INTRODUCTION
Mastitis, a ubiquitous disease with high incidence and prevalence worldwide, causes direct and indirect damage to the dairy industry, including decreased milk production and quality, discarded milk, additional costs for diagnostic tests and treatment, culling, death, and animal welfare problems [1]. Mastitis can be categorized into clinical and subclinical forms according to the visible milk condition of the mammary quarters [2,3]. Clinical mastitis, ranging from mild (abnormal milk only) to severe (abnormal milk, abnormal mammary glands, and systemic clinical signs), is characterized by visibly abnormal milk from the mammary quarters. Subclinical mastitis is characterized by normal-appearing mammary glands and visibly normal milk despite decreased milk production, lowered milk quality, and high somatic cell counts (SCC). Subclinical mastitis is approximately 40 times more prevalent than clinical mastitis, owing to the difficulty in its detection [4].
Field-applicable diagnostic methods have been used to detect subclinical mastitis on dairy farms. The California mastitis test (CMT) is conventionally used in farms because it is rapid, economical, and practical [5]. Automatic digital devices have been newly developed for on-farm subclinical mastitis detection [6]. Infrared thermography is effective in detecting quarters with subclinical mastitis [7]. Automatic milking systems (AMS) detect the occurrence of subclinical mastitis at an early stage by measuring SCC, milk electrical conductivity (EC), and lactate dehydrogenase levels in milk [7–9].
Previous studies have demonstrated an association between subclinical mastitis and hematological and serum biochemical parameters in dairy cows [10–13]. Studies have shown that dairy cows with subclinical mastitis have increased neutrophil and aspartate aminotransferase levels and decreased total protein levels [10,11,13]. However, findings regarding serum calcium and phosphorus levels differ [10, 12,13]. In addition, other hematological and serum biochemical parameters have not been investigated in association with subclinical mastitis.
Milk produced by quarters with subclinical mastitis has a composition different from that of milk produced by healthy quarters in cattle. Cows with subclinical mastitis produce milk with decreased lactose, solids-not-fat, calcium, iron, phosphorous, potassium, and zinc levels [10,14–17]. However, previous studies have reported varying results regarding the levels of fat, protein, and total solids in the milk of cows with subclinical mastitis [10,14,15,17].
The aims of this study were to explore changes in factors monitored via automatic health monitoring and AMS before the incidence of subclinical mastitis and to identify differences in hematological and serum biochemical parameters and milk composition between cows with subclinical mastitis and healthy cows. We hypothesized that unusuality monitored via automatic health monitoring and AMS may occur in cows with subclinical mastitis prior to the onset of the disease, and that blood and milk compositions in cows with subclinical mastitis may differ from those in healthy cows.
MATERIALS AND METHODS
Ethics approval
This work was approved by the Institutional Animal Care and Use Committee (IACUC) at the National Institute of Animal Science, the Republic of Korea (approved numbers: NIAS-2020127 and NIAS-2022102). All experimental procedures involving animals were conducted in strict accordance with relevant guidelines and regulations. All methods used for the present study were in accordance with REFLECT (Reporting Guidelines for Randomized Controlled Trials for Livestock and Food Safety) guidelines.
Animals and case definition
This observational study was conducted between February 2022 and May 2023 on a farm of the study’s host institution in Southeast Asia, in order to examine the factors associated with subclinical mastitis by analyzing the data from AMS and analyses of blood and milk samples. In total, 32 Holstein cows were included. The animals were raised in a loose barn with sawdust bedding that housed around 30 dairy cows at a time. They were milked on a Lely Astronaut A4 milking robot (Lely Industries NV, Maassluis, the Netherlands) with free cow traffic and fed total mixed rations comprising concentrates, soybean meal, corn silage, alfalfa hay, timothy hay, enzyme, minerals, and vitamin additives ad libitum.
We defined subclinical mastitis using three procedures: the alarms from the AMS, results from the CMT, and SCC obtained from milk composition analyses. The threshold for mastitis used in this study was 500×103 cells/mL since the AMS triggers an alert to indicate a mammary health problem when the system detects milk with 500×103 cells/mL. In addition, the cutoff level of 500×103 cells/mL is utilized when employing the CMT [18]. We examined alerts from the AMS (Lely Astronaut A4 milking robot) at 9 A.M. Once the AMS (Lely Astronaut A4 milking robot) alarmed cows with >500×103 cells/mL, we performed the CMT on milk obtained from a mammary quarter. Among cows of similar age, parity, and days in milk in the same herd, we randomly selected a control cow with <500×103 cells/mL and negative CMT results. Milk samples were immediately analyzed with a milk composition analyzer (Combiscope FTIR 300 HP; Delta Instruments B.V., Drachten, the Netherlands). Cows with <500×103 cells/mL, as measured by the AMS and milk composition analyzer, and negative CMT results were categorized in the control group (CON; n = 16), whereas those with >500×103 cells/mL and positive CMT results were categorized in the subclinical mastitis group (SCM; n = 16).
Measurement of body temperature and sample collection
On completion of CMT, body temperature was measured in the rectum using a thermometer. Then, blood and milk samples were obtained. Blood was drawn from the jugular vein and collected using ethylenediaminetetraacetic acid and serum-separating tubes from 32 cows (CON and SCM). Milk samples were obtained manually with clean hands using disposable conical tubes from each quarter in the morning. We obtained milk samples from the CON group before obtaining samples from the SCM group. In total, 121 milk samples (milk samples from the control group [mCON], n = 60; milk samples from the SCM group [mSCM], n = 61) were obtained from the groups, as four cows in the CON and three cows in the SCM groups had one non-functional quarter of the udder. In addition, mSCM (n = 61) was categorized as samples from non-inflamed quarters ([mNQ], n = 34, <500×103 cells/mL) and those from subclinically-inflamed quarters ([mIQ], n = 27, >500×103 cells/mL). The samples were immediately sent to a laboratory located on the farm.
Data collection from automatic systems
We collected data from the automatic health monitoring and AMS from 3 days prior to the incidence of subclinical mastitis. The rumination time and activity were measured using tags attached to the collars (HR-Tag; SCR Engineers Ltd., Netanya, Israel). Milk temperature, yield, and EC were measured using the Lely Astronaut A4 milking robot. Milk temperature and yield were measured for individual cows, whereas EC of the udder quarter was measured separately. Milk yield was aggregated daily. Milk yield per udder quarter was calculated by dividing the milk yield by the number of milk-producing udder quarters.
Laboratory analyses of blood and milk
The complete blood count (CBC) test and serum biochemical and mineral analyses were performed on 32 dairy cows (CON, n = 16; SCM, n = 16) as previously described [19]. The CBC profile included three types of parameters: erythrocytes (red blood cell count, hematocrit, hemoglobin, mean corpuscular volume, mean corpuscular hemoglobin, mean corpuscular hemoglobin concentration, red cell distribution width, and reticulocyte count); leukocytes (white blood cell, neutrophil, lymphocyte, monocyte, eosinophil, and basophil counts); and platelets (platelet count, mean platelet volume, platelet distribution width, and plateletcrit). The levels of glucose, non-esterified fatty acids, triglyceride, total cholesterol, total protein, albumin, blood urea nitrogen (BUN), creatinine, alanine transaminase, aspartate transaminase, alkaline phosphatase, gamma-glutamyl transferase, lactate dehydrogenase, creatine kinase, calcium, magnesium, and inorganic phosphorus were measured. The globulin level was calculated by subtracting the albumin level from the total protein level. The albumin-to-globulin (A/G) ratio was calculated by dividing the albumin level by the globulin level.
Analysis of 121 milk samples (mCONs, n = 60; mNQ, n = 34; mIQ, n = 27) was conducted to measure the SCC and levels of fat, protein, lactose, citrate, total solids, non-protein nitrogen, and solids-not-fat using a Combiscope FTIR 300 HP.
Bacteria isolation in milk samples from subclinically inflamed quarters
Bacterial examination for bovine mastitis from the sample with the SCC ≥500×103 cells/mL using a milk composition analyzer (Combiscope FTIR 300 HP; Delta Instruments B.V., the Netherlands) was conducted using standard laboratory techniques. Bacterial isolation in milk samples was performed as previously described [20]. Briefly, 10 μL of milk samples were streaked on 5% blood agar plates (Komed, Seongnam, Korea) and incubated at 37°C for 48 h. After incubation, colonies were identified by matrix-assisted laser desorption ionization–time of flight mass spectrometry (bioMérieux, Marcy l’Etoile, France), according to the manufacturer’s instructions.
Statistical methods
Statistical analyses were performed using the Statistical Package for the Social Sciences software (version 27.0; IBM Corp., Armonk, NY, USA). The Shapiro–Wilk and Levene tests were used for normality analysis and equality of variances for the independent t-test, paired t-test, and one-way analysis of variance (ANOVA). The Mann–Whitney U with Bonferroni’s method, Wilcoxon signed-rank test, and Kruskal–Wallis test were used for parameters that did not satisfy the normality analysis or equality of variances. A generalized linear mixed model with the Bonferroni correction was used to evaluate repeated measurements of milk temperature, yield, and EC. In the generalized linear mixed model with the Bonferroni correction, time and group were fixed effects, whereas cows nested within the group were random effects. Multiple comparison analyses among the groups were conducted using the Bonferroni post-hoc test for one-way ANOVA and the Mann–Whitney U test with Bonferroni’s method for Kruskal–Wallis tests. Data are expressed as the mean± standard deviation. A p-value of <0.05 was considered statistically significant. Statistical significance for the Mann–Whitney U test with Bonferroni’s method for the Kruskal–Wallis test was set at p<0.017 (0.05/3) for all three groups. Statistical significance for the Mann–Whitney U test with Bonferroni’s method for the Kruskal–Wallis test was divided by three, since the number of comparing the groups were three ways (mCON-mNQ, mCON-mIQ, and mNQ-mIQ).
RESULTS
Association of subclinical mastitis with data from automatic health monitoring and automatic milking systems
The CON group had age (5.2±1.6 years), parity (2.5±1.0), days in milk (146.7±89.5 days), and body temperature (38.9°C± 0.4°C) similar to those in the SCM group (5.3±2.0 years, 2.4±1.5, 176.2±112.1 days, and 39.0°C±0.4°C, respectively) (Table 1). Daily milking frequency and body weight did not differ between the CON and SCM groups. The rumination time, activity, and milk temperature in the CON group did not differ from those in the SCM group (Figure 1). Regarding milk yield, the CON group maintained higher milk production than did the SCM group (p<0.05). In the SCM group, milk yield per udder quarter decreased on day −1 (6.76±2.44 kg/d) and the day of incidence (6.50±2.06 kg/d) compared to that on day −3 (7.70±1.97 kg/d) (p<0.05). Regarding EC, the mIQ group showed a higher EC (70.5±3.9 unit) than did the mCON group (68.5±3.3 unit) 2 days prior to the incidence (p<0.05). The mNQ group had decreased EC from day −2 (68.8±3.3 unit) to the day of incidence (68.1±2.8 unit), whereas the mIQ group had increased EC from day −3 (69.8 ±3.2 unit) to the day of incidence (73.4±4.7 unit). On day −1 and the day of incidence, the mIQ group showed differences in EC compared to the mCON and mNQ groups (p<0.003).
Association of subclinical mastitis with hematological and serum biochemical parameters
The SCM group differed from the CON group in terms of hematological and serum biochemical parameters (Figure 2; Supplementary Figure S1 and S2). The SCM group had lower basophil count (0.009±0.009 K/μL); total cholesterol (186.1± 25.6 mg/dL), albumin (3.78± 0.27 g/dL), and BUN (17.6±3.7 mg/dL) levels; and A/G ratio (0.91±0.14) than did the CON group (0.019±0.014 K/μL, 219.4±36.1 mg/dL, 4.01±0.27 g/dL, 20.3±2.1 mg/dL, and 1.06±0.14, respectively) (p<0.05). However, the SCM group had a higher level of globulin (4.23±0.48 g/dL) than did the CON group (3.83± 0.38 g/dL) (p<0.05).
Association of subclinical mastitis with milk composition
The SCC, lactose, non-protein nitrogen, and milk solids-not-fat were associated with subclinical mastitis, whereas no association was found between subclinical mastitis and milk fat, protein, citrate, or total solids (Figure 3). Milk SCC was the lowest in the mCON group (97.8±118.5×103 cells/mL), followed by those in the mNQ (167.0±133.8×103 cells/mL) and mIQ (3,561.0±2,612.3×103 cells/mL) groups (p<0.017). The mIQ group had lower levels of lactose (3.58%±0.77%) and milk solids-not-fat (7.39%±0.83%) than did the mCON (4.66%±0.32% and 8.40%±0.61%, respectively) and mNQ (4.61%±0.30% and 8.38%±0.64%, respectively) groups (p< 0.017). In addition, the mCON group had higher levels of milk non-protein nitrogen (10.7±2.2 mg/100 g) than did the mNQ group (9.3±2.6 mg/100 g) (p<0.003).
Bacteriological findings from the udder quarter affected by subclinical mastitis
Mastitis pathogens were isolated from the mIQ group (Table 2) and no bacteria were detected in eight samples. In the remaining samples, microorganisms that normally inhabit the udders and are derived from the environment in which the cows live were isolated. The most common finding (n = 9, 33.3%) was coagulase-negative Staphylococci, followed by Corynebacterium spp. (n = 3, 11.1%), Streptococcus spp. (n = 3, 11.1%), and Psychrobacter pasteurii (n = 2, 7.4%). Among coagulase-negative Staphylococci, S. simulans (n = 4, 14.8%) was the most common pathogen, followed by S. epidermidis (n = 2, 7.4%) and S. xylosus (n = 2, 7.4%).
DISCUSSION
In the present study, we demonstrated the characteristics of automated sensing data, hematological and serum biochemical parameters, and milk composition detected using AMS, the CMT, and milk composition analyses in dairy cows with subclinical mastitis. Subclinical mastitis was significantly associated with milk yield, EC, basophil count, A/G ratio, and levels of total cholesterol, albumin, globulin, BUN, milk SCC, milk lactose, milk non-protein nitrogen, and milk solids-not-fat. Furthermore, opportunistic or environmental mastitis pathogens were isolated from milk samples collected from the udder quarters of cows with subclinical mastitis.
Methods to identify cows with subclinical mastitis and categorize their behaviors have been developed using AMS and automatic health monitoring systems. AMS indicate subclinical mastitis by detecting decreased milk yield and abnormal EC from quarters or repeated measures between milkings [21–23]. In the present study, we identified decreased milk yield in the SCM group, increased EC in the mIQ group compared to that in the mCON group, and reduced EC in the mNQ group prior to the incidence of subclinical mastitis. An interesting finding of this study is that milk had a lower EC on the day of incidence of subclinical mastitis than a few days before the incidence in mNQ. The increased EC produced in inflamed quarters can be attributed to an increase in Na+ and Cl− ions in milk owing to the leaky tight junctions in mammary epithelial cells [24]. Decreased concentrations of Na+ and Cl− ions in the blood due to the leakage of Na+ and Cl− ions in inflamed quarters may result in reduced EC in milk from mNQ. Further studies are required to elucidate why mNQ produce milk with decreased EC compared to that in the milk usually produced. In terms of automatic health-monitoring systems, a previous study reported that cows with subclinical mastitis spend less time ruminating [25]. However, in the present study, the rumination time between the CON and SCM groups did not differ, possibly because of the devices used. The device used in the present study recorded the rumination time by detecting sounds on the neck, whereas the device used in the previous study was based on the pressure of the jaw movements of cows [26,27].
Previous studies have reported increased white blood cell counts, particularly neutrophil counts, in dairy cows with subclinical mastitis [11,28,29]. However, no differences were observed in white blood cell and neutrophil counts between the CON and SCM groups in this study. The difference may be due to different experimental methods used in the studies. Other studies have used the CMT to classify cows with subclinical mastitis. However, in this study, we first detected subclinical mastitis using AMS and then determined the disease incidence using the CMT and SCC via milk composition analysis. The use of AMS and the CMT may help identify cows at an earlier stage of the disease than does the use of the CMT alone. This may have influenced the basophil counts in this study. The basophil count in the SCM group was lower than that in the CON group, although it was within the reference range and the difference was small. Decreased basophil counts are not considered clinically significant because cattle have very low basophil counts [30]. However, basophils are also involved in mastitis. During intramammary inoculation with Escherichia coli, dairy cows with lower circulating basophil counts reportedly developed severe mastitis [31]. Basophils might also be involved in subclinical mastitis. Further studies are required to elucidate the association between basophils and subclinical mastitis when detected by AMS.
Holstein cows respond systemically to subclinical mastitis, reflecting specific blood biochemical parameters related to inflammation and liver functions [32,33]. The inflammatory status may have contributed to the albumin and globulin levels and A/G ratio observed in this study. A decreased A/G ratio is observed when an impaired blood–milk barrier causes an increase in serum albumin levels in the milk, and globulin production is elevated in response to inflammation in cows with subclinical mastitis [33,34]. Moreover, decreased hepatic function can also influence the albumin concentration and A/G ratio, leading to decreased levels [35,36]. In addition, the total cholesterol and BUN levels indicate that subclinical mastitis may be associated with decreased hepatic functions. Cholesterol and BUN are mainly produced by the liver; therefore, decreased hepatic function results in decreased synthesis of cholesterol and BUN [35,36]. However, hepatic damage may not be associated with subclinical mastitis because hepatocellular leakage enzymes, including alanine aminotransferase, aspartate aminotransferase, and lactate dehydrogenase, which are sensitive indicators of hepatic damage, did not differ between the CON and SCM groups in this study [35].
The associations of subclinical mastitis with milk fat, protein, and total solids in previous studies [10,14,15,17] were not significant in the present study. As noted in previous studies, levels of milk lactose and solids-not-fat were lower in the mIQ group than in the mCON and mNQ groups [37,38]. Considering the differences in milk lactose and solids-not-fat between the groups, decreased lactose biosynthesis due to tissue damage in secretory cells may result in decreased levels of milk solids-not-fat [2,38]. Milk non-protein nitrogen is transferred from the blood to milk and consists of milk urea (about 50%), free amino acids, creatine, uric acid, peptides, organic acids, and phospholipids [39,40]. Interestingly, in our study, although the CON group had higher BUN levels than did the SCM group and the mCON group had higher milk non-protein nitrogen levels than did the mNQ group, no difference was observed between the mCON and mIQ groups. These findings indicated that the degree of inflammation may influence the transfer of urea nitrogen from the blood to milk.
Milk production in dairy cows can be influenced by various factors such as genetic potential, physiological status, environment, and management systems. We attempted to reduce the differences between the CON and SCM groups. We selected CON cows that had characteristics such as age, parity, and days in milk similar to those of Holstein cows raised in the same barn using the same AMS at the same time when subclinical mastitis was detected. Although milk production in the SCM group decreased due to subclinical mastitis, milk production in this group was lower than that in the CON group, regardless of the presence of the disease. A limitation of our study is that some factors that may influence milk production might not have been estimated. Future studies designed to overcome the limitations of the present study are required.
The present study was based on AMS using EC and milk production, which can help early detection of subclinical mastitis. Comparison of automated sensing data, hematological and serum biochemical parameters, and milk components between CON and SCM may provide insights into understanding the early stage of subclinical mastitis.
In conclusion, the present study demonstrated differences in subclinical mastitis using AMS. Milk yield and EC, as measured using AMS, changed before subclinical mastitis detection using SCC. On the day of disease incidence, Holstein cows with subclinical mastitis had reduced basophil count, A/G ratio, and levels of total cholesterol, albumin, and BUN and elevated globulin levels. Moreover, quarters with subclinical mastitis produced milk with elevated milk SCC and reduced milk lactose and solids-not-fat. AMS aid in the early detection of subclinical mastitis. Moreover, comparisons of automatic sensing data, hematological and serum biochemical parameters, and milk compositions between the CON and SCM groups may provide insights into the early stages of subclinical mastitis. However, further studies are required to elucidate how these variables change according to the subclinical mastitis stage and severity.
PDF Links
PubReader
ePub Link
Full text via DOI
Download Citation
Supplement1
Print
