J. Dairy Sci.

TBC
https://doi.org/10.3168/jds.2024-24897
© TBC, The Authors. Published by Elsevier Inc. on behalf of the American Dairy Science Association®.
This is an open access article under the CC BY license (https://creativecommons.org/licenses/by/4.0/).

The economic impact of purulent vaginal discharge

in dairy herds within a single lactation
O. A. Ojeda-Rojas,1 J. Pérez-Báez,2 S. Casaro,1 R. C. Chebel,1 F. Cunha,1 A. De Vries,3 J. E. P. Santos,3,4 F. S.
Lima,5 P. Pinedo,6 G. M. Schuenemann,7 R. C. Bicalho,8 R. O. Gilbert,9 S. Rodriguez-Zas,10 C. M. Seabury,11 G.
Rosa,12 W. W. Thatcher,3,4 and K. N. Galvão1,4*
1
Department of Large Animal Clinical Sciences, University of Florida, Gainesville, FL, 32610
2
Instituto Nacional de Investigaciones Agropecuarias, Facultad de Ciencias Agronómicas y Veterinarias, Universidad Autónoma de Santo
Domingo, Santo Domingo, Dominican Republic, 10904
3
Department of Animal Sciences, University of Florida, Gainesville 32610
4
D. H. Barron Reproductive and Perinatal Biology Research Program, University of Florida, Gainesville 32610
5
Department of Population Health and Reproduction, University of California, Davis 95616
6
Department of Animal Sciences, Colorado State University, Fort Collins 80521
7
Department of Veterinary Preventive Medicine, The Ohio State University, Columbus 43210
8
FERA Diagnostics and Biologicals Corp. College Station, 77845
9
School of Veterinary Medicine, Ross University, PO Box 334, Basseterre, St. Kitts, West Indies
10
Department of Animal Sciences, University of Illinois, Urbana-Champaign 61802
11
College of Veterinary Medicine, Texas A&M University, College Station 77843
12
Department of Animal Sciences, University of Wisconsin, Madison 53706

ABSTRACT inputs. Cows with PVD produced less milk (9,753.2 ±

333.6 vs. 9,994.6 ± 330.9 kg/cow), were less likely to
The goal of this study was to calculate the cost of pu- be pregnant (70.7 ± 1.7 vs. 78.9 ± 1.2%), and were more
rulent vaginal discharge (PVD) in dairy cows. The data likely to be culled by 305 DIM (34.6 ± 1.7 vs. 27.2 ±
set included 11,051 cows from 16 dairy herds located 1.3%) compared with cows without PVD. Consequently,
in 4 regions of the US. Purulent vaginal discharge was milk sales (4,744.7 ± 162.3 vs. 4,862.1 ± 161.0 $/cow)
characterized as a mucopurulent, purulent, or reddish- and residual cow value (1,079.6 ± 23.0 vs. 1,179.3 ± 20.3
brownish vaginal discharge collected at 28 ± 7 d in milk. $/cow) were lesser for cows with PVD. Replacement
Gross profit was calculated as the difference between (639.4 ± 26.4 vs. 526.0 ± 23.4 $/cow) and reproductive
incomes and expenses, and the cost of PVD was calcu- management costs (76.3 ± 2.5 vs. 69.0 ± 2.4 $/cow) were
lated by subtracting the gross profit of cows with PVD greater for cows with PVD. The mean cost of PVD was
from the gross profit of cows without PVD. Continuous $202. The stochastic analysis also showed a mean cost
outcomes such as milk production (kg/cow), milk sales of $202, ranging from $152 to $265. The robust data set
($/cow), cow sales ($/cow), feed costs ($/cow), repro- and the stochastic analysis strengthen both the external
ductive management cost ($/cow), replacement costs ($/ and internal validity of our findings, offering a deeper
cow), and gross profit ($/cow) were analyzed using lin- understanding of the economic consequences of PVD.
ear mixed-effects models. Pregnancy and culling by 305 In conclusion, PVD resulted in large economic losses
DIM were analyzed by generalized linear mixed-effects to dairy herds by being associated with decreased milk
models using logistic regression. Models included the yield, impaired reproduction, and greater culling.
fixed effects of PVD, metritis, parity, region, season of Key words: economics, uterine diseases, dairy cow
calving, and morbidity in the first 60 DIM, as well as the
interactions between PVD and metritis, PVD and parity
INTRODUCTION
group, and PVD and morbidity. Farm and the interaction
between PVD and farm were considered random effects The current scientific consensus is that uterine diseases
in all the statistical models. A stochastic analysis was have a negative association with the overall productive
conducted using 10,000 iterations with varying relevant and reproductive performance of dairy cows (Ribeiro et
al., 2013; Edelhoff et al., 2020; LeBlanc, 2023), which
results in significant economic losses (Bartlett et al.,
Received March 11, 2024. 1986; Lima et al., 2019; Pérez-Báez et al., 2021). It is re-
Accepted October 30, 2024. ported that the prevalence of purulent vaginal discharge
Declarations of interest: none
*Corresponding author: Klibs N. Galvão: galvaok@​ufl​.edu (PVD) among lactating dairy cows is approximately

The list of standard abbreviations for JDS is available at adsa.org/jds-abbreviations-24. Nonstandard abbreviations are available in the Notes.
 Ojeda-Rojas et al.: ECONOMICS OF PURULENT VAGINAL DISCHARGE

20%, with prevalence within herds varying from 5% to Diagnosis of Purulent Vaginal Discharge
> 30%, respectively (LeBlanc et al., 2002; Galvão et al.,
2009; Pinedo et al., 2020). The diagnosis of PVD was based on vaginal discharge
Purulent vaginal discharge has been associated with collected at 28 ± 7 DIM. The discharge within or adher-
delayed resumption of ovarian cyclicity (Santos et al., ent to the Metricheck device was scored using a 0 to 5
2010), reduced rate of pregnancy per artificial insemina- scoring system: 0 = no discharge, 1 = clear discharge,
tion (AI) (LeBlanc et al., 2002; Santos et al., 2010), lon- 2 = flecks of purulent material within otherwise clear
ger calving-to-pregnancy interval (Galvão et al., 2009; mucus, 3 = mucopurulent but < 50% purulent material,
Giuliodori et al., 2013), and increased risk of pregnancy 4 = mucopurulent with > 50% purulent material, and 5
losses (Galvão et al., 2009; Ribeiro et al., 2013). Further- = mucopurulent with > 50% purulent material and fetid
more, some studies have shown that PVD negatively af- discharge (McDougall et al., 2007). Cows with vaginal
fects milk yield (Paiano et al., 2019; Paiano et al., 2021; discharge ≥3 were considered to have PVD. The diag-
Husnain et al., 2023), all of which increase the risk of nostic criteria for PVD was based on the study’s data set.
culling (LeBlanc et al., 2002) when compared with cows A survival analysis was performed (Supplemental Figure
without PVD. S1; https:​/​/​doi​.org/​10​.6084/​m9​.figshare​.27310476) to
Despite the significant evidence showing a negative evaluate the time to pregnancy according to the vaginal
association between PVD and cow performance, the cost discharge score. Cows with vaginal discharge score ≥3
of PVD has not been investigated using cow-level data. had reduced (P < 0.05) hazard of pregnancy within 305
Hence, understanding the economic impacts of PVD DIM compared with scores 0, 1, and 2, and there was no
would contribute to informing management decisions difference (P > 0.05) between scores 0, 1, and 2.
that could increase the economic sustainability of dairy Study Population. To be included in the data analysis,
farms. Therefore, the objective of this study was to cal- cows must have had information about the diagnosis of
culate the cost of PVD in dairy cows within a single 305 metritis or PVD, milk yield information from the DHIA
DIM lactation using a data set from 16 herds in 4 regions test, type of breeding (AI after estrus detection or TAI)
of the US. at first service, and information regarding survival in the
herd.
Accordingly, from the 11,733 initial population, 682
MATERIALS AND METHODS
cows were excluded. For metritis, 41 cows did not have
This study derived from a previous study that aimed vaginal discharge scoring. Of those, 3 cows were sold
to identify genomic markers for fertility in Holstein and 2 died before diagnosis. For PVD, 588 cows did not
cows (Seabury et al., 2023). The initial data set included have vaginal discharge scoring. Of those, 228 cows were
11,733 cows (34.9% primiparous and 65.1% multipa- sold and 162 died before diagnosis. For milk yield, 53
rous cows) from 16 dairy farms located in the Northeast cows did not have milk production information, meaning
(22.5% of cows; 4 farms), Midwest (44.2% of cows; 6 none of the 10 monthly tests were recorded. Therefore,
farms), Southeast (10.0% of cows; 1 farm), and South- 11,051 cows (35.6% primiparous and 64.4% multiparous
west (23.3% of cows; 5 farms) of the US. Cows calved cows) were used in the current study.
between 2012 and 2014. The farms were visited once or Economic calculations. Incomes, expenses, and gross
twice weekly, and cows were enrolled in the study at par- profit were calculated for individual cows based on data
turition and monitored up to 305 DIM or until they were retrospectively collected by the authors and applying av-
removed (death/sale) from the herd, whichever happened erage prices collected over a 10-year period (2012–2021).
first. General information about cows such as cow num- Subsequently, income, expenses, and gross profit were
ber, calving date, lactation number, calving-related prob- statistically analyzed comparing cows with PVD and
lems (e.g., dystocia, stillbirth, twins), clinical mastitis, without PVD, adjusting for the effects of parity, region,
displacement of abomasum, pneumonia, type of AI (i.e., season of calving, and the occurrence of other diseases.
AI on estrus or timed AI, TAI), DIM at AI, pregnancy
diagnosis outcomes, sold or died, DIM leaving the herd, Income
and milk yield records from DHIA monthly tests were
collected up to 305 DIM from the farm herd management Income was calculated for each cow and included rev-
software. Cows were examined by the study personnel enues from milk sales by 305 DIM, the sale of the cow,
for the diagnosis of retained fetal membranes, subclinical or the residual cow value if the cow was not sold or died
ketosis, lameness, metritis, and PVD. Additional details by 305 DIM. Cows were tracked until 305 DIM and those
regarding the study population monitoring of cows and that were not pregnant by that time were considered to
disease definition and diagnosis are in Pinedo et al. have been sold. Cow sale value was calculated using the
(2020). following formula:

Journal of Dairy Science Vol. TBC No. TBC, TBC

 Ojeda-Rojas et al.: ECONOMICS OF PURULENT VAGINAL DISCHARGE

cowSaleValue = pricePerKg × BW, multiparous cows sold × average BW of multiparous

cows × pricePerKg],
where pricePerKg was $2.1/kg. This price was calculated
using historical (2012 to 2021) beef prices in the US (US- where % of primiparous and multiparous cows sold was
DA-NASS, 2022a). The nominal prices were deflated by 21.3 and 78.7%, respectively. The average BW of pri-
the gross domestic product price index of the US Bureau miparous and multiparous cows was 591.7 and 703.1 kg,
of Economic Analysis and converted into real values for respectively. The pricePerKg was $2.1/kg.
July 2022 (BEA, 2023). Body weight, expressed in kg, Milk production was measured monthly by the DHIA
was calculated using the following regression equations. for a maximum of 10 tests. Total milk yield per cow was
For primiparous (For the first 26 DIM): calculated from their monthly test using adjustment fac-
tors for milking interval, time of milking, and DIM as de-
BW = 584.4 – 3.91 × DIM + 0.081(DIM)2. scribed by Wiggans (1985). Cows with at least one DHIA
test milk yield were included in the analysis. Assuming
For primiparous (From 27 to 305 DIM): a 30-d interval between DHIA tests, the number of miss-
ing tests was calculated. In summary, 9,464 cows did not
BW = 520.6 – 0.42 × DIM + 0.00013(DIM)2. miss any monthly milk yield tests, 805 cows missed one
test, 642 cows missed 2 tests, 118 cows missed 3 tests,
For multiparous (For the first 26 DIM): 19 cows missed 4 tests, and 3 cows missed 5 tests. The
price per kg of milk was estimated using historical prices
BW = 701.6 – 3.47 × DIM + 0.055(DIM)2. over the previous 10 years (2012 – 2021; USDA-NASS,
2022b). Nominal values were deflated, like in cow sales
For multiparous (From 27 to 305 DIM): (BEA, 2023). Income from milk sales was calculated by
multiplying the total milk yield per cow (kg) by $0.5,
BW = 647.6 – 0.14 × DIM + 0.00085(DIM)2. which was the mean of the historical milk prices per kg
adjusted for inflation.
The regression equations were obtained using daily
BW for a 305-d lactation from 81 primiparous and 79 Expenses
multiparous from the University of Florida Dairy Unit
in Gainesville. The R2 in all equations was larger than Expenses were calculated for each cow, including
92.0% (Pérez-Báez et al., 2021). feed, reproductive management, and replacement costs
Cows pregnant by 305 DIM were retained in the herd of cows that were sold or died. Feed costs were calcu-
and a residual value was calculated using the following lated assuming all lactating cows were fed a TMR with
formula: a 1.75 Mcal of NEL/kg of DM. The net milk energy was
estimated to be 0.7 Mcal/kg based on milk containing
residualCowValue = replacementCost 3.5% fat, 5.0% lactose, and 3.2% true protein. Therefore,
– [lactationNumber × (replacementCost – each kg of marginal DM consumed supported 2.5 kg of
averageSalvageValue)/lactationCategories]. milk, and the net energy for lactation needs for mainte-
nance was calculated as 0.1 Mcal/kg of BW0.75. The BW
where replacementCost was $1,831.2, which was the was the mean of BW between the estimated BW at calv-
mean price for replacement heifers from 2012 to 2021 ing and the estimated BW on the last lactation day up to
(USDA-NASS, 2022b). lactationNumber was represented 305 DIM. The regression equations provided above for
as follows: 1, 2, and 3, for lactations 1 to 3, respectively, calculating BW at sale were used.
and 4 for lactation 4 or greater. lactationCategories cor- To estimate the DMI needed for maintenance, the daily
responds to the number of lactation categories, i.e., 4. maintenance requirements in kilograms of DM were mul-
This cut-off for lactation number was used because 94% tiplied by the number of days the cow was in lactation,
of the cows in our data set had up to 4 lactations, whereas considering the time until death, sale, or a cow reaching
the remaining 6% ranged from 5 to 9 lactations (Supple- 305 DIM, whichever happened first. The DMI needed to
mental Figure S2; https:​/​/​doi​.org/​10​.6084/​m9​.figshare​ support milk production was estimated by multiplying
.27310476). The averageSalvageValue was $1,438.8 and the total milk yield in the lactation by 0.7 Mcal/kg and
was calculated using the following formula: then dividing it by 1.75 Mcal/kg.
To summarize, the calculation of feed cost for each
averageSalvageValue = [%of primiparous cows sold × lactating cow involved multiplying the estimated DMI
averageBW of primiparous cows × pricePerKg] + [%of for maintenance and production by the cost of the TMR
for lactating cows during the period from 2010 to 2021
Journal of Dairy Science Vol. TBC No. TBC, TBC
 Ojeda-Rojas et al.: ECONOMICS OF PURULENT VAGINAL DISCHARGE

($0.3/kg of DM, USDA-ERS, 2023). The price per kg of subsequent AI was $23.4 as this was the average cost
of DM was adjusted for inflation using the previously between AI after estrous detection and TAI, excluding
described approach. the cost of the presynchronization protocol only used for
A detailed description of reproductive management the first AI.
was previously reported (Pinedo et al., 2020). Briefly, the The cost of performing a pregnancy diagnosis was
reproductive management program included a combina- $4.7/cow, considering that a veterinarian could perform
tion of estrus detection and TAI for the first and subse- 30 cows/h and the labor for pregnancy diagnosis was
quent AI; however, there were variations among farms. $140/h. Additionally, it was assumed that 50% of the in-
For reproductive management cost calculation, every seminated cows return spontaneously to estrus and would
cow was assumed to be enrolled in a Presynch-Ovsynch not receive a pregnancy diagnosis. For a pregnant cow,
program (Moreira et al., 2001). Briefly, the presynchro- the cost of 2 pregnancy diagnoses was considered (32 ± 3
nization program included 2 PGF2α injections given 14 and 60 ± 3 d after AI). Hence, the cost of pregnancy diag-
d apart, with the option of receiving AI in estrus after the nosis for a cow pregnant after one AI was $9.3. For cows
second PGF2α of the presynchronization program. The that received more than one AI and became pregnant,
TAI protocol included an injection of GnRH 12 d after the cost included 2 pregnancy diagnoses plus the cost
the second injection of PGF2α of the presynchronization of pregnancy diagnosis multiplied by 0.5 ($2.3) for each
program. Seven days later, cows received a PGF2α injec- additional AI that did not end in pregnancy. For cows that
tion, and 48 h later, a second injection of GnRH. Timed received one or more AI and did not become pregnant,
artificial insemination was performed 16 to 20 h after the the number of AI was multiplied by the cost of pregnancy
second GnRH. Records for the first service were available diagnosis multiplied by 0.5 ($2.3).
for 10,470 cows. Of those, 4,978 were AI after detected Cows that died or were sold were replaced by a first-
in estrus, and 5,492 were TAI. Data for 581 cows were lactation cow, which was assumed to cost $1,831.2
not available. Of those, 398 cows left the study by 68 ± (USDA-NASS, 2022d).
48 DIM (±SD) and 183 were classified as not eligible
for AI. Information about the type of service for the sec- Gross Profit Calculation
ond and subsequent AI was unavailable. Pregnancy was
diagnosed via ultrasonography at 32 ± 3 d after AI and Gross profit per cow was calculated as a difference
confirmed at 60 ± 3 d after AI. Pregnancy outcome at 32 between income and expenses. The primary source of
± 3 d after AI was considered for this study. A cow that income was milk sales. Additional income was generated
received AI but did not receive a pregnancy diagnosis by culled cow sales and the residual cow value at the end
was considered nonpregnant at the end of the collection of lactation for cows that survived to 305 DIM. Expenses
period. The cost of reproductive hormones was $2.7/dose were divided into feed, reproductive management, and
for PGF2α and $2.0/dose for GnRH, according to the replacement costs. All the calculations were done using
prices provided by Southeast Milk Inc. Cost of labor for a Microsoft Excel spreadsheet (Microsoft Corporation,
estrous detection, hormone administration, and AI was Redmond, WA). The following equation was used to es-
assumed to be $14.9/h. Labor price was the mean value timate the gross profit:
of 10 years (2012–2021; USDA-NASS, 2022c). The cost
of estrus detection was $0.1 per cow, assuming 120 cows grossProfit = (milkProduction + cowSaleValue +
checked per hour and that each cow had 24 estrus checks. residualCowValue) − (feedCost + reproductiveCost +
The cost per injection administration was $0.3, assuming replacementCost).
60 injections per hour, plus $0.05 for supplies. The cost
per insemination was $16.0, including $0.5/cow for la-
bor, considering that an AI technician can inseminate 30 Cost of Purulent Vaginal Discharge
cows/h at the cost of $14.9/h, with semen cost at $15.0/
dose and supplies (AI sheath, sleeve, semen applicator, The cost of PVD was the difference between the av-
water bath, and chalk) cost for each insemination of erage gross profit in cows with PVD and cows without
$0.50. Therefore, the cost of the first service in cows in PVD, adjusted by other postpartum diseases, parity
which estrous detection was used was $24.9, including group, region, and season of calving.
the cost of pre-synchronization, labor cost for hormonal
administration, cost of daily heat detection, labor cost of Statistical Analyses
AI, and cost of AI. In cows that were inseminated using
TAI, the cost for the first service was $32.6, including Assumptions of normality were evaluated by visual in-
the same items for cows inseminated after estrous de- spection of residuals, while homogeneity of variance was
tection, plus the cost of the Ovsynch program. The cost
Journal of Dairy Science Vol. TBC No. TBC, TBC
 Ojeda-Rojas et al.: ECONOMICS OF PURULENT VAGINAL DISCHARGE

assessed by Levene tests. The association of PVD with of calving, and time (month postpartum), as well as the
dependent variables was analyzed by ANOVA. Statistical interactions between PVD and metritis, PVD and time,
models included the fixed effects of PVD (yes or no), metritis and time, and the 3-way interaction among PVD,
metritis (yes or no), parity group (primiparous or multipa- metritis, and time. Farm and the interaction among PVD,
rous), region (Midwest, Northeast, Southeast, and South- metritis, and farm were considered random effects.
west), season of calving (cold, September to January;
warm, April to August), morbidity (yes or no), as well as Stochastic Analysis
the interactions between PVD and metritis, PVD and par-
ity group, and PVD and morbidity. Morbidity included To incorporate uncertainty, the stochastic Monte
diseases other than metritis or PVD, such as displaced Carlo simulation model was developed using @Risk 8.2
abomasum, lameness, mastitis, and respiratory disease, (Palisade Corporation, Ithaca, NY), an add-in software
all diagnosed in the first 60 DIM. Cows diagnosed with program for Microsoft Excel. A stochastic Monte Carlo
one or more of those 4 diseases were considered to have simulation accounted for the differences in gross profit
morbidity, whereas those without any of those diagnoses between cows with and without PVD under varying mar-
were considered without morbidity. A manual backward ket conditions for the most relevant inputs. This method
elimination procedure was used to build the final models, allows for a more comprehensive understanding of how
retaining only significant (P ≤ 0.10) predictors. Fixed ef- outcomes fluctuate due to changes in external prices,
fects with P > 0.10 were excluded from each model one allowing for a better understanding of the variability in
at a time, according to the largest P-value, until only pre- different scenarios. Outcomes were the distributions of
dictors with P ≤ 0.10 remained in the models. Farm and difference in gross profit between cows with PVD and
the interaction between PVD and farm were considered cows without PVD and the contribution of individual in-
random effects in all statistical models, and the latter was put costs to the total variation in gross profit differences.
the error term to test the association of PVD with the Fixed inputs were the cumulative milk produced (kg/
responses analyzed. cow) by 305 DIM, estimated DMI (kg/cow) by 305 DIM,
Continuous dependent variables such as milk produc- the percentage of cows sold by 305 DIM, the percentage
tion (kg/cow), milk sales ($/cow), cow sales ($/cow), feed of cows culled by 305 DIM, and the percentage of cows
costs ($/cow), reproductive management cost ($/cow), with reproductive management cost greater than zero.
replacement costs ($/cow), and gross profit ($/cow) were Stochasticity was incorporated into the milk price, re-
analyzed by ANOVA with linear mixed-effects models placement cost, cow sales, feed cost, residual cow value,
using the MIXED procedure of SAS (ver. 9.4, SAS In- and reproductive management cost.
stitute Inc., Cary, NC). Because the distribution of the Distributions of these items were modeled using default
residuals of reproductive management cost, replacement parameters of the batch fit option in @Risk. The Akaike’s
cost, cow sales, residual cow value, and gross profit were information criterion was used to select each variable's
either skewed to the left or right, we also analyzed them best distributional fit and distributional parameters.
using the non-parametric Wilcoxon rank sum test with Specifically, milk price was fitted with a log-logistic dis-
the NPAR1WAY procedure of SAS. Supplemental Table tribution, replacement cost a triangular distribution, and
S1 (https:​/​/​doi​.org/​10​.6084/​m9​.figshare​.27310476) de- cow sales with a Kumaraswamy distribution. Feed cost
picts the minimum, maximum, median, lower quartile, was fitted with a triangular distribution, residual cow
upper quartile, and interquartile range for each variable. value with a Pert distribution, and reproductive manage-
Analyzing non-normally distributed data with parametric ment cost with a Gamma distribution. These distributions
methods can lead to Type I or Type II errors, especially were then used to conduct risk simulations using 10,000
with small sample sizes; however, this concern is mini- iterations. Gross profit risk was assessed using the fol-
mized in our study (Vickers, 2005; Knief and Forstmeier, lowing equations:
2021). Dichotomous dependent variables, such as the
pregnancy by 305 DIM and removal (death/sale) from diffGrossProfit = (grossProfitCE) – (grossProfitNoCE),
the herd by 305 DIM, were analyzed using generalized
linear mixed-effects models using the GLIMMIX proce- where grossProfit was calculated with the same formula
dure of SAS. The results were reported using the LSM, used above for gross profit calculation.
SEM, and 95% confidence intervals. Significance was
determined when P < 0.05. RESULTS
Milk production for the first 10 monthly tests was
analyzed by ANOVA for repeated measures using the The prevalence of PVD in the study was 25.5%, being
MIXED procedure of SAS. Models included the fixed 27.2% in primiparous cows and 24.6% in multiparous
effects of PVD, metritis, parity group, region, season cows. The number of cows evaluated per herd ranged
Journal of Dairy Science Vol. TBC No. TBC, TBC
 Ojeda-Rojas et al.: ECONOMICS OF PURULENT VAGINAL DISCHARGE

from 281 to 1,195, and the prevalence of PVD among Stochastic simulation
herds varied from 11.8% to 34.7%.
Table 1 displays the productive, reproductive, and eco- Based on the 10,000 iterations simulated, the mean
nomic outcomes of cows with and without PVD. difference in gross profit between cows with PVD and
Cows with PVD had lesser (P < 0.01) total milk yield those without PVD was −$202, with 95% of the scenari-
by 305 DIM (Table 1), which resulted in $117 lesser (P os falling between −$265 and −$152. The minimum and
< 0.01) income from milk sales compared with cows maximum values of difference in gross profit for cows
without PVD. There was an interaction (P < 0.01) be- with and without PVD were −$339 and −$113, respec-
tween PVD and month postpartum on milk yield because tively (Figure 2A). The tornado plot (Figure 2B) shows
cows with PVD had less (P < 0.05) milk yield in monthly the contribution of select variables for the variability in
tests 1 (−1.08 kg/d), 2 (−0.82 kg/d), 4 (−0.83 kg/d), 5 gross profit difference between cows with PVD and those
(−0.77 kg/d), and 6 (−0.56 kg/d) compared with cows without PVD. The variability in gross profit difference
without PVD (Figure 1). There were no differences in was predominantly caused by replacement (48.7%) and
milk yield between cows with PVD and cows without milk price (37.1%). Cow sales, residual cow value, and
PVD in tests 3 and 7 to 10. Cows with PVD had lesser feed price contributed 7.9, 3.5, and 2.8% of the variation,
(P < 0.01) estimated DMI than cows without PVD, thus respectively (Figure 2B). Spearman correlation coeffi-
leading to almost $31 less (P < 0.01) feed costs compared cients were calculated between the inputs and the mean
with cows without PVD. Purulent vaginal discharge was cost of PVD. Replacement price had the greatest correla-
associated with a reduced (P < 0.01) proportion of cows tion (ρ = −0.70), followed by milk price (ρ = −0.56), cow
pregnant by 305 DIM (Table 1). Because of decreased sales (ρ = 0.22), residual cow value (ρ = −0.16), feed
reproductive performance, cows with PVD had greater price (ρ = 0.16), and reproductive cost (ρ = 0.03). The
(P < 0.01) reproductive management cost than cows spider plot in Figure 2C offers insights into the relation-
without PVD. The decrease in reproductive efficiency ship between the specific percentile range of the top 5
in cows with PVD affected culling dynamics, as cows input variables and the rate of change in the mean differ-
with PVD had an increased risk of leaving the herd by ence in gross profit between cows with and without PVD.
305 DIM (Supplemental Figure S3; https:​/​/​doi​.org/​10​ Inputs characterized by steeper lines in the spider plot
.6084/​m9​.figshare​.27310476). Thus, cows with PVD had had a greater impact on the response evaluated.
$113 greater (P < 0.01) cost associated with replacement.
Cows with PVD had greater (P < 0.01) income from cow DISCUSSION
sales than cows without PVD, but reduced (P < 0.01)
residual cow value. Altogether, PVD was associated with Cows with PVD had decreased pregnancy, milk yield,
a $202 reduction (P < 0.01) in gross profit per cow com- and survival in the herd by 305 DIM compared with cows
pared with cows without PVD. without PVD. As a result, the mean cost of PVD was
$202. The stochastic simulation showed that the cost of
PVD in 95% of the scenarios ranged from $152 to $265/
case. An essential component of this study was that all
responses were evaluated against the random effect of the

Table 1. Productive, reproductive, and economic results based on diagnosis of purulent vaginal discharge (LSM ± SEM)

Item PVD1 95% CI2 No PVD 95% CI Diff3 P-value

Cows, number 2,818 8,233 — —
Estimated DMI, kg 5,933.6 ± 146.2 5,624.2 - 6,242.9 6,035.9 ± 144.2 5,729.5 - 6,342.4 −102.3 <0.01
Milk by 305 DIM, kg 9,753.2 ± 333.6 9,046.1 - 10,460.0 9,994.6 ± 330.9 9,291.3 - 10,698.0 −241.4 <0.01
Pregnant by 305 DIM, % 70.7 ± 1.7 67.0 - 74.1 78.9 ± 1.2 76.2 - 81.4 −8.2 <0.01
Culled by 305 DIM, % 34.6 ± 1.7 31.2 - 38.3 27.2 ± 1.3 24.5 - 30.1 7.4 <0.01
Milk sales by 305 DIM, $/cow 4,744.7 ± 162.3 4,400.7 - 5,088.7 4,862.1 ± 161.0 4,519.9 - 5,204.2 −117.4 <0.01
Cow sales, $/cow 458.1 ± 21.2 414.9 - 501.3 371.9 ± 19.0 332.7 - 411.2 86.2 <0.01
Residual cow value, $/cow 1,079.6 ± 23.0 1,032.8 - 1,126.4 1,179.3 ± 20.3 1,137.2 - 1,221.4 −99.7 <0.01
Feed costs by 305 DIM, $/cow 1,789.3 ± 44.1 1,695.9 - 1,882.6 1,820.1 ± 43.5 1,727.7 - 1,912.6 −30.8 <0.01
Reproductive management cost, $/cow 76.3 ± 2.5 71.2 - 81.4 69.0 ± 2.4 64.1 - 73.9 7.3 <0.01
Replacement costs, $/cow 639.4 ± 26.4 585.8 - 693.0 526.0 ± 23.4 477.7 - 574.3 113.4 <0.01
Gross profit, $/cow 3,780.4 ± 114.5 3,539.4 - 4,021.4 3,982.7 ± 112.8 3,745.5 - 4,219.8 −202.3 <0.01
1
PVD = purulent vaginal discharge.
2
CI = confidence interval, lower - upper.
3
Diff = difference between cows that developed purulent vaginal discharge and those that did not.

Journal of Dairy Science Vol. TBC No. TBC, TBC

 Ojeda-Rojas et al.: ECONOMICS OF PURULENT VAGINAL DISCHARGE

negative association between PVD and the proportion of

cows pregnant by 305 DIM and their subsequent cull-
ing. This culling accounts for 25% of the mean cost and
is only surpassed by milk income, which accounted for
26% (Figure 3). The stochastic analysis conducted herein
showed that replacement price had the greatest impact on
the variation of the mean PVD cost, accounting for 48.7%
of the variance. Additionally, the analysis conducted by
Lhermie et al. (2018) did not consider the reduction in
milk production associated with PVD, as the literature
they used (LeBlanc, 2008; Dubuc et al., 2011) did not
report differences in milk yield between cows with and
without PVD. Furthermore, the feed costs calculation in
Lhermie et al. (2018) and the present study is divergent.

Figure 1. Milk yield in the first 10 mo postpartum for cows with

purulent vaginal discharge (PVD) and those without PVD (No PVD).
Definition of purulent vaginal discharge: vaginal discharge ≥3 (score 0 =
no discharge, 1 = clear discharge, 2 = flecks of purulent material within
otherwise clear mucus, 3 = mucopurulent but <50% purulent material,
4 = mucopurulent with >50% purulent material, and 5 = mucopurulent
with > 50% purulent material and fetid discharge; McDougall et al.,
2007) at 28 ± 7 d postpartum. * denotes statistical difference (P < 0.05).

interaction between PVD and farm. Thus, the observed

negative associations between disease and performance
and gross profit per cow were applicable to all 16 farms
investigated, which enhances the external validity of our
findings.
Extensive research has been conducted on the preva-
lence of, risk factors for, and treatment of PVD and its
associations with reproductive and productive perfor-
mances (LeBlanc et al., 2002; Williams et al., 2005;
Husnain et al., 2023). Nonetheless, it is worth noting that
only one previous publication addressed the economic
impact of PVD. Lhermie et al. (2018) estimated the cost
of PVD to range from $196 to 239 (inflation-adjusted)
per case as part of their economic analysis of the impact
of reducing antimicrobial usage in dairy herds. The cost
reported in the current study ($202/cow) falls within
this range. The present study, however, differs from
Lhermie et al. (2018) in several aspects. Their analysis
utilized simulation and a partial budgeting framework
exclusively based on data available from the literature
rather than observed values, such as in the current study.
Most notably, the value of a replacement first-lactation
cow in their study was $2,432 (inflation-adjusted). This
value was substantially larger than the one used herein
($1,831). In our study, if replacement cost were to be
increased from $1,831 to $2,432 per first lactation cow, Figure 2. (A) Relative frequency distribution for differences in gross
PVD cost would have to be increased from $202 to $261 profit between cows with purulent vaginal discharge (PVD) and those
per case. For the current analysis, replacement cost was without PVD for 10,000 iterations of stochastic simulation. (B) Tornado
graph showing the contribution to the variance of the variables included
a critical component of the PVD cost. This item was the in the model. (C) Spider graph showing the mean change across percen-
second largest component of PVD cost because of the tile range of the top 5 input variables included in the model.

Journal of Dairy Science Vol. TBC No. TBC, TBC

 Ojeda-Rojas et al.: ECONOMICS OF PURULENT VAGINAL DISCHARGE

The feed price used by Lhermie et al. (2018) was lower by reproductive performance (De Vries, 2006; Galvão
when compared with the one used in our study ($0.23 vs. et al., 2013; Cabrera, 2014;). Reproductive performance
$0.30 per kg). Although this difference did not specifi- directly affects factors such as the proportion of lactating
cally impact one specific group of cows, either cows with cows versus dry cows and the milk production per cow,
PVD or cows without PVD, it did result in lower overall which impact herd profitability (Galvão et al., 2013).
costs. Purulent vaginal discharge impairs reproductive perfor-
The economic impacts of uterine diseases depend on mance through several mechanisms, including reduced
the cost of treatment, milk yield losses, and infertility fertilization, impaired embryo quality, impaired elonga-
(Overton and Fetrow, 2008; Lima et al., 2019; Pérez- tion of the conceptus, and altered uterine environment
Báez et al., 2021). Still, milk income is the primary (Ribeiro et al., 2016; Dickson et al., 2020). In the current
source of income for dairy farms (White et al., 2002). In study, the proportion of cows pregnant by 305 DIM was
the current study, cows with PVD produced 241 kg less 8 percentage points smaller in cows with PVD compared
milk by 305 DIM than cows without PVD, and monthly with cows without PVD. It is worth noting that non-
differences were observed since the beginning of the lac- pregnant cows by 305 DIM were assumed to have been
tation. Because greater culling in cows with PVD would sold in the current study. Consequently, income from cow
contribute to lower total milk yield, we also compared sales among those with PVD was $86 greater than those
milk yield using an ANOVA for repeated measures ap- without PVD. This income increase was insufficient to
proach, which is not influenced by culling. Using this ap- compensate for the increase in replacement cost associ-
proach, we observed that both the differences in monthly ated with PVD, which was $113 more in cows with PVD
milk yield in early lactation and greater culling in late compared with cows without PVD. Moreover, residual
lactation accounted for the observed difference in total cow value decreased in PVD compared with cows with-
milk yield. Accordingly, lesser total milk yield in PVD out PVD. The difference in residual cow value between
cows led to decreased income from milk sales by $117 cows with PVD and cows without PVD is explained by
per case. According to our findings, more recent studies lower reproductive performance and higher culling ob-
have observed a decrease in milk yield in cows with PVD served in cows with PVD. Uterine diseases damage the
(Paiano et al., 2019; Paiano et al., 2021; Husnain et al., endometrium and uterine glands (Bonnett et al., 1991),
2023). In the context of this current study, approximately alter the endocrine function (Gilbert et al., 1990; Sheldon
74% of the total income was attributed to milk sales, et al., 2009), and perturb ovarian follicular activity (Wil-
making it the primary source of income and, therefore,
the major factor adding to the difference in gross profit
between cows with PVD and cows without PVD, which
accounts for 26% of the difference (Figure 3).
Although the DMI data presented herein is an estima-
tion, reduced DMI is commonly observed in cows with
uterine diseases and could contribute to decreased milk
production (Bell and Roberts, 2007; Huzzey et al., 2007;
Pérez-Báez et al., 2019). In the study conducted by
Pérez-Báez et al. (2019), the authors showed that cows
that developed metritis consumed less DM in the first
28 d postpartum compared with cows without metritis.
Interestingly, DMI recovered toward the end of the study
(28 d postpartum), but milk yield did not. This pattern
highlights the role of uterine infections in nutrient par-
titioning. In addition, based on the current findings, ap-
proximately 42% of cows diagnosed with metritis later
developed PVD. This finding implies that a portion of
the cows may still have an ongoing inflammatory process
attributed to PVD that could have impaired milk yield at
d 28 postpartum, as examined in the study by Pérez-Báez Figure 3. Pie chart displaying the components of the cost of puru-
lent vaginal discharge (PVD). The values represent the proportions of
et al. (2019). the total cost. A case of PVD costs $202. Definition of PVD: Vaginal
Purulent vaginal discharge has been associated with discharge ≥3 (score 0 = no discharge, 1 = clear discharge, 2 = flecks of
impaired reproductive performance (McDougall et al., purulent material within otherwise clear mucus, 3 = mucopurulent but
<50% purulent material, 4 = mucopurulent with >50% purulent material,
2007; Sheldon et al., 2009; Giuliodori et al., 2013). It and 5 = mucopurulent with > 50% purulent material and fetid discharge;
also has been shown that profitability is directly affected McDougall et al., 2007) at 28 ± 7 d postpartum.

Journal of Dairy Science Vol. TBC No. TBC, TBC

 Ojeda-Rojas et al.: ECONOMICS OF PURULENT VAGINAL DISCHARGE

liams et al., 2007). One crucial aspect explaining the dif- ing costs, subsequently impacting the difference in gross
ference between cows with metritis and PVD is the time profit between cows with and without PVD. This can be
of disease occurrence after calving. Metritis can develop an issue because profit margins in milk production op-
within the first 21 d postpartum; however, the peak inci- erations are heavily influenced by feed expenses, which
dence occurs at d 6 postpartum. Therefore, when treated constitute the highest variable cost. Lastly, the specific
or recovered spontaneously, cows with metritis had a details about the type of AI (i.e., AI after estrus detection
long interval between the disease occurrence and the or TAI) used for the second and subsequent AI were not
beginning of breeding. Conversely, the interval between available in the data set used in this study. Therefore, to
the occurrence of PVD and the beginning of breeding is estimate the cost of the second and subsequent AI, we
shorter; hence, cows with PVD have less time to recover. used the average cost between AI after estrus detection
The overall prevalence of PVD was 25.5%, ranging and TAI. This can lead to a miscalculation of the repro-
from 11.8 to 34.7%. Our findings align with those of Mc- ductive management cost and, consequently, the cost of
Dougall et al. (2007), who reported a prevalence of PVD PVD. Therefore, further research is warranted to address
of 21.2%, ranging from 10.1 to 31.1%. In their study, the limitations of our study and to evaluate potential
cows were diagnosed using the Metricheck device 1 to 8 long-term effects of PVD on herd dynamics and milk
weeks postpartum before the start of the breeding season. yield.
LeBlanc et al. (2002) showed a prevalence of PVD of
17% with a range of 5 to 26%. Therefore, based on an CONCLUSION
approximate prevalence rate of 20%, an average cost of
$202/case, and a milk cow inventory of approximately Cows diagnosed with PVD had reduced milk produc-
9.4 million milk cows, the annual cost of PVD in the tion and likelihood of pregnancy as well as a higher
US dairy industry could amount to approximately 380 culling rate by 305 DIM, which led to a significant
million dollars. This calculation shows the substantial decrease in income and increased expenses. The differ-
impact of PVD on both individual cows and the industry, ence in gross profit between cows with and without PVD
which emphasizes the economic significance of effec- amounted to $202. The variability in the difference of
tively managing and mitigating this condition. gross profit ranged from $152 to $265, according to the
The robust data set with 11,051 Holstein cows from 16 stochastic Monte Carlo simulation. Milk sales by 305
states across the US gives us confidence in the accuracy DIM explained 26% of the mean difference observed in
of our findings, whereas the stochastic analysis provides gross profit. To summarize, PVD negatively impacted
an estimation of the variability of our results. This dual the economic outputs of high-producing dairy cows, pri-
approach enhances both the external and internal validity marily attributed to its association with diminished milk
of our findings, contributing to a more comprehensive yield, reduced reproductive efficiency, and compromised
understanding of the economic consequences of PVD. herd survival.

Limitations ACKNOWLEDGMENTS

A significant limitation of the study is that cows were The authors gratefully acknowledge the USDA, Wash-
followed for a 305-d lactation, which means that milk ington, DC, for providing financial support through the
yield was not accounted for after 305 DIM, and assumes National Institute of Food and Agriculture (NIFA) and
that open cows would leave at 305 DIM. Ideally, a full Agriculture and Food Research Initiative (AFRI) Trans-
lactation or modeling of subsequent lactations would lational Genomics for Improved Fertility of Animals
need to be performed. If a modeling approach was to be grant (#2013-68004), which enabled the phenotyping of
used, it would be important to account for differences study animals. We also extend our gratitude to the partic-
in milk production efficiency and the effect of the tim- ipating dairy farms for permitting the continuous weekly
ing of pregnancy. Furthermore, the residual cow value monitoring of cows. The authors declare no competing
calculation used herein is not representative of the future interests.
performance, which could be addressed using a model-
ing approach. Another limitation is that the net energy REFERENCES
for lactation for maintenance and the DMI needed for
maintenance were estimated based on the mean of BW Bartlett, P. C., J. H. Kirk, M. A. Wilke, J. B. Kaneene, and E. C.
Mather. 1986. Metritis complex in Michigan Holstein-Friesian
between the estimated BW at calving and the estimated cattle: Incidence, descriptive epidemiology and estimated economic
BW on the last day of lactation up to 305 DIM or until impact. Prev. Vet. Med. 4:235–248. https:​/​/​doi​.org/​10​.1016/​0167​
they were removed (dead/sold) from the herd, whichever -5877(86)90026​-7​.
happened first. This can lead to a misestimation of feed-
Journal of Dairy Science Vol. TBC No. TBC, TBC
 Ojeda-Rojas et al.: ECONOMICS OF PURULENT VAGINAL DISCHARGE

BEA. 2023. The U.S. Bureau of Economic Analysis. The gross domestic postpartum clinical endometritis and its impact on reproductive per-
product price index. Accessed October 6, 2022. https:​/​/​www​.bea​ formance in dairy cows. J. Dairy Sci. 85:2223–2236. https:​/​/​doi​.org/​
.gov/​d ata/​p rices​- inflation/​g dp​- price​- index​# :​~ :​t ext​=​W hat​% 20is​% 20 10​.3168/​jds​.S0022​-0302(02)74302​-6.
the​%20GDP​%20Price​,t​%20part​%20of​%20this​%20index. Lhermie, G., L. W. Tauer, and Y. T. Gröhn. 2018. The farm cost of decreas-
Bell, M. J., and D. J. Roberts. 2007. The impact of uterine infection on ing antimicrobial use in dairy production. PLoS One 13:e0194832.
a dairy cow’s performance. Theriogenology 68:1074–1079. https:​/​/​ https:​/​/​doi​.org/​10​.1371/​journal​.pone​.0194832.
doi​.org/​10​.1016/​j​.theriogenology​.2007​.08​.010 https:​/​/​doi​.org/​https:​/​ Lima, F. S., A. Vieira-Neto, J. A. Snodgrass, A. De Vries, and J. E. P.
/​doi​.org/​10​.1016/​j​.theriogenology​.2007​.08​.010. Santos. 2019. Economic comparison of systemic antimicrobial
Bonnett, B. N., R. B. Miller, S. W. Martin, W. G. Etherington, and B. therapies for metritis in dairy cows. J. Dairy Sci. 102:7345–7358.
C. Buckrell. 1991. Endometrial biopsy in Holstein-Friesian dairy https:​/​/​doi​.org/​10​.3168/​jds​.2018​-15383.
cows. II. Correlations between histological criteria. Can. J. Vet. Res. McDougall, S., R. Macaulay, and C. Compton. 2007. Association be-
55:162–167. tween endometritis diagnosis using a novel intravaginal device
Cabrera, V. E. 2014. Economics of fertility in high-yielding dairy cows and reproductive performance in dairy cattle. Anim. Reprod. Sci.
on confined TMR systems. Animal 8:211–221. https:​/​/​doi​.org/​10​ 99:9–23. https:​/​/​doi​.org/​10​.1016/​j​.anireprosci​.2006​.03​.017.
.1017/​S1751731114000512. Moreira, F., C. Orlandi, C. A. Risco, R. Mattos, F. Lopes, and W. W.
De Vries, A. 2006. Economic Value of Pregnancy in Dairy Cattle. Thatcher. 2001. Effects of presynchronization and bovine somato-
J. Dairy Sci. 89:3876–3885. https:​/​/​doi​.org/​10​.3168/​jds​.S0022​ tropin on pregnancy rates to a timed artificial insemination protocol
-0302(06)72430​-4 https:​/​/​doi​.org/​https:​/​/​doi​.org/​10​.3168/​jds​.S0022​ in lactating dairy cows. J. Dairy Sci. 84:1646–1659. https:​/​/​doi​.org/​
-0302(06)72430​-4. 10​.3168/​jds​.S0022​-0302(01)74600​-0​.
Dickson, M. J., R. L. Piersanti, R. Ramirez-Hernandez, E. B. de Oliveira, Overton, M., and J. Fetrow. 2008. Economics of postpartum uterine
J. V. Bishop, T. R. Hansen, Z. Ma, K. C. C. Jeong, J. E. P. Santos, health. Pages 39–43 in Proc. Dairy Cattle Reprod. Counc. Annu.
I. M. Sheldon, J. Block, and J. J. Bromfield. 2020. Experimentally Meet., Omaha, Nebraska.
induced endometritis impairs the developmental capacity of bovine Paiano, R. B., J. Bonilla, A. M. Moreno, and P. S. Baruselli. 2021. Clini-
oocytes. Biol. Reprod. 103:508–520. https:​/​/​doi​.org/​10​.1093/​biolre/​ cal endometritis with Trueperella pyogenes reduces reproductive
ioaa069. performance and milk production in dairy cows. Reprod. Domest.
Dubuc, J., T. F. Duffield, K. E. Leslie, J. S. Walton, and S. J. LeBlanc. Anim. 56:1536–1542. https:​/​/​doi​.org/​10​.1111/​rda​.14017.
2011. Effects of postpartum uterine diseases on milk production and Paiano, R. B., C. G. P. Gonçalves, J. P. G. Mendes, J. Bonilla, D. B.
culling in dairy cows. J. Dairy Sci. 94:1339–1346. https:​/​/​doi​.org/​10​ Birgel, and E. H. Birgel Junior.. 2019. Comparative biochemical
.3168/​jds​.2010​-3758. profiles, production and reproduction status of the post-partum dairy
Edelhoff, I. N. F., M. H. C. Pereira, J. J. Bromfield, J. L. M. Vasconce- cows with and without purulent vaginal discharge. Reprod. Domest.
los, and J. E. P. Santos. 2020. Inflammatory diseases in dairy cows: Anim. 54:1188–1194. https:​/​/​doi​.org/​10​.1111/​rda​.13496​.
Risk factors and associations with pregnancy after embryo transfer. Pérez-Báez, J., C. A. Risco, R. C. Chebel, G. C. Gomes, L. F. Greco, S.
J. Dairy Sci. 103:11970–11987. https:​/​/​doi​.org/​10​.3168/​jds​.2020​ Tao, I. M. Thompson, B. C. do Amaral, M. G. Zenobi, N. Martinez,
-19070 https:​/​/​doi​.org/​https:​/​/​doi​.org/​10​.3168/​jds​.2020​-19070. C. R. Staples, G. E. Dahl, J. A. Hernández, J. E. P. Santos, and K. N.
Galvão, K. N., P. Federico, A. De Vries, and G. M. Schuenemann. 2013. Galvão. 2019. Association of dry matter intake and energy balance
Economic comparison of reproductive programs for dairy herds us- prepartum and postpartum with health disorders postpartum: Part I.
ing estrus detection, timed artificial insemination, or a combination. Calving disorders and metritis. J. Dairy Sci. 102:9138–9150. https:​/​
J. Dairy Sci. 96:2681–2693. https:​/​/​doi​.org/​10​.3168/​jds​.2012​-5982. /​doi​.org/​10​.3168/​jds​.2018​-15878.
Galvão, K. N., L. F. Greco, J. M. Vilela, M. F. Sá Filho, and J. E. P. Pérez-Báez, J., T. V. Silva, C. A. Risco, R. C. Chebel, F. Cunha, A.
Santos. 2009. Effect of intrauterine infusion of ceftiofur on uterine DeVries, J. E. P. Santos, F. S. Lima, P. Pinedo, G. M. Schuenemann,
health and fertility in dairy cows. J. Dairy Sci. 92:1532–1542. https:​ R. C. Bicalho, R. O. Gilbert, S. Rodrigez-Zas, C. M. Seabury, G.
/​/​doi​.org/​10​.3168/​jds​.2008​-1615. Rosa, W. W. Thatcher, and K. N. Galvão. 2021. The economic cost
Gilbert, R., W. Bosu, and A. Peter. 1990. The effect of Escherichia coli of metritis in dairy herds. J. Dairy Sci. 104:3158–3168. https:​/​/​doi​
endotoxin on luteal function in Holstein heifers. Theriogenology .org/​10​.3168/​jds​.2020​-19125.
33:645–651. https:​/​/​doi​.org/​10​.1016/​0093​-691X(90)90541​-Z​. Pinedo, P., J. E. P. Santos, R. C. Chebel, K. N. Galvão, G. M. Schuen-
Giuliodori, M. J., R. P. Magnasco, D. Becu-Villalobos, I. M. Lacau-Men- emann, R. C. Bicalho, R. O. Gilbert, S. Rodriguez Zas, C. M. Sea-
gido, C. A. Risco, and R. L. de la Sota. 2013. Clinical endometritis bury, G. Rosa, and W. W. Thatcher. 2020. Early-lactation diseases
in an Argentinean herd of dairy cows: risk factors and reproductive and fertility in 2 seasons of calving across US dairy herds. J. Dairy
efficiency. J. Dairy Sci. 96:210–218. https:​/​/​doi​.org/​10​.3168/​jds​ Sci. 103:10560–10576. https:​/​/​doi​.org/​10​.3168/​jds​.2019​-17951.
.2012​-5682. Ribeiro, E. S., G. Gomes, L. F. Greco, R. L. A. Cerri, A. Vieira-Neto, P.
Husnain, A., U. Arshad, M. B. Poindexter, R. Zimpel, M. N. Marinho, L. J. Monteiro Jr., F. S. Lima, R. S. Bisinotto, W. W. Thatcher, and J.
M. C. Perdomo, P. Fan, K. C. Jeong, C. D. Nelson, I. M. Sheldon, J. E. P. Santos. 2016. Carryover effect of postpartum inflammatory dis-
J. Bromfield, and J. E. P. Santos. 2023. Induced endometritis in early eases on developmental biology and fertility in lactating dairy cows.
lactation compromises production and reproduction in dairy cows. J. J. Dairy Sci. 99:2201–2220. https:​/​/​doi​.org/​10​.3168/​jds​.2015​-10337.
Dairy Sci. 106:4198–4213. https:​/​/​doi​.org/​10​.3168/​jds​.2022​-22846. Ribeiro, E. S., F. S. Lima, L. F. Greco, R. S. Bisinotto, A. P. A. Monteiro,
Huzzey, J. M., D. M. Veira, D. M. Weary, and M. A. G. von Keyserlingk. M. Favoreto, H. Ayres, R. S. Marsola, N. Martinez, W. W. Thatcher,
2007. Prepartum behaviour and dry matter intake identify dairy cows and J. E. P. Santos. 2013. Prevalence of periparturient diseases and
at risk for metritis. J. Dairy Sci. 90:3220–3233. https:​/​/​doi​.org/​10​ effects on fertility of seasonally calving grazing dairy cows supple-
.3168/​jds​.2006​-807. mented with concentrates. J. Dairy Sci. 96:5682–5697. https:​/​/​doi​
Knief, U., and W. Forstmeier. 2021. Violating the normality assumption .org/​10​.3168/​jds​.2012​-6335.
may be the lesser of two evils. Behav. Res. 53:2576–2590. https:​/​/​ Santos, J. E. P., R. S. Bisinotto, E. S. Ribeiro, F. S. Lima, L. F. Greco, C.
doi​.org/​10​.3758/​s13428​-021​-01587​-5​. R. Staples, and W. W. Thatcher. 2010. Applying nutrition and physi-
LeBlanc, S. J. 2008. Postpartum uterine disease and dairy herd reproduc- ology to improve reproduction in dairy cattle. Soc. Reprod. Fertil.
tive performance: a review. Vet. J. 176:102–114. https:​/​/​doi​.org/​10​ Suppl. 67:387–403. https:​/​/​doi​.org/​10​.5661/​RDR​-VII​-387​.
.1016/​j​.tvjl​.2007​.12​.019. Seabury, C. M., J. L. Smith, M. L. Wilson, E. Bhattarai, J. E. P. Santos,
LeBlanc, S. J. 2023. Review: Postpartum reproductive disease and fertil- R. C. Chebel, K. N. Galvão, G. M. Schuenemann, R. C. Bicalho, R.
ity in dairy cows. Animal 17:00781. https:​/​/​doi​.org/​10​.1016/​j​.animal​ O. Gilbert, S. L. Rodriguez-Zas, G. Rosa, W. W. Thatcher, and P. J.
.2023​.100781 https:​/​/​doi​.org/​https:​/​/​doi​.org/​10​.1016/​j​.animal​.2023​ Pinedo. 2023. Genome-wide association and genomic prediction for
.100781​. a reproductive index summarizing fertility outcomes in U.S. Hol-
LeBlanc, S. J., T. F. Duffield, K. E. Leslie, K. G. Bateman, G. P. Keefe, steins. G3 (Bethesda) 13:jkad043. https:​/​/​doi​.org/​10​.1093/​g3journal/​
J. S. Walton, and W. H. Johnson. 2002. Defining and diagnosing jkad043.

Journal of Dairy Science Vol. TBC No. TBC, TBC

 Ojeda-Rojas et al.: ECONOMICS OF PURULENT VAGINAL DISCHARGE

Sheldon, I. M., S. B. Price, J. Cronin, R. O. Gilbert, and J. E. Gadsby.

2009. Mechanisms of infertility associated with clinical and sub-
clinical endometritis in high producing dairy cattle. Reprod. Domest.
Anim. 44(s3):1–9. https:​/​/​doi​.org/​10​.1111/​j​.1439​-0531​.2009​.01465​
.x.
USDA-ERS. 2023. Feed price. National feed cost production. Accessed
January 6, 2023. https:​/​/​www​.ers​.usda​.gov/​data​-products/​milk​
-cost​- of​- production​- estimates/​m ilk​- cost​- of​- production​- estimates/​
#Recent ​ % 20Milk ​ % 20Cost ​ % 20of ​ % 20Production ​ % 20Estimates​
-2016​%20Base.
USDA-NASS. 2022a. Beef Price. Accessed September 19, 2022. https:​
/​/​q uickstats​. nass​. usda​. gov/​r esults/​4 EAF312F​- 59BA​- 3504​- 820C​
-415D1D7CECFE.
USDA-NASS. 2022b. Milk Price. Accessed September 19, 2022. https:​
/ ​ / ​ q uickstats ​ . nass ​ . usda​ . gov/ ​ r esults/ ​ 7 0826CF5 ​ - A4E4 ​ - 3264 ​ - 823D​
-FD0E20C531F7.
USDA-NASS. 2022c. Labor Price. Accessed November 4, 2022. https:​
/​/​q uickstats​. nass​. usda​. gov/​r esults/​C 036B1BE​- 653C​- 392A​- 865A​
-714795334E81.
USDA-NASS. 2022d. Replacement Price. Accessed September 19,
2022. https:​/​/​quickstats​.nass​.usda​.gov/​results/​98409749​-B223​
-3ECC​-ADDD​-42984391CB2E.
Vickers, A. J. 2005. Parametric versus non-parametric statistics in the
analysis of randomized trials with non-normally distributed data.
BMC Med. Res. Methodol. 5:35. https:​/​/​doi​.org/​10​.1186/​1471​-2288​
-5​-35​.
White, S. L., G. A. Benson, S. P. Washburn, and J. T. Green Jr.. 2002.
Milk Production and Economic Measures in Confinement or Pasture
Systems Using Seasonally Calved Holstein and Jersey Cows. J. Dairy
Sci. 85:95–104. https:​/​/​doi​.org/​10​.3168/​jds​.S0022​-0302(02)74057​
-5.
Wiggans, G. R. 1985. Procedures for calculating lactation records. Fact
sheet G-1. Agricultural Research Service-USDA, Beltsville, MD.
Williams, E. J., D. P. Fischer, D. E. Noakes, G. C. England, A. Rycroft,
H. Dobson, and I. M. Sheldon. 2007. The relationship between uter-
ine pathogen growth density and ovarian function in the postpartum
dairy cow. Theriogenology 68:549–559. https:​/​/​doi​.org/​10​.1016/​j​
.theriogenology​.2007​.04​.056.
Williams, E. J., D. P. Fischer, D. U. Pfeiffer, G. C. England, D. E. No-
akes, H. Dobson, and I. M. Sheldon. 2005. Clinical evaluation of
postpartum vaginal mucus reflects uterine bacterial infection and the
immune response in cattle. Theriogenology 63:102–117. https:​/​/​doi​
.org/​10​.1016/​j​.theriogenology​.2004​.03​.017.

Journal of Dairy Science Vol. TBC No. TBC, TBC