Salivary cortisol concentration in healthy dogs is affected by size, sex, and housing context

The aim of the article was to investigate the effect of site of sampling, size, and sex on the variations of salivary cortisol of healthy dogs. Samples of saliva were collected from dogs of private owners (n = 13), kennels (n = 4), and shelters (n = 2). For each dog, samples were collected at the first interaction of the day with man (T0) before the morning meal (6:00-8:00 am), 30 minutes after the meal (T1), and 30 minutes after the last interaction of the day with man (T2), when dogs were resting and apparently relaxed. A total of 92 dogs belonging to 17 different pure breeds or crossbred were eligible for the study, being 19 dogs privately owned, 47 recruited in kennels, and 26 hosted in shelters. Salivary cortisol concentrations of the dog population were not normally distributed, and data were transformed to natural logarithm (ln). The mean values ranged from −0.70 to 3.40 ln ng/mL, with an average of 0.90 ± 0.76 ln ng/mL, corresponding to 0.50, 30.00, and 3.48 ± 4.05 ng/mL. Mean salivary cortisol was significantly higher for dogs hosted in shelters than those privately owned or in the kennels (P < 0.05). Cortisol values from intact dogs did not differ between males and females, whereas for castrated males and spayed females, significantly lower values were found (P < 0.01 intact vs. castrated males; P < 0.05 intact vs. spayed females). Mean salivary cortisol concentration was significantly lower for giant and large-sized dogs than for small-sized dogs (P < 0.01), whereas mean cortisol for medium-sized dogs was not significantly different from the other sizes. The interaction of site with time of sampling was significant (P < 0.05), with the highest cortisol concentration at T2 for dogs privately owned and housed in the kennels and at T0 for dogs hosted in the shelters. This study, focused on healthy dogs, indicated that several factors can affect the concentration of salivary cortisol. Further studies also involving pathological conditions are required to identify critical values that can be used for clinical management settings.