Background and objective: Adrenal insufficiency is caused by primary adrenal failure or by impairment of the corticotropic axis. In both situations, cortisol secretion is deficient, and hydrocortisone is a logical replacement therapy. However, no consensus guideline for dosing has been published, and clinicians adapt the dose empirically after only a clinical evaluation. Under this regimen, some patients receiving an inappropriately high dose of cortisol feel comfortable and also have an increased risk of adverse effects. We performed a pharmacokinetic study of cortisol in patients with adrenal insufficiency to evaluate plasma concentrations when the dosing was based on clinical examination and to develop a model allowing optimization of drug dosing.
Study design: This was a prospective, open-label study in two endocrinology departments and a clinical investigation centre (Assistance Publique Hôpitaux de Marseille, Marseille, France).
Methods: Fifty patients with primary (n = 20) or secondary (n = 30) adrenal insufficiency were recruited. All patients were given their usual hydrocortisone replacement regimen. Blood samples for cortisol measurements were drawn at 0600, 0800, 1000, 1200, 1400, 1600, 1800, 2000, 2200 and 0000 h. The observed values were compared with the known physiological range throughout the day (0800, 1600 and 0000 h). A population pharmacokinetic analysis was performed using nonlinear mixed-effects modelling software (NONMEM). The final pharmacokinetic model was then used to simulate several hydrocortisone dosing scenarios.
Results: Thirteen different treatment regimens for 50 patients were observed. The cortisol plasma concentrations were compared with the physiological range and showed that 79%, 55% and 45% of patients were over- or under-treated at 0800, 1600 and 2400 h, respectively. The cortisol concentrations showed wide variability and were best described using a one-compartment model with zero-order input and first-order elimination. The pharmacokinetic parameters (intersubject variability) were the following: duration of absorption 0.54 hour, volume of distribution 38.7 L (39.7%) and clearance 12.1 L/h (23.2%). The proportional residual error was 32.3%. This final model was then used to simulate 18 different dosing regimens. The regimen with the highest proportion of simulated patients within the physiological targets was 10 + 5 + 5 mg at 0730, 1200 and 1630 h, respectively. However, even with this regimen, about 54%, 44% and 32% of patients would remain over- or under-treated at 0800, 1600 and 2400 h, respectively.
Conclusions: Most patients with adrenal insufficiency are imperfectly treated with hydrocortisone relative to their plasma cortisol concentrations. Using simulation, a standard dosing regimen is suggested, which increases the proportion of patients within the physiological target concentrations. However, an individualized dose adjustment would be more accurate.