- Email: [email protected]
Evaluation of mammary blood flow measurements in lactating goats using the ultrasound doppler principle
Camp. Eiochem. Physiot. Vol. 92A, No. 3, pp. 385-392, 1989
0300-9629189 $3.00+ 0.00 0 1989Pergamon Press plc
Printed in Great Britain
EVALUATION OF MAMMARY BLOOD FLOW MEASUREMENTS IN LACTATING GOATS USING THE ULTRASOUND DOPPLER PRINCIPLE K. CHRISTENSEN,* M. 0. NIELSEN, R. BAKER? and K. HILDEN~ The Royal Veterinary and Agricultural University, Department of Animal Physiology and Bi~hemistry and tDepartment of Physics, Thorvaldsensvej 40, DK-I871 Frederiksberg C, Denmark. Telephone: 06-65-2500 (Received
18 August
1988)
Abstract-l. Non-invasive methods were developed for measuring mammary blood flow in lactating goats. 2. A Doppler principle ultrasound device was equipped with an external detector measuring maximal blood velocity (V,,) and average blood velocity (V,,) was calculated as V,,/2. Volume flow then depended on determination of the angle of insonation and the cross-sectional area of the milk vein (the caudal superficial epigastric or subcutaneous abdominal vein). 3. Blood velocities were measured on the milk vein of either side of the animal while clamping the pudendal veins manually. Blood velocities ranged from 7-34cm/sec. 4. The milk vein diameter was measured by means of a slide gauge which, for clearly protruding veins, gave similar results to that measured by ultrasound scanning. In protruding veins the cross-section was circular. In non-protruding veins the cross-section was elliptical and the slide gauge significantly (P < 0.01) overestimated the cross-sectional area. The milk vein diameter of either side measured in 10 lactating goats was 8.8 f 1.1 mm (R + SD). 5. Blood flow ranged from 90-675 ml/min in a dry and a high-yielding (3.4 1 milk daily) goat, respectively. The reproducibility of the blood flow measurements was 12-16%. 6. It is concluded that the present method may be used for quantitative measurements of mammary blood flow in goats.
INTRODUCTlON Rates of mammary blood flow in ruminants have been measured by direct or indirect techniques as reviewed by Reynolds et al. (I 968), Linzell(1974) and Fleet and Mepham (1983), all of which require surgical incisions prior to the measurements, tk. division of the udder halves, preparation of a milk vein loop or placement of a flow transducer around the external pudendal artery. The recently available Transonic flowmeter (Transonic Systems Inc., New York) using ultrasound based on transit time principles is also for implantation. Ultrasonic methods exploiting the Doppler shift principle were also tried (Linzell, 1974), but apparently no results are available. The ultrasound Doppler principle has been used widely within human medicine for diagnostic and clinical purposes. Quantitative flow measurements have been difficult to perform because the vessels are normally deep-lying, whereby it is difficult to locate the vessel and measure the angle between the moving bloodstream and the ultrasound beam from the transducer. Quantitative measurements with the Doppler method in human fetuses have recently been carried out both with high accuracy and high precision (Rasmussen, 1987b) and the methodological *Present address: National Institute of Animal Science, Department of Animal Physiology and Biochemistry, Forsragsanlaeg Foulum, PO Box 39, DK-8833 0rum Sdrl., Denmark.
problems involved have been discussed in detail (Gill, 1985; Rasmussen, 1987a). The Doppler principle ultrasound equipment measures blood velocity, and volume flow is dependent on determination of the angle of insonation and the vessel cross-sectional area. Although this method so far has not fulfilled the requirements for exact mammary blood flow measurements in ruminants, it is still challenging because it is the only method which allows direct measurements without any surgical incisions. In lactating ruminants the milk vein (the caudal superficial epigastric or subcutaneus abdominal vein) is often easily observed, because it protrudes beneath the skin just anterior to the udder. Therefore, we decided to try the ultrasound Doppler principle to measure blood velocity through the udder of goats. Blood flow could then be calculated as the product of blood velocity and cross-sectional area of the milk vein. This method allowed measurements on both sides of the animal. We used commercially available hi-directional ultrasound Doppler equipment (Angiomatic AC2) originally built for non-invasive diagnosis of peripheral venous and supra-aorta1 diseases in humans. The equipment was improved for quantitative measurements by addition of an external detector. The cross-sectional area of the milk vein was measured in two ways. Vein diameter was measured with a slide gauge and cross-sectional area calculated assuming a circular cross-section, and cross-sectional area was calculated both by planimetry and by using the actually measured radia. The present paper presents the accuracy and precision of the methods.
385
K. CHRISTEWSENPI al.
386 MATERIALS
AND METHODS
Animals Norwegian lactating goats from the herd of the institute were used for the present methodological studies. The goats were in their 2nd or 3rd lactation. The animals were milked twice a day (778 a.m. and 3-S p.m.) and the milk yield was recorded. They were housed in individual stalls and fed concentrates according to milk yield in two equal portions at the time of milking and received hay czd libitum. They had free access to water. The goats were standing in their milking chairs during the measurements to avoid agitation. which might influence mammary blood flow (Linzell and Rasmussen, 1972). Movement of the body. bleating, rumination or rapid respiration should be minimized. Therefore, usual handling of the animals is essential. Measurement
of blood ce1ocit.v
The apparatus and the internal detector. The velocity of the blood was measured by the ultrasound continuous wave Doppler principle with an apparatus called the Angio~tic AC2, delivered by Medimatic, Copenhagen, Denmark. It was equipped with an 8.2 MHz transducer and a 3.5 Hz low pass filter. The signal was partly recorded in a memory oscilloscope with a capacity of 3 min continuous operation, partly on a two-channel hot pen recorder indicating blood flow to and from the probe. Calibration signals indicating blood velocities of 5-IO-2&40 cm/set were used at the end of each series of measurements. The apparatus was equipped with a loud speaker and headphones. It registered the average Doppler shift from the moving blood which can be transformed into average blood velocity (V,,) by the following expression:
where Afis the Doppler shift, FwUndthe velocity of ultrasound in blood (1550 mjsec), f; the emitted ultrasound frequency and 0, the angle between the ultrasound beam and the moving blood cells in the blood vessel (45”).
The apparatus had a lower cut-off frequency of I50 Hz and a high cut-off frequency of 5 kHz. The width of the near field of the ultrasound beam was 4mm. As shown above blood velocity is inversely proportional to cosine of the angle between the emitted ultrasound beam and the direction of the blood stream. The angle was fixated to 45 by placing the probe in a Plexiglas holder parallel to the vein as shown in Fig. I. This holder also improved contact between the transducer and the skin, as the gel used as contact medium was placed here. The Plexiglas holder was prepared by the Danish Institute of Biomedical Engineering, Glostrup, Denmark. In preliminary studies the detector (internal) was found to have some limitations for our purpose. As the diameter of the ultrasound beam of 4 mm was smaller than the diameter of the milk vein in lactating goats, which generally lies between 6-12 mm, the transducer actually registered a velocity between FX, and Y,,,. Furthermore, the noise level of the detector was found to be relatively high and the frequency filter caused erroneus measurements at low blood velocities. We therefore decided to improve the apparatus by connecting an external detector and only using the results from this detector. The esrernal detecfor. This was connected to the frequency analyser output of the equipment detecting the maximum Doppler shift from the blood. Anticipating a laminary flow profile of the blood in the milk vein we can then calculate PI,, as half the maximum velocity). This device is shown diagr~mmati~lly in Fig. 2. Block I consists of a variable gain amplifier which keeps the peak voltage in the Doppler shifted spectrum constant and a band pass filter, the center frequency of which can be moved between 100 100,000 Hz. This device performs a scan in the center frequency of the band pass filter. When the detector in Biock 2 during the frequency scan detects a level of 10% of maximum it reads out the frequency values in Volts and at the same time it resets the frequency scan. This device improved the signal to noise level by more than a factor of 5. Only results obtained with the external detector are presented in the following.
Fig. 1. The probe of the ultrasound Doppler for measurement of blood velocity through the udder of lactating goats was fixed in a Plexiglas holder at an angle (0) of 45’ There is a cavity in the Plexiglas above the probe to prevent the coupling gel from falling off. During the measurements the holder is kept in parallel with the vein and ample coupling gel is used so that the holder is positioned 3-5 mm beneath the skin, thus avoiding pressure from the holder on the vessel. The goats are standing in their milking chairs under normal conditions. No strain is put on the animals during the measurements.
Evaluation
INTERNAL
of mammary
blood
flow measurements
DETECTOR
in lactating
EXTERNAL
goats
387
DETECTOR
Ultrasound
I]--
Doppler
AC2
3. Reset
scan ss
Read
Fig. 2. Schematic
diagram
of the ultrasound Doppler set-up used in this work, showing external detection (see text for details).
A suitable place on the milk vein, as close to the udder as possible, was selected for the measurements. The skin was shaved and bathed in oil to improve contact with the scanning equipment and the ultrasound Doppler transducer. During all measurements of blood velocity and vessel cross-sectional areas the external pudendal veins were clamped manually as described by Reynolds et ul. (1968) to make sure that all mammary blood was leaving through the milk veins. Measurements of wssel cross-sectional area A suitable place for measuring blood flow on the milk vein was marked with a speed marker. Vein diameter was measured in two ways. In the first place a slide gauge with digital read-out, equipped with round tips, as shown in Fig. 3,
Fig. 3. Photo vein diameter
out
both internal
and
was used (Danish Institute of Biomedical Engineering, Glostrup, Denmark). The thickness of the skin around the vein was measured by pulling the skin free of the vein and measuring the double thickness of the skin (skin-fold) which was then subtracted from the measured diameter. This gives a measure of the outer diameter of the vein. The cross-sectional area of the vein was calculated as nr’, assuming a circular cross-section. In the second place we used an ultrasound scanning equipment for measurement of the cross-sectional area of the vein. This equipment (Aloka Echo Camera, model SSD-210 DX, Aloka Co., Ltd, Tokyo, Japan) was portable and easy to handle. It was equipped with a 5 MHz linear probe with 56mm scanning width, digital scan converter, 2-point dynamic focusing and freeze-frame function, distance measurement function and ID code display. On the
of the slide gauge with digital read-out and equipped with round tips for measurement of and skin-fold thickness above the vein. Outer vein diameter was calculated as width of vein plus skin less twice the skin thickness above the vein (skin-fold).
K. CHRISTENSEN et al.
388
Fig. 4. Cross-section of the milk vein of a goat photographed from the frozen ultrasound scanning Notice that the vessel wall is very thin and that no other vessel is visible.
frozen scanning image width and height of the vein could be measured with an accuracy of +O.S mm. The frozen image was photographed with a Polaroid camera. After exposure it was magnified and the cross-sectional area measured by electronic planimetry. A photo of the cross-section of the milk vein of a goat is shown in Fig. 4. Calculation of blood flow Blood flow was calculated as the product of blood velocity and cross-sectional area of the vein measured on the same spot of the animal. Thus, differences in blood flow could be evaluated for both sides, which is normally never
done with other methods. Also, total mammary blood flow could be related to total milk yield. Statistical analysis
The mean, standard deviation, and coefficient of variation (CL’, % = SD x loo/mean) were calculated for independent measurements. Comparisons of two independent measurements with two different methods or by two different persons were made using the paired r-test, as described by Box et al. (1978). The precision or reproducibility of the measurements was expressed as SDreprod x lOO/mean where (V = variance of two or three indepen8Dreprod was m dent repetitions and N = the number of observations). RESULTS BIood velocity
The external detector measures maximal blood velocity. Mean blood velocity (V,,) is calculated as
image.
half the maximal velocity anticipating a laminary stream in the veins. By means of the headphones or loudspeakers any changes in bloodstreaming can be detected and another place on the vein can be chosen until a smooth, clear sound is heard. A typical example of blood velocity measurements on a conscious standing goat is shown in Fig. 5. The accuracy of the readings corresponds to 2% of measured blood velocity. It is essential that the animal stands still and that the belly does not move too much as a consequence of rapid respiration, bleating or rumination. The external detector was calibrated in vitro with blood stabilized with 6 g EDTA per litre by means of a continuous pumping system, as described by Rasmussen (1987b). The internal diameter of the PVC tube was 7 mm. The amount of blood running through the tube within a certain time was weighed (VJ. The calibration curve is shown in Fig. 6. For all velocities the calibration curve can be used to calculate the true Va,. The deviation from the 45” line means that deviation from the parabolic velocity profile becomes more and more pronounced as the velocity increases, in accordance with the fact that deviation from the Poiseuilli laminar flow is expected as Reynold’s number increases (Reynold’s number equals density x diameter x velocity/viscosity). We have also calibrated with perspex tubes of two other
Fig. 5. Recording from blood velocity measurement by means of ultrasound Doppler standing goat (external detector). The paper moves in the same direction as the arrow with a speed of 5 cmjsec.
in a conscious. indicated below
Evaluation of mammary blood flow measurements in lactating goats
389
flow to the same Reynold numbers as represented in Fig. 6. This means a scaling with both voscosity and diameter (the calibrations have been made at 24°C where the viscosity of blood is 2.8 x 10.3 Pa x set, whereas the viscosity at 37°C is 2.1 x 10.3 Pa x set). Another factor which might influence blood velocity is the angle (0) between the emitted ultrasound beam and the blood stream, as this is inversely proportional to I’,,,. By means of a Plexiglas holder especially prepared for the purpose, the angle was aimed to be kept at 45”, when the holder was kept in parallel with the vein (see Fig. 1). The magnitude of the error introduced by changing this angle is + 1.7% for a deflection of f l”, but an underestimation viz. overestimation of the angle by +5” would result in an error in blood velocity of 8.3% and 9.1%, respectively. In our opinion we are able to keep the angle within 45” + lo during the measurements on the animal, whereby the error is kept below +2%. The headphones help us to hear, when the Doppler shift is at optimum. A comparison of blood velocities measured every hour during the day between the two milkings on two consecutive days by a trained person neither showed significant variations during the day nor from day to day (P > 0.05). A typical example of such measurements is shown in Table 1. It is seen from this table that the total mammary blood velocity is rather constant during the day, whereas blood velocity of the same side is more variable. Variations expressed as coefficients of variation (CV, %) were 12-19% for the left milk vein and 12-24% for the right milk vein, whereas the total mean blood velocity during the day was rather constant for the 2 days with a CV of 13%. As we did not notice systematic variations in blood velocities during the day, we routinely measured twice a day, viz. 9.30-11 a.m. and 1.3&3 p.m. A comparison of the measurements between two trained persons did not show significant differences. Training is, however, necessary in order to obtain a satisfactory repeatability of the measurements. The results from repeated measurements of blood velocities in two goats at different stages of the
Fig. 6. Velocity calibration of the external detector using whole blood. VWis the velocity as determined by weighing the amount of blood flow in a certain time interval and using the cross-section of the tube. V, is the Doppler measured average velocity assuming a Pouiseuilli flow pattern, i.e. VW= vn&.
diameters, of 9.8 and 22mm diameter. With these tubes we also tested the importance of the distance from the Doppler transducer to the water inlet between 0.1 and 1.2 m. No effect on the Doppler measured velocity relative to the velocity measured by weighing from different distances between the transducer and the point of water inlet could be observed. It was noted that different tube diameters gave different calibration curves. If, however, the velocities measured from the 9.8 and 22 mm diameter tubes were scaled down to the tube of 7mm by a factor 7/9.8 and 7/22, respectively, all three calibration curves could be virtually superimposed on the curve shown in Fig. 6. This is consistent with the fact that, for the same value of Reynold’s number, the flow pattern should be identical. The consequences of these observations are that for an actual flow in a vein one corrects the Doppler measured V,, by scaling the
Table I. Variations during the day and between two consecutive days in Doppler measured average blood velocity (V,,) and calculated blood flow through the left and right milk vein (the caudal superficial epigastric or subcutaneous abdominal vein) of a lactating goat Flow (ml/min)
V,, (cm/W Left
~__-~__-~ Right ~~__
Total
~~ Left
Right ___~
__Total
I
2
I
2
I
2
I
2
I
2
I
2
Time 8.30 9.30 10.30 II.30 12.30 13.30 14.30 15.30
12.0 16.5 II.0 12.0 II.0 12.5 8.5 10.5
Il.5 12.0 14.0 9.5 12.5 13.0 12.0 14.5
13.0 14.0 I I.0 II.0 II.0 12.5 Il.0 14.5
15.5 13.5 12.0 8.5 9.5 II.5 8.5 8.5
25.0 30.5 22.0 23.0 22.0 25.0 19.5 25.0
27.0 25.5 26.0 18.0 22.0 24.5 20.5 23.0
331 455 304 331 304 345 235 290
317 331 386 262 345 359 331 400
289 311 244 244 244 278 244 322
344 300 266 I89 211 255 I89 189
620 766 548 575 548 623 479 612
661 631 652 451 556 614 520 589
Mean SD cv (%)
II.8 2.28 19.4
12.4 I.55 12.5
12.3 I .46 Il.9
10.9 2.64 24.2
24.0 3.26 13.6
23.3 3.05 13.1
324 63 19.4
341 43 12.6
272 33 12.0
243 58 24.0
596 84 14.1
584 72 12.3
Dav no.
Blood flow was calculated as the product of mean velocity and cross-sectional area of by means of ultrasound scanning. Cross-sectional area was 0.46 + 0.04 and 0.37 f and right vein, respectively (mean values f SD, IV = 4/vein). The pudendal vein of was clamped manually during the measurements. Daily milk yield was I.61 and yield ratio 532: I.
each vein, measured 0.07 cm* for the left the appropriate side blood flow to milk
K. CHRISTENSENet al.
390
Table 2. Doppler measured blood velocities and calculated blood flow through the left and right milk vein (the caudal superficial epigastric or subcutanous abdominal vein) of two goats at different stages of the lactation ocriod Stage of lactation (weeks)
V,, (cm/se@ _______ Left Right Total
Blood flow (ml/min) ~---Left Right Total
Milk yield (ml/24 hr)
Mammary blood flow/milk (yield ratio) -_
3363 3623 3389 2200 2072 I988 1607
295:l 206: I 226: I 259: I 302: I 430: I 438: I
Goat I
I2 I3 I4 21 25 28 33
14.0 10.3 13.5 8.0 7.0 10.2 10.3
13.5 10.8 8.5 8.0 10.5 13.8 9.7
27.5 21.3 22.0 16.0 17.5 24.0 20.0
319 235 308 I82 160 233 235
356 285 224 211 277 364 256
675 520 532 393 437 597 491
Goat 2 386: I I515 258:l I677 l9l:l 1637 303:l 1I I4 325: I 903 dry Blood flow was calculated as the product of mean velocity (I’,,) and cross-sectional area of each vein measured by means of ultrasound scanning (Goat I: left vein 0.38 cm*, right vein 0.44 cm*; Goat 2: left vein 0.13 cm2, right vein 0.29 cm*). The pudenda1 vein of the appropriate side was clamped manually during the measurements. I2 I3 I5 21 25 36
17.5 II.5 Il.8 1.5 10.0 3.6
16.3 12.0 7.3 10.0 1.3 3.8
33.8 23.5 19.1 17.5 17.3 1.4
137 90 92 59 78 28
lactation period are shown in Table 2. As statistical analysis did not show significant differences between the measurements in the morning and the afternoon, the mean results of blood velocity measurements have been shown. The reproducibility of the measurements performed in the morning and the afternoon was 16.2%. Cross-sectional area of the milk vein
In order to calculate blood flow (ml/min) the cross-sectional area of the milk vein must be measured. The cross-sectional area may be assumed circular and in that case calculated as nr2 or it may be elliptical and calculated as nr,r2. The use of an especially prepared slide gauge (see Fig. 2) was based on the assumption that the crosssectional area was circular. The diameter of the vessel and the skin surrounding it was measured and the thickness of the skin fold was then subtracted to give a measure of the outer diameter (D,). The internal diameter (Q) could then be calculated as D,, less twice vessel wall thickness. As shown in Fig. 4 it is evident that the wall of the milk vein is thin and hardly visible. In our experience from autopsy, the thickness of the wall of the milk vein, including adhering adipose and fat tissue in lactating goats, is between 0.01 and 0.2 mm, but it may vary along the length of the vein. Some fat may be deposited in the connective tissue during the lactation period, especially in low yielding animals. In goats, the discrepancy between the outer and inner diameter would be 0.2-Y!! which would amount to an overestimation of the cross-sectional area of &lo%. In 10 lactating goats the mean vein diameter (D,) of the left and right milk vein was 8.8 f 1.l mm (r f SD, CV = 12.3%). We routinely measure diameter and skin-fold thickness five times, performed independently of each other, subtract the highest and lowest value and then use the mean of three measurements. Within each mean, based on three measurements the maximum difference was an average of 0.47 mm (range 0.14-l .28 mm) in lactating goats, which would correspond to an overestimation
284 209 126 174 127 65
421 299 218 233 205 93
of the cross-sectional area by 11.0% (range 3.231.2%). The mean reproducibility of the measurements of vein diameter was 2.9%. The thickness of the skin-fold over the milk vein of the above-mentioned goats was an average of 3.14 f 0.28 mm (CV = 9.0%) with a reproducibility of the measurements of 2.7%. By ultrasound scanning of the milk vein the crosssectional area was partly calculated as zr,r2 on the frozen images by means of the built-in calipers, partly on photos magnified and measured by electronic planimetry. A comparison of the two methods of calculations performed on 37 measurements in goats did not show significant differences (P > 0.05). To facilitate the measurements, we therefore decided to use the directly available data from the frozen scanning images for the calculations. A comparison of the cross-sectional area measured by means of the slide gauge and by ultrasound scanning in a total of 29 independent measurements in goats showed mean values + SD of 0.43 f 0.29 cm2 and 0.22 f 0.14 cm*, respectively, the difference being significant (P < O.Ol), as tested by the paired t-test. In goats, however, with clearly protruding veins, the two methods did not show significant differences (P > 0.05). If the veins are large and protruding, the scanning images show that the cross-sectional area is circular and that little connective and fat tissue surround the vein. Such veins are typical in high-yielding animals. In low-yielding animals, however, the cross-sectional area is elliptical, probably because the connective tissue stretches out the vein and tightens it to the belly. In our goats, r2 did not exceed r, by more than 0.5 mm, which, by assuming a circular cross-sectional area instead of an elliptical one, would result in an underestimation of the cross-sectional area by 12.5% for a vein radius of 4mm. Blood jo w
Blood flow was calculated as blood velocity (cm/set) times cross-sectional area (cm2) and was
Evaluation af mammary blood flow measurements in lactating goats
expressed in ml/min, as the specific gravity of the blood was taken into consideration in establishing the calibration curve. From the results of the measurements of blood velocity and cross-sectional area we concluded that we should make use of the external detector for measuring mean blood velocity (V,,) on both sides of the animal, while the external pudendal veins were clamped manually to make sure that all mammary blood was leaving through the milk veins. The crosssectional area of the milk veins should be measured by means of ultrasound scanning while clamping the milk veins. A typical result from repeated measurements in a goat with clearly protruding milk veins on both sides is shown in Table I. The measurements were performed every hour from half an hour after the morning milking to half an hour before the afternoon milking on two consecutive days by the same person. Mammary blood flow was expressed in mljmin and in relation to milk yield. The reproducibility of the measurements was greater for the total blood flow (CV = 12-14%) than for the blood flow of each side (CV = 12-24%). The ratio between mammary blood flow (MBF) and milk flow (MF) was 532: 1. As we did not find systematic changes in the blood flow measurements performed every hour from half an hour after the morning milking to half an hour before the afternoon milking, we routinely measure blood flow twice a day, ciz. 2 hr after the morning milking and 2 hr before the afternoon milking. This procedure gave a ~pr~u~biIity of 16.2% for the measurements performed on the two goats shown in Table 2, which was a slightly lower reproducibility than obtained when the goats were standing in their milking chairs during the measurements. Goat I (Table 2) had clearly protruding veins on both sides. It yielded from 3.4 1.6 I milk daily during the lactation period. Blood flow decreased from 675491 mI/min during the same period, while MBF:MF increased from 295: I to 438:l. in this goat, blood Row through each side of the mammary gland was almost identical and so was the milk ilow. We found that 49% of the daily milk flow ran through the left udder half and 51% through the right haif. Goat 2 (Table 2) had relatively small milk veins and the left milk vein diameter was only 41% of that of the right milk vein. The daily milk yield decreased from 1.6 I ta zero during the lactation period as shown in Table 2. The blood Bow was constantly lower through the left than the right milk vein as expected from their different sizes. Total mammary blood flow decreased from 421-93 mI/min and MBF:MF varied between 386: I-191:1. We found that on average 33% of the daily milk flow ran through the left half of the udder and 67% through the right half, which corresponded well with the different blood flows observed in this goat. DISCUSSION
Blood velocity through the mammary gland of goats and diameters, and hence cross-sectional areas, of the milk vein have not been determined by others, so we are not able to compare the present results with
391
those obtained by other authors. However, the results will be discussed indirectly via the calculated mammary blood flows, which have been determined with different methods, as already mentioned in the intr~uction. These methods, however, only measure blood flow of one side anticipating that that of the other side is identical. As shown in our studies, however, this may not necessarily be the case. The external detector which was introduced measured maximal blood velocity. Average blood velocity was cakcuhtted as V-/2, antici~ting laminary streaming in the veins. Measuring V,, instead of Paygives the following advantages: (1) The noise level at the maximal Doppler shift frequency is much less than the noise at the lower frequencies. (2) The lowest velocities are at the limit of electronic noise and this produces some uncertainty in the dete~ination of V_. (3) The dete~ination of V,,,_ does not depend on the vessel diameter in contrast to Vay,where velocities from flow outside the width of the ultrasound beam is cut off. The angle between the emitted ultrasound beam and the blood stream in the vessel is critical for accurate measurements of blood velocity as described in the ~ateriais and ~eth~s section. As the milk vein is superficial and clearly visible, the Plexiglas holder with the transducer (Fig. 1) secured that this angle could be kept at 45” rf: I’, whereby the error was kept below f 2%. The results obtained with the two methods of measuring milk vein diameter clearly demonstrated that the ultrasound scanning method is obIigatory for quantitative measurements. In case of clearly protruding veins, however, the slide gauge gave similar results to the scanning method, but in less protruding veins the slide gauge significantly (P < 0.01) overestimated the cross-sectional area because the vein diameter was overestimated. The ultrasound scanning of such veins showed that the cross-section was elliptical and not circular as in clearly protruding veins, and therefore the cross-sectional area should be calculated as xr,r2 and not as w*, which was anticipated in the calculations using the slide gauge. As shown in Fig. 4, the vessel wall of the milk vein is hardly visible and the error introduced by using the outer vein diameter instead of the inner vein diameter is to our experience between 0.2 and S!!, which would lead to an overestimation of the cross-sectional area of O-IO%. How do the present calculated blood flows correspond with blood flows reported for goats measured with traditional invasivemeth~s~Therepr~ucibiIity of the total mammary blood flow was 13%, when the goat was measured every hour from half an hour after the morning milking to half an hour before the afternoon milking (Table I), but 16% when the measurements were performed twice a day, niz. 2 hr after the morning milking and 2 hr before the afternoon milking, indicating that the animal may not have been fully adapted to the situation. These results are comparable to the findings of Burvenich (1980) who, in lactating goats, observed variations in arterial mammary blood Aow from day to day of 10% and from hour to hour of 16.5% as measured by the electromagnetic flow method. Reynolds et al. (1968) found that fluctuations in the mammary
392
K.
CHRISTENSEN et al.
arterial flow were seldom greater than 10% of the mean flow observed with the electromagnetic method. Gill (1984) reported a variation of blood flow through the umbilical vein in humans of 14%, while Rasmussen (1987b) observed a variation of 6.8%. The Doppler measuring technique is more sensitive to movements of the belly than the other methods. Consequently, in the conscious, standing goat a greater variation may be expected than in humans lying and breathing quietly. As reviewed by Linzell and Rasmussen (1972) mammary blood flow in goats ranges from about 20 ml/min in a virgin to just under 2 l/min in a high yielding animal at peak lactation. Rasmussen (1963) observed a blood flow of 40-96 ml/min during one gland of goats yielding 35&400 ml milk daily with the antipyrine absorption method. As seen in Table 2, our values for total mammary blood flow range from 93-675 ml/min, which, according to milk yield, seem to lie within the range observed by the above-mentioned authors. According to Linzell (1966) the ratio of mammary blood flow to milk yield over most of the lactation period was found to be 493 k 15 : 1 in goats yielding a mean of 135 ml/min/lOO g udder, which corresponds to milk yields of l-l .5 I daily. Recently, values between 590 and 410: 1 were published with the thermodilution method for goats yielding l-2 1 milk daily (Mepham et al., 1984). It can be calculated that in a goat yielding 1.5 1 milk daily and with a milk vein diameter of each side of 7 mm, one would expect a mean mammary blood flow of 521 ml/min, anticipating a blood flow to milk yield ratio of 500: I, which would result in a mean mammary blood velocity of 22.4 cm/set or 11.2 cmjsec through each side. This was achieved with the goat in Table I, who had clearly protruding, relatively large veins in each side. This was also the case for goat 1 in Table 2 in the later stages of lactation, when the milk yield was about 1.6 1 daily. However, in goat 2 of Table 2 yielding 1.6 1 milk daily, the ratio was between 200 and 400: 1. The reason for this finding, which is characteristic in goats with relatively small veins, may be that these veins are unable to carry all the blood, when the pudendal veins are clamped manually. The same may hold true for goats yielding more than 2 1 milk daily. As seen in Table 2, goat 1 yielded 3.6 1 milk daily at midlactation and the ratio of mammary blood flow to milk flows was about 300: 1. To our knowledge, no results are available for goats yielding more than 2 1 milk daily. Apparently, the blood flow to milk flow ratio is lower at higher yields, which may indicate a more efficient extraction of nutrients at higher yields. If the ratio of blood flow to milk yield should be 500: 1 in a goat yielding 3.5 1 milk daily and with a mean cross-sectional area of each milk vein of 0.41 cm’, the mean blood velocity would be a total of 49.4 cm/set or 24.7/set of each side (Table 2). Whether these values obtained by our Dopplerequipment are under-estimated, whether the clamping technique represents a limitation, or whether they are physiologically correct cannot be evaluated at the present. However, we believe the conclusion- to be justified through
that measurement of blood velocities the milk veins of lactating goats by means of
the ultrasound Doppler method here and measurement of the milk vein cross-sectional area by means of ultrasound scanning may be an alternative method of mammary blood flow measurements to the invasive methods commonly used. By selecting goats with large milk veins which are capable of carrying all the blood leaving the udder, when the pudendal veins are clamped manually, the precision and accuracy of the method may be further improved. Acknowledgements-The
present studies were supported by a grant from The Danish Agricultural and Veterinary Research Council. For fruitful discussions during the work, the authors wish to thank Dr Knud Rasmussen, M. D. Herlev Hospital and Prof. Dr med. vet. Folke Rasmussen, The Royal Veterinary and Agricultural University, Copenhagen. Mr Hans Busk BSc, The National Institute of Animal Science, Forssgsanlreg Foulum, kindly lent us his scanning equipment and has performed the planimetry of the photographs for which we are very grateful. REFERENCES
Box G. E. P., Hunter W. G. and Hunter J. S. (1978) Statistics for Experimenters, Chapt. 4. John Wiley and Sons, New York. Burvenich C. (1980) Variations of mammary artery blood flow and milk yield under normal conditions and during the oestrous cycle of the dairy goat. Z. Tierphysiol., Tiererniihrg. u. Futlermittelkde 43, 18-27.
Fleet I. R. and Mepham T. B. (1983) Physiological methods used in the study of mammary substrate utilization in ruminants. In Biochemistry of lactation (Edited by Mepham T. B.), pp. 469491. Elsevier, Amsterdam. Gill R. W. (1985) Measurement of blood flow by ultrasound: accuracy and sources of error. Ultrasound Med. Biol. 11, 625641.
Linzeli J. L. (1966) Measurement of venous flow by continuous thermodilution and its application to measurement of mammary blood flow in the goat. Circulation Res. 18, 745-754.
Linzell J. L. (1974) Mammary blood flow and methods of identifying and measuring precursors of milk. In Lactation (Edited by Larson B. L. and Smith V. R.), Vol. 1, pp. 143-225. Academic Press, New York. Linzell J. L. and Rasmussen F. (1972) Mammary blood flow and changes in circulation during lactation in goats and cows. In Hand&h der Tiererniihrung, II (Edited by Lenkeit W. and Breirem K.). Verlag Paul Parey, Hamburg. Mepham T. B., Lawrence S. E., Peters A. R. and Hart I. C. (1984) Effects of exogenous growth hormone on mammary function in lactating goats. Horm. metabol. Res. 16, 248-253.
Rasmussen F. (1963) The mammary blood flow in the goat as measured by antipyrine absorption. Acta vet. &and. 4, 271-280.
Rasmussen K. (1987a) Methodological problems related to measurement of quantitative blood flow with the ultrasound Doppler technique. &and. J. c/in. Lab. Invest. 47, 303-309.
Rasmussen K. (1987b) Precision and accuracy of Doppler flow measurements. In vitro and in viuo study of the applicability of the method in human fetuses. Stand. J. Clin. Lab. Invest 47, 311-318.
Rasmussen K. (1987~) Quantitative blood flow in the fetal descending aorta and in the umbilical vein in normal pregnancies. Longitudinal and cross-sectional studies. Stand. J. clin. Lab. Invesr. 47. 319-324.
Reynolds M. Linzell J. L. and Rasmussen F. (1968) Comparison of four methods for measuring mammary blood flow in conscious goats. Am. J. Physiol. 214, 1415-1424.
0300-9629189 $3.00+ 0.00 0 1989Pergamon Press plc
Printed in Great Britain
EVALUATION OF MAMMARY BLOOD FLOW MEASUREMENTS IN LACTATING GOATS USING THE ULTRASOUND DOPPLER PRINCIPLE K. CHRISTENSEN,* M. 0. NIELSEN, R. BAKER? and K. HILDEN~ The Royal Veterinary and Agricultural University, Department of Animal Physiology and Bi~hemistry and tDepartment of Physics, Thorvaldsensvej 40, DK-I871 Frederiksberg C, Denmark. Telephone: 06-65-2500 (Received
18 August
1988)
Abstract-l. Non-invasive methods were developed for measuring mammary blood flow in lactating goats. 2. A Doppler principle ultrasound device was equipped with an external detector measuring maximal blood velocity (V,,) and average blood velocity (V,,) was calculated as V,,/2. Volume flow then depended on determination of the angle of insonation and the cross-sectional area of the milk vein (the caudal superficial epigastric or subcutaneous abdominal vein). 3. Blood velocities were measured on the milk vein of either side of the animal while clamping the pudendal veins manually. Blood velocities ranged from 7-34cm/sec. 4. The milk vein diameter was measured by means of a slide gauge which, for clearly protruding veins, gave similar results to that measured by ultrasound scanning. In protruding veins the cross-section was circular. In non-protruding veins the cross-section was elliptical and the slide gauge significantly (P < 0.01) overestimated the cross-sectional area. The milk vein diameter of either side measured in 10 lactating goats was 8.8 f 1.1 mm (R + SD). 5. Blood flow ranged from 90-675 ml/min in a dry and a high-yielding (3.4 1 milk daily) goat, respectively. The reproducibility of the blood flow measurements was 12-16%. 6. It is concluded that the present method may be used for quantitative measurements of mammary blood flow in goats.
INTRODUCTlON Rates of mammary blood flow in ruminants have been measured by direct or indirect techniques as reviewed by Reynolds et al. (I 968), Linzell(1974) and Fleet and Mepham (1983), all of which require surgical incisions prior to the measurements, tk. division of the udder halves, preparation of a milk vein loop or placement of a flow transducer around the external pudendal artery. The recently available Transonic flowmeter (Transonic Systems Inc., New York) using ultrasound based on transit time principles is also for implantation. Ultrasonic methods exploiting the Doppler shift principle were also tried (Linzell, 1974), but apparently no results are available. The ultrasound Doppler principle has been used widely within human medicine for diagnostic and clinical purposes. Quantitative flow measurements have been difficult to perform because the vessels are normally deep-lying, whereby it is difficult to locate the vessel and measure the angle between the moving bloodstream and the ultrasound beam from the transducer. Quantitative measurements with the Doppler method in human fetuses have recently been carried out both with high accuracy and high precision (Rasmussen, 1987b) and the methodological *Present address: National Institute of Animal Science, Department of Animal Physiology and Biochemistry, Forsragsanlaeg Foulum, PO Box 39, DK-8833 0rum Sdrl., Denmark.
problems involved have been discussed in detail (Gill, 1985; Rasmussen, 1987a). The Doppler principle ultrasound equipment measures blood velocity, and volume flow is dependent on determination of the angle of insonation and the vessel cross-sectional area. Although this method so far has not fulfilled the requirements for exact mammary blood flow measurements in ruminants, it is still challenging because it is the only method which allows direct measurements without any surgical incisions. In lactating ruminants the milk vein (the caudal superficial epigastric or subcutaneus abdominal vein) is often easily observed, because it protrudes beneath the skin just anterior to the udder. Therefore, we decided to try the ultrasound Doppler principle to measure blood velocity through the udder of goats. Blood flow could then be calculated as the product of blood velocity and cross-sectional area of the milk vein. This method allowed measurements on both sides of the animal. We used commercially available hi-directional ultrasound Doppler equipment (Angiomatic AC2) originally built for non-invasive diagnosis of peripheral venous and supra-aorta1 diseases in humans. The equipment was improved for quantitative measurements by addition of an external detector. The cross-sectional area of the milk vein was measured in two ways. Vein diameter was measured with a slide gauge and cross-sectional area calculated assuming a circular cross-section, and cross-sectional area was calculated both by planimetry and by using the actually measured radia. The present paper presents the accuracy and precision of the methods.
385
K. CHRISTEWSENPI al.
386 MATERIALS
AND METHODS
Animals Norwegian lactating goats from the herd of the institute were used for the present methodological studies. The goats were in their 2nd or 3rd lactation. The animals were milked twice a day (778 a.m. and 3-S p.m.) and the milk yield was recorded. They were housed in individual stalls and fed concentrates according to milk yield in two equal portions at the time of milking and received hay czd libitum. They had free access to water. The goats were standing in their milking chairs during the measurements to avoid agitation. which might influence mammary blood flow (Linzell and Rasmussen, 1972). Movement of the body. bleating, rumination or rapid respiration should be minimized. Therefore, usual handling of the animals is essential. Measurement
of blood ce1ocit.v
The apparatus and the internal detector. The velocity of the blood was measured by the ultrasound continuous wave Doppler principle with an apparatus called the Angio~tic AC2, delivered by Medimatic, Copenhagen, Denmark. It was equipped with an 8.2 MHz transducer and a 3.5 Hz low pass filter. The signal was partly recorded in a memory oscilloscope with a capacity of 3 min continuous operation, partly on a two-channel hot pen recorder indicating blood flow to and from the probe. Calibration signals indicating blood velocities of 5-IO-2&40 cm/set were used at the end of each series of measurements. The apparatus was equipped with a loud speaker and headphones. It registered the average Doppler shift from the moving blood which can be transformed into average blood velocity (V,,) by the following expression:
where Afis the Doppler shift, FwUndthe velocity of ultrasound in blood (1550 mjsec), f; the emitted ultrasound frequency and 0, the angle between the ultrasound beam and the moving blood cells in the blood vessel (45”).
The apparatus had a lower cut-off frequency of I50 Hz and a high cut-off frequency of 5 kHz. The width of the near field of the ultrasound beam was 4mm. As shown above blood velocity is inversely proportional to cosine of the angle between the emitted ultrasound beam and the direction of the blood stream. The angle was fixated to 45 by placing the probe in a Plexiglas holder parallel to the vein as shown in Fig. I. This holder also improved contact between the transducer and the skin, as the gel used as contact medium was placed here. The Plexiglas holder was prepared by the Danish Institute of Biomedical Engineering, Glostrup, Denmark. In preliminary studies the detector (internal) was found to have some limitations for our purpose. As the diameter of the ultrasound beam of 4 mm was smaller than the diameter of the milk vein in lactating goats, which generally lies between 6-12 mm, the transducer actually registered a velocity between FX, and Y,,,. Furthermore, the noise level of the detector was found to be relatively high and the frequency filter caused erroneus measurements at low blood velocities. We therefore decided to improve the apparatus by connecting an external detector and only using the results from this detector. The esrernal detecfor. This was connected to the frequency analyser output of the equipment detecting the maximum Doppler shift from the blood. Anticipating a laminary flow profile of the blood in the milk vein we can then calculate PI,, as half the maximum velocity). This device is shown diagr~mmati~lly in Fig. 2. Block I consists of a variable gain amplifier which keeps the peak voltage in the Doppler shifted spectrum constant and a band pass filter, the center frequency of which can be moved between 100 100,000 Hz. This device performs a scan in the center frequency of the band pass filter. When the detector in Biock 2 during the frequency scan detects a level of 10% of maximum it reads out the frequency values in Volts and at the same time it resets the frequency scan. This device improved the signal to noise level by more than a factor of 5. Only results obtained with the external detector are presented in the following.
Fig. 1. The probe of the ultrasound Doppler for measurement of blood velocity through the udder of lactating goats was fixed in a Plexiglas holder at an angle (0) of 45’ There is a cavity in the Plexiglas above the probe to prevent the coupling gel from falling off. During the measurements the holder is kept in parallel with the vein and ample coupling gel is used so that the holder is positioned 3-5 mm beneath the skin, thus avoiding pressure from the holder on the vessel. The goats are standing in their milking chairs under normal conditions. No strain is put on the animals during the measurements.
Evaluation
INTERNAL
of mammary
blood
flow measurements
DETECTOR
in lactating
EXTERNAL
goats
387
DETECTOR
Ultrasound
I]--
Doppler
AC2
3. Reset
scan ss
Read
Fig. 2. Schematic
diagram
of the ultrasound Doppler set-up used in this work, showing external detection (see text for details).
A suitable place on the milk vein, as close to the udder as possible, was selected for the measurements. The skin was shaved and bathed in oil to improve contact with the scanning equipment and the ultrasound Doppler transducer. During all measurements of blood velocity and vessel cross-sectional areas the external pudendal veins were clamped manually as described by Reynolds et ul. (1968) to make sure that all mammary blood was leaving through the milk veins. Measurements of wssel cross-sectional area A suitable place for measuring blood flow on the milk vein was marked with a speed marker. Vein diameter was measured in two ways. In the first place a slide gauge with digital read-out, equipped with round tips, as shown in Fig. 3,
Fig. 3. Photo vein diameter
out
both internal
and
was used (Danish Institute of Biomedical Engineering, Glostrup, Denmark). The thickness of the skin around the vein was measured by pulling the skin free of the vein and measuring the double thickness of the skin (skin-fold) which was then subtracted from the measured diameter. This gives a measure of the outer diameter of the vein. The cross-sectional area of the vein was calculated as nr’, assuming a circular cross-section. In the second place we used an ultrasound scanning equipment for measurement of the cross-sectional area of the vein. This equipment (Aloka Echo Camera, model SSD-210 DX, Aloka Co., Ltd, Tokyo, Japan) was portable and easy to handle. It was equipped with a 5 MHz linear probe with 56mm scanning width, digital scan converter, 2-point dynamic focusing and freeze-frame function, distance measurement function and ID code display. On the
of the slide gauge with digital read-out and equipped with round tips for measurement of and skin-fold thickness above the vein. Outer vein diameter was calculated as width of vein plus skin less twice the skin thickness above the vein (skin-fold).
K. CHRISTENSEN et al.
388
Fig. 4. Cross-section of the milk vein of a goat photographed from the frozen ultrasound scanning Notice that the vessel wall is very thin and that no other vessel is visible.
frozen scanning image width and height of the vein could be measured with an accuracy of +O.S mm. The frozen image was photographed with a Polaroid camera. After exposure it was magnified and the cross-sectional area measured by electronic planimetry. A photo of the cross-section of the milk vein of a goat is shown in Fig. 4. Calculation of blood flow Blood flow was calculated as the product of blood velocity and cross-sectional area of the vein measured on the same spot of the animal. Thus, differences in blood flow could be evaluated for both sides, which is normally never
done with other methods. Also, total mammary blood flow could be related to total milk yield. Statistical analysis
The mean, standard deviation, and coefficient of variation (CL’, % = SD x loo/mean) were calculated for independent measurements. Comparisons of two independent measurements with two different methods or by two different persons were made using the paired r-test, as described by Box et al. (1978). The precision or reproducibility of the measurements was expressed as SDreprod x lOO/mean where (V = variance of two or three indepen8Dreprod was m dent repetitions and N = the number of observations). RESULTS BIood velocity
The external detector measures maximal blood velocity. Mean blood velocity (V,,) is calculated as
image.
half the maximal velocity anticipating a laminary stream in the veins. By means of the headphones or loudspeakers any changes in bloodstreaming can be detected and another place on the vein can be chosen until a smooth, clear sound is heard. A typical example of blood velocity measurements on a conscious standing goat is shown in Fig. 5. The accuracy of the readings corresponds to 2% of measured blood velocity. It is essential that the animal stands still and that the belly does not move too much as a consequence of rapid respiration, bleating or rumination. The external detector was calibrated in vitro with blood stabilized with 6 g EDTA per litre by means of a continuous pumping system, as described by Rasmussen (1987b). The internal diameter of the PVC tube was 7 mm. The amount of blood running through the tube within a certain time was weighed (VJ. The calibration curve is shown in Fig. 6. For all velocities the calibration curve can be used to calculate the true Va,. The deviation from the 45” line means that deviation from the parabolic velocity profile becomes more and more pronounced as the velocity increases, in accordance with the fact that deviation from the Poiseuilli laminar flow is expected as Reynold’s number increases (Reynold’s number equals density x diameter x velocity/viscosity). We have also calibrated with perspex tubes of two other
Fig. 5. Recording from blood velocity measurement by means of ultrasound Doppler standing goat (external detector). The paper moves in the same direction as the arrow with a speed of 5 cmjsec.
in a conscious. indicated below
Evaluation of mammary blood flow measurements in lactating goats
389
flow to the same Reynold numbers as represented in Fig. 6. This means a scaling with both voscosity and diameter (the calibrations have been made at 24°C where the viscosity of blood is 2.8 x 10.3 Pa x set, whereas the viscosity at 37°C is 2.1 x 10.3 Pa x set). Another factor which might influence blood velocity is the angle (0) between the emitted ultrasound beam and the blood stream, as this is inversely proportional to I’,,,. By means of a Plexiglas holder especially prepared for the purpose, the angle was aimed to be kept at 45”, when the holder was kept in parallel with the vein (see Fig. 1). The magnitude of the error introduced by changing this angle is + 1.7% for a deflection of f l”, but an underestimation viz. overestimation of the angle by +5” would result in an error in blood velocity of 8.3% and 9.1%, respectively. In our opinion we are able to keep the angle within 45” + lo during the measurements on the animal, whereby the error is kept below +2%. The headphones help us to hear, when the Doppler shift is at optimum. A comparison of blood velocities measured every hour during the day between the two milkings on two consecutive days by a trained person neither showed significant variations during the day nor from day to day (P > 0.05). A typical example of such measurements is shown in Table 1. It is seen from this table that the total mammary blood velocity is rather constant during the day, whereas blood velocity of the same side is more variable. Variations expressed as coefficients of variation (CV, %) were 12-19% for the left milk vein and 12-24% for the right milk vein, whereas the total mean blood velocity during the day was rather constant for the 2 days with a CV of 13%. As we did not notice systematic variations in blood velocities during the day, we routinely measured twice a day, viz. 9.30-11 a.m. and 1.3&3 p.m. A comparison of the measurements between two trained persons did not show significant differences. Training is, however, necessary in order to obtain a satisfactory repeatability of the measurements. The results from repeated measurements of blood velocities in two goats at different stages of the
Fig. 6. Velocity calibration of the external detector using whole blood. VWis the velocity as determined by weighing the amount of blood flow in a certain time interval and using the cross-section of the tube. V, is the Doppler measured average velocity assuming a Pouiseuilli flow pattern, i.e. VW= vn&.
diameters, of 9.8 and 22mm diameter. With these tubes we also tested the importance of the distance from the Doppler transducer to the water inlet between 0.1 and 1.2 m. No effect on the Doppler measured velocity relative to the velocity measured by weighing from different distances between the transducer and the point of water inlet could be observed. It was noted that different tube diameters gave different calibration curves. If, however, the velocities measured from the 9.8 and 22 mm diameter tubes were scaled down to the tube of 7mm by a factor 7/9.8 and 7/22, respectively, all three calibration curves could be virtually superimposed on the curve shown in Fig. 6. This is consistent with the fact that, for the same value of Reynold’s number, the flow pattern should be identical. The consequences of these observations are that for an actual flow in a vein one corrects the Doppler measured V,, by scaling the
Table I. Variations during the day and between two consecutive days in Doppler measured average blood velocity (V,,) and calculated blood flow through the left and right milk vein (the caudal superficial epigastric or subcutaneous abdominal vein) of a lactating goat Flow (ml/min)
V,, (cm/W Left
~__-~__-~ Right ~~__
Total
~~ Left
Right ___~
__Total
I
2
I
2
I
2
I
2
I
2
I
2
Time 8.30 9.30 10.30 II.30 12.30 13.30 14.30 15.30
12.0 16.5 II.0 12.0 II.0 12.5 8.5 10.5
Il.5 12.0 14.0 9.5 12.5 13.0 12.0 14.5
13.0 14.0 I I.0 II.0 II.0 12.5 Il.0 14.5
15.5 13.5 12.0 8.5 9.5 II.5 8.5 8.5
25.0 30.5 22.0 23.0 22.0 25.0 19.5 25.0
27.0 25.5 26.0 18.0 22.0 24.5 20.5 23.0
331 455 304 331 304 345 235 290
317 331 386 262 345 359 331 400
289 311 244 244 244 278 244 322
344 300 266 I89 211 255 I89 189
620 766 548 575 548 623 479 612
661 631 652 451 556 614 520 589
Mean SD cv (%)
II.8 2.28 19.4
12.4 I.55 12.5
12.3 I .46 Il.9
10.9 2.64 24.2
24.0 3.26 13.6
23.3 3.05 13.1
324 63 19.4
341 43 12.6
272 33 12.0
243 58 24.0
596 84 14.1
584 72 12.3
Dav no.
Blood flow was calculated as the product of mean velocity and cross-sectional area of by means of ultrasound scanning. Cross-sectional area was 0.46 + 0.04 and 0.37 f and right vein, respectively (mean values f SD, IV = 4/vein). The pudendal vein of was clamped manually during the measurements. Daily milk yield was I.61 and yield ratio 532: I.
each vein, measured 0.07 cm* for the left the appropriate side blood flow to milk
K. CHRISTENSENet al.
390
Table 2. Doppler measured blood velocities and calculated blood flow through the left and right milk vein (the caudal superficial epigastric or subcutanous abdominal vein) of two goats at different stages of the lactation ocriod Stage of lactation (weeks)
V,, (cm/se@ _______ Left Right Total
Blood flow (ml/min) ~---Left Right Total
Milk yield (ml/24 hr)
Mammary blood flow/milk (yield ratio) -_
3363 3623 3389 2200 2072 I988 1607
295:l 206: I 226: I 259: I 302: I 430: I 438: I
Goat I
I2 I3 I4 21 25 28 33
14.0 10.3 13.5 8.0 7.0 10.2 10.3
13.5 10.8 8.5 8.0 10.5 13.8 9.7
27.5 21.3 22.0 16.0 17.5 24.0 20.0
319 235 308 I82 160 233 235
356 285 224 211 277 364 256
675 520 532 393 437 597 491
Goat 2 386: I I515 258:l I677 l9l:l 1637 303:l 1I I4 325: I 903 dry Blood flow was calculated as the product of mean velocity (I’,,) and cross-sectional area of each vein measured by means of ultrasound scanning (Goat I: left vein 0.38 cm*, right vein 0.44 cm*; Goat 2: left vein 0.13 cm2, right vein 0.29 cm*). The pudenda1 vein of the appropriate side was clamped manually during the measurements. I2 I3 I5 21 25 36
17.5 II.5 Il.8 1.5 10.0 3.6
16.3 12.0 7.3 10.0 1.3 3.8
33.8 23.5 19.1 17.5 17.3 1.4
137 90 92 59 78 28
lactation period are shown in Table 2. As statistical analysis did not show significant differences between the measurements in the morning and the afternoon, the mean results of blood velocity measurements have been shown. The reproducibility of the measurements performed in the morning and the afternoon was 16.2%. Cross-sectional area of the milk vein
In order to calculate blood flow (ml/min) the cross-sectional area of the milk vein must be measured. The cross-sectional area may be assumed circular and in that case calculated as nr2 or it may be elliptical and calculated as nr,r2. The use of an especially prepared slide gauge (see Fig. 2) was based on the assumption that the crosssectional area was circular. The diameter of the vessel and the skin surrounding it was measured and the thickness of the skin fold was then subtracted to give a measure of the outer diameter (D,). The internal diameter (Q) could then be calculated as D,, less twice vessel wall thickness. As shown in Fig. 4 it is evident that the wall of the milk vein is thin and hardly visible. In our experience from autopsy, the thickness of the wall of the milk vein, including adhering adipose and fat tissue in lactating goats, is between 0.01 and 0.2 mm, but it may vary along the length of the vein. Some fat may be deposited in the connective tissue during the lactation period, especially in low yielding animals. In goats, the discrepancy between the outer and inner diameter would be 0.2-Y!! which would amount to an overestimation of the cross-sectional area of &lo%. In 10 lactating goats the mean vein diameter (D,) of the left and right milk vein was 8.8 f 1.l mm (r f SD, CV = 12.3%). We routinely measure diameter and skin-fold thickness five times, performed independently of each other, subtract the highest and lowest value and then use the mean of three measurements. Within each mean, based on three measurements the maximum difference was an average of 0.47 mm (range 0.14-l .28 mm) in lactating goats, which would correspond to an overestimation
284 209 126 174 127 65
421 299 218 233 205 93
of the cross-sectional area by 11.0% (range 3.231.2%). The mean reproducibility of the measurements of vein diameter was 2.9%. The thickness of the skin-fold over the milk vein of the above-mentioned goats was an average of 3.14 f 0.28 mm (CV = 9.0%) with a reproducibility of the measurements of 2.7%. By ultrasound scanning of the milk vein the crosssectional area was partly calculated as zr,r2 on the frozen images by means of the built-in calipers, partly on photos magnified and measured by electronic planimetry. A comparison of the two methods of calculations performed on 37 measurements in goats did not show significant differences (P > 0.05). To facilitate the measurements, we therefore decided to use the directly available data from the frozen scanning images for the calculations. A comparison of the cross-sectional area measured by means of the slide gauge and by ultrasound scanning in a total of 29 independent measurements in goats showed mean values + SD of 0.43 f 0.29 cm2 and 0.22 f 0.14 cm*, respectively, the difference being significant (P < O.Ol), as tested by the paired t-test. In goats, however, with clearly protruding veins, the two methods did not show significant differences (P > 0.05). If the veins are large and protruding, the scanning images show that the cross-sectional area is circular and that little connective and fat tissue surround the vein. Such veins are typical in high-yielding animals. In low-yielding animals, however, the cross-sectional area is elliptical, probably because the connective tissue stretches out the vein and tightens it to the belly. In our goats, r2 did not exceed r, by more than 0.5 mm, which, by assuming a circular cross-sectional area instead of an elliptical one, would result in an underestimation of the cross-sectional area by 12.5% for a vein radius of 4mm. Blood jo w
Blood flow was calculated as blood velocity (cm/set) times cross-sectional area (cm2) and was
Evaluation af mammary blood flow measurements in lactating goats
expressed in ml/min, as the specific gravity of the blood was taken into consideration in establishing the calibration curve. From the results of the measurements of blood velocity and cross-sectional area we concluded that we should make use of the external detector for measuring mean blood velocity (V,,) on both sides of the animal, while the external pudendal veins were clamped manually to make sure that all mammary blood was leaving through the milk veins. The crosssectional area of the milk veins should be measured by means of ultrasound scanning while clamping the milk veins. A typical result from repeated measurements in a goat with clearly protruding milk veins on both sides is shown in Table I. The measurements were performed every hour from half an hour after the morning milking to half an hour before the afternoon milking on two consecutive days by the same person. Mammary blood flow was expressed in mljmin and in relation to milk yield. The reproducibility of the measurements was greater for the total blood flow (CV = 12-14%) than for the blood flow of each side (CV = 12-24%). The ratio between mammary blood flow (MBF) and milk flow (MF) was 532: 1. As we did not find systematic changes in the blood flow measurements performed every hour from half an hour after the morning milking to half an hour before the afternoon milking, we routinely measure blood flow twice a day, ciz. 2 hr after the morning milking and 2 hr before the afternoon milking. This procedure gave a ~pr~u~biIity of 16.2% for the measurements performed on the two goats shown in Table 2, which was a slightly lower reproducibility than obtained when the goats were standing in their milking chairs during the measurements. Goat I (Table 2) had clearly protruding veins on both sides. It yielded from 3.4 1.6 I milk daily during the lactation period. Blood flow decreased from 675491 mI/min during the same period, while MBF:MF increased from 295: I to 438:l. in this goat, blood Row through each side of the mammary gland was almost identical and so was the milk ilow. We found that 49% of the daily milk flow ran through the left udder half and 51% through the right haif. Goat 2 (Table 2) had relatively small milk veins and the left milk vein diameter was only 41% of that of the right milk vein. The daily milk yield decreased from 1.6 I ta zero during the lactation period as shown in Table 2. The blood Bow was constantly lower through the left than the right milk vein as expected from their different sizes. Total mammary blood flow decreased from 421-93 mI/min and MBF:MF varied between 386: I-191:1. We found that on average 33% of the daily milk flow ran through the left half of the udder and 67% through the right half, which corresponded well with the different blood flows observed in this goat. DISCUSSION
Blood velocity through the mammary gland of goats and diameters, and hence cross-sectional areas, of the milk vein have not been determined by others, so we are not able to compare the present results with
391
those obtained by other authors. However, the results will be discussed indirectly via the calculated mammary blood flows, which have been determined with different methods, as already mentioned in the intr~uction. These methods, however, only measure blood flow of one side anticipating that that of the other side is identical. As shown in our studies, however, this may not necessarily be the case. The external detector which was introduced measured maximal blood velocity. Average blood velocity was cakcuhtted as V-/2, antici~ting laminary streaming in the veins. Measuring V,, instead of Paygives the following advantages: (1) The noise level at the maximal Doppler shift frequency is much less than the noise at the lower frequencies. (2) The lowest velocities are at the limit of electronic noise and this produces some uncertainty in the dete~ination of V_. (3) The dete~ination of V,,,_ does not depend on the vessel diameter in contrast to Vay,where velocities from flow outside the width of the ultrasound beam is cut off. The angle between the emitted ultrasound beam and the blood stream in the vessel is critical for accurate measurements of blood velocity as described in the ~ateriais and ~eth~s section. As the milk vein is superficial and clearly visible, the Plexiglas holder with the transducer (Fig. 1) secured that this angle could be kept at 45” rf: I’, whereby the error was kept below f 2%. The results obtained with the two methods of measuring milk vein diameter clearly demonstrated that the ultrasound scanning method is obIigatory for quantitative measurements. In case of clearly protruding veins, however, the slide gauge gave similar results to the scanning method, but in less protruding veins the slide gauge significantly (P < 0.01) overestimated the cross-sectional area because the vein diameter was overestimated. The ultrasound scanning of such veins showed that the cross-section was elliptical and not circular as in clearly protruding veins, and therefore the cross-sectional area should be calculated as xr,r2 and not as w*, which was anticipated in the calculations using the slide gauge. As shown in Fig. 4, the vessel wall of the milk vein is hardly visible and the error introduced by using the outer vein diameter instead of the inner vein diameter is to our experience between 0.2 and S!!, which would lead to an overestimation of the cross-sectional area of O-IO%. How do the present calculated blood flows correspond with blood flows reported for goats measured with traditional invasivemeth~s~Therepr~ucibiIity of the total mammary blood flow was 13%, when the goat was measured every hour from half an hour after the morning milking to half an hour before the afternoon milking (Table I), but 16% when the measurements were performed twice a day, niz. 2 hr after the morning milking and 2 hr before the afternoon milking, indicating that the animal may not have been fully adapted to the situation. These results are comparable to the findings of Burvenich (1980) who, in lactating goats, observed variations in arterial mammary blood Aow from day to day of 10% and from hour to hour of 16.5% as measured by the electromagnetic flow method. Reynolds et al. (1968) found that fluctuations in the mammary
392
K.
CHRISTENSEN et al.
arterial flow were seldom greater than 10% of the mean flow observed with the electromagnetic method. Gill (1984) reported a variation of blood flow through the umbilical vein in humans of 14%, while Rasmussen (1987b) observed a variation of 6.8%. The Doppler measuring technique is more sensitive to movements of the belly than the other methods. Consequently, in the conscious, standing goat a greater variation may be expected than in humans lying and breathing quietly. As reviewed by Linzell and Rasmussen (1972) mammary blood flow in goats ranges from about 20 ml/min in a virgin to just under 2 l/min in a high yielding animal at peak lactation. Rasmussen (1963) observed a blood flow of 40-96 ml/min during one gland of goats yielding 35&400 ml milk daily with the antipyrine absorption method. As seen in Table 2, our values for total mammary blood flow range from 93-675 ml/min, which, according to milk yield, seem to lie within the range observed by the above-mentioned authors. According to Linzell (1966) the ratio of mammary blood flow to milk yield over most of the lactation period was found to be 493 k 15 : 1 in goats yielding a mean of 135 ml/min/lOO g udder, which corresponds to milk yields of l-l .5 I daily. Recently, values between 590 and 410: 1 were published with the thermodilution method for goats yielding l-2 1 milk daily (Mepham et al., 1984). It can be calculated that in a goat yielding 1.5 1 milk daily and with a milk vein diameter of each side of 7 mm, one would expect a mean mammary blood flow of 521 ml/min, anticipating a blood flow to milk yield ratio of 500: I, which would result in a mean mammary blood velocity of 22.4 cm/set or 11.2 cmjsec through each side. This was achieved with the goat in Table I, who had clearly protruding, relatively large veins in each side. This was also the case for goat 1 in Table 2 in the later stages of lactation, when the milk yield was about 1.6 1 daily. However, in goat 2 of Table 2 yielding 1.6 1 milk daily, the ratio was between 200 and 400: 1. The reason for this finding, which is characteristic in goats with relatively small veins, may be that these veins are unable to carry all the blood, when the pudendal veins are clamped manually. The same may hold true for goats yielding more than 2 1 milk daily. As seen in Table 2, goat 1 yielded 3.6 1 milk daily at midlactation and the ratio of mammary blood flow to milk flows was about 300: 1. To our knowledge, no results are available for goats yielding more than 2 1 milk daily. Apparently, the blood flow to milk flow ratio is lower at higher yields, which may indicate a more efficient extraction of nutrients at higher yields. If the ratio of blood flow to milk yield should be 500: 1 in a goat yielding 3.5 1 milk daily and with a mean cross-sectional area of each milk vein of 0.41 cm’, the mean blood velocity would be a total of 49.4 cm/set or 24.7/set of each side (Table 2). Whether these values obtained by our Dopplerequipment are under-estimated, whether the clamping technique represents a limitation, or whether they are physiologically correct cannot be evaluated at the present. However, we believe the conclusion- to be justified through
that measurement of blood velocities the milk veins of lactating goats by means of
the ultrasound Doppler method here and measurement of the milk vein cross-sectional area by means of ultrasound scanning may be an alternative method of mammary blood flow measurements to the invasive methods commonly used. By selecting goats with large milk veins which are capable of carrying all the blood leaving the udder, when the pudendal veins are clamped manually, the precision and accuracy of the method may be further improved. Acknowledgements-The
present studies were supported by a grant from The Danish Agricultural and Veterinary Research Council. For fruitful discussions during the work, the authors wish to thank Dr Knud Rasmussen, M. D. Herlev Hospital and Prof. Dr med. vet. Folke Rasmussen, The Royal Veterinary and Agricultural University, Copenhagen. Mr Hans Busk BSc, The National Institute of Animal Science, Forssgsanlreg Foulum, kindly lent us his scanning equipment and has performed the planimetry of the photographs for which we are very grateful. REFERENCES
Box G. E. P., Hunter W. G. and Hunter J. S. (1978) Statistics for Experimenters, Chapt. 4. John Wiley and Sons, New York. Burvenich C. (1980) Variations of mammary artery blood flow and milk yield under normal conditions and during the oestrous cycle of the dairy goat. Z. Tierphysiol., Tiererniihrg. u. Futlermittelkde 43, 18-27.
Fleet I. R. and Mepham T. B. (1983) Physiological methods used in the study of mammary substrate utilization in ruminants. In Biochemistry of lactation (Edited by Mepham T. B.), pp. 469491. Elsevier, Amsterdam. Gill R. W. (1985) Measurement of blood flow by ultrasound: accuracy and sources of error. Ultrasound Med. Biol. 11, 625641.
Linzeli J. L. (1966) Measurement of venous flow by continuous thermodilution and its application to measurement of mammary blood flow in the goat. Circulation Res. 18, 745-754.
Linzell J. L. (1974) Mammary blood flow and methods of identifying and measuring precursors of milk. In Lactation (Edited by Larson B. L. and Smith V. R.), Vol. 1, pp. 143-225. Academic Press, New York. Linzell J. L. and Rasmussen F. (1972) Mammary blood flow and changes in circulation during lactation in goats and cows. In Hand&h der Tiererniihrung, II (Edited by Lenkeit W. and Breirem K.). Verlag Paul Parey, Hamburg. Mepham T. B., Lawrence S. E., Peters A. R. and Hart I. C. (1984) Effects of exogenous growth hormone on mammary function in lactating goats. Horm. metabol. Res. 16, 248-253.
Rasmussen F. (1963) The mammary blood flow in the goat as measured by antipyrine absorption. Acta vet. &and. 4, 271-280.
Rasmussen K. (1987a) Methodological problems related to measurement of quantitative blood flow with the ultrasound Doppler technique. &and. J. c/in. Lab. Invest. 47, 303-309.
Rasmussen K. (1987b) Precision and accuracy of Doppler flow measurements. In vitro and in viuo study of the applicability of the method in human fetuses. Stand. J. Clin. Lab. Invest 47, 311-318.
Rasmussen K. (1987~) Quantitative blood flow in the fetal descending aorta and in the umbilical vein in normal pregnancies. Longitudinal and cross-sectional studies. Stand. J. clin. Lab. Invesr. 47. 319-324.
Reynolds M. Linzell J. L. and Rasmussen F. (1968) Comparison of four methods for measuring mammary blood flow in conscious goats. Am. J. Physiol. 214, 1415-1424.