2.1. Farms, Animals, and Study Design

This study was conducted in Phnom Penh, Cambodia, on a train-smallholder dairy farm in a public university, using all milking cows from January 2023 to July 2023. The average milk yield ranged between 9.5 and 11 kg/cow/day. Three veterinarians were responsible for managing and preventing diseases, but the mastitis control program was initiated after the study began. The cows were crossbred Holstein-Friesian and were housed in a free-stall facility with concrete floors. They were milked twice a day using a 16-milking unit pipeline instrument in the separated milking parlor. The cow udder was washed with water and dried with a separate clean towel before milking. Due to the limited availability of dairy cow products, a local iodine disinfectant was used as a postdipping antiseptic agent, and no California Mastitis Test (CMT) was used to monitor subclinical mastitis in the herd. Acute and per-acute clinical mastitis were treated with systemic administration of antibiotics because there was a lack of intramammary antibiotic infusion in the market. There was no dry-cow therapy device available in the country. All of the milk from the farm was the sole source of raw milk for a self-pasteurizing milk company.

The study included all 21 milking cows. Trained veterinarians aseptically collected quarter milk samples every 2 weeks until the end of the study, as shown in Figure 1. The samples were transported with ice to the Laboratory of Bioagricultural Sciences Nagoya University-Cambodia Satellite Campus within 2 h for bacterial culture. Milk samples were cultured on a 5% blood agar plate at 37°C for up to 24 h. An IMI was identified when a single sample contained ≥1000 colony-forming units (cfu)/ml of one or two pathogens or when a sample of a clinical case contained ≥100 cfu/ml [3, 20]. Samples with more than three different colony morphology bacteria were considered contaminated and were not informative of IMI status. Each different morphology bacterial colony was subcultured in a 5% blood agar plate at 37°C for up to 24 h until its purification. The bacterial colonies were stored at −80°C until use.

2.2. MALDI-TOF MS for Bacterial Identification

The bacterial isolates were obtained from frozen archives and placed in tryptone soy broth (TSB). They were then incubated overnight at 37°C. After that, the samples were subcultured and purified on tryptone soy agar (TSA) and incubated at 37°C for 24 h. The bacterial colonies were submitted to the Office of Scientific Instrument and Testing at Prince of Songkhla University, Songkhla, Thailand, for bacterial identification using a MALDI-TOF mass spectrometer through the direct transfer protocol. The manufacturer’s recommendation was followed to identify bacterial fingerprints based on intact ribosomal proteins. The spectra were analyzed using Bruker Biotype software and Real-Time Classification 3.0 software, as described [21]. Only the isolates with a species-level identification score ≥2.00 and a log score ≥2.00 based on the manufacturer’s criteria were used.

2.3. Statistical Analysis

In a study by Leelahaponsathon et al. [4], an IMI episode is defined as a sequence of consecutive isolates of a specific bacterium from the same quarter. The duration of an episode is the time between the new IMI and the spontaneous cure. The spontaneous cure of an IMI episode is determined when an infected quarter does not show the same bacteria for at least 1 month or three immediate subsequent samplings, based on a 2-week sampling interval. The start and end dates of IMI were estimated halfway or 1 week between two adjacent sampling dates. The duration of infection was defined as the time between the midpoint between the last negative and the first positive sample and the midpoint between the last positive and the first negative sample, or the interval between the first and last IMI plus 14. The first positive sample was identified as the new infection time (New), and the subsequent samples until the last positive sample were identified as the persistent infection times (Persistent). There are two types of IMI: transient IMI and chronic IMI. Chronic IMI lasts for 28 days or more, according to the definition of chronic mastitis by Bonestroo et al. [22]. For statistical analysis, the main bacterial species must account for at least 10% of all isolates. Any other species were grouped as “Other” for analysis. Seasons were defined by sampling dates, including Cool (Jan–Feb), Hot (Mar–May), and Rainy (Jun–July).

The study presented the distribution of mastitis among different types of bacteria in terms of percentages. The proportions of uninfected, New, or Persistent infections were compared between seasons for each type of bacteria using Fisher’s Exact Chi-square tests. Fisher’s Exact Chi-square test was used to compare the proportion of transient and chronic episodes among bacterial groups. The duration of IMI among different types of bacteria was analyzed through survival analysis. The failed case was defined as the quarter that experienced spontaneous cure, while the censored case was defined as the quarter that was infected at the last observation. A logistic regression model was used to analyze the differences in pathogen ratios between transient and chronic episodes. The Kaplan–Meier survival function was used to describe the potential differences among different types of bacteria. The Cox proportional hazard model was used to compute the hazard ratio (HR) to compare the ratio of the spontaneous cure of the specified bacteria with the reference bacteria. The significance level was set at p  < 0.05, which indicates that the observed results were not likely due to chance. Additionally, a trend was observed with a p-value of less than 0.10, indicating a possible association between the variables that require further investigation to confirm any significant findings.

3.1. Bacteria Caused Mastitis and Proportions of Chronic Episodes

S. uberis, S. chromogenes, and S. gallolyticus were the main organisms responsible for mastitis on this farm. Out of a total of 174 episodes, the pathogens that had fewer than 17 episodes were grouped into a category called “Other.” This included other minor pathogens (n = 40), Gram-negative bacteria (n = 24), and other Gram-positive bacteria (n = 8). The minor pathogens included S. haemolyticus (13), S. saprophyticus (n = 6), S. borealis (n = 5), S. epidermidis (n = 4), S. xylosus (n = 2), C. amycolatum (n = 2), C. falseness (n = 2), and each of S. Barletta, S. cohnii, S. hominis, S. simulans, S. warneri, and S. xerosis. The Gram-negative bacteria included E. coli (n = 9), A. baumannii (n = 4), S. marcescens (n = 4), E. bugandensis (n = 2), A. towneri (n = 3), and each E. asburiae and B. diminuta, while the other Gram-positive bacteria were S. dysgalactiae (n = 2) and each of M. caseolyticus, W. cibaria, E. faecium, P. muralis, B. pumilus, and P. okeanokoites. Most pathogens in the “Other” category were transient except for 3, 1, and 1 chronic episodes of S. haemolyticus, C. amycolatum, and C. falsenii, respectively. The numbers of overall episodes (transient, chronic episodes) were 16 (9, 7), 40 (18, 22), 44 (28, 16), and 74 (69, 5) for S. gallolyticus, S. chromogenes, S. uberis, and Other, respectively. There were significant differences in the ratios of transient and chronic episodes among various pathogens (p  < 0.05). S. gallolyticus (odd ratio [OR] = 10.7), S. chromogenes (OR = 16.9), and S. uberis (OR = 7.9) had higher probabilities of chronic episodes than the pathogens in the “Other” category (p  < 0.05).

3.2. Effects of Seasons on Persistent and Transient IMI

The percentages of IMI isolates for each sampling period are visible in Figure 3. The highest overall IMI percentage was observed during the rainy season on days 182 and 196. This was due to the highest S. uberis IMI and Gram-negative IMI during this period. Table 1 shows the comparison of the New or Persistent IMI percentage of each pathogen among seasons. The New IMI of S. gallolyticus decreased among seasons (p  < 0.01), while the rest of the pathogens did not show significant association with seasons. The percentages of persistent isolates of S. chromogenes (22.1%) and S. uberis (9.7%) were significantly higher during the rainy season, indicating a significant association between IMI and season. On the other hand, the association between persistent IMI and seasons of S. gallolyticus was on the edge of being statistically significant (p = 0.06).

3.3. Mastitis Bacteria and Duration of IMI

Figure 4 displays a Kaplan–Meier curve that illustrates the distribution of IMI among different pathogens. The median time for a 50% spontaneous cure of IMI was found to be 14 days for S. uberis and Others, 21 days for S. gallolyticus, and 84 days for S. chromogenes. According to the results obtained from Cox’s model, pathogens were found to be correlated with spontaneous cures. In comparison to Others, S. chromogenes had a significantly lower rate of spontaneous cure with HR = 0.35, while no significant differences were observed for S. gallolyticus and S. uberis.

4.1. S. gallolyticus, an Emerging Mastitis Pathogen

In this study, 30 species of bacteria were found to be the causes of IMI. Still, only three species, S. gallolyticus, S. uberis, and S. chromogenes, were defined as the major cause in this herd. This study was the first report on finding S. gallolyticus and S. chromogenes as the main mastitis pathogens in a herd. For streptococci, the most relevant streptococcus species inducing bovine mastitis included S. agalactiae, S. dysgalactiae, and S. uberis [24]. Streptococcus gallolyticus belongs to the Group D streptococci, a large group of phenotypically diverse bacteria known as the S. bovis/S. equinus complex (SBSEC), which consist of safe-graded bacteria used in food-fermentation, commensal bacteria of the gut and opportunistic pathogens in both humans and animals [25]. S. gallolyticus infection in humans became a serious problem when several clinical studies demonstrated a strong association between invasive infections with S. gallolyticus and colon neoplasia in humans [25, 26]. S. gallolyticus, formerly known as S. bovis biotype I, is also part of the rumen biota but can cause disease in ruminants as well as in avian [27]. However, most reports indicated that S. gallolyticus was a rare mastitis pathogen. An isolate of S. gallolyticus has been reported as a mastitis pathogen in the milk of providers in Egypt [28] or one case of gangrenous mastitis caused by S. gallolyticus in Japan [29]. This study is the first report of a high prevalence of S. gallolyticus mastitis, which might be due to the poor mastitis control program on this farm in Cambodia, which is the starting dairy industry in the country. However, the use of conventional bacterial identification for mastitis [30, 31] might need to be improved in terms of the power of identifying bacterial species, especially in rare cases.

4.2. Emerging of the Mixed IMI

Our study revealed the dynamics of IMI, particularly a high number of mixed IMIs. Interestingly, most mixed IMIs included at least one chronic episode in those mixed IMIs. A poor mastitis control program can cause changes in regular teat scoring, resulting in increased teat-end hyperkeratosis and a higher risk of mastitis [32]. Cows with very rough teat end rings and dirty udders are more susceptible to IMI [33]. Our study also found that persistent IMI due to poor mastitis control programs could cause high mixed IMI. An in vitro study demonstrated that coculturing two mastitis pathogens resulted in one bacterium dominating the infection over time [14]. However, our study might have underestimated the prevalence of the mixed IMI in farms with high mastitis rates or poor control programs due to the exclusion of samples with three different bacteria, which were considered contaminated during the bacterial culture.

4.3. S. gallolyticus and S. chromogenes, Emerging Chronic Mastitis Pathogen

We found that certain groups of bacteria, including Gram-negative, NAS, and Corynebacterium, as well as other Gram-positive bacteria, had a lower risk of chronic IMI episodes compared to S. gallolyticus (OR = 10.7), S. chromogenes (OR = 16.9), and S. uberis (OR = 7.9). We also found that S. haemolyticus, a species in NAS, showed persistent episodes of IMI. Our study showed that S. gallolyticus had chronic IMI episodes, which is the first report of its kind (Figure 2). In humans, chronic colonization by S. gallolyticus is strongly associated with the occurrence of colorectal cancer, which may promote colonization of this bacteria in the gut by replacing commensal enterococci in their niche [34]. Therefore, the chronic episode of S. gallolyticus IMI might be due to changes in the mammary gland niche of this farm.

Our findings of chronic S. chromogenes IMI are supported by a previous longitudinal study in the Netherlands, which showed that 45% of IMI caused by S. chromogenes persisted for at least 28 days [35]. Bexiga et al. [36] found no significant differences in the mean duration of IMI caused by different NAS species. The different findings might be due to the differences in the phenotype of S. chromogenes strains on their survival within an environment with the other mastitis pathogens [14]. In addition, some S. chromogenes strains in this study appeared to be best adapted to survive in the udder gland. Persistent S. uberis has been reported in several studies, with some lasting more than 300 days of IMI [3, 4, 37]. Virulent S. uberis strains and low performance of the udder defense mechanism were suggested to be the cause of persistent S. uberis [4].

Table 1 reports that there were different types of IMI isolates in Cambodia during different seasons across all bacterial groups. During the study, the increased use of mastitis control programs, such as antiseptic teat dipping, dry cow therapy, and intramammary infusion, as well as subclinical mastitis monitoring with the CMT, suggested by an experienced veterinarian from Thailand, might have changed the environment of the mammary gland niche. This change might be the reason for the subsequent decrease of S. gallolyticus, while S. chromogenes and S. uberis became the main sources of mastitis pathogens on the farm. In general, the prevalence of mastitis varied at the international, national, and farm levels due to seasonal variations, farm management, and cow-related factors. Previous studies reported a seasonal variation in mastitis cases among different farms with high prevalence during the wet season (fall/winter) in comparison to the dry season (spring/summer) [38–40]. Other reports from tropical areas have also shown that mastitis from environmental pathogens was highest during the rainy season [41, 42].