1. Introduction
The three weeks before and after calving are called the “transition period” [
1,
2], during which dramatic changes occur in the cows. This means that the fetus grows significantly in the three weeks before calving, and after calving, more energy is required for milk synthesis. At this time, if the energy supply from feed intake is insufficient, cows can easily fall into a negative energy balance (NEB) [
3,
4]. The cow’s body attempts to adapt to the NEB by utilizing the carbohydrates, lipids, and proteins stored in the body. However, if the reaction is not sufficiently successful, it is assumed that the likelihood of peripartum diseases such as milk fever, retained placenta, and displaced abomasum increases.
Ketosis is known as a fundamental pathology closely linked to peripartum diseases [
5]. It is divided into clinical and subclinical ketosis (SCK). A subclinical state is a condition in which ketone body levels are elevated without clinical symptoms. The gold standard for SCK is serum or plasma β-hydroxybutyrate (BHBA) concentration of 1.2 mM or more [
6,
7]. Furthermore, SCK is classified into two types depending on the time of onset. Type I usually occurs 3 to 6 weeks postpartum (SCK I), and type II occurs within 2 weeks postpartum (SCK II) during early lactation [
8]. They were named because of their similarity in pathogenesis to human diabetes mellitus, but recent studies have indicated that their pathogenesis is not consistent, and Mann and McArt (2023) discourage the use of the terminology of type I and type II ketosis [
9]. On the other hand, SCK II, which occurs in the early postpartum period, is known to be the underlying disease for many perinatal diseases and does not have a favorable prognosis. But, SCK I is known to have a good response to treatments. We thought that grasping the prevalence of SCK I and SCK II would be useful information for understanding herd health in this study.
Previous studies have shown that the prevalence of SCK varies by country and time of occurrence. Briefly, in the United States, the prevalence within 3–16 days in milk (DIM) was calculated to be 43.2% (range from 26 to 56%), and the peak prevalence of SCK occurred at 5 DIM [
10]. A study in 10 European countries reported that the prevalence within 2–15 DIM was 21.8% (range from 11.2 to 36.6%) [
11]. In Slovakia, the prevalence of SCK within 20 DIM was calculated to be 18.5%, whereas it was 14.1% for 21–90 DIM [
12]. However, there are no reports investigating the prevalence of SCK in Japan.
Several individual factors are known to be associated with the occurrence of SCK, including the body condition score (BCS), parity, dystocia, twins, and calving season [
11,
13]. Walsh (2007) reported that the prevalence of SCK was 29.8% in tie stalls (TS) and 15.4% in two-row free stalls (FS) [
14]. Furthermore, Garzón-Audor and Oliver-Espinosa (2019) estimated that farms with 150 or more cows were at higher risk of SCK occurrence than farms with fewer cows (relative risk: 1.92; 95%CI: 1.04–3.57) [
15]. These findings suggest that the management system may be related to the incidence of SCK. Japanese dairy cows are kept in a wide variety of systems, ranging from TS of small herd size to FS and free barns (FB) of medium and large herd size. However, the relationship between the occurrence of SCK and management systems such as housing and feeding is unclear.
The aim of this study was to determine the prevalence of postpartum SCK in Hokkaido, Japan, and evaluate its characteristics at individual and herd levels through epidemiological surveys.
2. Materials and Methods
2.1. Animals and Their Information
This epidemiological study was conducted as an observational study between October 2012 and March 2014 in 18 regions of Hokkaido, Japan, through regular health examinations of blood profiles and physical condition by 11 veterinarians. The 108 farms (size range: 23–800 cows) that were insured by the Hokkaido Agricultural Mutual Aid Association and agreed to participate in this study were selected as survey farms. A total of 1407 clinically healthy Holstein cows, free of clinical disease for at least one week prior to blood sampling, were studied. Each veterinarian performed an average of 2.1 visits per farm (SE: 0.1, range: 1–10). The average number of sampling cows per farm was 12.9 (SE: 0.8, range: 1–63). Records for farm name, individual identification number, date of calving, and parity were obtained from the database of the Livestock Mutual Aid System of the Agricultural Mutual Aid Association. This accurately reflects the traceability information of cows as required by domestic law. Dairy farm information such as housing system (TS and FS/FB), feeding system (total mixed ration (TMR) and component feeding), and feeding frequency was also obtained through oral investigation conducted by veterinarian at the time of blood sampling.
Blood BHBA concentration ≥ 1.2 mmol/L was diagnosed as SCK II within 3–14 DIM and as SCK I from 15 DIM. On the other hand, cows with BHBA < 1.2 mM were classified as healthy controls. This study investigated the relationships between parity (1, 2, 3, 4, and 5 or more), BCS (≤3.25 or 3.50≤), RFS (≤2 or 3≤), housing and feeding systems, and number of cows with SCK.
Farms used for herd analysis were those with 12 or more cows sampled per farm [
7]. Herds with a combined percentage of SCK I and SCK II of less than 10% were classified as SCK-negative herds, those with percentages between 10 and 25% were classified as SCK-alert herds, and those with one of 25% or more were classified as SCK-positive herds [
7]. This study investigated the relationship between SCK-negative, -alert, and -positive herds and management systems (average number of cows, housing system, feeding system, and feeding frequency).
Table 1 shows the collection method, type of variable, and definition used in individual and herd analyses.
All animals were treated appropriately following the Laboratory Animal Control Guidelines of Rakuno Gakuen University, which essentially conforms to the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health in the United States [
16].
2.2. Blood Sampling and Physical Examinations
Blood samples were collected from the coccygeal vein only once per cow during 3–90 DIM. Plain collection tubes were used to collect serum for measurement of the concentrations of BHBA and nonesterified fatty acids (NEFA). Collection tubes with sodium fluoride were used for measuring the glucose concentration. Collected blood was stored immediately at 4 °C. For serum separation, centrifugation was performed at 2000×
g for 15 min within 6 h after collection, and sera were stored at −30 °C until analysis. The concentrations of serum BHBA, NEFA, and glucose were measured using an automatic analyzer (Bio Majesty JCA-BM2250; JEOL, Tokyo, Japan) with kits for them (N-assay NEFA; Nittobo Medical, Tokyo, Japan, 3-OHBA-TR; Kishimoto Clinical Laboratory, Sapporo, Japan, and GLU-TR; Kishimoto Clinical Laboratory) from the Kishimoto Clinical Laboratory. Cows were scored for BCS [
17] and the rumen fill score (RFS) [
18] together with blood collection by same veterinarian per farm. The BCS was rated on a 5-point scale with increments of 0.25, with higher scores indicating higher body fat stores. Plus, RFS was also assessed on a 5-point scale in increments of 1, with higher scores indicating better dry matter intake. All veterinarians used the same scoring system.
2.3. Statistical Analysis
All statistical analyses were performed using SPSS version 27.0 (IBM Japan Co. Ltd., Tokyo, Japan). Continuous data were transformed as necessary. A log10 transformation was applied to BHBA and NEFA. Other outcomes did not require transformation. Normality of data was assessed using the Shapiro–Wilk test. The comparison between groups in parametric data was analyzed by ANOVA and Games–Howell test. The analysis of non-parametric data was performed by Kruskal–Wallis test.
The chi-square test was used to compare categorical data (BCS, RFS, parity, housing system, feeding system, and feeding frequency). When significance was calculated to be present by the chi-square test, adjusted residuals below −1.96 were rated as significantly low frequency of occurrence and above +1.96 as significantly high frequency of occurrence [
19].
The analysis of blood metabolites in healthy and SCK cows was assessed by a linear mixed model. The statistical model was the following:
where
Ypq is the observed value (BHBA, NEFA, Glucose);
is the overall mean;
Typep is the fixed effect of the
pth class of type (
p = healthy control, SCK I, SCK II);
Herdq is the random effect of the
jth herd (
q = 1–108); and
epq is the residual error. Bonferroni’s multiple comparison test was used when comparing with groups.
Statistically significant differences were assessed at p < 0.05, and a trend was assessed at p < 0.10.