Professional Documents
Culture Documents
Veterinary Microbiology
journal homepage: www.elsevier.com/locate/vetmic
A R T I C L E I N F O A B S T R A C T
Keywords: Infection of the heifer mammary gland with common mastitis pathogens, particularly
Staphylococcus aureus staphylococci, prior to calving is well documented. Efforts to eliminate pre-partum
Coagulase-negative staphylococci intramammary infections (IMI) in heifers have focused primarily on intramammary
Mastitis antibiotic therapy shortly before or at the time of calving. Few studies have evaluated
Vaccination
vaccination of heifers against staphylococcal mastitis. The objectives of the present study
Humoral immunity
were to evaluate the efficacy of a commercially available Staphylococcus aureus bacterin in
protecting against staphylococcal IMI (S. aureus and coagulase-negative staphylococci
(CNS)), to study the effect of vaccination on milk SCC, and to evaluate the milk antibody
isotype response to vaccination using a lactating cow model. Ninety Holstein–Friesian
lactating dairy cows of various parities were systematically assigned to a vaccinated
(n = 44) or control (n = 46) group. Vaccinates received two 5 ml doses of the bacterin 14
days apart starting on day 0. Quarter milk samples for bacterial culture were collected
prior to each vaccination and approximately monthly thereafter for 6 months. Composite
milk samples were collected on days 0, 14, 28, 49 and 70 for IgA, IgG1, IgG2, and IgM
determinations and somatic cell count. No animals in either group developed a new S.
aureus IMI after vaccination. The numbers of mammary quarters that developed a new CNS
IMI, time to new CNS IMI, milk somatic cell count, and milk antibody isotype sample-to-
positive ratio did not significantly differ between groups (P > 0.05). In a herd with a 3%
prevalence of S. aureus IMI and a 30% prevalence of CNS IMI, the vaccine did not reduce the
new staphylococcal IMI rate. There may be insufficient vaccine-induced opsonizing
antibody in milk to facilitate phagocytosis and clearance of staphylococci from the
mammary gland.
ß 2008 Elsevier B.V. All rights reserved.
0378-1135/$ – see front matter ß 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.vetmic.2008.09.053
J.R. Middleton et al. / Veterinary Microbiology 134 (2009) 192–198 193
A common practice in the control of many infectious based on expected calving date. However, because the
diseases is vaccination. In very few studies vaccination of heifers were bull bred, time from vaccination to calving
heifers to prevent mastitis either before or after calving has varied. Hence, time from vaccination to intramammary
been evaluated. Commercially available R-mutant Gram- challenge with S. aureus and time to characterization of the
negative mastitis bacterins are widely used on dairy farms to immune response in milk varied between heifers. To
significantly reduce the severity of clinical mastitis in the eliminate this variability, the present study was performed
immediate post-partum period. However, research specifi- in lactating cattle so that the sampling intervals were the
cally focused on the efficacy of these vaccines in heifers is same in all cattle. The objectives of the study were to
lacking. Furthermore, heifer mammary glands are most evaluate (1) the efficacy of a Lysigin in protecting against
commonly infected with staphylococcal organisms around staphylococcal mastitis (S. aureus and CNS), (2) the effect of
the time of calving (Fox et al., 1995). A commercially vaccination on milk SCC, and (3) the milk antibody isotype
available Staphylococcus aureus mastitis bacterin (Lysigin, response to vaccination when administered at the labeled
Boehringer Ingelheim Vetmedica, Inc.) has been evaluated dose interval in a lactating cow model.
in heifers (Nickerson et al., 1999; Middleton et al., 2006;
Luby et al., 2007). In one study, the efficacy of commercially
2. Materials and methods
available Lysigin was compared with two experimental
formulations and unvaccinated controls in primiparous 2.1. Animals and treatments
heifers (Middleton et al., 2006). Heifers were vaccinated
twice, 28 days apart in late gestation with either a 3-isolate The study was conducted in the University of Missouri
experimental bacterin (Group I; n = 11), a 5-isolate experi- Foremost dairy research facility that milks approximately
mental bacterin (Group II; n = 11), or commercially available 180 cattle. The project was approved by the institutional
Lysigin (Group III; n = 14). Group IV consisted of 11 animal care and use committee. Approximately 3% of
unvaccinated control cattle. All groups (vaccinates and mammary quarters (7% of cows) were infected with S.
controls) were challenged with a heterologous strain of S. aureus and 30% of mammary quarters (58% of cows) were
aureus by intramammary infusion on days 6–8 of lactation in infected with CNS at the beginning of the study. Ninety
a single infection-free mammary quarter. All cattle became lactating Holstein–Friesian cattle free of S. aureus IMI were
infected with S. aureus after challenge. While three cattle in systematically assigned to one of the two groups based on
Group I, and one cow in Group III, cleared their S. aureus IMI ear tag number such that even-numbered cattle were
by the end of the study, there were no differences in S. aureus vaccinated (n = 44) and odd-numbered cattle served as
clearance rates between groups. Cattle vaccinated with unvaccinated controls (n = 46). Cattle were considered free
Lysigin had a lower mean duration of clinical mastitis and of S. aureus IMI based on the absence of S. aureus on three
lower total mastitis score post-challenge than controls. sequential weekly mammary quarter foremilk cultures,
Overall, there was no evidence that any of the vaccinated and a negative result on the S. aureus Antibody Test Kit
groups had a lower mean somatic cell count (SCC) than (SAATK, VMRD, Inc.) performed according to the manu-
control, and no evidence that vaccinates had greater milk facturer’s instructions. All study animals were at least 30
yield than controls post-challenge. Heifers vaccinated with days in milk and producing greater than 14 kg (30 lbs) milk
Lysigin had higher mean serum IgG1 and IgG2 sample-to- per day during the first 70 days of the study. Vaccinates
positive (S:P) ratios than controls against S. aureus strains of received two doses of Lysigin 14 days apart (study days 0
polysaccharide serotypes 5, 8, and 336. Milk total anti-S. and 14) subcutaneously in the neck. Control cattle were
aureus IgG S:P ratios were only different from controls not vaccinated. Composite samples containing milk from
against strains of serotype 8 and 336, and there were no all quarters of each cow were collected via an in-line
differences between groups for milk IgG1, IgG2, and IgM sampler (Westphalia Surge, Inc.) during the morning
(Luby et al., 2007). These data suggest that there may be milking on days 0, 14, 28, 49 and 70 for antibody
insufficient vaccine-induced opsonising antibody in milk to determinations and SCC enumeration. Quarter foremilk
facilitate clearance of S. aureus from the mammary gland. In samples were aseptically collected using standard techni-
contrast to our work, Nickerson et al. (1999) vaccinated ques (Hogan et al., 1999) immediately prior to each
heifers with commercially available Lysigin at 6 months of vaccination and at approximately 1-month intervals
age followed by a booster dose 2 weeks later and subsequent thereafter for 6 months for bacterial culture. Milk samples
vaccinations every 6 months until calving. Vaccinates had a for culture and antibody determinations were stored at
45% reduction in both new S. aureus IMI during pregnancy 20 8C without preservative until they were analyzed.
and new S. aureus IMI at calving relative to controls. In Milk samples for SCC enumeration were collected into
addition, vaccinates had a 30% reduction in new CNS IMIs non-sterile vials containing a 2-bromo-2-nitropropane-
which became chronic and a 31% reduction in new CNS IMI 1,3-diol preservative tablet (D & F Control Systems, Inc.),
at calving relative to controls, thus providing evidence that stored at 4 8C, and shipped overnight to an external
Lysigin may be of use in reducing staphylococcal mastitis in laboratory (Mid-South Dairy Records) within 72 h of
periparturient heifers (Nickerson et al., 1999). sample collection.
In order to further examine the efficacy of Lysigin against
S. aureus, and more specifically CNS, under field conditions, 2.2. Milk bacteriology and somatic cell count determination
we conducted a trial using a lactating dairy cow model. In
our previous studies (Middleton et al., 2006; Luby et al., Stored milk samples were thawed at room temperature
2007), primiparous heifers were vaccinated in late gestation and cultured using standard methods (Hogan et al., 1999).
194 J.R. Middleton et al. / Veterinary Microbiology 134 (2009) 192–198
Briefly, approximately 10 ml milk was applied to half a solution (1.5% sodium fluoride, pH 11.5) was added to
Columbia blood agar plate containing 5% sheep blood each well and optical densities (OD) were determined
(Remel) using a cotton-tipped swab. Plates were incu- using an automated plate reader (Thermolab Systems) at
bated at 37 8C for 24 h followed by further 24 h incubation 650 nm. Antibody S:P ratios were calculated using the
at room temperature (22 8C). Plates were examined for following formula: S:P ratio = (mean test sample
bacterial growth at 24 and 48 h. Staphylococci were OD mean negative control sample OD)/(mean positive
identified based on colony morphology, hemolytic pat- control sample OD mean negative control sample OD).
tern, catalase test, and tube coagulase test. A S. aureus IMI Each antibody isotype was run on its own plate with an
was defined as the recovery of a hemolytic coagulase- approximately even distribution of samples from vacci-
positive Staphylococcus on at least two out of three nates and controls on each plate. Each sample period
consecutive milk samples (Middleton et al., 2001). A CNS required a total of three plates per antibody isotype
IMI was defined as the recovery of CNS colonies of similar measured. Sample-to-positive ratios were calculated
morphology from a quarter not previously infected with a within plate such that mean test sample and positive
CNS in one of the following quantities: (1) >10 colonies/ and negative control ODs in the S:P ratio calculation
plate on a single culture, (2) 2–10 colonies/plate on two represented the mean of two wells tested on each plate.
out of three consecutive cultures or (3) 1 colony/plate on Hence, all assays were controlled within plate to
three consecutive cultures (adapted from Zadoks et al., eliminate plate-to-plate variation in assay results.
2001).
Milk SCC enumerations were performed by a commer- 2.4. Data analysis
cial laboratory (Mid-South Dairy Records) using an
automated counter (SomatoCount 300, Bentley Instru- Positive and negative control sample ODs for each
ments). Milk samples were tempered at 39–40 8C prior to antibody isotype were compared using a Wilcoxon Rank
analysis to facilitate homogenization of the sample. Sum test. Baseline data (milk antibody isotype S:P ratios
and SCC) collected prior to first vaccination were
2.3. Detection of different antibody isotypes investigated as a covariate for post-vaccination data
using the Kruskal–Wallis one-way analysis of variance
Antibody isotype (IgA, IgG1, IgG2, and IgM) S:P ratios (ANOVA). No differences were noted between groups for
against S. aureus in milk were determined using an any antibody isotype (P 0.5) or for composite milk SCC
antibody-capture ELISA on composite milk samples (P = 0.12) at baseline. Therefore, repeated measures
collected on days 0, 14, 28, 49, and 70. The ELISA was ANOVAs were performed on data collected after first
adapted from a commercially available ELISA (SAATK, vaccination, as per Ludbrook (1994). Several attempts
VMRD, Inc.). Microtiter plates coated with partially were made to transform the data to meet the assumption
purified exoprotein antigens from the Wood 46 strain of normality, but all attempts were unsuccessful. There-
of S. aureus were obtained from the manufacturer fore, since all data following baseline in its raw form met
(VMRD, Inc.). Assay positive control samples differed the equality of variances assumption, raw data were used
according to antibody isotype being detected. For the in the analyses, as per Ludbrook (1994). The statistical
milk IgG1 and IgG2 assays, the positive control was the model included group, time and the interaction of group
SAATK positive control. For the milk IgA and IgM assays, and time as fixed effects, and cow was nested within
the positive control was milk from a cow with a S. aureus group. The dependent variables were antibody isotype
IMI which had been experimentally infected with S. (IgA, IgG1, IgG2, and IgM) S:P ratio and milk SCC.
aureus. Milk from this cow had a significantly higher Regardless of F-test result, pair-wise comparisons of
optical density when tested for IgA and IgM than the means were conducted to test for improvement of
negative control sample (P < 0.001). Positive controls vaccinates over control. Separate analyses were per-
differed between assays due to differing reactivity for formed for each antibody isotype.
each antibody isotype ELISA. For all assays, the negative Incidence of new staphylococcal IMI during the
control was the SAATK negative control. All samples follow-up period was calculated as the number of new
were run in duplicate. Fifty microliters of each control or infections per quarter month at risk. A quarter was
test sample were added to each well and incubated for considered at risk for a new staphylococcal IMI if it was
20 min at room temperature (22 8C). All milk samples not infected with staphylococci at the start of the study
were run undiluted. Following incubation, each well was and was considered at risk until it became infected.
manually washed four times with 250 ml of the SAATK Staphylococcal IMI definitions are stated above in the
wash solution (0.01 M potassium phosphate, 0.35 M milk bacteriology section. The proportions of quarters
sodium chloride, 0.065% Tween 20 and 0.02% Thimer- which developed a new staphylococcal IMI were com-
osal). Subsequently, 50 ml of a 1:1000 dilution in pared between groups using the Chi-square test. Time to
phosphate buffered saline (pH 7.2) of either sheep new staphylococcal IMI was compared between groups
anti-bovine IgA, IgG1, IgG2, or IgM horseradish perox- using a Kaplan–Meier survival analysis on quarters at risk
idase conjugate were added to each well and incubated of new staphylococcal IMI at the time of vaccination. A
and washed as before. Fifty microliters of tetramethyl- quarter was censored from the data if it ceased to lactate
benzidine substrate (BioFx) were then added to each well during the follow-up period. Data for S. aureus and CNS
and plates were incubated for 20 min at room tempera- IMIs were analyzed separately. Significance for each
ture (22 8C). Following incubation, 50 ml of SAATK stop test was P < 0.05. All analyses were performed using
J.R. Middleton et al. / Veterinary Microbiology 134 (2009) 192–198 195
3. Results
Fig. 3. Mean IgG1 S:P ratios against S. aureus for vaccinated (n = 44) and
control (n = 46) cows. Vaccinates received Lysigin on days 0 and 14; milk
Fig. 1. Mean somatic cell counts (SCC) for vaccinated (n = 44) and control samples on these days were collected prior to vaccination. Baseline data
(n = 46) cows. Vaccinates received Lysigin on days 0 and 14; milk samples (day 0) were investigated as a covariate and no differences between
on these days were collected prior to vaccination. Baseline data (day 0) groups were detected; hence only post-vaccination data were analyzed
were investigated as a covariate and no differences between groups were (days 14–70). No significant differences were detected between groups
detected; hence only post-vaccination data were analyzed (days 14–70). and no group by time interaction was detected (P = 0.85 and P = 0.81,
Mean SCC did not differ between groups (P = 0.12) or study days (P = 0.71) respectively). Mean IgG1 S:P ratios differed between sample days
and no significant group by time interaction was detected (P = 0.27). Error regardless of group (P < 0.001; denoted by different letters above
bars represent standard error of the mean. bars). Error bars represent standard error of the mean.
196 J.R. Middleton et al. / Veterinary Microbiology 134 (2009) 192–198
4. Discussion
differentiate it from a sample containing no S. aureus consistent with our challenge trial data in heifers
antibody. vaccinated twice in late gestation (Middleton et al.,
Based on the reported results, significant differences in 2006; Luby et al., 2007). Collectively, our data suggest
mean milk antibody S:P ratios between vaccinates and that there may be insufficient vaccine-induced opsonizing
controls were not detected for any of the antibody isotypes antibody in milk to facilitate phagocytosis and clearance of
(P 0.8). In contrast to the current study, Williams et al. S. aureus from the mammary gland. In contrast to our work,
(1966) reported that cattle vaccinated with Lysigin had Nickerson et al.’s (1999) study suggests that early
increased total immunoglobulin against staphylococci in vaccination followed by multiple immunizations before
milk whey compared with unvaccinated controls. Simi- calving may be a useful adjunct in controlling staphylo-
larly, in a previous study, vaccination with Lysigin resulted coccal mastitis in heifers.
in a significant rise in total anti-S. aureus IgG in milk against
two strains of S. aureus found in the vaccine, but milk IgG1, Acknowledgements
IgG2 and IgM S:P ratios did not differ between vaccinates
and controls (Luby, 2007). The current study did not This project was funded by the Missouri Institute for
evaluate the total anti-S. aureus IgG response in milk Cattle. Vaccine was provided by Boehringer Ingelheim
following vaccination. Vetmedica, Inc., St. Joseph, Missouri, USA. Technical
While the inability of this study to detect differences in assistance was provided by Julie Holle, Ryan Bader,
milk immunoglobulin isotypes between vaccinates and Jonathon Bell and Jeff Dill.
controls may reflect the absence of the studied antibody
isotypes to the vaccine in milk, it may also reflect the fact Conflict of interest
that the SAATK coating antigen did not detect vaccine-
induced antibody. However, a previous study using heat- D.S. Adams is President and CEO of VMRD, Inc. who
killed S. aureus Wood 46, the same strain of S. aureus from manufactured the test kits that were the basis for the ELISA
which the SAATK coating antigens are derived, as an ELISA assays used in the study. Vaccine for the study was donated
coating antigen was capable of detecting significant to J.R. Middleton by Boehringer Ingelheim Vetmedica, Inc.
differences in colostral anti-S. aureus IgA and IgG1 between
vaccinates and controls when mammary quarters were References
inoculated with suspension containing heat-killed S.
aureus strains of serotypes 5 and 8 (Barrio et al., 2003). Borm, A.A., Fox, L.K., Leslie, K.E., Hogan, J.S., Andrew, S.M., Moyes, K.M.,
Similar to the present study, Barrio et al. (2003) did not Oliver, S.P., Schukken, Y.H., Hancock, D.D., Gaskins, C.T., Owens, W.E.,
Norman, C., 2006. Effects of prepartum intramammary antibiotic
detect differences in colostral IgM content between therapy on udder health, milk production, and reproductive perfor-
vaccinates and controls, and IgG2 could not be detected mance in dairy heifers. J. Dairy Sci. 89, 2090–2098.
in the colostrum of any animal (Barrio et al., 2003). Barrio, M.B., Rainard, P., Gilbert, F.B., Poutrel, B., 2003. Assessment of the
opsonic activity of purified bovine sIgA following intramammary
Differences in vaccine preparation and route of adminis- immunization of cows with Staphylococcus aureus. J. Dairy Sci. 86,
tration are possible explanations for the lack of similar 2884–2894.
results for IgA and IgG1 in the present study. Charlier, J., Clarebout, E., Duchateau, L., Vercruysse, J., 2005. A survey to
determine relationships between bulk tank milk antibodies against
Significant differences in antibody S:P ratios were Ostertagia ostertagi and milk production parameters. Vet. Parasitol.
detected between sample periods for milk anti-S. aureus 129, 67–75.
IgA, IgG1, and IgM regardless of group. Variations in serum Fox, L.K., Chester, S.T., Hallberg, J.W., Nickerson, S.C., Pankey, J.W., Weaver,
L.D., 1995. Survey of intramammary infections in dairy heifers at
IgG between sample periods were similarly observed in a
breeding age and first parturition. J. Dairy Sci. 78, 1619–1628.
previous study evaluating the same bacterin studied in the Hogan, J.S., Gonzalez, R.N., Harmon, R.J., Nickerson, S.C., Oliver, S.P.,
present report (Nickerson et al., 1999). The day-to-day Pankey, J.W., Smith, K.L., 1999. Laboratory Handbook on Bovine
variation in the absence of differences between groups Mastitis, 3rd ed. National Mastitis Council, Verona, WI, pp. 1–11,
31–39, 205–217.
suggests that the variability was due to a factor or factors Luby, C.D., Middleton, J.R., 2005. Efficacy of vaccination and antibiotic
other than vaccination. While the day-to-day variation therapy for Staphylococcus aureus mastitis. Vet. Rec. 157, 89–90.
might be explained by assay variability from one sample Luby, C.D., Middleton, J.R., Ma, J., Rinehart, C.L., Bucklin, S., Kohler, C.,
Tyler, J.W., 2007. Characterization of the antibody isotype response in
set to the next, this explanation seems unlikely because S:P serum and milk of heifers vaccinated with a Staphylococcus aureus
ratios were calculated within plate to eliminate plate-to- bacterin (LysiginTM). J. Dairy Res. 74, 239–246.
plate variation in results. Other potential explanations for Ludbrook, J., 1994. Repeated measurements and multiple comparisons in
cardiovascular research. Cardiovasc. Res. 28, 303–311.
the effect of sample period on milk antibody content Middleton, J.R., Fox, L.K., Smith, T.H., 2001. Management strategies to
include fluctuations in milk antibody content due to decrease the prevalence of mastitis caused by one strain of Staphy-
alterations in milk production or alterations in IMI status lococcus aureus in a dairy herd. J. Am. Vet. Med. Assoc. 218, 1615–
1618.
(Charlier et al., 2005).
Middleton, J.R., Timms, L.L., Bader, G.R., Lakritz, J., Luby, C.D., Steevens, B.J.,
2005. Effect of prepartum intramammary treatment with pirlimycin
5. Conclusion hydrochloride on prevalence of early first-lactation mastitis in dairy
heifers. J. Am. Vet. Med. Assoc. 227, 1969–1974.
Middleton, J.R., Ma, J., Rinehart, C.L., Taylor, V.N., Luby, C.D., Steevens, B.J.,
Results of this study suggest that two doses of Lysigin 2006. Efficacy of different Lysigin formulations in the prevention of
administered per label during lactation will not reduce the Staphylococcus aureus intramammary infection in dairy heifers. J.
incidence of new S. aureus or CNS IMI. Additionally, Dairy Res. 73, 10–19.
Nickerson, S.C., Owens, W.E., Tomita, G.M., Widel, P., 1999. Vaccinating
differences in milk IgA, IgG1, IgG2, IgM between vaccinates dairy heifers with a Staphylococcus aureus bacterin reduces mastitis at
and controls could not be detected. These data are calving. Large Anim. Pract. 20, 16–28.
198 J.R. Middleton et al. / Veterinary Microbiology 134 (2009) 192–198
Oliver, S.P., Mitchell, B.A., 1983. Intramammary infections in primigravid mammary infusion of pirlimycin hydrochloride or penicillin–
heifers near parturition. J. Dairy Sci. 66, 1180–1183. novobiocin at the first milking after parturition. J. Dairy Res. 74,
Oliver, S.P., Sordillo, L.M., 1988. Udder health in the periparturient period. 211–217.
J. Dairy Sci. 71, 2584–2606. Owens, W.E., Nickerson, S.C., Washburn, P.J., Ray, C.H., 1991. Efficacy of a
Oliver, S.P., Lewis, M.J., Gillespie, B.E., Dowlen, H.H., 1992. Influence of cephapirin dry cow product for treatment of experimentally induced
prepartum antibiotic on intramammary infections in primigravid Staphylococcus aureus mastitis in heifers. J. Dairy Sci. 74, 3376–
heifers during early lactation. J. Dairy Sci. 75, 406–414. 3382.
Oliver, S.P., Lewis, M.J., Gillespie, B.E., Dowlen, H.H., 1997. Antibiotic Pankey, J.W., Dreschler, P.A., Wildman, E.E., 1991. Mastitis prevalence in
residues and prevalence of mastitis pathogen isolation in heifers primigravid heifers at parturition. J. Dairy Sci. 74, 1550–1552.
during early lactation following prepartum antibiotic therapy. Zen- Trinidad, P., Nickerson, S.C., Alley, T.K., 1990a. Prevalence of intramam-
tralbl. Veterinarmed. B 44, 213–220. mary infection and teat canal colonization in unbred and primigravid
Oliver, S.P., Lewis, M.J., Gillespie, B.E., Dowlen, H.H., Jaenicke, E.C., dairy heifers. J. Dairy Sci. 73, 107–114.
Roberts, R.K., 2003. Prepartum antibiotic treatment of heifers: milk Trinidad, P., Nickerson, S.C., Alley, T.K., Adkinson, R.W., 1990b. Efficacy of
production, milk quality, and economic benefit. J. Dairy Sci. 86, intramammary treatment in unbred and primigravid dairy heifers. J.
1187–1193. Am. Vet. Med. Assoc. 197, 465–470.
Oliver, S.P., Gillespie, B.E., Ivey, S.J., Lewis, M.J., Johnson, D.L., Lamar, K.C., Williams, J.M., Mayerhofer, H.J., Brown, R.W., 1966. Clinical evaluation of
Moorehead, H., Dowlen, H.H., Chester, S.T., Hallberg, J.W., 2004. a Staphylococcus aureus bacterin (polyvalent somatic antigen). Vet.
Influence of prepartum pirlimycin hydrochloride or penicillin–novo- Med. Small Anim. Clin. 61, 789–793.
biocin therapy on mastitis in heifers during early lactation. J. Dairy Sci. Zadoks, R.N., Allore, H.G., Barkema, H.W., Sampimon, O.C., Wellenberg,
87, 1727–1731. G.J., Gröhn, Y.T., Schukken, Y.H., 2001. Cow- and quarter-level risk
Oliver, S.P., Headrick, S.I., Gillespie, B.E., Lewis, M.J., Johnson, D.L., Lamar, factors for Streptococcus uberis and Staphylococcus aureus mastitis. J.
K.C., Moorehead, H., Dowlen, H.H., Hallberg, J.W., 2007. Intramam- Dairy Sci. 84, 2649–2663.
mary infections in heifers during early lactation following intra-