
100%, 89.6%, 77.2%
97.4%, 81.9%, 75.6%
92.3%, 76.3%, 69.6%
75.0%, 57.0%, 48.8%
81.7%, 50.4%, 28.3%
p,0.001
92.9%, 76.1%, 66.5%
57.1%, 21.4%, 07.1%
p,0.001
94.5%, 78.3%, 67.7%
56.7%, 19.6%, 19.6%
p,0.001
95.0%, 90.0%, 83.1%
95.5%, 78.1%, 67.5%
65.3%, 53.3%, 31.5%
p50.672
93.3%, 75.9%, 64.0%
83.3%, 66.1%, 59.1%
88.0%, 66.0%, 66.0%
87.5%, 72.9%, 72.9%
p,0.001
99.1%, 87.9%, 81.5%
88.5%, 68.3%, 57.0%
p,0.001
93.2%, 77.2%, 69.5%
81.5%, 54.3%, 31.6%
p50.677
92.6%, 73.4%, 62.3%
90.5%, 74.7%, 66.6%

PRAE, percutaneous renal artery embolisation; yr, year.

Concerning the therapeutic effects of pre-operative


PRAE, the general opinion has remained deprecating
during the 35 years following the introduction of the
method. All studies that compared the survival of patients
undergoing pre-operative PRAE with the survival of
patients solely undergoing nephrectomy are retrospective
in design (Table 6) [7, 8, 1821]. Given the methodical
weakness of these studies, including varying patient
selection and missing or incorrect matching of the groups,
it is difficult to draw reliable conclusions (Table 6).
With regard to the methodical aspect, Zielinski et al [8]
presented the most conclusive study. In this retrospective study, patients who underwent PRAE before
nephrectomy (n5118) were compared with a casematched control group of patients undergoing nephrectomy alone (n5116) and a significant survival benefit in
the PRAE group was demonstrated [8]. However, this
survival benefit applied only to patients with pT2 and
pT3 disease and to patients with pT3N+ at the time of
surgery. Unfortunately, this study did not include
patients with pT1 tumours. The results of this selected
study group are similar to the data published by other
investigators [1719]. Some investigators have suggested
728

that the better prognosis in pre-operative PRAE patients


might be associated with a stimulation of certain
immune responses after embolisation [6, 8, 12, 13]. The
immunological benefits for patients undergoing preoperative PRAE might include augmentation of natural
killer cell activity and lymphoproliferative responses
[1214]. If these immunological phenomena were actually inducible by PRAE, as well as relevant for prognosis
of patients, further indications for selective embolisation
in nephron-sparing surgery could be defined other than
total pre-operative infarction of the renal artery.
Conversely, no improvement in survival was noted in
a study by Latal et al [21] of 84 pT13N0M0 patients with
RCC, including 39 patients treated with pre-operative
PRAE. However, owing to deficiencies in matching of
the study groups, the scientific basis of these results
seems quite weak [21].
Propensity score matching was used in our study as a
method to reduce the bias in the estimation of treatment
effects in the observational data set. With the use of this
method, baseline characteristics were summarised as a
single score that approximated the background characteristics for the individual patient. Table 1 and
The British Journal of Radiology, September 2009

Value of pre-operative renal artery embolisation in RCC


Table 3. Cancer-specific survival in the total group of matched patients with renal cell carcinoma (n5378): univariate and
multivariate Cox regression analysis
Variable

Univariate analysis
Age (#60 years vs .60 years)
Gender (male vs female)
pT stage (5 grades)
pN stage (pN0/pNx vs pN+)
cM stage (M0 vs M1)
Grading (3 grades)
Tumour size (#5 cm vs .5 cm)
MVI (absent vs present)
PRAE (Group 1 vs Group 2)
Multivariate analysis
pT stage (5 grades)
pN stage (pN0/pNx vs pN+)
cM stage (M0 vs M1)
Grading (3 grades)
Tumour size (#5 cm vs .5 cm)
MVI (absent vs present)

Hazard ratio

95% CI

p-Value

1.04
0.77
1.69
8.30
9.60
3.22
3.63
3.09
0.68

0.691.57
0.501.19
1.441.97
4.4915.34
5.6416.34
2.005.16
1.936.82
1.954.91
0.441.06

0.854
0.244
,0.001
,0.001
,0.001
,0.001
,0.001
,0.001
0.087

1.56
3.25
9.56
1.72
2.29
0.89

1.221.99
1.726.15
5.3117.24
1.042.84
1.194.41
0.451.76

,0.001
,0.001
,0.001
0.033
0.013
0.739

CI, confidence interval; PRAE, percutaneous renal artery embolisation.

Figure 1 demonstrate matching on the propensity score,


resulting in two treatment groups with balanced covariates. Thus, it was possible to compare both treatment
groups with regard to survival, similar to an unbiased
comparison achieved in a randomised controlled trial
[15].
Given the lack of data resulting from randomised
prospective trials, the present study represents the most
comprehensive and, by using a propensity score matching with nine relevant variables, also the most reasonable
study on this issue. Analysing the present results, we
conclude that pre-operatively performed PRAE does not
result in an improvement in post-operative survival
rates. There is even a trend towards a reduction in CSS in
patients undergoing preoperative PRAE (p50.085). In
order to report all post-operative complications, we
evaluated the complete study cohort (n5834) (Table 5).
Other than the number of post-operative blood transfu-

sions which were dispensed, no significant differences


between the groups could be shown. Interestingly, the
number of blood transfusions also differed significantly
in the matched study groups (58.5% vs 27.3%, p,0.01).
Thus, our data are not in concordance with other studies
that have described a reduction of the post-operative
transfusion rate in patients with pre-operative PRAE [9,
10].
The present study may include potential limitations.
Firstly, data were derived neither from a single centre
nor from a randomised clinical trial. However, propensity matching techniques were applied to make the study
groups comparable. Secondly, comparing a historical
group of patients (pre-operative PRAE) with current
patients (nephrectomy alone) resulted in a varying
median follow-up time (100 months vs 62 months,
p,0.01). All patients undergoing PRAE were operated
on between 1992 and 1997, whereas the surgical group

Table 4. Overall survival in the total group of matched patients with renal cell carcinoma (n5378): univariate and multivariate
Cox regression analysis
Variable

Univariate analysis
Age (#60 years vs .60 years)
Gender (male vs female)
pT stage (5 grades)
pN stage (pN0/pNx vs pN+)
cM stage (M0 vs M1)
Grading (3 grades)
Tumour size (#5 vs .5 cm)
MVI (absent vs present)
PRAE (Group 1 vs Group 2)
Multivariate analysis
pT stage (5 grades)
pN stage (pN0/pNx vs pN+)
cM stage (M0 vs M1)
Grading (3 grades)
Tumour size (#5 vs .5 cm)
MVI (absent vs present)

Hazard ratio

95% CI

p-value

1.35
0.76
1.57
5.75
7.29
3.12
2.63
2.72
0.93

0.961.89
0.531.09
1.391.78
3.2310.25
4.5611.67
2.114.61
1.674.15
1.844.01
0.661.32

0.086
0.131
,0.001
,0.001
,0.001
,0.001
,0.001
,0.001
0.678

1.44
2.40
6.77
1.87
1.74
0.94

1.181.76
1.324.36
4.0611.29
1.242.82
1.082.81
0.541.66

,0.001
0.004
,0.001
0.003
0.023
0.836

CI, confidence interval; MVI, microvascular invasion; PRAE, percutaneous renal artery embolisation.

The British Journal of Radiology, September 2009

729

M May, S Brookman-Amissah, S Pflanz et al


Table 5. Post-operative complications (30 days, except for renal dialysis) and intra-/post-operative blood transfusions after
nephrectomy with (n5227) and without (n5607) embolisation in 834 patients with renal cell carcinoma; post-embolisation
syndromes for 227 patients with pre-operative PRAE
Event

Study cohort (n5834)


(100%)

Mortality
8 (1.0%)
Wound infection
72 (8.6%)
Retroperitoneal abscess
10 (1.2%)
Cardiopulmonary event
77 (9.2%)
Secondary haemorrhage (re-operation)
24 (2.9%)
Blood transfusion
282 (33.8%)
Blood transfusion .2 units
120 (14.4%)
Thrombosis
8 (1.0%)
Cerebral apoplexy
5 (0.6%)
Ileus (conservative management)
6 (0.7%)
Ileus (re-operation)
0
Renal dialysis (full follow-up)
48 (5.8%)
Acute toxicity of PRAE (post-embolisation syndrome)
Lumbar pain
Hypertension
Fever
Haematuria (macroscopic)
Vomiting/nausea
Small groin haematoma

PRAE (n5227) (100%)

Nephrectomy alone
(n5607) (100%)

p-Value

3 (1.3%)
21 (9.3%)
3 (1.3%)
24 (10.6%)
8 (3.5%)
138 (60.8%)
63 (27.8%)
3 (1.3%)
1 (0.4%)
3 (1.3%)
0
11 (4.8%)

5 (0.8%)
51 (8.4%)
7 (1.2%)
53 (8.7%)
16 (2.6%)
144 (23.7%)
57 (9.4%)
5 (0.8%)
4 (0.7%)
3 (0.5%)
0
37 (6.1%)

0.46
0.68
1.00
0.42
0.49
,0.01
,0.01
0.45
1.00
0.35
1.00
0.62

189 (83.3%)
58 (25.6%)
112 (49.3%)
7 (3.1%)
92 (40.5%)
6 (2.6%)

PRAE, percutaneous renal artery embolisation.

included patients treated from 1992 to 2006. Thirdly, we


have to consider that there have been slight modifications both in surgical techniques, including an increased
use of nephron-sparing surgery, and in medical management during the period of investigation.
The results of the present study indicate that preoperative PRAE does not improve the survival of
patients after surgery and thus does not lead to a better
prognosis of patients with RCC. Further research, in
particular prospective randomised clinical trials, is
necessary to provide evidence for the pros and cons of
this procedure reliably.

References
1. Espey DK, Wu XC, Swan J, Wiggins C, Jim MA, Ward E, et
al. Annual report to the nation on the status of cancer, 19752004, featuring cancer in American Indians and Alaska
Natives. Cancer 2007;110:211952.
2. Janzen NK, Kim HL, Figlin RA, Belldegrun AS. Surveillance
after radical or partial nephrectomy for localized renal cell
carcinoma and management of recurrent disease. Urol Clin
North Am 2003;30:84352.
3. Wallen EM, Pruthi RS, Joyce GF, Wise M. Urologic diseases
in america project. kidney cancer. J Urol. 2007;177:200618;
discussion 201819.

Table 6. Studies on differences in survival between PRAE and radical nephrectomy alone in patients with renal cell carcinoma
Author (year)

Number of patients: preoperative PRAE vs


nephrectomy alone

Bono and Caresano (1983) 81 vs 40


[18]
Christensen et al (1985) [7]
32 vs 29

Differences in survival between the two groups


(significant findings)

Study problems

Significant survival benefit for PRAE


(particularly with regard to pT12)
No differences in survival, blood loss and
operative time between both groups
Decrease in tumour volume and survival
benefit for PRAE in patients with M1 or
post-operative metastases
Significant survival benefit for PRAE
(particularly with regard to pT2)

No matching

Masuda et al (1985) [19]

42 vs 42

Stosslein et al (1987) [20]

92 vs 66

Latal et al (1990) [21]

45 vs 39

Trend to improved survival for PRAE

Zielinski et al (2000) [8]

118 vs 116

Significant survival benefit for PRAE over


all stages

Own data

189 vs 189

No differences in survival between both


groups

Matching not
sufficiently
No matching

realised

Retrospective study of
five centres without
matching
Matching not realised
sufficiently
No pT1 tumours included
in analysis
No details to the time of
follow-up
Comparison of a historical and a current study
group

PRAE, percutaneous renal artery embolisation.

730

The British Journal of Radiology, September 2009

Value of pre-operative renal artery embolisation in RCC


4. Almgard LE, Fernstrom I, Haverling M, Ljungqvist A.
Treatment of renal adenocarcinoma by embolic occlusion of
the renal circulation. Br J Urol 1973;45:4749.
5. Turini D, Nicita G, Fiorelli C, Selli C, Villari N. Selective
transcatheter arterial embolization of renal carcinoma: an
original technique. J Urol 1976;116:41921.
6. Kaisary AV, Williams G, Riddle PR. The role of preoperative
embolization in renal cell carcinoma. J Urol 1984;131:6416.
7. Christensen K, Dyreborg U, Andersen JF, Nissen HM. The
value of transvascular embolization in the treatment of
renal carcinoma. J Urol 1985;133:1913.
8. Zielinski H, Szmigielski S, Petrovich Z. Comparison of
preoperative embolization followed by radical nephrectomy with radical nephrectomy alone for renal cell
carcinoma. Am J Clin Oncol 2000;23:612.
9. Bakal CW, Cynamon J, Lakritz PS, Sprayregen S. Value of
preoperative renal artery embolization in reducing blood
transfusion requirements during nephrectomy for renal cell
carcinoma. J Vasc Interv Radiol 1993;4:72731.
10. Singsaas MW, Chopp RT, Mendez R. Preoperative renal
embolization as adjunct to radical nephrectomy. Urology
1979;14:14.
11. Lopatkin NA, Mazo EB, Nabiev YN. Nephrectomy with
prior intraluminal balloon occlusion of the renal artery in
kidney cancer. Eur Urol 1979;5:2559.
12. Bakke A, Gothlin JH, Haukaas SA, Kalland T. Augmentation of natural killer cell activity after arterial embolization of renal carcinomas. Cancer Res 1982;42:38803.

The British Journal of Radiology, September 2009

13. Nakano H, Nihira H, Toge T. Treatment of renal cancer


patients by transcatheter embolization and its effects on
lymphocyte proliferative responses. J Urol 1983;130:247.
14. Schwartz MJ, Smith EB, Trost DW, Vaughan ED Jr. Renal
artery embolization: clinical indications and experience
from over 100 cases. Br J Urol Int 2007;99:8816.
15. Haviland A, Nagin DS, Rosenbaum PR. Combining propensity score matching and group-based trajectory analysis in an
observational study. Psychol Methods 2007;12:24767.
16. Kalman D, Varenhorst E. The role of arterial embolization
in renal cell carcinoma. Scand J Urol Nephrol 1999;33:
16270.
17. Maxwell NJ, Saleem Amer N, Rogers E, Kiely D, Sweeney P,
Brady AP. Renal artery embolisation in the palliative treatment of renal carcinoma. Br J Radiol 2007;80:
96102.
18. Bono AV, Caresano A. The role of embolization in the
treatment of kidney carcinoma. Eur Urol 1983;9:3347.
19. Masuda F, Nakada J, Takasaka S, Hishinuma H, Machida T.
The role of preoperative embolization in renal cell
carcinoma. Gan No Rinsho 1985;31:28992.
20. Stosslein F, Schwenke A, Munster W, Kohler H, Mebel M,
Raatzsch H, et al. Improvement in the prognosis of renal
cell carcinoma by percutaneous transvascular embolization.
Z Urol Nephrol 1987;80:57786.
21. Latal D, Kautzky W, Wanek R, Kumpan W. Effect of kidney
occlusion on survival in renal cell carcinoma: a 10-year
retrospective study. Z Urol Nephrol 1990;83:2731.

731