Applying a nonlinear regression model to characterize cortisol responses to corticotropin-releasing hormone challenge

We previously applied a pharmacokinetic model to characterize stress-induced cortisol curves. In the current study, we apply this pharmacokinetic model to characterize cortisol responses to CRH challenge in children aged 6-16. Two hundred eleven sessions were completed beginning between 4 and 5 pm. Sessions included 30-40 minutes preinfusion baseline, 1 mu g/kg CRH infusion, 90-180 minutes of postinfusion measures, and 9-10 plasma cortisol samples per session. A simpler version of the original oral dose model best-fit responses to the CRH challenge: Response = alpha + beta*Time + lambda*T*exp(-kappa*T). Time was standardized so that CRH infusion was at time 0. T = max (0, Time) such that the nonlinear departure from baseline only operated in the postinfusion phase. Alpha represents baseline cortisol at time of infusion, beta is the slope of the linear approximation of the diurnal rhythm, lambda controls magnitude of peak change from baseline, and kappa is reactivity rate. The model converged and appeared to provide a good fit to "realworld" data in healthy children. Our proposed model provides a detailed characterization of cortisol response to CRH infusion and should allow for increased flexibility in characterizing the influence of individual and situational variables on specific aspects of cortisol response curves.