Pre-partum straw-directed behaviour by sows in farrowing pens is positively associated with piglet survival

Non-confinement farrowing is generally associated with higher piglet mortality than farrowing crates, including increased risk from overlying by the sow (Baxter et al. 2012). These are significant welfare and economic issues that will undoubtedly influence producers’ consideration of adopting lower confinement housing for sows at farrowing and during lactation. However, public interest in achieving non-confinement during all stages of production is nevertheless a constant driver for change that cannot be ignored by industry. Large between-sow variation in piglet mortality in farrowing/lactation pens has been reported. For example, Andersen et al. (2005) investigated individual sow differences in pre-partum interaction with straw bedding to explain some of this variation. While research has identified the importance of straw in farrowing pens to promote maternal behaviour, the occurrence of variation also highlights the opportunity for selection of sows better-suited to farrowing in pens. Andersen et al. (2005) and Westin et al. (2015) reported positive associations between pre-partum straw-directed behaviour and careful behaviour by sows in the peri-partum period which indirectly was associated with lower mortality. In the present experiment we investigated the association between self-selection of straw by sows prior to farrowing in pens, performance of pre-partum sow behaviour and piglet survival. The hypothesis examined was that straw-directed behaviour would be positively associated with improved piglet survival in farrowing pens.

The pre-farrowing behaviour of 40 Large White-Landrace sows (parity 1–6) was collated from digital video records [M. Šafro & Co. Ltd. (MSH), Latvia]. Sows farrowed in pens measuring 2.4 m by 3.3 m. Each pen contained two areas: a ‘nest area’ (2.4 × 1.7 m) and a ‘non-nest’ area (2.4 × 1.6 m), separated by a 0.27 m high metal step-over barrier. The nest area incorporated internal sloped panels on the rear and one side wall to assist sow posture changing behaviour; the rear of the nest area also formed a heated piglet creep. A wire basket attached on the opposite sidewall was filled with 4 kg straw each morning before sows farrowed. The nest area floor consisted of a thin layer of wood shavings on a 30-mm thick rubber mat over solid concrete, sloped towards the barrier. The non-nest area contained the sow feeder and drinker, and the floor comprised both solid and slatted flooring. Four farrowing pens were located in a non-heated, partially insulated room. The study was conducted over 13 replicates in time, with 1–4 sows observed per replicate. The timing of piglet deaths was recorded, with cause of death confirmed by necropsy.

The video record of each sow’s farrowing event was collated using one-zero (binomial) sampling to record whether the sow performed any of 10 behaviours (see below) during 144, 10-min periods from 24 h pre-partum to the birth of the first-born piglet in the litter. Data were expressed as the mean probability that the specified behaviour was observed during any 10 min interval 24 h pre-partum, and analysed using correlation analysis (Genstat edn. 14.1; VSN International, UK). The listed behaviours were: (1) Take straw from rack; (2) Carry straw in mouth; (3) Root/nose straw on floor; (4) Paw at straw; (5) Root/nose pen walls; (6) Root/nose bare floor; (7) Feed; (8) Drink; (9) Defaecate; and (10) Urinate. Piglet mortality in the litters averaged 3.0 ± 2.71 (mean ± SD) piglets and ranged from 0 to 10 deaths per litter (0 to 83.3% of those born alive). There was an inverse association between the combined straw-directed behaviours (1 to 4) and piglet mortality (r = –0.328, P < 0.05), and piglet mortality due to overlying by the sow on d 1 post-partum (r = –0.352, P < 0.05). However, there were no associations between pre-partum nesting behaviour of the sow and piglet mortality due to overlying after d 1 post-partum (P > 0.05). Sows that spent more time in the nest area during 24 h pre-partum tended to have fewer piglet losses due to small/weak/chilled (r = –0.299, P = 0.061), and correspondingly, sows that spent more time outside the nest area pre-partum tended to have more piglet losses due to small/weak/chilled (r = 0.302, P = 0.059).

The results support the hypothesis and the findings of Andersen et al. (2005) and Westin et al. (2015), that increased straw-directed behaviour by sows in the 24 h pre-partum was associated with reduced piglet mortality in lactation, and specifically, due to reduced overlying by the sow within the first day of life.