Changes in Holstein Heifer Salivary Cortisol Concentrations and Behavior after Regrouping

: The objective of this study was to analyze the effect of regrouping on Holstein heifer salivary cortisol concentrations and behavior. Eighteen heifers (192.8 ± 13.6 days of age) were used during this study. Each of these heifers was introduced into a pen of older existing heifers. The heifers were assigned to four groups that corresponded to each of the four regroupings. Saliva samples were collected the day before regrouping (baseline; pre-regrouping), the day of regrouping, and the day after regrouping (post-regrouping). Video cameras continuously recorded from the hour before each regrouping through one day after each regrouping. Salivary cortisol concentrations were higher than the baseline for novel and existing heifers on the day of regrouping and one day post-regrouping ( p = 0.01). More aggressive and agnostic behaviors occurred during Regroupings 1 and 4 than during Regroupings 2 and 3. Novel heifers spent more time standing ( p = 0.05) and drinking ( p = 0.05) than the existing heifers and less time lying ( p = 0.05), but no other differences were observed between the behavior of existing and novel heifers. The salivary cortisol results of this study demonstrate that regrouping is a stressful event for both novel and existing dairy heifers.


Introduction
As the global population continues to increase, the importance of developing practices that maximize reproductive efficiency and milk production while minimizing the amount of cattle and space needed is crucial [1].Many management systems have been developed, but these will continue to be enhanced to encourage growth and proficiency in the coming years.As a result, much is required from a dairy cow's body every day.For instance, a dairy cow must have the ability to maintain pregnancy and an appropriate body condition score while producing high amounts of milk.In addition, dairy cows experience multiple potential stressors frequently, such as movement in and out of the milking parlor, handling for pregnancy diagnosis and health exams, and movement between pens, possibly introducing social interactions with other cows.Some of these demands indirectly expose dairy cattle to various stressors, such as social stress, which can have negative effects on their fertility as well as their overall health [2].Additionally, the presence of stress, as indicated by endocrine responses, is correlated with a decrease in milk production [3].Stress among animals is often demonstrated through both physiological and behavioral changes and can be broadly defined as an animal's inability to cope with a situation while maintaining its full genetic potential [4].Long-term stress inhibits dairy cattle from reaching their full genetic potential.In turn, this can lead to financial, economic, and production deficits.More importantly, it causes decreased animal welfare and, therefore, should be addressed.
Dairy cattle are naturally curious and social herd animals.They establish social hierarchies.For example, calves form social bonds that appear to be related to how long a particular group of cattle have been housed together [5].Every time novel cattle are introduced into a previously established group, this drastically changes the herd dynamic and new hierarchies are eventually formed.Dairy cattle may encounter forms of social stress due to social hierarchy patterns when being introduced to new social groupings [6].For example, during heifer regroupings, changes in behavior were observed to be statistically significant among the regrouped heifers [5].The regrouped animals spent more time standing without moving, less time lying, more time sniffing the pen and displayed more fight-oriented and aggressive behaviors than the control heifers [5].Similar results were obtained in another study utilizing lactating dairy cows [7].Furthermore, the study concluded that the amount of social regroupings that the heifers endured did not change their behavior or habituate them to the situation [5].Dominance hierarchies were not established any faster the more regroupings that the heifers endured [5].In an additional study examining heifers, when the heifers were separated from the older, more dominant cows, their feed intake increased, and the amount of time they spent lying also increased, indicating less competition for feed and water [8].Social regrouping among dairy heifers can cause social stress due to aggressive behaviors and dominance patterns that, due to increased competition for resources, can hinder the animal from proper and easy access to feed, water, and lying time.
Social stress has also been seen to affect reproductive efficiency in dairy cattle.Data from a behavioral study on herd dynamics found that the time from calving to conception was longer among cows that had a lower or submissive social status [4].Additionally, it generally took more artificial inseminations for submissive cows to become pregnant than for cows who had dominant status [4].Attributing this subfertility to stress can be supported by endocrine responses.While there is still much to be learned about the full function of the endocrine pathway in identifying stress, research clearly demonstrates that cortisol concentrations can be a key indication of stress and affect the animal's behavior [9].Cortisol is a steroid hormone secreted by the adrenal gland in response to stressors.It provides negative feedback to the hypothalamus, inhibiting the proper secretion of Gonadotropin Releasing Hormone (GnRH) and, therefore, the secretion of Luteinizing Hormone (LH) [4].GnRH and LH are vital to a cow's fertility and may have far-reaching consequences, especially regarding development and proper reproductive benchmarks in heifers that may be experiencing chronic stress.
There are several ways to measure cortisol concentrations in cattle, including plasma cortisol, hair cortisol, and salivary cortisol.Plasma is perhaps the most common method scientists use to determine cortisol concentrations.However, cortisol concentrations in the blood change quickly in response to stressors.Due to the potential stress caused by handling the animal, unless the blood is obtained almost immediately, the cortisol concentration may be misleading [10].Plasma collection is considered an invasive procedure while obtaining a saliva sample is generally believed to be a simple and low-stress collection method.Studies have shown that cortisol concentrations in plasma and saliva are directly correlated [11,12].While one study found there to be no lag time between peak concentrations of plasma and salivary cortisol, another study demonstrated that there is likely a 10 min lag from when plasma cortisol peaks to when salivary cortisol peaks [11,12].During an additional study with dairy cattle, researchers obtained a blood sample followed by a saliva sample that took three minutes to obtain and still observed a positive correlation between the two [10].Thus, while more research needs to be conducted to fine-tune this, it is logical to hypothesize that salivary cortisol concentrations tend to peak within 3-10 min following a stressful event.
High cortisol concentrations have been linked to immunosuppression, lack of fertility, and other issues in dairy cattle, making it an accurate partial indication of the overall well-being of the animals [13].Therefore, it is important to continue research to find the most accurate way to measure cortisol to properly assess dairy cattle welfare.Transitional periods, including moving cattle from one pen to another, are often coupled with a change in social grouping.The change in environment, as well as the change in social hierarchy, causes distress in cattle.The objective of this study was to analyze the effect of regrouping on Holstein heifer salivary cortisol concentrations and behavior.The results from this study will be used to develop heifer regrouping management practices that minimize distress during this transition.

Materials and Methods
All methods involving animals were approved by the Washington State University Institutional Animal Care and Use Committee (ASAF# 6218).This study was conducted from October to December 2020.No signs of thermal stress were observed in the heifers.The use of a sample size calculator (https://clincalc.com/stats/samplesize.aspx(accessed on 3 September 2020); 80% power; 0.05 alpha) led to eighteen Holstein heifers (192.8 ± 13.6 days of age at regrouping for novel heifers) being enrolled into this study.All animals were healthy upon enrollment into the study, and no health ailments were detected among the enrolled heifers.All heifers were housed in a building with three concrete walls, a feed bunk area with head gates along the front of the building, and access to an outdoor area (125 m 2 per pen).Each pen was 49 m 2 , with 30.89 m 2 of resting area bedded with wood shavings.All heifers had ad libitum access to water and a total mixed ration.Four novel groups of heifers were introduced into a pen of approximately 15 existing heifers (214.6 ± 16.5 days of age) throughout the course of this study.Two of the novel groups consisted of three heifers, and the other two novel groups consisted of six heifers.A novel group was introduced once every 14 ± 2.16 days, and each novel group of heifers was considered a new regrouping.These regroupings were labeled as Regrouping 1, Regrouping 2, Regrouping 3, and Regrouping 4. Heifers that were novel for Regrouping 1 were considered existing heifers for Regrouping 2, Regrouping 3, and Regrouping 4. Heifers that were novel for Regrouping 2 were considered existing heifers for Regrouping 3 and Regrouping 4. Heifers that were novel for Regrouping 3 were considered existing heifers for Regrouping 4. The day before a novel group was moved into the existing heifer pen was considered to be Day −1, and Day 0 was considered the first day that the novel heifers were introduced into the heifer pen.Heifers were regrouped by 1000 h on Day 0.
Saliva samples were obtained for the novel heifers for each regrouping on Day −1, Day 0, and Day 1, and samples were obtained for the existing heifers on Day 0 and Day 1 for each regrouping.All saliva samples were collected using the SalivaBio's Children's Swab (SCS) System for Animals (Salimetrics, State College, PA, USA).Saliva was collected from the novel heifers on Day −1 between 10:30 and 11:30 h.On Day 0, saliva collection began 30-40 min post-regrouping, and on Day 1m saliva was collected between 10:30 and 11:30 h.Careful care was taken to use a minimal amount of restraint (using head gates located at the feed bunk) and to complete the collection process within three minutes per heifer.Saliva was collected two hours after feeding.The swab was placed inside the corner of each heifer's mouth near pooling saliva by the gum for 90 s, which resulted in the collection of 1-2 mL of saliva.After saliva was collected for all of the heifers involved in each regrouping, the tubes were centrifuged within 30-60 min of collection at 1500 RPM for 15 min.After this, the swab was removed from the tube, and the saliva samples were frozen at −20 °C until the day of analysis.The cortisol concentrations were obtained using enzyme-linked immunosorbent assays (Expanded Range High Sensitivity Enzyme Immunoassay Kit, Salimetrics, State College, PA, USA).The ELISA kit's analytical sensitivity was 0.07 ng/mL.The intra-assay CV was 5.8%, and the inter-assay CV was 8.2%.
Four video cameras (Brinno TLC 200) were positioned within the heifer pen, with one facing the feed bunk, one facing the back wall, and two facing the front of the pen with views of the water trough and each side of the pen to provide a comprehensive view.The cameras were set to record continuously beginning one hour before each regrouping on Day 0 until the end of Day 1.The decision to only record the video footage for 48 h was based on a previous study that concluded that dairy cattle typically return to their baseline behaviors after two days [14].Behavior observations were recorded from the 48 h of video footage, but only during daylight hours.Individual animals were identified by using weatherproof animal paint on both sides of each heifer.
Two trained observers viewed the video footage using a VLC Media Player and documented behaviors.Their inter-observer reliability was 91.3%.To evaluate daily time budgets for the heifers, all-animal instantaneous scan sampling for specific behaviors was recorded at five-minute intervals for novel and existing heifers (Table 1).During the instantaneous scan sampling, the following behaviors were focused on: feeding, lying, standing, sniffing, locomotion, and drinking.All-animal scan sampling was determined to be the most efficient and accurate way to find the average percentage of animals observed performing these behaviors during any given observation period.It was decided that the above behaviors could be classified as a state defined as a continuous or ongoing behavior where duration should be considered [15].Scan sampling has been determined to be an effective way to analyze behaviors that are a state [15].Focal group, alloccurrence sampling was used to record specific social behaviors, with the focal group being the novel heifers for each regrouping (Table 2).The behaviors for the focal animal all-occurrence sampling was recorded continuously for the span of each day.The analyzed behaviors included head-butting actor (heifer performing the behavior) and reactor (heifer responding to the actor's behavior), charging actor and reactor, fighting, displacement actor and reactor, grooming actor and reactor, mounting, standing to be mounted, mounted run, chin rest actor and reactor, and stereotypic behaviors.Focal group all-occurrence sampling was determined as the best approach for the above behaviors because the frequency of each event for each of the novel groups was desired and, thus, the data would be considered nominal, which is appropriate for this sampling method [15].For scan sampling data, the proportion of existing heifers exhibiting the behavior each hour was calculated for further analysis.The proportion of novel heifers exhibiting the behavior each hour was also calculated.For focal group sampling, the number (frequency) of occurrences of the behavior exhibited each hour was calculated.Animal performs behaviors such as pacing (animal moves back and forth with no end destination and a continuation of motion), oral manipulation (an animal performs oral behaviors that do not serve an essential function), over-grooming (an animal continues to lick themselves until they start to lose hair)

SB (P, O, OG)
The data were tested for normal distribution using PROC UNIVARIATE in SAS (SAS 9.4, Cary, North Carolina, USA).Salivary cortisol concentrations were statistically analyzed using a mixed model ANOVA in SAS.Repeated measures were used in the model as multiple cortisol samples were obtained from each heifer (n = 18) over the course of the study.The experimental unit was each heifer.The independent variables were treatment (novel or existing) and day.Behavior data recorded using the all-animal scan sampling were analyzed using PROC GEN MOD in SAS for each behavior, with a Poisson distribution and log link.The dependent variable was the specific behavior being analyzed with the independent variables being treatment and day.The experimental unit was the regrouping number (n = 4).The significance was established to be p ≤ 0.05.The means are reported with the standard error.Behavior data recorded using the focal group and alloccurrence sampling was analyzed using PROC MIXED in SAS.The dependent variable was the specific behavior being analyzed, with the independent variable being day.The experimental unit was the regrouping number (n = 4).

Salivary Cortisol Concentrations
A significant interaction was detected between the treatment group and day (p = 0.01; Figure 1).On Day −1, which was considered the baseline cortisol concentration for novel heifers, the mean concentration was 1.18 ± 0.28 ng/mL (Figure 1).Salivary cortisol concentrations in novel heifers significantly increased between the baseline (Day −1) and Day 0 (p = 0.01) and Day 1 (p = 0.01).A significant increase in salivary cortisol concentrations occurred between the baseline cortisol concentration (Day −1) of the novel heifers and of the existing heifers on Day 0 (p = 0.001) and Day 1 (p = 0.005).
No significant difference was detected between existing and novel heifer cortisol concentrations on the same day.No significant difference occurred between existing heifer and novel heifer cortisol concentrations on Day 0 (p = 0.69), and no significant difference occurred between existing heifer and novel heifer cortisol concentrations on Day 1 (p = 0.76).No significant change was detected within the novel heifer cortisol concentrations on Day 0 and Day 1 (p = 0.65).

Heifer Time Budgets
Too few observations of sniffing and locomotion behaviors were recorded, so statistical analysis of those behaviors was not possible.For feeding behavior, no significant differences were detected among days (p = 0.81) or between treatments (p = 0.12).On average, 42.4 ± 2.4% of novel heifers and 42.0 ± 2.5% of existing heifers were observed feeding during the same observation period.
For lying behavior, a significant difference existed between treatment groups (p = 0.05), and a trend existed among days (p = 0.06; Figure 2).The mean percentage of existing heifers lying was 39.3 ± 2.6%, while the mean percentage of novel heifers lying was 36.5 ± 2.8% (Figure 2).A greater percentage of all (novel and existing) heifers was lying on Day 1 than Day 0, with the means per day as follows: Day 1 was 41.8 ± 2.5%, and Day 0 was 31.6 ± 2.7%.For standing behavior, significant differences were detected between treatment groups (p = 0.05) and among days (p = 0.05; Figure 2).The mean percentage of novel heifers standing was 37.5 ± 3.2%, and the mean percentage of existing heifers standing was 20.1 ± 1.2%.There was a greater percentage of all heifers standing on Day 0 than on Day 1, with the mean percentages per day as follows: Day 0 = 30.5 ± 2.9% and Day 1 = 27.1 ± 2.7%.
No significant differences were detected for drinking behavior among days (p = 0.49).A significant difference existed between treatment groups (p = 0.05).The mean percentage of novel heifers drinking was 25.1 ± 3.0%, and the mean percentage of the existing heifers drinking was 10.0 ± 0.5%.

Focal All Occurrence Sampling Behavior
For the number of occurrences of novel animals exhibiting head-butting behavior as an actor, no significance was observed among days (p = 0.70).As for fighting behavior exhibited by novel heifers, no significance was detected among days (p = 0.46).No significant difference was observed among days for novel heifers reacting to another heifer charging (p = 0.59).Too few observations of novel heifers charging at other heifers were recorded; therefore, no statistical analysis could be performed.
For chin resting behavior exhibited towards novel heifers, no significant difference was observed among days (p = 0.47).No significant difference was observed among days (p = 0.32) for novel heifers exhibiting chin resting behavior.In regard to novel heifers receiving grooming from another heifer, no significant differences were observed among days (p = 0.47).For the number of occurrences of novel heifers grooming another heifer, no significant differences were observed per day (p = 0.40).
For the number of occurrences in which novel heifers displaced another heifer, no significant differences were observed between the days (p = 0.77).
No significant difference was observed between days (p = 0.25) for the number of displacements during walking that the novel heifers received.As for the number of displacements that occurred while walking that novel heifers displayed towards other heifers, no significant differences were observed between days (p = 0.60).The number of dis-placements novel heifers received while lying down had no significant differences observed between days (p = 0.81).Too few observations of novel heifers displacing heifers that were lying down, mounting, standing to be mounted, mounted run and stereotypic behaviors were recorded; therefore, statistical analyses of these behaviors were not conducted.

Discussion
The results of this study show the short-term physiological and behavioral effects of regrouping on dairy heifers.To our knowledge, this is the first study to analyze the effects of heifer regrouping on salivary cortisol concentrations.All of the salivary cortisol concentrations in this study were within the range obtained for dairy cattle under normal conditions in a previous study, but the concentrations in the present study were generally lower than those observed under stressful conditions in both cows and calves [10].In the previous study, the mean baseline for salivary cortisol concentration for the calves was approximately between 0 and 1 ng/mL, and for the cows was approximately between 1 and 4 ng/mL [10].Following ACTH administration, the range for the salivary cortisol concentration of the calves was approximately 2 to 11 ng/mL, and during milking, the range for the salivary cortisol concentrations of the cows was approximately 5 to 15 ng/mL [10].However, the previous study subjected the calves to ACTH administration using a dosage intended to achieve what is known to be a maximum cortisol response similar to what would be expected after a severe stressor [10].When looking at the salivary cortisol concentrations of beef calves in painful and stressful situations, the salivary cortisol concentrations of the current study are of a very similar nature to what that study found when sampling within four hours of a stressful event (1.22-3.66 ± 0.53 ng/mL) [16].
The baseline salivary cortisol concentrations for the novel heifers in this study were significantly lower than the salivary cortisol concentrations for both the novel and existing heifers on the day of regrouping and the day after regrouping.This was to be expected as there is a known increase in aggressive behaviors and agnostic interactions on the day of regrouping, leading to social stress [5].We would have also expected there to be a difference between the salivary cortisol concentrations of the novel and existing heifers; however, no significant difference was observed.While no studies have analyzed salivary cortisol during regrouping or during regrouping by treatment groups (i.e., novel versus existing) in dairy cattle, a few studies have analyzed salivary cortisol concentrations during regrouping in swine.The results of one study saw higher cortisol concentrations in regrouped animals than in the control animals that were not regrouped [17].That study did not separate the subjects in the regrouped pens into novel and existing animals for analysis [17].An additional study in swine found a difference in salivary cortisol concentrations between the incoming (novel) sows and the resident (existing) sows two hours after regrouping, where the salivary cortisol concentrations greatly increased for the incoming sows [18].Perhaps, since heifers are young animals (the heifers in this study were between five and six months of age), they are not as accustomed to regrouping and reestablishing social hierarchies with social interactions as older animals.This may explain why the cortisol concentrations of both the existing and the novel heifers increased to similar concentrations on the day of regrouping and the day following regrouping as opposed to the baseline.
Regrouping 1 had significantly higher cortisol concentrations than the other three regroupings.In Regroupings 1 and 4, three novel heifers were introduced into a pen of existing heifers and in Regroupings 2 and 3, six novel heifers were introduced into a pen of existing heifers.The largest difference in mean salivary cortisol concentrations across novel and existing heifers was between Regrouping 1 and Regrouping 4. No significant differences were detected for mean salivary cortisol concentrations between Regrouping 2, Regrouping 3, and Regrouping 4. This indicates that moving a larger group of novel heifers together into a pen may also reduce stress during regrouping, as six heifers were moved during Regroupings 2 and 3.It is possible that with the movement of six heifers during these regroupings, there was generally less physiological stress than during regroupings, where only three heifers were moved.It should be noted that this study was not originally designed or intended to test the effect of group size on stress parameters in heifers.The goal was to conduct a field study that did not interfere with the management requirements of the cooperating dairy.Due to the spatial and operational needs of the dairy, three heifers needed to be moved during Regroupings 1 and 4, and six heifers needed to be moved during Regroupings 2 and 3.
However, the salivary cortisol concentration results are supported by the behavioral results of this study.Generally, more aggressive behaviors, chin rests, and displacements were exhibited by the existing heifers towards the novel heifers during the regroupings, with three novel heifers being introduced as opposed to six novel heifers being introduced.This presents an area where future research should be focused, as it appears that moving a larger group of novel animals together improves welfare and reduces physiological and psychological social stress by minimizing aggressive interactions.One study focused on analyzing the effects of moving three cows at a time into a new pen [14].They witnessed less aggressive and agnostic interactions than in a previous study, which only introduced one focal cow to a new pen at a time [7,14].The author of the current study would suggest that a future study be designed to evaluate the frequency of agnostic and aggressive interactions when moving one, three, and six novel focal dairy heifers into pens of existing dairy heifers.
For the majority of the above-listed behaviors, no significant differences occurred between the two regroupings with three heifers and between the two regroupings with six heifers.The increase in displacements from the feed bunk on the day of regrouping is consistent with the findings from another study of regroupings in dairy cows [7].Fewer displacements at the feed bunk occurred during Regroupings 2 and 3 than during Regroupings 1 and 4.
Although unexpected when taking into account the high prevalence of displacements at the feed bunk during Regroupings 1 and 4, no significant differences were observed between regroupings or between the percentage of novel and existing heifers feeding during any given observation period.These findings are relatively consistent with other studies of a similar nature.In the past, a decrease was found in feeding bouts and feeding time one hour after feeding on the day of regrouping, but no overall changes in feeding occurred when not taking specific times of day into account [7].An additional study found no difference in feeding time between novel and existing cows but found that feed intake or the feeding rate decreased among the novel animals, which they suggested may be due to an increase in vigilance behavior [14].The latter study was unable to replicate the decrease in feeding bouts and feeding time following the hour of regrouping in the previous study as their test subjects were fed at inconsistent times throughout the study [14].During the current study, the novel animals were introduced approximately an hour after morning feeding time and no statistical significance was observed for the percentage of animals feeding based on the time of day.
No differences were detected among regroupings for feeding, lying, standing, and drinking behaviors.Throughout all the regroupings, typically, more novel heifers were standing than existing heifers, and more novel heifers were drinking than existing heifers.During any given observation period, more existing heifers were lying than novel heifers.While this was considered statistically significant, the mean and standard errors of the mean for these percentages do not appear to have a significant difference.More animals across both treatment groups and all regroupings spent less time lying on the day of regrouping and more time lying on the day after regrouping, although this was not considered significant.Additionally, the heifers spent more time standing on the day of regrouping and less time standing on the day after regrouping.This is consistent with the findings of other studies analyzing regroupings, which also found lying time to decrease and standing time to increase on the day of regrouping [5,7].
The results of the current study as they pertain to lying and standing on the day of regrouping and the day post-regrouping may be indicative of more activity, movement, and general unrest in the pen on the day of regrouping.Data from an unpublished study at the WSU Knott Dairy Center found that novel heifers spent more time drinking during regrouping than existing heifers, which is consistent with the findings of the current study [19].The same unpublished study found no significant differences between novel and existing heifers for lying and standing behaviors, which differs from the results of the current study.A significant difference between the percentage of novel and existing heifers drinking was observed, and more research must be carried out to analyze the effect regrouping had on this behavior.To our knowledge, no published studies have analyzed heifer drinking behavior during regrouping, and the majority of the studies analyzing regrouping in dairy cattle did not assign the animals to novel and existing treatment groups.Research on heifer behavior in general is also limited.There are several studies that have analyzed the effects of regrouping in cows, but cows are older animals with generally more mature and established social interactions [7,14].In retrospect, allowing for multiple regroupings to increase our sample size may have led to more significant and conclusive behavioral results.

Conclusions
The baseline salivary cortisol concentrations for the novel heifers were significantly lower than the concentrations for both novel and existing heifers on the day of regrouping and the day post-regrouping.We can conclude that regrouping does cause physiological signs of distress in dairy heifers.We can also conclude that there are differences in behavior between the novel animals and the existing animals and that there is a large presence of aggressive and displacement behaviors during regrouping.

Figure 1 .
Figure 1.Salivary cortisol concentrations by day and treatment (mean ± standard error ng/mL) for individual heifers (n = 18).Gray bars represent existing heifers, and black bars represent novel heifers.

Figure 2 .
Figure 2. The mean percentage (mean ± standard error) of heifers performing the specified behavior during observation periods by treatment group (novel represented in blue bars and existing represented in orange bars), the regrouping number is the experimental unit (n = 4).

Table 1 .
Ethogram of behaviors used during analysis of video footage for all-animal instantaneous scan sampling at every 5 min interval.

Table 2 .
Ethogram of behaviors used during analysis of video footage for focal group all occurrence sampling.