Water buffalo (Bubalus bubalis) farming in Asian countries has grown exponentially over the past half century, and contributes around 13% of global milk production (Siddiky and Faruque, Reference Siddiky and Faruque2018). Water buffalo farming is also becoming popular in Bangladesh due to increasing demand for buffalo milk and milk products, greater resistance to many diseases compared with cows and lower management efforts and feeding costs (Hamid et al., Reference Hamid, Siddiky and Hossain2016). A significant impediment to milk production by water buffalo herds is mastitis, which affects the quantity, quality, and safety of milk, causes heavy economic loss, leads to increased use of antibiotics, and impairs animal welfare (Salvador et al., Reference Salvador, Beltran, Abes, Gutierrez and Mingala2012). Clinical mastitis (CM) can be diagnosed by visible changes in the milk, udder, and systemic condition of animals, but subclinical mastitis (SCM) remains undetectable in most cases due to lack of clinical signs (Patel et al., Reference Patel, Kunjadia, Koringa, Joshi and Kunjadiya2019). In water buffalo, SCM is around three-fold more common than CM (Ali et al., Reference Ali, Rahman, Qureshi, Hussain, Khan, Uddin, Iqbal and Han2014) and is responsible for declining milk production, deteriorating milk quality and reduced milk processability. Subclinical mastitis is also a milk safety concern because of the presence of pathogenic microorganisms (Sharma et al., Reference Sharma, Singh and Bhadwal2011). As the clinical signs remain unnoticed, affected animals may act as reservoirs that shed microorganisms continuously to the environment and affect their herd mates (Ali et al., Reference Ali, Rahman, Qureshi, Hussain, Khan, Uddin, Iqbal and Han2014). Persistent infection also limits the efficacy of antimicrobial treatment by creating a fibrous barrier between the organism and the antibiotic (Putz et al., Reference Putz, Palmer, Ma, Casas, Reinhardt and Lippolis2020).
Somatic cells, a combination of leucocytes and epithelial cells released during regeneration of udder secretory tissue, provide a second line of defense of the mammary gland, so their numbers increase in response to intramammary infection (IMI). These cells are present in high amounts in normal milk, but IMI or stress results in a significant increase in the quantity of somatic cells present in milk (Dang et al., Reference Dang, Kapila, Tomar and Singh2007; Alhussien and Dang, Reference Alhussien and Dang2018). Therefore, milk somatic cell count (SCC) is regarded as a reliable parameter to determine milk quality and identify SCM (Sahin et al., Reference Sahin, Yildirim, Ulutas and Ulutas2017). However, it has been suggested that a combination of direct assessment of SCC and indirect assessment by the California mastitis test (CMT) may be the best option to diagnose SCM (Preethirani et al., Reference Preethirani, Isloor, Sundareshan, Nuthanalakshmi, Deepthikiran, Sinha, Rathnamma, Nithin Prabhu, Sharada, Mukkur and Hegde2015). As there is no fixed standard for SCC in water buffalo bulk milk, the threshold level of 200 000 cells/ml for cows (Adkins et al., Reference Adkins, Middleton, Fox, Pighetti and Petersson-Wolfe2017) has been used as the cut-off value for SCM, for example in a study on water buffalo in Nepal (Dhakal, Reference Dhakal2006) and on Murrah buffalo in India (Dang et al., Reference Dang, Mukherjee, Kapila, Mohanty, Kapila and Prasad2010).
Several studies have investigated the prevalence of SCM in buffalo populations around the world, with most reporting a prevalence of between 36 and 48% (Ali et al., Reference Ali, Ahmad, Muhammad and Anjum2011; Salvador et al., Reference Salvador, Garcia, Abes and Mingala2013; Preethirani et al., Reference Preethirani, Isloor, Sundareshan, Nuthanalakshmi, Deepthikiran, Sinha, Rathnamma, Nithin Prabhu, Sharada, Mukkur and Hegde2015). The most important aetiological agents of SCM are reported to be non-aureus staphylococci (NAS), Staphylococcus aureus, Streptococcus spp. and Escherichia coli (Dhakal, Reference Dhakal2006; Patel et al., Reference Patel, Kunjadia, Koringa, Joshi and Kunjadiya2019). Among the staphylococci, NAS have been found to be dominant pathogens in IMI (Moroni et al., Reference Moroni, Sgoifo Rossi, Pisoni, Bronzo, Castiglioni and Boettcher2006; Ali et al., Reference Ali, Ahmad, Muhammad and Anjum2011; Guha and Guha, Reference Guha and Guha2012; Locatelli et al., Reference Locatelli, Piepers, De Vliegher, Barberio, Supré, Scaccabarozzi, Pisoni, Bronzo, Haesebrouck and Moroni2013) as well as in SCM and mild forms of clinical mastitis (Frey et al., Reference Frey, Rodriguez, Thomann, Schwendener and Perreten2013; Gindonis et al., Reference Gindonis, Taponen, Myllyniemi, Pyörälä, Nykäsenoja, Salmenlinna, Lindholm and Rantala2013; Xu et al., Reference Xu, Tan, Zhang, Xia and Sun2015). In contrast, S. aureus is considered a significant pathogen of buffalo IMI and a concern due to its greater resistance to β-lactam antibiotics (Turutoglu et al., Reference Turutoglu, Hasoksuz, Ozturk, Yildirim and Sagnak2009).
A few previous studies have examined SCM prevalence in water buffalo in Bangladesh (Talukder et al., Reference Talukder, Rahman, Mahmud, Alam and Dey2013; Islam et al., Reference Islam, Kabir, Rahman and Kabir2016; Biswas et al., Reference Biswas, Hanif, Rana and Anower2020). However, there are no data available on buffalo herd-level SCC and studies investigating the susceptibility to β-lactam antibiotics of staphylococci in water buffalo in Bangladesh are scarce. Identifying SCM-causing pathogens and investigating antimicrobial susceptibility in these pathogens are critical prerequisites for implementing effective control of buffalo mastitis. The present study was undertaken to (i) determine the occurrence of SCM and bulk milk SCC in water buffalo in Bangladesh, (ii) investigate the pathogens causing SCM, and (iii) evaluate penicillin resistance in isolated staphylococci strains.
Material and methods
Study area and population
Around 40% of the existing water buffalo population in Bangladesh is concentrated in coastal areas, including the Noakhali, Bagerhat, and Bhola regions (Faruque et al., Reference Faruque, Hasnath and Siddique1990). Two of these regions, Noakhali and Bagerhat, were selected for this study, based on ease of communication and time limitations. Bagerhat (22.8456°N, 89.5403°E) and Noakhali (22.8246°N, 91.1017°E) are located in the southern coastal part of Bangladesh and have buffalo populations of 21 570 and 13 168 head, respectively (Census, Reference Census2010; Huque and Khan, Reference Huque and Khan2017). In Bagerhat, buffalo are reared mainly on household farms, where 1–3 milking buffalo are tethered inside the residence boundaries with minimum roughage and concentrate supplements, or in a semi-intensive system with ≥7 buffalo in enclosed sheds, with or without wallowing facilities in nearby ponds, swamp/marshy land, or rivers. In Noakhali, buffalo are mainly reared in a free-range system (bathan), with 50–500 buffalo in each bathan. Several closely located bathans combine to form a ‘killa’, and there are several killas on each island in the Noakhali region. There are also scattered household farms (Habib et al., Reference Habib, Haque, Rahman, Aftabuzzaman, Ali and Shahjahan2017).
Study design and sampling strategy
A cross-sectional pilot study was performed during September−October 2019 to make a preliminary mapping of the SCM status of buffalo in the study regions. A total of 16 buffalo farms were included. The sampling population was 76 lactating buffalo with 299 functional udder quarters (five blocked quarters). In selection of participating farms, a list of buffalo farmers was obtained from Upazila Veterinary Hospital and 14 household/bathan farms (12 in Noakhali, 2 in Bagerhat) with at least one lactating buffalo were selected, based on accessibility and farmer willingness. Two semi-intensive buffalo farms were also selected, one in the Bagerhat region and one in Noakhali (online Supplementary Figure S1). A convenient sampling technique (1–3 lactating buffalo from each bathan/household and all lactating buffalo from the semi-intensive farms) was applied, due to time limitations and inadequate data on the location of the buffalo farms. In both study regions, the sampling period was seven days.
Data collection and processing
A structured questionnaire in English was used to collect data on farm location (Noakhali/Bagerhat region), education level of the farmer (no formal education/primary/secondary/veterinarian), herd size (total number of buffalo), rearing system (household/bathan/semi-intensive), and number of lactating buffalo on each farm. Participation was voluntary and participants were informed that withdrawal was possible at any time. Oral and written consent were obtained from each participant before applying the questionnaire. Data collection was performed in face-to-face interviews with the farmers (farm owner or acting manager) and through on-farm observations following a checklist. One of the authors (SS), administered the questionnaire and made observations on-farm. All interviews were conducted in Bengali, but the questionnaire was completed in English. The completed observation checklists were re-checked for potential inconsistencies at the end of each day.
Analysis and case definition of subclinical mastitis
The SCM status of each buffalo quarter was determined and categorized on a scale from 1 to 5 using the CMT, following the process described by Baloch et al. (Reference Baloch, Rind, Umerani, Bhutto, Abro, Rind, Abro, Rizwana, Kamboh and Baloch2016). A buffalo was considered positive for SCM if at least one of its four quarters had CMT ≥2. The same threshold was used to categorize SCM in individual udder quarters.
Sample collection and preservation
Samples (each 3–5 ml) of quarter milk for bacteriological culture were collected aseptically using 70% ethyl alcohol, following the NMC protocol (Adkins et al., Reference Adkins, Middleton, Fox, Pighetti and Petersson-Wolfe2017), from quarters with CMT score ≥2. CMT-negative samples were also randomly collected from 24 lactating buffaloes. Collected samples were immediately transferred to an insulated icebox, and transferred to a freezer (−10 to −15 °C) within two hours. Within seven days of field visits, the frozen milk samples were transported to the Udder Health Bangladesh laboratory at Chattogram Veterinary and Animal Sciences University. At the laboratory, the milk samples were stored at −20 °C and bacteriological culture was performed within 24 h.
Bulk milk somatic cell count
After aseptically collecting the composite morning milk sample on each farm, an on-farm bulk milk SCC was performed using a DeLaval cell counter (DeLaval Group, Stockholm, Sweden; sensitivity 73%, specificity 86%) (Adkins et al., Reference Adkins, Middleton, Fox, Pighetti and Petersson-Wolfe2017). This automated device determines SCC optically on a cassette loaded with 60 μl of milk stained with a DNA-specific fluorescent reagent. A digital camera takes pictures of the cell nuclei one-by-one and immediately displays the SCC results, expressed per μl milk.
Bacterial isolation and identification
A total of 146 quarter milk samples, comprising all CMT-positive samples (n = 108) and all randomly selected negative samples (n = 38) from 72 lactating buffalo (no negative samples were taken from the other four buffalo) were cultured in 5% bovine blood agar following the NMC protocol (Adkins et al., Reference Adkins, Middleton, Fox, Pighetti and Petersson-Wolfe2017). After 24 h of incubation at 37 °C, at least two morphologically similar bacterial colonies were re-cultured on bovine blood agar. The pure cultures were enriched in brain heart infusion broth and incubated at 37 °C, after which 300 μl 50% buffered glycerol were added and the samples were stored at −80 °C until further use. The pure cultures were also cultured on selective mannitol salt agar, to identify S. aureus and NAS. Bacterial colonies were identified by colony morphology, colony colour, and haemolytic pattern. All culture media and reagents were manufactured by Oxoid®, Oxoid Ltd, Basingstoke, United Kingdom.
Species confirmation
The isolates from pure cultures were transported to the National Veterinary Institute, Uppsala, Sweden, in Copan Transsystem® (Capon, Brescia, Italy) transport media tubes, for species confirmation using a MALDI-TOF Biotyper 3.0 (Bruker Daltonics GmbH, Bremen, Germany) for validation (score >1.8) of phenotypic identification accuracy.
Penicillinase test of Staphylococcus spp.
All staphylococci were tested for β-lactamase production, using the cloverleaf method described by Mee'aad et al. (Reference Mee'aad, Azhar and Ali2018). For quality control, the strains S. aureus ATCC 29213 and S. aureus ATCC 25923 were used.
Statistical analysis
Collected data were imported into Microsoft Excel 2019 (MicrosoftCorp., Redmond, Washington, USA). The data were then cleaned, sorted, coded, and checked for inconsistency before export into STATA-IC-13 (StataCorp, 4905, Lakeway Drive, College Station, Texas 77845, USA) for descriptive statistical analysis. Results are presented as frequency and percentage, with 95% confidence intervals (CI). Demographic characteristics of farmers (location, education, farming system) are presented as frequencies and percentages.
Results
Descriptive statistics on farms and farmers
Of the 16 buffalo farms studied, three were in the Bagerhat region and 13 were in the Noakhali region. Farm type was mainly household farms (n = 11), followed by bathan (n = 3) and semi-intensive farms (n = 2). Among the farmers, seven had primary-level education, four had no formal education, four had achieved secondary-level education, and one farm was managed by a registered veterinarian. Twelve of the farms had 1–5 lactating buffalo, two had 6–10 lactating buffalo and the remaining two farms had ≥7 lactating buffalo. The numbers of animals sampled on each farm are listed in online Supplementary Table S1.
Occurrence of SCM and intramammary infection
Estimated SCM occurrence in the buffalo at animal and quarter level was 81.6 and 42.5%, respectively. At quarter level, the estimated occurrence of IMI was 40.4% and the combined occurrence of SCM (CMT ≥2) and IMI was 43.5% (Table 1). Among the SCM-affected quarters (CMT score 2–5), a CMT score of 2 was the most frequent (Table 2).
Status of subclinical mastitis at animal and quarter level, based on 299 active udder quarters and 76 lactating water buffalo, on 16 buffalo farms in the Bagerhat and Noakhali regions of Bangladesh
SCM, Subclinical mastitis; CMT, California mastitis test (score range 1−5); IMI, Intramammary infection; CI, Confidence interval.
Status of subclinical mastitis according to the California mastitis test at quarter level, based on 299 active water buffalo udder quarters and cultures of 146 quarter milk samples from 16 buffalo farms in the Bagerhat and Noakhali regions of Bangladesh
CMT, California mastitis test (score range 1 to 5); IMI, Intramammary infection; CI, Confidence interval.
Bulk milk somatic cell count
Mean bulk milk SCC on the 16 farms was 195 000 cells/ml milk (range 47 000– 587 000 cells/ml milk). In quartile terms, the lowest 25, 50 and 75% of farms had a bulk milk SCC value of 129 000, 184 000, and 233 000 cells/ml milk, respectively. The individual bulk milk SCC count for each of the 16 farms is shown in online Supplementary Figure S2.
Pathogen distribution and penicillinase production by Staphylococcus spp.
Among the 146 quarter milk samples, 79 (54.1%) bacterial isolates were obtained and 60 were confirmed by the MALDI-TOF results. Ten isolates yielded no bacterial growth on re-culture after shipment and nine isolates were non-identifiable on a MALDI-TOF spectral scale of >1.8. The dominant bacteria were NAS, specifically S. hyicus (Table 3). All 36 NAS isolates were subjected to a penicillinase test, in which 13 isolates (36.1%) tested positive for β-lactamase production (Table 3).
Pathogen distribution and penicillin resistance identified in quarter milk samples collected from water buffalo in the Bagerhat and Noakhali regions of Bangladesh (confirmed by MALDI-TOF mass spectrometry)
NAS, Non-aureus Staphylococci. 108 subclinical mastitis-positive quarters and 38 negative quarter samples were tested. Frequency of 36 NAS species was below the overall NAS estimates in the table.
Discussion
The SCM status in two of Bangladesh's main buffalo farming regions was successfully estimated at animal- and quarter-level, and pathogen distribution and penicillin resistance by NAS were determined. The quarter- and animal-level occurrence of SCM was found to be rather high, which is consistent with previous reports from neighbouring countries and around the world (Sharma et al., Reference Sharma, Maiti and Sharma2007; Preethirani et al., Reference Preethirani, Isloor, Sundareshan, Nuthanalakshmi, Deepthikiran, Sinha, Rathnamma, Nithin Prabhu, Sharada, Mukkur and Hegde2015; Ahmed et al., Reference Ahmed, Straubinger, Hegazy and Ibrahim2018). The quarter-level occurrence of 42.5% in the present study was in line with the 44–46% reported for other sub-tropical countries such as India and Egypt, where water buffalo production systems resemble those in Bangladesh. Other studies in India have reported higher occurrence (59–78%) (Bhanot et al., Reference Bhanot, Chaudhri, Bisla and Singh2012; Charaya et al., Reference Charaya, Kumar, Sharma, Hoel, Kumar, Kumar, Kumar, Gupta and Mittal2015), while lower prevalence (28%) has been reported in a small-scale study in Bangladesh (Islam et al., Reference Islam, Islam, Jahan, Chaudhary and Anower2019). Compared with the quarter-level prevalence, the animal-level SCM prevalence was much higher in this study (81.6%), which is in line with previously reported values ranging from 70 to 78% (Bachaya et al., Reference Bachaya, Iqbal, Muhammad, Yousaf and Ali2005; Sentitula et al., Reference Sentitula, Yadav and Kumar2012). These differences in estimated SCM could be due to differences between study sites, study years, geographical areas, the criteria applied for SCM detection, climate variations, and underlying animal- or quarter-level risk factors (Hussain et al., Reference Hussain, Ahmad, Mushtaq, Chaudhry, Khan, Reichel, Hussain, Khan, Nisar and Khan2018). Investigating such differences was beyond the scope of this study, but further studies are needed to assess and quantify the farm- and animal-level factors that might influence SCM occurrence. Such knowledge could be valuable in controlling mastitis in water buffalo.
Mean bulk milk SCC estimated in the present study was 195 000 cells/ml milk. This is in line with Mujeeb et al. (Reference Mujeeb, Digraskar and Awaz2009) and within the acceptable threshold of 200 000– 500 000 cells/ml buffalo milk established in India (Panchal et al., Reference Panchal, Sawhney and Dang2016; Jhambh et al., Reference Jhambh, Dimri, Gopalakrishnan, Singh and Chhabra2017). However, it exceeds the cut-off value of 100 000 cells/ml milk recently proposed for Murrah buffalo by NDRI-India and barely meets the cut-off of 200 000 cells/ml milk suggested for neighbouring countries such as Nepal and India (Dhakal, Reference Dhakal2006; Alhussien and Dang, Reference Alhussien and Dang2018). The median level of bulk milk SCC was 184 000 cells/ml milk, representing slightly better status. Increased SCC is positively correlated with IMI in water buffalo (Moroni et al., Reference Moroni, Sgoifo Rossi, Pisoni, Bronzo, Castiglioni and Boettcher2006) but the increase in SCC level recorded here was still much lower than that in cows and goats (Moroni et al., Reference Moroni, Pisoni, Ruffo and Boettcher2005). However, there was large variation in SCC level between the individual farms (online Supplementary Figure S2), an issue which was not explored further in this study. The level of SCC increases with entry of IMI-associated pathogens into the udder through the teat canal and with stress conditions due to seasonal diurnal variations in climate conditions and parity of the lactating buffalo (Alhussien and Dang, Reference Alhussien and Dang2018; Bombade et al., Reference Bombade, Kamboj, Alhussien, Mohanty and Dang2018). To date, relevant data on SCM in water buffalo in Bangladesh have been insufficient. The present study therefore makes a timely contribution, especially considering the increasing consumer demand for buffalo milk.
The occurrence of IMI in buffalo was shown to be high in this study (40.4%). Relative IMI occurrence in healthy and SCM-affected buffalo quarters was similar (40 and 44%, respectively). In water buffalo in India, IMI occurrence in SCM quarters of 38–45% has been reported (Singh et al., Reference Singh, Sharma, Mittal and Charaya2014; Jhambh et al., Reference Jhambh, Dimri, Gopalakrishnan, Singh and Chhabra2017), which is consistent with our findings. The most common pathogens in this study were NAS, which is also consistent with findings in other studies on water buffalo (Bansal et al., Reference Bansal, Hamann, Lind, Singh and Dhaliwal2007; Sharma et al., Reference Sharma, Maiti and Sharma2007; Kumar, Reference Kumar2009; Singh et al., Reference Singh, Sharma, Mittal and Charaya2014). Staphylococcus hyicus was the most dominant species among the NAS and has been identified previously in IMI in water buffalo in India (Singh et al., Reference Singh, Sharma, Mittal and Charaya2014) and Brazil (Coimbra-e-Souza et al., Reference Coimbra-e-Souza, Brito, Chamon, Laport and Giambiagi-deMarval2017). Non-aureus staphylococci, which are commonly found in the teat canal, are an important group of SCM pathogens and are often resistant to one or more antimicrobials (Condas et al., Reference Condas, De Buck, Nobrega, Carson, Roy, Keefe, DeVries, Middleton, Dufour and Barkema2017; Stevens et al., Reference Stevens, Piepers, Supré and De Vliegher2018; Traversari et al., Reference Traversari, van den Borne, Dolder, Thomann, Perreten and Bodmer2019). In the present study, 36% of NAS exhibited penicillin resistance, which is in line with results reported in other studies in India and Brazil (Pizauro et al., Reference Pizauro, Silva, Santana, Clemente, Lara, Listoni, Vaz, Vidal-Martins, Ribeiro and Fagliari2014; Singh et al., Reference Singh, Sharma, Mittal and Charaya2014). This level of resistance is high in terms of public health safety and needs to be confirmed in future studies, which should also explore the underlying cause of dissemination of antibiotic resistance in buffalo herds in Bangladesh.
The discrepancies between the low bulk milk SCC in the buffalo herds studied (195 000 cells/ml milk) and the higher IMI at quarter level were not further assessed in the present study. However, a relatively low level of SCC and higher IMI in buffalo has been found in a previous study by Moroni et al. (Reference Moroni, Sgoifo Rossi, Pisoni, Bronzo, Castiglioni and Boettcher2006). Presence of contagious S. aureus or Streptococcus agalactiae plays a major role in increases in SCC level above 500 000 cells/ml milk (Moroni et al., Reference Moroni, Nydam, Ospina, Scillieri-Smith, Virkler, Watters, Welcome, Zurakowski, Ducharme, Yeager, Peek and Divers2018), so the absence of these pathogens in the present study could partly explain the lower SCC level. Moreover, the increase in SCC in IMI-associated quarters in buffalo may be lower than in dairy cows due to limited phagocytic activity. According to Sahoo et al. (Reference Sahoo, More and Singh1998), there is a core difference in lysosome containing hydrolases between cattle and buffalo, with lower concentrations of these enzymes in buffalo, resulting in lower phagocytic activity than in other ruminant species. The requirements of convenient sampling and small sample size imposed limitations in this pilot-scale study. However, the data obtained reflect the regional SCM situation in water buffalo in Bangladesh and can act as the basis for further national-level research.
In conclusion, high occurrence of SCM in water buffalo at quarter and animal level (42.5 and 81.6%, respectively) was found on buffalo farms in the Noakhali and Bagerhat regions of Bangladesh. A moderate rate of SCC was detected in buffalo bulk milk on these farms (7 out of 16 farms exceeded the SCC threshold), which is reasonably good for pre-harvesting buffalo milk in Bangladesh. These findings suggest that appropriate treatment and control strategies are needed to control mastitis in water buffalo in Bangladesh. NAS was the most frequently identified pathogen and the NAS strains detected frequently tested positive for penicillinase production, which emphasizes the importance of prudent use of antibiotics to maintain public health when consuming buffalo milk. A systematic assessment of the direct and indirect transmission pathways of these pathogens between water buffalo is urgently needed.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0022029921000698.
Acknowledgements
Some of the authors participated in the research project Udder Health Bangladesh (UHB). We are grateful to the Swedish Research Council (2018–03583) and Swedish International Development Corporation through the support to Minor Field Studies for funding this research project, and to the participating buffalo farmers for their cooperation. We also thank UHB bacteriology laboratory and the Chattogram Veterinary and Animal Sciences University and Swedish National Veterinary Institute laboratories for their technical support, and Prof. Dr Fabrizio Ceciliani, Department of Veterinary Medicine, University of Milan, Italy, for reading the manuscript.
