Content uploaded by Jan Tordoir
Author content
All content in this area was uploaded by Jan Tordoir on Dec 14, 2013
Content may be subject to copyright.
Journal of Human Hypertension (2002) 16, 501–507
2002 Nature Publishing Group All rights reserved 0950-9240/02 $25.00
www.nature.com/jhh
ORIGINAL ARTICLE
Renovascular disease in patients with
hypertension: detection with duplex
ultrasound
MW de Haan
1
, AA Kroon
3
, K Flobbe
1
, AGH Kessels
2
, JH Tordoir
4
,
JMA van Engelshoven
1
and PW de Leeuw
3
1
Department of Diagnostic Radiology, University Hospital, Maastricht, The Netherlands,
2
Department of
Clinical Epidemiology and Medical Technology Assessment, University Hospital Maastricht, The
Netherlands,
3
Department of Internal Medicine, University Hospital Maastricht, The Netherlands,
4
Department of Surgery, University Hospital Maastricht, The Netherlands
The aim of this study was to evaluate the accuracy of
duplex ultrasound for the diagnosis of renovascular dis-
ease in a cohort of hypertensive patients. In 78 patients
suspected of renovascular hypertension on clinical
grounds duplex ultrasound examination of the renal
arteries was performed. Renal angiography was used as
the standard of reference. Duplex ultrasound was incon-
clusive in 11 kidneys (7%). None of the supernumerary
renal arteries was detected with duplex ultrasound. The
overall prevalence of significant renovascular disease
(⭓50% stenosis) was 20%. Based on the combination of
parameters at thresholds commonly applied in current
literature: ie PSV
max
⬎180 cm/sec and RAR ⬎3.5 the
overall sensitivity of duplex ultrasound for detection of
haemodynamically significant renovascular disease
was 50.0% with a specificity of 91.3% (PPV: 87.9%; NPV:
Keywords: renovascular; renal arteries, stenosis or obstruction; renal arteries, US; ultrasound, Doppler studies
Introduction
Identification of renovascular disease in hyperten-
sive patients remains important as correction of
renal artery stenosis may improve blood pressure
control or stabilise renal function.
1–3
Conventional
X-ray angiography is still the most accurate method
to assess the presence and severity of renal artery
stenosis. However, its invasive nature and high costs
make this procedure unsuitable as a screening
modality in a population of patients with a low
prevalence of renal artery stenosis. Several less
invasive tests (including captopril renography) have
been proposed to select patients for angiography.
4–7
Magnetic resonance angiography (MRA) has shown
Correspondence: MW de Haan, Department of Diagnostic Radi-
ology, University Hospital Maastricht, P. Debyelaan 25, PO Box
5800, 6202 AZ Maastricht, The Netherlands.
E-mail: mdehaan얀radiology.azm.nl
Received 12 December 2001; revised and accepted 25 March 2002
59.1). Lowering the thresholds for both parameters
improved the test results at the cost of a significant
increase of false positive examinations. In a population
of hypertensive patients clinically suspected of renova-
scular hypertension, only limited results for duplex
ultrasound could be acquired in the detection of renova-
scular disease. This result, in combination with the wide
range of sensitivities and specificities published in
international literature and the relatively large number of
incomplete examinations does not support the general
application of duplex ultrasound as a screening pro-
cedure for detection and assessment of renovascular
disease.
Journal of Human Hypertension (2002) 16, 501–507. doi:
10.1038/sj.jhh.1001429
promising results as a non-invasive imaging
modality for assessing the renal arteries, but to-date
fails to depict the hilar and parenchymal renal
arteries. Also, CT angiography shows excellent
depiction of the first segments of the renal arteries,
at the expense of a relatively large amount of
(nefrotoxic) contrast material and radiation.
Duplex ultrasound has the potential to combine
direct visualisation of the renal arteries with acqui-
sition of flow velocity profiles, thus providing both
anatomical and functional information and has been
advocated as the optimal screening test. However,
the results of studies concerning the validity of
duplex ultrasound are sometimes conflicting and
there is no general agreement about its value as a
screening test for renovascular disease.
8,9
The aim of this study was to evaluate the accuracy
of duplex ultrasound for the diagnosis of renal artery
stenosis in a cohort of unselected hypertensive
patients with clinical suspicion of renovascular dis-
ease using conventional renal angiography as the
Renovascular disease in patients with hypertension
MW de Haan
et al
502
Journal of Human Hypertension
standard of reference. In addition, we tried to deter-
mine the most useful combination of thresholds for
the applied ultrasound parameters.
Materials and methods
Patients
Over a period of 3 years, 78 consecutive patients
with malignant, accelerated and/or treatment-resist-
ant hypertension (blood pressure ⬎160/100 mm Hg
despite treatment with two or more antihypertensive
drugs), loss of renal function after angiotensin-
converting enzyme (ACE) inhibition, unilateral loss
of kidney volume or unexplained deterioration of
renal function (change in serum creatinin
⬎20
mol/l within 12 months) were examined by
both colour duplex ultrasound and conventional X-
ray angiography of the renal arteries.
The results of seven patients were excluded from
analysis because duplex ultrasound was performed
only after endoluminal renal intervention. Thus,
data of 71 patients was available for evaluation and
comparison. Thirty-three (46%) of these patients
were men and 38 women; mean age was 55 years
(range 20–78 years).
Duplex ultrasound
The ultrasound studies were performed using a col-
our duplex scanner with a 2.5 and/or 5.0 MHz probe
(Aloka Company Europe, Hoofddorp, The
Netherlands) with the patient in a supine position.
Patients were either fasting or had received a clear-
liquid diet since midnight on the day before the
examination. The patients were examined in
anterior, lateral and anterolateral positions. The
renal length was recorded for each kidney. Doppler
samples were taken along the course of the renal
arteries as well as in the abdominal aorta at the level
of the renal arteries. Peak velocity readings were
obtained by means of angle correction with Doppler-
to-vessel angle of less than 60° and a 2 mm Doppler
gate width. Following published criteria, a peak sys-
tolic velocity in the renal artery (PSV
max
) of more
than 180 cm/sec and its ratio to the peak systolic
velocity in the abdominal aorta (RAR) of 3.5 or more
were used to discriminate a renal artery stenosis of
ⱖ50%.
10–12
A renal artery stenosis of ⬍50% was
defined as PSV
max
⬎180 cm/sec in combination with
a RAR ⬍3.5.
The grey-scale images were not used to confirm or
refute a diagnosis of renal artery stenosis. Renal
artery occlusion was diagnosed when there was no
flow signal in the renal artery and a low amplitude
velocity signal from the renal parenchyma. The
ultrasound studies were interpreted without knowl-
edge of the results of angiography. Two vascular
technicians experienced in the use of Doppler tech-
niques performed all examinations. The technicians
often sought the help of their colleague for patients
who were difficult to examine. No difference was
noted in the accuracy of the results or in the quality
of the studies among the two technicians who par-
ticipated in our study. The examination time lasted
between 45 and 60 min for both sides.
Conventional angiography
Angiography of the abdominal aorta and renal
arteries was performed in the anteroposterior view
with injection of 35 ml contrast material through a
5-F pigtail catheter (William Cook Europe, Bjerskov,
Denmark) positioned at the level of the renal
arteries. If the renal arteries were not adequately
depicted, additional angiographic series were
obtained in oblique projections and/or selective
renal angiography was performed with an end-hole
catheter (William Cook-Europe).
For the purpose of this study two independent
observers, who were unaware of the duplex ultra-
sound results and/or clinical history interpreted the
angiographic examinations again. The status of the
renal artery was graded into four categories: normal;
less than 50% stenosis; greater than 50% stenosis;
or occluded. Renal arteries with evidence of fibro-
muscular dysplasia (FMD) were graded as signifi-
cantly stenosed (⭓50%). The two observers
reviewed discordant cases together to come to a con-
sensus decision.
Statistical analysis
Sensitivity and specificity and positive and negative
predictive values regarding the detection of signifi-
cant renal artery stenosis were calculated at the level
of the separate kidneys as well as at the level of
patients with the results of conventional angiogra-
phy as the standard of reference. Determination of
the best threshold of the duplex ultrasound para-
meters included receiver-operator curve analysis
(ROC) and calculation of sensitivity, specificity and
positive and negative predictive values at various
thresholds.
Results
A total of 141 kidneys were examined in 71 patients.
One kidney could not be examined because it had
previously been removed surgically.
The duplex ultrasound study was unsuccessful in
11 kidneys (7%) due to the presence of bowel gas
and/or obesity. Of the remaining 130 kidneys, 106
(81.5%) showed a single renal artery. Twenty-four
kidneys (18.4%) had supernumerary renal arteries
on angiography, one of which showed a significant
stenosis. None of these accessory renal arteries were
detected with duplex ultrasound.
On the basis of angiographic findings, 104 kidneys
without stenotic lesions or with mild renal artery
stenosis (⬍50%) were identified. In 24 kidneys
(18.4%) significant renal artery stenosis (⭓50%) was
Renovascular disease in patients with hypertension
MW de Haan
et al
503
Table 1 Duplex ultrasound parameters recorded from the renal
arteries
Results Results angiography
duplex
Normal Stenosis Stenosis FMD Occlusion Total
⬍50% ⬎50% no.
PSV
max
199.0 139.8 256.5 167.9 150.0
(s.d.) (59.33) (46.98) (103.71) (52.56) (74.95)
RAR 1.75 2.13 4.66 2.41 2.8
(s.d.) (0.9) (0.88) (2.71) (1.06) (1.97)
Total 83 21 15 9 2 130
number
Data are presented as mean and standard deviations (s.d.’s). FMD
= fibromuscular dysplasia; PSV
max
= peak systolic velocity; RAR
= ratio of peak renal artery velocity to aortic velocity.
noted, including nine arteries with fibromuscular
dysplasia (7%). Total occlusion of the renal artery
was seen in two kidneys (1.5%), resulting in an
overall prevalence of significant renovascular dis-
ease of 20%.
The range of renal size in the group of kidneys
without or mild renal artery stenosis (⬍50%) was
8.4–12.9 cm (mean length 10.7 cm), whereas the
range in the group of kidneys with significant renal
artery stenosis was 8.0–12.2 cm (mean length
10.6 cm).
The mean values and standard deviations of a
peak systolic velocity (PSV
max
) and peak renal to
aortic velocity ratio (RAR) in relation to the angio-
graphy results are presented in Table 1. Based on
the combination of these parameters at thresholds
commonly applied in current literature: ie, PSV
max
⬎180 cm/sec and RAR ⬎3.5, the results of the
duplex ultrasound examination at the level of the
separate kidneys compared to the angiographic fin-
dings are displayed in Table 2.
11–13
Excluding incon-
clusive examinations and using a ⭓50% stenosis or
Table 2 Comparison between angiography and duplex ultra-
sound in relation to the percentage of renal artery stenosis in
hypertensive patients suspected of having renovascular disease
Results Results angiography
duplex
Normal Stenosis Stenosis FMD Occlusion Total
⬍50% ⬎50% no.
Normal 72 17 4 5 1 99
Stenosis 4 2 0 3 0 9
⬍50%
Stenosis 7 2 11 1 0 21
⭓50%
Occlusion 0 0 0 0 1 1
Total 83 21 15 9 2 130
number
Renal artery stenosis ⬍50%: PSV
max
⬎180 cm/sec and RAR ⬍3.5.
Renal artery stenosis ⭓50%: PSV
max
⬎180 cm/sec and RAR ⬎3.5.
FMD = fibromuscular dysplasia.
Journal of Human Hypertension
the presence of fibromuscular dysplasia on angio-
graphy as the diagnostic criterion for haemodynami-
cally significant renovascular disease, these results
show an overall sensitivity of 50% and a specificity
of 91%. The overall positive (PPV) and negative
(NPV) predictive values were 59% and 88% respect-
ively. At the level of the individual patient the val-
idity parameters show only marginal improvement
with a sensitivity of 55% and a specificity of 84%
(PPV 57%; NPV 82%). Receiver operating character-
istic (ROC) analysis of the applied ultrasound para-
meters show comparable curves for both PSV
max
and
RAR (Figure 1).
To improve the validity parameters of our duplex
ultrasound examination, the results were also ana-
lysed for both the PSV
max
and RAR at other thresh-
olds. At a constant threshold of the RAR the ROC-
curve and its area under the curve (AUC) for the
PSV
max
were determined. By calculating this AUC at
different values of the RAR threshold the optimal
RAR threshold was obtained. In the corresponding
ROC-curve the optimal threshold of the PSV
max
could be derived. This resulted in lower thresholds
for both duplex ultrasound parameters. Lowering
the peak systolic velocity threshold to 110 cm/sec
yielded a sensitivity of 85% and a specificity of
51%. Similarly, lowering the threshold for the renal
aortic ratio to 1.7 yielded a sensitivity of 81% and
a specificity of 56%. Combining these parameter
thresholds resulted in a sensitivity of 81% and a
specificity of 64% (Table 3).
Grading fibromuscular dysplasia (FMD) as sig-
nificant renovascular disease (⭓50%), duplex ultra-
sound reached a sensitivity of 11% with a specificity
of 81% in this subgroup of patients.
Figure 1 Receiver operating characteristic (ROC) curves showing
sensitivity for detection of renal artery stenosis (⭓50%) vs 1-
specificity for different PSV
max
and RAR cut-off values.
Renovascular disease in patients with hypertension
MW de Haan
et al
504
Journal of Human Hypertension
Table 3 Detection of renal artery stenosis: results for duplex
ultrasound parameters at different thresholds and combinations
Parameter Sensitivity Specificity PPV NPV
threshold (%) (%) (%) (%)
PSV ⬎180 cm/sec 61.5 86.5 53.3 90.0
RAR ⬎3.5 50.0 90.4 56.5 87.9
Combination 50.0 91.3 59.1 88.0
PSV ⬎180 cm/sec
and RAR ⬎3.5
PSV ⬎110 cm/sec 84.6 51.0 30.1 93.0
RAR ⬎1.7 80.8 55.8 31.3 92.1
Combination 80.8 63.5 35.6 93.0
PSV ⬎110 cm/sec
and RAR ⬎1.7
PSV = peak systolic velocity; RAR = renal aortic ratio; PPV = posi-
tive predictive value; NPV = negative predictive value.
Discussion
The results of this study, carried out in a population
of patients with clinically suspected renovascular
disease, show only moderate results of colour
duplex ultrasound in the detection of renal artery
stenosis.
The clinical presentation of renovascular disease
varies significantly between patients, ranging from
incidental findings of renal artery stenosis on angi-
ography to severe hypertension and progressive loss
of renal function. Identification of renal artery sten-
osis still seems useful since surgical or percutaneous
treatment of these lesions may improve blood press-
ure control and/or stabilise renal function, even
though the latter has been disputed in a recent pub-
lication.
14
Numerous diagnostic tests have been proposed to
identify patients with renovascular disease. How-
ever, none of them has received widespread clinical
acceptance. The ideal screening procedure should
be simple, non- (or minimally) invasive and highly
accurate both for the detection of renovascular dis-
ease and for assessment of its clinical significance.
Duplex ultrasound has been proposed as a promis-
ing modality in this respect, but reported data over
the last 10 years have shown contradictory results in
terms of its diagnostic accuracy. Some investigators
report remarkable success, while others have
reported dismal results in large unselected patient
groups.
15,16
A recent study declared duplex ultrasound to be
most valuable not only in the screening for renal
artery disease but also to estimate its functional sig-
nificance by means of determination of the resistive
index.
17
The duplex ultrasound performance in this
study, however (ie, sensitivity 97%; specificity
98%), was derived from a previous report by the
same authors in a selected group of patients with
a high prevalence of renovascular disease (58%).
18
Therefore, implicit adoption of these test perform-
ance parameters to a non-selected population with
a relatively low prevalence for renovascular disease
(2.3%) seems somewhat premature and raises
doubts about whether the resistive index has the
same prognostic significance for the remaining
patients with stenosis not detected by duplex ultra-
sound.
In the present study, duplex ultrasound para-
meters at thresholds commonly applied in inter-
national literature were used: peak systolic velocity
⬎180 cm/sec (PSV
max
) and peak renal to aortic velo-
city ratio ⬎3.5 (RAR), resulting in an overall sensi-
tivity of 50% and a specificity of 91% (PPV: 59%;
NPV: 88%). This relatively low positive predictive
value is acceptable for a screening test since it
reflects the fact that a degree of overcalling is desir-
able to minimise false-negative rates. However, the
false-negative predictive value, which is a more
important predictor of the usefulness of a parameter
in this respect, may prove to be too low for general
screening purposes.
Post-hoc improvement of our duplex ultrasound
results could be obtained by lowering both para-
meter thresholds. Based on ROC-curve analysis,
including calculation of the respective area’s under
the curve, the preferred combination of the PSV
max
and RAR could be derived for the detection of renal
artery stenosis. At thresholds for the PSV
max
and
RAR of 110 cm/sec and 1.7 respectively, this combi-
nation showed reasonable results in terms of sensi-
tivity, specificity and predictive values (Table 3).
This reduction in duplex parameter thresholds
contrasts with the otherwise wide range of estab-
lished criteria published in international litera-
ture.
13
The contrast can be explained by a systematic
underestimation of the peak: systolic velocity
resulting in relatively low threshold values for both
PSV
max
and RAR in our study group, which in itself
reflects the notorious operator dependency of the
ultrasound examination in general. By using the
combined parameters of peak systolic velocity and
renal aortic ratio at these lower thresholds, we could
have been 93% confident that a negative results
means that a renal artery is free from significant
stenosis with a consequential decrease in positive
predictive value (Table 3). In our study, this would
have resulted in a reduction of false-negative exam-
inations by eight patients at a cost of angiographic
examinations in 38 patients with a false-positive
outcome. Yet, the clinical implementation of screen-
ing tests such as duplex ultrasound is not determ-
ined by its validity results only. Other factors, such
as population structure, pre-test likelihood and cost-
effectiveness also influence the actual usefulness,
which may lead to disparity in appreciation of test
results in different clinical circumstances.
All incomplete examinations (11 kidneys) were
excluded from our results, whereas these data
should be added as a positive result since they
would require angiography in a screening setting.
Based on the angiographic results in this sub-group
of patients and on the traditional threshold para-
Renovascular disease in patients with hypertension
MW de Haan
et al
505
Table 4 Synthesis of studies concerning the detection of renovascular disease by duplex ultrasound vs angiography, using three differ-
ent duplex ultrasound strategies
Authors Year Kidneys Patients Data Sensitivity Specificity Prevalence Supernumerary Missing Odd’s
(n =)(n=) acquisition (%) (%) (%) arteries observations ratio
(%) (%)
(a) Duplex data acquisition using the traditional criteria (PSV
max
, RAR) in the renal artery only (RA)
Breitenseher
9
1992 34 17 RA 17 89 20 21 32 1.7
Miralles
19
1993 92 46 RA 90 91 37 11 0 91
Olin
20
1995 187 102 RA 98 98 66 9 8 2401
Strotzer
11
1995 110 55 RA 88 85 9 15 10 42
Miralles
21
1996 156 78 RA 87 91 37 12 8 68
Vigna
22
1998 196 98 RA 69/89 98/97 25 16 6 109/262
Kaplan
12
1998 56 28 RA 83 81 42 – 20 20.8
House
23
1999 125 63 RA 85 76 22 11 13 17.9
Claudon
13
2000 382 191 RA 79 98 38 13 24 184
This study 2001 141 71 RA 50 91 20 18 7 10.1
(b) Duplex data acquisition by spectral waveform analysis in the intrarenal vasculature (IR)
Bardelli
24
1992 98 49 IR 79 94 41 – 059
Schwerk
25
1994 142 72 IR 82 92 24 1 0 52
Strunk
26
1995 99 50 IR 69 90 17 42 4 20
Pedersen
27
1996 262 131 IR 75 76 35 – 0 9.5
Postma
28
1996 114 57 IR 47 97 37 – 9 28.7
Baxter
8
1996 143 73 IR 89 97 27 14 16 262
Lucas
29
1996 104 53 IR 68 91 21 13 3 21.5
Oliva
10
1998 135 71 IR 81 98 35 – 0 209
Johansson
30
2000 242 121 IR 84 94 19 – 0 82.3
(c) Duplex data acquisition by direct visualization of the renal artery in combination with spectral waveform analysis in the main
and segmental arteries (SARA)
Handa
34
1988 40 20 SARA 100 93 25 – 0 ⬁
Zoller
35
1992 172 86 SARA 84 99 13 – 8 520
Postma
36
1992 122 61 SARA 63 86 52 11 25 10.5
Dondi
37
1992 125 63 SARA 85 76 32 – 5 17.9
Kliewer
31
1993 93 36 SARA 57 69 30 15 0 3.0
Spies
38
1995 268 135 SARA 94 92 22 8 27 180
De Cobelli
39
2000 90 45 SARA 79 93 31 14 1 50
Odds ratio = sensitivity × specificity/{1 − sensitivity} × {1 − specificity}.
meters, sensitivity and specificity for duplex ultra-
sound would have been 57% and 86% respectively.
Supernumerary renal arteries were noted on angi-
ography in 18% of the kidneys (n = 24), but none of
these were detected with duplex ultrasound. Visual-
isation of a single normal renal artery does not
exclude the possibility of a stenotic supernumerary
renal artery. Although the influence of stenoses in
small supernumerary renal arteries is not com-
pletely understood, the examination should be
marked as inadequate, further limiting the outcome
of our data and, therefore, the usefulness of
duplex ultrasound.
Duplex ultrasound performed particularly poor in
patients with fibromuscular dysplasia (n = 5),
resulting in a sensitivity of 11% and specificity of
81%, with duplex parameters which fall within the
normal range of non-diseased vessels (Table 1). This
observation might be caused by fundamental differ-
ences in underlying pathology compared to atheros-
clerotic renovascular disease. Also, the prerequisite
to evaluate the entire length of the renal artery
including the more peripheral segment, which, even
in experienced hands is cumbersome, may be an
Journal of Human Hypertension
other explanation for this lack of discriminating
power. However, the number of patients with FMD
in our group is far too small to allow for a well-
founded discussion in this matter.
Compared with other studies in which only
patients screened for renovascular hypertension
were included, our results fall well within the wide
range of reported data in terms of specificity, sensi-
tivity and number of inadequate examinations. In
our study as well as in others (Table 4a), the duplex
ultrasound data of supernumerary renal arteries
were excluded from evaluation, thereby accepting a
number of inadequate examinations since these ves-
sels may be as large as the main renal arteries and
therefore of hemodynamic significance.
9,11–13,19–23
In order to overcome these limitations, others
have suggested the technique of duplex ultrasound
examination of the intrarenal vasculature.
8,10,24–30
Analysis of the Doppler spectral waveform, with
several criteria, ie mean resistive index, tardus-
parvus phenomena and acceleration time, has been
proposed as an alternative to detect upstream sten-
osis. With the lateral approach to the kidney, duplex
examination of the intrarenal arteries is easier to
Renovascular disease in patients with hypertension
MW de Haan
et al
506
Journal of Human Hypertension
perform resulting in a higher number of complete
examinations than with direct scanning of the renal
arteries. Furthermore, recording of changes in
Doppler velocity waveform pattern downstream in
the segmental arteries makes it possible, at least in
theory, to detect stenotic lesions in main, super-
numerary and segmental arteries, although this
hypothesis has been argued by others.
31
Despite
these theoretical advantages the range in sensi-
tivities and specificities remains relatively wide: 57–
100% and 69–99% respectively (Table 4b). These
discrepancies can be partly attributed to the varia-
bility in criteria and degree of stenosis. However, the
intrarenal parameters are probably influenced by
several, complex factors (ie, pulse, diastolic blood
pressure) which may also account for the differences
in the tabulated results.
32,33
Direct visualisation of the main renal arteries with
duplex ultrasound combined with Doppler spectral
waveform analysis in the main and segmental renal
arteries
31,33–39
does not seem to result in significantly
better results with sensitivities ranging from 47–
89% and specificities from 76–98% (Table 4c).
The range and variety of sensitivities and speci-
ficities in the quoted studies cannot be explained by
the application of different cut-off points. This is
demonstrated in Figure 2, as it is not possible to fit
one summary receiver-operating curve (SROC) over
Figure 2 Sensitivity vs 1-specificity in 25 studies concerning
detection of renovascular disease with duplex ultrasound vs angi-
ography, using three different duplex ultrasound strategies. Note:
Duplex data acquisition using the traditional criteria (PSV
max
,
RAR) in the renal artery only (쎲); or by direct visualisation of the
renal artery in combination with spectral waveform analysis in
the main and segmental arteries (+); or by spectral waveform
analysis in the intrarenal vasculature (왕). Present study using tra-
ditional parameter thresholds: PSV
max
180 cm/sec and RAR ⬎3.5
(#). Present study using lowered parameter thresholds: PSV
max
110 cm/sec and RAR ⬎1.7 (䉬).
the marks representing the different studies and
ultrasound strategies.
40
For quantitative analysis of the literature data the
diagnostic Odds Ratio (= sensitivity × specificity/{1
− sensitivity} × {1 − specificity}) was calculated for
each study (Table 4a–c). The Spearman rank corre-
lation coefficients between diagnostic Odds Ratio’s
of the separate studies and a number of study vari-
ables, ie number of patients, prevalence of renovas-
cular disease, year of publication, are less than 0.3,
indicating a lack of influence of these variables on
the diagnostic performance of duplex ultrasound. A
sound explanation for the wide range of odds ratio’s
(Table 4) cannot be deducted from the indicated
variables in the respective study protocols. Yet,
apart from differences in composition of the study
population, the different levels of experience with
duplex ultrasound in the institutions concerned is
likely to account for a substantial part of this variety
in results.
In conclusion, in a population of hypertensive
patients suspected of renovascular hypertension on
clinical grounds, duplex ultrasound proved to have
limited value for detection of renovascular disease
when commonly applied parameter thresholds:
PSV
max
180 cm/sec and RAR ⬎3.5 were used. In
addition, the relatively large number of incomplete
examination in combination with the wide range of
results published in international literature does not
support the general application of duplex ultra-
sound as a screening procedure for detection and
assessment of renovascular disease.
References
1 Jensen G et al. Treatment of renovascular hyperten-
sion: one year results of renal angioplasty. Kidney Int
1995; 48: 1936–1945.
2 Christensson A. Renovascular disease and renal insuf-
ficiency – diagnosis and treatment. Scand J Urol
Nephrol 1999; 33: 400– 405.
3 Erdoes LS, Berman SS, Hunter GC, Mills JL. Compara-
tive analysis of percutaneous transluminal angioplasty
and operation for renal revascularisation. Am J Kidney
Dis 1996; 4: 496–503.
4 Hillman BJ. Imaging advances in the diagnosis of reno-
vascular hypertension. AJR 1989; 153:5–14.
5 Kim D et al. Abdominal aorta and renal artery stenosis:
evaluation with MR Angiography. Radiology 1990;
174: 727–731.
6 Yucel KE. Magnetic resonance angiography of the
lower extremity and the renal arteries. Semin Ultra-
sound, CT and MRI 1992; 13: 291–302.
7 Davidson RA, Wilcox CS. Newer tests for the diagnosis
of renovascular disease, captopril renogram. JAMA
1992; 268: 3353–3358.
8 Baxter GM et al. Colour Doppler ultrasound in renal
artery stenosis: intrarenal waveform analysis. Br J
Radiol 1996; 69: 810–815.
9 Breitenseher von M et al. Screening of renal artery
stenosis: early results of colour Doppler sonography.
Fortschr Ro
¨
ntgenstr 1992; 156: 228–231.
10 Oliva V et al. Detection of renal artery stenosis with
Doppler sonography before and after administration of
Renovascular disease in patients with hypertension
MW de Haan
et al
507
captopril: value of early systolic rise. AJR 1998; 170:
169–175.
11 Strotzer M et al. Non-invasive assessment of renal
artery stenosis: a comparison of MR angiography, color
Doppler sonography, and intaarterial angiography.
Acta Radiologica 1995; 36: 143–247.
12 Kaplan S, Nadja C. Captopril renography and duplex
Doppler sonography in the diagnosis of renovascular
hypertension. Nephrol Dialysis Transplant 1998; 13:
313–317.
13 Claudon M et al. Renal arteries in patients at risk of
renal artery stenosis: multicenter evaluation of the
echo-enhancer SH U 508A at color and spectral
Doppler US. Radiology 2000; 214: 739–746.
14 van Jaarsveld BC et al. The effect of balloon angiopla-
sty on hypertension in atherosclerotic renal-artery
stenosis. Dutch Renal Artery Stenosis Intervention Co-
operative Study Group. N Engl J Med 2000; 342:
1007–1014.
15 Berland L, Koslin D, Routh W, Keller F. Renal artery
stenosis: prospective evaluation of diagnosis with
color duplex ultrasound compared with angiography.
Radiology 1990; 174: 421– 423.
16 Desberg A et al. Renal artery stenosis: evaluation with
color doppler flow imaging. Radiology 1990; 177:
749–753.
17 Radermacher J et al. Use of Doppler Ultrasonography
to predict the outcome of therapy for renal-artery sten-
osis. N Engl J Med 2001; 344: 410–417.
18 Radermacher J et al. Detection of significant renal
artery stenosis with color Doppler sonography: com-
bining extrarenal and intrarenal approaches to mini-
mize technical failure. Clin Nephrol 2000; 53: 333–
343.
19 Miralles M et al. Renal duplex scanning: correlation
with angiography and isotopic renography. Eur J Vasc
Surg 1993; 7: 188–194.
20 Olin JW et al. The utility of duplex ultrasound scan-
ning of the renal arteries for diagnosing significant
renal artery stenosis. Annals Int Med 1995; 122: 833–
838.
21 Miralles M et al. Value of Doppler parameters in the
diagnosis of renal artery stenosis. J Vasc Surg 1996; 23:
428–435.
22 Vigna C et al. Comparison of steerable continuous-
wave versus pulsed-wave Doppler ultrasound to renal
artery angiography in diagnosing renal artery stenosis.
Am J Cardiol, 1998; 81: 365–367.
23 House MK, Dowling RJ, King P, Gibson RN. Using
Doppler sonography to reveal renal artery stenosis: an
evaluation of optimal imaging parameters. AJR 1999;
173: 761–765.
24 Bardelli M, Jensen G. Vlokmann R, Aurell M. Non-
invasive ultrasound assessment of renal artery stenosis
by means of the Gosling pulsatility index. J Hypertens
1992; 10: 985–989.
25 Schwerk WB et al. Renal artery stenosis: grading with
image-directed Doppler US evaluation of renal
resistive index. Radiology 1994; 190: 786–790.
Journal of Human Hypertension
26 Strunk H, Jaeger U, Teifke A. Intrarenal colour-coded
duplex sonography in the diagnosis of renal artery
stenosis with special regard to multiple renal arteries:
analysis of Doppler waveform and “tardus-parvus”
phenomenon. Ultraschall in Med 1995; 16: 172–179.
27 Pedersen EB et al. Diagnosing renal artery stenosis: a
comparison between conventional renography, capto-
pril renography and ultrasound Doppler in a large con-
secutive series of patients with arterial hypertension.
Blood Press 1996; 5: 342–348.
28 Postma CT, Bijlsta PJ, Rosenbusch G, Thien Th. Pat-
tern recognition of loss of early systolic peak by
Doppler ultrasound has a low sensitivity for the detec-
tion of renal artery stenosis. J Hum Hypertens 1996;
10: 181–184.
29 Lucas P, Blome S, Roche J. Intra-renal Doppler wave-
form analysis as a screening test for renal artery sten-
osis. Austr Radiol 1996; 40: 276–282.
30 Johansson M et al. Evaluation of duplex ultrasound
and captopril renography for detection of renovascular
hypertension. Kidney Int 2000; 58: 774 –782.
31 Kliewer MA et al. Renal artery stenosis: analysis of
Doppler waveform parameters and parvus-tardus pat-
tern. Radiology 1993; 189: 779–787.
32 Malatino LS et al. Diagnosis of renovascular disease
by extra and intrarenal Doppler parameters. Angiology
1998; 49: 707–721.
33 Grenier N, Trillaud H. Comparison of imaging
methods for renal artery stenosis. BJU Int 2000; 86
(Suppl 1): 84 –94.
34 Handa N et al. A new accurate and non-invasive
screening method for renovascular hypertension: the
renal artery Doppler technique. J Hypertens 1988; 6:
S458–S460.
35 Zoller WG, Hermans H, Bogner JR, Hahn D. Middeke
M. Duplex sonography in the diagnosis of renovascu-
lar hypertension. Klinische Wochenschrift 1990; 68:
830–834.
36 Postma CT, Aalen van J. Boo de T, Rosenbusch G,
Thien Th. Doppler ultrasound scanning in the detec-
tion of renal artery stenosis in hypertensive patients.
Br J Radiol 1992; 65: 857–860.
37 Dondi M et al. Evaluation by captopril renal scintigra-
phy and echo-Doppler flowmetry of hypertensive
patients at high risk for renal artery stenosis. J Nuclear
Biol Med 1992; 36: 309–314.
38 Spies KP et al. Color-coded duplex sonography for
non-invasive diagnosis and grading of renal artery
stenosis. American J Hypertens 1995; 8: 1222–1231.
39 De Cobelli F et al. Renal artery stenosis: prospective
comparison of color Doppler US and breath-hold,
three-dimensional, dynamic, gadolinium-enhanced
MR Angiography. Radiology 2000; 214: 373–380.
40 Moses LE, Shapiro D, Littenberg B. Combining inde-
pendent studie of a diagnostic test in a summary ROC
curve: data-analytic approaches and some additional
considerations. Stat Med 1993; 12: 1293–1316.