Staphylococcus aureus in cow milk and milk products in Ambo and Bako towns, Oromia, Ethiopia: prevalence, associated risk factors, hygienic quality, and antibiogram

Staphylococcus aureus (S. aureus) is a foodborne bacterial pathogens that can cause staphylococcal food poisoning and contaminate food of animal origin worldwide. The current study was conducted to estimate the prevalence and assess risk factors, hygienic quality, and antibiogram of S. aureus in raw milk and milk products of cows in Ambo and Bako towns, Ethiopia. The overall prevalence of S. aureus in milk and milk products was 15.6% (94/601) with the highest prevalence in bulk tank raw milk (17.50%) and the lowest in “Ergo” (13.11%). High S. aureus contamination at farm level were associated with poor farm hygiene, extensive management system, medium farm size, loose housing, and less frequent removal of bedding. At the cow level, a high S. aureus isolation rate was observed in crossbred cows; cows with age equal to or greater than 5 years old, tick infestation, history of mastitis treatment, and udder washing were not practiced before milking. On the other hand, the type of container, hygiene of milk handler, and container were the major risk factors for bulk tank milk contamination with S. aureus. S. aureus counts ranging from 1.25 × 104 to 1.92 × 104 CFU/mL were detected in 28.33% of the bulk tank milk samples.. Antimicrobial susceptibility test showed higher resistance of S. aureus to amoxicillin (98.48%), oxacillin (98.48%), ampicillin (98.48%), cefoxitin (92.42%), and tetracycline (83.33%), with 43.94% of isolates showing multidrug resistance (MDR). The high prevalence of oxacillin and cefoxitin-resistant isolates, which is a possible indicator of the existence of methicillin-resistant Staphylococcus aureus (MRSA), was also noted in the current study. Higher prevalence of S. aureus and MDR isolates in milk and milk products was detected in study areas. Therefore, to make milk and milk products safe for human consumption, hygienic handling of milk and milk products, regular surveillance of antimicrobial resistance, and prudent use of drugs are recommended.


Background
Staphylococcus aureus is one of the bacterial pathogens contaminating food of animal origin (Nouichi and Hamdi 2009). It is known for its toxin-mediated virulence factors and its propensity for antimicrobial resistance (Argaw and Addis 2015).
The bacteria produce heat-stable and proteolytic enzymes and enterotoxins, which induce food poisoning in humans, manifested by vomiting, abdominal pain, and diarrhea (Busani et al. 2006;Argudín et al. 2010). S. aureus is also the cause of bovine mastitis, one of the most economically important diseases in high-yielding dairy cows (Tiwari et al. 2013).
Staphylococcal food poisoning (SFP) is one of the most common foodborne pathogens worldwide, with a high occurrence next to salmonellosis (Aycicek et al. 2005). Staphylococci cause mild skin infections to more severe illnesses such as pneumonia and septicemia (Ateba et al. 2010). Staphylococcal food poisoning occurs when there is an infection of the mammary gland, poor hygienic habits, such as coughing or sneezing, and failure to wash hands when handling milk, milk products, and storage equipment during or after milking (de Oliveira et al. 2011). In addition, prolonged storage of milk at room temperature before consumption could lead to further proliferation of the pathogen and the production of toxins by enterotoxigenic strains of Staphylococcus species. S. aureus is more prevalent during the warmest months of the year indicating the impact of temperature on the prevalence of the pathogen (Tarekgne et al. 2015). The origin of the samples and the acidic nature of cheese and yogurt reduce the prevalence of Staphylococcus species (Argaw et al. 2018).
Previous studies have shown that differences in milking practices, management systems, cow breeds, and location of the farm are among the risk factors associated with the occurrence of S. aureus in animals (Sudhanthiramani et al. 2015;Tafa et al. 2015). Antimicrobial resistance of S. aureus has also become a very essential public health issue worldwide (Abebe et al. 2016). Furthermore, strains of S. aureus have been observed to show resistance against multiple antimicrobial agents (Ventola 2015). Studies in Ethiopia also showed that the susceptibility of S. aureus to penicillin G and tetracycline is very low due to the regular use of these antibiotics for the treatment of cows which might result in the spread of resistant strains in most areas of the country (Pekana et al. 2017;Seyoum et al. 2017).
Consumption of raw milk and milk products such as cream, butter, traditionally fermented milk ("Ergo"), Ethiopian cottage cheese ("Ayib"), and yogurt are common in Ethiopia (Gonfa et al. 2001). However, the high incidence of SFP is likely to be due to poor hygiene in food production and preparation, scarcity of cooling facilities (Alemu 2014;Ararsa et al. 2014), and raw milk consumption practices in the country (Makita et al. 2012). Previous studies conducted in Toke Kutaye, Dendi, and Cheliya districts in West Shoa Zone, Oromia region, Ethiopia, showed a high isolation rate of Staphylococcus species from raw milk (Dabele et al. 2021) and a high prevalence of MDR S. aureus in raw milk and milk products from a study in Holeta, central Ethiopia (Gebremedhin et al. 2022). However, there is a scarcity of data on the prevalence and antimicrobial susceptibility profile of S. aureus in raw cow milk and milk products and risk factors contributing to contamination in Ambo and Bako towns, West Shoa Zone of Oromia region, Ethiopia. Therefore, the objectives of the current study were to estimate the prevalence, associated factors, hygienic quality, and antibiogram of S. aureus isolated from milk and milk products in Ambo and Bako towns, in the Oromia region, Ethiopia.

Study area
The study was conducted in Ambo and Bako towns in the West Shoa Zone of the Oromia Region, Ethiopia (Fig. 1). Ambo is found in the West Shoa Zone and is located 114 km west of Addis Ababa, the capital city of Ethiopia. The town had a total human population of 118,649, out of which 57,650 were women and 59,999 were men. Ambo has a cattle population of 110,952 (CSA 2007).
Bako town is located 260 km West of Addis Ababa and has an estimated total population of 18,641, of which 9370 are men and 9271 are women. Bako has a cattle population of 127,343. Cattle in both towns are reared either under extensive, intensive, or semi-intensive management systems. Veterinary services are provided by one governmental district veterinary clinic and several private veterinary pharmacies in each town. There was no documented statistical information on lactating cows in the study areas. The two towns have bimodal rainfall characterized by a small rainy season from February to May and a big rainy season from July to September. The dry season extends from October to January (CSA 2007). "Kebeles," the smallest administrative units under the towns, were purposively selected based on the availability and population of dairy cows. Accordingly, six urban "Kebeles" of Ambo and two "Kebeles" of Bako were considered for this study.

Study population and study animals
The study population was all lactating cows found in Ambo and Bako towns, Oromia region, Ethiopia. Lactating local breeds (Boran and Horro), Jersey (pure breed), and crossbreed cows of Boran with Holstein Friesian, from which samples were collected, were the study animals.

Study design and sample size
A cross-sectional study design was used, and the study was carried out from September 2018 to January 2020. The sample size was calculated using the formula described by Thrusfield (2005) at 5% precision and an expected prevalence of 47% S. aureus in raw milk from the previous study (Mekibib et al. 2010) and 5% in dairy products ("Eryo" and "Ayib") (Argaw et al. 2018).
where n is the sample size, Pexp is the expected prevalence, and d is the desired absolute precision. Accordingly, the minimum sample size is 383 raw cow milk and 73 milk products. However, to compensate for probable sample loss at different stages of processing and greater study power, 479 raw milk (399 from individual cows, 80 bulk tank milk samples from dairy farms and restaurants) and 122 milk product (61 "Ayib" and 61 "Ergo") samples were collected purposively, which makes the overall samples 601. A simple random sampling technique was used to select dairy farms, restaurants, and milk product sellers in the local market in the study areas.

Selection of herds, cows, and restaurants
There is no registration of dairy farmers in Ethiopia. Names and addresses of the dairy farm owners were collected from private and governmental artificial insemination records and district veterinary clinics (Livestock production and fishery office of Ambo, 2018; Livestock production and fishery office of Bako, 2018). A long list of 350 dairy farmers from Ambo and 251 from Bako was made. Computer-generated random numbers were assigned to each farm on the long list. From 601 farms identified in the two study towns, 20% (121 farms) was considered, and a proportional number was allocated for the two study areas (70 farms from Ambo and 51 farms from Bako). In addition, 22 restaurants (10 from Bako and 12 from Ambo town), known for serving milk and milk products for their customers, were selected purposively for bulk tank milk sample collection.

Questionnaire survey
A structured questionnaire survey was used to collect epidemiological information and risk factors including study area (Ambo, Bako), breed (Holstein Friesian cross bred with local Boran, pure Jersey, Horro, Boran), age of lactating cows (young when < 5 years, old when ≥ 5 years), parity (1-3 calves, 4-6 calves, ≥ 7 calves), stage of lactation (early, mid, late), management system (intensive, semi-intensive, extensive), and floor hygiene (poor when there was no drainage system and no daily manure cleaning and good when there is a drainage system and the cow manure is cleaned daily), farm hygiene (poor when the ground is wet, had a smell and gross dirt observed in farm compound and good if clean and dry), and farm size (small when the number of dairy cattle is ≤ 10, medium when it is > 10 but ≤ 40) (Abebe et al. 2016). Hygiene of milk product container (poor when washed just with water and good when washed with water and soap/Ajax and smoked) and hygiene of product handler (poor when not washed or just washed with water and dried using own unhygienic clothes and good if washed with lukewarm water and dried with clean towels) were also recorded. Other factors like the use of disinfectants, the practice of handwashing, types of housing (labeled as individual if the animal is kept in an individual pen and loose if the animal is allowed to move freely in the barn), frequency of removal of bedding, udder washing, use of individual towels, use of warm water, tick infestations, type of milking utensils, and previous history of mastitis treatment were also collected.

Sampling technique and transportation
The lactating cows were given an identification code. Dairy farms and lactating cows were selected using simple random sampling, and 250 mL of raw milk samples were collected from three spots, namely, raw milk from the milking pot of individual cows, bulk milk samples from milk collection containers at the farm level, and the restaurant. In addition to raw milk samples, around 100 g of cottage cheese / "Ayib"/ and 100 mL of "Ergo" were also sampled from the different retail dairy product vendors and local markets in the study areas. Until the required sample size was met, every other product seller along the way was added. An ice box containing an ice pack was used for the transportation of all samples from the field, and bacteriological analysis was conducted at the Zoonoses and Food Safety Research Laboratory of Ambo University within 1 to 2 h of arrival.

Bacteriological analysis
Upon arrival at the laboratory, samples were processed immediately or stored at refrigeration temperature until analyzed. The samples were processed as per the ISO standard (ISO 1999). The collected raw milk samples were cultured on 5% sheep blood agar (Oxoid, UK), and the plates were incubated aerobically at 37 °C and examined after 24 h of incubation for the growth of bacterial colonies. The colonies were identified on the bases of Gram's stain reaction (Gram-positive), colony morphology and pigmentation (creamy, greyish, white, or yellow), and hemolytic features on the blood agar plate. The Staphylococcus suspected colonies were subcultured on nutrient agar (Oxoid, UK) and incubated at 37 ℃ for 24 h under aerobic culture conditions to obtain pure colonies for further identification. The pure colonies were maintained in the nutrient slant at 4 ℃ for further analysis Radostits et al. (2007). The pure colonies from the nutrient agar plate (NAP) were subjected to the catalase test and tube coagulase test (Quinn et al. 2002). Accordingly, colonies that produce gas bubbles for the catalase test and any degree of visible clotting when colonies in a nutrient broth mixed with rabbit plasma (Tallent et al. 2001) were considered Staphylococcus species. Mannitol salt agar tests were done on the coagulase-positive Staphylococci to identify S. aureus, where the fermentation of mannitol by S. aureus results in yellowish discoloration of the medium (Quinn et al. 2002). All the bacterial isolates were subcultured on NAP with 20% glycerol at − 20 ℃ for further analyses (Radostits et al. 2007).

Enumeration of Staphylococcus aureus in raw bulk tank milk samples
Enumeration of S. aureus in raw bulk tank milk samples was done as per ISO 6888 (1999). Accordingly, 1 mL of raw bulk tank milk was taken and homogenized into 9 mL peptone water. Following this, serial dilutions were prepared from 10 3 up to 10 6 in normal saline water and cultured by transferring 1 mL of sample suspension aseptically to Baird parker agar base (SISCO, India) supplemented with egg yolk emulsion and potassium tellurite (England, Basingstoke). The inoculums were spread over the surface of the agar plate using a sterile bent glass streaking rod, and plates were retained in the upright position for about 10 min. Then the plates were inverted and incubated for 24-48 h at 37 °C. Finally, the plate containing colonies with a typical appearance of circular, smooth, convex, moist, and gray to jet black frequently with light colored margin surrounded by an opaque zone and frequently with an outer clear zone in the medium was counted as S. aureus. According to Regasa et al. (2019), plates containing 20-200 colonies were selected for S. aureus count, and total S. aureus colonies from two consecutive plates of each sample were counted and using the formula given by Cortimiglia et al. (2016), converted to colony forming unit per millimeter (CFU/mL) of milk.
where N is the number of bacterial colonies counted, ∑C is the sum of colonies identified on two consecutive dilutions, where at least one contained 20 colonies and less than 200 colonies, n1 is the number of plates counted at the first dilution, n2 is the number of plates counted at the second dilution, V is the volume of inoculums on each dish/plate in milliliter, and D is the dilution rate corresponding to the first dilution selected (the suspension is a dilution).

Antimicrobial susceptibility test
The S. aureus isolates were subjected to 12 antimicrobial discs, which were checked for expiry dates. Only unexpired antimicrobial discs were used for this experiment. Antimicrobials were selected based on the common usage of the drugs in the area. The antimicrobial discs were ampicillin (10 µg), amoxicillin (25 µg), cefoxitin (30 µg), oxacillin (1 µg), vancomycin (30 µg), tetracycline (30 µg), chloramphenicol (30 µg), nalidixic acid (30 µg), norfloxacin (10 µg), nitrofurantoin (300 µg), gentamicin (10 µg), and azithromycin (30 µg). The antimicrobial disk diffusion method was applied to check the antimicrobial susceptibility of the isolates following the guideline established by the Clinical and Laboratory Standards Institute (CLSI 2017). When isolates show a pattern of resistance to at least one agent from three or more antimicrobial classes, they were classified as multiple drug-resistant (Magiorakos et al. 2012). E. coli ATCC25922 and S. aureus ATCC6538 reference strains were used as quality control.

Data management and analysis
The data obtained from the field questionnaire survey and laboratory were entered in Excel 2010 Spreadsheet for storage, coded, and entered into the data editor view of STATA version 11 (STATACORP 2009). Descriptive statistical analysis was used to summarize the data into proportions, %, and mean ± SD. Pearson's chi-square test was used to analyze the proportions of categorical data. Univariable and multivariable logistic regression analyses were done to assess the magnitude of the association between the prevalence of S aureus and potential risk factors. Non-collinear variables with a P-value less than 0.25 in the univariable analysis were selected for the multivariable model. The S. aureus count data/mL of milk was first transformed to the logarithm of base ten (log counts/mL) before analysis. One-way ANOVA was used to assess the association between S. aureus count data and independent variables (potential risk factors). The odds ratio and 95% confidence interval (CI) were computed, and the results were considered significant at P < 0.05.
The higher S. aureus contamination in bulk milk compared to that of milk from individual cow levels in the current study agrees with the finding of Addis et al. (2011). This could be due to milk cross-contamination during co-mingling and inadequate handling during transportation from farms to collection centers and at milk collection facilities. The individual cow-level prevalence of S. aureus in the present study was almost in agreement with the 15.5% prevalence reported by Asrat et al. (2013) in Addis Ababa, Ethiopia. However, the current finding is higher than the 11.4% prevalence in Afar, Ethiopia, by Beyene (2016), 13.9% in Ambo and Guder town, Ethiopia, by Megersa (2015), and 8% in Bishoftu, Ethiopia, by Addis et al. (2011). On the other hand, the current value is lower than the previous prevalence of 19.6% by Ayele et al. (2017) in Sebeta, 19.3% by Abunna et al. (2016) in Asella, and 21.2% by Tessema and Tsegaye (2017) in Alage Agricultural Vocational Education and Training College Dairy Farm, Ethiopia. A much higher prevalence of 39.1% at Asella by Bedada and Hiko (2011), 46.5% in Asella, Ethiopia, by Seyoum et al. (2018), and 56.6% in Kombolcha by Asmelash et al. (2016) were also reported. The variation in the magnitude of prevalence in the different studies might be due to differences in the management practices at farms, the sample collection procedure, hygienic milking and handling practices, and the use of towels for udder drying. The prevalence of S. aureus (17.50%) in bulk tank milk in the present study was much higher than in the previous study by Wodajo et al. (2016) who reported 3.2% from Asella Dairy Union Ada Primary Dairy Cooperative, Ethiopia, lower than the 35.3% prevalence in Alage ATVET College Dairy Farm, Ethiopia, reported by Tessema and Tsegaye (2017), and 63.15% in Arsi Negele town, Ethiopia, by Tsige (2018). S. aureus is a contagious pathogen, which can spread from one animal to another or personnel by contact with cows during unhygienic milking procedures (Asmelash et al. 2016). The differences in the prevalence of S. aureus from bulk tank milk from various studies could be due to variations in the hygiene of udder and teats, milking utensils, bulk milk tanks, towels used for drying hands and utensils, and poor handling and transportation of milk to collections centers (Addis et al. 2011). Another factor that might contribute to the variation in S. aureus prevalence in the bulk tank milk would be the prevalence of mastitis in different areas, which is reported to vary from farm to farm and with localities in Ethiopia. Hence, when mastitic cows are present in the farm, S. aureus may be excreted directly from the mastitic udder into milk and upon mixing milk from different farms; the bulk tank milk could be contaminated resulting in higher contamination levels (Deddefo et al. 2022).
In the current study, the prevalence of S. aureus in raw bulk tank milk (17.50%) is higher than that of "Ayib" (18.03%) and "Ergo" (13.11%). Higher contamination with S. aureus in raw milk than in milk products, which is in agreement with the current study, was reported by Lemma et al. (2021). Ethiopian cottage cheese ("Ayib") is formed by heating fermented milk from which butter has been removed after churning, whereas traditionally fermented milk ("Ergo") is prepared by keeping the milk at room temperature until it ferments without pasteurization and adding microbes to start the fermentation process. Thus this processing of dairy products using heat and fermentation techniques might have potentially reduced S. aureus contamination in the milk products (Dhanashekar et al. 2012). The overall prevalence in the milk production of the current study 15.57% (19/122) is almost in agreement with the previous 11.3% prevalence of S. aureus by Lemma et al. (2021) in Addis Ababa and 14.3% prevalence by Argaw et al. (2018) from Jimma. However, the S. aureus contamination in "Ayib" in the current study is much higher than that in "Ayib" (5%) and lower than that in "Ergo" (17.5%) reported by Argaw et al. (2018) from Jimma, Ethiopia. This difference in the prevalence of the milk products in the different studies might be attributed to the difference in hygienic practices of dairy farmers and milk product processors and sellers/vendors in the different regions of the country. In addition, the presence of mastitic cows in the farms and milk and milk product transportation without cold chains and an absence of awareness of food-borne diseases in the farming community could be some of the reasons (Deddefo et al. 2022).

Risk factors of S. aureus prevalence at the farm level
Univariable analysis of potential farm-level risk factors showed that floor hygiene, farm hygiene, farm size, type of housing, and frequency of bedding cleaning were associated with the prevalence of S. aureus at the farm level. Higher prevalence was found in poor hygienic farms (OR = 3.60, 95% CI: 1.55, 8.38) and floors (OR = 2.58, 95% CI: 1.16, 5.77), medium-sized farms (OR = 5.41, 95% CI: 2.30, 12.72), loose housing systems (OR = 3.54, 95% CI: 1.40, 8.96), and in farms where bedding was removed less frequently (OR = 3.16, 05% CI: 1.41, 7.08) than their counterparts. The multicollinearity matrix revealed that except for farm hygiene and floor hygiene (r = 0.77), all independent variables were non-collinear with each other at the farm level. Of the collinear variables, farm hygiene was selected for the multivariable model. Accordingly, the management system, farm size, type of housing, frequency of removal of bedding, and farm hygiene were selected for the multivariable model. Thus, the analysis showed that farm hygiene, farm size, type of housing, and frequency of bedding removal were independent predictors of S. aureus prevalence at the dairy farm level in the study areas (Table 2). In agreement with the current finding, farm hygiene by Biffa et al. (2005) farm management system by Abera et al. (2013), and herd size by Abebe et al. (2016) reported to affect the prevalence of S. aureus at the farm level. This might be due to the fact that dirty and muddy shared barns with high herd sizes favor the proliferation and transition of pathogens like S. aureus in dairy farms.
The multicollinearity test showed that the use of individual towels and udder washing (r = 0.97) and udder disinfection and previous treatment (r = − 0.84) were collinear with each other. Among these collinear variables, udder washing and history of previous treatment of mastitis were selected for the multivariable model. Except for study area and parity, which were excluded due to univariable P-value > 0.25, the remaining variables were entered into the multivariable analysis. This analysis showed that breed, age, previous history of mastitis treatment, udder washing, use of warm water for utensils, and tick infestation were the independent predictors of S. aureus prevalence at the cow level. The full model has Hosmer-Lemeshow (HL) chi-square (8) = 10.49, P = 0.232, area under the ROC curve = 0.8105, sensitivity 23%, specificity 98.2%, positive predictive value 70%, and negative predictive value 87.6% (Table 3).
The current findings are in agreement with the previous study by Pittet et al. (2009), which reported that using detergents can decrease S. aureus contamination. Ayele et al. (2017) reported no practice of udder washing increases the prevalence of S. aureus, and Yilma et al. (2011) identified pre-milking udder preparation like udder washing and employing good milk handling practices play an important role in decreasing S. aureus in dairy farms. Even though not significant, the numerically higher prevalence of S. aureus observed in farms where milking containers were washed with water and soap could be due to the lukewarm water, which is not enough to kill the bacteria. Previous studies also showed that the position of the teat canal and udder anatomy of exotic and crossbred high milk-producing cows are responsible for injuries in these cows, resulting in infection of the mammary tissue with bacteria as reported by Radostits et al. (1994) and Elemo et al. (2017). The high prevalence of S. aureus in cow's teat lesions caused by tick infestation reported by Shiferaw and Telila (2017) is also in agreement with the current study. In addition, it was found that most of the time even if farmers wash their milking containers with hot or warm water, finally they rinse and/or wash it with their hands using cold water just before they use it for milking. Thus, such malpractice of washing the utensil with unwashed hands might be the cause of re-contamination of the milking utensils. Even though farmers keep milk and milk products for sale in relatively cool places in the current study, lack of cooling facilities and failure of keeping the cold chain during transportation and at milk collection centers might also be the other risk factors for the S. aureus contamination (Oliver et al. 2005;Muehlhoff et al. 2013).

Risk factors of S. aureus in bulk tank milk
The majority of bulk milk collectors in the study areas (70%) used plastic containers than stainless steel (30%). The prevalence of S. aureus in bulk tank milk showed a statistically significant association (P < 0.05) with the type of bulk milk collection container, handler hygiene, hygiene of milk collection tank, the habit of picking nose, handwashing after touching cash money, and covering the hair, mouth, and nose during milking (Table 4).
Similar to the current study, the association of the prevalence of S. aureus to the type of bulk tank milk was reported by Tegegne and Tesfaye (2017). This could be due to the inherent characteristic of plastics that makes them unsuitable for milk handling as they scratch easily, provide a hiding place for bacteria during cleaning, sanitation, and poor conductor heat, and thus will hinder effective sanitization by heat leading to bacterial contamination of milk (Omoe et al. 2005). On the other hand, the association of the prevalence of S. aureus with handler hygiene, hygiene of milk collection tank, the habit of picking nose, not washing hand after touching cash money, and covering the hair, mouth, and nose during milking in the current study is in agreement with the previous findings of Jorgensen et al. (2005), who reported the possibility of contamination of raw milk by S. aureus from human handling or the environment. Similarly, Fanta et al. (2012), also reported that pre-milking udder preparation and employing good milk handling practices play an important role in minimizing S. aureus contamination at the farm. In line with the present study, Ayandele and Adeniyi (2011) also reported high rates of microbial contamination arising from currency notes due to endemic bacterium families of S. aureus and arising from human normal skin flora.

Risk factors of S. aureus in "Ayib" and "Ergo"
The chi-square test revealed that the prevalence of S. aureus was significantly associated with the type of milk product container (P < 0.048), hygiene of product container (P < 0.003), hygiene of product handler (P < 0.001), environmental sanitation (P < 0.004), method of cleaning container (P < 0.009), hand wash after touching money (P < 0.015), and covering the nose, mouth, and hair while handling the products (P < 0.028) ( Table 5). These findings are in agreement with the report of Mamo et al. (2016), who reported that milk products sold in an open market were contaminated with S. aureus in Hawassa, Ethiopia. Similar to the current finding, the type of product container as a potential risk factor was reported by Omoe et al. (2005) and Soomro et al. (2003), who suggested stainless steel for handling products than plastic containers, which are prone to bacterial contamination of milk products as they are easily scratched during cleaning and cannot be effectively sanitized by heat.

Enumeration of S. aureus load in raw bulk tank milk
A total of 17 out of 61 (28.33%) of the samples contained S. aureus count that range from 1.25 × 10 4 CFU/mL to 1.92 × 10 4 CFU/mL (Supplementary table 2). There is no microbiological standard set for milk and milk products in Ethiopia. However, according to the European Commission's international microbiological criteria for dairy products, raw milk intended for manufacturing raw products (Directive 92/46/EEC), cow's milk intended for direct human consumption (Directive 92/46/EEC), and raw milk with greater than 2000 CFU/mL and 500 CFU/mL S. aureus, respectively, are of unacceptable quality Institute of Medicine (US) and National Research Council (US) (2003). This implies that 28.33% of the analyzed samples in the current study did not have recommended bacteriological quality both for direct human consumption as well as processing into raw milk products. The result of one-way analysis of ANOVA on the other hand showed that bacterial load in dairy farms and restaurants where no handwashing was practiced (mean 1.75 ± 2.12 SD; 95% CI: 0.94-2.55) was significantly higher (P < 0.05) than where handwashing practiced (mean 0.68 ± 1.59 SD; 95% CI: 0.10-1.27). Similarly, the bacterial count (mean 1.77 ± 2.12 SD; 95% CI: 0.99-2.55) in farms and restaurants where only water was used for handwashing was significantly higher (P < 0.05) than when soap and water were used (mean 0.58 ± 1.5 SD; 95% CI: 0.01-1.15) ( Table 6).
In line with the current finding, Reta et al. (2016), Yilma (2012), and Welearegay et al. (2012) reported that the dairy farms, where water and detergent were used for cleaning, had less contamination. This is because contamination can be avoided if the hand of milkers, utensils, or equipment is washed with water and detergents before use.

Antimicrobial susceptibility of S. aureus
The prevalence and level of antimicrobial resistance in veterinary medicine were increasing worldwide. Particularly, the spread of antimicrobial-resistant staphylococci is presenting a challenge to both human and animal health professionals (Roda and Fatema 2012). Of the total 66 S. aureus isolates subjected to 12 antimicrobials, the highest rate of susceptibility was recorded to norfloxacin 64 (96.97%) followed by  (Table 7). Nearly in agreement with the current finding, 100% (amoxicillin) and 64.7% (tetracycline) resistance was reported by Ayele et al. (2017). However, the 83.33% resistance against tetracycline in the current study is higher compared to the 32.5-82.2% previously reported resistances from various sources (Asrat et al. 2013;Elemo et al. 2017;Seyoum et al. 2018;Gebremedhin et al. 2022). The existence of an alarming level of S. aureus resistance to tetracycline, ampicillin, and amoxicillin was also documented in earlier studies in other countries by El-Jakee et al. (2008), Daka et al. (2012), andGentilini et al. (2002) suggesting a possible development of resistance from prolonged and indiscriminate usage of the antimicrobials. Cefoxitin-resistant isolates detected in the milk and milk products in the current study are in line with the findings of Elemo et al. (2017). The reason for the occurrence of antimicrobial-resistant S. aureus isolates could be due to the arbitrary use of antimicrobials, self-medication, and administration of sub-therapeutic doses of antimicrobials to livestock for prophylactic purposes and limited updating of the long-used drug groups Abraha et al. (2018). Moreover, the high prevalence of resistance to oxacillin and cefoxitin in the current study is suggestive of  the probable existence of MRSA, which contributes to the failure of empirical therapy due to their MDR nature and becoming a very serious public health concern (Lowy 1998).
Based on an analysis of antimicrobial resistance patterns of S. aureus isolates, 29 out of 66 isolates (43.94%) showed resistance to antimicrobials belonging to three or more classes of drugs, which is defined as multidrug resistance nature. All the MDR isolates (n = 29) showed resistance to the Beta-lactam and tetracycline class of drugs (Table 8). In contrast to this finding, a higher prevalence of 62.5% by Gebremedhin et al. (2022) and a lower prevalence of 2.4% by Regasa et al. (2019) and 6.2% by Kalayu et al. (2020) of multidrug-resistant S. aureus isolated from milk and milk products were reported, respectively. The emergence of resistance to numerous drugs represents public health hazard because foodborne outbreaks might be difficult to treat, and the group of multidrug resistance S. aureus in the food supply represents a reservoir for communicable resistant genes (Abraha et al. 2018).
The present study showed that a higher proportion of antimicrobial-resistant S. aureus, including the MDR isolates (72.41%), was detected in individual cow-level raw  milk followed by bulk tank milk (13.79%) than the milk products (Table 9). Since over 85% of the milk collected is consumed as fluid milk in Ethiopia CSA (2020), the detection of higher MDR S. aureus isolates in milk samples than the milk products implies a serious public health concern for the community.

Observation of hygienic practices
Personal observations at the different dairy farms and local markets have shown that farmers bring to open local markets only milk products, namely "Ayib" and butter for sale and not raw milk and "Ergo." "Ergo" was served to the consumers mostly at restaurants. Milking was done twice a day manually twice a day, early morning and in the evening. Transportation of milk to milk collection center and milk products to open local market for sale was done at room temperature, and no cooling facilities were used in both study area. Females were the ones engaged in milk product handling and selling. Different practices that can affect the hygienic quality of milk like shaking hands for greeting, counting cash money (both the coin and paper), and picking the nose and hair with the same hand that they use for handling the milk products were observed on the local markets. In addition, the habit of growing and decorating one's nails and milking without washing one's hands properly was frequently observed. Such malpractices might have contributed to the high prevalence of milk and milk products. Tick infestation with ticks belonging to the genera of Amblyomma and Hyalomma was very commonly seen in many of the dairy farms in the study areas.

Conclusions
The present study showed that S. aureus is widely prevalent in raw milk and milk products in Ambo and Bako towns. The potential risk factors for the prevalence of S. aureus were breed, age of cows, previous history of mastitis treatment, tick infestation, poor milkers' and farm hygiene, herd size, less frequent bedding removal, and loose housing system. There were variations in the hygienic status of the different farms, the experience of workers in the farms, the availability of dairy farm facilities, and farm inputs like detergents in the study areas. In most cases, the common hygienic actions taken during milking and milk handling practice were limited to udder washing. A higher proportion of antimicrobial resistance, particularly MDR, was reported in milk and milk products. Therefore, the high rate of isolation and high load of S. aureus necessitate a higher public health risk due to the widespread consumption of raw milk and its products in Ethiopia. Moreover, the large proportion of MDR S. aureus isolates may impede effective control of S. aureus udder infection in cows as well as present a public health risk due to the spread of drug-resistant zoonotic S. aureus. Thus, the use of antiseptics and disinfectants should be encouraged after washing hands and cleaning milk utensils, respectively. Awareness creation for dairy farm workers, milk product handlers, and milk collection center workers on the importance of good hygiene practices, control of tick infestation, rational use, and regular surveillance of antimicrobial resistance should be done. Moreover, to improve the milk production and milk processing quality in the study areas, improving the animal production system, veterinary services, and access to improved technologies are recommended. Further studies for phenotyping and molecular characterization should be done to evaluate the imminent danger posed by the microbes from milk and milk products.

Acknowledgements
The authors thank the staff members of Ambo University, Zoonosis, and Food Safety Laboratory for their technical support in the laboratory analysis. We also thank animal owners and animal health assistants of the different districts for their cooperation during sample collection.
Author contribution All authors made a significant contribution to the work including conception and study design. FT and LM collected data, and BM, EZ, and EJ supervised the data collection. FT and EZ performed the analyses and interpretation. FT, BM, and EZ drafted the manuscript. All authors read, critically reviewed, and approved the final manuscript and agreed on the journal to which the article will be submitted.
Funding Ambo University financially supported this research work. The funder had no role in the study design and collection, analysis, and interpretation of the data and in writing the manuscript.

Data availability
The datasets generated and analyzed during the current study are available from the corresponding author on reasonable request.

Consent to participate
The animal owners were informed about the purpose of this study, and informed consent was obtained from all individual participants included in the study. Participation in the study was on voluntary bases.

Consent to publication Not applicable.
Competing interests The authors declare no competing interests.