Association between somatic cell count or morbidity of chronic subclinical mastitis and longevity in dairy herds in Eastern Hokkaido, Japan: a cross-sectional study

Abstract

An epidemiological analysis was conducted on production records in Hokkaido, Japan, to investigate the potential association between improved milk quality and longevity outcomes. The study found significant variations in herd somatic cell count levels and chronic subclinical mastitis morbidity based on geographical area and herd size. The analysis also revealed a positive correlation between herd somatic cell count and chronic subclinical mastitis morbidity. Although the hypothesis of a causal link between milk quality and longevity was examined, no such association was found. However, intensive assistance for identified high-risk areas and farms is expected to enhance overall milk quality.

One way to achieve sustainable milk production is by extending the longevity of dairy cattle [9], which has led to a demand for cattle that can be bred for long periods and have high lifetime productivity in recent years. Reports suggest that extending the productive lifespan of dairy cattle has several benefits, including reduced replacement costs for cows and a decreased environmental impact [4, 17]. However, within the Hokkaido region of Japan, the average parity of culled dairy cows was 3.37 in 2020, which is a decrease of 0.24 from 3.61 in 2010 [12]. This result suggests a shortening in the productive lifespan of dairy cows situated within the region of Hokkaido.

According to Pinedo et al. [20], in 2,054 herds in the United States, the leading cause of culling was death, followed by reproductive disorders, injuries, low milk yield, and mastitis. In Hokkaido, the most common causes of culling are death and udder diseases such as mastitis, which is a major problem [12]. Moreover, there has been a global downtrend in herd somatic cell count (SCC) during recent years [2]; however, the improvement of SCC at the herd level has encountered a standstill in Hokkaido, maintaining a steady level at approximately 220,000 cells/mL over the preceding decade [12]. In dairy cows, the SCC is a common index used to assess inflammation, particularly in the diagnosis of mastitis. It acts as a surrogate for quantifying the concentration of neutrophils present in the milk, according to the reference [21]. The SCC is a common measure used to assess the physiological well-being of dairy cattle, the quality of the milk, and its safety. It is also a key determinant in the overall assessment of latent mastitis conditions [27]. Typically, an increased herd SCC more than 200,000 cells/mL is caused by the influx of milk with increased SCC from specified cows that manifest mastitis conditions within the herd [1]. Henceforth, it is imperative to acknowledge mastitis and diminished longevity as concerns within the Hokkaido.

Mastitis is highly related to the shortening of productive lifespan, such as reduced reproductive performance [22, 26, 29]. Moreover, there is a negative correlation between longevity and SCC [30], and increased age and the number of parities increase the risk of mastitis development [11]. Good udder health is crucial for high milk production and longevity in dairy cows, and improving udder health is an essential strategy in dairy management [25]. Therefore, to achieve milk quality collateral and extended longevity, dairy farmers and advisers must understand the actual situation in the fields and take necessary measures to improve milk quality. Although some studies related to animal breeding in the Japanese dairy herd have explored the relationship between longevity and milk quality [23], there have been no observational studies that can provide insight into the actual situation at production sites or suggest measures to improve existing problems. If the results of this study can be used to solve existing clinical problems, dairy farmers and their advisors will be able to adopt evidence-based strategies. In addition, targeted interventions, for example on farms with significant problems, could significantly improve productivity across the area.

This study aimed to comprehend the actual milk production of dairy farms in Eastern Hokkaido through descriptive analysis of production records, to test the hypothesis that “Herds exhibiting superior milk quality also demonstrate a superior level of longevity” regarding the causal relationship between milk quality and longevity, and to identify remedial measures for dairy farmers based on the results of the epidemiological analysis that would lead to longer production life and improved milk quality.

This study was conducted following the guidelines for the care and use of laboratory animals at the Obihiro University of Agriculture and Veterinary Medicine.

Currently, 1.37 million dairy cows are raised in Japan, with 840,000 of them kept in Hokkaido [16]. Hokkaido (Coordinates: 43°N 142°E) is the northernmost island in the Japanese archipelago, and dairy farming is prevalent in Eastern Hokkaido due to the cool climate. The present study utilized Dairy Herd Improvement (DHI) records collected by the Hokkaido Dairy Milk Recording and Testing Association [10]. The production records of 168 dairy farms (total of 874,771 cows) in the Eastern Hokkaido region with continuous subscribers in DHI from April 1, 2018, to March 31, 2021, were used for the analysis. A total of 36 tests were conducted once a month during the research period. Four farms, where the coefficient of variation (standard deviation per average) exceeded 20% in the average number of cows raised during the period, were excluded from the analysis. Finally, 164 farms were included in this study. The median number of feeding cows on the 164 farms was 95 (interquartile range: 66–170).

In this study, herd-level longevity and milk quality were investigated and analyzed using descriptive and cross-sectional studies. The longevity index was defined as the ratio of the number of cows with four or more parities to the total number of cows, and it was set as the dependent variable. The average parity in the herd is a general indicator of herd longevity [24], and since the average parity of culled cows across Hokkaido in 2020 was 3.37 [12], a standard of four or more parities was considered for longevity in this study. Further, the study herds were categorized into two groups: a long-lived (>median) and short-lived (≤median) group, based on the median of the ratio of four or more parities to the total number of cows.

The independent variables used for milk quality indices were herd SCC and morbidity of chronic subclinical mastitis (CSM). To calculate herd SCC, the total SCC of an individual was first calculated by multiplying its SCC by the milk yield of the individual (Total Individual SCC). Next, herd SCC was defined as the sum of total individual SCC divided by the total milk yield of the herd. Herd SCC was further categorized into four groups based on quartiles. Regarding CSM morbidity, the duration of subclinical mastitis (SM) may be short or long-lasting, with the latter usually considered as CSM [28]. Therefore, in this study, cows with 200,000/mL or more SCCs were defined as those with SM. Those with SCCs of 200,000/mL or more in the previous and current months for the second straight month were defined as cows with CSM [2, 15, 21]. CSM morbidity was calculated for each month during the study period as the number of CSM cows per total number of cows in each month. The median of the monthly value (35 values in total) during the study period was used as the representative value for the farms. In addition, the morbidity of CSM was divided into four groups on the basis of the quartiles.

Area (farm locate area), milking system, and herd size were used as confounding variables. There was no evidence of multicollinearity between any of the confounding variables. Area was classified into three groups: Shikaoi-area, Shihoro-area, and Kitami-area, with each area’s name shown anonymously using the alphabet. The milking system was classified into three groups: Pipeline (high line), Parlor (low line), and Robotic. The herd size was classified into three groups: small (first quartile group), middle (second and third quartile group), and large (fourth quartile group), based on quartiles of the median number of cows.

Results were summarized using summary statistics and were analyzed by multiple logistic regression models. For all variables, summary statistics were calculated for each group, and the normality of data distribution was tested. Parametric one-way analysis of variance was used followed by between-group comparisons using Tukey’s method for multiple comparisons when the variable data had a normal distribution. If the data on the variable did not follow a normal distribution, non-parametric one-way analysis of variance was then used, and the Steel-Dwass multiple comparison tests were used to make comparisons between groups. Correlation analysis was conducted to evaluate the magnitude of the relationship between the two variables. Furthermore, multiple logistic regression analysis with confounding variables was performed to examine the impact of independent variables on dependent variables. Values were considered statistically significant at P<0.05. All statistical analyses were performed using SAS version 9.4 (SAS Institute Japan Ltd., Tokyo, Japan).

Table 1 shows the summary statistics of the herd-level longevity and milk quality indexes. In Table 1, the distributions of all variables (average parity, rate of primiparous cows, rate of cows with 4 or more parities, herd SCC, CSM morbidity) were non-normal. Univariable analysis revealed no difference in longevity indices in comparisons by area. The median herd SCC was 164,000 cells/mL in area A, 211,000 cells/mL in B, and 161,000 cells/mL in C. Area B had significantly higher herd SCC than A and C in comparisons by area. The median morbidity of CSM was 0.09 in area A, 0.13 in B, and 0.09 in C. Area B had significantly higher morbidity of CSM in comparisons by area. There were no differences between the groups in terms of longevity, herd SCC, or CSM morbidity in the comparison of the milking system. In the comparison of herd size, while no differences were observed in longevity and herd SCC between groups, the median incidence morbidity of CSM was significantly higher in the small group (0.14) than in the middle and large groups (0.10).

Table 1. Summary statistics of the longevity and milk quality indexes for each variable.

Table 2 shows the correlation matrices of longevity, herd SCC, and CSM morbidity in the study herds. No significant correlation was observed between longevity and herd SCC (r=0.0563, P=0.4744), nor between longevity and CSM morbidity (r=0.1006, P=0.2000). However, a significant correlation was observed between herd SCC and CSM morbidity (r=0.9093, P<0.0001).

Table 2. Correlation matrix between the longevity index and milk quality indexes by Spearman’s rank method.

	Rate of cows with 4 or more parities (parity ≥4)	Herd SCC	CSM morbidity
Rate of cows with 4 or more parities (parity ≥4)	1
P value
Herd SCC	0.0563	1
P value	0.4744
CSM morbidity	0.1006	0.9093	1
P value	0.2000	<0.0001

SCC, somatic cell count (×10³ cells/mL); CSM, chronic subclinical mastitis.

A multivariable logistic regression analysis was conducted, including confounding variables. Table 3 shows the relationship between herd SCC and longevity, while Table 4 shows the relationship between CSM morbidity and longevity. As shown in Tables 3 and 4, no significant relationship was observed between the dependent and independent variables. Among the confounding variables, significant differences were found in the milking system. The results showed that the pipeline herds had a higher rate of cows with four or more parities compared to the robotic herds.

Table 3. Evaluation of the relationship between herd somatic cell count and the longevity index by multivariable logistic regression.

Variable			Number of farms	OR	95% CL		P value
Independent variable
	Herd SCC
		<130	39	1.231	0.477	3.175	0.6673
		130≤ to <170	42	0.635	0.242	1.664	0.3554
		170≤ to <235	42	1.528	0.601	3.881	0.3728
		≥235	41	Ref
Confounding variable
	Area
		A	75	1.843	0.711	4.778	0.2083
		B	63	1.739	0.632	4.788	0.2840
		C	26	Ref
	Milking system
		Pipeline	78	3.461	1.011	11.847	0.0479
		Parlor	67	1.993	0.655	6.063	0.2246
		Robotic	19	Ref
	Herd size
		Small (<66)	41	0.764	0.249	2.347	0.6385
		Middle (66≤ to <170)	80	0.714	0.307	1.660	0.4342
		Large (≥ 170)	43	Ref

SCC, somatic cell count (×10³ cells/mL); OR, odds ratio; CL, confidence limits, Herd SCC was categorized into four groups based on quartiles, and herd size was categorized into three groups based on quartiles.

Table 4. Evaluation of the relationship between chronic subclinical mastitis morbidity and the longevity index by multivariable logistic regression.

Variable			Number of farms	OR	95% CL		P value
Independent variable
	CSM morbidity
		<0.070	43	0.607	0.235	1.571	0.3035
		0.070≤ to <0.100	37	0.687	0.251	1.879	0.4646
		0.100≤ to <0.150	45	0.564	0.226	1.409	0.2203
		≥0.150	39	Ref
Confounding variable
	Area
		A	75	1.711	0.663	4.418	0.2669
		B	63	1.617	0.593	4.404	0.3477
		C	26	Ref
	Milking system
		Pipeline	78	3.533	1.045	11.949	0.0423
		Parlor	67	2.158	0.718	6.484	0.1706
		Robotic	19	Ref
	Herd size
		Small (<66)	41	0.685	0.221	2.128	0.5132
		Middle (66≤ to <170)	80	0.670	0.291	1.547	0.3484
		Large (≥170)	43	Ref

CSM, chronic subclinical mastitis; OR, odds ratio; CL, confidence limits, CSM morbidity was categorized into four groups based on quartiles, and herd size was categorized into three groups based on quartiles.

This study aimed to analyze production records to further improve milk quality and longevity. In Table 1, a notable disparity in herd SCC and CSM morbidity was observed among the three areas. Regarding herd size, the small herd exhibited a higher prevalence of CSM morbidity compared to the other two herds. Herd-level SCC exhibited variability across areas, which can be attributed to divergent climatic conditions and pathogen distribution prevalent in each area and differences in herd size and barn type [6, 18, 19]. Therefore, mastitis prevention and control programs should be tailored to farms based on herd size and management systems such as barn type. In addition, farmers in areas characterized by the presence of organizations committed to improving dairy production, such as agricultural cooperatives, have a wide range of support available to them. Conversely, in areas characterized by insufficient support, farmers face a lack of choice. It is therefore possible that differences in support systems may explain the regional differences in herd SCC and CSM. Barkema et al. [2] reported that should be made to ensure that the incidence of chronic mastitis in lactating bovines remains below 5%, with an upper threshold of 10%. Table 1 shows the presence of areas with CSM morbidity above 10%. In addition, it was noted that even in areas where CSM morbidity was below 10%, several herds had CSM morbidity above 10%. These results suggest that there is room for improvement in milk quality in the target areas. In the future, targeted support to areas or farmers identified through descriptive analysis, and strengthening the support system in the area could effectively improve productivity.

In Table 2, Lukas et al. [13] reported a robust correlation between bulk SCC and SM morbidity in herds, and our findings corroborate their results. CSM cows had a higher SCC than SM or healthy cows, and the changes observed in milk depended on the type of pathogen [15]. These results suggest that on-farm interventions that primarily target CSM cows and CSM morbidity in herds can effectively improve herd SCC. Regular monitoring of monthly records, and detection of individuals with persistently high SCC and causative pathogens can help in devising optimal management strategies for CSM cows. Farmers and advisors also need to invest time and effort in improving milk quality. Therefore, it is possible to effectively improve herd SCC by focusing on identifying CSM cows and their pathogens and reducing CSM morbidity. In addition, measures, such as early dry-off and culling in CMS cows, can significantly enhance the overall herd SCC and aid in effective mastitis prevention [2].

In Tables 3 and 4, the hypothesis “herds exhibiting superior milk quality also demonstrate a superior level of longevity” was tested, concerning the causal correlation between herd-level milk quality and longevity. No relationship was found between the independent and dependent variables. Among the confounding variables, the milking system showed a significant association. There are several previous reports on differences in longevity due to differences in milking systems, and it was reported that pipeline system herds have a longer productive life than robotics and parlors [5, 14]. In addition, it has been reported that herds managed with pipeline milking systems have lower stress levels than herds managed with intensive milking systems [5]. The differences in longevity observed between milking systems in this study may be due to differences in stress responses. Mastitis is a significant factor contributing to culling in dairy cows [7], and poor milk quality and udder health lead to culling and decreased longevity [3, 25]. Moreover, an increased age and parity elevate the risk of mastitis and culling in dairy cows [3, 14]. Despite numerous prior investigations indicating a strong relationship between udder health and productive longevity, this association was not observed in this study. This may be due to insufficient adjustment for confounding factors in the present study, which may have masked the association between milk quality and longevity. The production of heifers and the status of the holding vary considerably from farm to farm. Some farms do everything themselves, while others outsource everything. The risk of culling a dairy cow is significantly influenced by the self-breeding status of replacement heifers on a dairy farm [4]. Some farmers who have an abundance of replacement heifers deliberately choose to build young herds to avoid the risk of mastitis and poor fertility. The resale of pregnant cows as dairy cows to other farms is part of this selective culling. In the future, it may be possible to conduct analyses that more accurately reflect the actual situation in Hokkaido by adding more detailed factors that affect culling risk, such as heifer supply, as confounding factors. In addition, the cow-level associations of health and longevity may not hold true at the herd level [8].

The following limitations should be considered when interpreting these data. First, the data were predominantly collected from DHI subscribers; hence, the sample population may not fully represent all farms within the study area. Moreover, this study analyzed 164 herds using a cross-sectional design, which may limit the generalizability of the results. Further research with a larger dataset is necessary to enhance our comprehension of this subject matter.

In this study, the DHI records of 36 tests carried out over three years were used for analysis. In many cases in the fields, the herd level rather than the cow level is the primary unit for the first milk quality intervention by the advisors. We thought that it would be easier for the advisors to take action to improve milk quality if we had an introductory analysis at the herd level. As an initial study, we therefore carried out a study at the herd level. In the future, however, we believe there is a need for further research at the cow level. However, a relatively short period of three years was used to collect the production data in this study. We believe it will be necessary to collect production records over a longer period to carry out a longitudinal study at the cow level.

In summary, this study conducted an epidemiological analysis using production records to obtain new findings that could lead to improved milk quality and extended longevity. The descriptive analysis revealed a significant difference in herd SCC and CSM morbidity among herd area groups, as well as significant differences in CSM morbidity by herd size. There was also a positive correlation between herd SCC and CSM morbidity. However, analytical epidemiology did not find any association between milk quality and longevity. Targeted intervention support aimed at areas and farmers exhibiting problematic outcomes, specifically the identification of CSM individuals displaying persistently elevated SCC levels, as well as the identification of affected quarters and underlying pathogens, hold the potential to bring about comprehensive improvements in milk quality throughout the entire region. Additional research, including studies at the cow level, is needed in the future.

CONFLICT OF INTEREST

The authors declare no conflicts of interest associated with this manuscript.