Diagnostic role of new Doppler index in assessment of renal artery stenosis
© Chain et al; licensee BioMed Central Ltd. 2006
Received: 02 December 2005
Accepted: 25 January 2006
Published: 25 January 2006
Renal artery stenosis (RAS) is one of the main causes of secondary systemic arterial hypertension. Several non-invasive diagnostic methods for RAS have been used in hypertensive patients, such as color Doppler ultrasound (US). The aim of this study was to assess the sensitivity and specificity of a new renal Doppler US direct-method parameter: the renal-renal ratio (RRR), and compare with the sensitivity and specificity of direct-method conventional parameters: renal peak systolic velocity (RPSV) and renal aortic ratio (RAR), for the diagnosis of severe RAS.
Our study group included 34 patients with severe arterial hypertension (21 males and 13 females), mean age 54 (± 8.92) years old consecutively evaluated by renal color Doppler ultrasound (US) for significant RAS diagnosis. All of them underwent digital subtraction arteriography (DSA). RAS was significant if a diameter reduction > 50% was found. The parameters measured were: RPSV, RAR and RRR. The RRR was defined as the ratio between RPSV at the proximal or mid segment of the renal artery and RPSV measured at the distal segment of the renal artery. The sensitivity and specificity cutoff for the new RRR was calculated and compared with the sensitivity and specificity of RPSV and RAR.
The accuracy of the direct method parameters for significant RAS were: RPSV >200 cm/s with 97% sensitivity, 72% specificity, 81% positive predictive value and 95% negative predictive value; RAR >3 with 77% sensitivity, 90% specificity, 90% positive predictive value and 76% negative predictive value. The optimal sensitivity and specificity cutoff for the new RRR was >2.7 with 97% sensitivity (p < 0.004) and 96% specificity (p < 0.02), with 97% positive predictive value and 97% negative predictive value.
The new RRR has improved specificity compared with the direct method conventional parameters (RPSV >200cm/s and RAR >3). Both RRR and RPSV show better sensitivity than RAR for the RAS diagnosis.
Renovascular hypertension might account for 1–5% of all cases of hypertension [1, 2]. However, it affects 15–30% of patients who have clinical criteria suggestive of renovascular hypertension as refractory hypertension to an appropiate three-drug treatment associated either with moderate impairment of renal function or with carotid, peripheral or coronary atherosclerotic disease. The renovascular hypertension term focuses on the causal relationship between RAS and its clinical consequences (hypertension and renal failure) . Atherosclerosis accounts for 90% of cases of RAS, and usually involves the ostium and the proximal segment of the main renal artery and the perirenal aorta. Fibromuscular dysplasia accounts for less than 10% of cases of RAS. Fibromuscular dysplasia tends to affect girls and women who are between 15 and 50 years of age, and frequently involves the distal two segments of the renal artery and its branches . Atherosclerotic RAS is a common and progressive disease. Its prevalence increases with age, particularly in patients who have diabetes, aortoiliac occlusive disease, coronary artery disease or hypertension . Owing to improvements in the techniques used for screening, it is now recognized that 40–50% of patients with occlusive disease of the lower limb and 15–30% of patients with coronary artery disease have recognizable RAS [5, 6]. Reports of end-stage renal disease indicate a prevalence of RAS of 10–22% [7, 8].
A diagnosis of atherosclerotic RAS can be suspected on the clinical grounds mentioned, but can only be established through specific diagnostic procedures [4, 9, 10]. Because renovascular hypertension could by treated by percutaneous transluminal angioplasty, endovascular stent placement or surgical revascularization, several non-invasive methods have been advocated to screen for suspected renovascular disease [11, 12]. Renal color Doppler ultrasound (US) has been proposed as an effective modality for the diagnosis of RAS. The use of Doppler US in patients with hypertension has led to an increase in the diagnosis of RAS [13, 14].
The direct method Doppler parameters evaluate the renal artery peak systolic velocity (RPSV) and the aortic peak systolic velocity to determine the maximal RPSV and the renal aortic ratio (RAR). The new direct method Doppler parameter, the renal renal ratio (RRR), was defined as the rate between RPSV at the proximal or mid segment of the renal artery and RPSV measured at the distal segment of the renal artery.
The indirect method Doppler parameters evaluate the post stenostic "tardus-parvus" phenomenon in the intrarenal arterial Doppler waveforms . We did not evaluate this indirect method parameters in this research.
The aim of this study was to assess the sensitivity and specificity of a new renal Doppler US direct method parameter, the renal renal ratio (RRR), and to compare it with the sensitivity and specificity of another direct method conventional parameters, renal peak systolic velocity (RPSV) and renal aortic ratio (RAR), for the diagnosis of severe RAS.
Clinical features suggestive of renal artery stenosis
Epigastric or flank bruit (systolic or diastolic)
Accelerated or malignant hypertension
Unilateral small kidney discovered with any clinical study
Severe hypertension in a child or young adult or aged more than 50 years
Sudden development or worsening of hypertension at any age
Hypertension and unexplained impairment of renal function
Sudden worsening of renal function in a hypertensive patient
Hypertension refractory to an appropriate three-drug regimen
Impairment of renal function after treatment with an ACE inhibitor
Hypertension and extensive arterial occlusive disease (peripheral vascular and coronary arterial disease)
ACE: angiotensin-converting enzyme
Three signs were found in 11 patients, 4 signs in 10 and ≥5 signs in 13 patients
Demographic characteristics of the patients.
N = 34
54 (± 8.92)
30 kg/m2 (+/-5)
RRR = RPSV (proximal or mid renal artery)/RPSV (distal renal artery)
We have only evaluated the main renal arteries, we did not use the RRR to analize the accessory arteries.
Renal Doppler US was performed in our departments using a Toshiba Nemio 30 machine with low-frequency curve transducers (3–6 MHz) and low-frequency sectorial transducers (2–4 MHz) to allow greater penetration of the US beam.
All examinations were performed by a single physician with more than five-years experience in vascular sonography. The examinations were performed in the morning if possible and the patients were recommended to observe a 10-hour fasting period beforehand. The complete procedure was generally completed within approximately 30–35 minutes. All examinations were started with the patients in the supine position, in order to visualize the abdominal aorta and the origin and proximal course of the main renal artery. The patients were then kept recumbent in various different positions: the left decubitus position to explore the right renal artery, the ventral decubitus position to explore the left renal artery or the right decubitus position in a few cases.
The maximum diameter and the peak systolic velocity in the abdominal aorta were obtained in a longitudinal section of its proximal segment.
Epigastric transverse scans allowed us to identify the main renal arteries, which originate laterally (anterolaterally for the right artery and posterolaterally for the left artery) in the abdominal aorta about 1 cm under the emergence of the superior mesenteric artery. The right and left renal arteries also have a proximal course just posterior to the left renal vein under normal conditions. Therefore, the origin and the proximal segment of the renal arteries were initially assessed in a transverse section of the abdominal aorta in the epigastric region. Every patient had all segments of bilateral renal arteries interrogated: proximal, mid and distal. Depending on the anatomy of the patient, almost the entire length of the renal arteries could be assessed.
We performed direct imaging of the main renal artery with the color window and pulsed wave (PW) Doppler. When the renal artery image showed homogeneous color and standard velocity, we diagnosed normal renal artery patency.
For the diagnosis of RAS using the conventional technique, we performed direct imaging of the main renal artery. Color imaging revealed the presence of stenosis through signs of turbulence in the systolic phase, which was generally yellow and green. Subsequently, the PW Doppler was positioned in the area of interest in the center of the vessel, at the site identified by increased flow velocities and turbulence. The gain, filter and scales of the velocity curves of the PW Doppler were adjusted to provide an adequate curve for measuring the velocities. When the renal artery PW Doppler showed increased peak systolic velocity, we suspected RAS. The highest value of the three measurements of RPSV performed from different views was chosen in each case.
For the abdominal aorta and main renal arteries, the angle of incidence of the pulsed Doppler was maintained less than 60°, which was essential in order to obtain accurate measurements. We used the same or very similar angle correction at proximal, mid and distal segments of the renal artery. It was a very important technical detail. The velocities at proximal, mid and distal segments of the renal artery were assessed using obliqual sections.
Doppler samples were taken in apnoea, when the image of the artery was found to be optimal in the respiratory cycle.
We did not use contrast echo in case of difficult renal artery imaging.
Digital substraction angiography
Digital subtraction angiography was performed in all patients. This was performed using a 5-F pigtail catheter, with the tip positioned through the right or left femoral artery just proximal to the renal arteries. A non-ionic contrast material was injected and the images were obtained in the anteroposterior, left anterior oblique and right anterior oblique projections. The criterion for anatomically significant RAS at angiography was 50% or greater renal artery narrowing. The stenosis was assesed by calipers.
The sensitivity, specificity, positive predictive value and negative predictive value for the detection of significant RAS were calculated independently for the three parameters: RPSV, RAR and RRR, and its sensitivity and specificity were compared. The sensitivity for detecting stenosis was calculated as the number of true-positive findings according to color Doppler US divided by the number of positive findings by DSA. The specificity was calculated as the number of true-negative findings according to color Doppler US divided by the number of negative findings by DSA.
Additionally, receiver operating characteristic (ROC) curves were computed to compare the parameters, which provided a graphic description of the performance of the tested variables towards RAS detectability. The curves were generated from data obtained through sensitivity/specificity analysis of the variables.
The chi square test was used to evaluate the difference among the three direct method color Doppler US parameters. All statistical tests were two-tailed and performed at the 0.05 level of significance. The confidence intervals were two-sided with 95% intervals. The SPSS 10.0 statistical package was used for all the calculations. The protocol was approved by our institutional ethics committee and all the patients provided written informed consent.
N° of pts
RA without stenosis
RA with stenosis
RA with total oclusion
Pts without RAS
Pts with unilateral RAS
Pts with bilateral RAS
Sensitivity and specificity of the new RRR
Cutoff values for RRR sensitivity and specificity
Sensitivity and specificity of the conventional direct method parameters
The sensitivity and specificity of the conventional direct method parameters (RPSV and RAR) were evaluated. We found that RPSV >200 cm/s and RAR >3 were indicative of severe RAS as it is referred in the literature . RPSV >200 cm/s resulted in a sensitivity of 97%, a specificity of 72%, a positive predictive value of 81% and a negative predictive value of 95% in terms of the diagnostic accuracy for RAS. A severe RAS diagnosis with RAR >3 yielded a sensitivity of 77%, a specificity of 90%, a positive predictive value of 90% and a negative predictive value of 76%.
Sensitivity and specificity of the direct parameters
RPSV > 200 cm/s
RAR > 3
RRR > 2.7
The discriminatory capacities of the direct method parameters applied with color Doppler US were evaluated based on their statistical significance, as well as on the results of the sensitivity and specificity analysis.
Sensitivity and specificity of the direct method parameters
Hoffman et al. used an RPSV >180 cm/s to discriminate RAS with a sensitivity of 95% and a specificity of 90% . Souza de Oliveira et al. showed a sensitivity of 83.3%, a specificity of 89.5% and a cutoff value of approximately 150 cm/s . In our population, the best cutoff value was 200 cm/s with a sensitivity of 97% and a specificity of 72%.
Hoffman et al. using an RAR of 3.5 identified a RAS >60% with a sensitivity of 92% and a specificity of 62% . Similar values of sensitivity (ranging from 92 to 95%) and higher specificity results (ranging from 88 to 90%) were obtained by others authors, such as Rabbia et al. and Conkbayir et al. [17, 18]. In Souza de Oliveira's study population, the best cutoff value for this index was considered to be approximately 1.8, which was much less than the 3.5 cutoff value, and showed satisfactory results for sensitivity (83.3%) and a low specificity (78.9%) . In our study population, the best cutoff value was 3, with a sensibility of 77% and a specificity of 90%. These variable results are probably related to differences in the estimated cutoff values and the different angiographic degrees of stenosis (ranging from 50 to 70%) that were considered to be significant by some previous authors [14, 16–18].
The new RRR
Our new proposed diagnostic index, RRR, is based on the fact that increased blood flow velocity across the stenosis and the immediate post-stenotic segments, and the observed decrease in blood flow velocity distal to the stenosis is proportional to the degree of stenosis [14, 19].
In our current study, the RRR was particularly usefull because produced specificity results for RAS diagnosis that were superior to those of RPSV. RRR uses comparative velocity and it increases the diagnostic assurance to the operator. An advantage of our proposed index over RAR was that the diameter of the main renal artery in the distal segment was similar to those of the proximal and middle segments; by contrast, RAR uses the abdominal aorta, which has a much larger diameter. This enables a better comparison of the flow velocity between proximal and distal segments .
The most commonly used indirect method parameters are the systolic acceleration ratio, the acceleration time, and the intraparenchymal resistance ratio difference between the right and left kidneys [14, 21]. We did not use these indirect method parameters in our research because they do not show better results than direct method parameters in the literature . Unlike the indirect methodologies, the new index RRR is not affected by changes in parenchymal stiffness, not based on the waveform and allows beam-angle corrections. Souza de Oliveira et al. showed that the renal segmental ratio (RSR), which was defined as the proximal RPSV divided by the segmental artery peak systolic velocity, was able to predict the presence of significant stenosis at the main renal artery. The RSR showed satisfactory results for sensitivity (83.3%) and low specificity (78.9%) for all renal units. This study analyzed intrarenal segmental velocities that were subjeted to more variable factors, particularly the elastic properties of the arterial vessel walls, changes in peripheral resistance within renal vascular circuits and changes in parenchymal stiffness . In our current study, the RRR results for sensitivity and specificity were 97 and 96%, respectively. These data might be explained by the fact that we employed the RPSV in the extraparenchymal segment, which was not subjet to variability in the factors mentioned above .
Van der Hulst et al. found that the ratio between the maximum peak systolic velocity at the renal artery and the minimum value at the arcuate artery, analysed using an endovascular procedure, gave good results for RAS diagnosis (a sensitivity of 94% and a specificity of 100%). However, their study used an invasive approach and had a higher risk than the RRR proposed in our current research .
Our best cutoff value for RRR was 2.7 in accordance with an angiographic RAS >50%; however, we found an angiographic RAS >50% with an RRR >2.5. We did not find RAS when the RRR was <2, so we proposed this cutoff to be the normal value.
Body mass index (BMI) was calculated for each patient. The mean BMI was 30 (± 5) kg/m2. (Table 2). The 97% of our patients were successfully examined with color Doppler US, even in those with a high BMI.These favourable results might have been related to the strict bowel preparation and the fact that the examinations were performed by a trained specialist.
This index was assessed in all patients with only one exception. This patient had RAS at the end of the distal renal artery segment, so we could not take a distal velocity to compare with the RPSV at the RAS. We therefore used to replace the distal renal artery segment velocity, the measurement at non-stenotic distal branch of the main renal artery at the extraparenchymal level.
The main limitation of this study was that we evaluated only the main renal arteries because these vessels have a more important role at the renovascular disease and were able to undergo endovascular treatment. Neither we compared the color Doppler US with other diagnostic non-invasive diagnostic methods as computed tomographic, magnetic resonance angiography or captopril renography for the diagnosis of RAS. We did not considerate these diagnostic methods in this study because their high cost and not widespread availability of all of them may somehow avoid a routine use in clinical practice .
The new RRR, with a cutoff value >2.7, shows significantly improved specificity compared with conventional direct method Doppler parameters (RPSV > 200 cm/s and RAR >3). Both RRR and RPSV show better sensitivity than RAR for RAS diagnosis. The results indicate that RRR and RPSV were the best criteria for RAS diagnosis. The RRR improves the diagnostic effectiveness of the color Doppler US.
List of abbreviations
renal peak systolic velocity
renal aortic ratio
renal artery stenosis
- Renal color Doppler ultrasound =:
color Doppler US
digital subtraction arteriography
- ROC curve:
receiver-operating characteristic curve
- PW Doppler:
pulsed wave Doppler
We thank Perla Feldman, Carlos Castañeda, Juan José Ruiz, Alejandro Zerdán and Sebastián Pata for their valuable contributions.
- Olin JW, Melia M, Young JR, Graor RA, Risius B: Prevalence of atherosclerotic renal artery stenosis in patients with atherosclerosis elsewhere. Am J Med 1990, 88: 46N-51N. 10.1016/0002-9343(90)90262-CView ArticlePubMedGoogle Scholar
- Mann SJ, Pickering TG: Detection of renovascular hypertension: state of the art-1992. Ann Intern Med 1992, 117: 845-853.View ArticlePubMedGoogle Scholar
- Safian RD, Textor SC: Renal artery stenosis. N Engl J Med 2001, 344: 431-442. 10.1056/NEJM200102083440607View ArticlePubMedGoogle Scholar
- Crowley JJ, Santos RM, Peter RH, Puma JA, Schwab SJ, Phillips HR, Stack RS, Conlon PJ: Progression of renal artery stenosis in patients undergoing cardiac catheterization. Am Heart J 1998, 136: 913-918. 10.1016/S0002-8703(98)70138-3View ArticlePubMedGoogle Scholar
- Choudhri AH, Cleland JGF, Rowlands PL, Tran TL, Mc Carty M, Al-Kutoubi MA: Unsuspected renal artery stenosis in peripheral vascular disease. BMJ 1990, 301: 1197-1198.View ArticlePubMedPubMed CentralGoogle Scholar
- Uzu T, Inoue T, Fujii T, Nakamura S, Inenaga T, Yutani C, Kimura G: Prevalence and predictors of renal artery stenosis in patients with myocardial infarction. Am J Kidney Dis 2000, 29: 733-738.View ArticleGoogle Scholar
- Beutler JJ, Van Ampting JM, Van de Ven PJ, Koomans HA, Beek FJ, Woittiez AJ, Mali WP: Long-term effects of arterial stenting on kidney function for patients with ostial atherosclerotic renal artery stenosis and renal insufficiency. J Am Soc Nephrol 2001, 12: 1475-1481.PubMedGoogle Scholar
- Ramos F, Kotliar C, Alvarez D, Baglivo H, Rafaelle P, Londero H, Sanchez R, Wilcox C: Renal function and outcome of PTRA and stenting for atherosclerotic renal artery stenosis. Kidney Int 2003, 63: 276-282. 10.1046/j.1523-1755.2003.00734.xView ArticlePubMedGoogle Scholar
- Zucchelli PC: Hypertension and atherosclerotic renal artery stenosis: diagnostic approach. J Am Soc Nephrol 2002, 13: S184-S186. 10.1097/01.ASN.0000032547.12173.5EView ArticlePubMedGoogle Scholar
- De Cobelli F, Venturini M, Vanzulli A, Sironi S, Salvioni M, Angeli E, Scifo P, Garancini MP, Quartagno R, Bianchi G, Del Maschio A: Renal arterial stenosis: Prospective comparison of color Doppler US and breath-hold, three-dimensional, dynamic, gadolinium-enhanced MR angiography. Radiology 2000, 214: 373-380.View ArticlePubMedGoogle Scholar
- Blum U, Krumme B, Flugel P, Gabelman A, Lehnert T, Buitrago-Tellez C, Schollmeyer P, Langer M: Treatment of ostial renal artery stenoses with vascular endoprostheses after unsuccessful balloon angioplasty. N Engl J Med 1997, 336: 459-465. 10.1056/NEJM199702133360702View ArticlePubMedGoogle Scholar
- Radermacher J, Chavan A, Bleck J, Vitzthum A, Stoess B, Gebel MJ, Galansky M, Koch KM, Haller H: Use of Doppler Ultrasonography to predict the outcome of therapy for renal-artery stenosis. N Engl J Med 2001, 344: 410-417. 10.1056/NEJM200102083440603View ArticlePubMedGoogle Scholar
- Krumme B, Blum U, Schwertferger E, Flugel P, Hollstin F, Schollmeyer P, Rump LC: Diagnosis of renovascular disease by intra and extrarenal Doppler scanning. Kidney Int 1996, 50: 1288-1296.View ArticlePubMedGoogle Scholar
- Souza de Oliveira I, Widman A, Lazlo J, Fukushima J, Praxedes J, Cerri G: Colour Doppler ultrasound: a new index improves the diagnosis of renal artery stenosis. Ultrasound Med Biol 2000, 26: 41-47. 10.1016/S0301-5629(99)00119-2View ArticlePubMedGoogle Scholar
- Claudon M, Plouin PF, Baxter GM, Rohban T, Devos DM: Renal arteries in patients at risk of renal arterial stenosis: multicenter evaluation of the echo-enhancer SH U 500A at color and spectral Doppler US. Levovist Renal Artery Stenosis Study Group. Radiology 2000,214(3):739-746.View ArticlePubMedGoogle Scholar
- Hoffmann U, Edwards JM, Carter S, Goldman ML, Harley JD, Zaccardi MJ, Strandness DE Jr: Role of duplex scanning for the detection of atherosclerotic renal artery disease. Kidney Int 1991, 39: 1232-1239.View ArticlePubMedGoogle Scholar
- Rabbia C, Valpreda S: Duplex scan sonography of renal artery stenosis. Int Angiol 2003,22(2):101-115.PubMedGoogle Scholar
- Conkbayir I, Yucesoy C, Edguer T, Yanik B, Yasar Ayaz U, Hekimoglu B: Doppler sonography in renal artery stenosis. An evaluation of intrarenal and extrarenal imaging parameters. Clin Imaging 2003,27(4):256-260. 10.1016/S0899-7071(02)00547-8View ArticlePubMedGoogle Scholar
- Van der Hulst VP, Van Baalen J, Kool LS, van Bockel JH, van Erkel AR, Ilgun J, Pattynama PM: Renal artery stenosis: endovascular flow wire study for validation of Doppler US. Radiology 1996, 200: 165-168.View ArticlePubMedGoogle Scholar
- Otto CM, Pearlman AS, Gardner CL, Kraft CD, Fujioka MC: Simplification of the Doppler continuity equation for calculating stenotic aortic valve area. J Am Soc Echocardiogr 1988,1(2):155-157.View ArticlePubMedGoogle Scholar
- Stavros AT, Parker SH, Yakes WF, Chantelois AE, Burke BJ, Meyers PR, Schenck JJ: Segmental stenosis of the renal artery: pattern recognition of tardus and parvus abnormalities with duplex sonography. Radiology 1992, 184: 487-492.View ArticlePubMedGoogle Scholar
- Nelemans PJ, Kessels AG, De Leeuw P, De Haan M, van Engelshoven J: The cost-effectiveness of the diagnosis of renal artery stenosis. Eur J Radiol 1998, 27: 95-107. 10.1016/S0720-048X(97)00158-7View ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.