noninvasive ultrasonic estimation of brachial artery blood flow

NONINVASIVE ULTRASONIC ESTIMATION OF BRACHIAL ARTERY BLOOD FLOW

by Paul A. Thomas, Jr., MD

ABSTRACT Ultrasonic instrumentation based on the Doppler phenomenon fulfflls the requirement that a blood flowmeter should sense a variable which is a function of the volume of blood moved through a blood vessel in situ. The blood flow velocity waveform recorded noninvasively from the brachial artery in the antecubital fossa was used as basis for quantita- tively estimating a stroke flow index. The method used in these experiments was reproducible in 31 volunteer subjects studied; the first and second brachial artery stroke flow indices determined on different days had a correlation coefficient of 0.909. The calculated minute flow index ranged from 32 t o 95 ml in these subjects. A clinical application was explored by making serial measurements before and after operation in 64 patients submitting to 68 open heart operations. A significant reduction in the brachial artery blood flow velocity was recorded postoperatively in 21 of these patients. The brachial artery stroke flow index has potential value as an objective measure of cardiovascular instability.

Indexing Words Ultrasonic Blood Flowmeter Doppler Velocimeter Brachial Artery Blood Flow

Brachial Artery Stroke Flow

An ideal blood flowmeter should be capable of sensing a variable which is a predictable function of the volume of blood moved per unit of time through the blood vessel in situ. Ultrasonic instrumentation based on the Doppler principle fulfills this requirement. In 1961 Franklin et a1 (1) described a method of measuring blood flow by determining the sound wave frequency change from incident to reflected sound waves based on the assumption that a linear relation- ship exists between the observed frequency change and the velocity of moving elements in the stream of blood. Subsequently, Vatner et a1 (2) experimentally compared and calibrated an ultrasonic blood flowmeter with an electromagnetic blood flowmeter and demonstrated a linear response to the arterial flow ranges studied using both instruments. The flow probes for both instruments were chronically implanted around the arteries of their experimental animals. They concluded that accurate measurement of

From the Division of Research, The Lankenau Hospital,

Received December 17 , 1975; revision accepted August 25.

For reprints contact: VA West Side Hospital. 8 2 0 South

Philadelphia, Pennsylvania 191 51 .

1976.

Damen Avenue, Chicago, Illinois 6 0 6 1 2 .

regional blood flow was possible with an ultrasonic flowmeter. Benchimol et a1 ( 3 ) simultane- ously recorded identical ultrasonic brachial artery flow velocity waveforms from both implanted and transcutaneously placed probes in a patient studied for aortic valve stenosis. Their study confirmed the potential for noninvasive estimation of blood flow. Ram and Bhimani ( 4 ) calibrated an ultrasonic flow probe in uitro and made transcutaneous clinical measurements over paired vessels of the extremities. They arrived at quantitative estimates of blood flow by evaluating the flow velocity waveforms and calculating a ratio for paired vessels. They concluded that the velocity signal of sound frequency shift was directly proportional to the volume of blood flow, provided the size of the vessel remained constant. Flax et al (5) described design and testing specifications for Doppler ultrasonic blood flowmeters to assure reliability. Reagan et a1 (6) calculated the mean velocity of femoral artery blood flow from the recorded ultrasonogram and estimated the artery cross-sectional lumen area by ultrasonic echo ranging to determine volume flow. Al- though their calculated femoral artery volume

7 2 JOURNAL O F CLINICAL ULTRASOUND

FIGURE 1. The recorded peak deflection i s measured f rom the f l ow velocity tracing and the peak velocity IS calculated as a rat io o f the standard signal deflection (Equation 1) The average area beneath the prlmary waveform deflect ion IS measured planimetrically and the area beneath the callbratlon signal, pro- jected in t ime for the period o f the primary waveform deflection (shaded area), i s determined F rom the rat io o f these areas, a brachial artery stroke f l ow index i s calculated (Equation 3)

flows were comparable to those determined by other investigators using indicator dye dilution methods, they were cognizant of the probe placement difficulties inherent in the transcutaneous method. We modified their method for clinical application. In this report we describe a noninvasive method of recording the brachial artery flow velocity waveform and the deriva- tion of a quantitative estimate of the stroke flow volume. Application of the method is illustrated by measurements made serially in cardiac surgical patients.

MATERIAL AND METHODS

Instrumentation: Brachial artery flow velocity measurements were made using a nondirectional, multipurpose, continuous wave 5 MHz Doppler instrument.* This instrument is sensitive to changes in blood flow velocity but not the direc- tion of blood flow. Velocity frequency shifts produced by vessel wall motion are reported to be less than 100 Hz ( 5 ) . Since these fre- quencies are below the range of linear response for the instrument used, they were filtered out electronically and presumably did not contri- bute to the shifted frequency signals recorded as blood flow velocity. The frequency shift from incident to reflected sound was detected by the transducer housing transmitting and receiving. piezoelectric crystals. The frequency shift signal * Smith Kline Instruments

was passed through a zero-crossing circuit t o a calibrated recorder.* this ultrasonic instrument has characteristics consistent with the specifications described by Flax et a1 (5 ) which overcome objections expressed regarding reliability and accuracy of ultrasonic blood flowmeters. The flow velocity waveform recorded from the nondirectional instrument was qualitatively compared with the waveform recorded from a directional ultrasonic flowmeter.**

Calculations: The recorded flow velocity waveform produced by the passage of a bolus of blood through the incident sound beam directed into the brachial artery was used t o estimate a mean peak velocity and a stroke flow index (Fig. 1). A calibration signal of 100 mv representing a piezoelectric crystal excitation of 1000 cycles/ sec equivalent to 15.4 cm/sec velocity was recorded with each brachial artery flow velocity determination. For convenience the recorder was adjusted t o provide a needle deflection of 20 mm from the baseline in respolise to the calibration signal. The mean peak velocity was computed from the ratio:

15.4cm/sec x MPRmm 1. MPV = 20mm

where MPV = mean peak velocity 15.4cmlsec = velocity of calibration signal

MPR = amplitude of mean peak recorded

20mm = amplitude of recorded calibration signal

signal

The mean peak velocity recorded was estimated by inspection of sequential flow velocity waveforms displayed on the brachial artery flow velocity tracing. The formula derived for the cal- culation of a quantitative brachial artery stroke flow index was based on the equation relating the Doppler shift in sound wave frequency to the velocity of a moving object:

2. v =

where V =

A f =

ft =

a = c =

A f . C 2 f t c o s a

velocity of the object frequency shift, transmitted t o reflected sound frequency, transmitted sound angle, acoustical axis with blood stream velocity of sound in tissue

However, the moving object, a bolus of blood, accelerates and decelerates with each propagated

*Hewlett Packard Viso-Cardiette Model 500 using direct current input mode.

Delalande Electronique. **Directional Ultrasonic Flowmeter, continuous wave 4MHz.

VOLUME 5. NUMBER 2 73

pulse wave in an artery; therefore, it was neces- sary to determine the mean velocity. This was done by measuring the areas under three to five consecutive velocity waveforms using a planim- eter and taking the average area to estimate the mean frequency shift. Since both the velocity of sound in tissue, 1.54 x lo5 cm/sec, and the frequency of the incident sound for the instrument used, 5 x lo6 cycleslsec, are constant, the mean velocity is proportional to the average area beneath the recorded waveforms as a function of the calibration signal (Fig. 1). The brachial artery stroke flow index (SQ) was calculated as the product of the mean blood flow velocity of each bolus of blood passing through the incident sound beam and the vessel cross-sectional lumen area (A):

3. SQ = VA average area beneath

- recorded waveforms X 15.4cm/sec X A calibration signal area for time cosa average waveform

The angle between the incident sound beam and the brachial artery was not determined precisely in these experiments. For noninvasive clinical measurements, the absolute anatomical course of the brachial artery in the antecubital fossa of a particular subject was not known. The ultrasound transducer was placed over the artery and was adjusted to the point at which the loudest, highest pitched frequency shift was audible over the loudspeaker of the instrument. Trial- and-error positioning of the transducer probed directing the incident sound beam into the brachial artery consistently resulted in an angle of less than 45" with the skin surface. In so doing, the value of cos a approached 1.

The brachial artery cross-sectional lumen area was determined by direct measurement in 95 patients who required arteriotomy and retrograde catheterization during the period of this investigation. This group included some, but not all, subjects in the brachial artery flow velocity study since arteriotomy was not neces- sary in all. Brachial artery sizing was done by inserting into the artery progressively larger calibrated obturators to visually determine the largest size obturator the vessel would accommodate. A probable cross-sectional lumen area was calculated for each size obturator used. The calculated areas ranged from 2.2 x 10-2 cm2 to 7.2 x 10-2 cm2. All patients were adults and as expected the obturator size correlated with the sex and body habitus of the patients. Smaller brachial arteries were encountered in women and small men. The calculated cross-

74

PRIMARY

BASELINE

FIGURE 2. Brachial artery flow velocity tracing using a directional Doppler instrument. The secondary deflection below the baseline is retrograde flow. The tertiary deflection is antegrade flow of a comparable magnitude. The tracing was made at a paper speed of 50mm/sec.

sectional lumen area was used for brachial artery stroke flow index determination in those study subjects who submitted to arteriotomy although flow velocity measurements were made over the contralateral vessel to avoid surgically induced flow disturbances. A value for the lumen area was assigned for subjects not requiring arteriotomy . Assignment of area was based on an average lumen size for similar individuals of comparable body size and the same sex. Computation of the brachial artery stroke flow index required an assumption that the lumen area remained acceptably constant throughout the pulse propagation cycle. Flow velocity studies were repeated on one or more occasions for each subject and the same value for lumen area was used for all flow calculations for that individual.

The brachial artery flow velocity tracings recorded for each clinical determination were inspected and the section with the least transducer placement artifact was selected for analysis. The area beneath the primary velocity deflection of three to five consecutive waveforms was measured planimetrically and an average area per waveform was determined. Qualitative brachial artery blood flow. velocity tracings made with the directional flowmeter (Fig. 2) support the assumption that secondary and tertiary oscillations represent reverse and forward flow respectively of comparable magnitude. The secondary and tertiary oscillations were excluded as noncontributory to forward flow of blood. The average time elapsed to transcribe the primary velocity waveform was noted to determine the area for the calibration signal per stroke. These data were inserted into Equation 3 along with the measured or assumed cross-sectional lumen area for that individual

JOURNAL OF CLINICAL ULTRASOUND

TABLE I Patients Studied Before and After Cardiac Surgery

Number of Number Incidence Patients of 50 Per cent

ODeration Studied Procedures Flow Reduction ~ ~~

1. Myocardial Revascularization Coronary Artery Bypass

2. Aortic Valve Replacement

3. Mitral Valve Replacement Valvuloplasty

4. Closure Atrial Septa1 Defect

5. Bivalve Replacement Mitral and Aortic

6. Miscellaneous Procedures* TOTAL:

23

10

10 4

9

4

4 64 -

25 9

12 5

10 2 4

9

4 2

2 68 21 - - 4

*Includes closure ventricular septa1 defect, resection of atrial myxoma and ventricular aneurysm, and correction oi anomalous pulmonary venous drainages.

and the brachial artery stroke flow index was computed.

Clinical Measurement Method: Subjects selected for the clinical investigation were ex- amined while relaxing, supine, in bed to esta- blish resting circulatory dynamics before each determination of brachial artery flow velocity. A colloidal water coupling gel was applied to the skin surface of the antecubital space to eliminate air spaces between the transducer crystals and the brachial artery. The sound transducer was positioned in line with the anatomical course of the artery directing the incident sound beam into the blood stream. The transducer was adjusted by hand to the position and angle that resulted in the loudest, highest pitched audible sound. Small adjustments in the position of the transducer during brachial artery flow velocity recording did not produce critical changes.

Thirty -one apparently heal thy volunteers were studied and restudied on separate days by the same individual to standardize the measurement method and to determine the technique’s reproducibility. A feasibility study of one potential clinical application was made in 64 patients submitting to 68 cardiac surgical procedures (Table 1). Brachial artery flow velocity measurements were made preoperatively and serially for 7 t o 14 days postoperatively in these patients.

RESULTS

Secondary and tertiary oscillations in the nondirectional brachial artery flow velocity tracing

(Fig. 1) were identified as retrograde and antegrade flow by qualitative comparison with a brachial artery flow velocity tracing recorded using a directional flowmeter (Fig. 2). Although the output signal from the directional instrument was not calibrated for this study, the amplitude of the secondary retrograde and tertiary antegrade velocity changes were of similar magnitude. The estimated probable error introduced by excluding the algebraic sum of areas beneath the secondary and tertiary oscillations in calculating the brachial artery stroke flow index is G3 per cent.

A good correlation between the first and second brachial artery flow velocity measurements was demonstrated for the 31 volunteer subjects studied. The calculated brachial artery stroke flow indices (Fig. 3) have a correlation coefficient of 0.909 with no significant difference between the first and second determinations (p<O.OOl). As mentioned, the acoustical angle of incident sound beam to blood stream as well as the cross-sectional lumen area of the artery were considered to remain constant for both determinations. The pulse rate was counted from the brachial artery flow velocity tracing and a minute flow index was calculated which ranged from 32 to 95 ml (mean 59 ml).

A majority of the cardiac surgical patients studied experienced an uneventful recovery following operation. A significant reduction in an -individual’s brachial artery blood flow, de- fined as a 50 per cent or greater reduction in

VOLUME 5. NUMBER 2 1 5

STROKE INDEX

1.601 140- -

r 2 120- - y1

\ 0

2 100-

980-

0 c 0

L

c 0)

2 060-

L

li:

040-

C O R O W ARTERY BYPASS (3 vessels)

020J, I I I I I

020 040 060 000 100 120 140 160 Second Determinationb/sfroke)

FIGURE 3. The first determination o f the brachial artery stroke f l ow index i s p lo t ted against the second determination for each of the 31 volunteer subjects studied Although a variation be tween subjects is evident, the determinations fall close t o the line of 1 1 correspondence for each subject The correlation coeff ici ent for these data i s 0 909 w i th the probabil i ty of error p<OOOl

both peak velocity and stroke flow index, was recorded at some time during the immediate postoperative period in 21 patients (Table 1). Peak velocity was used as one of two criteria for characterizing a low flow state because it is an objective observation independent of the as- sumptions made to estimate the stroke flow index. Six of the patients studied died; serial brachial artery flow velocity tracings recorded from one such patient illustrates a persistent flow deficit (Fig. 4). The decreased brachial artery stroke flow index observed immediately after surgery frequently improved after three to five days as illustrated graphically by quantitative estimates of the peak velocity and stroke flow index in a patient submitting to myocardial revascularization (Fig. 5).

DISCUSSION

The ultrasonic Doppler instrument provides noninvasive access to blood flow phenomena and senses a variable which is a function of the volume of blood moved through the incident sound beam per unit of time. Therefore, the brachial artery stroke index is the volume of blood passing through the incident sound beam coupled into the blood stream per heart beat. The moving acoustical interfaces responsible for the backscattered sound waves are presumably the cells suspended in the blood. Since blood flow is laminar, the suspended cells move at different velocities presenting a parabolic velo-

I54 crn/rc

Reoperative I

I A J A A Day of surgery

One Day Fostopemtive

Two Doys Postoperative I U

I U Three Guys Postoperative

FIGURE 4. Serial brachial artery f l o w velocity tracings made before and after myocardial revascularization demonstrated a persistently l o w peak velocity postoperatively fo r the three days this patient (No. 24) survived. The lowest oo int of each tracing i s the baseline and the calibration signal o f 15.4 cm/sec was recorded w i th each determination. The calculated stroke f l ow indices were 1.16, 0.62, 0.65 and 0.62, respectively.

CORONARY ARTERY BYPASS (3 vessels)

0-I

Stroke Index cc/st

0 2 4 6 8 Days Postoperative

FIGURE 5 . Quantitative estimation o f brachial artery b lood f l ow b y graphically displaying calculated peak velocity and the stroke f l ow index illustrates recovery f rom a " low f l ow" state in this patient (No 3 1 1 after myocardial revascularlzatton

city profile. The Doppler ultrasonogram of pul- satile arterial blood flow represents some average velocity of the accelerating and decelerating axial volume of blood at any selected moment of pulse wave propagation (7). Apparently, the

76 JOURNAL OF CLINICAL ULTRASOUND

density of specific frequency backscattered sound waves is not critical since hematocrit variations within ranges usually encountered in clinical patients do not produce any appreci- able alteration in the ultrasonogram (4 , 8).

The brachial artery cross-sectional lumen area used t o calculate the stroke flow index was estimated in this study. Determination of absolute flow requires precise knowledge of the vessel lumen diameter in both systole and diastole. McDonald (9) estimated that maximal changes in the radius of the femoral artery .with pulsa- tion were 5 per cent or less, of short duration, and could be ignored. On the other hand, Olson and Cooke (10 ) reported that the systolic diameter of the carotid artery was 8 to 13 per cent greater than the diastolic diameter measured by ultrasonic pulsed echo-ranging methods. We are not aware of any studies that estimate similar changes in cross-sectional lumen area of the brachial artery. Although our estimates of the brachial artery cross-sectional lumen area, obtained by obturator sizing and speculation, are not reliably accurate, it should be pointed out that the same value was used for all flow calculations for any given subject. A systematic error of consistent magnitude does not prohi- bit quantitative comparison of serial data ac- cumulated by repeating flow velocity determinations periodically in the same individual. The additional error introduced by overestima- tion of the brachial artery lumen area in the low flow states associated with vasoconstriction has not precluded recognition of significantly reduced stroke flow. This error can be obviated in the future by using available instrumentation to estimate noninvasively and with greater accuracy the cross-sectional lumen area in v i m .

For the purpose of this investigation we assumed that the secondary retrograde and tertiary antegrade velocity shift oscillations recorded for each arterial pulse wave represented insigni- ficant net forward blood flow. This was done to simplify the waveform analysis because the only instrument available for these studies was nondirectional. Alternative forearm and brachial artery blood flow measurements reported by other investigators are of a comparable magnitude. Andres et a1 ( 1 1 ) determined that the forearm blood flow ranged from 27 t o 78 ml/ min (mean 46 ml/min for 10 subjects) based on indicator dye dilution studies. Gault e t a1 (12) measured brachial artery blood flow by placing an electromagnetic flow probe around the vessel exposed for cardiac catheterization. They recorded flows ranging from 22 t o 107 ml/min

(mean 78 ml/min for five patients without left ventricular disease). In our series of normal volunteers the brachial artery minute flow index calculated as the product of stroke flow index and pulse rate was 32 to 95 ml/min (mean 59 ml/min for 31 subjects). However, the assumption that retrograde flow is insignifi- cant may not be valid for all abnormal flow conditions. Rittenhouse and Strandness (1 3 ) demonstrated significant retrograde blood flow in the femoral arteries of patients with aortic valve disease. Also, at times of decreased brachial artery flow, as is frequently encountered early after cardiac surgery, the secondary retrograde waveform becomes relatively more promi- nent in some instances. Under these conditions the calculated brachial artery stroke flow index overestimates the true blood flow. Some of these disadvantages may be overcome by using a directional instrument.

Standardized placement of the ultrasound transducer, pro be at a predetermined angle of incidence with the column of blood was im- practical in this clinical investigation. Freedom of probe placement over the brachial artery was constrained by the character and variety of other physiologic monitoring devices employed in these patients. Therefore, placement was accomplished by trial and error guided by a subjective determination of sound loudness and pitch. Ram and Bhimani ( 4 ) concede that the acoustical angle between the incident sound beam and the column of blood is important, but they found that by adjusting the probe until a maximal signal was obtained satisfactory results were achieved. Bernstein et al (14 ) attempted to standardize probe placement for experimental quantitative calibration of Doppler devices. They noted that in some clinical appli- cations, the standard angle did not produce the most audible sounds or largest velocity tracings. Their observations suggest that probe placement using anatomical surface landmarks is probably not as accurate as probe adjustment to a maximal sound signal. The most sophisticated method of probe placement involves locating the plane of the artery by ultrasound echo-ranging t o fix the angle of the incident sound beam at a reproducible 45" (1 0). This system has been used successfully t o study carotid artery blood flow. It may be fortuitous that the system used in our study did not respond with critical sensitivity to small adjustments in probe position during the flow velocity recording procedure. Furthermore, individual determinations were reproducible in the volunteers and the patients studied when

V O L U M E 5 , NUMBER 2 I7

measurements were repeated by both the same and different people.

The pathophysiological significance of reduced brachial artery blood flow and quantitative estimation of the severity of the reduction must be interpreted. Reduced flow may result from obtunded myocardial contractility, in- sufficient ventricular filling, proximal obstruction to flow, or constricted distal perfusion bed. On each occasion, the serial brachial artery flow velocity tracings recorded from the illus- trative case (Fig. 4) were made by chance at approximately the same pulse rate. There was no reason to suspect flow obstruction; therefore, the reduced stroke flow could be attributed to either myocardial contractility, a constricted distal perfusion bed, or both in combination, which would be an indication of pump power reduction. Cardiac rhythm disturbances result in obvious brachial artery flow velocity waveform aberrations which have been recorded in some of the patients studied in this series. Benchimol et al (15) studied a series of 33 patients with various spontaneous or induced cardiac arrhythmias recording flow velocity curves from the thoracic aorta. They describe beat-to-beat changes in peak velocity. These changes were minimized in our clinical determinations by the method used, which included selection of the three-to-five best brachial artery flow velocity waveforms for analysis.

The potential value of an ultrasonic method of determining the brachial artery flow velocity lies in the simple, direct, noninvasive nature of the measurement technique. Quantitative estimates of the brachial artery flow are feasible, as demonstrated in this clinical investigation. With the development of more sophisticated equipment incorporating directional flow capa- bility, a continuous digital summation of volume flow similar to that provided with electromagnetic flowmeters is possible. Probe design with precision placement and noninvasive determination of lumen area are possible by adaptation of available technology. Ultrasonogram waveform analysis may yield additional clinically valuable information about the acceleration of a bolus of blood (16). Continuous monitoring of regional arterial blood flow as a noninvasive method of instantaneously observing evolving cardiovascular events would find broad clinical application. Such information is potentially available by this method.

ACKNOWLEDGMENTS The author wishes to recognize the profes-

sional advice provided by Denis M. Abelson.

78

MD, the cooperation of George J. Haupt, MD and Joseph C. Donnelly, MD whose patients were studied, the assistance of Daniel J. Callahan, MD who made numerous clinical measurements, Miss Heather Neal who did the studies on volunteer subjects, and Mr. John Nylund for his a s sistance with the instrumentation. 'a

REFERENCES 1. Franklin DL, Schlegel W, and Rushmer RF: Blood

flow measured by Doppler frequency shift of backscattered ultrasound. Science 134:564,1961.

2. Vatner SF, Franklin D, and Van Citters RL: Simul- taneous comparison and calibration of the Doppler and electromagnetic flowmeters. J Appl Physiol 29:907,1970.

3. Benchimol A, Maia IG, Gartlan Jr JL, and Franklin D: Telemetry of arterial flow in man with a Doppler ultrasonic flowmeter. Am J Cardiol22:75,1968.

4. Ram MD and Bhimani BK: Studies on velocity blood flow with the ultrasonic flowmeter. Surg Gynec & Obstet 1332315,1971.

5. Flax SW, Webster JG, and Updike SJ: Pitfalls using Doppler ultrasound to transduce blood flow. IEEE Trans Bio-Med Eng 20:306,1973.

6. Reagan RR, Miiler CW, and Strandness Jr DE Transcutaneous measurement of femoral arteg flow. J Surg Res 11:477,1971.

7. Franklin DL: Techniques for measurement of blood flow through intact vessels. Med Electron Biol 3:27,1965.

8. Michie DD and Cain CP: Effect of hematocrit upon the shift in Doppler frequency. Proc Soc Exp Biol Med 138:768,1971.

9. McDonald DA: Blood Flows In Arteries. London Edward ARnold LTD, 1960 p. 95.

10. Olson RM and Cooke JP: Human carotid artery diameter and flow by a noninvasive technique. Med Instrum 9:99,1975.

11. Andres R, Zierlet KL, Anderson HM, et al: Mea- surement of blood flow and volume in the forearm of man; with notes on the theory of indicator- dilution and on production of turbulence, hemoly- sis, and vasodilation by intravascular injection. J Clin Invest 33:482, 1954.

12. Gault JH, Ross J Jr, and Mason DT: Patterns of brachial arterial blood flow in conscious human subjects with and without cardiac dysfunction. Circulation 345333,1966.

13. Rittenhouse EA and Strandness Jr DE: Oscillatory flow patterns in patients with aortic valve disease. Am J Cardiol28:568,1971.

14. Bernstein EF, Murphy J r AE, Shea MA, and Hous- man LB: Experimental and clinical experience with transcutaneous Doppler ultrasonic flowmeters. Arch Surg 101:21,1970.

JOURNAL OF CLINICAL ULTRASOUND

15. Benchimol A, Stegall HF, Maroko PR, et al: Aortic 16. Fronek A, Johansen KH, Dilley RB, and Bernstein flow velocity in man during cardiac arrhythmias EF: Noninvasive physiologic tests in the diagnosis measured with the Doppler catheter-flowmeter and characterization of peripheral arterial occlusive system. Am Heart J 78:649,1969. disease. Am J Surg 126:205, 1973.

VOLUME 5. NUMBER 2 7 9

noninvasive ultrasonic estimation of brachial artery blood flow

Documents