Saliva cortisol and heart rate measurements of nurse sows during lactation compared to control sows

Nurse sows are used in piggeries with hyper-prolific sows to manage large litters. It is not known if nurse sows experience prolonged stress by having to stay in farrowing crates beyond the normal weaning time (Baxter et al. 2013). Our aim was to quantify long-term saliva cortisol as a measurement of stress of nurse sows compared to sows weaning their piglets at d 25 of lactation (control) and compare heart rate responses to d 20. A method called ‘cascade fostering’ using two lactating sows is normally performed in Denmark. In this method, the first nurse sow (N1) has her own piglets removed after a week and receives surplus newborn piglets that she fosters until weaning. The second nurse sow (N2) weans her own litter after 21 days and receives the litter from N1, which she rears until weaning. It was hypothesised that N1 and 2 sows would have increased saliva cortisol throughout lactation compared to control sows.

In total, 60 sows (n = 20) were randomly allocated to become a control, N1 or N2 sow in the same section over two time periods (summer 2013 and winter 2013/2014). Saliva was collected on d 6, 13, 20 and 24 at 1000 h, 1300 h and 1600 h for all sows and pooled on a daily basis for analysis. Additional saliva samples were taken on d 31 for N1 and N2 and d 38 for N2 for long-term measurements. Saliva samples were analysed for cortisol using a Salivary Cortisol kit (Salimetrics, UK). Pulse belts (model RS800CX, Polar Electro Oy, Finland) were placed around the chest of the sow from Monday to Wednesday to measure heart rate. Recordings were measured from the morning and continued to late afternoon. Specific time points (1000 h, 1300 h, 1600 h and 1900 h) were chosen to compare mean heart rate (HR), in 5 min intervals. Data were analysed using PROC MIXED (SAS^®; USA). Cortisol data was not normally distributed and therefore logarithmically transformed before analysis. Results presented here are arithmetic back-transformed data.

Results showed that there was no effect of treatment on saliva cortisol, but an effect of day (P < 0.001) with saliva cortisol declining throughout lactation (Fig. 1a). The N1 sows tended to have lower cortisol values (8.3 nmol/l) on d 31 than on d 24 (11.5 nmol/l; P = 0.08), and N2 sows had lower cortisol values on d 38 (7.4 nmol/l) and on d 31 (7.5 nmol/l) than on d 24 (11.1 nmol/l; P < 0.05). Heart rate values increased throughout lactation (P < 0.001) but remained unaffected by treatment (Fig. 1b).

Fig. 1.  (a) Pooled salivary cortisol for control (black), N1 (grey) and N2 (dotted) sows presented as arithmetic back-transformed values and sows differing significantly on day *(P < 0.05), **(P < 0.01) when compared to themselves. (b) Mean heart rate (± pooled SE) during lactation, for control (), N1 () and N2 () sows. Letters denote effect of time. ^a,bSignificant effect of time (P < 0.05).

These data indicated that saliva cortisol levels declined throughout lactation with no differences in saliva cortisol levels between control and NURSE sows. Heart rate increased throughout lactation to d 20 probably due to the increase in milk production. Salivary cortisol levels as indicators of stress, suggested no additional long-term effects of being selected as a nurse sow.