Noninvasive assessment of aortocoronary bypass graft patency using pulsed Doppler echocardiography

Noninvasive Assessment of Aortocoronary Bypass Graft Patency Using Pulsed Doppler Echocardiography

BENOIT DIEBOLD, MD* PIERRE THEROUX, MD, FACC’ MARTIAL G. BOURASSA, MD, FACC’ PIERRE PERONNEAU, PhDt JEAN-LEON GUERMONPREZ, MD+ Montreal, Quebec, Canada Paris, France

To evaluate noninvasively aortocoronary bypass graft patency, pulsed Doppler echocardiography was performed at the time of postoperative coronary angiography in 120 consecutive patients. Ultrasonic examination of 163 vein grafts was possible. One hundred twenty-seven patent and 14 occluded grafts were correctly identified. Eleven patent grafts could not be recorded, and 11 occluded grafts were falsely diagnosed as patent. The method had an overall sensitivity of 92 percent and a specificity of 56 percent. This high sensitivity level may be increased to almost 100 percent by enhanced technical skill and experience. The low specificity level, although the method must be tested in a larger number of bypass grafts, stresses the importance of correctly identifying other sources of diastolic blood flow. Diastolic flows from the superior vena cava, internal mammary veins, tricuspid valve, mitral valve and right ventricle may be eliminated by careful adjustment of the depth, site and size of the pulsed Doppler electronic sampling gate. Standard echocardiographic landmarks for avoiding confusion with the coronary arteries are also described.

The only currently reliable method for assessing aortocoronary bypass graft patency is cardiac catheterization and angi0graphy.l A noninvasive technique allowing transcutaneous evaluation of graft patency would be invaluable for clinical diagnosis and long-term follow-up of graft outcome. The pulsed Doppler flowmeter was described in 19692 and was applied to aortocoronary bypass graft detection by Gould et al.3 in 1972. In this study we describe a method of graft flow recording using an external pulsed Doppler flowmeter and evaluate its reliability by comparing the results obtained with those of angiography. Method Doppler Principle and Instrumentation

From the Department of Medicine, Montreal Heart Institute, University of Montreal Medical School, Quebec, Canada,* and Clinique Cardiologiquer and Centre National des Recherches Scientifiques,t Hopital Broussais, Paris, France. This work was supported in part by the J. L. Levesque Foundation. Manuscript received May 2, 1978; revised manuscript received August 1, 1978, accepted August 16, 1978. Address for reprints: Pierre Theroux, MD, Montreal Heart Institute, 5000 East, Belanger Street, Montreal, Quebec, HlT lC8, Canada.

10

January 1979

The directional range-gated pulsed Doppler instrument used in this study was a prototype developed by the Centre National des Recherches Scientifiques, Paris, France. An improved unit is now commercially available (Alvar Electronic, Paris, France). According to the Doppler principle, ultrasonic waves are backscattered with a shifted frequency by moving erythrocytes. The difference in frequency, called the Doppler shift, is proportional to the emitted frequency, the velocity of the target erythrocytes and the angle cosine between the emitted ultrasonic beam and the direction of moving cells; it is inversely proportional to a constant expressing the speed of ultrasound in the medium. The frequency range of the Doppler shift is within the audible range. In this study, we did not evaluate the angle between the beam of ultrasound and the target bypass graft and we did not attempt a quantification of flow. The unit has a range-gated system that is adjustable in depth and in width.

It transmits at a frequency of 4 megahertz and at a pulse repetition rate adjustable from 5 to 20 kilohertz. The maximal depth of exploration is 14 cm. The duration of the acoustic pulse generated by the piezoelectric crystal ranges from 2 to 4 Msec.After a delay, the crystal is allowed to receive for 1 wsec or more. The

The American Journal of CARDIOLOGY

Volume 43

PULSED DOPPLER STUDY OF BYPASS

GRAFTS-DIEBOLD

ET AL.

FIGURE 1. Schematic diagrarrls illustrating the method of graft location in the transverse (left) and lateral (right) planes. The grafts take off from the aorta (AO) to join the right (RCA), left circumflex (CCA) and left anterior descending (LAD) coronary arteries. The transverse lines on the axis of the transducer indicate the ultrasonic beam. The aorta and, anteriorly, the right ventricular outflow tract (RVOT) are easily recognized on A and M mode echocardiography. The Doppler unit has a range-gated system that is adjustable in depth and in width allowing sampling of the Doppler signal at any level along the ultrasonic beam; this sample volume is illustrated on the figure by the lines parallel to the transducer on the’ultrasonic pathway. To identify grafts to the left coronary artery, the transducer is positioned in the second or third left intercostal space and the aorta is localized with standard echocardiograpliy and its Doppler flow signal. As illustrated, the sample volume is then moved slightly to the left, up and down by tilting the transducer and anteriorly by changing the depth of exploration of the Doppler electronic sampling gate. Graft flow is identified in front of or to the left of the right veniricular outflow tract. For a graft to the right coronary artery, the transducer is positioned in the right parasternal space and the aorta is first localized, Slight angulation of the probe with anterior displacement of the electronic sampling gate allows recording of the graft flow. Note that the internal mammary artery (IMA) and veins (IMV) cross the ultrasonic pathway. LV = left ventricle; SVC = superior vena cava.

depth at which the velocity is detected is controlled by the duration of the delay in thee receiver; the minimal size of the sample volume is determined by the duration of the burst of ultrasound. A high pass filter with a cutoff frequency of 500 hertz eliminates low frequencies from cardiac and vessel wall motion. The site of the sample volume is displayed in the simultalieous echocardiogralm and therefore can be exactly located and selected. This range-gated system separates the flow being analyzed from other moving structures on the axis of the transducer. This property of the pulsed Doppler technique allows a discrimination in depth and an estimation of the width of the flow as opposed to continuous wave Doppler measurements whose sample volume starts at the transducer and ends with degradation of the signal. The analog curve of velocity is derived from a zero crossing detector and transcribed on a strip chart recorder with the electrocardiogram at a paper speed of 25 and 50 cm/set. Material and Methods Pulsed Doppler studies vvere performed without knowledge of the status of the graft. The only information available to the investigators at the time of study was the number and location of the aortocoronary bypass grafts as described in the operative protocol. Only flows well documented on strip chart recordings were retained fi3r the diagnosis of graft p.atency. Cardiac catheterization was performed the following day according to techniques already described.4T5 The Doppler study

January

and the angiogram were then correlated. When discordant results were observed, the Doppler study was repeated to identify the sources of misinterpretation. Cardiac catheterization was performed as part of a longterm prospective study designed to evaluate graft outcome. The study was attempted in 120 consecutive patients. Eight patients were studied within 4 weeks of surgery; the remaining 112 were studied 1 to 6 l/2 years postoperatively (mean 4.7 years). Adequate Doppler ultrasonic evaluation as judged by identification of the aortic flow was possible in 108 of the 120 patients. The total number of grafts studied was 163. Technique for identifying and recording graft flow: Grafts to the left coronary artery were examined with the patient in the left lateral position and the probe in the second or third left intercostal space. The aorta was first localized, then anteriorly and to the left, the right ventricular outflow tract. Moving the sample volume laterally by tilting the transducer or anteriorly by changing the depth of exploration allowed identification or graft flow in front of or to the left of the right ventricularautflow tract. The operative report was used to aid identification of grafts to the left anterior descending artery or to the left circumflex artery, or both. When both vessels had grafts, flow in the left anterior descending artery was usually more anterior and to the right than that in the circumflex artery. Bypass grafts to the right coronary artery were examined with the patient lying in the right lateral position and the probe located in the second or third right

1979

The American

Journal of CARDIOLOGY

Volume 43

11

PULSED DOPPLER STUDY OF BYPASS

GRAFTS-DIEBOLD

ET AL

FIGURE 2. Identification of right coronary bypass graft. Left, recording of the Doppler signal from the graft. The flow by reference to the electrocardiogram (ECG) is diastolic. The Doppler apparatus in its process of signal analysis introduces a delay of about 65 msec. 0 = zero flow; T = fhe time axis; V = velocity axis. The angiographic image of the graft is illustrated at right.

FIGURE 3. Illustrativeexample of a left anterior descending coronary bypass graft. A diastolic flow is recorded with the transducer in the left parasternal space and the Doppler sample volume on the left side of the pulmonary outflow tract (left). The angiographic image of the graft is illustrated on the right. Abbrevi-

ationsas in Figure 2.

intercostal space. The aorta was first localized, and a slight angulation of the probe with anterior displacement of the electronic sampling gate allowed recording of right coronary artery bypass grafts on the right side of the aorta (Fig. 1). Before a graft was judged occluded, the adequacy of the ultrasonic examination was confirmed by recording flows from the aorta. For the purpose of this study, the Doppler ultrasonic evaluation was considered not possible if the aortic blood flow signal was not adequately recorded. This occurred in 12 of 120 consecutive cases. If aortic flow was well recorded and a diastolic flow was not recorded in the presumed position of the graft, the graft was judged occluded. Sensitivity was calculated as the percent of patent grafts detected by the Doppler anal-

TABLE I Noninvasive Assessment of Aortocoronary Bypass Grafts (108 patients, 183 grafts examined)

Doppler Patent graft Occluded graft

LAD

Angiography Patent Graft Occluded Graft LCx RCA Total LAD LCx RCA Total

56

22

47

127

6

1

4

11

4

2

5

11

4

5

5

14

P <0.0001 Sensitivity 92 percent;

specificity

56 percent

LAD = grafts to the left anterior descending coronary artery; LCx = grafts to the circumflex coronary artery; P = probability value calculated with the Fisher probability test: RCA = grafts to the right coronary artery.

12

January

1979

The American Journal of CARDIOLOGY

ysis and specificity as the percent of occluded grafts correctly identified.

Results Graft Flow bypass grafts are recorded over a small width as diastolic flows. The high pass filter eliminates the low amplitude systolic component of coronary flow. Figures 2 and 3 show illustrative examples of flows from saphenous vein grafts connected, respectively, to the right and the left circumflex coronary arteries. The corresponding angiographic images of the grafts are shown; by referring to the electrocardiographic tracing, the diastolic pattern of flow can be recognized. The Doppler apparatus introduces a delay of about 65 msec in its process of analysis. At times, the diastolic flow is associated with a definite systolic component. If this systolic component is directionally the same as the diastolic component, the curve obtained has a bifid appearance and is referred to as a bifid flow pattern (Fig. 4); if the systolic component is directionally opposite, the curve is referred to as a biphasic flow pattern (Fig. 5). Overall results are shown in Table I. There were 138 Bypass

Aortocoronary

positive test results; that is, an unquestionable diastolic flow was recorded at a depth and location compatible with that of a patent graft in the angiogram. The grafts with a positive test included 64 left anterior descending, 24 left circumflex and 52 right coronary bypass grafts.

Volume 43

PULSED DOPPLER STUDY OF BYPASS

No flow could be recorded in 25 grafts, including 10 to the left anterior descending, 6 to the left circumflex and 9 to the right coronary artery. Among the 25 grafts occluded at angiography, 14 were correctly identified as occluded with pulsed Doppler analysis and 11 were falsely diagnosed as patent; that, is, a definite diastolic flow was recorded in the presumed position of the graft but the graft was found occluded at angiography. Among the 138 angiographically patent grafts, 127 were recorded by Doppler analysis; 11 were not. found. All occluded grafts were closed proximally, and the only remaining visible structure was usually a small aortic stump (less than 1 cm long) in the aortic angiogram. A diastolic blood flow signal from the graft, proximal to the occlusion, could not thus be recorded. Four grafts had narrowing of 90 percent or more. One was on the proximal portion, one on the mid portion and two on the distal portion of the graft. However, flow was recorded in these grafts and they were considered pat,ent. The Fisher.exact probability test showed a significant statistical difference (P
Because of the low specificity, a major aim of this study was to analyze in greater detail the different patterns of intrathoracic flows obtained by external

FIGURE 5. Biphasic flow recorded from the right ventricle. This figure illustrates the complexity of flows in the right ventricular cavity. As in Figure 4, the transducer was held fixed and the Doppler sample volume displaced posteriorly from the anterior part of the right ventricle to the aorta. A negative systolic flow is recorded during the first cardiac cycle; during the following three cycles, the flow appears biphasic, and further back it is systolic. The sample volume is then displaced through the aortic wall and a positive systolic flow is recorded from the aorta in the two last cardiac cycles. This maneuver demonstrates that this biphasic floi originates from the right ventricle. Abbreviations as in Figure 2.

GRAFTS-DIEBOLD

ET AL.

FIGURE 4. Bifid flow recorded from a left anterior descending bypass graft. The transducer was held fixed and the Doppler electronic sampling gate moved posteriorly. A bifid flow pattern is first observed over a few millimeters (corresponding to the first two cardiac cycles on the figure); then a systolic flow is recorded over a few centimeters. The diastolic component of the bifid Ilow is from the graft and the systolic component from the right ventricular outflow tract. This maneuver allows differentiation of graft flow from the normally occurring bifid flow in other portions of the right ventricle (see text and Fig. 7). Abbreviations as in Figure 2.

examination and to define better diagnostic criteria. During this study, sources of diastolic flows other than those of the grafts were idetitified and correctly interpreted. Patent grafts: The flow curves recorded from patent grafts were diastolic in most cases. However, they were biphasic in seven grafts, bifid in nine and either biphasic or bifid, depending on orientation of the probe, in three additional grafts. The flow misdiagnosed as originating from a graft had a pure diastolic flow pattern on eight occasions and a biphasic flow pattern in three. Internal mammary veins: In the left lateral position, pure diastolic flows were recorded from bypass grafts, internal mammary veins, the left anterior de-

ECG

January

1979

The American

Journal of CARDIOLOGY

Volume 43

13

PULSED DOPPLER STUDY OF BYPASS GRAFTS-DIEBOLD

Internal

mammary

ET AL

vein

ht.

FIGURE 6. Recordings of internal mammary vessels. At left, a diastolic flow is recorded. It is not affected by deep inspiration, showing that it is fixed to the chest wall. The Valsalva maneuver abolishes it, suggesting a low pressure venous flow. In addition, a systolic flow coursing in an opposite direction is identified from the internal mammary artery (Int. mam. art.). Abbreviations as in Figure 2.

ECG

Basal

Inspiration

Valsalva

scending coronary artery and the mitral valve orifice. Internal mammary veins can be recognized in all patients. More superficial than aortocoronary bypass grafts, these veins are fixed to the chest wall and remain in the Doppler sample volume during deep inspiration. They are little affected by heart motion and can be recorded without the high pass filter. Moreover, their low pressure venous flow is blocked by the Valsalva maneuver. At their side, a systolic flow is recorded coursing in an opposite direction and representing flow from the internal mammary artery (Fig. 6). The internal mammary veins were identified as a potential source of error in our preliminary study and were easily avoided in the present study. Left coronary artery and mitral valve: The native coronary arteries exhibit the same flow characteristics as bypass grafts, but the adjacent structures are different. A mode and Doppler echocardiography allow recognition of the left anterior descending coronary artery by identification of a systolic flow in the left ventricular outflow tract behind it and, further back,

mam. art.

of the diastolic filling flow through the mitral valve (Fig. 1). Flow from the mitral valve is diastolic and recorded over a wide area; it is recognized by its pattern and by echocardiographic identification of valve leaflets. Four of the seven false positive recordings of left coronary bypass grafts were pure diastolic flows and were thought to originate from the’left anterior descending artery. Right ventricle: A biphasic flow pattern in the left lateral position was recorded from bypass grafts and from the right ventricle. Three of the false positive recordings were biphasic and from the right ventricle. A bifid flow pattern was recorded from the left side of the sternum at the apical region of the right ventricle (Fig. 7). It could be recorded over a wide depth and was not a source of misinterpretation in this study. Tricuspid valve and right coronary artery: In the right lateral position, a pure diastolic flow was found from the bypass grafts, the tricuspid valve and the right coronary artery. The tricuspid valve flow was recorded over a width of a few centimeters in the vicinity of the tricuspid valve leaflets on A mode echocardiography.

FIGURE 7. Bifid flow recorded from the right ventricle. The usual right ventricular flow pattern is illustrated. As opposed to the maneuver described in Figure 4, the depth of the sample volume was kept constant while the transducer was swept from the apex to the pulmonary valve. A bifid flow is recorded in the apical region corresponding to filling toward the pulmonary valve and ejection in the same direction. Higher -in the outflow tract a systolic flow is found. The direction and intensity’of recorded flows are related to the angle between the ultrasonic beam and moving erythrocytes. Initially it is less than 90’; the intensity of flow decreases when the angle apbroaches 90’; then, with a wider angle, tiflow curve is inverted. Abbreviations as in Figure 2.

14

January 1979

The American Journal of CARDIOLOGY

Volume 43

PULSED DOPPLER STUDY OF BYPASS

It had two diastolic components, and sounds from the valve motion were heard. Right coronary arterial flow was thought to be the cause of three of the false positive flow recordings and the tricuspid valve of a fourth one. Superior vena cawa: A bifid pattern of flow was found in the right lateral position from the superior vena cava. This flow was easily identified because it was recorded beside the aorta over a few centimeters. It was not a source of error in this study. Diiscussion A large incidence of postoperative symptomatic improvement after aortocoronary bypass surgery has resulted in a widespread use of this form of treatment in patients with obstructive coronary artery disease. So far, objective evaluation of graft outcome has been assessed only with cardiac catheterization and angiography.’ Long-term results will have to be reassessed periodically because of the evolution in surgical techniques and different experiences among institutions. Obviously, a method that would alla’w noninvasive, reliable and reproducible results in thle study of graft patency would be of much interest. The Doppler principle was first applied to blood flow studies by Franklin et al6 Pisko-Dubniesky et al.,7 using a contitiuous wave Doppler, studied 226 aortocoronary bypass grafts in the weeks after surgery. They reported a detection rate of more than 90 percent for grafts to the right or the left anterior descending coronary artery; the study of grafts to the left circumflex coronary artery was considered unreliable. Few patients had angiograms in their study. Moreover, considering the complexity of intrathoracic flows, their results may be open to question. Information derived from continuous wave Doppler is an average of velocity of structures crossed by the ultrasonic beam without distance discrimination, and it does not allow identification of sources of misinterpretation. Gould et al.:’ described pulsed Doppler ultrasonic arteriography and applied this method to the study of aortocoronary bypass grafts. They identified patency in 24 of 41 grafts studied and a sensitivity level of 59 percent. They attributed their low detection rate to the limits of their instrument, particularly to its signal to noise ratio; indeed, if they retained for analysis flows that were heard but not recorded, their sensitivity level increased to 83 percenl;; specificity was not examined in their study. Sensitivity and advarntagesof method: Our method is simpler and easier to perform because it does not require visualization of the takeoff of the graft from the aorta. The use of a more modern apparatus may have also contributed to the better sensitivity. This sensitivity could be further increased to almost 100 percent with use of a control unblinded Doppler study. Instrumentation now available with an improved signal to noise ratio should allow applicability of the method to more patients. The specificity level appears low; however, the method has to be tested on a larger number of grafts before the specificity data will be meaningful. Because 11 false positive reponses were found among

GRAFTS-DIEBOLD

ET AL.

the 25 occluded grafts, it is likely that in some cases of patent grafts, the Doppler signal originated from other structures and not actually from the graft. Thus, the actual sensitivity is lower than the calculated one. However, a significant statistical difference was found between the groups of occluded and patent grafts, thereby validating the Doppler capability of detecting graft patency. Differentiation of various causes of diastolic flow: Our diagnostic criterion was initially similar to that of previous studies-that is, identification of a diastolic flow. It soon became clear that other sources of diastolic flow could be confused with that of aortocoronary bypass grafts. Our results thus stress the importance of correctly identifying intrathoracic flow patterns. Explanations for these patterns were derived from anatomic and physiologic considerations and must be controlled by further studies. The diastolic pattern of internal mammary vein flow was identified in most patients; its dependence on chest wall structures and its low pressure characteristics made the diagnosis quite obvious, and it should not be a source of misinterpretation. Similarly, flows from the tricuspid and mitral valves were easily recognized using A mode and Doppler echocardiography. Their diastolic flow had two components similar to the motion of the valve leaflets in the echocardiogram. The superior vena cava in this study was identified by its location to the right of the aorta and also by its pattern of flow similar to the one already described for the right atrium9 and jugular veins.lO Patterns of flow in the right ventricle were more complex. Depending on the cosine of the angle between

the ultrasonic beam and the direction of flows, diastolic, systolic, biphasic or bifid patterns could be recorded. In the vicinity of the tricuspid valve a pure diastolic flow was found. By moving the electronic sampling gate from the apex toward the pulmonary valve, a bifid flow with a large diastolic component was first recorded; then the diastolic component decreased progressively and finally disappeared, whereas the systolic component increased (Fig. 7). The geometry of the right ventricle favors diastolic filling toward the pulmonary valve and systolic ejection in the same direction. Such a bifid flow was described in the left ventricle by Benchimol et al.ll using a catheter-tip nondirectional Doppler flow probe. Occasionally, flow sampled in the right ventricular outflow tract appeared biphasic, some flow current filling the outflow tract during diastole in a direction opposite to that of ventricular ejection (Fig. 5). The use of an empirically determined depth limit above which all or most grafts lie did not improve specificity and lowered the sensitivity. Indeed the depth of the recorded flow varied not only with the chest wall thickness but also with the recording site on the graft and its angle with the transducer. Diastolic flow recorded in the graft was sometimes associated with a systolic flow. The low frequency

systolic component of graft flow itself is usually eliminated with use of the high pass filter. Associated systolic flow represented contamination from other structures projecting anteriorly in the Doppler sample volume

January 1979

The American Journal of CARDIOLOGY

Volume 43

15

PULSED DOPPLER STUDY OF BYPASS

GRAFTS-DIEBOLD

ET AL.

during systole. The right ventricular outflow tract and aorta have such an anterior motion. Penetration of their flows in the Doppler sample volume during systole resulted in the recording of a bifid or biphasic flow pattern, depending on their orientation relative to the ultrasonic probe. Maneuvers for confirming patency or occlusion of graft: Before concluding that a graft is patent, one must be able to recognize the different flow patterns in the vicinity of the graft. At the left sternal border, patent grafts are recorded as diastolic flows over a small width, facing the right ventricle at a site where its flow is purely systolic; behind the right ventricle the systolic flow of aorta should be identified. Because coronary bypass grafts have the same flow pattern as the coronary arteries,Q care must be taken to avoid confusion. Left coronary arterial flow will be avoided if one is careful to record the graft above the left ventricle using echocardiographic landmarks. At the right sternal border, the tricuspid valve is easily recognized. Because the graft and right coronary artery often run parallel in the atrioventricular groove they may be difficult to distinguish from each other. These grafts should be examined as close as possible to the aorta. On the other hand, before judging that a graft is occluded, one has to make sure that all areas in front and

the left of the right ventricular outflow tract have been carefully explored for left coronary arterial grafts and to the right side of the aorta for right coronary arterial grafts. In so doing it is useful to remove the high pass filter that might eliminate the low velocity flow of some bypass grafts. With use of this thorough methodical technique, developed while the study was ongoing, our results gradually improved. Other maneuvers, such as injection of a selective coronary vasodilator or indocyanine green or a saline solution, could also be useful to identify, respectively, flow from the right side of the heart and from the bypass grafts. Clinical implication: Because of its low specificity, the method cannot be recommended as a clinical tool for evaluating graft patency in individual patients. However, it appears promising for the future and exciting as a clinical research tool. Its development should proceed with improvement of the performance of the apparatus and with enhanced clinical skill and experience in identifying and eliminating sources of error. to

Acknowledgment The technical assistance of Michel Xhaard and Alain Barbet and the secretarial work of Diane Roy are gratefully acknowledged.

References 1.

2. 3.

4.

5.

6.

16

Bourassa MG,Lespbrance

J, Campeau L, Grondin CM: Serial angiographic studies after aortocoronary bypass surgery. Postoperative patency rates. In, Coronary Angiography and Angina Pectoris, Symposium of the European Society of Cardiology, Hanover, March 1975 (Lichtlen PR, ed). Stuttgart, Georg Thieme, 1976, p 199-205 Pdronneau P, L&ger F: Doppler ultrasonic pulsed blood flowmeter. Proc 8th lnternat Conf Med Biol Eng Chicago, 1969, p 10-l 1 Gould KL, Mozersky DJ, Hokanson DE, Baker DW, Kennedy JW, Summer DS, Strandness ED Jr: A noninvasive technic for determining patency of saphenous vein coronary bypass grafts. Circulation 46:595-600, 1972 Bourassa MG, LespBrance J, Campeau L, Bois MA, Saltiel J: Selective coronary angiography using a percutaneous femoral technique. Can Med Assoc J 102:170-173, 1970 LespCrance J, Saltiel J, Bourassa MG: Angulated views in the sagittal plane for improved accuracy of cinecoronary angiography. Am J Roentgen01 121:565-574, 1974 Franklin DL, Schlegel W, Rushmer RF: Blood flow measured by Doppler frequency shift or back-scattered ultrasound. Science

January 1979

The American Journal of CARDIOLOGY

134:564-565, 1961 7. Pisko-Dubiensky ZA, Baird RJ, Wilson DR: Non-invasive assessment of aorta-coronary saphenous vein bypass graft patency using directional Doppler. Circulation 51, 52:Suppll:l-188-l-196, 1975 8. Kalmanson D, Bernier A, Veyrat C, Wiichetr S, Savier CH, Chiche P: Normal pattern and physiological significance of mitral valve flow velocity recording using transseptal directional Doppler ultrasound catheterization. Br Heart J 37:249-256, 1975 9. Benchimol A, Baretts EC, Gartlan JL: Right atrialflow velocity in patients with atrial septal defect. Am J Cardiol 25:381-388, 1970 10. Kalmanson D, Veyrat C, Derai C, Savier CH, Berkman M, Chiche P: Non-invasive technique for diagnosing atrial septal defect and assessing shunt volume using directional Doppler ultrasound. Correlations with phasic flow patterns of the shunt. Br Heart J 34:981-991. 1972 11. Benchimol A, Desser KB, Gartlan JL Jr: Left ventricular blood flow velocity in man studied with the Doppler ultrasonic flowmeter. Am Heart J 85:294-301,1973

Volume 43