Weaning can be defined and understood in many ways. For example, it can be seen as the change of nutritional source from the mother’s milk to entirely using other species-specific food sources [1, 2]. Weaning can also be understood as every change the offspring undergoes during the time of parental deprivation resulting in the development into an independent adult [1]. In today’s most common way of breeding and raising horses (Equus caballus), weaning takes place at a precise date or age. Although it concerns the foal’s adaptation to the new situation, it should be seen as a process rather than a certain point in time. The weaning itself is in most cases just one of the stressors facing the young horse. In addition to separation from the mare, there is a change of diet, integration into a new social group, change of location, or a change of management procedures.

Although existing for other animals, there are no binding legal regulations for appropriate husbandry of horses in Germany. Only the “guidelines for the assessment of horse husbandry with the aspect of animal welfare” [3] published in 2009 by the Federal Ministry of Food and Agriculture portray a preferable way of horse keeping. The guidelines cover all aspects of horse husbandry, such as social needs, free movement and exercise, resting behavior, dietary recommendations, and management. Concerning foals and young horses, it is stated as mandatory to keep them in groups with same-aged horses. It is also suggested to keep an older horse within the group for educational reasons and stress reduction, which has been confirmed by the work of Erber et al. [4]. Furthermore, foals and young horses should have access to a pasture or free run as often as possible and be familiarized towards humans and handling. Specific weaning methods are not addressed in the guidelines, but weaning a foal into single housing is not consistent with the guidelines summarized above.

There are different approaches for weaning foals, one of which includes the abrupt and total separation of foal and mare [5–7]. The foal then goes into single stall housing, paired stall housing with a same-aged companion, or into a group of same-aged foals. Other strategies seek to reduce the stress for mare and foal through slow weaning. One approach is to gradually take the mares away from a herd until only the foals are left [4]. A different approach is to adapt the horses to being separated by isolating mares and foals for short periods at first or preventing direct contact, such as separation by a fence [8, 9]. This method eventually transitions to complete separation. The welfare of horses is often measured in comparison with the possibility of performing the natural behaviors of a free-ranging horse. Heleski et al. [7] summarized the time budget for behavior patterns of free-ranging horses as 30–70% spent with feeding, 15–50% standing, 2–10% lying, and 4–10% moving about. An approximation to this natural behavior distribution during the slow weaning processes and the weaning into groups suggests these methods to be beneficial. Nevertheless, being a member of a social group includes hierarchic encounters, which can induce stress and raise the risk of injuries [5, 7].

Most authors who compared single against paired or group weaning share the opinion that foals benefit from companion animals throughout the weaning period [7, 10–12]. However, Hoffman et al. [5] described a better behavioral response to weaning in singly housed than in paired housed foals. Their finding was based on higher cortisol levels after adrenocorticotrophic hormone (ACTH) challenge and higher aggression in the paired stalled foals. On the other hand, social deprivation can have a negative long-term effect on the development, especially for young horses [13]. Sex-specific differences might also exist. Górecka-Bruzda et al. [14] found a significantly higher increase in fecal glucocorticoid levels in fillies than in colts. However, in most studies, no difference between colts and fillies is detected [5, 6, 8, 12]. Foals with access to a paddock and companions display a wider range of behaviors such as grazing, locomotor play, and contact to herd members [7]. This conduct is stated as more natural because it is closer to the time budget of a free-ranging horse.

The slow weaning process is described as the more natural way because under natural conditions the offspring decreases the contact to its mother gradually [1, 2]. The natural weaning is also encouraged by the mare in preparation for the new foal [4, 10, 15]. In terms of behavioral changes, foals seem to adapt better to gradual weaning than to abrupt weaning [6, 8, 9]. The immediate response to weaning was observed by McCall et al. [8], comparing total and partial separation of foal and mare as well as different feeding plans. In their study, the foals showed aggression against companion foals, non-nutritional sucking, and no play behavior as signs of stress, regardless of treatment group. The higher stress level was evidenced by higher vocalization rates and no lying sideways shown by the abruptly weaned foals within the first five hours of weaning. A study by Holland et al. [6] did not find differences concerning the serum cortisol level after an ACTH challenge between abruptly and gradually weaned foals. Because the observation was made up until 48 hours after weaning, the results are not useful to predict the long-term effect of the gradual weaning process. The foals in their trial did not undergo a change of location. It is possible that the change of location may be an even greater stressor than the separation from the mare. Transportation causes stress in horses, as well as the adaptation to a new environment [16–18]. For example, Dubcová et al. [15] studied the long-term effect of relocating foals to a rearing farm sometime after weaning had occurred. They found that concerning the long-term effects on the foals, separating both events in time has a more negative effect than simultaneously segregating and relocating the weanlings.

Using glucocorticoids, especially cortisol, as a measure of stress is a well-studied and often used method when it comes to assessing animal welfare [19]. Cortisol is a steroid hormone produced in the zona fasciculata of the adrenal gland. Its main purpose is to preserve the existential body functions during times of distress by raising the blood sugar levels. Therefore, it interferes with the effects of insulin.

When an organism experiences stress (chemical, physical, or emotional) it causes the activation of the hypothalamic-pituitary axis. The stimulated hypothalamus secretes the corticotropin-releasing hormone, which then stimulates the pituitary gland to secrete ACTH. The ACTH gets resorbed in the cortex of the adrenal gland where cortisol is produced through an intracellular pathway. The metabolism occurs mainly in the liver and kidneys, and it involves different pathways to convert the steroid hormone into a water-soluble form [20]. Because the breakdown product is water soluble, cortisol metabolites can be found in many body fluids, such as plasma, feces, saliva, or urine [21]. Most of the metabolites are excreted through feces and urine [20]. So, if the organism experiences stress, it undergoes an increase in cortisol levels and therefore an increase in glucocorticoid metabolites, which then can be analyzed.

Cortisol may not be the correct indicator to measure stress in horses performing in sports because the activation of the pituitary gland has an individual dependence on the exercise level of a horse [22]. Likewise, no relationship was found between stereotypic behavior in horses and a stimulation of glucocorticoid secretion [23]. However, other stressors such as pain cause a significant increase in cortisol production and excretion in horses [24].

Cortisol has an impact on multiple organ systems and plays an important role in how the body responds to stress. The effects of high cortisol levels in foals on bone growth, muscle and fat tissue, the inflammatory and immune system, and the brain [20] must be considered for animal welfare. As an antagonist of vitamin D, cortisol lowers the plasma calcium levels by decreasing the resorption in the duodenum as well as the reabsorption of calcium and phosphate in the kidneys. Cortisol also causes an increase in bone resorption and a reduction in the synthesis of collagen type I. These effects can lead to disruption of the length growth of the bones, which is of concern in a growing horse. To promote the gluconeogenesis in the liver, the body provides substrates such as fatty acids, amino acids, or ketone bodies through the breakdown of muscle and fat tissue. Therefore, long-term stress can lead to weight loss and disturbance in growth. Cortisol weakens the immune system on the cellular as well as on the humoral level. Cortisol also affects the central nervous system by increasing the appetite, decreasing the rapid eye movement sleep phase, and by modulating the excitability, mood, and behavior of the animal [20]. If an organism experiences these effects chronically, the negative effect can be severe. Therefore, it is in the best interest for the horse and the caretaker to decrease the possibilities of long-term stress resulting from weaning.

The methods originally used to measure the stress hormone levels include directly determining the cortisol level in the blood of the animal or undertaking an ACTH challenge with the horses [5, 6, 10, 12, 25]. The sampling of blood is an invasive procedure and therefore constitutes as stressor itself. For this reason, professionals have sought noninvasive methods to detect cortisol levels in many species, especially wildlife. The two methods most validated for horses are the detection of cortisol metabolites in saliva [26] and in the feces [27]. Both methods have been used successfully to detect the stress level of foals during weaning [4, 7, 9, 14, 15, 28]. Because detecting the cortisol metabolites in feces involves no force on the horse through sampling, we chose this method over the saliva glucocorticoid detection. We applied the 11-oxoetiocholanolone enzyme immunoassay, which was first established and validated by Palme and Möstl in 1997 [29] for representation of blood cortisol levels in the feces of ruminants. The use of this method was validated for horses in 1999 [27] and is an often used procedure [30].

This study has the aim to demonstrate the midterm effects (three weeks) of weaning on foals’ welfare within rearing conditions that meet the standards of the “guidelines for the assessment of horse husbandry with the aspect of animal welfare” [3]. We assumed that foals finish the acclimation phase in the new situation within three weeks after weaning. As aspects of the study, the changes in behavior and fecal levels of cortisol metabolites were evaluated.

The study took place at the facilities of the “Haupt- und Landgestüt Marbach”, the state stud farm of Baden-Wuerttemberg in Germany. Until weaning, all foals were raised in a group with other dams and their foals, and the housing method was an open stabling with daily pasture time depending on weather conditions. The feeding consisted of grass on the pasture, hay ad libitum renewed three times a day, and for the dams concentrated feed twice a day in the stable. The foals also had access to the concentrated feed of their mothers. The foals were divided into three weaning blocks based on the stud farm’s management process, which took the age and developmental stage of the foals into account. The first in September, the second in October, and the last in November. On the day of weaning, a veterinarian sedated the foals before they were transported to a breeding station 18 kilometers from the stud farm. The breeding station continued the known daily feeding schedule with concentrated feed twice a day, hay three times a day, and grass on the pasture depending on weather conditions.

In this study, 10 foals were observed during the weaning process to determine behavioral patterns and stress, the latter measured through cortisol metabolites in fecal samples. The group consisted of nine Arabian foals including six colts and three fillies and one warmblood filly (Table 1). Each foal was observed for eight hours per day on five days: on the day before weaning in the known environment, the day after weaning in the new environment, four and eight days after weaning, and the last observation day was within the third week after weaning. The observations in the third week after weaning took place at the 18th or the 20th day after weaning. The day of weaning itself could not be included because the foals were transported at a variable time of the day and the sedation would have falsified the results. For simplification, a timeline was created in which the day before weaning is defined as day 1 and the following days are consecutively numbered (Fig 1). Resulting from that, the first measurement (M1) equals day 1, measurement 2 (M2) equals day 3, M3 equals day 6, M4 equals day 10, and the last measurements (M5) are combined to day 20 for the statistics.

The foals of block 1 (foals F1–F4) were weaned on September 2, 2019, into a mixed gender group with known foals from growing up, unknown foals that were weaned the same day, and a lead mare. Before the second weaning date, the breeding farm divided the first group by gender, leaving the lead mare with the fillies. Block 2 foals (F5–F7) were weaned on October 14, 2019, into gender-divided groups. Block 3 foals (F8–F10) were weaned on November 24, 2019; the two gender groups were placed in the same stable but separated by a wooden partition that allowed some contact between fillies and colts.

The behavioral observation was done in an eight-hour period between 7:00 a.m. and 5:00 p.m., including a break of one hour maximum per day. The observer documented the exact behavior shown by the foal every five minutes within the eight hours. If the behavior changed in the moment of observation, the new displayed behavior was noted. The observation interval of five minutes was chosen based on existing literature [7, 31]. To define and evaluate the observed behavior, we used the time budget recording sheet found in Heleski et al. [7] and divided play behavior into object play, play sexual behavior, locomotor play, and play fighting as done by McDonnell and Poulin [32]. Each observed behavior and its description are listed in Table 2.

To scale the stress experienced by the foal throughout the observation period, we measured glucocorticoid metabolites. The analysis was performed by staff members of the Department of Biomedical Sciences at the University of Veterinary Medicine Vienna (Unit of Physiology, Pathophysiology and Experimental Endocrinology). The 11-oxoetiocholanolone enzyme immunoassay, which detects a group of glucocorticoid metabolites (11,17-dioxoandrostanes), was used to assess the change of the foals’ plasma cortisol levels through fecal samples. This measurement method was validated for fecal samples of horses by Möstl et al. [27].

Fecal samples were collected at some point during the eight-hour observation period, but not earlier than 10:30 a.m. owing to the circadian rhythm of natural cortisol levels [4, 20]. Because 11,17-dioxoandrostanes have a delay of 24 hours to be excreted into the feces of the horse [27], the results allow a conclusion to the blood plasma cortisol levels of the days 0, 2, 5, 9, 19, and 21 (Fig 1, green numbers). On day 3 (i.e., the day after weaning), it was made sure to take the feces samples in 24-hour delay from the point the sedation had worn off, in order to be able to see the accurate levels of the change within the cortisol at the time when the foal experienced the separation from its mother. We have refrained from combining day 19 and day 21 as in the behavioral observations because this two-day window seems to be showing an interesting point of time. For the sample, the whole pile of feces was collected, transferred into a plastic bag, and kneaded until homogenic; afterwards, a portion of about 10 g feces was taken. The samples were stored on dry ice or in a freezer with a temperature of at least −20°C until analysis [30].

The displayed behavior of observation day 1 (i.e., the day before weaning) is regarded as the normal behavior for foals this age because the foals were in the presence of the mother as well as in a known environment and herd. The changes in behavior over the timeframe of three weeks after weaning will be compared in relation to the behavior on day 1. In addition, the cortisol metabolite values of day 0 are regarded as relation to the normal cortisol level for foals before weaning.

Weaning marks a stressful period in a horse’s life that results in increasing fecal cortisol metabolite levels and behavioral changes. The cortisol metabolite values demonstrated a significant curve during three weeks after weaning, despite the broad individual range. Overall, the foals in this study spent most of the daytime in movement while with their mare. After weaning, their posture distribution during the following three weeks took a shift towards mainly standing. The analysis of resting behavior showed that after separation from the mare, the foals increased the frequency of resting while standing significantly. Our results furthermore suggest a benefit of companion animals throughout the weaning process. Even though hierarchic encounters naturally cause stress in the foal, companions provide the opportunity to perform care-soliciting behaviors such as mutual grooming. Moreover, horses evolved as social herd animals, and adult horses are often housed in groups at least for some time during the day. Thus, one may assume that foals learning to integrate into social structures will have better adaptation skills as adult horses. Our results could not determine statistical support for the illustrated “signs of stress” and “coping behaviors” because of the small sample size and large individual variations. However, the findings pave the way for future studies that should, for example, include more foals and narrow the focus to fewer behaviors. Furthermore, it seems worthwhile to observe the foals’ changes in lying and sleeping behaviors after weaning to reassess the hypothesis that weaned foals sleep more during the day than during the night. It is not possible to accomplish weaning without producing stress in the foal. The goal must be to determine the process that provides the best long-term welfare for the foal.