Edited by: Edward Narayan, The University of Queensland, Australia
Reviewed by: Sarah Halina Ison, World Animal Protection, United Kingdom; Keelin Katherine Mary O'Driscoll, Teagasc, Ireland
This article was submitted to Animal Behavior and Welfare, a section of the journal Frontiers in Veterinary Science
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Confinement to farrowing crates is known to prevent sows from performing natural behavior, impairing animal welfare and possibly causing chronic stress. Hair cortisol analyses are increasingly used to detect chronic stress in animals. In the present study, hair samples were collected in the neck of sows kept either in farrowing crates (FC,
The subject of animal welfare in intensive pig farming has become increasingly important for the public in recent years (
A widely used method to quantify stress is to measure the cortisol level in body fluids or excreta as a biomarker. Cortisol is the main glucocorticoid in most mammals (
Koren et al. (
Considering all these possible origins, the question remains to what extent measured hair cortisol concentrations (HCC) are influenced by systemic cortisol levels and whether they actually reflect the HPA axis activity. Some studies have shown that HCC increased in times of higher plasma cortisol levels, or when ACTH was applied to the organism experimentally (
Even if some doubts remain, and there is a need for further research, detection of hair cortisol is increasingly considered a useful marker to determine chronic stress in animals. Therefore, it may be suitable for assessing long-term stress caused by different housing systems for farm animals. Hence, the aim of the present study was to explore the applicability of hair cortisol measurement to detect chronic stress in sows kept in two different farrowing systems. Moreover, factors which affected the sows' stress levels in the farrowing systems should be analyzed as well. Since physical damage in pigs can also influence chronic stress levels (
The study was conducted between June 2018 and January 2019 as part of a larger research project at the research farm of the Lower Saxony Chamber of Agriculture in Wehnen, Bad-Zwischenahn, Germany. The animals were kept in accordance with the European Directive 2008/120/EC and the corresponding German national law (Tierschutzgesetz and Tierschutz-Nutztierhaltungsverordnung). The experiments did not include any invasive procedure involving the animals. The study was reviewed and received approval from the Animal Welfare Officer of the University of Veterinary Medicine Hannover, Hannover, Foundation, Germany.
In the study, two different farrowing systems for sows were compared: conventional pens with farrowing crates (FC) and a loose-housing system without farrowing crates (LH). Both systems were installed in adjacent rooms, and in both systems the sows were single-housed. The LH system had six pens per room and the FC system was equipped with eight pens. Both systems were provided by the same manufacturer (Big Dutchman International GmbH, Vechta, Germany). A single LH pen (
Single loose housing pen (LH). CrA, creep area; Sw, swing gate; A, anti-crushing bars. © Big Dutchman International GmbH, Vechta, Germany.
The FC pen (
Pen with farrowing crate (FC). CrA, creep area. © Big Dutchman International GmbH, Vechta, Germany.
Before entering the farrowing systems, pregnant sows were housed in groups of three to five animals. Five days before the expected farrowing date, sows were moved to the farrowing pens and thus, were single housed either in FC or in LH pens. Six sows were housed in each farrowing system per batch. At the beginning of the study, the sows were randomly assigned to the two housing systems and thereafter were always allocated to the same housing system. Before entering their pens, the sows were washed and weighed using digital scales (82-b2, RHEWA WAAGENFABRIK, August Freudewald GmbH & Co. KG, Mettmann, Germany). The weight of the piglets was individually recorded within 24 h after birth (scale: SC-A, T.E.L.L. Steuerungssysteme GmbH & Co. KG, Vreden, Germany) and an ear tag was immediately applied to identify the individual animals. The teeth (canines) of the piglets were shortened at the same time. To obtain litter sizes that were as homogeneous as possible, cross-fostering was carried out within the same housing system between three and 72 h after birth. After 28 days, piglets were weaned, reweighed individually and then transferred to the farm's own rearing unit.
While LH sows were never confined during the entire housing period (free farrowing), FC sows were permanently fixed in the crate. Feeding-management in the two systems was the same: sows received a commercial lactation diet twice a day (07:30 and 16:30). The amount of feed was rationed on the days before farrowing (maximum 5 kg per day) and on the day of parturition (maximum 2 kg per day). After parturition, the feed amount was increased by about 0.5 kg per day to reach an ad libitum feeding level after about 14 days (8–9 kg per day).
The sows left the farrowing pens after a period of 33 days, were weighed for the second time and entered the service center for the following reproduction cycle.
In both farrowing systems, temperature and air humidity were measured every 2 min in the respective rooms. The sensors (DOL 114 and DOL 12, dol-sensors A/S, Aarhus, Denmark and 135pro, Big Dutchman International GmbH, Vechta, Germany) were placed at the animals' body height in a farrowing pen in the middle of the room.
In a total of six batches, data of 69 sows (Landrace x Large White, db.Vicoria, BHZP GmbH, Dahlenburg, Germany) from first to seventh parity (LH: 3.8 ± 1.8, FC: 4.3 ± 1.8) were obtained. In five batches, data on all recorded parameters were collected for all sows (
Cameras (Everfocus ez.HD, Everfocus Electronics Corp., New Taipei City, Taiwan) were installed above each pen to record the animals' behavior. They were arranged at the cable duct, directly above the middle of the pen, to observe the entire area from a top view. The cameras were connected to a digital video recorder (Everfocus ECOR FHD 16 × 1, Everfocus Electronics Corp., New Taipei City, Taiwan), which recorded continuously on hard drives over the entire experimental period. The behavior of 60 sows in five consecutive batches was analyzed regarding the occurrence of stereotypies. The associated ethogram is shown in
Overview of the sows' stereotypies analyzed in the present study.
Head waving | Sow moves its head up and down |
Bar biting | Sow bites into the bars of the pen |
False chewing | Sow chews independently of feed intake, formation of foam at the mouth |
The total results of the behavioral analyses are planned to be published in a following paper.
In five batches, the sows (
Scoring scheme for skin injuries (
0 | No injuries |
1 | A small number (<5) of superficial scratches |
2 | A mean number (5–10) of superficial or a small number of deep scratches (<5) |
3 | A high number (>10) of superficial or a mean up to a high number of deep scratches (>5) |
Using electronical clippers, a bilateral symmetric area of 20 × 30 cm was shaved in the transition between neck and shoulder blades (
Localization of the shaving area in the transition between neck and shoulder blades with subsections: (
In order to take samples from the newly grown hair, representing the period being in the farrowing systems, the surface of interest was shaved twice. Considering a depth of the hair shaft in the skin of 3–4 mm (
Shaving regime for hair sampling:
Hair sample washing and the extraction of the hair cortisol based on Davenport's methods (
The hair cortisol concentration was finally analyzed using a commercially available immunoassay with chemiluminescence detection (CLIA, IBL International GmbH, Hamburg, Germany). Due to low concentrations of cortisol in the hair, the protocol “RE62019” for ultra-sensitive detection was followed. This included the preparation of an additional standard by diluting standard B 1:3 with standard A. One hundred microliters of each standard, control and sample were pipetted into the respective wells of the microtiter plate. The enzyme conjugate was diluted at 75% and 50 microliters of this was added into each well. The plate was then incubated for 3 h at room temperature on an orbital shaker (400–600 rpm). After washing the plate four times with 250 microliters of diluted wash buffer, 50 microliters of prepared substrate solution mixture were pipetted into each well. The measurement of the relative luminescence units was performed after 10 min.
The assay precision in this study, indicated by the intra- (variation within plates) and interassay (variation between plates) coefficient of variance, was below 10 and 12%, respectively.
Statistical analyses were performed using the R statistics software (
Data were tested for normal distribution by using histograms.
A linear mixed effects model was used for hair length analysis in different body regions by implementing the R package
In the following, hair lengths are stated as means ± standard deviations (SD).
First, the measured cortisol values were logarithmized to approach a normal distribution. A linear model was used with hair cortisol concentration as the dependent variable and the following fixed effects: farrowing system, sows' parity, number of piglets born alive, number of weaned piglets, total piglet loss, mean temperature in the experimental period, sows' weight loss in the farrowing system, occurrence of stereotypies, BLS and ULS on the third day of investigation (
Prior to the analysis, the potential effects were prioritized according to their potential influence on HCC. Thereafter, they were included stepwise in the model before the final model was developed.
The effects could potentially affect the stress level of sows and were investigated for the following reasons:
The housing system to unveil potential environmental effects on stress level
Parity, to show the effects of age and life experience on stress level
The number of born, weaned, piglet loss to uncover the effects of litter sizes on the sows' stress level
Temperature, to illuminate heat or cold stress
Weight loss of sows, to discover possible links between the physical conditions and the sows' stress level
Stereotypies, to show any link between behavior and stress
Lesions scoring results, to determine if body/udder lesions were stress-related.
The effects of the model were examined for significance using
The stated mean values and standard deviations were calculated from the measured, not the modeled (logarithmized), HCC values.
A logarithmic mixed-effects regression model was used in conjunction with the R package
The effects could potentially affect the lesion score of sows and were chosen for the following reasons:
Farrowing system, in order to reveal environmental effects on the lesion score
Time of investigation, in order to reveal the effect of the housing period on the lesion score
Number of weaned piglets, in order to reveal the effects of litter sizes on lesions
Sows' parity, in order to reveal the effects of age and life experience on the lesion score
Sows' body weight at the day of entering the systems, in order to reveal the effects of weight/force on the lesion score
Sows' weight loss in the farrowing system to show any link between body weight or nutrition and the lesion score.
Based on the final model,
Highly significant differences were found in the hair growth rate between different regions of the shaving area. While the hair in both lateral parts of the shaving area grew almost identically in length within 30 days (left side: 7.48 ± 3.52 mm, right side: 7.44 ± 3.24 mm,
Overall, HCC were measured from a minimum of 0.49 pg/mg to a maximum of 8.92 pg/mg with a mean of 1.99 ± 1.23 pg/mg for all analyzed samples. Mean HCC did not differ significantly between the farrowing systems (LH: 1.85 + 0.82 pg/mg, FC: 2.13 + 1.53 pg/mg,
Descriptive results of hair cortisol concentrations (pg/mg) in the two farrowing systems (LH, loose housing; FC, farrowing crate).
LH | 30 | 1.735 | 1.853 | 0.817 | 0.490 | 3.730 |
FC | 31 | 1.630 | 2.125 | 1.526 | 0.940 | 8.920 |
Mean hair cortisol concentrations and standard deviations depending on the sows' parity.
In both housing systems, the mean individual BLS declined from the beginning to the end of the housing period (
Mean body lesion score (BLS) and standard deviations of the mean in the two systems (pens with farrowing crate = FC, loose housing pens = LH, each
In order to use HCC as a retrospective calendar of stress, knowledge about the speed of hair growth is important. Only then can the period reflected by HCC be narrowed down, as this is how the sampling interval is determined (
In human studies, a mean hair growth rate of scalp hair of 1 cm/month is generally accepted. Nonetheless, scalp hair growth varies according to the region, with the posterior vertex region accepted to be the one with the most uniform growth rates, resulting in less intra-individual variation of HCC. Thus, samples are typically obtained from this region (
It is assumed that chronic stress can cause both increased and decreased reaction of the HPA axis at different time points during a stressful situation: an HPA activation and elevated cortisol levels at the beginning can be followed by a counter-regulatory response over time, with rebounded glucocorticoid levels below normal ones by feedback mechanisms (
In the current paper, we explore if methods adapted from previous successful hair cortisol extraction studies (
To the best of our knowledge, this is the first study using hair cortisol analyses for evaluating different farrowing systems for sows.
In addition, it was investigated whether the housing systems had physical impacts on the sows using a lesion score, and whether these were related to measured HCC.
Further data obtained in this study concerning animal behavior and performance will be presented in another paper.
However, the results of the present study do not reveal any effects of either the housing system (farrowing crate or loose housing) or of all other investigated parameters on hair cortisol levels of sows. Thus, the results of the present study seem to be partly contradictory to those obtained by previous studies. Trevisan et al. (
The sampling procedure could also be a cause of conflicting results in different studies. The hair sampling method in the current study was an attempt to improve over previous methods by considering thoroughly the hair growth (which was included for analysis). Due to a standardized procedure in the present study, it is assumed that meaningful results were achieved. In the present study, care was taken to ensure that only hair that was newly formed during the study period was used for hair cortisol measurement. Thus, it was necessary to consider the time delay between incorporating cortisol in the hair and the appearance of this hair section on the skin's surface (
Bacci et al. (
Most earlier studies dealing with stress levels of crated and loose-housed sows in farrowing systems used cortisol measurement in saliva or blood and are, therefore, not directly comparable to this study (
Hair cortisol is supposed to be an indicator of the previous weeks' or months' adrenocortical activity (
Furthermore, the present study did not reveal any other influencing factors on hair cortisol levels in sows, such as the number of suckling or weaned piglets. This could be explained by the habituation to the constant stressors during the housing period. Suckling was proven to be a stress factor for sows as higher cortisol plasma levels were found the day before weaning compared to the day after weaning. Shortly after weaning, an increase in plasma cortisol was also measured (
Finally, it may also be assumed that measuring hair cortisol in pigs is not the appropriate method to determine stress levels. Heimbürge et al. (
As lesion scores are an important animal welfare indicator, they were further investigated.
Although the farrowing crate itself may cause injuries to the sows (
The results of the present study revealed that the use of hair cortisol measurements in sows around farrowing seems to be limited. This may be due to the constant stressful conditions in farrowing systems, such as suckling bouts or single housing, to which the sows' HPA axis may adapt over time, resulting in decreased cortisol levels. However, it is also possible that measuring hair cortisol is not the appropriate method for determining stress in pigs. Thus, a meaningful use of HCC in sows for comparing the effect of different farrowing systems on animal welfare remains questionable. Further research on the time course of cortisol levels mapped in the HCC seems necessary to validate the measured values. Also, the hair growth rate should be considered in further studies when measuring HCC. Only if the collected hair grows more or less homogeneously can the time period of hair analysis be defined. Regional differences in hair growth rate within the same shaving area should, therefore, be considered in future studies.
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.
The animal study was reviewed and approved by Animal Welfare Officer of the University of Veterinary Medicine Hannover, Hannover, Foundation, Germany. Written informed consent was obtained from the owners for the participation of their animals in this study.
D-HW and MF designed the experiments. D-HW collected the data. SB and SH analyzed the data and performed the statistical analysis. D-HW, MF, NK, SB, and SH wrote and revised the paper. All authors contributed to the article and approved the submitted version.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The authors wish to thank the Lower Saxony Chamber of Agriculture for being given the opportunity to carry out this study on the research farm in Wehnen, Bad-Zwischenahn, Germany. They would also like to thank Big Dutchman International GmbH for providing the farrowing systems. Special thanks also go to the technical staff of the research farm for its assistance in this project. Furthermore, sincere thanks go to TU Dresden and especially Prof. Kirschbaum and his laboratory team for preparing the hair samples and performing the hair cortisol measurements.
The Supplementary Material for this article can be found online at: