In conventional and automatic milking systems (AMS), mastitis is a major challenge for the dairy industry (Bradley, Reference Bradley2002). Cows with clinical mastitis (CM) show signs of inflammation (pain, hyperaemia, local or systemic temperature increase and oedema) and changes in milk (e.g. flakes and presence of blood) that are easily identified by experienced milkers (Ruegg, Reference Ruegg2017). However, in AMS, detection of CM relies on sensors, which may pose a challenge for reliable disease diagnosis.
Because CM can rapidly become severe or chronic, in commercial dairies treatment is usually applied without the diagnosis of the causal pathogen (Oliveira et al., Reference Oliveira, Hulland and Ruegg2013). Mastitis is commonly caused by bacteria, which require microbiological evaluation (e.g. culture tests) that may take too long for results to arrive (Vasquez et al., Reference Vasquez, Nydam, Capel, Eicker and Virkler2017). This dilemma has led to the development of rapid on-farm microbiological tests for mastitis pathogen detection (Royster et al., Reference Royster, Godden, Goulart, Dahlke, Rapnicki and Timmerman2014).
Using mastitis pathogen identification prior to treatment has reduced case duration and probability of subsequent treatments for the same case (Griffioen et al., Reference Griffioen, Velthuis, Koop and TJGM2021) reducing the chance of generating antimicrobial resistance (Lago et al., Reference Lago, Godden, Bey, Ruegg and Leslie2011) leading to overall reduced use of antimicrobials (Vasquez et al., Reference Vasquez, Nydam, Capel, Eicker and Virkler2017).
The objective of this study was to evaluate the use of a previously validated rapid on-farm microbiological test on CM case duration and the probability of requiring more than one treatment to cure CM in an AMS herd.
Material and methods
This study was approved by the Animal Use Ethics Committee of the Pontifical Catholic University of Paraná (PUCPR) – protocol 020343. Data were collected from August 2019 to July 2021.
Housing and management
The lactating herd comprised 58 dairy cows housed in a freestall barn divided into two pens, measuring 362 m2 (19.6 × 18.5 m) each. The cows had an average milk production of approximately 11,000 kg per 305 d, with 2.7 ± 0.2 (SD) milkings/cow/d. The AMS settings allowed between 2 and 5 milkings per day, depending on days in milk (DIM) and individual milk production. The stalls were equipped with rubber mattresses covered by a 2–5 cm layer of sawdust and were cleaned daily. Each pen included a feed alley of 84 m2 (21 × 4 m) and 35 stalls; stall stocking density was always less than 100%.
Cows were provided with free access to water troughs and automatic brushes, and manure removal was carried out using automatic scrapers. Additionally, farm staff manually removed faeces from the bedding daily. The cows were milked using a voluntary milking system (DeLaval VMS™ V300), using the milk-first layout. This allowed cows to voluntarily go for milking before accessing the feeding alley. In the AMS, milk preparation begins with a pre-dipping spray applied to each teat individually, followed by the attachment of a dedicated cleaning unit from the milking cluster. This unit cleans the teats using pulses of warm water and dries them under vacuum. Subsequently, the teat cups are attached and removed individually, quarter by quarter, and a post-dipping spray is applied at the end of milking.
The cows’ diet consisted of a partial mixed ration (PMR) provided twice a day in the feed alley. They also received a commercial pellet (approximately 4 kg/day) during milking. The formulation of the PMR followed the recommendations of the National Research Council (NASEM (National Academies of Sciences, Engineering, and Medicine), 2021), using RLM 3.3 Software. Cows were fed at approximately 08:00 and 16:00 h daily.
Mastitis case definition
Cows that had variation of one or more AMS milking parameters in the last milking were flagged for veterinary inspection. Milking parameter criteria used to flag cows included: electrical conductivity >6.44 mS/cm per quarter (Norberg et al., Reference Norberg, Hogeveen, Korsgaard, Friggens, Sloth and Lovendahl2004); quarter milk yield below 80% of expected production; values >2 in the MDi (Delaval Mastitis Detection index); presence of blood in milk; >200,000 cells/ml in this milking (OCC – online cell counter, Delaval, Jaguariúna, Brazil). The MDi consists of electrical conductivity (EC), milking interval, and the presence of blood (ppm) at the quarter level, with values ranging from 0.8 to 4.0. Values below 1.4 indicate a low probability of mastitis, while values around 2.0 or higher suggest a high risk of mastitis (Lusis et al., Reference Lusis, Antane and Laurs2017). This data is provided through the farm management platform (DeLaval DelPro®). The MDi index used in our study also incorporate somatic cell count values. The specific influence of each variable on the MDi value is proprietary to DeLaval (DeLaval DelPro®).
For each flagged cow, a physical examination was performed by a trained veterinarian. Physical examination included the assessment of rectal temperature, heart rate, respiratory frequency, rumen movement and vulvar mucous, followed by udder examination: palpation, detection of lesions, changes in the colour and physical aspects of the milk (watery, small or large flakes etc.), and assessment of pain (Rosenberger, Reference Rosenberger, Dirksen, Gründer and Stöber1990). Cows that showed signs of CM were diagnosed with CM. All CM cases were monitored daily while the antibiotic/anti-inflammatory treatment was applied and then weekly through veterinary examinations. A case was considered clinically cured when no clinical signs of mastitis were observed in any lactating quarter for at least 7 days after onset (i.e. each case had a minimum of 7 days to cure).
Mastitis cases were classified into three scores based on severity (Wenz et al., Reference Wenz, Barrington, Garry, Dinsmore and Callan2001). Mild cases showed abnormal milk with clots, flakes and changes in colour and consistency but no external signs of udder inflammation. Moderate cases included abnormal milk and mammary gland inflammation, such as redness, heat, pain and swelling. Severe cases featured abnormal milk, mammary gland issues and signs of systemic illness; elevated body temperature, altered rumination, appetite and hydration (Rosenberger, Reference Rosenberger, Dirksen, Gründer and Stöber1990).
In this study, severe mastitis cases were also included, since their management followed a specific protocol established at the research unit (details below). Cows with severe mastitis initially received anti-inflammatory and analgesic therapy while awaiting the 24-hour culture results, and antimicrobial treatment was initiated earlier only when the attending veterinarian considered the animal’s systemic condition to be critically compromised (such cases were not included in the study).
During the study, an average of 58 cows was milked in the AMS. However, because the studied some cows were dried off and some entered the lactating herd a total of 65 cows had at least one CM case included in the study. Out of a total of 106 CM cases included, 78 were solved with a single treatment, while 28 required two or more treatments. Regarding the number of treatments: 78 CM cases required only one treatment, 21 CM required two treatments, 5 CM required three treatments and 2 CM required four treatments.
Eighty-one cases were assigned to the non-selective farm protocol and 25 cases were assigned to a selective treatment protocol based on results for the mastitis-causing pathogen obtained from chromogenic culture (details below). Case assignment was random, with no other specific criteria. A proportion of approximately 3 to 1 was used to increase the study’s power (see Dohoo et al., Reference Dohoo, Wayne and Stryhn2012).
The non-selective farm protocol was developed based on the severity of mastitis. The selection of antibiotics used in each protocol was based on the farm’s history and the results of on-farm microbiological culture, along with the antibiogram performed at a commercial laboratory every 6 months. Mild cases were treated exclusively with the anti-inflammatory NSAID – injectable Meloxicam, 2% (Maxicam 2%, Ouro Fino; dosage: 0.5 mg/kg) for 3 days. Mild cases were typically associated with Gram-positive bacteria. Moderate cases were treated with intramammary antibiotics for 3 days, using the commercial product Ubrecilin® (Vansil), composed of sodium cloxacillin, and NSAID – Meloxicam, 2% (Maxicam 2%, Ouro Fino) for 3 days. Moderate cases were typically associated with Gram-positive bacteria. Severe cases were treated with the intramammary antibiotic Mastijet Forte® (MSD), composed of tetracycline 200 mg, neomycin 250 mg, neomycin sulphate 365 mg, bacitracin 2000 IU for 5 days. Severe cases were typically associated with Gram-negative bacteria. There were cows that, even after receiving the standard protocol treatment, still showed signs of discomfort, pain and abnormalities in the physical exam. In such cases, there was the possibility of including analgesic medication, such as dipyrone (Sodium Dipyrone 500 mg; dosage: 25 mg/kg) for 3–5 days, or extending the duration of anti-inflammatory treatment for up to 7 days with NSAIDs, such as Meloxicam 2% (Maxicam 2%, Ouro Fino).
The rapid OnFarm® microbiological test used in this study was a validated chromogenic culture test (Granja et al., Reference Granja, Fidelis, Garcia and Santos2021) that is commercially available in Brazil (OnFarm®, Rúmina, Brazil) and used on-farm for mastitis-causing pathogen diagnosis and to aid in mastitis treatment decisions. To collect the sample for the chromogenic culture, the teat of the infected quarter was disinfected, dried, and three streams of milk were discarded. Then, the milk sample (10 ml) was collected in a sterile tube and inoculated onto the plate using a sterile swab. The plate was then incubated in a controlled temperature oven at 36.6°C. After 24 h, the plate was read using the manufacturer's software by taking a photograph of the plate in an illuminated chamber. The software analysed the colour of each colony that grew in each section of the plate and suggested the most likely pathogens (Granja et al., Reference Granja, Fidelis, Garcia and Santos2021)
OnFarm® uses a tri-plate with selective media for the main mastitis pathogens (13 specific bacteria). The first section of the plate allowed the growth of the following bacteria: Enterococcus spp., Lactococcus spp., Streptococcus uberis, Streptococcus agalactiae, or Streptococcus dysgalactiae, and other gram-positive bacteria. The second section allowed the growth of the following bacteria: Klebsiella spp. or Enterobacter spp., Escherichia coli, Pseudomonas spp., Serratia spp., Prototheca spp. The third section allowed the growth of Staphylococcus aureus and non-aureus staphylococci and mammaliicocci (NASM).
Antibiotics selection after reading the plates took into consideration the results of routine antibiogram evaluations conducted on the farm (every 6 months). Therefore, a microbiological milk culture and an antibiogram were performed biannually; thus, the antibiotics chosen were the ones that showed effectiveness against the pathogens in these routine antibiograms. This protocol decision was adopted because, in certain cases of mastitis, the severity of clinical signs, along with the impact on milk production and the cow's health, was so significant that waiting for the time required to perform a conventional laboratory culture and antibiogram for each case becomes unfeasible. Thus, using the farm’s historical data on pathogens and effective antibiotics was a way to expedite the treatment process and reduce the impact on the cow’s health.
Statistical analysis
Statistical analyses were performed in RStudio version 1.2.5019 using R 3.5.3 (R Core Team, 2019).The following variables were categorized: parity was divided into two categories [primiparous and multiparous], days in milk (DIM) divided into three categories [DIM_1, DIM_2 and DIM_3; 0 to 100, 101 to 200 and ≥200 DIM, respectively]; mastitis severity was categorized into three categories [mild, moderate and severe]; number of treatments per case was categorized into one treatment or two or more treatments.
We used a seven-day period to consider the duration of each case of mastitis, as evaluations were conducted weekly. For cases that took >7 days to cure, the number of weeks until the last veterinary assessment before cure was noted and converted to number of days. To measure the association of use of the on-farm test with CM time to cure, a linear regression mixed model was used (Bates et al., Reference Bates, Mächler, Bolker and Walker2015). Likelihood of receiving two or more treatments (i.e. completing more than one treatment protocol within the same case) before CM cure was assessed by mixed logistic regression; results are presented as odds ratio and 95% confidence interval. Confounding variables were selected based on causal diagrams and included as fixed variables in the models: parity, DIM, mastitis severity and whether the cow had a previous case of mastitis in the dataset.
Results
The 25 CM cases treated with a selective protocol including information from the chromogenic culture are presented in Table 1. The time-to-cure of CM cases ranged from 7 to 35 days, averaging 9.34 ± 5.34 days. The use of the on-farm test was associated with a 2.78-day reduction in CM time-to-cure compared to cases in animals treated using the standard mastitis treatment protocol (without chromogenic culture) (Table 2). Primiparous cows’ cases were 1.77 days shorter compared with cases in multiparous cows. There was a trend for parity to be associated with CM time to cure (Table 2). There was a difference among the three categories of mastitis severity, with severe cases of mastitis lasting longer compared to mild and moderate cases. Cows with prior mastitis cases tended to experience an average increase of 2.75 days (p = 0.07) in recovery time compared to cows encountering their first case. The annual mean data for this experimental herd showed a sub-CM prevalence of 10% and a CM prevalence of 4% per year.
Culture analysis in 106 mastitis cases from dairy cows on an AMS-freestall dairy herd in Fazenda Rio Grande, Brazil. Cases are categorized based on the number of detected pathogens (one, two, or three) and the specific pathogen combinations. The severity of mastitis cases (mild, moderate, or severe) is also summarized. One case was excluded from the table due to being considered a contaminated result (Klebsiella spp. or Enterobacter spp., NASM and other gram-positive bacteria)
Least square means (LSM) from the linear mixed regression model assessing clinical mastitis (CM) time-to-cure (days) of 106 cases of 65 Holstein dairy cows on an AMS-freestall dairy herd in Fazenda Rio Grande, Brazil
a Standard error. bConfidence interval.
The results from the model built to measure the association of use of the on-farm culture test and the covariates and the number of antimicrobial treatments before clinical cure are presented in Table 3. In brief, the use of the on-farm culture test was not associated with the number of antimicrobial treatments before cure. Primiparous cows were less likely to be treated two or more times before cure compared to multiparous cows. DIM and having a previous case of mastitis were not associated with the likelihood of being treated two or more times before a cure. Mastitis severity was associated with the likelihood of being treated two or more times before cure; less severe cases were less likely to be treated more than once for CM.
Parameters from the models evaluating the likelihood of being treated two or more times before cure. Data from 106 cases of 65 Holstein dairy cows on an AMS-freestall dairy herd in Fazenda Rio Grande, Brazil
a Reference category. bConfidence interval.
Discussion
The use of the on-farm culture test to support CM case treatment decision was associated with reduced CM case duration. Even with the 24 h wait for culture results, which by default delayed antibiotic treatment, on-farm testing reduced time-to-cure by two to three days. However, we emphasize that antibiograms should be performed periodically throughout the year, rather than for every case, to evaluate whether the antibiotic treatment protocol remains effective or if it may be contributing to antimicrobial resistance, in which case the active compounds used should be reconsidered.
A study using similar methods compared the use of an on-farm culture test to a farm standard protocol and concluded that it reduces the use of intramammary antibiotics by half and tended to decrease milk retention time by 1 day (Lago et al., Reference Lago, Godden, Bey, Ruegg and Leslie2011). However, they found no significant differences in days to clinical cure, risk of bacteriological cure and risk of new intramammary infection. In another study Vasquez et al. (Reference Vasquez, Nydam, Capel, Eicker and Virkler2017), time to cure was also not affected by the treatments chosen based on the results of a rapid on-farm diagnostic test, however the study only included cases of mild and moderate CM. In our study, the low number of cases in each severity category does not allow for assessing the effect by category, but it is likely that our results are skewed by including severe cases of CM.
In fact, the inclusion of severe mastitis cases may be questioned, as treatment in such cases is often initiated before culture results are available. In this study, however, their inclusion was based on a management protocol that allowed the use of rapid on-farm culture with 24-h results. Cows initially received anti-inflammatory and analgesic therapy, and antibiotics were administered earlier only when the veterinarian considered the systemic condition critically compromised. This approach enabled the evaluation of selective therapy even in severe cases, providing a more comprehensive assessment of rapid culture applicability.
As expected, we found that primiparous cows tended to have shorter CM cases and were more likely to cure in the first treatment compared to multiparous cows. This may be due to age that being associated with a weaker immune system (Gilbert et al., Reference Gilbert, Gröhn, Miller and Hoffman1993). Similar to previous research, the majority of CM cases were of mild to moderate severity (Oliveira et al., Reference Oliveira, Hulland and Ruegg2013). All factors related to cows, such as higher parity, lactation stage and a history of mastitis, showed the expected negative association with cytological cure, consistent with findings from prior studies (Pinzón-Sánchez and Ruegg, Reference Pinzón-Sánchez and Ruegg2011).
A prior history of mastitis is a consistent predictor of longer CM case duration and of recurrence of CM episodes (Schukken et al., Reference Schukken, Bar, Hertl and Gröhn2010; Barkema et al., Reference Barkema, Green, Bradley and Zadoks2009). The mechanisms underlying reinfection and prolonged case duration are not fully understood, largely because CM is a multifactorial disease, but they are thought to relate to pathogen persistence within the mammary gland (Swinkels et al., Reference Swinkels, Lam, Green and Bradley2013) and to an ineffective local immune response at the mammary gland level, which delays infection resolution and increases the likelihood of repeated treatments (Wenz et al., Reference Wenz, Garry, Lombard, Elia, Prentice and Dinsmore2005; Abureema et al., Reference Abureema, Smooker, Malmo and Deighton2014). Other factors also influence CM recurrence, such as parity (older cows) and cows within the first weeks of lactation, which are associated with a higher risk of CM (Elghafghuf et al., Reference Elghafghuf, Dufour, Reyher, Dohoo and Stryhn2014). Although in our study we were not able to retrieve the full history of mastitis cases of the enrolled cows our results indicate that having a second case during the study tended to be positively associated with longer time to cure.
A higher frequency of CM cases included in this study were caused by gram-negative pathogens, with the commonly isolated pathogens being Klebsiella spp., Enterobacter spp. and Escherichia coli. The pathogen distribution observed in this study is similar to that of other studies that included milk samples from CM cases on dairy farms in the United States (Lago et al., Reference Lago, Godden, Bey, Ruegg and Leslie2011). These data corroborate the severity results of mastitis cases, where there was a higher proportion of moderate and severe mastitis cases (40% and 14% of cases, respectively). This pattern can be explained by the fact that the isolated pathogens are often associated with environmental mastitis cases, where the environment represents the primary source of contamination for the cows (Bradley, Reference Bradley2002).
However, the presence of these pathogens in microbiological cultures, particularly E. coli, does not directly assist in treatment, as the efficacy of treatment for cows with mastitis caused by E. coli is debated due to its high spontaneous cure rate (Lago et al., Reference Lago, Godden, Bey, Ruegg and Leslie2011). Some studies have reported no difference in the bacteriological cure rates for untreated cows compared to those treated for mastitis caused by gram-negative pathogens, and most antimicrobials labelled for treating mastitis have limited activity against these organisms (Harris et al., Reference Harris, Lane and Swanson2021; Ruegg, Reference Ruegg2017).
Additionally, it was found that 48% of mastitis cases were related to mild cases of mastitis. These cases may be associated with gram-positive pathogens, as according to Ruegg (Reference Ruegg2017), gram-positive pathogens typically do not cause severe clinical signs. Thus, classifying the severity of the disease is crucial for developing effective treatment protocols and assisting farmers in identifying cows at higher risk of developing bacteraemia and determining the need for supportive therapy (Wenz et al., Reference Wenz, Barrington, Garry, Dinsmore and Callan2001). Consequently, it is expected that the reduction in antibiotic use associated with on-farm culture test for the selective treatment of CM will lead to decreased antibiotic use and a reduction in the potential for antimicrobial residue occurrences (Lago et al., Reference Lago, Godden, Bey, Ruegg and Leslie2011).
Additionally, the study did not focus on specific bacterial strains associated with bovine mastitis which may have different clinical presentation, case length and severity. Moreover, the OnFarm test (Granja et al., Reference Granja, Fidelis, Garcia and Santos2021) does not differentiate certain pathogens, such as S. agalactiae and S. dysgalactiae, which possess different pathogenic characteristics but are identified as group (Harris et al., Reference Harris, Lane and Swanson2021). Consequently, the identification of the range of mastitis-causing pathogens and their pathogenicity in our study remains limited. Although the use of chromogenic culture has reduced the turnaround time, the study did not include a comparative analysis with other diagnostic methods or treatments, such to per-form monthly antibiograms, which could provide a better basis for antibiotic selection.
This study has some limitations that should be considered. The relatively low number of cases limited the possibility of stratifying outcomes by severity category. The chromogenic on-farm culture test employed does not differentiate certain pathogens with distinct pathogenic characteristics, such as S.agalactiae and S. dysgalactiae. The lack of routine antibiogram comparisons restricted the evaluation of antimicrobial resistance dynamics. Despite these limitations, the results provide valuable insights into the feasibility and applicability of rapid culture-based selective therapy under field conditions.
Further research with more herds and animals is encouraged to enhance external validity of our results. The assessment of production and behavioural parameters, along with CM data, is vital to understand the impact of microbiological-based treatment decisions on production recovery and animal comfort levels. Developing smart monitors for predicting mastitis cure using clinical, production and behavioural data is also seen as highly valuable.
In summary, using the information provided by a chromogenic culture methodology for detecting mastitis in an AMS herd reduced the mastitis case duration by 2.78 days compared to treatment without chromogenic culture. Further work with a larger number of animals and herds is needed to confirm the positive impact of chromogenic culture as a decision-making tool for CM management.
Acknowledgements
We thank the staff of the Experimental Farm Gralha Azul at PUCPR for their cooperation throughout the study.
Funding statement
The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.
Competing interests
The authors declare no conflicts of interest.
Author contributions
INHP, AO and RRD contributed to the study conception and design. Material preparation and data collection were performed by INHP, ISP and LVC. Data analyses were performed by RRD. The first draft of the manuscript was written by INHP and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. Translation to English was done by RRD.
Institutional review board statement
This study was approved by the Animal Use Ethics Committee of the Pontifical Catholic University of Paraná (PUCPR) – protocol 020343.
Data availability statement
The datasets generated during the current study are available from the corresponding author on reasonable request.
