J. Dairy Sci.

96:3621–3631
http://dx.doi.org/10.3168/jds.2012-5922
© American Dairy Science Association®, 2013.

Metritis in dairy cows: Risk factors and reproductive performance

M. J. Giuliodori,* R. P. Magnasco,† D. Becu-Villalobos,‡ I. M. Lacau-Mengido,‡ C. A. Risco,§
 HOD6RWD‫ۅ‬1
DQG5/G
*Cátedra de Fisiología, Facultad de Ciencias Veterinarias – Universidad Nacional de La Plata (FCV-UNLP), La Plata, B1900AVW, Argentina
†Estudio Magnasco, Canals, Córdoba, X2650BXC, Argentina
‡Instituto de Biología y Medicina Experimental – Consejo Nacional de Investigaciones Científicas y Técnicas (IByME – CONICET), Buenos Aires,
C1428ADN, Argentina
§Large Animal Clinical Sciences, College of Veterinary Medicine, University of Florida (LACS, CVM-UF), Gainesville 32611
#Cátedra y Servicio de Reproducción Animal, FCV-UNLP, La Plata, B1900AVW, Argentina
‫ۅ‬CONICET, Buenos Aires, C1033AA-, Argentina

ABSTRACT for pregnancy by 150 DIM (hazard rate: 0.753, 95% CI:
0.621–0.911), and took longer to get pregnant (129 vs.
The objectives of this study were to assess the risk 111 vs. 109 d, for puerperal metritis, clinical metritis,
factors for metritis, its effects on milk yield and on and healthy cows, respectively). Ceftiofur treatment
reproductive performance, and the efficacy of ceftio- was not associated with cure rate or milk yield but
fur therapy in Holstein dairy cows. Cows (n = 303) was related to increased risk for pregnancy at timed
from a commercial dairy herd in Argentina were stud- artificial insemination (AOR = 2.688, 95% CI: 0.687–
ied. Cows were scored for body condition, and blood 10.832), and for lower risk of reproductive cull (AOR =
samples were collected on d −14, 7, 21, 31, 41, and 50 0.121, 95% CI: 0.014–1.066). In conclusion, abnormal
relative to parturition. Cows having a watery, purulent, calving and negative energy balance are associated with
or brown, and fetid vaginal discharge (VD) and rectal increased risk for metritis. Metritis, especially puerperal
temperature ≤39.2°C were diagnosed as having clinical metritis, correlates with reduced milk production and
metritis, and those having a similar VD and rectal tem- poor reproductive performance. Finally, the likelihood
perature >39.2°C were diagnosed as having puerperal for having a normal VD (indicative of cure) increased
metritis. Both clinical and puerperal metritis cows were 2.6% for every day of increase in postpartum time and
randomly assigned to control (no treatment) or ceft- was 2 times higher for cows with clinical metritis than
iofur group (2.2 mg/kg × 3 consecutive days). Cure for those with puerperal metritis.
was declared if clear VD was observed at 21 d in milk Key words: dairy cow, puerperal metritis, risk factor,
(DIM). Blood samples were analyzed for nonesterified reproductive performance
fatty acids, β-hydroxybutyrate, and blood urea nitro-
gen using commercial kits, and for insulin-like growth
factor-1, insulin, and leptin by RIA. Data were ana- INTRODUCTION
lyzed with PROC MIXED, GENMOD, PHREG, and
LIFETEST from SAS (SAS Institute Inc., Cary, NC). In postpartum dairy cows, uterine diseases such as
The risk for metritis increased with dystocia, retained metritis and endometritis affect a large proportion of
fetal membranes, and dead calf [AOR (adjusted odds the population and are associated with substantial pro-
ratio) = 2.58, 95% CI: 1.189–5.559], and as prepartum ductive losses (Fourichon et al., 1999, 2000). Despite
nonesterified fatty acids levels increased (AOR = 1.001, the importance of uterine disease in the dairy system,
95% CI: 0.999–1.002). Conversely, risk decreased as case definitions were standardized only recently. Thus,
prepartum insulin-like growth factor-1 increased (AOR cows having enlarged uterus, fetid watery red-brown
= 0.65, 95% CI: 0.349–1.219). Cows having either clini- vaginal discharge (VD), fever, and signs of systemic
cal or puerperal metritis produced less milk by 90 DIM illness within 21 DIM are defined as having puerperal
than did healthy cows (2,236 ± 172 vs. 2,367 ± 77 vs. metritis, whereas cows having enlarged uterus and fetid
2,647 ± 82 kg, respectively). Cows with puerperal me- watery red-brown VD without signs of systemic illness
tritis had lower risk for pregnancy by 100 DIM (AOR within 21 DIM are defined as having clinical metritis;
= 0.189, 95% CI: 0.070–0.479) and a lower hazard rate and cows having pus in VD, 21 DIM or more, without
systemic signs of illness are defined as having clinical
endometritis (Sheldon et al., 2006). While there is no
Received July 9, 2012.
Accepted February 18, 2013. gold standard for the diagnosis of these uterine dis-
1
Corresponding author: dairydoc82@gmail.com eases, evaluation of VD is the most useful procedure,
3621
3622 GIULIODORI ET AL.

because pus in VD is correlated with pathogenic bac- Lactating dairy cows were milked 3 times a day (0400,
teria load in the uterus (LeBlanc et al., 2002; Williams 1200, and 2000 h). After a voluntary waiting period of
et al., 2005). Among risk factors for metritis are calv- 40 d, healthy cows having either a clear normal or clear
ing problems (dystocia, twins, retained placenta, and with flecks of pus vaginal discharge were estrous syn-
stillbirth; Gröhn et al., 1990; Correa et al., 1993), low chronized and received a timed artificial insemination
DMI (Urton et al., 2005; Huzzey et al., 2007), and high (TAI) 50 DIM. Pregnancy was checked by palpation
prepartum serum NEFA concentration (Dubuc et al., per rectum 35 d post-TAI. Calving history and milk
2010; Ospina et al., 2010). The consensus is that cows yield were obtained from dairy records.
with metritis require systemic antibiotic treatment be-
cause of severe illness and risk for death but criteria for Sampling
assessing treatment success are inconsistent (LeBlanc,
2008). Treatments of choice include ceftiofur (1 to 2 We estimated a sample size of 300 cows to detect
mg/kg i.m., 1/d) or procaine penicillin (21,000 IU/kg a 15-percentage-points reduction in metritis incidence
i.m., 2/d) for 3 to 5 d (Smith et al., 1998; Drillich et (from 40% in control cows to 25% in ceftiofur-treated
al., 2001; Chenault et al., 2004). The remission of fever cows; Chenault et al., 2004). The sample size estima-
is approximately 70% by 5 to 10 d after treatment, but tion was calculated with WinEpiscope 2.0, by using a
elimination of fetid VD is much lower (Drillich et al., 95% CI and 80% of power (Thrusfield et al., 2001). A
2001; Chenault et al., 2004). Therefore, there is a lack prevalence of metritis of 40% was obtained from the
of information about postpartum self-cure (because dairy farm records of the last 3 yr. Cows were scored for
most studies have used positive controls) and about body condition (5-point scale; Edmonson et al., 1989;
the efficacy of treatment for prevention of subsequent Ferguson et al., 1994) and were tail bled between 1600
related diseases or for improvement of eventual produc- and 1800 h on 14 ± 3 d prepartum, and on 7, 21, 31,
tive and reproductive performance (LeBlanc, 2008). 41, and 50 DIM. Blood samples were collected in 10-mL
The objectives of this study were to assess (1) the polystyrene vials containing 20 mg of Na2EDTA and
clinical and metabolic risk factors for metritis, (2) the kept in an ice bath during sampling. Plasma was har-
use of rectal temperature and BCS as diagnostic tools vested within 2 h after sampling and stored at −20°C
for metritis, (3) the effects of metritis on reproductive until analysis. Rectal temperature was measured on
performance, milk yield, and metabolic status, (4) the 5, 6, and 7 DIM by inserting the glass thermometer
efficacy of ceftiofur therapy on cure rate, milk produc- (Franklin, Termómetros Argentinos S.A., Buenos Aires,
tion, and reproductive performance, and (5) the self- Argentina; accuracy and resolution of 0.1°C between
cure rate during postpartum period. 32 and 43.9°C measured between 18 and 28°C) at a
45° angle approximately 8 to 10 cm into the rectum to
ensure that the bulb made full contact with the rectal
MATERIALS AND METHODS
wall. Subsequently, a sample of VD was obtained with
Animals and Treatments a gloved hand and observed by direct inspection on
5 to 7, 21, 31, and 41 DIM (Williams et al., 2005).
The study was conducted in a commercial dairy herd The following scale was used: VD-0 = normal clear
(32°49cS, 62°52cW, Argentina) with 1,600 Holstein cows discharge, VD-1 = clear discharge with pus flecks,
with a milk yield of 8,500 kg/305 d, where autumn-calv- VD-2 = mucopurulent not fetid discharge, and VD-3 =
ing cows (n = 303) were enrolled. Ambient temperature watery, purulent or brown-colored, and fetid discharge
and relative humidity were monitored at a weather sta- (adapted from Chenault et al., 2004).
tion located 16 km from the farm. Average temperature-
humidity index (THI) during the study period (March Diagnostic Criteria
to August) was 56.7 ± 0.82. Prepartum transition cows
that were within 3 wk of expected calving date were Abnormal calving included cows having dystocia
maintained on dry lots, fed a low DCAD diet, and moni- (defined in a 3-point scale as follows: 1 = no assistance,
tored for signs of calving by trained personnel through 2 = slight to moderate assistance, and 3 = extreme as-
visual observation. After calving, cows were sent for 3 sistance or veterinary assistance, retention of placenta
d to the fresh herd and kept on a dry lot. At 4 DIM, (fetal membranes not expelled by 24 h), or delivering a
healthy cows were moved to a lactating herd kept in a dead calf. Metritis was defined as cows having VD-3 on
dry lot and fed 4 times daily. Diets for both pre- and 5 to 7 DIM independently of their rectal temperature
postpartum transition cows were a TMR formulated to (Sheldon et al., 2006). Clinical metritis was defined as
meet or exceed the requirements of lactating dairy cows cows having VD-3 and a rectal temperature ≤39.2°C
according to guidelines established by the NRC (2001). on 5 to 7 DIM and puerperal metritis as cows having
Journal of Dairy Science Vol. 96 No. 6, 2013
 METRITIS AND REPRODUCTIVE PERFORMANCE 3623

VD-3 and a rectal temperature >39.2°C on 5 to 7 DIM (DS Labs, Webster, TX; Becú-Villalobos et al., 2007).
(Smith et al., 1998). Clinical endometritis was defined The minimum detectable concentration of leptin was
as cows having VD-1, 2, or 3 between 21 and 41 DIM 0.02 nM and intra- and interassay CV were 6.7 and
(Sheldon et al., 2006). Cows not becoming pregnant 9.0%, respectively. Metabolic hormone concentrations
and diagnosed with persistent clinical endometritis (pu- were reported in nanograms per milliliter.
rulent and fetid vaginal discharge) together with those
that aborted were categorized as reproductive culling. Statistical Analysis

Ceftiofur Trial Data are shown as least squares means ± standard

errors (LSM ± SE). Statistical significance was set at
Cows having metritis (VD-3 on 5–7 DIM, indepen- P < 0.05, and a trend for significance was set at P <
dently of rectal temperature) were included in the 0.10 unless otherwise stated. The cow was considered
ceftiofur trial and randomly allocated to the ceftiofur the experimental unit.
treatment group (ceftiofur hydrochloride, 2.2 mg/kg of Metritis and Reproduction. The risks for metri-
BW i.m. for 3 consecutive days following the manu- tis, cure, and reproductive outcomes (e.g., pregnancy,
facturer’s recommendation; Ceobiotic, Tecnofarm SRL, culling) were analyzed with PROC GENMOD (SAS In-
Buenos Aires, Argentina), or to the untreated control stitute, 2003) with binomial distributions and logit link
group. Cure rate was assessed by inspection of VD, as functions. The logistic model for the clinical risk of me-
noted above, on 21 DIM. Cure was declared when cows tritis included the fixed effect of parity (1 vs. ≥2), calv-
had a clear, normal VD on 21 DIM. ing (normal vs. abnormal; abnormal calving included
dystocia, retention of placenta, and dead calf), and their
Laboratory Analysis second-order interaction. The model for metabolic risk
factors included the fixed effects of BCS, metabolites
Subsamples of cows were randomly obtained and (NEFA, BHBA, and BUN), and metabolic hormones
their respective plasma samples were used for metabo- (IGF-1, insulin, and leptin) at 14 ± 3 d prepartum.
lite (cows, n = 110; samples, n = 660) and metabolic These predictor variables were included as continuous
hormone (cows, n = 50; samples, n = 300) analyses. variables. The logistic model for the risk for cure at 21
Metabolites. Blood plasma samples were ana- DIM only included the fixed effect of ceftiofur (yes vs.
lyzed for NEFA, BHBA, and BUN using commercial no). Finally, the logistic model for the effect of metritis
kits (NEFA-HR(2), Wako Chemicals, Richmond, VA; on reproductive outcomes included the fixed effect of
Ranbut RB1007, Randox Laboratories, Antrim, UK; parity (1 vs. ≥2), calving (normal vs. abnormal), metri-
and Urea Color-2, Wiener Lab, Rosario, Argentina, tis (no vs. clinical vs. puerperal), clinical endometritis
respectively). Intraassay CV were 5.7, 6.3, and 6.5%, (no vs. yes), and their second-order interactions. Mod-
respectively, and interassay CV were 7.1, 7.6, and 7.8%, eling was performed using a manual backward elimina-
respectively. Metabolite concentrations were expressed tion method with an exclusion criterion set at P > 0.2.
in micromoles per liter (μM) for NEFA and BHBA, and Calving-to-conception and calving-to-culling intervals
milligrams per deciliter for BUN. by 150 DIM were analyzed with PROC PHREG (SAS
Metabolic Hormones. Hormones were measured Institute, 2003). Cox proportional hazards regression
by RIA as previously described (Lacau-Mengido et al., models included the fixed effect of parity (1 vs. ≥2),
2000; Díaz-Torga et al., 2001; Becú-Villalobos et al., calving (normal vs. abnormal), metritis (no vs. clini-
2007). Briefly, the IGF-1 RIA was performed, after acid- cal vs. puerperal), clinical endometritis (no vs. yes),
ethanol extraction, with the IGF-1 antibody (UB2-495, and their second-order interactions. Modeling was per-
Hormone Distribution Program of the National Insti- formed using a manual backward elimination method
tute of Diabetes and Digestive and Kidney Diseases, with an exclusion criteria set at P > 0.2. Time intervals
Bethesda, MD). The minimum detectable concentration (median ± 95% CI) for days open and survival function
was 0.33 nM, and intra- and interassay CV were 7.2 and estimates were obtained from PROC LIFETEST (SAS
8.8%, respectively. Insulin RIA was performed by using Institute, 2003).
anti-bovine insulin antibody (Sigma, St. Louis, MO) Diagnostic Tools. Body condition score and rec-
and standard human insulin (Laboratorios Beta, Buenos tal temperature as diagnostic tools for metritis were
Aires, Argentina). Minimum detectable concentration analyzed with receiver operator characteristics (ROC)
was 0.08 ng/mL, and intra- and interassay CV were 6.8 curves by using Sigmaplot 10.0 (Systat, 2006). Vaginal
and 8.9%, respectively. Leptin RIA was performed by a discharge with fetid odor (VD-3) was used as a gold
double antibody method with ovine-specific antiserum standard for metritis in the ROC analysis. Sensitiv-
(Delavaud et al., 2002) and recombinant bovine leptin ity (Se) was the proportion of animals diagnosed with
Journal of Dairy Science Vol. 96 No. 6, 2013
3624 GIULIODORI ET AL.

metritis (having VD-3) that were above a given thresh- for mastitis (n = 10), lameness (n = 6), or pneumonia
old, and specificity (Sp) was the proportion of animals (n = 3) before the recheck for cure (VD evaluation)
diagnosed as without metritis (not having VD-3) that on 21 d postpartum. Therefore, 303 cows were used in
were below a given threshold (Greiner et al., 2000). the incidence rate analysis and 284 were used in the
The critical threshold was the point on the ROC curve remainder of the analyses.
that had the highest combined Se and Sp. Interpreta-
tion of critical threshold was based on the area under Incidence of and Risk Factors for Metritis
the curve. Likelihood ratio positive (LR+) was the
probability that a test result at or above the threshold Incidence of puerperal metritis was 29.7% (90/303)
would be more likely to come from an animal with me- and of clinical metritis was 9.6% (29/303). Multiparous
tritis (having VD-3). cows had lower odds for metritis than primiparous herd-
Milk Yield and Metabolic Status. Milk, metabo- mates [adjusted odds ratio (AOR) = 0.646, 95% CI:
lites and metabolic hormones were analyzed with PROC 0.371–1.125; Table 1). Cows with abnormal calving had
MIXED (SAS Institute, 2003) as repeated measures. higher odds than herdmates with normal calving (AOR
The model for milk yield included the random effect of = 2.576, 95% CI: 1.189–5.559; Table 1). High levels of
the cow and the fixed effects of time (−14 ± 3 vs. 7 vs. prepartum NEFA were associated with increased risk
21 vs. 31 vs. 41 vs. 50 d relative to parturition), parity (AOR = 1.001, 95% CI: 0.999–1.002; Table 1), whereas
(1 vs. ≥2), calving (normal vs. abnormal), metritis (no high levels of prepartum IGF-1 (AOR = 0.652, 95% CI:
vs. clinical vs. puerperal), clinical endometritis (no vs. 0.349–1.219; Table 1) were related to reduced risk for
yes), and their second-order interactions. The model metritis. Prepartum NEFA (n = 110 cows) had a ROC
for the effect of ceftiofur on milk yield included the area of 0.52 ± 0.07 (P = 0.77) and a cut-off of 431
random effect of the cow and the fixed effects of time μM (Se = 0.57, Sp = 0.63, and LR+ = 1.51), whereas
(−14 ± 3 vs. 7 vs. 21 vs. 31 vs. 41 vs. 50 d related to prepartum IGF-1 (n = 50 cows) had a ROC area of
parturition), ceftiofur (no vs. yes), and their second- 0.56 ± 0.12 (P = 0.63) and a cut-off of 286 ng/L (Se =
order interaction. The model for metabolites and meta- 0.33, Sp = 0.93, and LR+ = 4.99).
bolic hormones included the random effect of the cow
and the fixed effects of time (−14 ± 3 vs. 7 vs. 21 vs. Diagnostic Tools for Metritis
31 vs. 41 vs. 50 d related to parturition), metritis (no
vs. yes), clinical endometritis (no vs. yes), and their Rectal temperature at the time of metritis diagnosis
second-order interactions. The covariance structure (7 DIM, n = 303 cows) had a ROC area of 0.604 ±
having the smallest Akaike information criterion and 0.036 (P = 0.003) and a cut-off of 39.2°C (Se = 0.566,
Schwarz’s Bayesian criterion was used (Littell et al., Sp = 0.638, and LR+ = 1.567), whereas BCS (7 DIM,
2002). A polynomial contrast was used to test the lin- n = 303 cows) had a ROC area of 0.636 ± 0.029 (P <
ear and quadratic effects of time on response variables, 0.001) and a cut-off of 2.6 (Se = 0.504, Sp = 0.706, and
and another contrast was used to compare prepartum LR+ = 1.718).
versus postpartum values (for metabolites and meta-
bolic hormones). The effect of time was removed from Metritis and Reproduction
the models used to analyze milk yield by 90 DIM.
Vaginal Discharge. Cows diagnosed with metritis Pregnancy rate by 100 DIM was affected by metritis
were used in this analysis, and their data on vaginal (P < 0.001), with cows having puerperal metritis show-
discharges were analyzed with PROC GENMOD (SAS ing lower odds than healthy herdmates (AOR = 0.219,
Institute, 2003) by using a multinomial distribution 95% CI: 0.095–0.502), whereas cows with clinical me-
(i.e., VD-0, 1, 2, and 3) and a cumulative logit link tritis had similar odds to healthy cows (AOR = 0.984,
function. This repeated multinomial model included 95% CI: 0.376–2.577). Nonpregnancy rate by 200 DIM
the fixed effect of time (21, 31, and 41 DIM), parity was affected by metritis (P = 0.095), with puerperal
(1 vs. ≥2), calving (normal vs. abnormal), metritis metritis cows having higher odds than healthy herd-
(clinical vs. puerperal), ceftiofur (yes vs. no), and their mates (AOR = 2.096, 95% CI: 0.968–4.537) and clinical
second-order interactions. Modeling was performed us- metritis cows having similar odds to healthy herdmates
ing a manual backward elimination method with an (AOR = 0.599, 95% CI: 0.129–2.779). The hazard for
exclusion criteria set at P > 0.2. pregnancy per day at risk was affected by metritis (P
= 0.005), with cows having puerperal metritis showing
RESULTS
a lower hazard rate than cows with clinical metritis
A total of 303 cows were enrolled in the study period, or healthy herdmates (hazard rate = 0.753, 95% CI:
and 19 cows were not included because they were culled 0.621–0.911; P = 0.014). Cows with puerperal metritis
Journal of Dairy Science Vol. 96 No. 6, 2013
 METRITIS AND REPRODUCTIVE PERFORMANCE 3625
Table 1. Clinical and metabolic risk factors for metritis in Holstein dairy cows (parity, calving, and BCS, n =
303; metabolites, n = 110; hormones, n = 50)

Incidence Risk for metritis1

Adjusted
Risk factor % No./total odds ratio 95% CI P-value
Clinical risk factors
Parity 0.08
Primiparous 48.9 44/90 Referent
Multiparous 35.2 75/213 0.646 0.371–1.125
Calving 0.008
Normal 34.6 90/260 Referent
Abnormal 65.1 28/43 2.576 1.189–5.559
Metabolic risk factors2
NEFA3 — 110 1.001 0.999–1.002 0.18
IGF-14 — 50 0.652 0.349–1.219 0.14
1
Metritis = cows having watery, purulent or brown-colored, and fetid vaginal discharge 5 to 7 DIM. Parity by
calving interaction had no effect (P > 0.20).
2
Metabolic risk factors with P > 0.2 were excluded from the table (i.e., BCS, BHBA, BUN, insulin, and leptin).
3
Concentration (μM) at 14 ± 3 d prepartum.
4
Concentration (ng/mL) at 14 ± 3 d prepartum.

had longer calving to conception interval (median = respectively; Table 2, Figure 3). Metritis had no effect
141.0, 95% CI: 120.0–150.0) than cows having clinical on insulin (P = 0.85) or on leptin (P = 0.35; Table 2).
metritis (median = 120.0, 95% CI: 71.0–145.0) or being
healthy (median = 104.5, 95% CI: 94.0–121.0, log-rank Ceftiofur and Reproduction
test = 8.33, df = 2, P = 0.015; Figure 1).
Ceftiofur had no effect on cure at 21 DIM: the percent-
Metritis and Milk Yield age of cured cows was similar for treated and untreated
control cows [14% (8/56) and 16% (7/44), respectively;
Time by parity affected milk yield (P < 0.001), with P > 0.10]. Ceftiofur-treated cows had greater odds for
multiparous cows producing more milk than primipa-
rous herdmates from early to mid lactation and less
milk than primiparous cows at the end of lactation.
Time by metritis also affected milk yield (P < 0.001,
Figure 2), given that cows having (clinical and puerper-
al) metritis produced less milk than healthy herdmates
during early lactation, whereas they produced more
milk than healthy cows during late lactation (Figure 2).
The remaining second-order interactions had no effect
(P > 0.1).
Multiparous cows produced more milk by 90 DIM
than primiparous herdmates (LSM ± SE: 2,640.97 ±
77.36 vs. 2,191.94 ± 88.19 kg; P < 0.001). The healthy
cows had higher production by 90 DIM than both the
clinical metritis and puerperal metritis cows (2,646.56
± 82.10 vs. 2,235.62 ± 172.11 vs. 2,367.20 ± 77.45 kg,
respectively; P = 0.009). Interactions had no effect on
milk yield by 90 DIM (P > 0.10).
Figure 1. Survival function curves of calving to conception inter-
vals by 150 DIM in Holstein dairy cows (n = 303). Clinical metritis was
defined as cows having watery, purulent, or brown, and fetid vaginal
Metritis and Metabolic Status discharge with a rectal temperature <39.2°C; puerperal metritis was
defined as cows having watery, purulent, or brown, and fetid vaginal
Cows with metritis had lower BCS than healthy discharge with a rectal temperature ≥39.2°C at 5 to 7 DIM. Cows with
herdmates (2.58 ± 0.02 vs. 2.67 ± 0.02; P = 0.004). puerperal metritis had longer calving to conception intervals by 150 d
(median = 141.0, 95% CI: 120.0–150.0) than cows with clinical metri-
Time by metritis had an effect on NEFA, BHBA, BUN, tis (median = 120.0, 71.0 to 145.0) and healthy herdmates (median =
and IGF-1 levels (P = 0.022, 0.009, 0.010, and 0.007, 104.5, 95% CI: 94.0 to 121.0, log-rank test = 8.33, df = 2; P = 0.015).

Journal of Dairy Science Vol. 96 No. 6, 2013

3626 GIULIODORI ET AL.

Vaginal Discharge

The evolution of VD during postpartum period is

shown in Figure 4. The likelihood for having a clear,
normal VD increased with time during postpartum
(AOR = 1.114, 95% CI = 1.084–1.145; P < 0.001).
Hence, the risk for cure increased 11% per day during
postpartum [(AOR – 1) × 100 was used to transform
AOR into percentage; Allison, 1999]. As the AOR for
time (i.e., day) seemed to be misleading (i.e., 11%
increase in the likelihood for normal VD per 1-d of in-
crease in time during postpartum), the following equa-
tion: β × pi (1− pi) was used to address this concern
(Allison, 1999), where β is the regression coefficient
for day [0.1079 (95% CI = 0.0808–0.1351)] and pi is
the probability for normal VD (0.7604). The corrected
Figure 2. Mixed model effects on milk yield (kg/d) in Holstein
dairy cows with clinical metritis, puerperal metritis, or no metritis change in probability for normal VD was 0.020 (0.015
(n = 303). Clinical metritis was defined as cows having watery, puru- to 0.025) per 1-d increase during postpartum (between
lent, or brown, and fetid vaginal discharge with a rectal temperature 7 and 41 DIM). In addition, the risk for normal VD
<39.2°C; puerperal metritis was defined as cows having watery, puru-
lent, or brown, and fetid vaginal discharge with a rectal temperature was higher for cows with clinical metritis than for those
≥39.2°C at 5 to 7 DIM. There was a time by metritis effect on milk having puerperal metritis (AOR = 1.956, 95% CI:
yield (P < 0.001). *P < 0.05, **P < 0.01. 1.151–3.324; P = 0.013). The remaining predictors and
their second-order interactions had no effect (P > 0.1).

TAI and for pregnancy at TAI than control cows (AOR DISCUSSION
= 2.126, 95% CI: 0.983–4.597, P = 0.055, and AOR =
2.688, 95% CI: 0.667–10.832, P = 0.14, respectively). The objectives of the study were to assess the risk
Ceftiofur also reduced reproductive culling from 13.6% factors for metritis; the effects of metritis on milk,
(6/44) to 1.8% (1/56). Thus, treated cows showed much reproductive performance, and metabolic status; and
lower odds for culling than control cows (AOR = 0.121, finally, the efficacy of ceftiofur therapy on cure rate,
95% CI: 0.014–1.066; P = 0.057). milk, and reproductive performance. The observed
prevalence of metritis in this herd was within the range
Ceftiofur and Milk Yield of reported data (Fourichon et al., 1999, 2000; Mejía
and Lacau-Mengido, 2004).
Ceftiofur had no effect on milk yield (26.78 ± 0.75
vs. 26.41 ± 0.72 kg/d; P = 0.72, for control and treated Risk Factors
cows) or on milk produced by 90 DIM (2,287.37 ±
72.98 vs. 2,418.23 ± 72.98 kg; P = 0.21, for control and Our main findings were that primiparous cows, cows
treated cows). with abnormal calving, and cows in poor energy bal-

Table 2. Effect of metritis on the metabolic status in Holstein dairy cows

Metritis1 Contrast (P-value)

No Yes Time ×
Item No. (LSM) (LSM) SE Time2 Metritis Metritis Linear Quadratic Pre-Post3
BCS (5-point scale) 110 2.67 2.58 0.02 <0.001 0.004 0.99 <0.001 <0.001 <0.001
NEFA (μM) 110 676.14 563.21 30.52 0.06 0.02 0.02 0.27 0.02 0.003
BHBA (μM) 110 525.54 482.44 30.44 0.05 0.35 0.009 0.82 0.002 0.11
BUN (mg/dL) 110 10.60 9.54 0.55 <0.001 0.17 0.01 <0.001 0.05 <0.001
IGF-1 (ng/mL) 50 233.09 224.75 21.34 <0.001 0.80 0.07 <0.001 <0.001 <0.001
Insulin (ng/mL) 50 1.47 1.43 0.12 0.18 0.85 0.67 0.78 0.30 0.97
Leptin (ng/mL) 50 1.22 1.83 0.45 0.56 0.35 0.51 0.73 0.46 0.37
1
Metritis was diagnosed in cows having watery, purulent, or brown, and fetid vaginal discharge 5 to 7 DIM.
2
Time = days related to parturition (d −14, 6, 21, 31, 41, and 50).
3
Pre-Post = prepartum versus postpartum.

Journal of Dairy Science Vol. 96 No. 6, 2013

 METRITIS AND REPRODUCTIVE PERFORMANCE 3627

ance during prepartum (i.e., high NEFA and low IGF-

1) have increased risk for metritis. This is in agreement
with previous studies (Gröhn et al., 1990; Correa et al.,
1993; Kaneene and Miller, 1995; Dubuc et al., 2011).
Primiparous cows have greater risk for metritis than
multiparous herdmates because of a greater need for
calving assistance (Bruun et al., 2002; Bell and Rob-
erts, 2007; Ghavi Hossein-Zadeh and Ardalan, 2011).
Abnormal calving usually requires human intervention,
which increases the likelihood for bacterial contamina-
tion and uterine lesions, both leading to the develop-
ment of metritis (Kaneene and Miller 1995; Sheldon
and Dobson, 2004). In addition, our finding that high
prepartum NEFA increases the risk for metritis sup-
ports 2 recently published studies (Ospina et al., 2010;
Dubuc et al., 2011). These researchers suggested NEFA
thresholds of 300 μM (Ospina et al., 2010) and 600
μM (Dubuc et al., 2011) for predicting metritis, and
we found a cut-off lying between (prepartum NEFA of
431 μM ). We also found that high prepartum IGF-1
reduces the risk for metritis, and to our knowledge, this
is the first report for an effect of IGF-1 on the risk for
metritis. In short, our findings and those previously
reported provide clear evidence that negative energy
balance plays a key role in the development of metritis.

Diagnostic Tools

According to Swets (1988), a diagnostic test is ac-

curate if the ROC area is 0.5 to 0.7. Hence, the cut-off
values obtained in this study for rectal temperature and
BCS (ROC areas ~0.6) would be accurate diagnostic
tools for metritis. Because many other diseases, such
as mastitis, could increase rectal temperature, rectal
temperature would be useful for the diagnosis but not
diagnostic in itself. The same happens with BCS, given
that many situations can lead to low BCS. Therefore,
BCS would be useful for the diagnosis but not diagnos-
tic in itself. Our cut-off value for temperature (39.2°C)
was lower than that of 39.5°C originally reported by
Drillich et al. (2001) and recently suggested for the
case definition by Sheldon et al. (2006). Whereas the
cut-off of 39.5°C reported by Drillich et al. (2001) was
arbitrarily set, the cut-off of 39.2°C in this study was
obtained using a ROC analysis. In fact, most of previ-
ous studies have arbitrarily set the cut-off rectal tem-
perature for diagnosis of metritis in early postpartum
(Smith et al., 1998; Drillich et al., 2001, 2007; Chenault
et al., 2004; Benzaquen et al., 2007), with the excep-
Figure 3. Mixed model effects on NEFA, BHBA, BUN, and IGF-1 tion of one study where the cut-off was set at the 75%
in Holstein dairy cows (n = 110). Metritis was defined as a cow hav- quartile (Sheldon and Dobson, 2004). A recent study
ing watery, purulent, or brown, and fetid vaginal discharge 5 to 7 d by Wenz et al. (2011) reported that the rectal tem-
in milk. There was a time (day) by metritis effect on NEFA, BHBA,
BUN, and IGF-1 levels (P = 0.022, 0.009, 0.010, and 0.007, respec- perature of cows with metritis peaked on the day of
tively). *P < 0.05, **P < 0.01. diagnosis at 39.6 ± 0.1°C and was 0.8°C higher than
Journal of Dairy Science Vol. 96 No. 6, 2013
3628 GIULIODORI ET AL.

expect no effect of ambient temperature and humidity

on rectal temperature measurement made in this study.
Regarding BCS, the cut-off of 2.6 reported here is,
to our knowledge, the first reported cut-off of BCS
for metritis. Finally, the limiting factor regarding the
usefulness of rectal temperature and BCS as diagnostic
tools for metritis is their low sensitivity (0.57 and 0.50,
respectively).

Metritis and Reproduction

Metritis, especially puerperal metritis, has a nega-

tive effect on reproductive performance (i.e., increased
calving to conception interval). To our knowledge,
this is the first study to compare the effect of clinical
versus puerperal metritis, and the first to show that
puerperal metritis has a much greater negative effect on
Figure 4. Cure rate between 21 and 41 postpartum days in Holstein
dairy cows with metritis (n = 119). Metritis was defined as a cow hav-
production outcomes than clinical metritis. Fourichon
ing watery, purulent, or brown, and fetid vaginal discharge at 5 to 7 et al. (2000) reviewed some studies and found impaired
DIM. VD-0 = normal, clear vaginal discharge; VD-1 = clear vaginal reproduction in cows with metritis. The negative effects
discharge with flecks of pus; VD-2 = muco-purulent non-fetid vaginal
discharge; and VD-3 = watery, purulent, or brown, and fetid vaginal
of metritis could affect fertility in many ways: first,
discharge. A multinomial cumulative logit model showed that time through a delay in the return to cyclicity after parturi-
(postpartum day) had an effect on the likelihood for having a VD-0 (P tion (Mateus et al., 2002; Sheldon et al., 2002); second,
< 0.001), and that the probability for VD-0 increased 2.0% (1.5 to 2.5)
per every 1 d of increase during postpartum.
through a clear disruption of the uterine environment
(BonDurant, 1999; Sheldon and Dobson, 2004); and
third, through an impairment of embryo development
(Soto et al., 2003).
that of parity-matched healthy cows during the first
10 DIM. Although only 2.5% of normal cows compared Metritis and Milk Yield
with 34% of cows with mastitis, metritis, or pneumonia
had rectal temperature >39.5°C (mean + 2 SD; Wenz Clinical and puerperal metritis both have negative
et al., 2011), no reference was made to the percent- effects on peak milk yield. This is in agreement with
age of cows with mastitis, metritis, or pneumonia that previous reports that found lower milk production dur-
had rectal temperature ≤39.5°C or to changes in this ing early lactation in cows with metritis (Rajala and
percentage when the cut-off was increased or decreased. Grohn, 1998; Bell and Roberts, 2007; Dubuc et al.,
When the cut-off point is increased, the number of cows 2011; Wittrock et al., 2011). One reason for this ef-
with fever that are actually healthy is reduced (type I fect could be that reduced feed intake in affected cows
error), but the number of cows with no fever that are decreases the energy available for milk synthesis (Bell
actually diseased is increased (type II error; Kristula and Roberts, 2007; Wittrock et al., 2011). In fact, that
et al., 2001). Using ROC curves allows us to maximize reduced intake has been detected even before parturi-
sensitivity and specificity for a given cut-off of rectal tion (Urton et al., 2005; Huzzey et al., 2007).
temperature, avoiding treating too many healthy ani-
mals or not treating too many diseased animals. Metritis and Metabolic Status
Several factors such as operator, thermometer, pen-
etration depth, and defecation may affect repeatability In the current study, cows with metritis had lower
of rectal temperature measurement in dairy cows in BCS than healthy herdmates and we found a metritis
early postpartum (Burfeind et al., 2010). In addi- by time interaction on NEFA and BHBA concentra-
tion, Wenz et al. (2011) reported diurnal and seasonal tions. Cows with metritis have higher BHBA concen-
variations in rectal temperature. The present study was trations than healthy cows immediately after parturi-
done between May and August (fall and winter in the tion, but later on in lactation (31 to 41 DIM) they have
Southern Hemisphere). During that time of the year, lower levels than normal herdmates. Therefore, cows
the average THI was 56.5 ± 0.85. A reduction in milk with metritis would have a deeper negative energy bal-
yield as an indicator of cow discomfort starts when the ance around parturition, probably caused by low DMI
THI is >72 (Igono et al., 1992); therefore, we would prepartum (Urton et al., 2005; Huzzey et al., 2007).
Journal of Dairy Science Vol. 96 No. 6, 2013
 METRITIS AND REPRODUCTIVE PERFORMANCE 3629

Then, as lactation progresses, healthy cows would have Benzaquen et al., 2007), whereas others have evaluated
a deeper negative energy balance than cows with me- a 5-d treatment (Smith et al., 1998; Chenault et al.,
tritis because of the heavier demands imposed by the 2004). Regarding the use of 3-d protocols, Zhou et al.
udder to support their higher milk production in early (2001) reported that ceftiofur-treated cows (1 mg/kg)
lactation. had higher cure rates (defined as remission of fever and
absence of vaginal discharge) than untreated control
Ceftiofur cows (56 vs. 29%, respectively). Drillich et al. (2001)
compared a 3-d therapy with 1.1 mg/kg ceftiofur with
Ceftiofur did not affect the cure of metritis or peak an ampicillin plus cloxacillin intrauterine treatment
milk production. Ceftiofur increased the risk for preg- and found similar cure rates (defined as remission of
nancy at TAI and reduced the risk for reproductive fever, ~80%). Drillich et al. (2007) also found similar
culling. The usefulness of ceftiofur therapy for the cure rates (defined as remission of fever, ~89%) in
treatment of metritis has been studied (Smith et al., cows receiving 1.1 mg/kg ceftiofur for 3 d (which was
1998; Drillich et al., 2001; Zhou et al., 2001; Chenault extended for 2 extra days if fever was present on treat-
et al., 2004). Some works included positive controls ment d 3) alone or in combination with flunixin meglu-
(Smith et al., 1998; Drillich et al., 2001), whereas oth- mine. Regarding the use of 5-d protocols, Smith et al.
ers used negative controls (Zhou et al., 2001; Chenault (1998) found that 2 mg/kg ceftiofur was as effective as
et al., 2004). In addition, most studies used body tem- procaine penicillin G or procaine penicillin G plus in-
perature (i.e., <39.5°C) to define a successful therapy; trauterine infusion of oxytetracycline for the treatment
that is, when temperature returned to within normal of metritis. Chenault et al. (2004) studied the effect of
range, animals were declared cured (Smith et al., 1998; a 5-d therapy with ceftiofur at 2 doses (2.2 and 1.1 mg/
Drillich et al., 2001; Zhou et al., 2001) even though kg) with an untreated control and found cure rates (de-
many of them were still having a fetid VD (Drillich fined as remission of fever and absence of a fetid vaginal
et al., 2001). As a result, rectal temperature as the discharge) of 77, 65, and 62%, respectively, 14 d after
only criterion for cure may overestimate the response initial treatment. Unfortunately, to our knowledge, no
to therapy (Chenault et al., 2004). Considering normal published clinical trial has compared the effects of a
temperature and non-fetid VD as criteria for cure, the 3-d versus 5-d therapy with ceftiofur for the treatment
reported cure rates (14 d after initial treatment) are of metritis. Therefore, the lack of response to ceftiofur
62% for control cows, 65% for cows treated with 1.1 mg treatment in the current study might be explained in
of ceftiofur/kg of BW, and 77% for those receiving 2.2 part by the treatment duration (3 d instead of 5 d).
mg of ceftiofur/kg of BW (Chenault et al., 2004). In our
cows, 14 to 16 d after initial treatment, 70% of control Self-Cure Rate
and 68% of ceftiofur-treated cows had a non-fetid VD.
Therefore, our results were similar to those of Chenault We found that the likelihood of having a normal
et al. (2004). Even more importantly, the efficacy of VD (indicative of cure rate of metritis) increased 2.6%
the ceftiofur treatment, based on production interest for every day of increase postpartum. As the effect of
variables such as pregnancy rate at TAI or calving to time on cure rate could be considered as an indicative
conception interval, has not previously been reported. of self-cure of metritis, this would be the first study
Only one study (using positive controls) reported no to report an objective measure of the relationship be-
effect of treatment on measures of reproductive per- tween postpartum time and cure rate in dairy cows
formance such as pregnancy rate at first AI, calving with metritis. We found that cows with clinical metritis
to conception interval, and culling rate (Drillich et al., have twice the odds for having a normal VD compared
2001). In our study, ceftiofur tended to increase the risk with those with puerperal metritis. The fact that cows
for pregnancy at TAI and greatly reduced the risk for having clinical metritis had a greater chance for cure
culling. Ceftiofur is used for the treatment of mastitis, than herdmates having puerperal metritis is expected
laminitis, and respiratory diseases (Zwald et al., 2004); on the basis that the latter is a life-threatening entity
therefore, one reason for our positive findings regarding accompanied by systemic signs of illness (i.e., fever and
this therapy could be that it minimizes the negative depression), whereas the former is restricted to the
effect of any of these pathologies on the reproductive uterus (Sheldon et al., 2006).
performance of the cows. Therefore, this is the first
study reporting the effect of ceftiofur on production CONCLUSIONS
interest variables using a negative control.
Some studies have assessed the effect of 3-d ceftiofur A deeper prepartum negative energy balance and
therapy (Drillich et al., 2001, 2007; Zhou et al., 2001; calving problems were associated with an increased risk
Journal of Dairy Science Vol. 96 No. 6, 2013
3630 GIULIODORI ET AL.

for metritis. Clinical metritis and puerperal metritis Dubuc, J., T. F. Duffield, K. E. Leslie, J. S. Walton, and S. J. LeBlanc.
2010. Risk factors for postpartum uterine diseases in dairy cows.
were associated with lower milk yield in early lactation. J. Dairy Sci. 93:5764–5771.
Puerperal metritis was related to lower early pregnancy Dubuc, J., T. F. Duffield, K. E. Leslie, J. S. Walton, and S. J. LeBlanc.
rate and extended calving to conception interval. A 3-d 2011. Effects of postpartum uterine diseases on milk production
and culling in dairy cows. J. Dairy Sci. 94:1339–1346.
therapy with ceftiofur was not associated with cure rate Edmonson, A. J., I. J. Lean, L. D. Weaver, T. Farve, and G. Webster.
but was related to increased pregnancy rate at TAI 1989. A body condition scoring chart for Holstein dairy cows. J.
and reduced risk for reproductive culling. Cure rate of Dairy Sci. 72:68–78.
Ferguson, J. D., D. T. Galligan, and N. Thomsen. 1994. Principal
metritis increased 2.6% per day of increase during the descriptors of body condition score in Holstein cows. J. Dairy Sci.
postpartum period and was double when the metritic 77:2695–2703.
cow was not febrile. Fourichon, C., H. Seegers, N. Bareille, and F. Beaudeau. 1999. Effects
of disease on milk production in the dairy cow: A review. Prev.
Vet. Med. 41:1–35.
ACKNOWLEDGMENTS Fourichon, C., H. Seegers, and X. Malher. 2000. Effect of disease on
reproduction in the dairy cow: A meta-analysis. Theriogenology
53:1729–1759.
This study was financed by a grant to R.L.S. Greiner, M., D. Pfeiffer, and R. D. Smith. 2000. Principles and practi-
(V11/134, Proyecto de Incentivos, Universidad Nacio- cal application of the receiver-operating characteristic analysis for
nal de La Plata). The authors thank the farm owner diagnostic tests. Prev. Vet. Med. 45:23–41.
Gröhn, Y. T., H. N. Erb, C. E. McCulloch, and H. S. Saloniemi. 1990.
and personnel for their support. Epidemiology of reproductive disorders in dairy cattle: Associa-
tions among host characteristics, disease and production. Prev.
Vet. Med. 8:25–39.
REFERENCES Ghavi Hossein-Zadeh, N., and M. Ardalan. 2011. Cow-specific risk fac-
tors for retained placenta, metritis and clinical mastitis in Holstein
Allison, P. D. 1999. Logistic Regression Using SAS: Theory and Ap-
cows. Vet. Res. Commun. 35:345–354.
plication. SAS Institute Inc., Cary, NC.
Huzzey, J. M., D. M. Veira, D. M. Weary, and M. A. G. v. Keyserlingk.
Becú-Villalobos, D., I. García-Tornadú, G. Shroeder, E. E. Salado, G.
2007. Prepartum behavior and dry matter intake identify dairy
Gagliostro, C. Delavaud, Y. Chilliard, and I. M. Lacau-Mengido.
cows at risk for metritis. J. Dairy Sci. 90:3220–3233.
2007. Effect of fat supplementation on leptin, insulin-like growth
Igono, M. O., G. Bjotvedt, and H. T. Sanford-Crane. 1992. Environ-
factor I, growth hormone, and insulin in cattle. Can. J. Vet. Res.
mental profile and critical temperature effects on milk production
71:218–225.
of Holstein cows in desert climate. Int. J. Biometeorol. 36:77–87.
Bell, M. J., and D. J. Roberts. 2007. The impact of uterine infection on
Kaneene, J. B., and R. Miller. 1995. Risk factors for metritis in Michi-
a dairy cow’s performance. Theriogenology 68:1074–1079.
gan dairy cattle using herd- and cow-based modelling approaches.
Benzaquen, M. E., C. A. Risco, L. F. Archbald, P. Melendez, M. J.
Prev. Vet. Med. 23:183–200.
Thatcher, and W. W. Thatcher. 2007. Rectal temperature, calv-
Kristula, M., B. I. Smith, and A. Simeone. 2001. The use of daily
ing-related factors, and the incidence of puerperal metritis in post-
postpartum rectal temperatures to select dairy cows for treatment
partum dairy cows. J. Dairy Sci. 90:2804–2814.
with systemic antibiotics. Bovine Pract. 35:117–125.
BonDurant, R. H. 1999. Inflammation in the bovine female reproduc-
Lacau-Mengido, I. M., M. E. Mejıa, G. S. Díaz-Torga, A. Gonzalez-
tive tract. J. Anim. Sci. 77:101–110.
Iglesias, N. Formia, C. Libertun, and D. Becu-Villalobos. 2000. En-
Bruun, J., A. K. Ersbøll, and L. Alban. 2002. Risk factors for metritis
docrine studies in ivermectin-treated heifers from birth to puberty.
in Danish dairy cows. Prev. Vet. Med. 54:179–190.
J. Anim. Sci. 78:817–824.
Burfeind, O., M. A. G. von Keyserlingk, D. M. Weary, D. M. Veira,
LeBlanc, S. J. 2008. Postpartum uterine disease and dairy herd repro-
and W. Heuwieser. 2010. Short communication: Repeatability
ductive performance: A review. Vet. J. 176:102–114.
of measures of rectal temperature in dairy cows. J. Dairy Sci.
Littell, R. C., W. W. Stroup, and R. J. Freund. 2002. SAS for Linear
93:624–627.
Models. 4th ed. SAS Institute Inc., Cary, NC.
Chenault, J. R., J. F. McAllister, S. T. Chester, K. J. Dame, F. M.
Mateus, L., L. L. da Costa, F. Bernardo, and J. R. Silva. 2002. Influ-
Kausche, and E. J. Robb. 2004. Efficacy of ceftiofur hydrochloride
ence of puerperal uterine infection on uterine involution and post-
sterile suspension administered parenterally for the treatment of
partum ovarian activity in dairy cows. Reprod. Domest. Anim.
acute postpartum metritis in dairy cows. J. Am. Vet. Med. Assoc.
37:31–35.
224:1634–1639.
Mejía, M. E., and I. M. Lacau-Mengido. 2004. Endometritis treatment
Correa, M. T., H. Erb, and J. Scarlett. 1993. Path analysis for seven
with a PGF2α analog does not improve reproductive performance
postpartum disorders of Holstein cows. J. Dairy Sci. 76:1305–
in a large dairy herd in Argentina. Theriogenology 63:1266–1276.
1312.
NRC. 2001. Nutrient Requirements of Dairy Cattle. 7th rev. ed. Natl.
Delavaud, C., A. Ferlay, Y. Faulconnier, F. Bocquier, G. Kann, and
Acad. Sci., Washington, DC.
Y. Chilliard. 2002. Plasma leptin concentration in adult cattle: Ef-
Ospina, P. A., D. V. Nydam, T. Stokol, and T. R. Overton. 2010. Eval-
fects of breed, adiposity, feeding level, and meal intake. J. Anim.
uation of nonesterified fatty acids and β-hydroxybutyrate in tran-
Sci. 80:1317–1328.
sition dairy cattle in the northeast United States: Critical thresh-
Díaz-Torga, G. S., M. E. Mejía, A. González-Iglesias, N. Formía, D.
olds for prediction of clinical diseases. J. Dairy Sci. 93:546–554.
Becú-Villalobos, and I. M. Lacau-Mengido. 2001. Metabolic cues
Rajala, P. J., and Y. T. Grohn. 1998. Effects of dystocia, retained
for puberty onset in free grazing Holstein heifers naturally infected
placenta, and metritis on milk yield in dairy cows. J. Dairy Sci.
with nematodes. Theriogenology 56:111–122.
81:3172–3181.
Drillich, M., O. Beetz, A. Pfutzner, M. Sabin, H. J. Sabin, and P.
SAS Institute. 2003. SAS/STAT Software for Windows 9.1. SAS Insti-
Kutzer. 2001. Evaluation of a systemic antibiotic treatment of
tute Inc., Cary, NC.
toxic puerperal metritis in dairy reproductive cows. J. Dairy Sci.
Sheldon, I. M., and H. Dobson. 2004. Postpartum uterine health in
84:2010–2017.
cattle. Anim. Reprod. Sci. 82–83:295–306.
Drillich, M., D. Voigt, D. Forderung, and W. Heuwieser. 2007. Treat-
Sheldon, I. M., G. S. Lewis, S. LeBlanc, and R. O. Gilbert. 2006.
ment of acute puerperal metritis with flunixin meglumine in addi-
Defining postpartum uterine disease in cattle. Theriogenology
tion to antibiotic treatment. J. Dairy Sci. 90:3758–3763.
65:1516–1530.

Journal of Dairy Science Vol. 96 No. 6, 2013

 METRITIS AND REPRODUCTIVE PERFORMANCE 3631
Sheldon, I. M., D. E. Noakes, A. N. Rycroft, D. U. Pfeiffer, and H. Wenz, J. R., D. A. Moore, and R. Kasimanickam. 2011. Factors associ-
Dobson. 2002. Influence of uterine bacterial contamination af- ated with the rectal temperature of Holstein dairy cows during the
ter parturition on ovarian dominant follicle selection and follicle first 10 days in milk. J. Dairy Sci. 94:1864–1872.
growth and function in cattle. Reproduction 123:837–845. Williams, E. J., D. P. Fischer, D. U. Pfeiffer, G. C. England, D. E.
Smith, B. I., G. A. Donovan, C. A. Risco, R. Littell, C. Young, and Noakes, H. Dobson, and I. M. Sheldon. 2005. Clinical evaluation of
L. H. Stanker. 1998. Comparison of various antibiotic treatments postpartum vaginal mucus reflects uterine bacterial infection and
for cows diagnosed with toxic puerperal metritis. J. Dairy Sci. the inflammatory response to endometritis in cattle. Theriogenol-
81:1555–1562. ogy 63:102–117.
Soto, P., R. P. Natzke, and P. J. Hansen. 2003. Actions of tumor ne- Wittrock, J. M., K. L. Proudfoot, D. M. Weary, and M. A. G. von
crosis factor-α on oocyte maturation and embryonic development Keyserlingk. 2011. Metritis affects milk production and cull rate
in cattle. Am. J. Reprod. Immunol. 50:380–388. of Holstein multiparous and primiparous dairy cows differently. J.
Swets, J. A. 1988. Measuring the accuracy of diagnostic systems. Sci- Dairy Sci. 94:2408–2412.
ence 240:1285–1293. Zhou, C., J. F. Boucher, K. J. Dame, M. Moreira, R. Graham, J. Nan-
Systat. 2006. Sigmaplot User’s Guide, version 10.0. Systat Software tel, S. Zuidorf, L. Arfi, R. Flores, G. Neubauer, and J. Olson. 2001.
Inc., Chicago, IL. Multilocation trial of ceftiofur for treatment of postpartum cows
Thrusfield, M., C. Ortega, I. de Blas, J. P. Noordhuizen, and K. Fran- with fever. J. Am. Vet. Med. Assoc. 219:805–808.
kena. 2001. Win EPISCOPE 2.0: Improved epidemiological soft- Zwald, A. G., P. L. Ruegg, J. B. Kaneene, L. D. Warnick, S. J. Wells,
ware for veterinary medicine. Vet. Rec. 148:567–572. C. Fossler, and L. W. Halbert. 2004. Management practices and
Urton, G., M. A. G. v. Keyserlingk, and D. M. Weary. 2005. Feeding reported antimicrobial usage on conventional and organic dairy
behavior identifies dairy cows at risk for metritis. J. Dairy Sci. farms. J. Dairy Sci. 87:191–201.
88:2843–2849.

Journal of Dairy Science Vol. 96 No. 6, 2013