Welfare of Sows Housed in Stalls during Gestation

Livestock Update, April 2003

Mark J. Estienne, Swine Research Physiologist and Allen F. Harper, Extension Animal Scientist-Swine, Tidewater AREC

Introduction
Individual housing of pregnant females is a common practice in the swine industry. Barnett et al. (2001) estimated that at least 60 to 70% of U.S. sows are housed in stalls throughout gestation.

The use of gestation stalls is currently one of the most contentious welfare issues facing pork producers. Typical gestation stalls are approximately 2' wide and 7' long and physically limit sows to standing, sitting, and lying. This severe restriction of freedom of movement has been robustly criticized by many animal right and welfare activists.

Public perceptions and misconceptions of welfare issues have the potential to dramatically impact swine production if governments, the swine industry, or consumers react to these issues by banning housing systems, changing husbandry practices, or boycotting pork. For example, last November, an amendment to the Florida Constitution was passed, prohibiting pregnant sows in that State from being confined in gestation stalls. In reality, Florida is a small state in terms of pork production and the ban will impact only two commercial farms. Never-the-less, the amendment could set a precedent for pork producers if other states with larger industries follow suit.

In determining whether or not a sow's welfare is compromised, individuals may act emotively and perhaps without factual biological information. The objective of this paper is to provide the reader with a brief review of some of the more pertinent scientific literature focusing on sows housed in gestation stalls. Currently, the most obvious and indeed, most common, alternative to gestation stalls is housing various numbers of pregnant sows together in a common pen that allows freedom of movement and physical interaction with one another; Sows are fed on the floor or in troughs. Thus, in general, various indicators of welfare will be compared for stall-housed and traditional group-housed sows.

Why Use Gestation Stalls?
As mentioned above, gestation stalls are routinely used on U.S. swine farms. It would, therefore, seem appropriate to briefly discuss why this method of housing is popular. From the perspective of the swine producer, housing gestating swine in stalls offers a number of advantages compared with traditional group housing systems. Building and labor costs are reduced. Care-taking is simpler and because sows are not mixed there are fewer injuries due to fighting. Signs of morbidity such as feed refusal or discharge from the vulva are more easily detected.

Perhaps the most compelling advantage of gestation stalls compared with traditional group housing systems is that sows can all feed at the same time and each sow is certain to have the full ration of feed available to her. In modern swine production systems, appropriate feeding of the gestating sow and gilt is paramount to enhancing biological performance. In traditional group housing systems, however, "boss" sows will inevitably over-consume pen feed resources at the expense of more timid, submissive sows. This point was illustrated by the results of an experiment we recently conducted at the Tidewater Agricultural Research and Extension Center in Suffolk.

The demonstration study employed 27 Yorkshire x Landrace pregnant sows housed in standard gestation stalls (2' x 7' floor space). Each sow was given 5.5 pounds of feed and the time each female required to consume the ration was recorded. The rate of feed consumption (lbs/min) was calculated for each sow. On average, sows consumed feed at a rate of .27 lbs/min with a range of .14 to .44 lbs/min. As a group, the 27 sows consumed 7.18 lbs feed/min.

Given the group's feed consumption rate, it can be determined that less than 21 minutes would be required for the total amount of feed allowed (27 sows x 5.5 lbs/sow = 148.5 lbs) to be consumed (148.5 lbs ÷ 7.18 lbs/min = 20.7 min). Assume now that the sows were housed together in a single pen and given access to 148.5 lbs of feed. The theoretical amounts of feed that would be consumed by each sow in 20.7 min can be calculated. Given the variation in individual sows' feed consumption rates, average feed intake would be 5.5 lbs/sow, but the range would be 3.0 to 9.3 lbs/sow. Assuming the 5.5 lbs of feed is meeting National Research Council (NRC) (1998) recommendations for the various nutrients, in this example, 14 sows (52%) would be over-fed, and 13 sows (48%) would be under-fed. No sows would receive the desired 5.5 lbs of feed.

Because the issue of social hierarchy and other factors that impact feed consumption rates have been ignored in this example, caution must be exercised when extrapolating feed consumption data gathered for individually housed sows to that which would be generated if the animals were group-penned. The point, however, is clear. In traditional group housing systems some sows will be underfed and some will be overfed.

There are a number of periods during gestation when over or underfeeding can compromise reproduction. For example, gilts consuming excess concentrations of energy early in gestation (prior to Day 30 post-mating) have decreased embryo survival rates (Dyck and Strain, 1983). Moreover, it is well established that the greater the energy intake during pregnancy, the lower the feed intake during lactation (Estienne et al., 2000, 2003). Cole (1989) reported that when feed intake during pregnancy exceeds 4.4 lbs/day, feed intake during the subsequent lactation is reduced. Xue et al. (1997) demonstrated that reduced voluntary feed intake and excessive loss of weight and subcutaneous fat during lactation resulted in a delay in the post-weaning return-to-estrus.

Estienne et al. (2000) and Harper and Estienne (1999) determined the last rib backfat thickness for 24 primiparous, Yorkshire x Landrace sows using a Lean-meater® ultrasound unit (Renco Corporation, Minneapolis, MN). During gestation, sows were group-housed and fed, which contributed to differences in body condition at farrowing. Based on backfat thickness, sows were classified as FAT (over 25 mm), MEDIUM (20 to 25 mm) or THIN (less than 20 mm) immediately after farrowing. Loss of backfat from farrowing to weaning at Day 22 of lactation was greatest for the FAT sows and FAT sows consumed the least amount of feed during lactation. The percentage of sows in estrus by Day 10 after weaning was lowest for the FAT group. Although this experiment involved a limited number of sows, it serves to illustrate an important point. Sows that are over-conditioned at farrowing are more likely to experience poor feed consumption during lactation. Under this catabolic state the sow loses more stored body fat and is likely to have poorer rebreeding performance.

How Do We Define Sow Welfare?
Before delving into a review of research conducted to ascertain if gestation stalls compromise sow well-being, a working definition of "sow welfare" must first be presented. Barnett et al. (2001) suggested, and the authors of this review concur, that the "homeostasis" approach offers the best assessment of the welfare of animals. This approach involves comparing housing or husbandry systems, and risks to welfare are assessed on the basis of changes in behavior and physiology and corresponding decreases in fitness. The ability to grow, survive, and reproduce are measurements of fitness. There are a number of reports in the scientific literature which validate the homeostasis approach. For example, studies have demonstrated that pigs made fearful due to rough handling have sustained increases in blood cortisol concentrations. The consequences of this chronic stress response include depressed growth rates and poor reproductive performance (Barnett et al., 2001).

Welfare of Sows Housed in Gestation Stalls or Penned in Groups
Behavior. Stereotypies are behaviors that do not vary in presentation, are regularly repeated, and do not have an obvious function. Although there is controversy regarding the cause and function of stereotypies, it has been suggested that the behaviors develop as a consequence of boredom, restraint, or hunger (Lawrence and Terlouw, 1993), and may indicate a past problem for the animal in coping with its conditions (Barnett et al., 2001).

Vieuille-Thomas et al. (1995) reported the results of an experiment conducted under commercial conditions during which stereotypies were compared for pregnant Large White x Landrace sows individually housed in stalls (15 ft2 floor space/sow) or group-housed (5 to 9 sows/pen; 34 ft2 floor space/sow). The proportion of sows developing stereotypies was lower in group-housed, compared with stall-housed sows (66.2 vs. 92.6%). Stereotypies described for sows housed in gestation stalls included biting or chewing on the bars of the stall and sham chewing (i.e., chewing without anything in the mouth). The most frequent stereotypies observed for group-housed sows were licking of concrete walls and sham chewing.

An important conclusion can be drawn from these data: stereotypies develop in sows that are housed either individually or in groups. However, the proportion of sows exhibiting stereotypic behavior is greater for stall-housed females. Stereotypies that result in physical damage, such as the development of lesions in stall-housed sows that constantly rub stall fittings; have obvious and immediate animal welfare implications.

Blood chemistry. When an animal is "stressed", compensatory and protective mechanisms are invoked to restore stability of the internal environment. For example, in stressed animals, increased levels of adrenocorticotropic hormone (ACTH) are released into the circulation from the anterior pituitary gland, a small organ located just below the brain. The ACTH stimulates secretion of cortisol from the adrenal glands. Elevated blood concentrations of cortisol favor hyperglycemia (i.e., elevated blood glucose) by stimulating the production of glucose and by antagonizing glucose utilization in many tissues. In essence, glucose is "spared" for use in critical tissues during the period of stress. During the stress response, functions important to optimum swine production such as reproduction are compromised until the stress is alleviated. Indeed, Barb et al. (1982) reported that injections of hydrocortisone, a compound with similar actions to cortisol, blocked ovulation in gilts.

Therefore, blood cortisol concentrations have often been used as an indicator of stress in farm animals. Barnett et al. (1989) reported that sows housed in stalls had a moderate, but statistically significant increase in cortisol concentrations compared with group-housed sows. However, blood glucose concentrations were not elevated. Moreover, Broom et al. (1995) reported similar concentrations of cortisol for stall- and group-housed sows. Finally, von Borell et al. (1992), reported no difference in the cortisol response to an injection of ACTH for Yorkshire gilts housed individually in gestation stalls (2.5' x 6.9' floor space) or group-housed (6 gilts/pen; 21.3 ft2 floor space/gilt) in a pen serviced by electronic feeders.

Immune Function. Immunity refers to the physiological mechanisms that allow animals to recognize materials as foreign or abnormal and to neutralize or eliminate them. In recent years, immunological status has been used more frequently as a method of assessing the impacts of housing and husbandry techniques on swine welfare. In general, research has shown that the immune systems in sows housed in gestation stalls or group penned are not differentially activated.

For example, Sorrells et al. (2001) reported similar hematocrit levels (percentage of blood that is red blood cells) and blood levels of lymphocytes (white blood cells that produce antibodies) and granulocytes (white blood cells that engulf foreign particles) in Landrace x Yorkshire gilts that were individually stalled (2' x 7'3" floor space/stall) or group-housed (4 gilts/pen; 25.8 ft2 floor space per gilt; four separate, free-access feeding stalls). In that study there were also no treatment differences for _1-acid glycoprotein or fibrinogen, which are indicators of inflammation. However, there was a trend for higher blood concentrations of haptoglobin, likewise an indicator of inflammation, in gilts housed in stalls compared to gilts housed in groups. The authors of that study hypothesized that movement restriction in stalls may have caused some inflammation in joints, shoulders, and legs that caused haptoglobin levels to increase.

Injuries. Perhaps due to a lack of exercise, muscle mass and bone strength are reduced in sows housed in stalls over successive parities (Barnett et al., 2001). However, group-housed sows display a number of vices such as vulva biting (van Putten and van de Burgwal, 1990). Indeed, an advantage often cited by advocates of gestation stalls is that this method of housing prevents fighting between sows and potential injuries. Research by Harris et al. (2001) supports this notion. In their experiment, Yorkshire x Landrace gilts were allocated to either an individual stall (2' x 7'3" floor space/stall) or a group pen (4 gilts/pen; 25.8 ft2 floor space per gilt; four separate, free-access feeding stalls). During gestation six regions of the head and body and five areas of the feet and legs were inspected for injuries. As gilts walked to the farrowing house, lameness was scored using a 6-point gait system (0 = even strides and normal gaits; 5 = unable to stand or move unaided).

Throughout pregnancy, group-housed females had more scratches, cuts and wounds on their head, face and body than did sows housed in gestation stalls. At Day 91 of gestation, the feet and legs of gilts housed in groups were also in poorer condition compared with those of stall-housed gilts. Although some of the lesions were undoubtedly a result of fighting, injuries may also have been caused by individuals being stepped on or by contact with sharp pen fittings. The majority of gilts (63%) walked normally and although mean lameness were numerically higher for group-housed compared to stall-housed sows (.64 vs. .29), this difference was not statistically significant.

Reproductive Performance. Results from experiments that have compared reproductive performance of sows housed during gestation in stalls or group pens are equivocal. Meticulously controlled studies employing large numbers of sows are needed before definitive conclusions can be drawn. For example, Schmidt et al. (1985) reported 15 % lower pregnancy rates at Day 35 after mating for Duroc x Yorkshire sows housed in small gestation stalls (1.6' x 5.6' floor space) compared to sows housed in groups (4 to 5 sows/pen; 16.6 to 20.8 ft2 floor space per sow). In contrast, in an Australian commercial study, 220 sows were housed in stalls for 5 weeks post-mating and housed in groups for the remainder of gestation. A total of 450 sows were housed in groups throughout gestation. The stall-housed sows had more pigs born alive compared with the group-housed counterparts (11.6 vs. 10.8) (Barnett et al., 2001).

Den Hartog et al. (1993) reported the results of a large field trial during which farrowing data was collected for sows that had been housed in stalls (933 litters) or group pens (951 litters) during gestation. The number of live-born pigs per sow per year (23.1 vs. 22.4) and average pig birth weight (3.38 vs. 3.32 lbs) were higher for sows housed in stalls than for group-housed sows. In a four-parity study involving 215 litters, McGlone et al. (1989) reported farrowing rates of 78.5, 80.0, 83.1, and 65.3% for sows gestated on pasture, in group pens, in stalls, or with neck-tethers. Finally, von Borell et al. (1992), reported no differences in pigs born alive or the weaning-to-estrus interval for Yorkshire gilts housed individually in gestation stalls (2.5' x 6.9' floor space) or group-housed (6 gilts/pen; 21.3 ft2 floor space/gilt) in a pen serviced by electronic feeders.

Potential Alternatives to Gestation Stalls
Given the controversy involving stalls for housing pregnant sows, it is not surprising that a great deal of research is currently underway examining alternatives to this method of housing. One alternative that could perhaps be considered a compromise is to house sows in stalls for a specific period during gestation (for example, the first 30 days after mating when developing embryos are especially venerable) rather than the entire length of gestation. The advantages and disadvantages of this type of housing system warrant scientific scrutiny. Other potential alternatives to stalls (and traditional group-pens employing floor or trough feeding as described above) for gestating swine are briefly described below.

Outdoor Paddocks. Low capital costs are incurred for a system in which gestating sows are kept in pastures contained by electric fencing and housed in simple huts or sheds. This method is generally perceived by consumers to be welfare friendly, however, unsuitable soil types and extreme climatic conditions limits its widespread applicability.

Group Pens with Individual Feeding Stalls. With this type of system, sows are group penned in buildings but each animal has access to a separate stall that can be closed, allowing simultaneous and protected feeding. This system is costly because of the space requirement associated with having group pens and separate feeding stalls, and aggressive interactions between sows may still occur.

In the Hurnik-Morris, or "loop-feeding" system (von Borell et al., 1992) sows are housed in small groups (approximately 6 sows) and each group is fed separately in a common feeding pen. In the feeding pen, sows have electronically controlled access to separate feeding compartments. After consuming their rations, sows return to their own pen and the next group can be fed.

Group Pens with Free-Access Feeding Stalls. In this system, stalls serve as both lying and feeding space. Aggressive interactions between sows may still occur, particularly at feeding time.

Group Pens with Short-Stall Feeders. In this system, only head- or shoulder-length partitions are placed between feeding spaces. Aggressive interactions between sows may still occur, particularly at feeding time. Research in Texas is focusing on this type of system used in combination with "trickle feeding". Feed is metered slowly to each feeding place to prevent different levels of consumption due to inequality of eating speeds.

Electronic Sow Feeders. With this system, individual sows are electronically identified by a transponder on a collar, ear tag, or implant, and allocated a specified daily ration in a computer-controlled feeding station. Aggressive interactions between sows still occur, particularly when new sows are added to the pen serviced by the electronic feeder.

Summary and Conclusions
Stalls for individually housing gestating swine will continue to be a hotly debated topic. A review of the scientific literature suggests that overall welfare is not poorer for sows housed in stalls compared with traditional group pens. Stereotypies develop for sows housed in either gestation stalls or group pens; however, the frequency is greater for individually housed animals. Blood levels of cortisol, used as an indication of stress, are either similar or slightly elevated for sows in stalls. In general, research has shown that the immune systems in sows housed in gestation stalls or group penned are not differentially activated and there appear to be fewer injuries in sows individually housed. Reproductive performance is in general better in stall-housed versus group-penned sows, although more research is needed in this area. There are a number of alternatives to gestation stalls being evaluated, and they must not only meet sow requirements for welfare and health but also pork producers requirements for high biological performance, low labor input, ease of management, acceptable capital cost, and acceptable financial return.

Literature Cited
Barb, C.R., R.R. Kraeling, G.B. Rampacek, E.S. Fonda, and T.E. Kiser. 1982. Inhibition of ovulation and LH secretion in the gilt after treatment with ACTH or hydrocortisone. J. Reprod. Fertil. 64:85-92.

Barnett, J.L., P.H. Hemsworth, G.M. Cronin, E.C. Jongman, and G.D. Hutson. 2001. A review of the welfare issues for sows and piglets in relation to housing. Aust. J. Agric. Res. 52:1-28.

Barnett, J.L., P.H. Hemsworth, E.A. Newman, T.H. McCallum, and C.G. Winfield. 1989. The effect of design of tether and stall housing on some behavioural and physiological responses related to the welfare of pregnant pigs. Appl. Anim. Behav. Sci. 24:1-12.

Broom, D.M., M.T. Mendl, and A.J. Zanella. 1995. A comparison of the welfare of sows in different housing conditions. Anim. Sci. 61:369-385.

Borell, von E., J. R. Morris, J. F. Hurnik, B. A. Mallard, and M. M. Buhr. 1992. The performance of gilts in a new group housing system: Endocrinological and immunological functions. J. Anim. Sci. 70:2714-2721.

Cole, D.J.A. 1989. Sow nutrition-The key to profitable pig production- More piglets in less time. Biotechnology in the Feed Industry. Proceedings of Alltech's Fifth Annual Symposium, 1989. p.107-120.

Dyck, G.W., and J.H. Strain. 1983. Post-mating feeding level effects on conception rate and embryonic survival in gilts. Can. J. Anim. Sci. 63:579-585.

Estienne, M.J., A.F. Harper, C.R. Barb, and M.J. Azain. 2000. Concentrations of leptin in serum and milk collected from lactating sows differing in body condition. Domest. Anim. Endocrinol. 19:275-280.

Estienne, M.J., A.F. Harper, D.M. Kozink, and J.W. Knight. 2003. Serum and milk concentrations of leptin in gilts fed a high- or low-energy diet during gestation. Anim. Reprod. Sci. 75: 95-105.

Harper, A.F., and M.J. Estienne. 1999. Nutrition and feeding management strategies on large hog farms. Proceedings Annual Symposium, Efficient Pork Production through Advanced Technology. College of Animal Resources, Kangwon National University, Chunchon, Korea. p. 91-108.

Harris, M.J., A.D. Sorrells, S.D. Eicher, B.T. Richert, and E.A. Pajor. 2001. Effects on production and health of two types of housing for pregnant gilts. Purdue University Swine Day Report. p. 115-119.

Hartog, den L.A., G.B.C. Backus, and H.M. Vermeer. 1993. Evaluation of housing systems for sows. J. Anim. Sci. 71:1339-1344.

Lawrence, A.B., and E.M.C. Terlouw. 1993. A review of behavioral factors involved in the development and continued performance of stereotypic behaviors in pigs. J. Anim. Sci. 71:2815-2825.

McGlone, J.J., E.W. Fugate, J.R. Clark, and R.J. Hurst. 1989. Reproductive performance of sows over four parities in four housing systems. Texas Tech University Agricultural Sciences Technical Report, Lubbock. p. 67.

National Research Council. 1998. Nutrient Requirements of Swine (10th Ed.). National Academy Press, Washington, DC.

Putten, van G., and J.A. van de Burgwal. 1990. Vulva biting in group-housed sows: Preliminary report. Appl. Anim. Behav. Sci. 26:181-186.

Schmidt, W.E., J. S. Stevenson, and D.L. Davis. 1985. Reproductive traits of sows penned individually or in groups until 35 days after breeding. J. Anim. Sci. 60:755-759.

Sorrells, A.D., S.D. Eicher, M.J. Harris, E.A. Pajor, and B.T. Richert. 2001. Evaluating housing stress in gestating gilts using immunological measures. Purdue University Swine Day Report. p. 112-114.

Vieuille-Thomas, C., G. Le Pape, and J.P. Signoret. 1995. Stereotypies in pregnant sows: indications of influence of housing system on the patterns expressed by the animals. Appl. Anim. Behav. Sci. 44:19-27.

Xue, J., Y. Koketsu, G.D. Dial, J. Pettigrew, and A. Sower. 1997. Glucose tolerance, luteinizing hormone release, and reproductive performance of first litter sows fed two levels of energy during gestation. J. Anim. Sci. 75:1845-1852.

Visit Virginia Cooperative Extension