Evaluation of Automatic
Mastitis Detection
Equipment A. GEBRE-EGZlABHER 1, H. C. WOOD 2, J. D. ROBAR 2, and G. BLANKENAGEL 1 Department of Dairy and Food Science University of Saskatchewan and SED Systems, Ltd. Saskatoon, Saskatchewan S7N 2R6
one of the most costly diseases encountered in dairy farming. The disease not only decreases An electronic sensor was evaluated as the quantity but also reduces the quality of the an instrument for early detection of milk. These changes may be serious enough to mastitis. This method involved measuring prevent the marketing of the milk, thus resulting the conductivity of milk continuously in significant losses to the dairy farmer. Considerthroughout the milking process and then ing only the decrease in milk production, authoriestablishing a conductivity ratio. The ties estimate that mastitis costs dairymen in the lowest conductivity measurement of the United States approximately 225 to 500 four quarters was a basis for assessing the million dollars per year (13). The magnitude of degree of mastitis in the other quarters. the cost of mastitis and the modern conditions This assumed that at least one of the with large dairy herds and universal machine quarters was normal at examination and milking emphasize the need for a more efficient the lowest reading was normal conducand direct method for detection of mastitis. tivity. The conductivity ratio was evaluMethods for detecting mastitis require ated by comparison with the leukocyte removal of milk and subsequent testing for concentration and combined leukocyte leukocyte concentration or bacterial culture. concentration and cultural examinations These methods require skilled technicians and of milk samples from 1028 quarters. In are too time consuming for routine screening of healthy cows conductivities of milk from milk from individual cows. each of the quarters were similar. If, An electrical conductivity instrument could however, one or more quarters were be incorporated into the milking system and infected, this milk showed higher conducmight be useful as an indicator of mastitis. tivity compared to the noninfected Sodium and chloride ion concentration is quarter of the same cow. The conductivity increased in milk as a result of the destruction ratio correctly identified 69% of the of udder tissue by mastitis (16, 18). A problem established cases of mastitis. For the with this test is that there is a normal variation Wisconsin Mastitis Test, 93.2% of the in the conductivity of milk from cow to cow normal quarters were detected correctly due to nonpathological differences. However, by the conductivity ratio. Leukocyte to minimize these problems between individual counts were frequently high when there animals, Davis (3) suggested the difference in was no other evidence of mastitis. We conductivity between the four quarters of each believe the conductivity ratio is effective cow as an indicator of mastitis. Jones (8) using in detecting mastitis at an early stage of Davis' method did not find agreement with the infection caused by most of the pathocell count. However, he reported that milk genic microorganisms. samples were stored at room temperature for INTRODUCTION 24 to 28 h before analysis. Later, this suggestion was pursued by Greatrix et al. (3), and they Mastitis, an inflammation of the udder, is found a correlation between conductivity and cell count of milk from individual quarters. Received October 9, 1978. Wolf et al. (19) described a method based on 1Department of Dairy and Food Science, Univerthe same principle. A thorough study of the sity of Saskatchewan, Saskatoon, Saskatchewan. conductivity of milk from individual quarters 2SED Systems Ltd., Saskatoon, Saskatchewan. ABSTRACT
1979 J Dairy Sci 62:1108-1114
1108
MASTITIS DETECTION has been by Linzell et al. (10) and Linzell and Peaker (11). The objective of this study was to evaluate the conductivity ratio as a method for detection of mastitis at its early stage of development.
1109
I N D I V I D U A L QUARTER MILK CHAMBER
:
:
"': :
O V E R F L O W TUBE
:
,/"_ : : : ~'.:c: : 5 ~ : : : : " -
~
.~
~
::
.~ ::. . :"
CLEAR PLASTIC J,~ S~RUCTUREStRUCTURE
:~ - : :: :
ELECTRICAL CONNECIlON :
10
RECOVERY
M A T E R I A L S AND METHODS
The experiment was at the University of Saskatchewan dairy herd consisting of 51 Holstein cows ranging in age from 2 to 11 yr and in various stages of lactation. Seventeen percent of these cows had been treated for mastitis at one time or another. The total number of quarters tested was 1,028 since some of them were examined several times. An automatic mastitis detection instrument was installed in the milking parlor. During evening milking, the conductivity of milk from each quarter was measured while the cow was being milked. From the same quarters, foremilk samples were collected for the enumeration of somatic cells and for the identification of bacterial infections. The leukocyte numbers in milk were determined by the Wisconsin Mastitis Test (WMT) according to the method of Thompson and Postle (17) with a millimeter measuring square. This method was chosen since it is a simple, rapid, and sensitive screening test for handling a large number of samples. Tests were immediately after collection without refrigerating the samples. The bacteriological examinations were by the Department of Veterinary Microbiology and consisted of detection and identification of
Streptococcus agalactiae, Streptococcus uberis, Streptococcus dysgalactiae, Staphylococcus aureus, Staphylococcus epidermidis, coliforms, and Corynebacterium spp. (5, 7, 9, 14, 19). INSTRUMENTATION
To obtain data on the electrical conductivity of milk from individual quarters during milking, SED Systems designed and built a plastic milking claw which collected milk from each quarter separately into chambers. From each of these chambers, the milk flowed by gravity through a sensor assembly installed between the entrance chamber and a common collection chamber below. Figure 1 schematically illustrates the construction of the SED milking machine, showing only one of the four entrance
Figure 1. A cross sectional view of the SED Systems milk conductivity instrument for individual quarters.
chambers. The overflow tubes allowed for an even vacuum throughout the machine, and allowed milk flows greater than those possible through the sensor, to pass to the collection chamber. The sensor tube was constructed with a size that assured a continuous flow of milk during and between pulsations until flow rate near the end of milking was low. An electronics module was sealed within the plastic structure of the milking claw with electrical connection to a low voltage power supply and the data collecting devices through a cable. Colinear cylindrical metal electrodes were placed around the sensor tube with a central active electrode and upper and lower grounded electrodes. The inner surface of the electrode assembly was covered with a thin insulating plastic and sealed, preventing any contact between the milk and the metal electrodes. A high frequency alternating voltage was applied to the active electrode, and the current flowing to the electrode was measured with a sensitive circuit. The current flowing to the electrode is proportional to the electrical conductivity of the milk in the sensor that is between the active and the ground electrodes; hence, a signal is produced from the sensor proportional to the conductivity of the milk. The signals from the four quarters were transferred to a digital data collection device every second, and the data were processed later. The quarter with the milk of minimum conductivity was determined, and the ratio of the conductivity of milk from each of the other quarters was calculated with respect to the minimum value. Calibration of the device was by preparing water solutions of sodium chloride of known conductivity and pouring them through the sensor assemblies. All four channels Journal of Dairy Science Vol. 62, No. 7, 1979
1110
GEBRE-EGZ1ABHER ET AL.
were set to equal gain and to equal absolute output value for equal calibrating solutions so any deviations from equal conductivity could be determined from the data. Four of the SED instrumented milking claws were built and installed in the University of Saskatchewan Dairy parlor, replacing the normal milking machines for the duration of the experiment. Separate cables connected each machine to a central automatic power and data collecting unit. The milking procedure was normal with the University personnel operating the milking machines with a minimum of disruption to the standard parlor routine.
COW 87
10
"~ E ~fi
8 6
>_ ~
4
wO
2
0 ANALYSIS
OF DATA
The efficacy of conductivity as an indicator of mastitis was determined by classifying the udder quarters according to their health status. On the basis of the WMT and bacteriological examinations, it was assumed that the milk sample was mastitic if a pathogenic organism was identified in 50% or more of the milk samples (regardless of the leukocyte count) or if the WMT value was consistently greater than 20 mm, and it was considered negative if no pathogens were cultured or if the WMT was consistently less than 10 mrn. In evaluating the conductivity data we decided to use the conductivity ratio, the conductivity of the milk from each quarter divided by the lowest conductivity reading of the four quarters. The essence of this method was that the lowest conductivity of milk from the four quarters was a basis for assessing the degree of infection in other quarters. This assumes that on rare occasions all four quarters might be affected simultaneously. The conductivity ratio is believed to overcome the problems of variations in conductivity due to physiological differences among individual cows, This approach has been taken by Wolf et al. (19) and Linzell and Peaker (11).
----a----~--• - -
L
810
TIME (s
~
/ 160
Left front Left hind Right front Right hind
I
i 240
I
during single milking)
Figure 2. Conductivity measurements during one milking for an uninfected cow.
during a single milking. The conductivity of milk from each of the four quarters was similar although all measures tended to increase during milking. Figure 3 illustrates a similar trend for three of the normal quarters of cow 41, although the values were higher than those of cow 87. One quarter which was infected with Streptococcus uberis, however, showed a higher
C O W 41
5
10
E >_>
6
4
Left front Left hind Right front Right hind
--+--- -
Z 0 u
RESULTS AND DISCUSSION
In healthy cows conductivity of milk from the four quarters was similar. If, however, one or more quarters were infected with mastitis, this milk showed higher conductivity. Examples are in Figures 2 to 6. Figure 2 shows the variation in conductivity of a normal cow 87 Journal of Dairy Science Vol. 62, No. 7, 1979
0
I
l
80 TIME (s
I
I
160
I
I
240
during single milking)
Figure 3. Conductivity measurements during the milking of cow 41 showing an infection in the left front quarter.
MASTITIS DETECTION c o n d u c t i v i t y t h r o u g h o u t the milking. The WMT value f r o m this quarter was > 30 m m whereas WMT of o t h e r quarters was < 10 ram. The variation o f conductivities in the four quarters o f two normal cows t h r o u g h o u t 4 wk is in Figure 4. Cow 80 was in her first and cow 112 was in her second lactation. Results of the WMT and culture tests of these cows showed normal milk in all quarters. Figure 5 d e m o n strates high conductivities for quarters with high cell counts in cow 90. Milk f r o m these quarters also had high l e u k o c y t e counts, but culture tests o f milk samples were negative at all times. Both Staphylococcus aureus and Streptococcus dysgalactiae were isolated f r o m the milk of the right front quarter o f cow 29. The conductivity is in Figure 6. The WMT for the milk f r o m this quarter was consistently > 2 0 m m which indicates a persistent subclinical mastiffs. The variation of absolute c o n d u c t i v i t y in milk f r o m healty cows in day to day sampling also was observed by Little et al. ( 1 2 ) a n d Linzell and Peaker (9). They c o n c l u d e d t h a t it
Left front ~ _ _ Left hind • Rightfront - -
Right hind
1111
I0
> Left front ÷ - - Left hind
Right fronl Right Mnd
;
,;
,'5
CONSECUTIVE DAYS THROUGHOUT
2'0
;s
3o
TRIAL PERIOD
Figure 5. Conductivity data for cow 90 showing infection of the left front and right hind quarters.
was difficult to establish values for normal milk. The f r e q u e n c y histogram of c o n d u c t i v i t y for the three WMT groups is in Figures 7, 8, and 9. In the normal group, the m a x i m u m f r e q u e n c y occurred b e t w e e n 7 and 7.5 m m h o / c m , in the mildly infected groups b e t w e e n 8 and 8.5 m m h o / c m , and in the severely infected category b e t w e e n 9 and 9.5 m m h o / c m . In analyzing the conductivity data for comparison with o t h e r test results, the variation in milk conductivity c o m m o n to all quarters has been eliminated by calculating the ratio of the conductivity for each quarter to the minim u m c o n d u c t i v i t y o f the four quarters. Table 1 shows a comparison of the conductivity ratio with the MWT. A CR of 1.2 was optimal, giving the smallest n u m b e r of false positive and false negative determinations. Samples with WMT values o f > 1 0 m m but conductivity ratios of 20 mm.
loo
50
4
5
6
7
8
9
10
11
12
CONDUCTIVITY (m mho/cm)
Figure 7. Distribution of quarter samples on conductivity results when the WMT was 10 mm and CR >1.2 WMT
Mastitis Detection
Equipment A. GEBRE-EGZlABHER 1, H. C. WOOD 2, J. D. ROBAR 2, and G. BLANKENAGEL 1 Department of Dairy and Food Science University of Saskatchewan and SED Systems, Ltd. Saskatoon, Saskatchewan S7N 2R6
one of the most costly diseases encountered in dairy farming. The disease not only decreases An electronic sensor was evaluated as the quantity but also reduces the quality of the an instrument for early detection of milk. These changes may be serious enough to mastitis. This method involved measuring prevent the marketing of the milk, thus resulting the conductivity of milk continuously in significant losses to the dairy farmer. Considerthroughout the milking process and then ing only the decrease in milk production, authoriestablishing a conductivity ratio. The ties estimate that mastitis costs dairymen in the lowest conductivity measurement of the United States approximately 225 to 500 four quarters was a basis for assessing the million dollars per year (13). The magnitude of degree of mastitis in the other quarters. the cost of mastitis and the modern conditions This assumed that at least one of the with large dairy herds and universal machine quarters was normal at examination and milking emphasize the need for a more efficient the lowest reading was normal conducand direct method for detection of mastitis. tivity. The conductivity ratio was evaluMethods for detecting mastitis require ated by comparison with the leukocyte removal of milk and subsequent testing for concentration and combined leukocyte leukocyte concentration or bacterial culture. concentration and cultural examinations These methods require skilled technicians and of milk samples from 1028 quarters. In are too time consuming for routine screening of healthy cows conductivities of milk from milk from individual cows. each of the quarters were similar. If, An electrical conductivity instrument could however, one or more quarters were be incorporated into the milking system and infected, this milk showed higher conducmight be useful as an indicator of mastitis. tivity compared to the noninfected Sodium and chloride ion concentration is quarter of the same cow. The conductivity increased in milk as a result of the destruction ratio correctly identified 69% of the of udder tissue by mastitis (16, 18). A problem established cases of mastitis. For the with this test is that there is a normal variation Wisconsin Mastitis Test, 93.2% of the in the conductivity of milk from cow to cow normal quarters were detected correctly due to nonpathological differences. However, by the conductivity ratio. Leukocyte to minimize these problems between individual counts were frequently high when there animals, Davis (3) suggested the difference in was no other evidence of mastitis. We conductivity between the four quarters of each believe the conductivity ratio is effective cow as an indicator of mastitis. Jones (8) using in detecting mastitis at an early stage of Davis' method did not find agreement with the infection caused by most of the pathocell count. However, he reported that milk genic microorganisms. samples were stored at room temperature for INTRODUCTION 24 to 28 h before analysis. Later, this suggestion was pursued by Greatrix et al. (3), and they Mastitis, an inflammation of the udder, is found a correlation between conductivity and cell count of milk from individual quarters. Received October 9, 1978. Wolf et al. (19) described a method based on 1Department of Dairy and Food Science, Univerthe same principle. A thorough study of the sity of Saskatchewan, Saskatoon, Saskatchewan. conductivity of milk from individual quarters 2SED Systems Ltd., Saskatoon, Saskatchewan. ABSTRACT
1979 J Dairy Sci 62:1108-1114
1108
MASTITIS DETECTION has been by Linzell et al. (10) and Linzell and Peaker (11). The objective of this study was to evaluate the conductivity ratio as a method for detection of mastitis at its early stage of development.
1109
I N D I V I D U A L QUARTER MILK CHAMBER
:
:
"': :
O V E R F L O W TUBE
:
,/"_ : : : ~'.:c: : 5 ~ : : : : " -
~
.~
~
::
.~ ::. . :"
CLEAR PLASTIC J,~ S~RUCTUREStRUCTURE
:~ - : :: :
ELECTRICAL CONNECIlON :
10
RECOVERY
M A T E R I A L S AND METHODS
The experiment was at the University of Saskatchewan dairy herd consisting of 51 Holstein cows ranging in age from 2 to 11 yr and in various stages of lactation. Seventeen percent of these cows had been treated for mastitis at one time or another. The total number of quarters tested was 1,028 since some of them were examined several times. An automatic mastitis detection instrument was installed in the milking parlor. During evening milking, the conductivity of milk from each quarter was measured while the cow was being milked. From the same quarters, foremilk samples were collected for the enumeration of somatic cells and for the identification of bacterial infections. The leukocyte numbers in milk were determined by the Wisconsin Mastitis Test (WMT) according to the method of Thompson and Postle (17) with a millimeter measuring square. This method was chosen since it is a simple, rapid, and sensitive screening test for handling a large number of samples. Tests were immediately after collection without refrigerating the samples. The bacteriological examinations were by the Department of Veterinary Microbiology and consisted of detection and identification of
Streptococcus agalactiae, Streptococcus uberis, Streptococcus dysgalactiae, Staphylococcus aureus, Staphylococcus epidermidis, coliforms, and Corynebacterium spp. (5, 7, 9, 14, 19). INSTRUMENTATION
To obtain data on the electrical conductivity of milk from individual quarters during milking, SED Systems designed and built a plastic milking claw which collected milk from each quarter separately into chambers. From each of these chambers, the milk flowed by gravity through a sensor assembly installed between the entrance chamber and a common collection chamber below. Figure 1 schematically illustrates the construction of the SED milking machine, showing only one of the four entrance
Figure 1. A cross sectional view of the SED Systems milk conductivity instrument for individual quarters.
chambers. The overflow tubes allowed for an even vacuum throughout the machine, and allowed milk flows greater than those possible through the sensor, to pass to the collection chamber. The sensor tube was constructed with a size that assured a continuous flow of milk during and between pulsations until flow rate near the end of milking was low. An electronics module was sealed within the plastic structure of the milking claw with electrical connection to a low voltage power supply and the data collecting devices through a cable. Colinear cylindrical metal electrodes were placed around the sensor tube with a central active electrode and upper and lower grounded electrodes. The inner surface of the electrode assembly was covered with a thin insulating plastic and sealed, preventing any contact between the milk and the metal electrodes. A high frequency alternating voltage was applied to the active electrode, and the current flowing to the electrode was measured with a sensitive circuit. The current flowing to the electrode is proportional to the electrical conductivity of the milk in the sensor that is between the active and the ground electrodes; hence, a signal is produced from the sensor proportional to the conductivity of the milk. The signals from the four quarters were transferred to a digital data collection device every second, and the data were processed later. The quarter with the milk of minimum conductivity was determined, and the ratio of the conductivity of milk from each of the other quarters was calculated with respect to the minimum value. Calibration of the device was by preparing water solutions of sodium chloride of known conductivity and pouring them through the sensor assemblies. All four channels Journal of Dairy Science Vol. 62, No. 7, 1979
1110
GEBRE-EGZ1ABHER ET AL.
were set to equal gain and to equal absolute output value for equal calibrating solutions so any deviations from equal conductivity could be determined from the data. Four of the SED instrumented milking claws were built and installed in the University of Saskatchewan Dairy parlor, replacing the normal milking machines for the duration of the experiment. Separate cables connected each machine to a central automatic power and data collecting unit. The milking procedure was normal with the University personnel operating the milking machines with a minimum of disruption to the standard parlor routine.
COW 87
10
"~ E ~fi
8 6
>_ ~
4
wO
2
0 ANALYSIS
OF DATA
The efficacy of conductivity as an indicator of mastitis was determined by classifying the udder quarters according to their health status. On the basis of the WMT and bacteriological examinations, it was assumed that the milk sample was mastitic if a pathogenic organism was identified in 50% or more of the milk samples (regardless of the leukocyte count) or if the WMT value was consistently greater than 20 mm, and it was considered negative if no pathogens were cultured or if the WMT was consistently less than 10 mrn. In evaluating the conductivity data we decided to use the conductivity ratio, the conductivity of the milk from each quarter divided by the lowest conductivity reading of the four quarters. The essence of this method was that the lowest conductivity of milk from the four quarters was a basis for assessing the degree of infection in other quarters. This assumes that on rare occasions all four quarters might be affected simultaneously. The conductivity ratio is believed to overcome the problems of variations in conductivity due to physiological differences among individual cows, This approach has been taken by Wolf et al. (19) and Linzell and Peaker (11).
----a----~--• - -
L
810
TIME (s
~
/ 160
Left front Left hind Right front Right hind
I
i 240
I
during single milking)
Figure 2. Conductivity measurements during one milking for an uninfected cow.
during a single milking. The conductivity of milk from each of the four quarters was similar although all measures tended to increase during milking. Figure 3 illustrates a similar trend for three of the normal quarters of cow 41, although the values were higher than those of cow 87. One quarter which was infected with Streptococcus uberis, however, showed a higher
C O W 41
5
10
E >_>
6
4
Left front Left hind Right front Right hind
--+--- -
Z 0 u
RESULTS AND DISCUSSION
In healthy cows conductivity of milk from the four quarters was similar. If, however, one or more quarters were infected with mastitis, this milk showed higher conductivity. Examples are in Figures 2 to 6. Figure 2 shows the variation in conductivity of a normal cow 87 Journal of Dairy Science Vol. 62, No. 7, 1979
0
I
l
80 TIME (s
I
I
160
I
I
240
during single milking)
Figure 3. Conductivity measurements during the milking of cow 41 showing an infection in the left front quarter.
MASTITIS DETECTION c o n d u c t i v i t y t h r o u g h o u t the milking. The WMT value f r o m this quarter was > 30 m m whereas WMT of o t h e r quarters was < 10 ram. The variation o f conductivities in the four quarters o f two normal cows t h r o u g h o u t 4 wk is in Figure 4. Cow 80 was in her first and cow 112 was in her second lactation. Results of the WMT and culture tests of these cows showed normal milk in all quarters. Figure 5 d e m o n strates high conductivities for quarters with high cell counts in cow 90. Milk f r o m these quarters also had high l e u k o c y t e counts, but culture tests o f milk samples were negative at all times. Both Staphylococcus aureus and Streptococcus dysgalactiae were isolated f r o m the milk of the right front quarter o f cow 29. The conductivity is in Figure 6. The WMT for the milk f r o m this quarter was consistently > 2 0 m m which indicates a persistent subclinical mastiffs. The variation of absolute c o n d u c t i v i t y in milk f r o m healty cows in day to day sampling also was observed by Little et al. ( 1 2 ) a n d Linzell and Peaker (9). They c o n c l u d e d t h a t it
Left front ~ _ _ Left hind • Rightfront - -
Right hind
1111
I0
> Left front ÷ - - Left hind
Right fronl Right Mnd
;
,;
,'5
CONSECUTIVE DAYS THROUGHOUT
2'0
;s
3o
TRIAL PERIOD
Figure 5. Conductivity data for cow 90 showing infection of the left front and right hind quarters.
was difficult to establish values for normal milk. The f r e q u e n c y histogram of c o n d u c t i v i t y for the three WMT groups is in Figures 7, 8, and 9. In the normal group, the m a x i m u m f r e q u e n c y occurred b e t w e e n 7 and 7.5 m m h o / c m , in the mildly infected groups b e t w e e n 8 and 8.5 m m h o / c m , and in the severely infected category b e t w e e n 9 and 9.5 m m h o / c m . In analyzing the conductivity data for comparison with o t h e r test results, the variation in milk conductivity c o m m o n to all quarters has been eliminated by calculating the ratio of the conductivity for each quarter to the minim u m c o n d u c t i v i t y o f the four quarters. Table 1 shows a comparison of the conductivity ratio with the MWT. A CR of 1.2 was optimal, giving the smallest n u m b e r of false positive and false negative determinations. Samples with WMT values o f > 1 0 m m but conductivity ratios of 20 mm.
loo
50
4
5
6
7
8
9
10
11
12
CONDUCTIVITY (m mho/cm)
Figure 7. Distribution of quarter samples on conductivity results when the WMT was 10 mm and CR >1.2 WMT
