This study included saliva and sera from sows sampled in a Danish sow farm as previously described by Kaiser et al. [26, 27, 37]. Briefly, PDS + sows (n = 38) and PDS ÷ sows (n = 38) were sampled once a day from 2 days before parturition until the development of PDS, or for a maximum of 2 days after parturition. Before each morning feeding, saliva was sampled by additive-free cotton swabs (Salivette® Cortisol, Hounissen, Denmark), immediately centrifuged for 5 min at 1,000 × g, and stored at ÷ 80oC until analysis for CgA and cortisol as previously described . After the morning feeding, each sow underwent a thorough clinical examination that included measurements of the body temperature and the heart rate. Thereafter, blood was sampled from v. jugularis using tubes with the addition of EDTA (BD, New Jersey, US) for the analyses of total WBC, lymphocyte and neutrophil counts, and additive-free tubes (BD, New Jersey, US) were centrifuged at 3,000 × g for 10 min. and used in the analyses of serum cortisol, Fe, ALB, IL-1, IL-6, TNF-α, SAA, Hp, CRP, and 8-epi-PGF2α, as previously described . In the statistical analyses, the samples were retrospectively divided into seven different periods in relation to parturition of the first piglet (0 h): A. (÷ 60 to ÷ 36 h); B. (÷ 36 to ÷ 24 h); C. (÷ 24 to ÷ 12 h); D. (÷ 12 to 0 h); E. (0 to 12 h); F. (12 to 24 h); and G. (24 to 36 h p.p.). Based on the clinical examinations, the sow was diagnosed with PDS (PDS+) if she fulfilled two out of three criteria: 1) reduced appetite (the trough were not emptied within 30 minutes after feeding); 2) inflammation of the udder (redness, swelling and increased skin temperature of the udder); 3) rectal temperature ≥ 39.5°C.
Based on the same criteria, the PDS + sows were retrospectively matched with sows that remained healthy (PDS÷) according to Kaiser et al.
Saliva and serum were analysed for ADA1, ADA2 and tADA activities using a commercially available, spectrophotometric, automated assay (Adenosine Deaminase assay kit, Diazyme Laboratories, Poway, CA), based on the method described by Galanti et al. . In the analyses of ADA2 and tADA, the measurements were performed in the presence and absence, respectively, of erythro-9-(2-hydroxy-3-nonyl) adenina (EHNA), a specific ADA1 inhibitor  and ADA1 was calculated as the difference between the two measurements. The method was adapted to an automated analyzer (Olympus AU400, Olympus Diagnostica GmbH, Ennis, Ireland) following the manufacturer’s instruction with some modifications for its use in porcine saliva .
Two autoregressive linear regression models (A and B) were used in the statistical analysis by the PROC MIXED procedure of Statistic Analytical Software, Enterprise Guide 7.1 (SAS® Institute, Cary, North Carolina, USA).
Model A: OUTCOME PARAMETERij
= µ + TIMEi
where OUTCOME PARAMETERij
indicates ADA1, ADA2, tADA values for salivary and serum; µ was the value of the observations at time 0; TIMEi
was the explanatory variable time intervals (A to G); GROUPj
was the explanatory variable “PDS+/PDS÷”; TIME*GROUPij
was the interaction between “PDS+/PDS÷” and time intervals (A to G), and εij
was the random residual error term. If significant interaction occurred by the use of model A, differences between PDS + and PDS ÷ and differences between the seven different time intervals were accepted. In case of non-significant interaction, model B was used instead of model A: OUTCOME PARAMETERij
= µ + TIMEi
. The effect was considered non-significant if changes in TIMEi
were non-significant. The significance level was p
< 0.05. The sows body condition score and parity were incorporated as explanatory variables. To improve normality of residuals plots, natural logarithm transformation was used in the analysis of ADA1 (salivary and serum), ADA2 (salivary) and tADA (salivary and serum).
The association between the values of ADA1, ADA2, tADA (salivary and serum) and WBC, neutrophils, lymphocytes, TNF-α, IL-6, SAA, CRP, Hp, Fe, ALB, 8-epi-PGF2α, cortisol (salivary and serum), CgA, rectal temperature and heart rate was tested by a regression analysis in the PROC MIXED procedure of Statistic Analytical Software, Enterprise Guide 7.1 (SAS® Institute, Cary, North Carolina, USA).