If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Production Animal Health, Faculty of Veterinary Medicine, University of Calgary, Calgary, AB, T2N 4N1, CanadaDepartment Community Health Sciences, Cumming School of Medicine, University of Calgary, Calgary, AB, T2N 4N1, Canada
Clinical mastitis affects 3% of primiparous dairy cattle (heifers) in the first month after calving. Additionally, the prevalence of intramammary infection (IMI) in the months before first calving is high, resulting in a high prevalence of heifers calving with IMI. Precalving therapy is an accepted recommendation for reducing mastitis in multiparous cows, but prophylactic treatment for heifers is uncommon in North America. Objectives of this study were to (1) quantify changes in postcalving udder health in heifers following application of a precalving treatment; (2) compare effectiveness among various types of treatments; and (3) compare effectiveness of various types of treatments against specific pathogens. A systematic review was conducted comparing interventions aimed at improving udder health in heifers. Of 62 included studies, 48 clinical trials were used in a meta-analysis. Data were synthesized using a random effects model for meta-analysis, followed by sub-group analyses comparing treatment types, and specific pathogens with statistical testing using meta-regression. Occurrence of mastitis (defined as elevated somatic cell count, clinical mastitis, and IMI) was reduced in treated heifers compared with untreated controls with a pooled risk ratio of treated to untreated heifers of 0.56 (95% confidence interval: 0.47 to 0.67). Upon stratification by treatment types, teat sealants and combination therapies (vaccines and antimicrobials; antimicrobials and teat sealants; and all 3) were most effective at improving udder health with pooled risk ratios of 0.40 (95% confidence interval: 0.30 to 0.52) and 0.34 (95% confidence interval: 0.25 to 0.45), respectively. Antimicrobials and vaccines also reduced occurrence of IMI and subclinical and clinical mastitis when compared with untreated heifers. Although variation was observed in the pathogen-specific effectiveness of treatments at reducing rates of disease, antimicrobials, teat sealants, and combinations of vaccines or teat sealants with antimicrobials were consistently effective, whereas vaccines were only effective for contagious pathogens. Recommendations for use of antibiotics should consider their relative benefit while also considering potential for increasing antimicrobial resistance.
) as it has implications not only for animal health and welfare, but also for milk quality and production. The occurrence of mastitis is not only high in lactating cows, but also in heifers in the first weeks after calving (
). Heifers with (sub)clinical mastitis, IMI, or both are often not identified because they are not milked, and therefore their teats and milk typically are not examined (
), the NMC published a 10-point plan specifically for controlling heifer mastitis. In this new set of recommendations, only 2 were specific to heifers, whereas the others were similar to recommendations in the first 10-point plan. Some commonly used precalving treatments are vaccines, non-antimicrobial teat sealants, short- and long-term antimicrobials, and combinations of these treatments. Many studies have been conducted on effects of a single type of treatment in heifers (
Effect of teat dipping with a germicide barrier teat dip in late gestation on intramammary infection and clinical mastitis during the first 5 days post-partum in primiparous cows.
Effect of systemic antibacterial administration during prepartum period on coagulase negative staphylococcal intramammary infection in Holstein heifers.
Pathogen group-specific risk factors for intramammary infection in treated and untreated dairy heifers participating in a prepartum antimicrobial treatment trial.
), as well as characterizing pathogens causing ensuing infections. Systematic reviews have been conducted on the effectiveness of antimicrobial treatments in heifers (
). However, no review has focused on quantifying treatment effectiveness in heifers, or compared efficacy of precalving treatment types using a meta-analysis to estimate overall effect size. Objectives of this study were therefore to (1) synthesize results from previous studies to determine and compare effectiveness of various types of precalving therapies aimed at improving udder health; and (2) determine and compare effectiveness of various types of precalving therapies against types of pathogens, grouped by etiology.
MATERIALS AND METHODS
This systematic review and meta-analysis was conducted following a pre-specified protocol created using the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols guidelines (
Online databases searched were CAB Abstracts, Web of Science (complete collection), AGRICOLA, MedLine, and SCOPUS, from inception to September 2017. In addition, 2 individuals (SAN and DBN) hand-searched references of existing reviews on the subject. Gray literature consisted of a search through all available conference proceedings in relevant conferences, as well as websites of pharmaceutical companies and the Proceedings of the World Buiatrics Congress (2000–2014) on The International Veterinary Information Service website. In addition, proceedings of the American Dairy Science Association (1957–2017) and National Mastitis Council (NMC; 1989–2017) annual meetings were searched. Experts in the field were also contacted for information about potential ongoing or unpublished studies, and were identified through the review process or through direct communications at conferences (Mastitis Research Worker's meeting in November 2016, and NMC annual meeting in January 2017). No publication date or language filter was applied to the search, and the search was conducted as recommended (
The search strategy was split into 3 main components corresponding to the population of interest, intervention type, and study outcome. The population of interest included primigravid to primiparous dairy cattle, with the intervention having to be applied during pregnancy and udder health measured after calving. The interventions of interest were any kind of treatment applied precalving aimed specifically at improving udder health. This was intended to exclude interventions such as feed additives intended to improve overall health and growth, as it did not address the primary question of targeted and individual treatment of heifers. Full details of the search terms used are provided (Appendix).
Study Eligibility and Selection
An initial screen of all titles was performed by 2 individuals (SAN and DBN) to determine if manuscripts were eligible for full text review for inclusion in the systematic review. If the title was not detailed enough to decide about moving forward with full-text screening, the abstract was scrutinized by one of the reviewers (SAN) to determine whether it was eligible. A manuscript needed to meet 2 conditions to be considered for a full text review: (1) The manuscript had to report on primary data; and (2) the manuscript had to report on an intervention aimed at improving udder health (no vaccines for other conditions such as leucosis, or treatment with anthelminthic products), although it may not necessarily be the primary aim of the study (comparing mastitis in 2 regions where a particular product may not be allowed).
This initial screen was designed to be fairly broad to encompass as many potentially relevant manuscripts for full text review. A full text review was then conducted on all of these screened manuscripts to ensure that some aspect of intervention was addressed in the manuscript. The reason for this full text screening step in the systematic review was that the second condition described above could have resulted in manuscripts whose primary goal was not assessing intervention effectiveness, but was included as a covariate in their analysis. As it was not the primary goal of the study, it may not have been reported in the abstract and a full text screening would capture these manuscripts.
Identification of Manuscripts for Eligibility in the Systematic Review and Meta-Analysis
The same 2 individuals (SAN and DBN) reviewed full text manuscripts identified after the 2 screening stages described in the previous section using the following criteria to select manuscripts eligible for inclusion in the systematic review and meta-analysis.
d
s
•
Population = primigravid/primiparous dairy cattle (if the primary study population was not primiparous dairy cattle, the manuscript must stratify and report on them separately from older animals).
•
Intervention = any sort of treatment given after confirmed pregnancy, before calving or on the day of calving that was aimed specifically at improving udder health. Studies that did not assess treatment effectiveness as a primary goal (described in the previous section) were excluded from the systematic review and meta-analysis because they would not be designed around that variable and pooling would not be appropriate.
•
Comparison = a treatment must be compared with either another treatment, or no treatment at all.
•
Outcome = postcalving prevalence or incidence of (1) clinical mastitis (CM); (2) subclinical mastitis (SCM; elevated SCC); or (3) IMI (presence of bacteria in the milk) reported within 2 mo after first calving.
•
Study design = randomized controlled trials or observational cohort studies.
The screening of manuscripts was designed to capture studies comparing precalving treatments in heifers to each other or to untreated heifers in their association with incidence of udder disease postcalving. Only randomized controlled trials were meta-analyzed, whereas results from observational cohort studies were included descriptively in the discussion. Agreement between reviewers (SAN and DBN) regarding study inclusion was quantified using a kappa score and percent agreement, and any disagreements were resolved by the last author (HWB).
Study Quality Assessment and Data Extraction
Study quality and risk of bias was assessed using the Systematic Review Centre for Laboratory Animal Experimentation (SYRCLE) tool (
), with some items modified or added to fit better with risk of bias in experimental trials in animals. Briefly, the SYRCLE tool consists of 10 items, corresponding to 5 types of bias: selection, performance, detection, attrition, and reporting bias. Risk of selection bias is evaluated through considering allocation sequence generation, subject baseline characteristics, and allocation concealments. Performance bias is evaluated by considering random housing of heifers and blinding when assigning treatments. Detection bias is assessed by evaluating randomization of subjects during outcome assessment and blinding of outcome assessors. Attrition bias is assessed by evaluating the study's reporting on incomplete outcome data. Lastly, reporting bias is assessed by evaluating outcomes on which studies reported, as well as outcomes that were excluded. A study was considered to be at high risk of a given type of bias if it had failed to address over half of the sources of error contributing to that given type; moderate risk of bias if they failed to address at least 1, but not more than half; and low or no risk of bias if they did not have any sources of error from a given type of bias.
Data were extracted using a single electronic form using Microsoft Excel (Microsoft Corp., Redmond, WA). Study characteristics extracted were (1) country of study, (2) year of publication, (3) breed, (4) intervention type(s), and (5) time of intervention application. As management decisions are based on anticipated calving dates, the intended or median of intended application times was recorded in the extraction form (e.g., for 2 to 3 wk precalving, the number taken was 17.5 d). Outcome characteristics recorded, often multiple for a study, were (1) type of outcome defined as cure rate of existing IMI, SCM, IMI (and its definition), and CM; (2) time of outcome measurement or follow-up in the case of CM; and (3) numbers of individuals (animals or quarters depending on the study) and cases per treatment arm of the study. For studies where postcalving samples were collected on multiple days, the earliest sample that was not the day of calving was used for outcome extraction. Numbers in treatment arms were pooled by intervention or control and type of intervention depending on possible stratification in meta-analyses.
Statistical Analyses
All statistical analyses were performed using the “metafor” package (
) in R version 3.4.2 (R Core Team 2017, R Foundation for Statistical Computing, Vienna, Austria). For all analyses, a P-value <0.05 was considered significant, and a 0.05 < P-value < 0.10 was considered a tendency for significance. Results from randomized clinical trials were pooled using a random-effects meta-analysis, as no observational studies were included in the systematic review. Randomized clinical trials that did not report numbers of individuals and only ratios were not included in the meta-analysis but were included only descriptively. A mixed effects model was selected given the underlying heterogeneity in intervention and population characteristics, and the fact that studies often had multiple treatment types and outcomes. Effect sizes were pooled as risk ratios (RR) because it is most intuitively interpretable in terms of detecting a benefit/harm. All pooled estimates and stratified meta-analyses were visualized using forest plots, and homogeneity was assessed using the Cochrane Q statistic for significance and I2 statistic to quantify the amount of variation due to study heterogeneity. The primary question concerning therapy type was statistically tested using meta-regression, and heterogeneity was then explored by stratifying on relevant variables of practical interest: type of outcome and bacterial species of infection. Variation in effectiveness by pathogen type was addressed through a combination of meta-regression and subgroup analyses. Sensitivity and specificity to detect IMI varies considerably among IMI definitions (
Diagnosing intramammary infections: Comparison of multiple versus single quarter milk samples for the identification of intramammary infections in lactating dairy cows.
), so a meta-regression was conducted to determine whether effect size estimates with rates of IMI differed between studies that described their IMI definition and those that only provided details for culturing. Questions related to effects of study quality or design (abstract versus full text, disease definitions) were also addressed using meta-regression. Publication bias for the primary outcome of interest (mastitis of any form) was assessed using a funnel plot of the RR with the standard error of the RR. Asymmetry was tested using Egger's regression test for small study effects, and visualized with a graph of study variance against the natural log of the RR.
RESULTS
Description of Studies
Search Results
Numbers of studies included at various stages of the review process are shown (Figure 1). Briefly, a total of 974 unique records were identified, with most coming from the database search (n = 1,245 before duplicate removal) and a small number being identified through the described gray literature review (n = 44 before duplicate removal). After screening abstracts, 82 peer-reviewed manuscripts were identified for a more thorough full-text review, along with 31 records from conference proceedings. A small number of non-English manuscripts were identified: Portuguese (n = 1), German (n = 3), Dutch (n = 1), Czech (n = 1), and Italian (n = 1). These manuscripts were reviewed using the pre-specified criteria by colleagues with fluency in the respective languages.
Figure 1Flow diagram of the manuscript selection process. ADSA = American Dairy Science Association; NMC = National Mastitis Council.
Studies Eligible for Inclusion in the Systematic Review
Forty studies from published manuscripts were included in the systematic review, of which 38 were eligible for inclusion in the meta-analysis. All included peer-reviewed manuscripts reported results from clinical trials; 2 could not be included in the meta-analysis because they did not report numbers of outcomes, rather relative measures such as RR and odds ratios. In addition, 22 studies from conference proceedings or abstracts were included in the systematic review. Ten conference proceedings provided enough detail in their results to be eligible for meta-analysis, whereas others were discussed only descriptively. Records from 48 studies were included in the meta-analysis, although multiple records were included for many studies because they compared multiple treatment types (Table 1). A detailed description of the specific products used in each study is provided (AppendixTable A1).
Table 1Study characteristics of the 62 included studies
Type of intervention studied: AM = antimicrobial; TSE = teat sealant; TD = teat dips/sprays; VAC = vaccine; COM = combination of multiple types. Intended application time for intervention was based on expected parturition date; actual time relative to calving was variable. Outcomes reported on the following: CM = clinical mastitis; SCM = subclinical mastitis; CR = cure rate of existing infections.
Reference
Country
Heifer breed
Intervention type
Time of intervention application (d before calving)
Acuña, C. N., and R. E. Chertcoff. 2003. Mastitis control program on dairies with high incidence in first calving heifers: Part 2. In Proc. 42nd NMC Annu. Mtg.
Use of dry cow therapy with internal teat sealant for the prevention of clinical mastitis occurring after calving in heifers with Staphylococcus aureus.
in: Proc. 54th NMC Annu. Mtg. NMC Proceedings Library,
New Prague, MN2015: 171-172
Effect of systemic antibacterial administration during prepartum period on coagulase negative staphylococcal intramammary infection in Holstein heifers.
Prevalence of infection in mammary quarters of nulliparous heifers and efficacy of intra-mammary infusion of antibiotics three weeks before parturition.
Effect of teat dipping with a germicide barrier teat dip in late gestation on intramammary infection and clinical mastitis during the first 5 days post-partum in primiparous cows.
Heifer teats sprayed in the dry period with an iodine teat sanitizer have reduced Streptococcus uberis teat-end contamination and less Streptococcus uberis intra-mammary infections at calving.
Effect of application of an external teat sealant and/or oral treatment with a monensin capsule precalving on the prevalence and incidence of subclinical and clinical mastitis in dairy heifers.
Effect of teat dipping with a germicide barrier teat dip in late gestation on intramammary infection and clinical mastitis during the first 5 days post-partum in primiparous cows.
Heifer teats sprayed in the dry period with an iodine teat sanitizer have reduced Streptococcus uberis teat-end contamination and less Streptococcus uberis intra-mammary infections at calving.
Effect of application of an external teat sealant and/or oral treatment with a monensin capsule precalving on the prevalence and incidence of subclinical and clinical mastitis in dairy heifers.
Intramammary infections in heifers during early lactation following intramammary infusion of pirlimycin hydrochloride or penicillin-novobiocin at the first milking after parturition.
Quarter-level analysis of subclinical and clinical mastitis in primiparous heifers following the use of a teat sealant or an injectable antibiotic, or both, precalving.
Systemic prepartum treatment of end-term dairy heifers with penethamate hydriodide: Effect on udder health, milk yield, and culling until 120 days in milk.
Partial budget analysis of prepartum antimicrobial therapy and Escherichia coli J5 vaccination of dairy heifers and their effect on milk production and milk quality parameters.
1 Type of intervention studied: AM = antimicrobial; TSE = teat sealant; TD = teat dips/sprays; VAC = vaccine; COM = combination of multiple types. Intended application time for intervention was based on expected parturition date; actual time relative to calving was variable. Outcomes reported on the following: CM = clinical mastitis; SCM = subclinical mastitis; CR = cure rate of existing infections.
2 Presented if the study differentiated infectious pathogens. Cont. = contagious; Env. = environmental.
3 Inclusion and analysis based on abstract or conference proceedings.
Design characteristics of all full-text manuscripts that are important for risk of bias are described (Table 2). Based on the SYRCLE guidelines, all studies were at high risk for at least one type of bias. The majority of studies were at moderate risk of selection bias (n = 27), with 2 being at low/no risk of selection bias. Allocation sequence generation was often systematic rather than completely random, and baseline animal characteristics were described in most studies (n = 30). The largest type of bias that this pool of studies was at risk for was detection/performance bias, as most did not describe blinding or randomized animal housing. Most (n = 31) studies adequately addressed incomplete data by providing numbers of heifers at each stage of the study, with reasons provided for their removal or exclusion from the trial minimizing risk of attrition bias. No studies were at risk of reporting bias, as they all either provided justification for outcomes measured or they were selected in part based on their reporting of outcomes.
Table 2Risk of bias as assessed by Systematic Review Centre for Laboratory Animal Experimentation (SYRCLE) guidelines (Hooijmans et al., 2013) from full-text manuscripts included in systematic review (n = 40)
Effect of systemic antibacterial administration during prepartum period on coagulase negative staphylococcal intramammary infection in Holstein heifers.
Effect of teat dipping with a germicide barrier teat dip in late gestation on intramammary infection and clinical mastitis during the first 5 days post-partum in primiparous cows.
Heifer teats sprayed in the dry period with an iodine teat sanitizer have reduced Streptococcus uberis teat-end contamination and less Streptococcus uberis intra-mammary infections at calving.
Effect of application of an external teat sealant and/or oral treatment with a monensin capsule precalving on the prevalence and incidence of subclinical and clinical mastitis in dairy heifers.
Intramammary infections in heifers during early lactation following intramammary infusion of pirlimycin hydrochloride or penicillin-novobiocin at the first milking after parturition.
Quarter-level analysis of subclinical and clinical mastitis in primiparous heifers following the use of a teat sealant or an injectable antibiotic, or both, precalving.
Systemic prepartum treatment of end-term dairy heifers with penethamate hydriodide: Effect on udder health, milk yield, and culling until 120 days in milk.
Partial budget analysis of prepartum antimicrobial therapy and Escherichia coli J5 vaccination of dairy heifers and their effect on milk production and milk quality parameters.
Another aspect of variability among studies and a possible source of bias was how they defined outcomes (Table 3), IMI in particular. Many studies (n = 16) only described culturing details for their milk samples, without a strict definition of what was considered an IMI. Risk ratio for IMI did not differ between treated and untreated heifers when comparing the 2 groups of studies (P = 0.93). Definition of IMI was not used in any subsequent analyses, or considered for stratification.
Table 3Outcome definitions in full-text manuscripts included in the systematic review (n = 40)
Effect of systemic antibacterial administration during prepartum period on coagulase negative staphylococcal intramammary infection in Holstein heifers.
Effect of teat dipping with a germicide barrier teat dip in late gestation on intramammary infection and clinical mastitis during the first 5 days post-partum in primiparous cows.
Heifer teats sprayed in the dry period with an iodine teat sanitizer have reduced Streptococcus uberis teat-end contamination and less Streptococcus uberis intra-mammary infections at calving.
Effect of application of an external teat sealant and/or oral treatment with a monensin capsule precalving on the prevalence and incidence of subclinical and clinical mastitis in dairy heifers.
Intramammary infections in heifers during early lactation following intramammary infusion of pirlimycin hydrochloride or penicillin-novobiocin at the first milking after parturition.
Quarter-level analysis of subclinical and clinical mastitis in primiparous heifers following the use of a teat sealant or an injectable antibiotic, or both, precalving.
Systemic prepartum treatment of end-term dairy heifers with penethamate hydriodide: Effect on udder health, milk yield, and culling until 120 days in milk.
Partial budget analysis of prepartum antimicrobial therapy and Escherichia coli J5 vaccination of dairy heifers and their effect on milk production and milk quality parameters.
The overall RR of udder disease of any sort in treated compared with untreated heifers was 0.60 (95% CI: 0.47 to 0.67), demonstrating a reduction in udder diseases after calving in treated compared with untreated heifers (P < 0.0001). Effect size estimates from abstracts/proceedings tended to be slightly lower and had more variance than estimates from full-text manuscripts (P = 0.08; Figure 2), but their inclusion did not have an effect on overall effect size (RR = 0.60 for full-text manuscripts alone; RR = 0.57 for a combined meta-analysis). All further analyses were conducted on results from both full-text manuscripts and abstracts/conference proceedings without stratification.
Figure 2Forest plot of effect size estimates from all (n = 48) studies included in the meta-analysis assessing effectiveness of precalving therapy at improving postcalving udder health in dairy heifers. Size of boxes reflects relative size of effect estimate; diamonds represent pooled estimates within strata. CM = clinical mastitis; new IMI = IMI that was not present precalving; SCM = subclinical mastitis (defined as elevated SCC).
The majority of records (n = 34) used IMI as the outcome, whereas 15 records assessed CM and 2 assessed SCM. Effect size estimates for CM tended to be lower than those for IMI (P = 0.07), but tended to be higher than those for SCM (P = 0.06; Table 4). For subsequent meta-analyses and regressions, all outcomes other than cure rate were combined through a random-effects term included for outcome type within studies to account for any within-study correlation. Risk of cure of existing IMI following precalving treatment with antimicrobials was higher than spontaneous cure with no treatment [RR = 1.63 (95% CI: 1.19 to 2.25); P = 0.003].
Table 4Meta-regression of records included in a systematic review precalving therapies for improving udder health in primiparous dairy heifers on the outcome used to assess udder health
Heterogeneity refers to residual heterogeneity from meta-regression model attributed to differences among studies. I2 = statistic to quantify the amount of variation due to study heterogeneity; Q = Cochrane Q statistic for significance.
Heterogeneity refers to residual heterogeneity from meta-regression model attributed to differences among studies. I2 = statistic to quantify the amount of variation due to study heterogeneity; Q = Cochrane Q statistic for significance.
Subclinical mastitis defined as a heifer having an elevated SCC (measured by electrical conductivity of milk, the California mastitis test or a quantified SCC >120,000 or 250,000 cells/mL).
0.30 (0.18–0.50)
2
P < 0.0001
1 Ratio of risk of infection in treated heifers to risk of infection in untreated controls.
2 Heterogeneity refers to residual heterogeneity from meta-regression model attributed to differences among studies. I2 = statistic to quantify the amount of variation due to study heterogeneity; Q = Cochrane Q statistic for significance.
3 Subclinical mastitis defined as a heifer having an elevated SCC (measured by electrical conductivity of milk, the California mastitis test or a quantified SCC >120,000 or 250,000 cells/mL).
A meta-regression on treatment type using CM, SCM, and IMI as outcomes resulted in a difference in effectiveness among treatment types (Table 5). Teat sealants and combination therapies were the most effective treatments (P < 0.0001), although antimicrobials, vaccines, and other treatments still reduced risk of CM, SCM, or IMI compared with untreated controls (P < 0.0001). Teat disinfection before calving did not reduce udder disease compared with untreated controls. Records not included in the meta-analysis also reported similar patterns—antimicrobials were effective at preventing CM and IMI, as well as curing existing IMI. Farms observed longitudinally had decreasing prevalence of IMI and incidence of CM around calving after incorporating precalving antimicrobial treatment into heifer management (
Acuña, C. N., and R. E. Chertcoff. 2003. Mastitis control program on dairies with high incidence in first calving heifers: Part 2. In Proc. 42nd NMC Annu. Mtg.
Table 5Meta-regression of records included in a systematic review of precalving treatment for improving udder health in primiparous dairy heifers on the type of treatment assessed in the study
Antimicrobial includes antibiotics as well as antiseptic compounds such as chlorhexidine. Combinations in included studies consisted of teat sealants and antimicrobials; vaccines and antimicrobials; teat sealants, vaccines and antimicrobials. Other includes insecticides, and management practices such as prepartum milking (Appendix Table A1).
Residual heterogeneity from meta-regression model attributed to differences in studies. I2 = statistic to quantify the amount of variation due to study heterogeneity; Q = Cochrane Q statistic for significance.
Risk ratio different (P < 0.05) from studies assessing antimicrobial effectiveness.
(0.74–3.1)
3
Q = 369.78
Vaccine
0.55 (0.35–0.86)
12
P < 0.0001
Combination
0.34 (0.25–0.45)
4
Other
0.50 (0.11–1.1)
4
1 Antimicrobial includes antibiotics as well as antiseptic compounds such as chlorhexidine. Combinations in included studies consisted of teat sealants and antimicrobials; vaccines and antimicrobials; teat sealants, vaccines and antimicrobials. Other includes insecticides, and management practices such as prepartum milking (Appendix Table A1).
2 Ratio of risk of infection in treated heifers to risk of infection in untreated controls.
3 Residual heterogeneity from meta-regression model attributed to differences in studies. I2 = statistic to quantify the amount of variation due to study heterogeneity; Q = Cochrane Q statistic for significance.
4 Risk ratio different (P < 0.05) from studies assessing antimicrobial effectiveness.
Although studies differentiated culture results to varying levels of detail, pathogens were grouped into 3 categories: contagious (Staphylococcus aureus, Streptococcus agalactiae), environmental (Escherichia coli, non-agalactiae streptococci), and CNS. Subgroup analyses including meta-regression on treatment type were conducted using the 3 categories as outcomes (Table 6). Antimicrobials, teat sealants, and vaccines all presented similar reductions (RR = 0.51, 0.40, and 0.56, respectively) in contagious pathogen CM and IMI compared with untreated controls. Only 1 study reported on teat dips/sprays (
Effect of teat dipping with a germicide barrier teat dip in late gestation on intramammary infection and clinical mastitis during the first 5 days post-partum in primiparous cows.
Effect of application of an external teat sealant and/or oral treatment with a monensin capsule precalving on the prevalence and incidence of subclinical and clinical mastitis in dairy heifers.
); in both studies, the 2 therapies did not change occurrence of CM or IMI compared with controls. Manuscripts not included in the pooled estimate also reported a reduced incidence of CM and IMI by contagious pathogens when heifers received precalving treatment with antimicrobials, and also when fly control was implemented for precalving heifers on pasture.
Table 6Meta-regression of records included in a systematic review of precalving treatment for improving udder health in primiparous dairy heifers on the type of treatment assessed in the study, stratified by type of pathogen
Contagious pathogens include Staphylococcus aureus and Streptococcus agalactiae. Environmental pathogens include Escherichia coli and non-agalactiae streptococci.
Antimicrobial includes antibiotics as well as antiseptic compounds such as chlorhexidine. Combinations in included studies consisted of teat sealants and antimicrobials; vaccines and antimicrobials; and teat sealants, vaccines, and antimicrobials. Other includes insecticides and management practices such as prepartum milking (Appendix Table A1).
Residual heterogeneity from meta-regression model attributed to differences in studies. I2 = statistic to quantify the amount of variation due to study heterogeneity; Q = Cochrane Q statistic for significance.
Antimicrobial includes antibiotics as well as antiseptic compounds such as chlorhexidine. Combinations in included studies consisted of teat sealants and antimicrobials; vaccines and antimicrobials; and teat sealants, vaccines, and antimicrobials. Other includes insecticides and management practices such as prepartum milking (Appendix Table A1).
Intervention type within pathogen type stratum did not have enough records for meta-analysis. Risk ratio presented was calculated from the single record within that stratum.
Intervention type within pathogen type stratum did not have enough records for meta-analysis. Risk ratio presented was calculated from the single record within that stratum.
Antimicrobial includes antibiotics as well as antiseptic compounds such as chlorhexidine. Combinations in included studies consisted of teat sealants and antimicrobials; vaccines and antimicrobials; and teat sealants, vaccines, and antimicrobials. Other includes insecticides and management practices such as prepartum milking (Appendix Table A1).
Antimicrobial includes antibiotics as well as antiseptic compounds such as chlorhexidine. Combinations in included studies consisted of teat sealants and antimicrobials; vaccines and antimicrobials; and teat sealants, vaccines, and antimicrobials. Other includes insecticides and management practices such as prepartum milking (Appendix Table A1).
Intervention type within pathogen type stratum did not have enough records for meta-analysis. Risk ratio presented was calculated from the single record within that stratum.
(0.19–3.34)
1
1 Contagious pathogens include Staphylococcus aureus and Streptococcus agalactiae. Environmental pathogens include Escherichia coli and non-agalactiae streptococci.
2 Antimicrobial includes antibiotics as well as antiseptic compounds such as chlorhexidine. Combinations in included studies consisted of teat sealants and antimicrobials; vaccines and antimicrobials; and teat sealants, vaccines, and antimicrobials. Other includes insecticides and management practices such as prepartum milking (Appendix Table A1).
3 Ratio of risk of infection in treated heifers to risk of infection in untreated controls.
4 Residual heterogeneity from meta-regression model attributed to differences in studies. I2 = statistic to quantify the amount of variation due to study heterogeneity; Q = Cochrane Q statistic for significance.
5 Intervention type within pathogen type stratum did not have enough records for meta-analysis. Risk ratio presented was calculated from the single record within that stratum.
6 Risk ratio was different (P < 0.05) from studies assessing antimicrobial effectiveness.
Teat sealants and combination therapies reduced occurrence of CM and IMI with environment pathogens over untreated controls (RR = 0.27 and 0.25, respectively; Table 6). Antimicrobials, teat disinfection, and vaccines did not have any benefit compared with untreated controls, but 2 studies not included in the meta-analysis (
) reported a reduction of clinical signs during E. coli IMI following precalving vaccination. Fly control for heifers on pasture reduced Streptococcus spp. IMI. Studies reporting on CM and IMI where CNS were isolated reported a reduction compared with controls in heifers treated with antimicrobials, teat-sealants, and combination therapies, with combination therapies having the greatest reduction (RR = 0.34), and similar conclusions were drawn from studies not included in the meta-analysis. Vaccines did not reduce CM and IMI with CNS occurrence compared with untreated controls. Due to only single studies assessing other therapies and teat dips/sprays, pooling was not possible, but neither demonstrated a reduction in occurrence of CM and IMI with CNS compared with untreated controls.
Meta-Analysis Statistics
Heterogeneity
Significant heterogeneity was present in all analyses. Mixed effects meta-analysis using all records with random effects for outcome type and intervention type had an overall I2 = 0.87 (Q = 439.28, P < 0.0001). Introducing outcome type as a moderator did not reduce heterogeneity noticeably; I2 = 0.85 (Q = 343.54, P < 0.0001), nor did conducting a meta-regression using outcome type affect heterogeneity; I2 = 0.85 (Q = 369.78, P < 0.0001). Sub-group analyses by pathogen type led to reductions in residual heterogeneity, except for the CNS analysis: contagious pathogen analysis I2 = 0.29 (Q = 39.36, P = 0.03); environmental pathogen analysis I2 = 0.57 (Q = 71.77, P < 0.0001); CNS analysis I2 = 0.88 (Q = 212.71, P < 0.0001).
Publication Bias
Based on visual assessment of the funnel plot of standard error compared with the effect size estimate (Figure 3) from each study, we inferred that some asymmetry due to missing studies showing an increased risk following treatment with small sample sizes (bottom right quadrant). Egger's regression small study effects indicated a possibility of publication bias as an increasing standard error was associated with a larger effect size estimate (P < 0.0001; Figure 4).
Figure 3Egger's regression for small-study effects in a meta-analysis of studies assessing effectiveness of precalving therapies at improving udder health in dairy heifers plotted as the natural log of the risk ratio from a given study against the estimate of variance from the same study. Vertical bars indicate 95% confidence intervals at the intercept (variance = 0; largest possible studies).
Figure 4Funnel plot to assess asymmetry and publication bias in meta-analysis of studies assessing effectiveness of precalving therapies at improving udder health in dairy heifers. Relative risk = 1 indicates equal risk of udder disease in treated and untreated animals (no effect of treatment). Darkest triangle indicates area of 90% confidence, light gray triangle indicates area of 95% confidence, and white triangle indicates area of 99% confidence.
Management of mastitis on dairy farms often starts after the heifer's first calving, too often resulting in less than optimal udder health for first lactation heifers (
). Many housing and management practices associated with udder health of older cows are also associated with improved udder health in primiparous heifers (
), but the level of implementation in nonlactating heifers is unknown. Prophylactic treatments of the udder before calving (e.g., antimicrobial dry cow therapy or using internal teat sealants) are widely adopted for use in the lactating herd (e.g.,
), but these treatments are less commonly used in primiparous heifers. This systematic review and meta-analysis of studies assessing effectiveness of precalving therapies for improving postcalving udder health in heifers demonstrated an overall benefit to udder health in heifers when using any kind of precalving prophylactic treatment. Treatments varied in both cost to purchase as well as costs and labor associated with application, so an understanding of the effectiveness of treatments is important to inform management decisions. Effectiveness of treatments varied considerably and was also dependent on the type of bacteria causing the IMI.
This is the first review to meta-analyze results from studies of multiple types of precalving therapies in primiparous heifers. Recent reviews have considered heifer mastitis as a whole (
), so results from studies on older lactating cattle may not be generalizable to heifers. A synthesis of studies aimed at heifers in particular is therefore necessary when trying to implement more targeted management strategies. Findings from this review were in line with most individual studies and reviews on lactating cows that demonstrated a net positive benefit of any type of treatment for improving udder health. Previous reviews have demonstrated differential effectiveness of antimicrobials (
) or E. coli (Enviracor and StartVac). Compared with the present review focusing on heifers specifically, variation in treatment effectiveness by pathogen type was similar to previous reports from studies focusing on older cows.
As mentioned above, management of nonlactating heifers is less than ideal, with practices such as comingling with older cattle and moving cows close to calving and increasing stress being common on most dairy farms (
). This could be an important reason why treatments were effective in improving udder health. It is therefore likely that effectiveness of treatments, particularly antimicrobial treatment, will be less on farms that devote more attention to nonlactating heifers.
Although the effectiveness of antimicrobials was consistent across gram-positive pathogens, their use needs to be balanced with concerns about the rising prevalence of antimicrobial resistance (
Restricting the use of antibiotics in food-producing animals and its associations with antibiotic resistance in food-producing animals and human beings: A systematic review and meta-analysis.
). To prevent overuse, the NMC's 10-point plan for heifer mastitis management recommends prepartum antibiotic treatment only under very specific conditions. Other treatments can be just as effective, if not more so, than antimicrobials in heifers, and current research is beginning to focus on their effectiveness in older cows (
). Antimicrobial treatment of nonlactating heifers improved udder health more than other treatments when contagious (gram-positive) bacteria caused udder health problems after calving. In contrast, teat sealants and combination therapies were more effective than antimicrobial treatment when environmental pathogens were involved, suggesting that antimicrobials are not always the best option. It is therefore important when udder health problems after calving are an issue on a farm and antimicrobial treatment is considered, to culture quarter milk of heifers after calving.
In addition to selecting the best treatment for each farm, personnel must also be properly trained in its application. For example, improper application of intramammary treatments, either causing trauma by inserting the cannula too far or using poor hygiene, can actually increase a quarter's risk of infection (
). It is difficult to quantify the effect of iatrogenic infections in studies included in this review. Although some studies mentioned disinfection/preparatory procedures, no studies detailed training for personnel in treatment application. Safety of treatment application and ease of application may also vary by treatment type and should be considered when selecting a treatment. For example, feed supplements are much easier and safer to administer compared with prepartum intramammary infusions of antimicrobials. A comparison of systemic antibiotic treatment with those applied intramammarily could further inform decisions if they differ in effectiveness or ease of application. This comparison was not possible for the present review because intramammary infusions of antibiotics (n = 31) greatly outnumbered systemically applied antimicrobials (n = 6), resulting in a lack of power due to the unbalanced groups. Similarly, a comparison of antimicrobial treatments based on classes may have resulted in differences in effectiveness; however, because the vast majority of antibiotic products used in the included studies were macrolides or macrolide combinations, lack of power was again an issue.
provided a more comprehensive and in-depth review specific to antimicrobial treatments for heifers.
In comparing all types of treatments for heifers, this review faced the challenge of significant between-study variability, not only in treatments used, but also in other aspects of study design and setting. Although we tried to explore heterogeneity through subgroup analyses, many other study characteristics still varied considerably to allow for valid subgroup analyses (e.g., timing of treatment application, study location, timing of data collection, and geographic location). As a result, significant residual heterogeneity was present even after stratification. Although the significant between-study variability may cast some doubt on the precision of the pooled results, the overall message remains unchanged, as the direction of the effect remained consistent among studies, even if the effect size did not. Specifically, precalving treatment of heifers improved udder health after calving, and the degree of improvement varied by treatment type and pathogen. Variation in characteristics such as breed, geographic location, and sampling season or timing of treatment application and specific treatment are inherently linked, making it difficult to be consistent across studies from various countries and studying a variety of treatments.
An assessment of the risk of bias focuses on characteristics of study design and reporting in a way that is less dependent on factors outside of the researchers' control. The assessment using the SYRCLE tool (
) highlighted some areas with significant risk of bias. Although randomization of housing and outcome assessment order are less likely to contribute to detection or performance bias in livestock studies (SYRCLE was designed for laboratory trials using animals), blinding of both treatment allocator or outcome assessor was not mentioned in most of the studies, nor was allocation concealment. For some outcomes (e.g., SCM or IMI), blinding may be inherent, because milk samples are collected by farm personnel and the outcome (SCC or bacterial culture) is assessed by laboratory technicians. Despite the possible inherent blinding of outcome assessors, lack of blinding of the treatment allocator could still contribute to detection bias, and both types of blinding should be reported, with justification provided if blinding was not done. Similarly, randomization and allocation sequence generation were only described in 6 studies, and presented a possible source of selection bias. Clear outcome definitions, especially for IMI, were also inconsistently reported; therefore, different definitions could contribute to residual heterogeneity. Although just defining IMI did not change effect size from studies that did not, looking at specific definitions would likely provide different results because each would have its own operating characteristics (
Diagnosing intramammary infections: Comparison of multiple versus single quarter milk samples for the identification of intramammary infections in lactating dairy cows.
). This likely is not a reflection of study quality, but more on inconsistency in reporting standards for studies on livestock. The Reporting Guidelines for Randomized Controlled Trials for Livestock (REFLECT;
) were developed to address this issue, but are still not widely implemented. Although this review reports on a substantial number of studies, an updated review in the future once the REFLECT statement is more widely implemented would likely detect more consistency among studies, leading to more precise pooled estimates of protective effect size.
As the first review to focus on the effectiveness of all existing precalving therapies at improving udder health in primiparous dairy heifers, the present study effectively synthesized results from a large number of studies. Heterogeneity was explored through subgroup analyses and meta-regression, focusing primarily on the main goals of the study: understanding variation in therapy success by treatment type and pathogen. Despite this exploration of heterogeneity, even stratification by both treatment type and pathogen still yielded significant residuals and unexplained study variability. As discussed, this was likely a reflection of differences in study settings, design characteristics, and reporting. Although this degree of study variability affected precision of pooled estimates, based on the direction of the effect, there was a positive effect on udder health when precalving treatments were applied. Publication bias in a review is another aspect that could affect the reliability of findings, and Egger's regression indicated that there were significant small-study effects in the review. Based on examination of the funnel plot, no studies had a negative effect of precalving treatment and used a small sample size. Although this may be a reflection of publication bias, in this case it is more clearly reflective of the true heterogeneity in the included studies, attributed primarily to treatment intensity and differences in underlying risk (
). Treatments varied considerably in timing of application and dosage, both of which would increase variability. Differences in geographical location, affecting housing, nutrition and other management factors, and animal breed could have contributed to differences in regional risks for mastitis, or differences in susceptibilities of heifers to disease. Asymmetry in the funnel plot may also be an artifact (
), as the majority of interventions under study were very unlikely to increase the risk of udder diseases over untreated heifers, making it very unlikely that small studies would show harm to udder health by precalving treatments. Although statistical evidence suggested publication bias, this was more likely a reflection of study variability as quantified by the residual heterogeneity rather than a bias in study selection.
CONCLUSIONS
In this first review of precalving treatments for improving udder health in primiparous dairy heifers, we assessed the findings of 60 studies, 48 of which were meta-analyzed for pooled estimates. Overall, any type of precalving therapy improved udder health by reducing occurrence of IMI, CM, and SCM compared with untreated controls. Variation in success by therapy type demonstrated that studies assessing antimicrobials, teat sealants, vaccines, and combination therapies demonstrated a significant reduction in the risk of udder disease, whereas those using teat dips/sprays or other types of therapy did not reduce risk compared with untreated heifers. There was also significant variation by pathogens that the studies examined: antimicrobials, teat sealants, vaccines, and combination therapies were effective at reducing risk of udder disease caused by contagious pathogens; antimicrobials, teat sealants, and vaccines were effective at reducing risk of udder disease caused by environmental pathogens; and finally, antimicrobials, teat sealants, vaccines, and combinations were effective at reducing udder disease caused by CNS. Once rigid reporting guidelines are more widely implemented by journals and authors, an updated literature review on udder health of dairy heifers should yield more precise pooled estimates, with decreased between-study variability.
ACKNOWLEDGMENTS
This study was supported by the Industrial Research Chair in Infectious Diseases of Dairy Cattle, funded by Canada's Natural Sciences and Engineering Research Council (NSERC) Industrial Research Chair Program (Ottawa, ON, Canada), with industry contributions from Alberta Milk (Edmonton, AB, Canada), the Dairy Farmers of Canada (Ottawa, ON, Canada), Westgen Endowment Fund (Milner, BC, Canada), the BC Dairy Association (Burnaby, BC, Canada), Canadian Dairy Network (Guelph, ON, Canada), CanWest DHI (Guelph, ON, Canada), and Dairy Farmers of Manitoba (Winnipeg, MN, Canada). The first and second authors were supported by a NSERC CREATE in Milk Quality Graduate student scholarship from the Canadian Bovine Mastitis and Milk Quality Research Network (Saint-Hyacinthe, QC). The first author was also supported by a Queen Elizabeth II scholarship from the province of Alberta. We also thank Todd Wilson and Karen Tang (teaching assistants for MDCH 741: Systematic Reviews and Meta-Analysis at the University of Calgary) for their help in planning the review and refining research questions, Lorraine Toews (librarian at the Health Sciences Library at the University of Calgary) for her assistance in developing the database and gray literature search strategies, and John Kastelic (University of Calgary) for editing the manuscript.
APPENDIX
The following search terms were used in database searches. The MeSH search is integrated in the Web of Science, so the same terms were employed there.
The search terms used were separated into 3 separate components corresponding to the study population, intervention, and outcome type. Each of these was combined with the Boolean operator AND. Subject heading searches were conducted separately from the text search as the available subject headings were not comprehensive.
1)
To identify studies with a study population looking at heifers and not just all cows grouped together, the search was split into 2 parts (to ensure that the age of animals in the study was appropriate, as well as restrict our search to dairy cows only), which were combined using the Boolean operator AND:
a.
The terms “dairy” and “milk*” combined with the Boolean operator OR. The corresponding subject headings were “dairy herds,” “dairy cows,” “dairy cattle,” and “milking” combined with the operator OR.
b.
The terms “heifer*,” “primiparous,” “primigravid,” “nulliparous” combined with the Boolean operator OR. The corresponding subject headings were “heifers” and “bred heifers” combined with the Boolean operator OR.
2)
To identify studies with the relevant interventions, the search was again split into 2 parts (timing of intervention had to be precalving, as well as specify all types of intervention), which were again combined using the Boolean operator AND:
a.
The terms “precalving,” “prepartum,” and “calving” combined with the Boolean operator OR. The corresponding subject heading search terms were “calving” and “prepartum period” combined using the operator OR.
b.
The terms “therap*,” “treat*,” “antimicrob*,” “antibiotic,” and “seal*” combined using the operator OR. The corresponding subject headings were “drug therapy,” “therapy,” “treatment,” and “combination therapy” combined using the Boolean operator OR.
3)
To identify all possibly relevant outcomes the search terms used were “mastit*,” “infect*,” “clinical mastit*,” “bacteri*,” “inflam*,” “pathogen,” and “microorganism” combined using the Boolean operator OR. The corresponding subject headings used were “mastitis,” “summer mastitis,” “subclinical mastitis,” “mycotic mastitis,” “bovine mastitis,” and “inflammation” combined using the Boolean operator OR.
Table A1Details on product used, primary therapeutic ingredient(s), method of application, and categorization of intervention for each of the 62 studies
Acuña, C. N., and R. E. Chertcoff. 2003. Mastitis control program on dairies with high incidence in first calving heifers: Part 2. In Proc. 42nd NMC Annu. Mtg.
Use of dry cow therapy with internal teat sealant for the prevention of clinical mastitis occurring after calving in heifers with Staphylococcus aureus.
in: Proc. 54th NMC Annu. Mtg. NMC Proceedings Library,
New Prague, MN2015: 171-172
Effect of systemic antibacterial administration during prepartum period on coagulase negative staphylococcal intramammary infection in Holstein heifers.
Prevalence of infection in mammary quarters of nulliparous heifers and efficacy of intra-mammary infusion of antibiotics three weeks before parturition.
Effect of teat dipping with a germicide barrier teat dip in late gestation on intramammary infection and clinical mastitis during the first 5 days post-partum in primiparous cows.
Heifer teats sprayed in the dry period with an iodine teat sanitizer have reduced Streptococcus uberis teat-end contamination and less Streptococcus uberis intra-mammary infections at calving.
Effect of application of an external teat sealant and/or oral treatment with a monensin capsule precalving on the prevalence and incidence of subclinical and clinical mastitis in dairy heifers.
Intramammary infections in heifers during early lactation following intramammary infusion of pirlimycin hydrochloride or penicillin-novobiocin at the first milking after parturition.
Quarter-level analysis of subclinical and clinical mastitis in primiparous heifers following the use of a teat sealant or an injectable antibiotic, or both, precalving.
Systemic prepartum treatment of end-term dairy heifers with penethamate hydriodide: Effect on udder health, milk yield, and culling until 120 days in milk.
Partial budget analysis of prepartum antimicrobial therapy and Escherichia coli J5 vaccination of dairy heifers and their effect on milk production and milk quality parameters.
Use of dry cow therapy with internal teat sealant for the prevention of clinical mastitis occurring after calving in heifers with Staphylococcus aureus.
in: Proc. 54th NMC Annu. Mtg. NMC Proceedings Library,
New Prague, MN2015: 171-172
Acuña, C. N., and R. E. Chertcoff. 2003. Mastitis control program on dairies with high incidence in first calving heifers: Part 2. In Proc. 42nd NMC Annu. Mtg.
Effect of systemic antibacterial administration during prepartum period on coagulase negative staphylococcal intramammary infection in Holstein heifers.
Prevalence of infection in mammary quarters of nulliparous heifers and efficacy of intra-mammary infusion of antibiotics three weeks before parturition.
Diagnosing intramammary infections: Comparison of multiple versus single quarter milk samples for the identification of intramammary infections in lactating dairy cows.
Effect of teat dipping with a germicide barrier teat dip in late gestation on intramammary infection and clinical mastitis during the first 5 days post-partum in primiparous cows.
Heifer teats sprayed in the dry period with an iodine teat sanitizer have reduced Streptococcus uberis teat-end contamination and less Streptococcus uberis intra-mammary infections at calving.
30224-8&id=gr1.jpg){kind=link}
30224-8&id=gr2.jpg){kind=link}
30224-8&id=gr3.jpg){kind=link}
30224-8&id=gr4.jpg){kind=link}