Financial analysis of alternative treatments for clinical mastitis associated with environmental pathogens.

Abstract

For two large California dairy herds with twice daily milking, 171 infected quarters of lactating cows with mild clinical mastitis were randomly assigned to one of three treatment groups. Group A (50 cows) was treated with 62.5 mg of intramammary amoxicillin every 12 h for three milkings. Group C (50 cows) was treated with 200 mg of intramammary cephapirin every 12 h for two milkings. Group O (71 cows) was treated with 100 U of intramuscular oxytocin every 12 h for three milkings. Clinical cure rates did not differ among treatment groups. The cost per episode of clinical mastitis was higher ($54.47) for group C than for groups A ($38.53) or O ($34.88). Group O had a higher incidence of relapse (41%), and a greater percentage of group O cows (65%) experienced an additional mastitic event in the enrolled lactation than did cows in groups A and C. Treatment had no effect on the mean number of nonsalable milkings (24.7) associated with mastitis in the lactation interval subsequent to enrollment on the trial. Twenty-two percent of the cows accrued more than 30 nonsalable milkings and produced 55% of the total nonsalable milkings associated with mastitis. There was no treatment effect on total milk production, fat production, or time to removal of the enrolled cows from the herd.

Full text

Page 1

Financial Analysis of Alternative Treatments for Clinical
Mastitis Associated with Environmental Pathogens
A. L. VAN EENENNAAM,lf* 1. A. GARDNER.3 J. HOLMES,J L. PERANI,4
R. J. ANDERSONP J. S. CULLOR.4 and W. U. GUTERBOCKS
University of California
Davis 95616
ABSTRACT
For two large California dairy herds
with twice daily milking, 171 infected
quarters of lactating cows with mild clin-
ical mastitis were randomly assigned to
one of three treatment groups. Group A
(50 cows) was treated with 62.5 mg of
intramammary amoxicillin every 12 h for
three milkings. Group C (50 cows) was
treated with 200 mg of intramammary
cephapirin every 12 h for two milkings.
Group 0 (71 cows) was treated with 100
U of intramuscular oxytocin every 12 h
for three milkings. Clinical cure rates did
not differ among treatment groups. The
cost per episode of clinical mastitis was
higher ($54.47) for group C than for
groups A ($38.53) or 0 ($34.88). Group
0 had a higher incidence of relapse
(41%), and a greater percentage of group
0 cows (65%) experienced an additional
mastitic event in the enrolled lactation
than did cows in groups A and C. Treat-
ment had no effect on the mean number
of nonsalable milkings (24.7) associated
with mastitis in the lactation interval
subsequent to enrollment on the trial.
Twenty-two percent of the cows accrued
more than 30 nonsalable milkings and
produced 55% of the total nonsalable
milkings associated with mastitis. There
was no treatment effect on total milk
Received December 2, 1994.
Accepted March 28, 1995.
'Depamnent of Animal Science.
2Reprint requests.
3Department of Medicine and Epidemiology, School of
Veterinary Medicine.
4Dep~ment of Pathology, Microbiology, and I m u -
nology; School of Veterinary Medicine.
5Veterinary Medicine Teaching and Research Center,
18830 Road 112, Tulare, CA 93274.
production, fat production, or time to
removal of the enrolled cows from the
herd.
(Key words: antibiotics, mastitis treat-
ment, environmental pathogen, cost)
INTRODUCTION
Treatment of clinical mastitis is the major
source of antibiotic contamination of dairy
products in the United States (11). Because of
concern over drug residues in the milk supply,
the routine use of intramammary antibiotic
tubes for the treatment of clinical mastitis
should be examined. The decision to treat a
mastitic cow must be cost-effective. The cost
of discarded milk and medication must be less
than the cost of leaving the quarter untreated.
Intramammary tubes must therefore assist the
host defense system to eliminate the pathogen,
decrease the deleterious effects of the infection
on the gland, or both (5). The lack of efficacy
data on intramammary antibiotic preparations
against environmental pathogens, the short du-
ration of such infections, and the potential
economic loss from contamination of milk and
meat resulting from antibiotic residues, are
valid reasons to evaluate the economic impli-
cations of using intramammary antibiotics to
treat mild clinical mastitis caused by environ-
mental pathogens.
This paper details a financial analysis of
alternative treatments for mild cases of clinical
mastitis associated with environmental patho-
gens. The analysis is based, in part, on previ-
ously published data establishing the efficacy
of two commercially available intramammary
antibiotic preparations compared with controls
given intramuscular oxytocin for the treatment
of mild environmental clinical mastitis (6).
That study found no difference in the bacterial
(absence of primary pathogen isolated pretreat-
ment) or clinical (return of quarter and milk to
normal) cure rates between treatments. The
1995 J D a i r y Sci 78:2086-2095
2086

Page 2

OUR INDUSTRY TODAY
2087
current paper examines the financial outcomes
and the effects on milk production of alterna-
tive therapies for treatment of mild clinical
mastitis caused by environmental pathogens.
MATERIALS AND METHODS
Study Population and Trial Design
Methods were detailed previously (6). The
financial analysis was conducted on data col-
lected from two commercial dairy herds in the
San Joaquin Valley of California. These large
herds (700 and 1200 milking cows) were on
DHIA test, had computerized herd records
(Dairycomp 305@; Valley Agricultural Soft-
ware, Tulare, CA), and were milked twice
daily in herringbone parlors. Both herds were
free of the contagious mastitis pathogens,
Staphylococcus aureus, Streptococcus agalac-
tiae, and Mycoplasma species. Somatic cell
counts of bulk tank milk taken prior to the
beginning of the trial were <200,000/ml for
both herds. Eighty-five cows in one herd and
86 cows in the other were diagnosed with mild
clinical mastitis and randomly assigned to one
of three treatment groups. Mild clinical masti-
tis was defined as the occurrence of visually
abnormal milk (flakes or clots in normal or
watery foremilk stripped on the floor) with or
without swelling of the affected quarter. Group
A (25 cows in each herd) received one in-
tramammary tube containing 62.5 mg of amox-
icillin as amoxicillin trihydrate (Amoxi-Mast?
Smith Kline Beecham, Exton, PA) in the af-
fected quarter after milking for the first three
milkings following diagnosis. Group C (25
cows in each herd) received one intramammary
tube containing 200 mg of cephapirin as
cephapirin sodium (Cefa-Lak? Aveco, Fort
Dodge, IA) in the affected quarter after milking
for the first two milkings following diagnosis.
Group 0 (35 cows in one herd and 36 cows in
the other) received an intramuscular injection
of 100 I U of oxytocin (Anthony Products, Ar-
cadia, CA) prior to each of the first three
milkings following diagnosis.
To facilitate adherence to the trial protocol,
only one antibiotic therapy at a time was in use
on the dairy. All mastitic cows were housed in
the hospital string for the milk withdrawal
period specified for the antibiotic being ad-
ministered. Group 0 cows were kept in the
hospital string for the same number of milk-
ings as the concurrent antibiotic therapy
groups to facilitate sample collection. If an
infected quarter did not make satisfactory pro-
gress or worsened, the cow was treated by
dairy personnel according to their usual proce-
dures, and these nonprotocol treatments were
recorded.
Classification of Outcome
Single aseptic foremilk samples were col-
lected from the affected quarter prior to milk-
ing. Samples were taken before the first treat-
ment (milking 1) and at milkings 4 and 9
(group A and contemporaneous group 0 quar-
ters) or at milkings 1, 4, and 11 (group C and
contemporaneous group 0 quarters). A final
sample was taken on d 21 after enrollment.
Samples were evaluated for mastitis-causing
bacteria (6). A clinical cure was defined as the
return of the secretion and quarter to normal
appearance, as assessed by visual observation
of foremilk stripped on the floor and by palpa-
tion at or before milking 9. A clinical failure
was defined as failure of the quarter and secre-
tion to return to normal by milking 9. If a
quarter did not appear to be making satisfac-
tory progress toward recovery and received
nonprotocol treatments during the observation
period, the case was defined as a clinical fail-
ure.
Cows were monitored for the duration of
the enrolled lactation, and all Occurrences of
mastitis and milkings in the hospital string
were recorded. Additional mastitic events were
categorized as one of three types. A relapse
was defined as a retreatment or relapse of the
enrolled quarter following an apparent clinical
cure in the 21 d following the initial enroll-
ment of that quarter. A recurrence was defined
as a clinical mastitis event that occurred in the
enrolled quarter >21 d after the date of initial
enrollment, but prior to the completion of the
enrolled lactation. Infection in an unenrolled
quarter was defined as a clinical mastitis event
in the same lactation as the original clinical
mastitis event in a quarter other than the one
enrolled on the trial. The DHIA production
records were available for all of the cows
enrolled on the trial and for all of their con-
temporary herdmates completing a lactation
between January 1991 and August 1993.
Journal of Dairy Science Vol. 78, No. 9, 1995

Page 3

208 8
VAN EENENNAAM ET AL.
Data Analysis
Clinical cure, relapse, and recurrence rates;
mastitis events in unenrolled quarters; and the
percentage of cases that were withdrawn from
the trial were tabulated by treatment. Chi-
squared analysis was used to determine
whether any difference was significant (P <
.05) among these percentages for the different
treatment groups.
Financial Analysis
The cost of treatment was calculated by
adding the cost of the therapy to the value of
the milk withheld. Cost per treatment was
calculated by multiplying the retail cost of the
product by the number of times the treatment
was administered. Costs for the amoxicillin
(three intramammary tubes), cephapirin (two
intramammary tubes), and oxytocin (one sy-
ringe, three needles, and 300 IU of oxytocin)
treatments were $3.23. $2.37, and $1.00,
respectively. All treatments in this study were
designed to be used within the context of a
modern large-scale dairy operation. No addi-
tional expense was allowed for veterinary serv-
ices or treatment labor because all treatments
were administered in the hospital string by the
milker or herdsperson as a part of the usual
daily routine.
The milk that had been withheld from clini-
cally cured cases was calculated for each medi-
cation used. The number of milkings that were
withheld from the market for amoxicillin and
cephapirin treatments was 8 and 10, respec-
tively. Half of the milk production recorded by
DHIA on the test day prior to enrollment on
the trial was multiplied by the number of
milkings discarded, thus estimating the amount
of milk withheld from the bulk tank (8). For
group 0, no official withholding was required
for the milk, and so the milk could have been
returned to the bulk tank as soon as it appeared
to be visually normal. This number of milkings
was used in the cost estimate. The actual num-
ber of milkings that was discarded was more
than estimated because the cows that received
oxytocin were required by the experimental
protocol to remain in the hospital string for the
same number of milkings as the concurrent
antibiotic therapy group.
The milk loss that occurred from a cow that
was not clinically cured by the end of the
withdrawal period was estimated by multiply-
ing the total number of milkings in the hospital
string until the milk appeared to be visually
normal by one-half of the daily milk produc-
tion. The cost of milk that had been discarded
as a result of a relapse, recurrence, or infection
in another quarter was not included in this
analysis of the cost of therapy per episode of
clinical mastitis. All calculations were based
on a milk price of $11.00/45 kg. Only cows
that had a DHIA test day prior to the day of
enrollment were used in this part of the analy-
sis because we had no way to estimate the
daily milk production of cows prior to their
first test day. Treatment costs were compared
by one-way ANOVA with pairwise compari-
sons by Scheffd’s method when the ANOVA
was significant (P < .05).
Nonsalable Milkings for the Lactation
The total number of milkings that each cow
produced in the hospital string as a result of
mastitic events was tabulated by treatment
group for the lactation interval subsequent to
enrollment on the trial. The percentage of
cases in which all nonsalable milkings for the
entire lactation resulted only from the one
mastitis episode that triggered enrollment on
the trial was calculated for each treatment
group. The nonsalable milkings associated
with all of the mastitic events in the enrolled
lactations were tabulated by lactation number,
and a chi-squared analysis was used to deter-
mine whether there was a lactation number
effect (P < .OS) on the total number of nonsala-
ble milkings per lactation in which a mastitic
event occurred.
Milk Production
Mature equivalent milk production for a
305-d lactation was compared among the three
treatment groups by multiple linear regression.
For this analysis, 2 Jersey cows and 18 cows
that had milked for >305 d before the mastitic
event were excluded. For independent varia-
bles with k categories, k - 1 dummy variables
were created. The model included fixed effects
for group (A, C, or 0), herd (1 or 2), and days
in lactation at the time of enrollment on the
trial (1 to 60 d, 61 to 120 d, or 121 to 304 d).
Goodness of fit of the model was assessed by
R2 values. Assumptions of normality and
Journal of Dairy Science Vol. 78, No. 9, 1995

Page 4

OUR INDUSTRY TODAY
2089
homoscedasticity were checked by plotting
residuals.
Production of 305-d milk, fat, and FCM
was compared among the three treatment
groups by multiple linear regression using a
model that included fixed effects for group (A,
C, or 0), herd (1 or 2). and lactation (1.2, or 2
3). The 2 Jersey cows and the 18 mastitic cows
that had been milked for >305 d prior to the
mastitic event were again excluded from this
analysis. Lactations of <305 d and 2155 d
were adjusted to 305 d using the National
DHIA method for the extension of lactation
records (12). Cows that had lactations of 4 5 5
d were excluded because of the inaccuracies
involved in extending their lactation record to
305 d.
Mastitic cows were matched with control
herdmates that calved at approximately the
same time (f20 d) and were in the same lacta-
tion grouping (1, 2, or 23). Herdmates with a
DHIA CAR (condition affecting record) code,
meaning that some irregular condition affected
their production record, were excluded as
potential control cows. The mean milk and fat
production of all controls that matched the
criteria for the case cows was used in the
analysis. An ANOVA model that included
herd, pathogen, and treatment was used to
analyze the difference between case production
and the mean production of the concurrent
herdmates for 305-d milk, fat, and FCM.
T h e to Removal from the Herd
Survival analysis was used to determine
whether time to removal (culling or death)
from the herd differed among the three treat-
ments. Records for cows that completed their
lactation (137 cows) or were still milking (5
cows) at the termination of the follow-up
period were considered censored observations
for this analysis. Estimates obtained by the
Kaplan-Meier product-limit method were
stratified by herd and pretreatment bacterial
isolate at enrollment to determine whether
these variables might have influenced treat-
ment response. Breslow and Mantel-Cox non-
TABLE 1. Clinical cure by treatment group and bacterium isolated at pretreatment sampling in a randomized field trial of
treatments for mild clinical mastitis in California during 1991 to 1992.
Pretreatment
bacterial isolate’
Treatment
Oxytocin Amoxicillin Cephapirin
P
(no./no.)z
(%)
(no./no.)
(%)
(noho.)
(%)
Protocol treatment’
Coliform
Streptococcus spp.
Other
No bacteria isolated
Subtotal
Nonprotocol treatment4
14/16
16/20
316
15/16
48/58
as
80
50
94
83
7/10
11/14
516
718
30138
70
79
83
88
79
11/12
9/12
6f7
1011 1
36/42
92
75
86
91
86
.34
.95
.28
.87
.73
Coliform
Streptococcus spp.
Other
No bacteria isolated
Subtotal
T o t a l enrolled
on
w4
w1
0/1
0/13
48t71
016
0/4
010
0/2
0/12
30150
014
013
010
01 I
018
36/50 68 60 72 .43
IOf the 55 coliforms, 44 (80%) were Escherichia coli. Of the 57 Streptococcus species, 27 (47%) were Streprococcus
uberis, and 14 (25%) were Streptococcus dysgalactiae. Of the 20 “other” bacteria, 6 (30%) were mixed infections, 11
(55%) were Staphylococcus species, 1 (5%) was Bacillus species, 1 (5%) was Corynebacterium species, and 1 was a
contaminated sample.
*Number of cases clinically cured by milking 9 out of the number of cases treated.
3Cows in this group received mastitis treatment in accordance with the experimental protocol.
Tows removed from the trial because of the administration of nonprotocol treatments were classified as clinical
failures (see text).
Journal of D a i r y Science Vol. 78, No. 9, 1995

Page 5

2090
VAN EENENNAAM ET AL.
parametric linear rank statistics were used to
test for homogeneity of survival curves across
strata. The Breslow statistic gives greater
weight to earlier observations, and the Mantel-
Cox statistic gives equal weight to early and
late observations.
RESULTS
Table 1 shows the rates for clinical cures
and withdrawal, classified by treatment group
and pretreatment bacterial isolate. Thirty-three
of the 171 cows were removed from subse-
quent analyses because the commercial herd
managers chose to administer nonprotocol
treatments prior to the 21-d sample. Clinical
cure rates and the proportion of cows with-
drawn from the trial did not differ by treatment
(P > .05). Relapse, recurrence, and I M I rates in
unenrolled quarters for the enrolled lactation
categorized by the three treatment groups are
summarized in Table 2. Relapse rate differed
among the three groups (P = .02). Relapse rate
was higher for cows in group 0 than for those
in groups C (P = .01) or A (P = .13). Many of
the relapses and recurrences for cows in group
0 occurred when the mastitic event was as-
sociated with the pretreatment isolation of an
environmental Streptococcus sp. from the a€-
fected quarter.
The financial analysis of the mean cost of
each therapy per episode of clinical mastitis is
shown in Table 3. The cost per episode ranged
from $5.84 to $194.96 @= $41.31). The mean
cost differed significantly (P = .006) among the
three treatment groups; mean cost for group C
was significantly higher than that for groups 0
(P < .Ol) or A (P c .01). The legal milk
withdrawal time for cephapirin (96 h) is longer
than for amoxicillin (60 h), necessitating the
TABLE 2. Incidence of relapse.' recurrence? and infection in an unenrolled quarter,3 by treatment group and
pretreatment bacterial isolate. in the lactation interval subsequent to diagnosis of clinical mastitis and the enrollment of
the affected quarter i n a randomized field trial of treatments for mild clinical mastitis in California during 1991 to 1992.
Both relapse
and
recurrence
Mastitis in
unenrolled
quarte+
Pretreatment
bacterial isolate Relapse4 Recurrences
( n o . )
18
4
6
4
4
6
(%)
3Ia
( n o . )
7
1
6
0
0
7
(46)
12c
(no.)
6
2
3
0
1
4
(%I
1Od
(no.)
16
(%)
Oxytocin (n = 58)
Coliform
Streptococcus spp.
Other
No bacteria isolated
Arnoxicillin (n = 38)
Coliform
Streptococcus spp.
Other
No bacteria isolated
Cephapirin (n = 42)
2
1
1
2
5
2
0
1
4
8
1
2
0
1
2
Coliform
Streptococcus spp.
1
1
0
2
0
1
Other
No bacteria isolated
3
0
3
3
1
0
16ab
18C
lld
7
18ef
12b
19E
5d 4
1 Of
gbC.d.efhfeans within columns with no common superscripts differ (P c .05).
'A relapse was defmed as a retreatment or relapse of the enrolled quarter following an apparent clinical cure in the 21
d following the initial enrollment of that quarter on the trial.
2A recurrence was defined as a clinical mastitis event that occurred in the enrolled quaaer >21 d after the date of
initial enrollment, but prior to the completion of the enrolled lactation.
31nfection in an unenrolled quarter was defined as a clinical mastitis event in the same lactation as the original clinical
m a s t i t i s event in a quarter other than the one enrolled on the trial.
"is category does not include quarters that had a recurrence.
This category does not include quarters that had a relapse.
6This category includes some quarters that had a relapse or recwence or both.
Journal of D a i r y Science Vol. 78, No. 9, 1995

Page 6

OUR INDUSTRY TODAY
209 1
TABLE 3. Cost (therapy and nonsalable milk) per episode of clinical mastitis, number of milkings until the affected
quarter secretion had a n o d appearance, and the DHlA daily milk production at the test prior to the diagnosis of
mastitis in a randomized field trial of treatments for mild clinical mastitis in California during 1991 to 1992.
Treatment EDisode of clinical mastitis
-
X
SD
34.88* 34.78
n o d secretion
-
X
Min.
8b 3
at last test dav
-
X
37.w
Cost per episode Milkings to Daily milk production
Min.
5.84
Max.
194.96
Max.
41
Min.
9.9
Max
57.4
Amoxicillin
Cephapirin
Mean
36
33
38.531 14.27
54.47b 24.79
41.3 1
15.99
26.74
89.75
153.92
7b
8b
3
3
15
34
34.2C
37.7C
13.0
19.9
64.3
58.1
Oxytocin
52
*b*Means within columns with no common superscripts differ (P < .05).
discard of additional milk. Most quarters ap-
peared to be clinically cured by milking 9;
therefore, the discard of milking 9 and 10 is a
real cost associated with using an intramam-
mary treatment product with a longer (96 h)
withdrawal period. The oxytocin treatment
costs were as low as $5.84 per episode, but, on
average, were not significantly lower t h a n for
amoxicillin because of the increased number of
milkings required for some of the affected
quarters of group 0 cows to produce milk with
a normal appearance.
Table 4 summarizes the mean number of
nonsalable ("hospital")
with mastitis for enrolled cows during the lac-
tation interval that remained following enroll-
ment on the trial. The value for the cows in
group 0 includes the milkings during which
cows were producing grossly normal milk fol-
lowing their recovery from mastitis, but those
cows remained in the hospital string to allow
sample collection along with their contem-
porary antibiotic treatment group. Although the
milk was salable at this time, producers likely
milkings associated
TABLE 4. Mean number of mastitis-associated nonsalable milkings for enrolled cows in each of the three treatment
groups and a comparison of the mean number of nonsalable milkings related to relapses,' recurrences,2 and unenrolled
quarter infections3 in the lactation interval subsequent to enrollment in a randomized field trial of treatments for mild
clinical mastitis in California during 1991 to 1992.
Lactations with
no relapse,
recurrence, or
Treatment
X
Lactations
with norelapse
with no relapse
or recurrence
unenrolled
quarter infection
Lactations
Treatment
(no. milkings)
16.1c
10.36
12.4cd
(nolno.)4
19/54
20135
22/34
(%)s
35e
57'
65f
Oxytocin
Amoxicillin
Cephapirin
24.la
18.71
27.1a
20.4b
13.8b
17.9
9
8
10
~b~.d.e&leans within columns with no common superscripts differ (P < .05).
*A relapse was defined as a retreatment or relapse of the enrolled quarter following an apparent clinical cure in the 21
d following the initial enrollment of that quarter on the trial.
*A recurrence was defined as a clinical mastitis event that occurred in the enrolled quarter >21 d after the date of
initial enrollment, but prior to the completion of the enrolled lactation.
31nfection in a unenrolled quarter was defined as a clinical mastitis event in the same lactation as the original clinical
mastitis event in a quarter other than the one enrolled on the trial.
4Number of cases with no additional mastitis events following enrollment on the trial out of number of cases with full
mastitis event lactation data in each treatment group.
sPercentage of cases with no additional mastitis events following enrollment on the trial out of number of cases with
complete mastitis event lactation data in each treatment group.
Journal of Dairy Science Vol. 78. No. 9. 1995

Page 7

2092
VAN EENENNAAM ET AL.
would keep the cows in the hospital string for
a short period following the secretion of milk
with normal appearance to ensure complete
recovery. There was no significant difference
(P > .OS) in the mean number of nonsalable
milkings produced by each treatment group.
The treatment mean for group C was elevated
by the prolonged period needed for the
relapsed quarters in this group to achieve milk
with normal appearance. A significantly higher
percentage (65%) of group 0 lactations had an
additional mastitic event subsequent to enroll-
ment.
Table 5 shows that the number of nonsala-
ble milkings associated with the initial clinical
mastitis episode and withdrawal period (8 or
10 milkings, depending upon the product label
withdrawal time) constituted only 17% of the
total number of nonsalable milkings produced
by the mastitic cows during the enrolled lacta-
tions. The mean number of milkings that was
discarded per enrolled mastitic lactation across
all therapy groups was 24.7. If mean produc-
tion per milking is 18 kg, then the mean cost
of discarded m i l k per mastitic lactation would
be $108.68, which is considerably higher than
the mean cost per episode of clinical mastitis
($41.31) calculated in Table 3. Table 5 also
shows that older cows tended to spend more
days in the hospital string than did heifers (P =
.06) following the first mastitic event.
Milk Production
Linear regression showed no significant ef-
fect of treatment on 305-d mature equivalent
milk production of the enrolled cows (Table 6).
There was a herd effect and an effect of days
in lactation at the time of enrollment on the
trial. Cows that had a mastitic event late in
their lactation (after 120 d) had significantly
higher mature equivalent milk production over
305 d (P = .Ol) than did cows that were
enrolled prior to the completion of the first 120
d of their lactation. There was a significant
herd and parity effect, but no treatment effect
on 305-d production of milk, fat, and FCM for
the mastitic cows. Comparison of the DHIA
lactation records of the mastitic cows and the
mean of their contemporary herdmates,
matched on the basis of lactation number and
calving date, showed no significant difference
between the 305-d production of milk or fat
between the mastitic and control cows. The
305d production of milk and fat by the
mastitic cows was 9554 and 359 kg, respec-
tively, and the means for production by their
contemporary herdmates was 9539 and 353 kg,
respectively. There was no significant effect of
treatment, pathogen, or herd on the difference
between the 305-d production of milk and fat
of the mastitic cows and their contemporary
herdmates (data not shown).
Time to Removal from the Herd
Median time to removal from the herd after
initiation of treatment did not differ signifi-
cantly among treatments with either the Bres-
low (P = 57) or Mantel-Cox (P = .72) statistic.
Stratification of the survival curves by herd
and bacterial type at enrollment revealed no
significant effect of these two variables on the
treatment response. The Mantel-Cox P values,
adjusted for strata, were .72 and .84, respec-
tively. For all three treatments combined, me-
dian time to removal after treatment was lower
(P = .05) for one herd than for the other, but
did not differ (P = .63) by bacterial type at
initial sampling.
TABLE 5. Number of all nonsalable milkings associated with mastitis by parity in the lactation interval subsequent to the
diagnosis of mild clinical m a s t i t i s i n two commercial d a q herds in California during 1991 to 1992.
Lactation
Range of nonsalable
milkings per lactation
Total nonsalable
milkings
1
2
23
All
1-1 1
12-30
31-60
%O
Total
Mean per lactation
17
8
I
1
27
15.9
18
15
I I
3
47
25.7
31
31
8
10
80
27.2
66
54
20
14
154
24.7
43
35
13
9
100
652
1061
866
1232
3811
17
28
23
32
100
(no.) -
(no. cases)
(96)
(no.)
(W
Journal of Dairy Sciencc Vol. 78, No. 9, 1995

Page 8

OUR INDUSTRY TODAY
2093
DISCUSSION
Oxytocin treatment may appear to be eco-
nomically appealing as a clinical mastitis ther-
apy because no milk withdrawal time is re-
quired after the use of this product, thereby
allowing the milk to be returned to the bulk
tank as soon as it has a normal appearance.
There was, however, no cost advantage to
using oxytocin over amoxicillin, despite the
fact that the clinical cure rates of the three
treatment groups were similar by the ninth
milking. This cost similarity was due, in part,
to the long time it took for some of the group
0 cows to produce milk with a normal appear-
ance following enrollment on the trial. Many
of the relapses (38%) and recurrences (69%) in
group 0 occurred for cows with Streptococcus
species isolated in the pretreatment milk sam-
pling. The higher cost per episode of clinical
mastitis associated with group C can be partly
explained by the longer (96 h) milk withdrawal
period that was required with the use of this
treatment. Oxytocin treatment may be more
cost effective for herds in which coliform in-
fections predominate and few environmental
streptococcal infections occur. The oxytocin
treatment was used in this study to make the
trial protocol more acceptable to the cooperat-
ing commercial dairy producers, who may
have objected to an untreated control group.
The standard deviation of the cost per episode
of clinical mastitis for the oxytocin treatment
was higher than that associated with either of
the two antibiotic treatments, and risk-averse
producers may choose to use products with
less outcome variability.
TABLE 6. Regression coefficients (kilograms) for 305-d mature equivalent @E)
production for cows enrolled in a randomized field trial of treatments for mild clinical mastitis in California during 1991
to 1992.
-
milk, milk, fat, and 3.5% FCM
Variable
X
SE
P
X
-
SE
P
Intercept
Herd 1
Herd 2
DIL4 1-60
DIL 61-120
DlL >120
Lactation 1
Lactation 2
Lactation 23
Oxytocin
Amoxicillin
Cephapirin
ME Milk1 MilkZ
8782
*3
3060
*
56
960
7510
*
2940 287 <.001 279 <.001
368
350
.88
.01
*
660
1199
*
-292
240
416
379
.12
<.001
*
-317
397
351
340
.37
.25
336
335
FCM6
.39
.48
Fats
Intercept
Herd 1
Herd 2
Lactation 1
Lactation 2
Lactation 23
Oxytocin
Amoxicillin
Cephapirin
283
*
114
*
13
44
*
-2
7840
*
3125
*
500
1238
*
-155
233
10
<.001 274 <.001
15
14
.39
<.001
409
372
.22
<.001
13
12
.89
.52
33 1
330
.64
.48 8
1Rz = .49.
2R2 = .48.
3Category used as the reference for the appropriate fixed effect i n the model.
4DIL = Days in lactation a t the time of the mastitis event that triggered enrollment on the trial.
5Rz = .51.
6RZ = .52.
Journal of Dairy Science Vol. 78, No. 9, 1995

Page 9

2094
VAN EENENNAAM ET AL.
Discarded or nonsalable m i l k per episode of
clinical mastitis has been estimated at 5 (11) to
8 d of production (3). The protocol of the
present study precluded using the intramam-
mary antibiotic treatments for longer than two
(group Cj or three (group A) milkings, which
limited the number of nonsalable milkings tal-
lied in the cost per episode of clinically cured
mastitis associated with the two antibiotic ther-
apies. Clinical recovery occurred within 3 d in
53% of the cases and within 4 to 6 d in 38% of
the cases; the remaining 9% of the cows took
>6 d to recover clinically from clinical masti-
tis. If producers had been free to follow their
own treatment protocol, then intramammary
treatment would most likely have been ex-
tended until the infected quarter appeared to be
clinically normal. In a study o f 629 cases of
clinical mastitis, Kirk found that the duration
of intramammary antibiotic therapy for clinical
mastitis was 5 3 d for 551 (88%) of the cases,
4 to 6 d for 36 (6%) of the cases, and >6 d for
the remaining 42 (6%) of the cases (8). Ex-
tended use of intramammary antibiotic tubes
would increase the cost associated with the
antibiotic therapy because of the need to pur-
chase a greater number of intramammary tubes
and the cost of additional nonsalable milk as-
sociated with the milk withdrawal time of the
extra antibiotic treatments. Our figures for cost
per episode of clinical mastitis underestimate
the costs that would be associated with the
extended (extralabel) use of intramammary an-
tibiotic tubes.
The present results indicate that the mastitic
milk discarded over an entire lactation is con-
siderably more than that associated with the
treatment of an initial mastitis event. Morse et
al. (9) reported that, for a large Florida herd,
mean milk discarded was 8.6, 11.9, 15.9, 18.2,
and 20.6 d for lactations 1 through 5, respec-
tively, from lactations that were treated for
mastitis. The estimated mean cost of discarded
milk, at a price of $1 1.50/45 kg, ranged from
$29.73 to $166.18 per lactation for lactations 1
through 5. These figures are higher than our
results, which showed a mean of 8.0, 12.8, and
13.6 d of discarded milk for lactations 1, 2,
and 3 or greater, respectively. This difference
in results may be because Morse et al. (9)
examined complete lactations, but we exam-
ined partial lactations only for the period re-
maining in the lactation subsequent to the
mastitic event that triggered enrollment on the
trial. Older cows produce more milk than heif-
Journal of D a i r y Science Vol. 78, No. 9, 1995
ers, and so the opportunity cost associated with
nonsalable milk produced by mature cows in
the hospital string for extended periods would
be higher than our estimate.
Relapses and recurrences increased the
mean number of nonsalable milkings to 24.7,
and this value did not differ significantly
among treatment groups. From episodes 2
through 5, the probability of another occur-
rence of clinical mastitis was >.75 (10). The
high percentage of cases that had numerous
nonsalable milkings from repeat occurrences of
mastitis in the present study was therefore not
surprising. The Florida study (9) found that
6.1% of lactations accrued >28 d of discarded
milk over the lactation, accounting for 52.7%
of discarded milk. We also found that a dis-
proportionately large number of the nonsalable
milkings associated with mastitis (55%) were
contributed by 22% of the cows enrolled on
this trial. The repeated occurrences of mastitis,
particularly in older cows, added a substantial
opportunity cost for nonsalable milk to lacta-
tions that involved a mastitic event.
The decreased production and the milk
withheld from market have been estimated to
account for approximately 88% of the total
monetary loss associated with clinical mastitis
(1). Bartlett et al. (2) reported that 70% of lost
marketable milk resulting from clinical masti-
tis was due to withholding, as opposed to 30%
from decreased production. Other factors
generally considered as contributing to the
economic loss attributable to mastitis include
increased culling and mortality, leading to a
decreased rate of genetic progress and more
replacement cows, more labor demands, in-
creased veterinary costs, and decreased milk
quality. We did not try to estimate these other
factors in this trial.
Lactation milk production was not affected
by a mastitic event. Quantification of lost
production following clinical mastitis has been
approached in a variety of ways. Most esti-
mates of lost production following clinical
mastitis have used a case-control design to
compare mastitic cows with unaffected herd-
mates on the basis of monthly DHIA milk
production tests, as was done in the present
trial. The sensitivity of this method is reduced
by the long interval between DHIA test days.
Temporary reductions in production associated
with clinical mastitis can often go undetected
with monthly milk tests. Use of the DHIA
CAR code, adjusting for the effects of mastitis

Page 10

OUR INDUSTRY TODAY
2095
on production, also makes it difficult to esti-
mate accurately from DHIA records the lacta-
tion milk production by mastitic cows (2, 8).
The positive correlation between high milk
production and the development of clinical
mastitis (2, 7) can also lead to the conclusion
that clinical mastitis has a beneficial effect on
305-d milk, which may also explain why we
found no effect of clinical mastitis on overall
milk production.
We found no difference among treatment
groups or among pretreatment bacterial iso-
lates in the survival times of mastitic cows
after the initial enrollment. Clinical mastitis is
an important predictor of culling in heifers and
cows. Cows with clinical mastitis were 2.1
times more likely to be culled than cows with-
out clinical mastitis, and clinical mastitis was
related to the likelihood that the cow would be
culled at 4 5 0 d in lactation (4). Another study
(8) reported that cows with clinical mastitis
were culled (40.9%) at a significantly higher
rate (P = .01) than unaffected herdmates
(32.7%). The high cost of discarded milk as-
sociated with repeated mastitis events in a
lactation with at least one mastitic event indi-
cates that culling may be a rational response to
clinical mastitis, especially for older cows con-
tracting mastitis in the first 120 d of lactation.
Further analysis of other costs, such as replace-
ment costs, would be necessary to determine
whether and when culling is the most econom-
ically optimal response to clinical mastitis.
CONCLUSIONS
The antibiotics studied did not improve
clinical cure rates over oxytocin, but the cost
per episode of mastitis associated with the use
of the labeled number of intramammary
cephapirin tubes was higher than the other two
treatments partially because of the longer milk
withdrawal time (96 h) associated with the use
of this product. The percentage of relapses was
higher for cows in the oxytocin treatment
group, especially when the mastitis-associated
pathogen was an environmental Streptococcus
species. There was no difference among ther-
apy groups in the mean number of nonsalable
milkings per mastitic lactation. Only 17% of
the nonsalable milkings were associated with
the initial episode of clinical mastitis. Most
nonsalable hospital milkings were associated
with the return of the enrolled cows to the
hospital string with a relapse or recurrence of
mastitis in the infected quarter or a new case
of mastitis in another quarter. The additional
nonsalable milkings associated with mastitis
that occurred in the lactation interval subse-
quent to the initial clinical mastitis event added
a substantial opportunity cost to the lactations.
ACKNOWLEDGMENTS
The authors thank the California Milk Ad-
visory Board for its financial support of this
research. In addition, sincere appreciation is
expressed to the owners, herd veterinarians,
and personnel of the commercial dairy opera-
tions that cooperated on this trial,
REFERENCES
1 Bartlett, P. C., G. Y. Miller, C. R. Anderson, and J. H.
Kirk. 1990. Milk production and somatic cell count in
Michigan dairy herds. J. Dairy Sci. 73:2794.
2 Bartlett, P. C.. J. Van Wijk, D. J. Wilson. C. D. Green,
G. Y. Miller, G. A. Majewski, and L. E. Heider. 1991.
Temporal patterns of lost milk production following
clinical mastitis in a large Michigan Holstein herd. J.
D a i r y Sci. 74:1561.
3 Bennett, R. H. 1990. Clinical mastitis from environ-
mental pathogens: analysis of a large commercial
dauy. Page 181 in Proc. lnt. Symp. Bovine Mastitis.
Am. Assoc. Bovine Pract., Indianapolis, IN.
4Erb. H. N., R. D. Smith, P. A. Oltenacu, C. L. Guard,
R. B. Hillman, P. A. Powers, M. C. Smith, and M. E.
White. 1985. Path model of reproductive disorders
and performance, milk fever, mastitis, milk yield, and
culling in Holstein cows. J. Dairy Sci. 68:3337.
5Erskine. R. J. 1991. Therapy of clinical mastitis:
successes and failures. Page 40 in Proc. 30th Annu.
Mte.. Natl. Mastitis Counc. Natl. Mastitis Counc.,
In< '~r~ington,
VA.
6Guterbock. W. M.. A. L. Van Eenennaam, R. J.
Anderson, I. A. Gardner. J. S. Cullor, and C. A.
Holmberg. 1993. Efficacy of intramammary antibiotic
therapy for treatment of clinical mastitis caused by
environmental pathogens. J. Dairy Sci. 76:3437.
7Hoblet. K. H., G. D. Schnitkey, D. Arbaugh, J. S.
Hogan, and K. L. Smith. 1991. Economics of clinical
mastitis. Page 24 in Proc. 30th Annu. Mtg., Natl.
Mastitis Counc. Natl. Mastitis Counc., Inc., Washing-
ton, DC.
8 Kirk, J. H., and P. C. Bartlett. 1988. Economic impact
of m a s t i t i s in Michigan Holstein dairy herds using a
computerized record system. Agri-Practice 9:3.
9 Morse, D., M. A. DeLorenzo, R. P. Natzke, and D. R.
Bray. 1987. Factors affecting days of discarded milk
due to clinical mastitis and subsequent cost of dis-
wded milk. J. Dairy Sci. 70:2411.
IOMorse, D., M. A. DeLorenzo, C. J. Wilcox, R. P.
Natzke, and D. R. Bray. 1987. Occurrence and reoc-
currence of clinical mastitis. J. Dairy Sci. 70:2168.
11 Reneau, J. K., and V. S. Packard. 1991. Monitoring
mastitis. milk quality, and economic losses in dauy
fields. Dairy Food Environ. Sanit. 11:4.
12 Wiggans, G. R., and R. L. Powell. 1980. Projection
factors for milk and fat lactation records. USDA DHI
Lett. 56(1):1.
Journal of Dairy Science Vol. 78, No. 9, 1995