Salivary cortisol is an unreliable correlate of serum cortisol in adult pet dogs and assistance dog puppies

Cortisol is widely used in mammals as a measure of HPA axis response. To estimate response to an acute stressor, minimize pain and ease sample collection, salivary cortisol has become preferred over serum cortisol across a variety of species. This includes domestic dogs in which research with laboratory dogs initially demonstrated the predicted relationship between cortisol concentrations in serum and salivary levels sampled within minutes of each other. The Model Population Hypothesis suggests a laboratory dog should be physiologically representative of all dogs. We provide a critical test of this idea by providing the first validation of salivary cortisol against serum measures in healthy puppies less than six months of age (n = 34; 8 to 20-week-old Labrador x Golden Retrievers) as well as a group of healthy adult pet dogs (n = 38). Following previously established methodology, blood and saliva were collected within 4 min of each other. Puppies were sampled multiple times while adults were sampled once. We found that salivary and serum cortisol are poorly correlated in our puppies r(216) = - 0.092, p = 0.178, and adult dogs (r(36) = 0.092, p = 0.582). Our results suggest that previously validated methods with laboratory dogs may not translate to healthy puppies and pet dogs, particularly those less than six months of age. Further research is now needed to identify a salivary sampling method that might allow for this less invasive sampling technique to be used in puppies and pet dogs living in a range of real-world contexts.