Andrea Richard
L Desberg, MD #{149} David B. Troy, MD #{149} Andrew
Renal with
M. Paushter, MD #{149} Gary C. Novick, MD #{149} Joseph
Index
terms:
Kidney,
Renal Renal
arteries, arteries,
stenosis or obstruction, US studies, 96.12984 1990;
MD MS
#{149}
Artery Stenosis: Evaluation Color Doppler Flow Imaging’
Renal artery stenosis (RAS) is the most common correctable cause of hypertension. The current study was undertaken to evaluate the usefulness of color Doppler flow imaging as a screening examination in the detection of significant RAS. Fifty-five kidneys in 30 patients were examined with aortography and color Doppler flow imaging in a double-blind fashion. The peak systolic velocity (PSV) in the renal artery, the renal-aortic ratio (RAR) (ie, the ratio of the PSV in the renal artery to the PSV in the aorta), and the renal artery resistive index were determined and compared with the percentage of stenosis as determined with angiography. Ultrasound (US) criteria used to diagnose RAS were (a) an RAR of 3.5 or greater and/or (b) a renal artery NV of greater than 100 cm/sec. Doppler tracings were obtained in 25 (69%) of 36 kidneys with a patent single renal artery. RAR and PSV each yielded a sensitivity of 0% in the diagnosis of RAS. Doppler tracings were obtained in three (50%) of six occluded renal arteries. Accessory arteries were present in 13 (24%) of 55 kidneys, but none were visualized with color Doppler flow imaging. The authors conclude that with current technical capability, color duplex US is not an adequate screening method for the detection of RAS.
Radiology
K. Lammert, MD #{149} Jonathan C. Hale, V. Nally, Jr, MD #{149} Ann M. Weltevreden,
US studies,
177:749-753
81.1298 96.72.
#{149}
H
is present
YPERTENSION
timated
58 million
in an es-
Americans
(1). In the majority of patients, the cause is unknown; however, a secondary cause is present in approximately of these
artery in less
10%
of patients. Less than half are curable (2). Renal stenosis (RAS), while present than 2% of adults with hyper-
cases
tension (3), remains the single most common correctable cause (2). Currently, an adequate noninvasive screening examination for the detection of significant RAS in the hypertensive population is not available, and angiography remains the standard (4). Previous authors, employing various parameters, have suggested the utility of duplex sonography in the detection of significant RAS (5-12). A peak systolic velocity (PSV) in the menal artery greater than 100 cm/sec has been used as a diagnostic criterion for the presence of RAS (6). Two recent articles have compared the menal-aortic
ratio
(RAR)
(ie,
the
ratio
of
the PSV in the renal artery to the PSv in the aorta) with the percentage of stenosis as determined with angiography.
tides
The
authors
of these
two
am-
demonstrated between
PATIENTS
an
I From the Departments of Radiology (A.L.D., D.M.P., C.K.L., J.C.H.), Urologic Surgery (R.B.T., A.C.N.), Hypertension-Nephrology (J.V.N.), and Biostatistics (A.M.W.), Cleveland Clinic Foundation, 1 Clinic Center Dr, Cleveland, OH 44195-5103. From the 1989 RSNA scientific assembly. Received April 3, 1990; revision requested May 14; revision received and accepted July 20. Address reprint requests to D.M.P. 0 RSNA, 1990
AND
METHODS
Between December 20, tember 14, 1989, 55 kidneys
1988,
undergoing
were
aortography
and
Sep-
in 30 patients examined
with color Doppler flow imaging. The 14 female and 16 male patients ranged in age from 19 to 75 years (mean, 50.3 years). The
patient
population
included
(a) po-
tential renal donors (26 kidneys in 13 patients), (b) patients with suspected renovascular hypertension (25 kidneys in 13 patients), and (c) patients with a known renal mass in whom the contralateral kidney was studied (four kidneys in four patients). The only criterion for patient inclusion in the study was that an aomtogram had been obtained within 10 weeks of the US examination. All US examinations were performed with an Acuson 128 US unit 3.5-MHz
a positive correBAR of 3.5 or greater and RAS of 60% or greater (8,12). A third, more recent study evaluated the use of color duplex ultrasound (US) as a possible screening tool, comparing both BAR and renal artery PSV with the percentage of stenosis as determined with angiography. The sensitivity of RAR and PSV could not be assessed, since the US study was unsuccessful in all of the significantly stenotic arteries. lation
With the utilization of PSV as the major criterion for the diagnosis of RAS, a 31% false-positive rate was obtamed (13). We used color duplex US to determine whether this modality may currently play a role in the detection and grading of RAS.
(Mountain (imaging
View, Calif) and Doppler)
with a phased-
array transducer. Patients were either fasting or had received a clear-liquid diet since
midnight
on the
night
before
the
examination. All studies were performed by one of several US technologists, all cxperienced in the use of Doppler techniques.
Each
graphic
findings
amination.
was
blinded
to the angio-
at the time
In several
of the US cx-
instances
of
technically difficult studies, a radiologist (either of the two who intempreted the studies)
assisted
in performing
the exami-
nation. This occurred in only a limited number of patients and was not recorded. The following measurements were obtamed, when technically achievable: PSV in the midabdominal aorta, PSV and enddiastolic velocity (EDV) in both the pmoximal and distal portions of the renal am-
Abbreviations: RAS = renal artery PSv = peak systolic velocity, RAR tic ratio, EDV end-diastolic velocity, sistive
stenosis, renal-aor-
RI
re-
index.
749
tery,
renal
(RI) (RI mum
parenchymal resistive index [PSV - EDV]/PSV), and maxi-
=
renal
length.
The
velocities
aorta and the renal artery mined from their respective ings.
The
the
proximal
were
proximal
and
60#{176} was
used
Patients supine posterior
and
artery
RIs
were
necessary.
typically
portions
in the
decubitus
or
was
when
used
artery
left
and
distal
usually
neces-
tion. in
US examination
the
following
was
manner:
od, no time
restraint
technologist
for
time If the
tery was not satisfactorily within 15-20 minutes, artery was terminated; nal artery was visualized
study
was
imal
considered
or distal
adequate
artery
from
ar-
on the renal
of the US those of the and the US
successful
renal
was obtained
renal
was placed
obtaining
Similarly,
a single the
experienced
renal
US
in US imaging
techniques.
Neither
had
of the angiographic
and Doppler any
results
the US interpretation. The used to diagnose significant based on previous studies: 3.5
or greater
was
knowledge
at the time major
of
criteria
RAS were (a) an BAR
considered
of
to indicate
stenosis
of 60% or greater and/or (b) a mePSV greater than 100 cm/sec was considered to indicate stenosis of 50% nal
artery
or greater. A diagnosis made if the renal artery
ized
but no Doppler
of occlusion was well
signal
could
was visual-
be ob-
tamed.
Thirty-six intraarterial phy;
19,
kidneys digital with
aortography.
were studied with subtraction aortogma-
conventional
of eight
with
diatrizoate meglumine dium (Renografin-60; Brunswick, NJ) (11.7 at a rate of 20 mL/sec 30-40 mL of two-thirds
zoate (5.9-7.8
750
meglummne g of iodine)
#{149} Radiology
gion that
Neither
1(5) 1(20)
7(32) 1(20)
8(36) 3(60)
6(27) 0(0)
of tracings,
.
percentages
was
adjacent
normal
renal
of the
diameter
mated. graphic
diameter
stenosis
the
One
no knowledge the aortoin
artery,
and
of the
with
or distal
the
percentage
stenosis
was
In all cases, the correlative and color Doppler flow were
obtained
me-
within
esti-
angioexamina-
a period
of 9
weeks. Each kidney was placed into one of the following categories of percentage of stenosis, according to the angiographic results:
normal,
71%-99%,
50%
and
or
less,
51%-70%,
..
.
...
2(22) 1(17)
Among the 30 patients involved in the study, one had a solitary kidney, and four had a unilateral renal mass, leaving 25 pain
whom
both
Among had
the one
kidneys
latter
were
group,
kidney
that
two renal arteries, and two had dual renal arteries to both These kidneys were subsequently ed from statistical analysis for described in the “US Results” Results of intraclass correlations and
percentage 15 patients
both renal arteries could dicated that a relationship the left and right kidneys patient. Thus, ity determinations
for
pa-
supplied
by
US parameters in the remaining
exam-
eight
was
patients kidneys. excludthe reasons section. for the of stenosis whom
in
be examined inexists between within a single
sensitivity and and statistical
ses, one kidney was randomly from each of these 15 patients.
specificanaly-
selected The data
values for each of the parameters within each stenosis group were not normally distributed,
so the
kal-Wallis ences
indexes
nonparametric
test was used
among
with
tiple comparisons termine which were contributing cance.
stenosis
groups.
significant
Krus-
to detect For
differences,
differthose
mul-
were performed to depairs of stenosis groups to the overall signifi-
RESULTS
and diatrizoate soSquibb, New g of iodine) injected (conventional) or strength diatri-
Angiography demonstrated that 42 kidneys were supplied by a single renal artery. These were categorized by percentage of stenosis as follows: normal (n 22), 50% or less (n = 5), 51%-70% (n 0), 71%-99% (n 9),
sodium
at a rate
of
and
occluded
neys
were
5(56) 3(50)
Results
(n
6).
supplied
Although evaluated
was
Thirteen
by two
each separately
renal
renal for
kidam-
artery percent-
age of stenosis as determined with angiography, these data are not included, since no accessory renal artery was visualized with US.
US
Results
The Doppler
average time for the color flow examination was 25 per kidney, including both
minutes
occluded.
Angiographic
injected
.
in parentheses.
teries.
the
compared
proximal
of
the use of 40 mL of
and diatrizoate
.
1(11) 0(0)
with
subtraction).
who had interpreted
luminal
of the in
kid-
neys obtained at outside institutions were reviewed. At our institution, aortography
was performed
Artery Both
screen-film
Angiograms
of Renal
Artery
Distal
...
(digital
The
tients
tracing
patent
grams.
med.
if a prox-
Doppler
mL/sec
tients
examination was considered successful if no tracing was obtained from one occluded renal artery or if tracings were obtained from each menal artery in a kidney with multiple renal arteries. All results of US studies were meviewed by one of two radiologists, both
artery.
Renal
Proximal
1(11) 2(33)
are numbers
angiographer, the US results,
tions
limited
visualized the study of that however, if the mewithin this peri-
artery tracings. The findings study were compared with angiographic examination,
a Single
re-
sitated two separate views. Data from the proximal renal artery were typically obtamed from a transverse image in which the aorta and the origin of the renal artery were identified. Sampling of the distal renal artery was performed at the renal hilum, proximal to the first bifurcaThe
with
was the
proximal
renal
of Stenosis
Note-Values
as
15-20
of the
of the
data.
positioning
helpful imaged.
Visualization
than
scanned
however, positioning
Nonsupine
more frequently nal artery was
these
of US in Kidneys
0(n22) 50(n5) 51-70(n0) 71-99(n9) 100(n6)
technicalof less
in obtaining
position; oblique
Percentage
RARs
renal angle
Rate
Portion
distal
Whenever
Success
trac-
and
a Doppler
1
Technical
the
were deterDoppler
distal
also calculated.
ly possible,
in
Table
successful
and
tions.
The
nation
in
unsuccessful
examina-
range of times for patients in whom both kidneys were exammed was 22-97 minutes. One patient had a solitary kidney, and the examination time was 15 minutes (unsuccessful study). If all patients with Unsuccessful bilateral examinations are excluded, the average examination time per kidney was approximately 25 minutes. Table 1 shows the technical success rate of the color Doppler flow examithose
kidneys
with
a single
renal artery. A proximal and/or distal tracing was obtained from the menal artery in 25 (69%) of 36 kidneys with a patent single renal artery. In 11
(31%)
tery
of
36
kidneys,
the
renal
am-
was
not sufficiently visualized obtain a Doppler tracing proximally or distally within the vessel. These latter
examinations
were
to
considered
technically unsuccessful. The success rate did not differ significantly when results in the right and left sides were compared (69% vs 70%). As the percentage of stenosis increased, our ability to obtain Doppler tracings diminished (78% for the combined 0%-
50% stenosis 7l%-99% tion,
(50%)
groups stenosis
tracings
were
of six occluded of the latter
In two proximal tracing third, a tracing
mally
and
distally.
vs 44% group). obtained
for the In addiin three
renal arteries,
arteries. only a
was obtained; was obtained
Of these
one was typical of that renal artery, while the
in the proxi-
tracings,
of a normal remainder December
1990
S
were
atypical
S
stolic
flow
C’)
In 1
0’
S
II
ft
u
H .5. UH
N. t
C)
5
‘a
.-
c’
-
,-,
_
0’
I
0
-
II
II
I
5)
a’ c’
C’)
‘0
c’
m
‘0
8
C?
In
In
N.
.
C’)
-
‘0 C)
II
II
H
II
.-..
5) (‘I C?
00
m
C’) 0
4 In
m
-
‘0
I
F,
I
‘0
9
C’4
C’)
I
I
I
8
‘0
0’
0’
)< CS
0C’)In
N
0
;
-
0’ ‘0 0’
C’)
ble
N.
-
a)
U S a)
the
a)
II
In
0
0
C’)
H
.-.
.
In
‘0
N H
0’
1’
t,.
0.! CS 5)!
U
‘0
In
U
I,..
8
0
0 I
I
C’)
t% ‘
0’
‘0
‘0
I
so 5)
0’
CS
U
‘
II
II
‘0!
C’)
N.
5) a)
C’)
‘
In
OH
U)
II
H
CS
E x
0
N
0
‘
0.
0’
2
1A00 ‘0 00
0.
CS
I
In ‘0
C’) C’)
0
0’ -
00
0
0.
I 00 CO In
0
‘-
.-
0’
CS
4:
c’1
In N. -
CS
0’ ‘0 I
00
.5C
II
C’)
N
CS
U)
00
C’)
II
0
5)
t .5’
-
0
.5’
‘0
‘0 N.
C’)
N
II
II
II
‘0
N.
‘0
I ‘-4
-
‘0
‘0
0 ‘0
C’)
N.
In ‘0
In
C’)
C)
4:
II 0.
z
0’
0’
-
‘5 In
I N.
.5)
0
‘0
a) t
5;
I
.0
CS
0’
S 5)
C’)
II
II
CS
II
E
.5,
x
N.
2
0’ 00
I N. .5’
0.
0’ In
z
0 N. I
I
0
00
‘0
In
.5’ N.
N.
S a)
0. 0.
CS
II 0.
CS 0 S
CS
8
4:
N
z
‘0
.8
.c
sample
to be marginally significant (P .071) and insignifi(P = .123). However, sizes in the various stenosis
the
groups were small, resulting in low statistical power for detecting relatively small differences in parameters. Table renchymal
percentage ly selected statistical
Volume
pro-
US
proximally cant distally
II
0.
I
-
H
‘0
4:
C’)
In
made
of the occluded
statistical analyses. Table 3 relates RAR,
99% stenosis group compared with the
I It)
0
0.
a diag-
not
parameters for the randomly group of kidneys utilized
ed a statistically
.5’
In -
arteries
the significant stenosis group (0% sensitivity, 100% specificity). In addition, comparison of the median RAR values proximally and distally in those arteries in the normal, 50%-stenosis-or-less, and 7i%-99% stenosis groups did not reveal an increase in the RAR with increasing stenosis. In fact, the median proximal RAR in the 71%-99% stenosis group demonstrat-
CS U
In -
the
of kidneys. were below
8
U
was
by a single
nosis
S
z
In
diaNone
occluded
on the basis in any of the
RI, and
C’)
0
low
Consequently,
3 delineates
same lected
S
II
three
occlusion
supplied
0
In
CS 5)
5) 0 5) S 0 U
0
C’)
I 0’ In
Dt
very
demonstrated.
sels. Table 2 delineates the descriptive data relating to all measured US parametems for all 42 kidneys that were
0
In
of
spectively findings
x a)
.5
other
visualized.
nosis
VI 0.
8U
r
I0
S 5) ‘5
S In
CS
I
U
was
0’
‘0
In
E
of the
in that was
3 also relates the median RI and renal length
pato the
of stenosis in the randomgroup of kidneys used for analysis.
No
specific Radiolov
trend
751
#{149}
in median RI was noted with increasing stenosis, and no significant difference among groups was found (P = .267). The median length of those kidneys with occluded renal amteries was decreased, as expected; however, this difference was not statistically significant (P .152). Thirteen kidneys were supplied by two renal arteries. No accessory arteries were imaged with color Doppler flow examination. Visualization of a single normal renal artery did not cxdude the presence of a nonvisualized, possibly stenotic accessory artery. These US examinations were therefore considered inadequate, yielding a success rate for these 13 kidneys
this
of 0%.
group
Further
was
not
analysis
of
performed.
DISCUSSION Hypertension remains one of the most important risk factors in the development of stroke and coronary artery disease (1). Although renovasculam disease is the cause of hypertension in a minority of patients (3), it is the most common correctable cause (2). Therefore, it remains an important goal to find an accurate, noninvasive
screening
detection
examination
of significant
hypertensive
for
RAS
the
in the
population.
Various authors have evaluated the utility of duplex sonography in the detection of significant RAS. Attempts to use PSV as a major criterion have met with variable success (6,13). Kohier et al (8) and Taylor et al (12) have measured the RAR, have cornpared this value to the percentage of stenosis as determined with angiography, and have shown a positive correlation
between
an
RAR
of 3.5
or
greater and RAS of 60% or greater. High sensitivity and specificity were obtained in both studies (91% and 95%, respectively, in the former study and 84% and 97%, respectively, in the latter). In addition, both groups of investigators achieved a high percentage of technically adequate examinations (90% and 87.6%, respectively), similar to the results in other reported series (5-7, 9-11). We hoped to reproduce these data in our study, which included the added capability anticipated
of color that
flow imaging. We the added dimension
of color would facilitate visualization of the renal arteries and would allow more accurate Doppler angle correction and therefore more accurate yelocity determination. However, even with color flow imaging, we were unable to successfully reproduce the 752
Radiology
#{149}
findings of these previous studies. Our results more closely parallel those of Berland et al (13), who achieved a success rate of only 58% in identifying
the
main
renal
artery
with color duplex US. They were unable to evaluate the sensitivity of RAR or PSV as a diagnostic criterion, since all US examinations in the group of arteries with significant stenosis were technically inadequate. However, with use of renal artery PSv, they obtained a specificity of only 37%. In our study, technically adequate examinations
were
achieved
in only
51% of kidneys (69% of kidneys with a patent single renal artery, 0% of kidneys with two renal arteries). We believe that two main factors contributed to the relatively low number of technically successful studies. The US examinations were performed in a time-limited fashion to examine the true feasibility of this modality as an effective screening tool in a busy US department. Most of the previous studies cited either required a greater amount of time to perform the examination or did not note the time mequired to obtain a satisfactory study (5-8,10,12). Berland et al (13) employed a time-restricted US examination, as we did, and achieved a cornparably low percentage of successful studies. The second factor concerns those kidneys supplied by more than one renal artery. In our study, 24% of the kidneys fell into this category. This number correlates well with reports in the literature, which state that 20%-28% of kidneys have multiple renal arterial supply (14). Because of the possibility of discrepant stenoses among the multiple renal arteries, visualization of a single normal artery does not exclude the possibility of a significantly stenotic accessory artery. Therefore, if all arteries to a kidney are not visualized during the color Doppler flow examination, it should be deemed inadequate as a screening study for that particular kidney. In our study, all of the color Doppler flow examinations performed on kidneys supplied by two renal arteries were deemed technically unsuccessful on this basis. This group of kidneys accounted for a significant percentage of the total number in the study, which contributed greatly to the low success rate. Most previous investigators have not addressed the issue of multiple renal arteries (5-10,12), except for Berland et al (13), who found a similar percentage (2.2%) of kidneys with accessory
arteries and were also unable to visualize any accessory arteries with color Doppler US. Among the kidneys with adequate US examinations, no elevation of the median RAR with increasing stenosis was noted, as was reported in previous studies. In fact, for proximal RAR, a statistically significant decrease in the median value was found in the 71%-99% stenosis group compared with the normal group and 50%-stenosis-or-less group, an obvious contradiction to the results of Kohiem et al (8) and Taylor et al (12); however, the reason for this discrepancy is unclear. Similarly, our results obtained with use of renal artery PSV are somewhat contradictory to those of Avasthi et al (6), who achieved a sensitivity of 42% and a specificity of 92% with use of this parameter to diagnose RAS. We obtained 0% sensitivity proximally and distally and only 50% specificity proximally and 79% specificity distally. Our results more closely correlate with those of Berland et al (13), who reported a specificity of 37%. We also calculated the renal artery RI both proximally and distally, when technically achievable. Theometically, we anticipated that a critical stenosis of the renal artery may be manifested by an increased renal amtery RI due to a decrease in the EDV proximal to the stenosis. Most stenoses in our study were orificial or within the 1st centimeter of the main renal artery (17 of 20), thereby making it difficult or impossible to obtain a renal artery tracing proximal to the stenosis. Only one of the remaining three renal arteries that demonstrated a more distal stenosis was severely stenotic, but the examination in this case was technically inadequate. Consequently, we were unable to assess
the
effect
of severe
stenosis
on
the renal artery RI obtained proximal to the area of narrowing. We did, however, still compare median values of renal artery RI proximally and distally to the percentage of stenosis. No
specific
trend
could
be
detected
with increasing stenosis, and no statistically significant differences were found. The final two parameters measured were the parenchymal RI and renal length. Parenchymal RI is an indicatom of small-vessel resistance in the kidney. Increased RI signifies increased resistance within the small intrarenal vessels and is often associated with significant parenchymal disease and poor function. Elevated RI and decreased renal length may be December
1990
secondary to a variety of causes and are nonspecific findings. Therefore, these parameters were evaluated for use as possible minor criteria to support a diagnosis of RAS in the face of marginally positive major criteria. As noted previously, no meaningful trend in median parenchyrnal RI was noted with increasing percentage of stenosis of the renal artery; therefore, this was not a helpful parameter, nor was renal length. The only detectable difference in median renal length was a decrease in the kidneys with occluded arteries; however, this did not prove to be statistically significant. In 50% of the occluded arteries, a Doppler tracing was obtained; other authors have reported the same prob1cm (7,8,1 1). These “false” tracings were probably obtained from a large collateral vessel or from a meconstituted segment of the main renal artery. Clearly, this problem may serve as a source of false-negative examination results in the clinical setting.
of the most recent studies (13) and our own study, both utilizing color Doppler flow Imaging, yielded mesults contradictory to those previously reported in the literature. These discrepancies indicate the need for continued prospective studies with larger numbers of patients to clarify the accuracy of the proposed objective diagnostic criteria. However, as evidenced by the low number of technically adequate examinations we achieved, duplex sonogmaphy of the renal artery is difficult to perform, especially in a time-limited setting. With current technical capability, color Doppler flow US is not an adequate screening examination for the detection of significant RAS. U References 1.
2.
CONCLUSION 3.
During the past 10 years, many studies have been performed to evaluate the mole of duplex US in the detection of RAS, and most have yielded promising results. However, one
6.
7.
8.
9.
59:225-229.
10.
Saunders,
1988;
4.
Treadway hypertension.
KK, Slater Annu
5.
Greene ER, Venters MD, Avasthi PS, Conn RL, Jahnke RW. Noninvasive characterization of renal artery blood flow. Kidney tnt 1981; 20:523-529.
Jenni
R, Vieli A, Luscher TF, Schneider E, W, Anliker M. Combined two-dimensional ultrasound Doppler technique: new possibilities for the screening of renovascular and parenchymatous hypertension? Nephron 1986; 44(suppl l):2-4. Robertson R, Murphy A, Dubbins PA. Renal artery stenosis: the use of duplex ultrasound as a screening technique. Br J Radiol 1988; 61:196-201. Taylor DC, Kettler MD, Moneta GL, et al.
Vetter
11.
12.
Frohlich ED. Evaluation and management of the patient with essential hypertension. In: Parmley WW, Chatterjee K, eds. Cardiology. Vol 2. Philadelphia: Lippincott, 1989; 1-15. Cressman MD, Gifford RW. Evaluation of secondary forms of hypertension. In: Parmley WW, Chatterjee K, eds. Cardiology. Vol 2. Philadelphia: Lippincott, 1988; 1-22. Kaplan NM. Systemic hypertension: mechanisms and diagnosis. In: Braunwald E, ed. Heart disease: a textbook of cardiovascular medicine. 3rd ed. Philadelphia:
Avasthi PS, Voyles WF, Greene ER. Noninvasive diagnosis of renal artery stenosis by echo-Doppler velocimetry. Kidney Int 1984; 25:824-829. Norris CS, Pfeiffer JS, Rittgers SE, Barnes RW. Noninvasive evaluation of renal artery stenosis and renovascular resistance. VascSurg 1984; 1:192-201. Kohler TR, Zierler RE, Martin RL, et al. Noninvasive diagnosis of renal artery stenosis by ultrasonic duplex scanning. Vasc Surg 1986; 4:450-456. Dubbins PA. Renal artery stenosis: duplex Doppler evaluation. Br J Radiol 1986;
Duplex nosis
13.
artery
scanning stenosis:
in the diaga prospec-
tive evaluation. J Vasc Surg 1988; 7:363369. Berland LL, Koslin DB, Routh WD, Keller
FS.
14.
ultrasound of renal
Renal
artery
stenosis:
prospective
evaluation of diagnosis with color duplex US compared with angiography. Radiology 1990; 174:421-423. Davis GD. Variations in renal blood supply. In: Kincaid OW, ed. Renal angiography. Chicago: Year Book Medical, 1966; 83-87.
819-861.
EE. Rev
Renovascular Med 1984;
35:665-692.
Volume
177
Number
#{149}
3
Radiology
753
#{149}
L Desberg, MD #{149} David B. Troy, MD #{149} Andrew
Renal with
M. Paushter, MD #{149} Gary C. Novick, MD #{149} Joseph
Index
terms:
Kidney,
Renal Renal
arteries, arteries,
stenosis or obstruction, US studies, 96.12984 1990;
MD MS
#{149}
Artery Stenosis: Evaluation Color Doppler Flow Imaging’
Renal artery stenosis (RAS) is the most common correctable cause of hypertension. The current study was undertaken to evaluate the usefulness of color Doppler flow imaging as a screening examination in the detection of significant RAS. Fifty-five kidneys in 30 patients were examined with aortography and color Doppler flow imaging in a double-blind fashion. The peak systolic velocity (PSV) in the renal artery, the renal-aortic ratio (RAR) (ie, the ratio of the PSV in the renal artery to the PSV in the aorta), and the renal artery resistive index were determined and compared with the percentage of stenosis as determined with angiography. Ultrasound (US) criteria used to diagnose RAS were (a) an RAR of 3.5 or greater and/or (b) a renal artery NV of greater than 100 cm/sec. Doppler tracings were obtained in 25 (69%) of 36 kidneys with a patent single renal artery. RAR and PSV each yielded a sensitivity of 0% in the diagnosis of RAS. Doppler tracings were obtained in three (50%) of six occluded renal arteries. Accessory arteries were present in 13 (24%) of 55 kidneys, but none were visualized with color Doppler flow imaging. The authors conclude that with current technical capability, color duplex US is not an adequate screening method for the detection of RAS.
Radiology
K. Lammert, MD #{149} Jonathan C. Hale, V. Nally, Jr, MD #{149} Ann M. Weltevreden,
US studies,
177:749-753
81.1298 96.72.
#{149}
H
is present
YPERTENSION
timated
58 million
in an es-
Americans
(1). In the majority of patients, the cause is unknown; however, a secondary cause is present in approximately of these
artery in less
10%
of patients. Less than half are curable (2). Renal stenosis (RAS), while present than 2% of adults with hyper-
cases
tension (3), remains the single most common correctable cause (2). Currently, an adequate noninvasive screening examination for the detection of significant RAS in the hypertensive population is not available, and angiography remains the standard (4). Previous authors, employing various parameters, have suggested the utility of duplex sonography in the detection of significant RAS (5-12). A peak systolic velocity (PSV) in the menal artery greater than 100 cm/sec has been used as a diagnostic criterion for the presence of RAS (6). Two recent articles have compared the menal-aortic
ratio
(RAR)
(ie,
the
ratio
of
the PSV in the renal artery to the PSv in the aorta) with the percentage of stenosis as determined with angiography.
tides
The
authors
of these
two
am-
demonstrated between
PATIENTS
an
I From the Departments of Radiology (A.L.D., D.M.P., C.K.L., J.C.H.), Urologic Surgery (R.B.T., A.C.N.), Hypertension-Nephrology (J.V.N.), and Biostatistics (A.M.W.), Cleveland Clinic Foundation, 1 Clinic Center Dr, Cleveland, OH 44195-5103. From the 1989 RSNA scientific assembly. Received April 3, 1990; revision requested May 14; revision received and accepted July 20. Address reprint requests to D.M.P. 0 RSNA, 1990
AND
METHODS
Between December 20, tember 14, 1989, 55 kidneys
1988,
undergoing
were
aortography
and
Sep-
in 30 patients examined
with color Doppler flow imaging. The 14 female and 16 male patients ranged in age from 19 to 75 years (mean, 50.3 years). The
patient
population
included
(a) po-
tential renal donors (26 kidneys in 13 patients), (b) patients with suspected renovascular hypertension (25 kidneys in 13 patients), and (c) patients with a known renal mass in whom the contralateral kidney was studied (four kidneys in four patients). The only criterion for patient inclusion in the study was that an aomtogram had been obtained within 10 weeks of the US examination. All US examinations were performed with an Acuson 128 US unit 3.5-MHz
a positive correBAR of 3.5 or greater and RAS of 60% or greater (8,12). A third, more recent study evaluated the use of color duplex ultrasound (US) as a possible screening tool, comparing both BAR and renal artery PSV with the percentage of stenosis as determined with angiography. The sensitivity of RAR and PSV could not be assessed, since the US study was unsuccessful in all of the significantly stenotic arteries. lation
With the utilization of PSV as the major criterion for the diagnosis of RAS, a 31% false-positive rate was obtamed (13). We used color duplex US to determine whether this modality may currently play a role in the detection and grading of RAS.
(Mountain (imaging
View, Calif) and Doppler)
with a phased-
array transducer. Patients were either fasting or had received a clear-liquid diet since
midnight
on the
night
before
the
examination. All studies were performed by one of several US technologists, all cxperienced in the use of Doppler techniques.
Each
graphic
findings
amination.
was
blinded
to the angio-
at the time
In several
of the US cx-
instances
of
technically difficult studies, a radiologist (either of the two who intempreted the studies)
assisted
in performing
the exami-
nation. This occurred in only a limited number of patients and was not recorded. The following measurements were obtamed, when technically achievable: PSV in the midabdominal aorta, PSV and enddiastolic velocity (EDV) in both the pmoximal and distal portions of the renal am-
Abbreviations: RAS = renal artery PSv = peak systolic velocity, RAR tic ratio, EDV end-diastolic velocity, sistive
stenosis, renal-aor-
RI
re-
index.
749
tery,
renal
(RI) (RI mum
parenchymal resistive index [PSV - EDV]/PSV), and maxi-
=
renal
length.
The
velocities
aorta and the renal artery mined from their respective ings.
The
the
proximal
were
proximal
and
60#{176} was
used
Patients supine posterior
and
artery
RIs
were
necessary.
typically
portions
in the
decubitus
or
was
when
used
artery
left
and
distal
usually
neces-
tion. in
US examination
the
following
was
manner:
od, no time
restraint
technologist
for
time If the
tery was not satisfactorily within 15-20 minutes, artery was terminated; nal artery was visualized
study
was
imal
considered
or distal
adequate
artery
from
ar-
on the renal
of the US those of the and the US
successful
renal
was obtained
renal
was placed
obtaining
Similarly,
a single the
experienced
renal
US
in US imaging
techniques.
Neither
had
of the angiographic
and Doppler any
results
the US interpretation. The used to diagnose significant based on previous studies: 3.5
or greater
was
knowledge
at the time major
of
criteria
RAS were (a) an BAR
considered
of
to indicate
stenosis
of 60% or greater and/or (b) a mePSV greater than 100 cm/sec was considered to indicate stenosis of 50% nal
artery
or greater. A diagnosis made if the renal artery
ized
but no Doppler
of occlusion was well
signal
could
was visual-
be ob-
tamed.
Thirty-six intraarterial phy;
19,
kidneys digital with
aortography.
were studied with subtraction aortogma-
conventional
of eight
with
diatrizoate meglumine dium (Renografin-60; Brunswick, NJ) (11.7 at a rate of 20 mL/sec 30-40 mL of two-thirds
zoate (5.9-7.8
750
meglummne g of iodine)
#{149} Radiology
gion that
Neither
1(5) 1(20)
7(32) 1(20)
8(36) 3(60)
6(27) 0(0)
of tracings,
.
percentages
was
adjacent
normal
renal
of the
diameter
mated. graphic
diameter
stenosis
the
One
no knowledge the aortoin
artery,
and
of the
with
or distal
the
percentage
stenosis
was
In all cases, the correlative and color Doppler flow were
obtained
me-
within
esti-
angioexamina-
a period
of 9
weeks. Each kidney was placed into one of the following categories of percentage of stenosis, according to the angiographic results:
normal,
71%-99%,
50%
and
or
less,
51%-70%,
..
.
...
2(22) 1(17)
Among the 30 patients involved in the study, one had a solitary kidney, and four had a unilateral renal mass, leaving 25 pain
whom
both
Among had
the one
kidneys
latter
were
group,
kidney
that
two renal arteries, and two had dual renal arteries to both These kidneys were subsequently ed from statistical analysis for described in the “US Results” Results of intraclass correlations and
percentage 15 patients
both renal arteries could dicated that a relationship the left and right kidneys patient. Thus, ity determinations
for
pa-
supplied
by
US parameters in the remaining
exam-
eight
was
patients kidneys. excludthe reasons section. for the of stenosis whom
in
be examined inexists between within a single
sensitivity and and statistical
ses, one kidney was randomly from each of these 15 patients.
specificanaly-
selected The data
values for each of the parameters within each stenosis group were not normally distributed,
so the
kal-Wallis ences
indexes
nonparametric
test was used
among
with
tiple comparisons termine which were contributing cance.
stenosis
groups.
significant
Krus-
to detect For
differences,
differthose
mul-
were performed to depairs of stenosis groups to the overall signifi-
RESULTS
and diatrizoate soSquibb, New g of iodine) injected (conventional) or strength diatri-
Angiography demonstrated that 42 kidneys were supplied by a single renal artery. These were categorized by percentage of stenosis as follows: normal (n 22), 50% or less (n = 5), 51%-70% (n 0), 71%-99% (n 9),
sodium
at a rate
of
and
occluded
neys
were
5(56) 3(50)
Results
(n
6).
supplied
Although evaluated
was
Thirteen
by two
each separately
renal
renal for
kidam-
artery percent-
age of stenosis as determined with angiography, these data are not included, since no accessory renal artery was visualized with US.
US
Results
The Doppler
average time for the color flow examination was 25 per kidney, including both
minutes
occluded.
Angiographic
injected
.
in parentheses.
teries.
the
compared
proximal
of
the use of 40 mL of
and diatrizoate
.
1(11) 0(0)
with
subtraction).
who had interpreted
luminal
of the in
kid-
neys obtained at outside institutions were reviewed. At our institution, aortography
was performed
Artery Both
screen-film
Angiograms
of Renal
Artery
Distal
...
(digital
The
tients
tracing
patent
grams.
med.
if a prox-
Doppler
mL/sec
tients
examination was considered successful if no tracing was obtained from one occluded renal artery or if tracings were obtained from each menal artery in a kidney with multiple renal arteries. All results of US studies were meviewed by one of two radiologists, both
artery.
Renal
Proximal
1(11) 2(33)
are numbers
angiographer, the US results,
tions
limited
visualized the study of that however, if the mewithin this peri-
artery tracings. The findings study were compared with angiographic examination,
a Single
re-
sitated two separate views. Data from the proximal renal artery were typically obtamed from a transverse image in which the aorta and the origin of the renal artery were identified. Sampling of the distal renal artery was performed at the renal hilum, proximal to the first bifurcaThe
with
was the
proximal
renal
of Stenosis
Note-Values
as
15-20
of the
of the
data.
positioning
helpful imaged.
Visualization
than
scanned
however, positioning
Nonsupine
more frequently nal artery was
these
of US in Kidneys
0(n22) 50(n5) 51-70(n0) 71-99(n9) 100(n6)
technicalof less
in obtaining
position; oblique
Percentage
RARs
renal angle
Rate
Portion
distal
Whenever
Success
trac-
and
a Doppler
1
Technical
the
were deterDoppler
distal
also calculated.
ly possible,
in
Table
successful
and
tions.
The
nation
in
unsuccessful
examina-
range of times for patients in whom both kidneys were exammed was 22-97 minutes. One patient had a solitary kidney, and the examination time was 15 minutes (unsuccessful study). If all patients with Unsuccessful bilateral examinations are excluded, the average examination time per kidney was approximately 25 minutes. Table 1 shows the technical success rate of the color Doppler flow examithose
kidneys
with
a single
renal artery. A proximal and/or distal tracing was obtained from the menal artery in 25 (69%) of 36 kidneys with a patent single renal artery. In 11
(31%)
tery
of
36
kidneys,
the
renal
am-
was
not sufficiently visualized obtain a Doppler tracing proximally or distally within the vessel. These latter
examinations
were
to
considered
technically unsuccessful. The success rate did not differ significantly when results in the right and left sides were compared (69% vs 70%). As the percentage of stenosis increased, our ability to obtain Doppler tracings diminished (78% for the combined 0%-
50% stenosis 7l%-99% tion,
(50%)
groups stenosis
tracings
were
of six occluded of the latter
In two proximal tracing third, a tracing
mally
and
distally.
vs 44% group). obtained
for the In addiin three
renal arteries,
arteries. only a
was obtained; was obtained
Of these
one was typical of that renal artery, while the
in the proxi-
tracings,
of a normal remainder December
1990
S
were
atypical
S
stolic
flow
C’)
In 1
0’
S
II
ft
u
H .5. UH
N. t
C)
5
‘a
.-
c’
-
,-,
_
0’
I
0
-
II
II
I
5)
a’ c’
C’)
‘0
c’
m
‘0
8
C?
In
In
N.
.
C’)
-
‘0 C)
II
II
H
II
.-..
5) (‘I C?
00
m
C’) 0
4 In
m
-
‘0
I
F,
I
‘0
9
C’4
C’)
I
I
I
8
‘0
0’
0’
)< CS
0C’)In
N
0
;
-
0’ ‘0 0’
C’)
ble
N.
-
a)
U S a)
the
a)
II
In
0
0
C’)
H
.-.
.
In
‘0
N H
0’
1’
t,.
0.! CS 5)!
U
‘0
In
U
I,..
8
0
0 I
I
C’)
t% ‘
0’
‘0
‘0
I
so 5)
0’
CS
U
‘
II
II
‘0!
C’)
N.
5) a)
C’)
‘
In
OH
U)
II
H
CS
E x
0
N
0
‘
0.
0’
2
1A00 ‘0 00
0.
CS
I
In ‘0
C’) C’)
0
0’ -
00
0
0.
I 00 CO In
0
‘-
.-
0’
CS
4:
c’1
In N. -
CS
0’ ‘0 I
00
.5C
II
C’)
N
CS
U)
00
C’)
II
0
5)
t .5’
-
0
.5’
‘0
‘0 N.
C’)
N
II
II
II
‘0
N.
‘0
I ‘-4
-
‘0
‘0
0 ‘0
C’)
N.
In ‘0
In
C’)
C)
4:
II 0.
z
0’
0’
-
‘5 In
I N.
.5)
0
‘0
a) t
5;
I
.0
CS
0’
S 5)
C’)
II
II
CS
II
E
.5,
x
N.
2
0’ 00
I N. .5’
0.
0’ In
z
0 N. I
I
0
00
‘0
In
.5’ N.
N.
S a)
0. 0.
CS
II 0.
CS 0 S
CS
8
4:
N
z
‘0
.8
.c
sample
to be marginally significant (P .071) and insignifi(P = .123). However, sizes in the various stenosis
the
groups were small, resulting in low statistical power for detecting relatively small differences in parameters. Table renchymal
percentage ly selected statistical
Volume
pro-
US
proximally cant distally
II
0.
I
-
H
‘0
4:
C’)
In
made
of the occluded
statistical analyses. Table 3 relates RAR,
99% stenosis group compared with the
I It)
0
0.
a diag-
not
parameters for the randomly group of kidneys utilized
ed a statistically
.5’
In -
arteries
the significant stenosis group (0% sensitivity, 100% specificity). In addition, comparison of the median RAR values proximally and distally in those arteries in the normal, 50%-stenosis-or-less, and 7i%-99% stenosis groups did not reveal an increase in the RAR with increasing stenosis. In fact, the median proximal RAR in the 71%-99% stenosis group demonstrat-
CS U
In -
the
of kidneys. were below
8
U
was
by a single
nosis
S
z
In
diaNone
occluded
on the basis in any of the
RI, and
C’)
0
low
Consequently,
3 delineates
same lected
S
II
three
occlusion
supplied
0
In
CS 5)
5) 0 5) S 0 U
0
C’)
I 0’ In
Dt
very
demonstrated.
sels. Table 2 delineates the descriptive data relating to all measured US parametems for all 42 kidneys that were
0
In
of
spectively findings
x a)
.5
other
visualized.
nosis
VI 0.
8U
r
I0
S 5) ‘5
S In
CS
I
U
was
0’
‘0
In
E
of the
in that was
3 also relates the median RI and renal length
pato the
of stenosis in the randomgroup of kidneys used for analysis.
No
specific Radiolov
trend
751
#{149}
in median RI was noted with increasing stenosis, and no significant difference among groups was found (P = .267). The median length of those kidneys with occluded renal amteries was decreased, as expected; however, this difference was not statistically significant (P .152). Thirteen kidneys were supplied by two renal arteries. No accessory arteries were imaged with color Doppler flow examination. Visualization of a single normal renal artery did not cxdude the presence of a nonvisualized, possibly stenotic accessory artery. These US examinations were therefore considered inadequate, yielding a success rate for these 13 kidneys
this
of 0%.
group
Further
was
not
analysis
of
performed.
DISCUSSION Hypertension remains one of the most important risk factors in the development of stroke and coronary artery disease (1). Although renovasculam disease is the cause of hypertension in a minority of patients (3), it is the most common correctable cause (2). Therefore, it remains an important goal to find an accurate, noninvasive
screening
detection
examination
of significant
hypertensive
for
RAS
the
in the
population.
Various authors have evaluated the utility of duplex sonography in the detection of significant RAS. Attempts to use PSV as a major criterion have met with variable success (6,13). Kohier et al (8) and Taylor et al (12) have measured the RAR, have cornpared this value to the percentage of stenosis as determined with angiography, and have shown a positive correlation
between
an
RAR
of 3.5
or
greater and RAS of 60% or greater. High sensitivity and specificity were obtained in both studies (91% and 95%, respectively, in the former study and 84% and 97%, respectively, in the latter). In addition, both groups of investigators achieved a high percentage of technically adequate examinations (90% and 87.6%, respectively), similar to the results in other reported series (5-7, 9-11). We hoped to reproduce these data in our study, which included the added capability anticipated
of color that
flow imaging. We the added dimension
of color would facilitate visualization of the renal arteries and would allow more accurate Doppler angle correction and therefore more accurate yelocity determination. However, even with color flow imaging, we were unable to successfully reproduce the 752
Radiology
#{149}
findings of these previous studies. Our results more closely parallel those of Berland et al (13), who achieved a success rate of only 58% in identifying
the
main
renal
artery
with color duplex US. They were unable to evaluate the sensitivity of RAR or PSV as a diagnostic criterion, since all US examinations in the group of arteries with significant stenosis were technically inadequate. However, with use of renal artery PSv, they obtained a specificity of only 37%. In our study, technically adequate examinations
were
achieved
in only
51% of kidneys (69% of kidneys with a patent single renal artery, 0% of kidneys with two renal arteries). We believe that two main factors contributed to the relatively low number of technically successful studies. The US examinations were performed in a time-limited fashion to examine the true feasibility of this modality as an effective screening tool in a busy US department. Most of the previous studies cited either required a greater amount of time to perform the examination or did not note the time mequired to obtain a satisfactory study (5-8,10,12). Berland et al (13) employed a time-restricted US examination, as we did, and achieved a cornparably low percentage of successful studies. The second factor concerns those kidneys supplied by more than one renal artery. In our study, 24% of the kidneys fell into this category. This number correlates well with reports in the literature, which state that 20%-28% of kidneys have multiple renal arterial supply (14). Because of the possibility of discrepant stenoses among the multiple renal arteries, visualization of a single normal artery does not exclude the possibility of a significantly stenotic accessory artery. Therefore, if all arteries to a kidney are not visualized during the color Doppler flow examination, it should be deemed inadequate as a screening study for that particular kidney. In our study, all of the color Doppler flow examinations performed on kidneys supplied by two renal arteries were deemed technically unsuccessful on this basis. This group of kidneys accounted for a significant percentage of the total number in the study, which contributed greatly to the low success rate. Most previous investigators have not addressed the issue of multiple renal arteries (5-10,12), except for Berland et al (13), who found a similar percentage (2.2%) of kidneys with accessory
arteries and were also unable to visualize any accessory arteries with color Doppler US. Among the kidneys with adequate US examinations, no elevation of the median RAR with increasing stenosis was noted, as was reported in previous studies. In fact, for proximal RAR, a statistically significant decrease in the median value was found in the 71%-99% stenosis group compared with the normal group and 50%-stenosis-or-less group, an obvious contradiction to the results of Kohiem et al (8) and Taylor et al (12); however, the reason for this discrepancy is unclear. Similarly, our results obtained with use of renal artery PSV are somewhat contradictory to those of Avasthi et al (6), who achieved a sensitivity of 42% and a specificity of 92% with use of this parameter to diagnose RAS. We obtained 0% sensitivity proximally and distally and only 50% specificity proximally and 79% specificity distally. Our results more closely correlate with those of Berland et al (13), who reported a specificity of 37%. We also calculated the renal artery RI both proximally and distally, when technically achievable. Theometically, we anticipated that a critical stenosis of the renal artery may be manifested by an increased renal amtery RI due to a decrease in the EDV proximal to the stenosis. Most stenoses in our study were orificial or within the 1st centimeter of the main renal artery (17 of 20), thereby making it difficult or impossible to obtain a renal artery tracing proximal to the stenosis. Only one of the remaining three renal arteries that demonstrated a more distal stenosis was severely stenotic, but the examination in this case was technically inadequate. Consequently, we were unable to assess
the
effect
of severe
stenosis
on
the renal artery RI obtained proximal to the area of narrowing. We did, however, still compare median values of renal artery RI proximally and distally to the percentage of stenosis. No
specific
trend
could
be
detected
with increasing stenosis, and no statistically significant differences were found. The final two parameters measured were the parenchymal RI and renal length. Parenchymal RI is an indicatom of small-vessel resistance in the kidney. Increased RI signifies increased resistance within the small intrarenal vessels and is often associated with significant parenchymal disease and poor function. Elevated RI and decreased renal length may be December
1990
secondary to a variety of causes and are nonspecific findings. Therefore, these parameters were evaluated for use as possible minor criteria to support a diagnosis of RAS in the face of marginally positive major criteria. As noted previously, no meaningful trend in median parenchyrnal RI was noted with increasing percentage of stenosis of the renal artery; therefore, this was not a helpful parameter, nor was renal length. The only detectable difference in median renal length was a decrease in the kidneys with occluded arteries; however, this did not prove to be statistically significant. In 50% of the occluded arteries, a Doppler tracing was obtained; other authors have reported the same prob1cm (7,8,1 1). These “false” tracings were probably obtained from a large collateral vessel or from a meconstituted segment of the main renal artery. Clearly, this problem may serve as a source of false-negative examination results in the clinical setting.
of the most recent studies (13) and our own study, both utilizing color Doppler flow Imaging, yielded mesults contradictory to those previously reported in the literature. These discrepancies indicate the need for continued prospective studies with larger numbers of patients to clarify the accuracy of the proposed objective diagnostic criteria. However, as evidenced by the low number of technically adequate examinations we achieved, duplex sonogmaphy of the renal artery is difficult to perform, especially in a time-limited setting. With current technical capability, color Doppler flow US is not an adequate screening examination for the detection of significant RAS. U References 1.
2.
CONCLUSION 3.
During the past 10 years, many studies have been performed to evaluate the mole of duplex US in the detection of RAS, and most have yielded promising results. However, one
6.
7.
8.
9.
59:225-229.
10.
Saunders,
1988;
4.
Treadway hypertension.
KK, Slater Annu
5.
Greene ER, Venters MD, Avasthi PS, Conn RL, Jahnke RW. Noninvasive characterization of renal artery blood flow. Kidney tnt 1981; 20:523-529.
Jenni
R, Vieli A, Luscher TF, Schneider E, W, Anliker M. Combined two-dimensional ultrasound Doppler technique: new possibilities for the screening of renovascular and parenchymatous hypertension? Nephron 1986; 44(suppl l):2-4. Robertson R, Murphy A, Dubbins PA. Renal artery stenosis: the use of duplex ultrasound as a screening technique. Br J Radiol 1988; 61:196-201. Taylor DC, Kettler MD, Moneta GL, et al.
Vetter
11.
12.
Frohlich ED. Evaluation and management of the patient with essential hypertension. In: Parmley WW, Chatterjee K, eds. Cardiology. Vol 2. Philadelphia: Lippincott, 1989; 1-15. Cressman MD, Gifford RW. Evaluation of secondary forms of hypertension. In: Parmley WW, Chatterjee K, eds. Cardiology. Vol 2. Philadelphia: Lippincott, 1988; 1-22. Kaplan NM. Systemic hypertension: mechanisms and diagnosis. In: Braunwald E, ed. Heart disease: a textbook of cardiovascular medicine. 3rd ed. Philadelphia:
Avasthi PS, Voyles WF, Greene ER. Noninvasive diagnosis of renal artery stenosis by echo-Doppler velocimetry. Kidney Int 1984; 25:824-829. Norris CS, Pfeiffer JS, Rittgers SE, Barnes RW. Noninvasive evaluation of renal artery stenosis and renovascular resistance. VascSurg 1984; 1:192-201. Kohler TR, Zierler RE, Martin RL, et al. Noninvasive diagnosis of renal artery stenosis by ultrasonic duplex scanning. Vasc Surg 1986; 4:450-456. Dubbins PA. Renal artery stenosis: duplex Doppler evaluation. Br J Radiol 1986;
Duplex nosis
13.
artery
scanning stenosis:
in the diaga prospec-
tive evaluation. J Vasc Surg 1988; 7:363369. Berland LL, Koslin DB, Routh WD, Keller
FS.
14.
ultrasound of renal
Renal
artery
stenosis:
prospective
evaluation of diagnosis with color duplex US compared with angiography. Radiology 1990; 174:421-423. Davis GD. Variations in renal blood supply. In: Kincaid OW, ed. Renal angiography. Chicago: Year Book Medical, 1966; 83-87.
819-861.
EE. Rev
Renovascular Med 1984;
35:665-692.
Volume
177
Number
#{149}
3
Radiology
753
#{149}
