The course of parturition in two sow genotypes and two hut designs under free-range conditions

Accepted Manuscript Title: The course of parturition in two sow genotypes and two hut designs under free-range conditions Authors: S.-L.A. Schild, L. Rangstrup-Christensen, C.K. Thorsen, M.K. Bonde, L.J. Pedersen PII: DOI: Reference:

S0168-1591(18)30484-2 https://doi.org/10.1016/j.applanim.2019.02.005 APPLAN 4767

To appear in:

APPLAN

Received date: Revised date: Accepted date:

30 August 2018 30 January 2019 6 February 2019

Please cite this article as: Schild S-LA, Rangstrup-Christensen L, Thorsen CK, Bonde MK, Pedersen LJ, The course of parturition in two sow genotypes and two hut designs under free-range conditions, Applied Animal Behaviour Science (2019), https://doi.org/10.1016/j.applanim.2019.02.005 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

The course of parturition in two sow genotypes and two hut designs under free-range conditions

S.-L. A. Schild1*, L. Rangstrup-Christensen2, C.K. Thorsen1, M.K. Bonde3, and L.J.

of Animal Science - Behaviour and Stress Biology, Aarhus University, Blichers

SC R

1Department

IP T

Pedersen1

Allé 20, DK-8830 Tjele, Denmark 2Department

of Animal Science - Epidemiology and management, Aarhus University, Blichers

of Development for Outdoor Livestock Production, Marsvej 43, DK-8960 Randers,

N

3Center

U

Allé 20, DK-8830 Tjele, Denmark

M

A

Denmark

ED

*Corresponding author: Sarah-Lina A. Schild Email: [email protected]

Highlights

Genotype and parity affected behaviour during parturition but not parturition

PT



CC E

duration

Genotype partly influenced maternal protectiveness



Duration of parturition affected stillbirth and postnatal mortality



Number of posture changes during parturition did not affect early postnatal

A



mortality

Abstract

Sida 1 av 42

High piglet mortality remains a major challenge in outdoor pig production. The use of hyper prolific sow genotypes giving birth to large litters may prolong parturition and cause parturition-related problems under outdoor conditions where birth assistance cannot easily be provided. Therefore, the aim of the current study was to investigate the course of parturition and maternal protectiveness in two sow genotypes and two

IP T

farrowing hut designs. Secondly, the aim was to investigate if the course of parturition (parturition duration and posture changes during parturition) constituted a significant

SC R

risk for pre- and postnatal piglet mortality. Sows gave birth in four batches of 10 to 11 Danbred Landrace × Yorkshire crossbred sows (in first and then second parity) and 10 to 12 Topigs Norsvin TN70 sows (in first and then second parity). Within batch, six

U

sows of each genotype gave birth in a communal hut, and the remaining sows in each

N

batch in standard A-frame huts. Neither genotype nor hut design significantly affected

A

parturition duration. Sows spent the majority of time lying in lateral position during

M

the first 3 hours of parturition (median 120.0 min, range 0-180 min) although they

ED

changed posture 23 times (median, range 0-110) during these three hours. The behaviour (time spent in lateral and sternal position, standing and number of posture

PT

changes) of the sows during the first 3 hours of parturition depended on both sow genotype and parity. Furthermore, Topigs Norsvin sows stayed closer to their piglets

CC E

during piglet handling, but sow agitation score was unaffected by genotype. A longer parturition increased the risk of having at least one stillborn piglet in the litter (P=0.003) and the risk of postnatal mortality (P=0.051). The hourly number of posture

A

changes during parturition did, however, not influence postnatal mortality.

Keywords: outdoor pig production, organic pig production, animal welfare, course of parturition, behaviour

Sida 2 av 42

1 Introduction A recent Danish study shows that piglet mortality remains a challenge in organic sow herds with outdoor farrowing systems (Rangstrup-Christensen et al., 2018). In their study Rangstrup-Christensen et al. (2018) found that 29.5 % of the total born piglets died before weaning, the main causes of mortality being stillbirth and crushing. Thus,

IP T

the findings suggest that a long parturition and posture changes by the sow constitute

risk factors for piglet mortality. Accordingly, 82 % percent of the stillborn piglets died

SC R

during parturition (Rangstrup-Christensen et al., 2018), suggesting prolonged

parturition as a main cause of stillbirths. Furthermore, Baxter et al. (2009) and Thorsen et al. (2017) showed that a long parturition increase the risk of stillborn piglets

U

in free-range and organic sow herds, respectively, and Thorsen et al. (2017) recorded

N

parturitions lasting up to 28.9 h illustrating that undesirably long parturitions do occur

A

in outdoor systems. Aside from the negative consequences for prenatal mortality, long

M

parturitions may also affect postnatal mortality as they increase the risk of piglets

(Herpin et al., 1996).

ED

experiencing hypoxia during birth, a condition related to reduced postnatal viability

PT

Litter size and parturition duration are related (e.g. Baxter et al., 2009) and so use of sows giving birth to smaller litters may decrease the risk of prolonged parturitions and

CC E

the related negative consequences for pre-weaning mortality. In countries like Denmark, France and Germany, organic systems, with outdoor as well as indoor farrowing, use the same hyper prolific sow genotypes as those used in conventional

A

indoor production (Früh et al., 2014). Different sow genotypes are bred with different breeding goals. For instance, Danbred Landrace Yorkshire crossbreds are bred for number of live piglets at day five postpartum (pp) whereas the Topigs Norsvin TN70 (Landrace Z-line crossbred) sow is bred for being able to nurse her own litter. These breeding goal differences impact e.g. litter size (Pedersen et al., 2018) why differences Sida 3 av 42

in the course of parturition may also be expected. Therefore, use of alternative sow genotypes may hold a potential for influencing sow parturition and through that lower piglet mortality in outdoor production. The Norsvin Landrace part of the TN70 sow has been selected under loose-housing conditions. It therefore seems likely that these housing conditions could promote sows, which are more careful and reactive towards

IP T

their piglets.

In addition to the use of alternative sow genotypes, implementation of birth assistance

SC R

could also hold a potential for reducing piglet mortality in free-range and organic pig

production systems. Within the conventional indoor pig production, birth assistance is routinely used to lower piglet mortality and White et al. (1996) found a substantially

U

decreased proportion of stillborn piglets for sows subjected to birth assistance. In

N

countries such as Denmark and the UK, free-range and organic parturient and lactating

A

sows are housed on pasture and give birth in huts, commonly in traditional farrowing

M

huts (e.g. A-frame huts) with only one, sow sized opening, making birth assistance

ED

during prolonged parturitions very difficult to conduct. Therefore, the first step to implement management initiatives to assist sows during long parturitions is to

PT

introduce farrowing huts that allow birth supervision and assistance to be conducted; such huts could potentially also ease the handling of piglets pp.

CC E

The primary aim of the current study was to investigate the course of parturition and maternal protectiveness in two sow genotypes (Danbred and Topigs Norsvin TN70) and two farrowing hut designs (A-frame and communal). The two hut designs

A

represented a traditional A-frame hut for housing a single sow and a communal hut with improved opportunities for supervision and intervention, housing four sows individually but under the same roof. The second aim was to investigate if the course of parturition (parturition duration and posture changes during parturition) constituted a significant risk for pre and postnatal piglet mortality. Sida 4 av 42

Giving the difference in breeding goals (litter size) it was hypothesized that parturition duration would be longer for the Danbred compared to the Topigs Norsvin sows. Sows seek isolation prior to parturition (Jensen, 1986). Sows housed in communal huts were able to hear and smell their neighbours and it was therefore expected that sows housed in these huts would be more restless (display more posture changes) during

IP T

parturition when compared to sows giving birth in A-frame huts. Furthermore, due to the different breeding goals and to the fact that the landrace part of the TN70 sows

SC R

were selected under loose-housed conditions, Topigs Norsvin sows were expected to

be more protective of their piglets during piglet handling than the Danbred sows. Lastly, it was hypothesized that longer parturitions would result in both higher pre-

U

and postnatal mortality and that a higher frequency of posture changes would increase

ED

2.1 Animals and housing

M

2 Materials and methods

A

N

postnatal mortality.

The study was conducted at the experimental herd at the Department of Animal

PT

Science, Aarhus University from November 2016 to August 2017. The study comprised 87 parturitions from 48 sows followed during their first and second parity. The

CC E

parturitions took place in four batches of 10 to 11 Danbred Landrace x Yorkshire crossbred sows and 10 to 12 Topigs Norsvin TN70 (Norsvin landrace x Z-line (Large white) crossbred) sows. Details on litter size and mortality rate for the two genotypes

A

are presented in Pedersen et al. (2018). Batches 1 and 2 represented sows during first parity, while batches 3 and 4 represented second parity. The sows were housed outdoors in paddocks throughout the entire project period, no birth assistance was provided.

Sida 5 av 42

Prior to parturition sows were housed in groups of 10-14 sows in the gestation field. Gestation paddocks measured 40 x 100 m and provided access to two communal huts (Length 234 x widthtop 110 x widthbottom190 x Height 110 cm) per paddock. Sows were grouped according to protein level in the diet (normal or low; normal, gestating: 7315kJ/kg, raw protein 11.1%; Green So Drægtig, Vestjyllands Andel A.m.b.a.,

IP T

Ringkjøbing, Denmark; low, diluted gestating: 7338kJ/kg, raw protein 10.2%;

Vestjyllands Andel A.m.b.a., Ringkjøbing, Denmark) and had ad libitum access to

SC R

water (in a through) and roughage (grass cover in the paddocks during summer and

spring and clover grass silage during autumn and winter). Ten days prior to expected parturition of the first sow per batch sows were moved to the farrowing field where

N

warm season (batches 3 and 4), access to a wallow.

U

they were kept in individual paddocks (18 × 25 m) with access to a hut, and, in the

A

Twelve sows per batch (six of each genotype) had access to a communal hut (Center of

M

Development for Outdoor Livestock Production, Marsvej 43, DK-8960 Randers,

ED

Denmark; Figure 1). Each communal hut had four pens for individually housed sows, and each of the four compartments housing one sow measured L:240 cm, W:250 cm.

PT

The pen separations in the communal hut consisted of solid walls that were 119 cm high, meaning, sows were unable to see one another, but they could hear and smell the

CC E

other animals in the hut and see personnel present in the middle aisle between the two sides of the hut. The remaining sows in the batch were housed in standard A-frame huts (Figure 2) with room for only one sow (L:230 cm, Wbottom:190 cm, Wtop:110 cm,

A

H:110 cm). In the sow entrance to each A-frame hut a board, and in the entrance to the communal hut a roller, prevented piglets from leaving the hut during the first 10 days after birth. Additionally, plastic curtains were placed in the sow entrance of both communal and A-frame huts during autumn and winter (batches 1 and 2).

Sida 6 av 42

All huts were supplied with barley straw prior to sow relocation to the farrowing field. In summer, approximately 7.6 kg of straw per square metre was provided, in autumn and spring approximately 10 kg/m2 and in winter 12.6 kg/m2. Between parturition batches, huts were relocated to new paddocks that had been without sows for at about 5 months.

IP T

Sows were fed once daily in the morning, and, as part of another study not reported further here, the sows were fed different protein levels (normal or low) balanced within

SC R

genotype and hut design. From the sows entered the farrowing field, they were fed a dry feed for lactating sows (sows on normal protein: 7623kJ/kg, raw protein 14.4%; sows on low protein: 7554kJ/kg, raw protein 12.5%; Vestjyllands Andel A.m.b.a.,

U

Ringkjøbing, Denmark). During spring and summer (batches 3 and 4), the clover grass

N

cover on the paddocks constituted ad libitum roughage, whereas in autumn and winter

A

(batches 1 and 2), when the grass cover was sparse, sows were provided ad libitum

2.2 Data collection

ED

M

clover grass silage. Sows had ad libitum access to water from a large trough.

PT

2.2.1 Behavioural recordings 2.2.1.1 Video recordings

CC E

All parturitions were monitored continuously by use of video cameras (IPCHDBW4100EP-0360B, Dahua Technology Co., Queen Anne House. 25-27 Broadway, UK) with a fitted wide-angle lens. Video recordings were saved digitally and analysed

A

by use of S-VIDIA Client MegaPixel (M. Shafro & Co., Riga, Latvia). Recordings were made from birth of first until birth of last piglet in the litter (defined as the parturition duration), irrespective of whether the pigs were stillborn or liveborn. Additionally, the time of birth of each piglet was recorded as well as sow postures during parturition (ethogram described in Table 1). Sida 7 av 42

2.2.1.2 Direct observations On day one (24 to 36 hours after birth of first piglet) and three postpartum sow protectiveness was scored during piglet processing (earmarked day 1 pp or castrated day 3 pp). For safety reasons the sows were locked out of the hut before collection of

IP T

piglets. Sows were lured out of the hut after which the entrance to the hut was blocked

and the piglets collected. A few sows were rather reluctant to leave the hut, therefore,

SC R

to avoid major risk for the working staff, piglets were captured using a net on a long

stick while the sows were still in the hut. Afterwards, the sow was forced out of the hut. For all sows, once the sows were outside their hut, the hut entrance was closed to

U

prevent sows from re-entering. For sows housed in communal huts, piglet handling

N

took place inside the hut. For sows housed in A-frame huts handling was, for safety

A

reasons, conducted inside a tractor transport box with solid sides. The transport box

M

was driven up close to the hut entrance once the sow had been lured out of the hut.

ED

Maternal protectiveness was scored every 30 sec for the first 10 min after the first pig was being processed by use of scan sampling. A sow agitation score of 1 referred to:

PT

sow not reacting to the piglet handing, sow is eating, drinking or performing another activity not directed towards the farrowing hut/transport box; 2 referred to: sow lifts

CC E

her head and turns it towards the farrowing hut/transport box but remains calm, the sow may be vocalizing; 3 sow walks in the paddock while vocalizing; 4 sow runs around the farrowing hut/transport box while vocalizing loudly; 5 sow is pushing or

A

biting the farrowing hut/transport box while vocalizing loudly. When noting the agitation score, sow distance to the farrowing hut/transport box was also recorded.

2.3.2 Piglet recordings

Sida 8 av 42

The day after parturition (24 to 36 hours after birth of first piglet), all piglets were counted, inspected, weighed and ear tagged for individual recognition. Piglets weighing less than 700 g were considered non-viable (e.g. Quiniou et al. (2002) and Rangstrup-Christensen et al. (2018) found that piglets with a birth weight <1 kg have low viability) and euthanized by blunt force trauma. At 3 days of age, piglets were once

IP T

again weighed and inspected. Furthermore, male piglets received 0.04 to 0.05 mL

Flunixin 50 mg and were castrated. Cross fostering was done within genotype once

SC R

within 3 days of parturition when sows gave birth to a surplus of piglets (more piglets

than available teats). Piglets were always added to an evenly aged or younger litter. Due to the loose-housing and the lack of a ‘sow-free area’ in e.g. A-frame huts, feeding with

U

artificial milk replacer, as is used in conventional production (De Vos et al., 2014), is

N

not possible under outdoor conditions. Complete artificial rearing on milk replacer is

A

prohibited in Denmark (Anonymous, 2016). Thus, in the current study, when cross

M

fostering was not possible (no sows with spare teats), the smallest of the surplus piglets

ED

were euthanized for ethical reasons by blunt force trauma. Humane endpoints for both sows and piglets during the study were defined as follows:

PT

unable to stand on own accord, severe injury (broken bone, deep wound). An additional humane endpoint for piglets was emaciation (conspicuous ribs, backbone

CC E

and hip bones).

2.3.3 Necropsies

A

Dead piglets from each litter were collected from parturition until day 3 postpartum. The date of collection and the sow from which they were taken were recorded. In addition, it was recorded whether piglets were euthanized or died of their own accord. The collected piglets were kept frozen and subsequently thawed at room temperature prior to necropsy. At necropsy, the ultimate cause of death of each piglet was recorded. Sida 9 av 42

The cause of death was divided into five categories: stillborn, crushed, euthanized, other and unknown. Stillborn piglets were identified by testing whether lung tissue sank when suspended in water. In addition, stillborn piglets were examined to determine whether time of death occurred before the onset of parturition or during parturition. Stillborn piglets with general autolysis of internal organs (rendering all

IP T

organs a homogeneous brick-red colour) were classified as having died prior to onset of parturition. Stillborn piglets presenting internal organs with distinctively different

SC R

colours (a brown/reddish liver, white and pale stomach and small intestines, and greenish large intestines) with no sign of autolysis were classified as having died during parturition. Crushed piglets were identified by subcutaneous oedema, internal and/or

U

external lacerations and/or fractures. At necropsy, euthanized piglets were identified

N

by the cause of death noted during collection. Piglets for which the cause of death could

A

not be determined were categorised as unknown. Piglets diagnosed with a cause of

M

death other than stillborn, crushed or euthanized were grouped together as “other”.

PT

2.4 Statistical analysis

ED

This category included piglets dying from infection, starvation and malformation.

For the statistical analyses, SAS (SAS 9.3, SAS Institute Inc., Cary, NC, USA) was used.

CC E

Parturition duration was calculated as the time interval between birth of first and last piglet in the litter irrespective of the last piglet being stillborn or alive. One parturition was an outlier in the dataset and lasted 35 hours, a sensitivity analysis was conducted

A

by removing this parturition. This did not change any results, thus the parturition was included in the analyses.

2.4.1 The duration of parturition

Sida 10 av 42

Parturition duration was square root transformed to achieve normally distributed residuals. The following final linear mixed model was used, 𝑌𝑖𝑗𝑘 = 𝐿𝑖𝑗𝑘 + 𝛿𝑖𝑗𝑘 + 𝜀𝑖𝑗𝑘

(model 1)

L is the parameter describing the effect of litter size at birth (i in Danbred, Topigs

IP T

Norsvin, j in parity 1 and 2, and k refers to sow identity). δ and ε are normal distributed. δ is the effect of (index i, j, k) of sow identity, and ε is the residual error term. The initial

SC R

model also included the variables genotype (G), hut design (HD), parity (P) and the two-way interactions P*G, G*HD, G*L, but these were removed by elimination of statistically non-significant effects (P>0.05). Results of the analysis are presented as

2.4.2 Sow behaviour

M

2.4.2.1 Behaviour during parturition

A

N

U

LS means ± Standard error.

ED

The median parturition duration was 4 h 18 min (range 66 min to 35 h. Twenty-two parturitions lasted >4 hours and 76 of 83 parturitions lasted at least 3 hours).

PT

Therefore, to include as many parturitions as possible in the analyses, only sow behaviour during the first 3 hours of parturition were investigated. The 3-hour time

CC E

interval was assumed to be representative for the average course of parturition. Furthermore, studies suggest that sows are most responsive in the beginning of the parturition in that sows are more active in this period compared to later during

A

parturition (Petersen et al., 1990; Jarvis et al., 1999; Thodberg et al., 1999; Pedersen et al., 2003). Thus, influences of e.g. hut design and sow genotype on behaviour during parturition were expected to be seen within this time-period. However, in order to account for the potential effect of parturition duration on sow behaviour the variable ‘total parturition duration’ was created and referred to the total duration of parturition Sida 11 av 42

in whole hours. The last commenced hour of each parturition was removed from the data set, as this was never a complete hour, and so total parturition duration was a variable from 1 to 11 (excluding one parturition that lasted 35 hours). For the description of sow behaviour during the first 3 hours of parturition, the categorical variable hour of parturition was also created where 1 referred to the first hour after

IP T

birth of first piglet, 2 referred to the second hour after birth of first piglet, and 3, to the third hour after birth of first piglet. In model 2 to 5 compound symmetry was used to

SC R

account for correlation between repeated observations within parturition.

Below, the final models, after elimination of statistically non-significant effects (P>0.05), are shown.

U

As most sows spent the majority of the parturition lying in lateral position, data were

N

left skewed and had a large proportion at or near the right limit of 3600 seconds per

A

hour. The response was therefore converted to the duration per hour, in minutes, not

M

spent lying in lateral position calculated as 3600-time spent lying in lateral position/60

ED

(termed transformed lateral). This response was analysed by the following final negative binomial generalised linear model

(model 2)

+ 𝛽5 𝐺 ∗ 𝑃

CC E

PT

log(µ) = β0 + β1 𝐻𝑃1 + β2 𝐻𝑃2 + β3 𝐺 + β4 P

β0 is the intercept referring to the reference combination of parturition hour 3, Topigs Norsvin genotype and parity 2. The effect of hour of parturition (HP) was modelled by

A

two indicator functions HP1 and HP2 being one for parturition hour 1 and 2, respectively, and zero for hour 3. The corresponding two coefficients are β1 and β2. The parameters β3, β4 and β5 describe the effects of genotype (G), parity (P) and their interaction (G*P), where G and P are indicator functions being 1 for Danbred and parity 1, respectively, and 0 otherwise. The initial model also included the variables hut Sida 12 av 42

design (HD), total parturition duration (PD), and the two-way interactions HP*G, HP*P, and HD*P.

The time spent lying in sternal position was analysed by the following final zero-

model included hour of parturition, genotype, parity and hut design. (model 3)

SC R

log(µ) = β0 + β1 𝐻𝑃1 + β2 𝐻𝑃2 + β3 𝐺 + β4 P

IP T

inflated negative binomial generalised linear model. The zero-inflation part of the

β0 refers to the intercept for the combination of parturition hour 3, genotype in Topigs Norsvin and parity 2. Hour of parturition (HP) was again modelled by the two indicator

U

functions HP1 and HP2 being 1 for parturition hour 1 and 2, respectively, and 0

N

elsewhere. The corresponding two coefficients are β1 and β2. The parameters β3 and β4

A

describe the effects of genotype (G) and parity (P), where G and P are indicator

M

functions being 1 for Danbred and parity 1, respectively, and 0 otherwise. The initial model also included the variables hut design (HD), total parturition duration (PD) and

PT

ED

the two-way interactions HP*G, HP*P, G*P and HD*P.

The time sows spent standing during parturition was analysed by the following zero-

CC E

inflated negative binomial generalised linear model. The zero-inflation part of the model included hour of parturition, genotype, parity and hut design.

A

log(µ) = β0 + β1 𝐻𝑃1 + β2 𝐻𝑃2 + β3 𝐺 + β4 P + β5 𝐻𝐷 + 𝛽6 𝐺 ∗ 𝑃 (model 4)

β0 refers to the intercept for hour of parturition in 3, genotype in Topigs Norsvin, parity in 2 and hut design in communal hut. β1 and β2 are the parameters describing the effect of hour of parturition (HP) in 1 and 2, respectively, in relation to the reference (HP in 3). HP1 and HP2 were again modelled as indicator functions. The parameters β3, β4, β5 Sida 13 av 42

and β6 describe the effects of genotype (G), parity (P), hut design (HD) and the G*P interaction. G, P and HT are indicator functions being 1 for Danbred, parity 1 and Aframe hut, respectively, and 0 otherwise. The initial model also included the variable total parturition duration (PD) and the two-way interactions HP*G, HP*P, and HD*P.

IP T

Sow posture changes were defined as changes between standing, sitting, lying in lateral position and lying in sternal position. For the analysis of the number of posture

SC R

changes during parturition, the following final negative binomial generalised linear model was used.

(model 5)

U

log(µ) = β0 + β1 𝐻𝑃1 + β2 𝐻𝑃2 + β3 𝐺 + β4 P

N

β0 is the intercept referring to the reference combination of parturition hour 3, Topigs

A

Norsvin genotype and parity 2. The effect of the three levelled categorical variable hour

M

of parturition (HP) was again modelled by the two indicator functions HP 1 and HP2, and the corresponding two coefficients were β1 and β2. The parameters β3 and β4

ED

describe the effects of genotype (G) and parity (P) where G and P are indicator

PT

functions being 1 for Danbred and parity 1, and 0 otherwise. The initial model also included the variables hut design (HD), total parturition duration (PD) and the two-

CC E

way interactions HP*G, HP*P, G*P and HD*P. Results are presented as rate ratio (RR) with 95% confidence interval (CI).

A

2.4.2.1 Maternal protectiveness The sow agitation (mean of the 20 recordings per sow) was analysed by a mixed model including genotype, hut design, parity, and day postpartum as class variables and an interaction between mother number and parity as a random effect. However, none of the variables significantly impacted sow protectiveness and so no final model is shown. Sida 14 av 42

For the analysis of sow distance (mean of the 20 recordings, in meters) during the protectiveness the following mixed model was used, 𝛾𝑖𝑗𝑘 = 𝛼𝑖𝑗 + 𝜁𝑖𝑗𝑘 + 𝜀𝑖𝑗𝑘

(model 6)

where α is the parameter describing the effect of sow genotype (index i in Danbred,

IP T

Topigs Norsvin, j in day 1,3), ζ is the normal distributed random effect accounting for

repeated measures within sow and genotype (index i,j,k where j in parity 1,2) and ε is

SC R

the normal distributed residual error. The initial model also included the variables: hut design, parity and day postpartum.

U

2.4.3 Piglet mortality

N

The variable inter birth interval between birth of piglets was created and calculated as

A

the mean duration between the birth of two consecutive piglets. Afterwards the

M

standard deviation of the mean inter birth interval was calculated. For the analyses of

ED

piglet mortality the number of sow posture changes (changes between standing, sitting, lying in lateral position and lying in sternal position) during the entire parturition

PT

(birth of first till birth of last piglet) was considered (termed total number of posture changes). The variable hourly number of posture changes was created and calculated

CC E

as the total number of posture changes divided by parturition duration. The continuous variables (parturition duration, inter birth interval, standard deviation of the inter birth interval, and litter size) were tested for correlation, and as parturition

A

duration, inter birth interval and standard deviation of the inter birth interval correlated significantly with each other, only parturition duration and litter size were included in the analyses.

2.4.3.1 Stillborn piglets Sida 15 av 42

Of the 87 parturitions, 55 were without stillbirths, 20 with one stillborn in the litter and 9 with two, while 3 had more than two stillborn piglets. The variable “stillbirth” was converted to a binomial variable given a value of 0 to litters with no stillborn and the value of 1 to litters with one or more stillborn piglets.

following binomial generalised linear model. (model 7)

SC R

𝑙𝑜𝑔𝑖𝑡(𝜋𝑖 ) = 𝛽0 + 𝛽1 𝑃𝐷

IP T

The odds of having litters with at least one stillborn piglet were analysed by use of the

β0 refers to the intercept, and β1 is the parameter describing the effect of parturition duration (PD). The initial model also included, birth litter size (L), parity (P) and their

U

possible two-way interactions (including with PD), but these effects were all removed

N

by elimination of statistically non-significant (P>0.05) effects. Results are presented

M

2.4.3.2 Early liveborn mortality

A

odds ratio (OR) with 95% CI.

ED

For the analysis of postnatal mortality, no euthanized piglets were included; only

PT

piglets dying of their own accord were considered. Cross-fostered piglets (N=10) that died (N=1) were counted as dead by their genetic mother – not the foster mother.

CC E

The proportion of piglets dying until day 3 pp (postnatal deaths / (live) litter size on day 1 pp (after euthanasia of non-viable piglets)) was analysed by a binomial

A

generalised linear mixed model with sow number as random effect. 𝑙𝑜𝑔𝑖𝑡(𝜋𝑖𝑗 ) = 𝛽0 + 𝛽1𝑃𝐷 + 𝛽2 𝐿 + 𝛽3 𝑃

(model 8)

β0 refers to the intercept for parity 2. β1 is the parameter describing the effect of parturition duration (PD), and β 2 is the parameter describing the effect of birth litter size (L). β3 describes the effect of parity 1 (P) in relation to the reference (parity 2). Sida 16 av 42

Parity was included as an indicator function, implying P is 1 (if parity in 1) or 0 (if parity in 2). The initial model also included, hourly number of posture changes and the twoway interactions P*PD, PD*L, P*L. Results of the analyses are presented as OR with 95 % CI. 3 Results

IP T

3.1 The duration of parturition

The median parturition duration was 4.3 h (range 66 minutes to 35 hours). The

SC R

parturition duration increased slightly with increasing litter size (F1,37=3.9, P=0.055), see Table 5, but was not significantly affected by genotype, hut design or parity. Figure 3 shows a scatter plot of the parturition duration against the litter size (total born).

U

The median duration of parturition in first parity was for Danbred 3.2 h (range 0.8-11)

N

and for Topigs Norsvin 4.2 h (range 2.0-10.8). In second parity the median duration

A

for Danbred was 6.4 hours (range 0.7-35) while it was 4.7 hours (range 2.0-8.9) Topigs

3.2 Sow behaviour

ED

M

Norsvin.

PT

3.2.1 Behaviour during parturition

Seventy-six of the 83 sows had a parturition lasting at least 3 hours. During the first 3

CC E

hours of parturition, these sows spent 103±29 minutes (raw mean±standard deviation) lying in lateral position, 4.0±10.1 minutes in sternal position, 1.5±2.8 minutes standing and changed posture 4.7±7.9 times. All sow behaviour depended significantly on hour

A

of parturition (see Table 2) and sows increased their lateral lying as parturition progressed while decreasing all other behaviour (see Table 2 and 3). Both the rate of lying in lateral position (Χ2 (1df)=5.7, P=0.017) and the rate of standing (Χ2 (1df)=5.9, P=0.015) were affected by an interaction between sow genotype and parity (Table 2). Topigs Norsvin were lying less in lateral position in first (34±18 min) compared to Sida 17 av 42

second parity (44±18 min) while the opposite was true for standing (7±1 and 3±1 min, respectively, for first and second parity). Furthermore, in first parity Topigs Norsvin sows were lying less in lateral position compared to first parity Danbred sows (51±12 min). The rate ratio of lying in sternal position and of posture changes were affected by genotype (Χ2 (1df)=29.9, P<0.001 and Χ2 (1df)=21.5, P<0.001, respectively for sternal

IP T

position and posture changes) and parity (Χ2 (1df)=7.4, P=0.007 and Χ2 (1df)=5.9,

P=0.015, respectively for sternal position and posture changes). Both the rate of lying

SC R

in sternal position and of posture changes were higher for Topigs Norsvin than Danbred sows and higher in first compared to second parity (Table 2). Hut design only

had a significant impact on the rate of standing (Χ2 (1df)=7.6, P=0.006) and sows stood

N

U

more in communal huts (6±9 min) compared to A-frame huts (3±5 min).

A

3.2.2 Maternal protectiveness

M

None of the investigated factors (genotype, hut design, parity, day pp) significantly

ED

affected the sow agitation score. Danbred sows were given a mean score of 4.1±0.2 (LS mean±standard error) and Topigs Norsvin a score of 3.8±0.1. Contrary, the distance

PT

of the sows during the agitation score depended on sow genotype (F1,47=14.72, P=0.0004). Topigs Norsvin sows were closer to the hut/transport box during scoring

CC E

(5.3±0.9 m) than Danbred sows (10.5±1.0 m).

3.1 Piglet mortality

A

Forty-six sows gave birth in first parity (22 Danbred and 24 Topigs Norsvin) and 41 of these again in their second parity (20 Danbred and 21 Topigs Norsvin). The median total litter size was 17.0 (range 6-27) piglets; 16 (range 6-23) for first-parity sows and 19 (range 8-27) for second-parity sows, and 19 (median, range 7-27) for Danbred and

Sida 18 av 42

16 (range 6-21) for Topigs Norsvin. The results of the piglet necropsies are listed in

A

CC E

PT

ED

M

A

N

U

SC R

IP T

Table 4.

Sida 19 av 42

3.1.1 Stillborn piglets Thirty-two litters included stillborn piglets, and 55 did not. On average, 0.6±0.1 (raw mean±standard error, range 0-7) piglets per litter were stillborn and the necropsies showed that approximately 71% of the stillborn piglets died during parturition. The odds of litters with stillborn piglets significantly increased with increasing

IP T

duration of parturition (Χ2 (2df)=8.7, P=0.003, Table 3). A 60-minute increase in parturition increased the risk of litters with stillbirths by 29%, corresponding to an

SC R

odds ratio of 1.29 (95% CI[1.05;1.57]).

3.1.2 Early mortality of liveborn piglets

N

per litter (not including euthanized piglets).

U

Within 3 days pp, 2.4±0.2 (raw mean±standard error, range 0-8) liveborn piglets died

A

The odds of early mortality increased slightly with increasing duration of parturition

M

(OR 1.05 95% CI[1.00-1.11], F1,35=4.1, P=0.051, Table 3). Both litter size (F1,35=5.2,

ED

P=0.028) and parity (F1,35=15.1 P<0.001) affected the risk of postnatal deaths. Firstparity sows had 115% higher odds of postnatal mortality compared to sows in their

PT

second parity (CI [1.44;3.21]). Within the two parities, the mean number of postnatal deaths was 2.5±0.3 (raw mean±standard error, range 0-8) in first parity and 2.2±0.3

CC E

(range 0-7) for second-parity sows. The odds of liveborn piglets dying increased with 7% for every extra piglet in the litter, corresponding to an OR of 1.07 (CI[1.01-1.14]). Hourly number of posture changes did not influence the risk of liveborn piglets dying

A

within 3 days pp, taking into account that non-viable and surplus piglets had been euthanized.

Sida 20 av 42

4 Discussion 4.1 The duration of parturition Contrary to the expected, genotype did not affect parturition duration. Due to the difference in breeding goals and reported litter sizes (Pedersen et al., 2018) Danbred sows were expected to have longer parturitions than Topis Norsvin sows, but this could

IP T

not be confirmed. Both sow genotypes had parturition durations, which may be

considered long (parturitions lasting < 240 min Lucia Jr et al. (2002); <180 minutes

SC R

Borges et al. (2005); <300 minutes Oliviero et al. (2010)). In second parity, Topigs Norsvin sows (282 min) had a 102 min shorter parturition duration compared to

second parity Danbred sows (384 min). The difference was not statistically significant,

U

but numerically the results indicated an increasing parturition duration in Danbred

N

sows with increasing parity. This increase could be linked to the increased litter size in

A

second parity. Since litter size and parity are correlated (Koketsu et al., 1999; Quesnel

M

et al., 2008; Hales et al., 2014) it is possible that greater differences in parturition

ED

duration would arise between the two sow genotypes in their later parities. In the present study the only factor affecting parturition duration was litter size, which

PT

corresponds with earlier studies (e.g. Baxter et al., 2009).

CC E

4.2 Behaviour during parturition 4.2.1 Sow behaviour In accordance with previous studies (Jensen, 1986; Petersen et al., 1990; Jarvis et al.,

A

1999), the results showed that sows spent the majority of the parturition process lying in lateral position; although only considering the first 3 hours after birth of first piglet. The sows generally displayed an increased rate of lying in lateral position and a decrease in all other observed behavioural measures (sternal position, standing, changing posture) the further parturition progressed. Sida 21 av 42

Sow posture and the number of posture changes were both affected by sow genotype and parity or an interaction between the two. Topig Norsvin sows, especially in first parity, displayed less lying in lateral position, more sternal lying, standing and posture changes than Danbred sows. These differences in behaviour could suggest that the Topigs Norsvin sows are more restless or perhaps more responsive to their piglets

IP T

during parturition, particularly in first parity. Parturition duration was numerically higher for second parity Danbred sows compared to both first and second parity Topigs

SC R

Norsvin sows. Furthermore, the Danbred sows gave birth to, numerically, larger litters than Topigs Norsvin sows. Thus, the increased duration of lateral lying for Danbred

sows could also indicate stronger birth contractions caused by enlarged hormonal

U

signals from the larger litter.

N

The influence of parity on sow posture and posture changes was likely caused by the

A

sows’ experience with the birthing process and presence of piglets. First-parity sows

M

may have reacted more strongly to the parturition process and been more responsive

ED

to the newborn piglets, and for example gotten up more to sniff their piglets. First parity sows are known to react more strongly to the presence of piglets and studies

2008).

PT

have shown more savaging in gilts compared to sows (Harris et al., 2003; Chen et al.,

CC E

The negative effects of a prolonged parturition on piglet survival and the fact that the majority of stillborn piglets die during parturition (Leenhouwers et al., 1999; Marchant et al., 2000; Rangstrup-Christensen et al., 2018), suggest an opportunity to save viable

A

piglets suffering hypoxia during prolonged parturitions by providing birth assistance. Within the indoor pig-production, initiatives are taken to assist sows with prolonged parturitions, and studies show that total mortality can be lowered with as much as 50% (White et al., 1996; Christison et al., 1997; Andersen et al., 2009) by implementing systematic birth supervision and caretaking of less viable piglets. The communal hut Sida 22 av 42

used in the current study allowed supervision and access for caretakers to each sow from a sow-protected aisle dividing the pens on each side, although the possibility for birth assistance was not used in the present study due to the obvious bias it would introduce in the results on parturition duration and behaviour. Still, the aforementioned hut design features would allow both birth supervision and assistance,

IP T

without jeopardising the safety of stockmen. However, sows seek isolation prior to (and

during) parturition (Jensen, 1986) and therefore the communal hut, where sows were

SC R

able hear and smell their neighbours, could compromise sows need for isolation. Contrary, to the expected, sows housed in the communal huts did not display more posture changes than sows giving birth in A-frame huts. The only behaviour affected

U

by hut design was standing. Sows in communal huts had a higher rate of standing

N

during parturition compared to sows housed in A-frame huts. The increased rate of

A

standing could be due to sows reacting to disturbance from their neighbours. Another

M

cause of the higher rate of standing could be that in the A-frame huts, sows had less

ED

floor space available compared to sows in communal huts and further the inward sloping sidewalls restricted sow movement, which might have affected the sows’

PT

willingness/motivation to spend time standing. Cronin et al. (1998) showed that sows housed indoors in pens (total area 9.1 m2) with a small nesting area (3.4 m2) spent

CC E

more time standing compared to sows housed in pens (total area 9.1 m 2) with a larger nesting area (4.3 m2), which is opposite to the current findings concerning floor space availability.

A

Since hut design only appeared to impact sow behaviour to a minor extent the communal hut may hold a potential for implementing birth supervision and assistance in outdoor production. However, development of hut design with similar possibility for supervision but incorporating fewer sows may be preferred. It is also worth considering that, outdoor-housed sows are not as habituated to the presence of Sida 23 av 42

humans as sows housed indoors. Therefore, birth supervision and assistance could cause disturbance, increasing the risk of piglet mortality; for example, if sows get up, it will increase the risk of piglet crushing. Thus, whether birth assistance could be beneficial outdoors needs further investigation.

IP T

4.2.2 Maternal protectiveness

The Norsvin part of the TN70 sow has been selected under loose-housing conditions

SC R

and the TN70 crossbred is bred with the goal of being able to care for its own litter. For this reason, it was hypothesized that Topigs Norsvin sows would be more protective of their piglets compared to Danbred sows. Maternal protectiveness has been suggested

U

to be beneficial for piglet welfare as indicated by an increased growth rate among

N

piglets belonging to litters of highly aggressive gilts (Marchant Forde, 2002). Increased

A

protectiveness among Topigs Norsvin sows was only partly confirmed. Topigs Norsvin

M

sows stayed closer to their piglets (the hut/transport box) during handling procedures

ED

than did Danbred sows, indicating higher levels of maternal protectiveness for Topigs Norsvin sows. Yet this was not reflected in the sow agitation test where no significant

PT

differences could be shown between the two sow genotypes, which payed only little

CC E

attention to their piglets during piglet processing.

4.1 Piglet mortality

4.1.1 Stillborn piglets

A

The results showed higher odds of litters with stillborn piglets when parturition duration increased, which is in accordance with previous findings for both indoor (Pedersen et al., 2011; Panzardi et al., 2013) and outdoor-housed (Baxter et al., 2009; Thorsen et al., 2017) parturient sows. A prolonged parturition will increase the risk of piglets experiencing, or even dying from, hypoxia during birth. The present results Sida 24 av 42

found that 71% of the stillborn piglets died during parturition. This is slightly lower than what has been shown in previous studies where 78-88% of the stillborn piglets died during parturition when identified by necropsy (Edwards et al., 1994; Marchant et al., 2000; Rangstrup-Christensen et al., 2018). The difference may be explained by the young sows used in the current study.

IP T

Surprisingly, litter size did not influence the risk of stillbirth in the present study, which

is in contrast to several previous studies (KilBride et al., 2012; Rangstrup-Christensen

SC R

et al., 2017; Schild et al., 2018). The lacking effect of litter size as well as parity in this study may be due to the correlation between parturition duration, litter size and parity. Consequently, some of the variation caused by litter size and parity may be explained

U

by parturition duration. In accordance, the present results showed a longer parturition

N

duration with increasing litter size. In the present study, parity was confounded with

A

time of year, as first-parity sows gave birth during winter, and second-parity sows gave

M

birth during spring/summer. Studies on outdoor-housed sows have suggested that

ED

time of year may influence the risk of stillborn piglets, when sows give birth in insulated huts (Schild et al.; Rangstrup-Christensen et al., 2017). Hence, it cannot be excluded

PT

that time of parturition (winter or spring/summer) may explain some of the parity

CC E

variation seen in the current study.

4.1.2 Postnatal mortality Postnatal mortality tended to increase with parturition duration. Hypoxia during birth

A

results in the birth of less viable piglets (Herpin et al., 1996) and so, prolonged parturitions were expected to increase postnatal mortality. Surplus piglets were euthanized, in the present study, starting with the smallest piglets and piglets that appeared weak or were injured. Thus, if prolonged parturitions affect piglet viability

Sida 25 av 42

then the effect of parturition duration on postnatal mortality may have been underestimated. Baxter et al. (2009) did not find an effect of parturition duration on postnatal mortality when parturient sows gave birth outdoors in huts. The parturition duration and the litter sizes (13.95 (surviving) and 14.96 (stillborn)) recorded by Baxter et al. (2009) are

IP T

both low compared to those found in the current study, which may explain the

difference in results between the studies. Although litter sizes in the present study were

SC R

larger than what previous studies report, they do resemble those found in other recent

studies using Danbred LxY sows (total born e.g. ~17.9 Hales et al. (2015); 16.1 (1st parity) and 18.7 (2nd and 3rd parity) Hansen et al. (2017); 18.4 Thorsen et al. (2017))

U

and the Danish national average recorded in 2017 (18.6 piglets, Hansen (2018)).

N

Postnatal mortality was unaffected by the hourly number of posture changes during

A

parturition despite the main cause of mortality being crushing. Crushing is one of the

M

main causes for postnatal mortality, indoors (Marchant et al., 2000; Pedersen et al.,

ED

2006) as well as outdoors (Baxter et al., 2009; KilBride et al., 2012; RangstrupChristensen et al., 2018). Posture changes resulting in the highest risk of crushed

PT

piglets include e.g. lying down from standing (Weary et al., 1996; Wechsler and Hegglin, 1997; Marchant et al., 2001) and rolling (Weary et al., 1996; Wechsler and

CC E

Hegglin, 1997; Marchant et al., 2001). Particularly during and immediately after parturition, piglets may be in risk of crushing as the vigilance or mobility of the piglets is limited (English and Smith, 1975). Weary et al. (1996) found the highest incidence

A

of crushing and near crushing during parturition compared to after parturition when sows were housed indoors in crates and pens and gave birth to around 10 total born piglets (197 piglet from 20 litters). The differences between the present results and those by Weary et al. (1996) could be caused by differences in housing and/or litter size. Sida 26 av 42

Conclusion In conclusion, no effect of sow genotype could be shown for the duration of parturition. However, sow posture changes during the first 3 hours of parturition depended on both sow genotype and parity while hut type did not show an effect. Sows in communal huts stood more than sows housed in A-frame huts suggesting, the presence of other sows

IP T

or the space available inside the hut affect behaviour during parturition.

Parturition duration affected stillbirth risk while no significant influence on liveborn

SC R

mortality was found. Despite crushing, being the main cause of early postnatal death, no effect of posture changes during partition could be shown for postnatal mortality.

N

U

There are no conflicts of interest to declare

A

Acknowledgements

M

This work was funded as part of the VIPiglets project under the Organic RDD 2

ED

programme, which is coordinated by International Centre for Research in Organic Food Systems (ICROFS). It has received grants, J.nr. 34009-13-0679, from the Green

PT

Growth and Development programme (GUDP) under the Danish Ministry of Food, Agriculture and Fisheries.

CC E

The authors gratefully acknowledge the stockmen at AU-Foulum and John Misa Obidah, Carsten Kjærulff Christensen, Maria Eskildsen, Uffe Krogh, Connie Hårbo Middelhede, Mads Ravn Jensen, Lars Bilde Gidlbjerg and Betty Skou for technical

A

assistance.

Sida 27 av 42

References Andersen, I.L., Haukvik, I.A., Boe, K.E., 2009. Drying and warming immediately after birth may reduce piglet mortality in loose-housed sows. Animal 3, 592-597. Anonymous, 2016. BEK nr 17 af 07/01/2016 Announcement on the protection of pigs [Bekendtgørelse om beskyttelse af svin], Ministry of Environment and Food of Denmark.

IP T

Baxter, E.M., Jarvis, S., Sherwood, L., Robson, S.K., Ormandy, E., Farish, M., Smurthwaite, K.M., Roehe, R., Lawrence, A.B., Edwards, S.A., 2009. Indicators of piglet survival in an outdoor farrowing system. Livestock Science 124, 266-276.

SC R

Borges, V.F., Bernardi, M.L., Bortolozzo, F.P., Wentz, I., 2005. Risk factors for stillbirth and

foetal mummification in four Brazilian swine herds. Preventive Veterinary Medicine 70, 165176.

U

Chen, C.Y., Gilbert, C.L., Yanga, G.C., Guo, Y.M., Segonds-Pichon, A., Ma, J.W., Evans, G.,

N

Brenig, B., Sargent, C., Affara, N., Huang, L.S., 2008. Maternal infanticide in sows: Incidence

M

Animal Behaviour Science 109, 238-248.

A

and behavioural comparisons between savaging and non-savaging sows at parturition. Applied

Christison, G.I., Wenger, I.I., Follensbee, M.E., 1997. Teat seeking success of newborn piglets

ED

after drying or warming. Canadian Journal of Animal Science 77, 317-319. Cronin, G.M., Dunsmore, B., Leeson, E., 1998. The effects of farrowing nest size and width on

PT

sow and piglet behaviour and piglet survival. Applied Animal Behaviour Science 60, 331-345.

CC E

De Vos, M., Che, L., Huygelen, V., Willemen, S., Michiels, J., Van Cruchten, S., Van Ginneken, C., 2014. Nutritional interventions to prevent and rear low-birthweight piglets. Journal of animal physiology and animal nutrition 98, 609-619.

A

Edwards, S.A., Smith, W.J., Fordyce, C., MacMenemy, F., 1994. An analysis of the causes of piglet mortality in a breeding herd kept outdoors. The Veterinary Record 135, 324-327. English, P.R., Smith, W.J., 1975. Some causes of death in neonatal piglets. Vet ann 15, 95-104. Früh, B., Bochicchio, D., Edwards, S., Hegelund, L., Leeb, C., Sundrum, A., Werne, S., Wiberg, S., Prunier, A., 2014. Description of organic pig production in Europe. Organic Agriculture 4, 83-92. Sida 28 av 42

Hales, J., Moustsen, V.A., Devreese, A.M., Nielsen, M.B.F., Hansen, C.F., 2015. Comparable farrowing progress in confined and loose housed hyper-prolific sows. Livestock Science 171, 64-72. Hales, J., Moustsen, V.A., Nielsen, M.B., Hansen, C.F., 2014. Higher preweaning mortality in free farrowing pens compared with farrowing crates in three commercial pig farms. Animal 8,

IP T

113-120. Hansen, C., 2018. National average for productivity | the pig production 2017

(Landsgennemsnit for produktivitet | Svineproduktionen 2017). Seges Svineproduktion,

SC R

https://svineproduktion.dk/publikationer/kilder/notater/2018/1819 accessed 22/01/2019.

Hansen, C.F., Hales, J., Weber, P.M., Edwards, S.A., Moustsen, V.A., 2017. Confinement of sows 24h before expected farrowing affects the performance of nest building behaviours but

U

not progress of parturition. Applied Animal Behaviour Science 188, 1-8.

A

journal of animal science 83, 819-821.

N

Harris, M.J., Li, Y.Z., Gonyou, H.W., 2003. Savaging behaviour in gilts and sows. Canadian

M

Herpin, P., Le Dividich, J., Hulin, J.C., Fillaut, M., De Marco, F., Bertin, R., 1996. Effects of the level of asphyxia during delivery on viability at birth and early postnatal vitality of newborn

ED

pigs. Journal of Animal Science 74, 2067-2075. Jarvis, S., McLean, K.A., Calvert, S.K., Deans, L.A., Chirnside, J., Lawrence, A.B., 1999. The

PT

responsiveness of sows to their piglets in relation to the length of parturition and the involvement of endogenous opioids. Applied Animal Behaviour Science 63, 195-207.

CC E

Jensen, P., 1986. Observations on the maternal behaviour of free-ranging domestic pigs. Applied Animal Behaviour Science 16, 131-142. KilBride, A.L., Mendl, M., Statham, P., Held, S., Harris, M., Cooper, S., Green, L.E., 2012. A

A

cohort study of preweaning piglet mortality and farrowing accommodation on 112 commercial pig farms in England. Preventive Veterinary Medicine 104, 281-291. Koketsu, Y., Takahashi, H., Akachi, K., 1999. Longevity, Lifetime Pig Production and Productivity, and Age at First Conception in a Cohort of Gilts Observed over Six Years on Commercial Farms. Journal of Veterinary Medical Science 61, 1001-1005. Sida 29 av 42

Leenhouwers, J.I., van der Lende, T., Knol, E.F., 1999. Analysis of stillbirth in different lines of pig. Livestock Production Science 57, 243-253. Lucia Jr, T., Corrêa, M.N., Deschamps, J.C., Bianchi, I., Donin, M.A., Machado, A.C., Meincke, W., Matheus, J.E.M., 2002. Risk factors for stillbirths in two swine farms in the south of Brazil. Preventive Veterinary Medicine 53, 285-292.

IP T

Marchant Forde, J.N., 2002. Piglet- and stockperson-directed sow aggression after farrowing and the relationship with a pre-farrowing, human approach test. Applied Animal Behaviour Science 75, 115-132.

SC R

Marchant, J.N., Broom, D.M., Corning, S., 2001. The influence of sow behaviour on piglet mortality due to crushing in an open farrowing system. Animal Science 72, 19-28.

Marchant, J.N., Rudd, A.R., Mendl, M.T., Broom, D.M., Meredith, M.J., Corning, S., Simmins,

U

P.H., 2000. Timing and causes of piglet mortality in alternative and conventional farrowing

N

systems. Veterinary Record 147, 209-214.

A

Oliviero, C., Heinonen, M., Valros, A., Peltoniemi, O., 2010. Environmental and sow-related

M

factors affecting the duration of farrowing. Animal Reproduction Science 119, 85-91. Panzardi, A., Bernardi, M.L., Mellagi, A.P., Bierhals, T., Bortolozzo, F.P., Wentz, I., 2013.

ED

Newborn piglet traits associated with survival and growth performance until weaning. Preventive Veterinary Medicine 110, 206-213.

PT

Pedersen, L.J., Berg, P., Jorgensen, G., Andersen, I.L., 2011. Neonatal piglet traits of importance for survival in crates and indoor pens. Journal of animal science 89, 1207-1218.

CC E

Pedersen, L.J., Damm, B.I., Marchant-Forde, J.N., Jensen, K.H., 2003. Effects of feed-back from the nest on maternal responsiveness and postural changes in primiparous sows during the first 24 h after farrowing onset. Applied Animal Behaviour Science 83, 109-124.

A

Pedersen, L.J., Jørgensen, E., Heiskanen, T., Damm, B.I., 2006. Early piglet mortality in loosehoused sows related to sow and piglet behaviour and to the progress of parturition. Applied Animal Behaviour Science 96, 215-232.

Sida 30 av 42

Pedersen, L.J., Schild, S.L.A., Bonde, M., Serup, T., 2018. Sow maternal characteristics in an outdoor farrowing system of two genetic selection lines, Book of Abstracts No. 24 (2018) of 69th EAAP, Dubrovnik, Croatia. Petersen, V., Recén, B., Vestergaard, K., 1990. Behaviour of sows and piglets during farrowing under free-range conditions. Applied Animal Behaviour Science 26, 169-179.

IP T

Quesnel, H., Brossard, L., Valancogne, A., Quiniou, N., 2008. Influence of some sow characteristics on within-litter variation of piglet birth weight. animal 2, 1842-1849.

Quiniou, N., Dagorn, J., Gaudré, D., 2002. Variation of piglets’ birth weight and consequences

SC R

on subsequent performance. Livestock Production Science 78, 63-70.

Rangstrup-Christensen, L., Krogh, M.A., Pedersen, L.J., Sørensen, J.T., 2017. Sow-level risk factors for stillbirth of piglets in organic sow herds. Animal 11, 1078-1083.

U

Rangstrup-Christensen, L., Schild, S.-L.A., Pedersen, L.J., Sørensen, J.T., 2018. Causes of

N

preweaning mortality in organic outdoor sow herds. Research in Veterinary Science 118, 171-

A

180.

M

Schild, S.L.A., Foldager, L., Bonde, M.K., Andersen, H.M.L., Pedersen, L.J., Does hut climate matter for piglet survival in organic production? animal, doi:10.1017/S175173111800201X.

ED

Thodberg, K., Jensen, K.H., Herskin, M.S., Jørgensen, E., 1999. Influence of environmental stimuli on nest building and farrowing behaviour in domestic sows. Applied Animal Behaviour

PT

Science 63, 131-144.

Thorsen, C.K., Schild, S.-L.A., Rangstrup-Christensen, L., Bilde, T., Pedersen, L.J., 2017. The

CC E

effect of farrowing duration on maternal behavior of hyperprolific sows in organic outdoor production. Livestock Science 204, 92-97. Weary, D.M., Pajor, E.A., Fraser, D., Honkanen, A.-M., 1996. Sow body movements that crush

A

piglets: a comparison between two types of farrowing accommodation. Applied Animal Behaviour Science 49, 149-158. Wechsler, B., Hegglin, D., 1997. Individual differences in the behaviour of sows at the nest-site and the crushing of piglets. Applied Animal Behaviour Science 51, 39-49.

Sida 31 av 42

White, K.R., Anderson, D.M., Bate, L.A., 1996. Increasing piglet survival through an improved

A

CC E

PT

ED

M

A

N

U

SC R

IP T

farrowing management protocol. Canadian Journal of Animal Science 76, 491-495.

Sida 32 av 42

Table 1 List of sow postures recorded during farrowing (after Thorsen et al. (2017)). A sow was scored as having changed to a new posture when she had rested in the given posture for more than 2 seconds. Table 1 List of sow postures recorded during farrowing (after Thorsen et al. (2017)). A sow was scored as having changed to a new posture when she had rested in the given posture for more than 2 seconds. Definition

Standing/walking

Maintaining an upright position on extended legs

Sitting

Body weight supported by hind quarters and both front legs stretched and front hoofs in the ground

Lateral position

Lying on one side with the udder exposed and with one shoulder touching the ground, her head raised or resting

Sternal position

Lying on the belly with the udder hidden, or partly hidden, and without a shoulder touching the ground, her head raised or resting

Outside

The sow is not in the hut

A

CC E

PT

ED

M

A

N

U

SC R

IP T

Posture

Sida 33 av 42

Table 2 Summary of the output from the final models of behaviour during the first 3 hours of parturition. First output is from the final model 4 (transforemed lateral position). Second output is from the final model 5 (sternal position). Third output is from the final model 6 (standing). Fourth output is from the final model 7 (posture changes). N refers to the number of observations, provided in brackets. Rate ratio (RR) for each variable with corresponding 95% confidence intervals (CI), X2 and p values are presented for the overall effect of each variable. DB refers to Danbred, TN to Topigs Norsvin, and 1 and 2 to first and second parity, respectively.

A

CC E

PT

ED

M

A

N

U

SC R

IP T

1

Sida 34 av 42

Table 2 Summary of the output from the final models of behaviour during the first 3 hours of parturition. First output is from the final model 4 (transforemed lateral position). Second output is from the final model 5 (sternal position). Third output is from the final model 6 (standing). Fourth output is from the final model 7 (posture changes). N refers to the number of observations, provided in brackets. Rate ratio (RR) for each variable with corresponding 95% confidence intervals (CI), X2 and p values are presented for the overall effect of each variable. 95% CI

Χ2

Hour of parturition

1 (83) 2 (78) 3 (67)

2.55 1.43 1

1.8; 3.6 1.0; 2.0

(2df)=26.7

Parity*Genotype1

DB 1 v. 2 TN 1 v. 2 1 DB v. TN 2 DB v. TN

0.53 1.75 0.28 0.91

0.3; 1.1 1.1; 2.9 0.2; 0.5 0.4; 1.9

(1df)=5.7

0.017

Hour of parturition

1 (83) 2 (78) 3 (67)

1.82 1.42 1

1.9; 2.8 0.9; 2.2

(2df)=7.0

0.030

Parity

1 (120) 2 (108)

1.63 1

1.2; 2.3

(1df)=7.4

0.007

Genotype

Danbred (101) Topigs Norsvin (127)

0.32 1

0.2;0.5

(1df)=29.9

<0.001

Hour of parturition

1 (83) 2 (78) 3 (67)

2.02 1.14 1

1.4;3.0 0.7; 1.8

(2df)=15.7

<0.001

A-frame (99) Communal (129)

0.63 1

0.5; 0.9

(1df)=7.6

0.006

DB 1 v. 2 TN 1 v. 2 1 DB v. TN 2 DB v. TN

0.78 1.71 0.82 1.80

0.5; 1.3 1.2; 2.5 0.6; 1.2 1.1; 3.0

(1df)=5.9

0.015

Hour of parturition

1 (83) 2 (78) 3 (67)

2.19 1.25 1

1.7; 2.9 0.9; 1.7

(2df)=30.6

<0.001

Parity

1 (120) 2 (108)

1.53 1

1.1; 2.1

(1df)=5.9

0.015

Hut design Standing

CC E

PT

Parity*Genotype1

Posture changes

p

N

U

SC R

IP T

RR

A

Sternal

Levels (N)

M

Transforme d lateral

Variable

ED

Model

<0.001

Danbred (101) 0.42 0.3;0.6 (1df)=21.5 <0.001 Topigs Norsvin (127) 1 1 DB refers to Danbred, TN to Topigs Norsvin, and 1 and 2 to first and second parity, respectively.

A

Genotype

Sida 35 av 42

Table 3 Sow behaviour recorded during the first three hours of parturition provided with respect to sow genotype (Dan: Danbred, ToN: Topigs Norsvin), parity (1st: first parity, 2 nd: second parity), hut design (A: A-frame hut, Com: communal hut) and hour of parturition, which reffers to the 1st (1), 2nd (2) or 3rd (3) hour after the birth of the first piglet in the litter. Values are presented as raw means ± standard deviation. Table 3 Sow behaviour recorded during the first three hours of parturition provided with respect to sow genotype (Dan: Danbred, ToN: Topigs Norsvin), parity (1st: first parity, 2 nd:

IP T

second parity), hut design (A: A-frame hut, Com: communal hut) and hour of parturition,

which reffers to the 1st (1), 2nd (2) or 3rd (3) hour after the birth of the first piglet in the litter.

Main effects

Interaction

Parity

Dan ToN

1st

2nd

Hut design

Hour of parturition

Genotype*parity

A

1

1stDan 2ndDan 1stToN 2ndToN

Com

2

3

U

Genotype

36±1845±19 49±16 10±12 4±7

9±5

7±12

A

Sternal 2±4 11±12 position (min)

N

Lateral position (min)

3±5 6±9 11±10 7±7

51±12 46±20 34±18 44±18

4±9

8±8 3±8

2±4

13±10 8±8

6±7

4±1

4±1

7±1

A

CC E

PT

ED

5±6 12±10

M

Standing (min) Posture changes (no)

SC R

Values are presented as raw means ± standard deviation.

Sida 36 av 42

3±1

Table 4 Results of the necropsies. The primary causes of piglet death listed together with the number of piglets dying from each cause. 1 Crushed

includes all piglets dying form sow inflicted injuries (over lain, trampled, bitten). live born piglets that were not crushed or euthanized. Includes death due to infection, starvation, non-viable piglets, malformations etc. 3 Euthanized due to: weight (piglets with a body weigh <700g were considered non-viable), injury or lack of teats available during cross fostering (surplus). Eighty-eight Danbred piglets were euthanized (56 from first parity sows and 32 from second parity sows) and 23 Topigs Norsvin (15 from first parity sows and 8 from second).

A

CC E

PT

ED

M

A

N

U

SC R

IP T

2 All

Sida 37 av 42

Table 4 Results of the necropsies. The primary causes of piglet death listed together with the number of piglets dying from each cause.

During parturition

30

Before parturition

11

Time unknown

5

Crushed1

154

Other2

14

Euthanized3

111

Unknown

28

IP T

Stillborn

Number of piglets

SC R

Primary cause of death

1 Crushed

includes all piglets dying form sow inflicted injuries (over lain, trampled, bitten). live born piglets that were not crushed or euthanized. Includes death due to infection, starvation, non-viable piglets, malformations etc. 3 Euthanized due to: weight (piglets with a body weigh <700g were considered non-viable), injury or lack of teats available during cross fostering (surplus). Eighty-eight Danbred piglets were euthanized (56 from first parity sows and 32 from second parity sows) and 23 Topigs Norsvin (15 from first parity sows and 8 from second).

A

CC E

PT

ED

M

A

N

U

2 All

Sida 38 av 42

Table 5 Summary of the output from the final models of piglet mortality and the duration of parturition. First output is from the final model 1 (stillbirth). Second output is from the final model 2 (postnatal mortality). Third output is from the final model 3 (parturition duration). N refers to the number of observations, provided in brackets. F and p values are presented for the overall effect of each variable. OR refers to odds ratio and 95% CI to the 95% confidence interval.

IP T

Table 5 Summary of the output from the final models of piglet mortality and the duration of parturition. First output is from the final model 1 (stillbirth). Second output is from the final refers to the number of observations, provided in brackets.

SC R

model 2 (postnatal mortality). Third output is from the final model 3 (parturition duration). N F and p values are presented for the overall effect of each variable. OR refers to odds ratio and 95% CI to the 95% confidence interval. Variable

Levels (N)

OR

Stillbirth

Parturition duration

Continuous (86)

1.29

Postnatal mortality

Parturition duration

Continuous (86)

Litter size

Continuous (86)

Parity

1 (45) 2 (41)

Litter size

Continuous (86)

1.05;1.57

Χ2 (1df)=8.7

0.003

1.05

1.00;1.11

F1,35=4.1

0.051

1.07

1.01;1.14

F1,35=5.2

0.028

2.15 1

1.44;3.21

F1,35=15.1

<0.001

F1,37=3.9

0.055

N

A

M

Test value p

A

CC E

PT

ED

Parturition duration

95% CI

U

Model

Sida 39 av 42

Figure 1. Sketch of the communal hut (Center of Development for Outdoor Livestock Production, Marsvej 43, DK-8960 Randers, Denmark). a) seen from the outside and b) seen from the insiside. Each hut had room for four individually housed sows. From the human walkway in the centre of the hut, a view of each sow compartment was available, and gates allowed access to each compartment from the walkway. Dotted lines show where farrowing rails were located. Each piglet creep area was heated with an eHEAT PLUS Heater version 1.0 (Orbital A/S, Trykkerivej 5, 6900 Skjern, Denmark). There was no flooring in the farrowing

A

CC E

PT

ED

M

A

N

U

SC R

IP T

hut or in the covered piglet creep area.

Sida 40 av 42

A

CC E

PT

ED

M

A

N

U

SC R

IP T

Figure 2 A sketch of the A-frame hut that housed one sow per hut (after Schild et al., 2018).

Sida 41 av 42

Figure 3 The parturition duration plotted against the total litter size at birth (including

A

CC E

PT

ED

M

A

N

U

SC R

IP T

stillborn).

Sida 42 av 42