Ethical statement
This study was approved by the Board of Advanced Studies and Research (BASR) of the National University of Sciences and Technology (NUST), Islamabad. In addition, written consent was taken from the farm owners before sampling. It was made sure that milk samples for microbiological analysis were only taken from the routine milking performed at the farms, and besides milking no experimental work was conducted on the farm animals.
Study area
The study was conducted from March 2019 to March 2020 in dairy herds of the Rawalpindi district. Rawalpindi district is in the Potohar plateau in the northernmost part of the Punjab province of Pakistan. It lies between longitudes 73° 15′ 0″ east and latitude 33° 20′ 0″ north. The area of the district is around 5,286 km2 (2,041 sq mi) and is situated on the southern slopes of the north-western boundaries of the Himalaya mountain series. It is rich with mountain tracts, valleys, and rivers and is known for its moderate climate due to its closeness to the foothills (31). Rawalpindi district has seven autonomous tehsils; including Rawalpindi, Taxila, Kahuta, Murree, Gujar Khan, Kallar Syedan, and Kotli Sattian. Rawalpindi city is the capital of the district and is the 4th most populous metropolitan city of Pakistan. Punjab province is the most livestock-populated province of Pakistan (32). According to Veterinary Research Institute (VRI); the population of cows in the Rawalpindi district is around 252,298 cows, out of which 104,950 (40% of the total population) are lactating cows.
Farm type and management practices
A total of 40 farms were randomly selected from the study area. The herd size of the selected farms varied from 1 to 150 animals, of which 1 to 100 were lactating. Herd type included both single type (15%) and mixed herds (85%). For management of the farm, 34 (85%) of the herds were group barns and only 4 (10%) of the herds had bedding material. Around 52.5% of herds (21) were managed intensively and rest were managed semi-intensively. In intensive management system, animals are kept indoors all the time and given concentrated foods including crop residues and hay, etc while in semi-intensive management system, animals are free to graze in pastures and given supplementary food only in morning and evening, before milking). Moreover, all of the farms included in the study were also checked for milking method, hand and udder washing, floor type, feed sharing, post teat dipping, hygiene of the farms, use of hormones, tick control activity, manure removal, and others (Table 3) before performing mastitis detection test on cows.
Sample size determination
The sample size for the desired study was calculated according to Thursfield (33) at 5% precision, 95% confidence interval, and with 23% expected mastitis prevalence of cows (According to PRI, Rawalpindi) in the Rawalpindi district.
Based on the formula, the total number of lactating cows required for the prevalence study was 272 cows. But, for better and precise results, 432 cows and 1728 teat quarters were considered in the study.
Data collection
A questionnaire comprised of close-ended questions was used to collect the data about various animal and herd-level factors that are thought to be associated with the occurrence of mastitis in cattle. A total of 28 factors were analyzed in this study and were divided into herd-level and animal-level factors (Table 1). Leg and udder hygiene score was recorded in the form of slightly dirty, moderately dirty, and very dirty (34) (Table 2).
Table 1
Various herd-level and animal-level factors thought to be associated with the occurrence of mastitis
Sr. No
|
Associated with Mastitis
|
Description
|
1.
|
Herd-level Factors
|
1. Herd location, (2) Herd Size, (3) Herd type, (4) Use of bedding material, (5) Milking practices (Machine/Manual), (6) Frequency of milking, (7) Housing, (8) Floor type, (9) Herd management system, (10) Washing practices of udder, (11) Drying of udder before milking, (12) Use of towel, (13) Use of hormones, (14) Pre/post teat dipping in sanitizing agents, (15) Standing position after milking, (16) Milking mastitic cow last, (17) Sharing of feed, (18) Manure removal routine
|
2.
|
Animal-level Factors
|
2. Breed, (2) Age, (3) Lactation stage, (4) Udder position, (5) Udder condition, (6) Presence of ticks, (7) Teat end lesion, (8) History of mastitis, (9) Leg and udder hygiene score, (10) Number of people attending cow
|
Table 2
Criteria to assess the leg and udder hygiene score of cattle
Sr. No
|
Criteria
|
Description
|
1.
|
Slightly dirty
|
2-10% area of the hind legs and udder is covered in dirt
|
2.
|
Moderately dirty
|
10-30% area of the hind legs and udder is covered in dirt
|
3.
|
Very dirty
|
>30% area of the hind legs and udder is covered in dirt
|
Clinical examination of cow udder and milk for clinical mastitis
During farm visits, clinical examination of milk and cow udder was conducted. For milk examination; milk from each quarter was separately examined by visual inspection for the presence of any blood clots, flakes, coagulates of milk, smell, and color change. For clinical examination of the udder; the udder of each cow was first examined visually followed by palpation to detect any inflammatory swelling, atrophy of udder tissue, or fibrosis. The consistency and size of the udder were also inspected for any abnormalities such as firmness, disproportional symmetry, and blindness.
Preparation of udder and teats
The udder and the teats were thoroughly cleaned and dried before milk sample collection for the detection of mastitis. A dry towel was used to brush the surface of teats for the removal of dirt, particles of bedding material, and other filth. The teats were then swabbed with cotton and 70% ethanol (35). Teats on the far side of the udder were scrubbed first to avoid re-contamination.
Milk sample collection
When clinical mastitis was detected in the cow, the milk sample was collected from the infected teat by a standard milk sampling technique (36). The time chosen for milk sample collection was before milking. To prevent contamination during sampling, rear-end teats were sampled first. The first three milk streams were discarded and then approximately 10 ml milk sample was collected in a sterile falcon tube. The tube was labeled properly and kept in the icebox. Samples were transferred to the laboratory for bacterial identification. Milk samples were stored at 4ᵒ C for a maximum of 24 hours or at -20ᵒ C till further use.
Surf Field Mastitis Test (SFMT)
SFMT was used as the screening test for subclinical mastitis. It was carried out by the previously described procedure (37). A squirt of milk, about 1 ml from each quarter, was taken in each of the four shallow cups of the Mastitis detection paddle. An equal amount of the SFMT reagent (3% surf solution) was added to each cup. The mixture was gently rotated in the horizontal plan for 20-30 s. The mixture was then examined for thickening or any other visible change. The SFMT results were scored based on gel formation; 0 (negative), 1 (weakly positive), 2 (distinct positive), and 3 (strongly positive). All SFMT scores of 0 were considered as negative while SFMT scores of 1, 2, and 3 were considered as indicators of sub-clinical mastitis. Mastitis-positive cows were defined as having at least one quarter with an SFMT score of 1+ (Table: 7). The theory of gel formation (38) is summarized in Figure 1. Milk sampling was done from mastitis positive cows as described above, for bacteriological analysis. SFMT negative cows were not included in milk sampling. During sampling, milk was collected only from those quarters that were SFMT positive.
Processing of milk samples and bacteriological Assays
Bacteriological analysis was done according to National Committee for Clinical Laboratory Standards (39) and Quinn (36). Refrigerated milk samples were warmed at room temperature for 1-2 hours and vortexed thoroughly to disperse the fat and bacteria in the milk samples and left for some time to disperse the foam before inoculation. Milk sample (0.01 ml) was streaked on blood agar enriched with 5% sheep blood (HiMedia, India), plates were incubated aerobically at 37∘ C for 24-48 h to rule out slow-growing bacteria. The plates were examined for growth, hemolytic characteristic, and morphological features including colony shape, size, opacity, and color. Suspected colonies were subcultured on nutrient agar plate (Oxoid, UK) to obtain pure cultures for further investigation. Pure colonies were subjected to primary investigation through Gram staining and cellular morphology (rods or coccus). The purified strains were further identified on genus level based on (1) their growth patterns on different selective and differential media such as; MacConkey agar (Oxoid, UK), EMB agar (Oxoid, UK), Cetrimide agar (Himedia, India), and Mannitol salt agar (Himedia, India), and (2) their biochemical tests such as catalase test, oxidase test, coagulase test, the “IMViC” tests (Indole, Methyl-Red, Vogas Proskaur, and Citrate utilization), TSI, oxidation fermentation test and CAMP test (40).
Awareness of the farm handlers regarding mastitis
Farm owners and the farmers were asked about the general concepts and information regarding mastitis, to analyze their knowledge and awareness level of this disease Figure 6. To avoid biases, the same person from the team collected the information from all farm handlers.
Preprocessing of data
All the collected data were recorded in the Microsoft Excel spreadsheet and checked for any missing value. The data was coded, and statistical analyses were performed using IBM SPSS Statistics (Version 25) software. The prevalence of clinical and subclinical mastitis was calculated as the proportion of the mastitis-positive cows against the total number of investigated cows. A dairy herd was considered as mastitis positive if at least a single animal with clinical mastitis or SFMT positive result was detected.
Bias variable identification
The association between the categorical independent variables (n=28) and dependent variable, Mastitis (0 = negative and 1 = positive) was calculated by univariable logistic regression analyses. The independent variables assessed were described in Table 1. All variables with a p-value < 0.05 in the univariable analysis were assessed for binary logistic regression model construction.
Variable selection for model construction
All variables with a p-value < 0.05 in the univariable analysis were analyzed for multicollinearity using Kruskal gamma statistics and the variables whose gamma value ranged between −0.6 and +0.6 were considered in a multivariable logistic regression analysis. The final model was built in a backward stepwise elimination procedure with reference to wald statistics. In this analysis, statistical significance was set at p < 0.05 and the variables having higher values were eliminated from the final model (Table 9).
Performance evaluation of model
The model was assessed for goodness-of-fit using the Hosmer-Lemeshow method (41). To further evaluate the performance of the binary logistic regression model; sensitivity, specificity, accuracy, and precision of the model were calculated by the given formulas: