Blood Velocity Profiles in the Human Renal Artery by Doppler Ultrasound and Their Relationship to Atherosclerosis
Abstract Blood velocity profiles were measured in the renal branch (diameter 5.9±1.3 mm) of the aortorenal bifurcation using a 20-MHz 80-channel pulsed Doppler velocimeter during retroperitoneal surgery in 10 patients. The peak Reynolds number was 1145±140 and the frequency parameter (Wormersley parameter) was 3.0±0.8. Immediately distal to the ostium of the renal artery, reverse flow, indicating flow separation, was observed near the cranial wall mainly during the first part of the cardiac cycle. There were flows from the cranial to the caudal side of the artery at this location, indicating the presence of strong secondary flows. Two diameters downstream of the ostium, the velocity profiles were skewed to the caudal side in all patients. Four diameters downstream, the flow profile was symmetrical (3 patients) or only slightly skewed (7 patients) and virtually parabolic throughout the cardiac cycle. These observations mean that the flow on the cranial side of the renal branch of the human aortorenal bifurcation is characterized by (1) a bidirectional oscillation of the flow, (2) separation of the flow during systole, and (3) low time-averaged shear rate. These blood velocity patterns may be related to the localization and development of atheromatous plaque that occurs preferentially in this region of the renal artery. Conversely, the unidirectional, axisymmetrical flow found in more distal parts of the renal artery are associated with a very low incidence of lesions.
- Received June 15, 1995.
- Accepted October 16, 1995.
In the renal artery of humans, atherosclerotic plaque occurs primarily in the region 1 to 2 cm downstream from the renal ostium.1 2 3 When plaque occurs and develops sufficiently to disturb renal blood flow and reduce renal perfusion pressure, renal hypertension can occur, making plaque development in this region very important clinically. According to the response-to-injury hypothesis for atherosclerosis, many different risk factors can lead to endothelial dysfunction,4 and a large number of growth factors, cytokines, and vasoregulatory molecules may participate in its development.5
It is widely recognized that hemodynamic factors play an important role in vascular biology and in the localization and development of atherosclerotic lesions. Despite its importance, the characteristics of blood flow near the renal ostium have been studied only in model tube flows6 7 and by a computer simulation.8 Recently, advances in Doppler ultrasound techniques have enabled us to measure velocity profiles in vessels up to ≈7 mm in diameter.9 10 11 Using a high-frequency, range-gated Doppler ultrasound device, we have measured the velocity profiles at various sites around the aortorenal bifurcation in dogs.12 Dogs, however, do not develop atherosclerosis spontaneously, although atheromatous lesions can be produced by diets with very high lipid levels or by hypothyroidism.13
The high susceptibility of the ostial region of the renal artery in humans1 2 3 and the very low susceptibilities of more distal regions provide the opportunity for assessing which hemodynamic features are associated with lesion development. In particular, distinctions can be drawn between the effects of reversing and nonreversing blood flow close to the vessel wall. To characterize the flow patterns, we have used high-resolution multigated Doppler ultrasound to measure the instantaneous velocity profiles throughout the cardiac cycle at three different locations near the ostium. Because of the complexity of the local hemodynamics at the bifurcation, it was necessary to make measurements at different beam angles to the axis of the renal artery to determine the different components of the velocity vectors.
We studied a total of 10 patients (7 men, 3 women) who were admitted to the Department of Urology, Kawasaki Medical School (see Table 1⇓). Their clinical diagnoses were renal cancer (4 men, 3 women), retroperitoneal lymph node metastases of testicular cancer (2 men), and neuroblastoma (1 man). None of the patients had abnormal renal function tests, hypercholesterolemia, or thoracic or abdominal arterial calcification. The geometry of the aortorenal bifurcation was examined by DSA, CT, and/or MRI. In none of the patients was an atherosclerotic change in the aortorenal bifurcation region discernible.
The investigations were approved by the Medical Ethics Committee of the Kawasaki Medical School, and informed written consent was obtained from all of the patients studied. During the abdominal surgery, none of the patients had bleeding for which blood transfusion was required. Radical nephrectomy, enucleation of tumor, or retroperitoneal lymph node dissection was performed under modified neuroleptoanesthesia (with pentazocine and diazepam) because of its minimal influence on renal flows.14 This anesthesia was induced by thiopental (4 mg/kg) followed by succinylcholine chloride (1 mg/kg). After tracheal intubation, pancuronium (4 mg), pentazocine (30 mg), and diazepam (10 mg) were administered. The lungs were ventilated mechanically with a mixture of nitrous oxide and oxygen (FIO2=0.3). Operation postures were supine position (3 patients) or left or right side up (7 patients). The angle of the operating table for the side-up position was 30° to 40° from a horizontal plane. The angle of the aortorenal bifurcation was assessed from the DSA, CT, and MRI scans, and the blood velocities were obtained before nephrectomy, enucleation, or dissection. The intraoperative velocity measurements were accomplished within 10 minutes without any disturbance of the operative procedure. The blood pressure was measured in the left renal artery, and the electrocardiogram was monitored in lead II.
Blood velocity profiles were measured with a high-frequency (20 MHz) pulsed Doppler velocimeter, which was developed in the Department of Medical Engineering, Kawasaki Medical School, in collaboration with Fujitsu Lab Co and has been described previously with detailed validation of the velocity patterns it measures.9 10 11 In brief, this device measures the Doppler shift from 80 channels with sample volumes with a radius of ≈0.5 mm and a thickness of ≈0.2 mm. Being extravascular, the measuring probe does not influence the intravascular blood flow. The Doppler signals are analyzed simultaneously by a zero-cross method and by FFT. The high-pass filter had a cutoff frequency of 375 Hz, which corresponds to a velocity of 1.5 cm s−1 with a probe angle of 60° to the vessel axis.
In all the patients, the velocity profiles were measured at four sites in the renal artery, as indicated in Fig 1⇓. For comparison of results in different patients, all measurements were made relative to D measured in each patient just downstream from the ostium from the DSA, CT, or MRI scans. Site 1 was 0.5 D, site 2 was 2 D, and site 3 was 4 D distal to the ostium, where the artery was somewhat narrower. Measurements were made at all three sites with the probe placed on the cranial side of the vessel at an angle of 60° to the axis of the vessel. We called the main forward flow component in the first part of the cardiac cycle the systolic flow wave and the smaller flow component in the latter part the diastolic flow wave (see Figs 2⇓, 3⇓, and 4⇓), although there was some time delay from cardiac systole and diastole due to pressure wave propagation. To determine the direction of the velocity in the more complex flows that were observed at site 1, measurements were also made with the probe at 90° to the vessel axis (site 1P). The exposed vessels were bathed in ultrasound gel, and the probe was held at the required angle with a probe holder. The lateral position of the probe was adjusted until the velocity profiles were of maximum width, indicating that the ultrasound beam was traversing the midline of the vessel. Care was taken not to alter the geometry of the vessels being measured.
Two hydrodynamic parameters that characterize the flow are the Reynolds number (Re) and frequency parameter (α), which can be calculated as follows:
where D is the diameter of the vessel, Up the peak velocity, ω the frequency of the cardiac cycle, and ρ the density and μ the coefficient of the viscosity of blood. These parameters were calculated for each patient, the Reynolds number being calculated for peak velocity measured at site 2. This was in the single channel displaying the maximum velocity. The Reynolds number based on the peak velocity rather than the more common mean velocity is quoted because it provides an upper bound for these highly pulsatile flows and is therefore more useful in assessing the probability of turbulent flows during the whole of the cardiac cycle.
Bifurcation Geometry and Hemodynamic Parameters
The proximal portion of the renal artery lies more ventral than the abdominal aorta and then passes to the most dorsal surface of the abdominal cavity. The diameter of the abdominal aorta, measured just upstream of the aortorenal junction, ranged from 18 to 24 mm. The diameter of the renal artery, measured half a diameter distal to the ostium, ranged from 3.0 to 7.0 mm. The angle between the aorta and the renal artery ranged from 60° to 90°. In the whole patient group, the mean blood pressure was 96±10 mm Hg, the mean peak Reynolds number was 1145±140, and the mean frequency parameter was 3.0±0.8. Turbulent flows are unlikely to be found at these Reynolds numbers and frequency parameters.15 All of the results for the geometry and hemodynamic parameters are given in Table 2⇓.
Blood Velocity Profiles
Fig 2⇑ shows the velocity profiles measured in one patient at site 1, immediately distal to the orifice of the renal artery. Site 1 refers to measurements taken with the probe at 60° to the vessel. The upper panel is the velocity as a function of time, determined from the FFT of the Doppler signal. The lower panel shows velocity profiles across the vessel, calculated by the zero-cross method measured at 20-millisecond intervals during the portion of the waveform indicated by the vertical lines. The inbuilt algorithm used by the ultrasound device to compute the velocities is based on the assumption that the velocity vectors are parallel to the direction of the vessel’s axis. We have attempted to overcome this problem by using the technique of vector addition of signals obtained with the probe oriented at different angles to the vessel axis. The FFT velocity waveform was obtained from the channel containing the maximum velocity (as assessed by the zero-cross method) at each measuring time. The velocity waveforms indicate that there is continuous forward flow throughout the cardiac cycle. The velocity profiles showed that in all the patients, reverse velocities were present at the cranial side of the vessel during the period of systolic flow wave and also during the period of diastolic flow wave, but the duration of reverse flow in a cardiac cycle at the cranial side wall was variable. The maximum reverse/forward velocity ratio, defined as the ratio of the maximum value of reverse velocities across the vessel in a cardiac cycle to the maximum value of forward velocities, for 10 patients was 0.3±0.1. In most of the period of the diastolic wave, flow was maintained in the forward direction, though velocities were below threshold close to both walls.
The velocity vector measured at site 1P with the probe perpendicular to the vessel had a component directed from the cranial to the caudal side of the vessel, particularly during the period of the systolic flow wave, indicating the presence of secondary flow. An example is shown in the upper left panel of Fig 5⇓. This flow component toward the caudal direction occurred between the cranial wall and the point within the vessel where the peak velocity was measured at site 1. It became smaller and reversed near the caudal wall. Vector addition of the velocities measured at site 1 (V1) and site 1P (V1P) gives the two-dimensional velocity vector (V) in the medial plane of the vessel at this site (upper right panel of Fig 5⇓). The technique for vector addition and the assumptions made have been previously described.12 These vectors, calculated at the beginning of the systolic wave, at the time of peak forward flow, and in the early and late diastolic waves (A, B, C, and D in Fig 5⇓), are shown in the bottom panel of Fig 5⇓. A rotation of the flow, typical of separated flows, is apparent, and the velocities are skewed, mostly toward the caudal wall.
The velocity waveform and velocity profiles measured at site 2, 2 D downstream of the bifurcation, are shown in Fig 3⇑. The velocity is still greatest on the caudal side of the vessel in all patients, but the reversed flow near the cranial side wall seen at site 1 was not detectable with our cutoff frequency. The results measured at site 3, 4 D downstream of the bifurcation, are shown in Fig 4⇑. The velocity profiles are much more axisymmetrical; a slight skewing toward the caudal side was apparent during the period of peak systolic flow wave in seven patients, but the other three patients showed a completely symmetrical pattern. Two of seven patients with skewed velocity profiles at site 3 were in the supine position and the other five patients were side up. One of three patients with symmetrical profiles was supine and the other two were side up. Accordingly, the patient’s position did not seem to affect the results, although the number of data was limited. There was no noticeable difference in the velocity profiles at site 3 among the seven patients with renal cancer (two symmetrical and five slightly skewed) and the three without (one symmetrical and two slightly skewed).
Early atherosclerotic lesions tend to be found at bifurcations and curving regions of arteries, and it has been suggested that hemodynamic factors play an important role in their development. The current study was undertaken to investigate this problem by measuring the blood velocity profiles, with high spatial resolution, of the renal arteries of humans during retroperitoneal surgery. The measurements were necessarily performed on a group of surgical patients with renal abnormalities. Those with discernible atherosclerosis were excluded. The similarity of the velocity profiles in patients presenting with markedly different clinical conditions suggested that these particular abnormalities were not important in modifying the hemodynamics. The major findings of this study are that the velocity pattern near the cranial wall at the renal ostium, at which atherosclerotic lesions are prone to develop, is characterized by (1) separation of the flow mainly in systole, with velocity vectors directed toward the caudal wall and (2) a time-varying oscillation of flow. Both of these conditions will give rise to a low time-averaged shear rate at this site.
There are many reports on the occurrence of atherosclerosis in the renal artery, although the frequency of occurrence of lesions in this vessel appears to be much less than in the aorta or coronary arteries.16 17 Most reports lack precision about the location of the lesions within the vessel, but DeBakey et al18 found that in patients requiring operation for atherosclerotic obstruction of the renal artery, 74% of the lesions were localized at a proximal site close to the renal ostium. Yutani et al3 found that in vessels from subjects who had hypercholesterolemia, intimal thickening and foam cell infiltration were more prominent in the proximal wall than in the more distal wall in a variety of ostia, including that of the renal artery. More recently, Nguyen and Haque19 carried out a survey of lesion distribution in the abdominal aorta and its major branches in 18 human autopsy specimens. The autopsy specimens were chosen by the absence of diffuse atherosclerotic involvement, which might mask the focal distribution of the disease, and the number of lesions in each of 35 different segments throughout this region was reported. Plaques were found on the cranial wall of the renal artery, within one diameter of the ostium, in 14 of 18 specimens on the left side and in 12 of 18 on the right side. The incidence of lesions at these locations was among the highest of all the 35 sites studied. On the caudal wall of the renal artery, within one diameter of the ostium, the incidences were 7 of 18 and 8 of 18 on the left and right sides, respectively. In keeping with earlier reports on the paucity of renal artery lesions, this study showed plaques further downstream in the vessels in only 1 of 18 cases on the left side and 1 of 18 on the right side.
The variation in velocity patterns close to the vessel wall found in this study may be related to these pathological observations. The site at which lesions have been most commonly observed, the cranial wall of the proximal regions, is that which showed a transient region of flow separation in all 10 patients. In the early part of the cycle, velocities close to the wall, and therefore the local wall shear stresses, were in the reverse direction, while in the remainder of the cycle, velocity and shear stress were in the forward direction. At the caudal wall of the proximal region, where lesions occur less frequently, there were strong secondary flows in all patients, indicating that the shear stress would have a circumferential component and then change in both magnitude and direction during a cardiac cycle. In the distal part of the vessel, the most notable feature was the axisymmetry and unidirectionality of the profiles throughout the cycle. In pathological studies, this region has been found to be remarkably free from atherosclerotic lesions. The velocity patterns at this distal site are in marked contrast to the reversing patterns of flow seen in various parts of the aorta and particularly in the epicardial coronary arteries, sites where the prevalence of lesions is very much greater. The local dependence of lesion development and the presence of variation in direction of shear found in the renal artery appear similar to that found in the carotid sinus.20
From this study, it would appear that blood flow separation, with its attendant reversing shear stress, may be a strong correlate of atherogenesis, lesions being more prevalent at sites where there is flow separation than at those with unidirectional flow. We have not attempted to quantitate the wall shear stresses applied to the wall from the velocity profiles, because even with the very high resolution of our technique, there is still uncertainty about the position of the vessel wall, which moves during the cardiac cycle. Inspection of the velocity profiles, however, indicates that shear rates close to the wall and hence shear stresses are likely to be higher on the caudal wall of the proximal region of the renal artery than at more distal sites. The lower incidence of atherosclerosis in the more distal regions suggests that low shear without flow reversal may be associated with fewer lesions, while higher shear stresses may not be completely protective.
Previous authors have drawn attention to the potential importance of branching angle.21 Nguyen and Haque19 tried to correlate the distribution of lesions in the different specimens to the branching angle between the renal arteries and the abdominal aorta. They found a weak inverse correlation between the angle and the incidence in the part of the aortic wall between the renal artery and the aortoiliac bifurcation but did not draw attention to any correlation between that angle and incidence of lesions in the renal artery itself. Our patients exhibited a range of angles between 60° and 90°, but this did not appear to alter the main features of the velocity profiles. These features were also retained in an earlier series of experiments on dogs when the aortorenal angle was actively changed between 90° and 30°.12 The angle of the aortorenal bifurcation is altered by respiratory movements of the diaphragm, by changes in posture, and during movement, but our results would suggest that these factors would not markedly alter the local shear forces.
These studies on human patients shed further light on the relationship between hemodynamic factors and atherosclerosis. Since the lesions within the renal artery appear to be highly localized, it is possible to demonstrate that flow separation, which causes a transient reversal of wall shear stresses, appears to be the most important causative factor, while unidirectional flow appears protective. It has been suggested that the stagnation of blood, which is associated with regions of flow separation, may be an important factor in this association.2 22 23 Our study shows, however, that the separation zone persists only for about one third of the cardiac cycle, so that the residence time of blood components close to the wall is relatively short. Even in regions with no flow separation or reversal, such as in the more distal regions of the renal artery, where lesions are rarely found, the velocity profiles suggest that there may be stagnation of blood close to the vessel wall during the latter part of the cycle because of the very low axial velocity.
Selected Abbreviations and Acronyms
|D||=||diameter of the proximal portion of the renal artery|
|DSA||=||digital subtraction angiography|
|FFT||=||fast Fourier transform|
|MRI||=||magnetic resonance imaging|
This research was supported by grant-in-aid 05454278 for General Scientific Research (B) and 06671366 (C) from the Ministry of Education, Science and Culture, Japan. We thank Chikako Tokuda and Mieko Hayashi for their assistance in manuscript preparation.
Osmundson PJ, Bernatz PE. Occlusive disease of abdominal visceral arteries. In: Juergens JL, Spittell JA Jr, Fairbairn JF II, eds. Peripheral Vascular Disease. 5th ed. Philadelphia, Pa: W B Saunders Co; 1980:295-325.
Yutani C, Imakita M, Ishibashi-Ueda H, Yamamoto A, Takaichi S. Localization of lipids and cell population in atheromatous lesions in aorta and its main arterial branches in patients with hypercholesterolemia. In: Yoshida Y, Yamaguchi T, Caro CG, Glagov S, Nerem RM, eds. Role of Blood Flow in Atherogenesis. Tokyo, Japan: Springer-Verlag; 1988:25-31.
Liepsch D, Poll A, Moravec S. Flow studies in true-to-scale models of human renal arteries. In: Yoshida Y, Yamaguchi T, Caro CG, Glagov S, Nerem RM, eds. Role of Blood Flow in Atherogenesis. Tokyo, Japan: Springer-Verlag; 1988:91-96.
Ogasawara Y, Hiramatsu O, Kagiyama M, Tsujioka K, Tomonaga G, Kajiya F, Yanashima T, Kimura Y. Evaluation of blood velocity profile by high frequency ultrasound pulsed Doppler velocimeter by a multigated zerocross method together with a Fourier transform method. IEEE Comput Cardiol. 1984;84:447-450.
Kajiya F, Tsujioka K, Ogasawara Y, Wada Y, Matsuoka S, Kanazawa S, Hiramatsu O, Tadaoka S, Goto M, Fujiwara T. Analysis of flow characteristics in poststenotic regions of the human coronary artery during bypass graft surgery. Circulation. 1987;76:1092-1100.
Fujiwara T, Kajiya F, Kanazawa S, Matsuoka S, Wada Y, Hiramatsu O, Kagiyama M, Ogasawara Y, Tsujioka K, Katsumura T. Comparison of blood-flow velocity waveforms in different coronary artery bypass grafts: sequential saphenous vein grafts and internal mammary artery grafts. Circulation. 1988;78:1210-1217.
Yamamoto T, Tanaka H, Jones CJH, Lever MJ, Parker KH, Kimura A, Hiramatsu O, Ogasawara Y, Tsujioka K, Caro CCG, Kajiya F. Blood velocity profiles in the origin of the canine renal artery and their relevance in the localization and development of atherosclerosis. Arterioscler Thromb. 1992;12:626-632.
Mazze RF. Renal physiology and the effects of anesthesia. In: Miller RD, ed. Anesthesia, Vol 2. 2nd ed. New York, NY: Churchill Livingstone; 1986:1223-1248.
Nerem RM, Seed WA. An in vivo study of aortic flow disturbances. Cardiovasc Res. 1972;6:1-14.
Roberts JC Jr, Moses C, Wilkins RH. Autopsy studies in atherosclerosis, I: distribution and severity of atherosclerosis in patients dying without morphologic evidence of atherosclerotic catastrophe. Circulation. 1959;20:511-519.
Zarins CK, Giddens DP, Bharadvaj BK, Sottiurai VS, Mabon RF, Glagov S. Carotid bifurcation atherosclerosis: quantitative correlation of plaque localization with flow velocity profiles and wall shear stress. Circ Res. 1983;53:502-514.