Effects of carbamazepine on pituitary-adrenal function in healthy volunteers

Carbamazepine (CBZ) is a widely used therapeutic agent in seizure, pain, and mood disorders. Although CBZ has been shown to inhibit hypothalamic CRH secretion in vitro, limited data suggest that systemic CBZ induces pituitary-adrenal activation. Few data are available to reconcile these effects or clarify their mechanism(s), particularly in healthy human subjects. We report here a study of basal ACTH and cortisol secretion and their responses to ovine CRH administration in nine healthy volunteers, studied both during repeated (2-3 weeks) administration of CBZ and while medication free. CBZ significantly increased mean 24-h urinary free cortisol (mean +/- SE, 197 +/- 17 vs. 137 +/- 24 nmol/day; P less than 0.02) and evening basal total plasma cortisol (113 +/- 17 vs. 83 +/- 14 nmol/L; P less than 0.05) as well as cortisol-binding globulin-binding capacity (497 +/- 36 vs. 433 +/- 28 nmol/L; P less than 0.01). Despite the CBZ-induced hypercortisolism, plasma ACTH responses to CRH during CBZ treatment remained robust, rather than being suppressed by basal hypercortisolism. In fact, during CBZ treatment, we noted a positive correlation between the increase in basal plasma cortisol and the increase in the plasma ACTH response to CRH (r = 0.65; P less than 0.05). We also observed a reduction in cortisol-binding globulin-binding capacity after CRH administration (315 +/- 25 vs. 433 +/- 28 nmol/L; P less than 0.001), which was accentuated by CBZ treatment (342 +/- 19 vs. 497 +/- 36 nmol/L; P less than 0.001; magnitude of fall, -155 +/- 22 nmol/L on CBZ vs. -118 +/- 11 nmol/L off CBZ; P less than 0.05). We conclude that CBZ increases plasma cortisol secretion in healthy volunteers independent of its effect on plasma cortisol-binding capacity. This pituitary-adrenal activation seems to reflect a pituitary, rather than a hypothalamic, effect of CBZ. Hence, despite CBZ-induced hypercortisolism, the ACTH response to CRH remained robust in direct proportion to the CBZ-induced rise in basal plasma cortisol. Thus, we propose that the increased cortisol secretion observed during CBZ treatment reflects a relative inefficacy of glucocorticoid negative feedback at the pituitary. This pituitary-driven increase in cortisol secretion combined with the expected reduction in centrally directed CRH secretion could contribute to the anticonvulsant properties of CBZ.