Abstract
Context
Prenatal dexamethasone therapy is used in female foetuses with congenital adrenal hyperplasia to suppress androgen excess and prevent virilisation of the external genitalia. The traditional dexamethasone dose of 20 µg/kg/day has been used since decades without examination in clinical trials and is thus still considered experimental.
Objective
As the traditional dexamethasone dose potentially causes adverse effects in treated mothers and foetuses, we aimed to provide a rationale of a reduced dexamethasone dose in prenatal congenital adrenal hyperplasia therapy based on a pharmacokinetics-based modelling and simulation framework.
Methods
Based on a published dexamethasone dataset, a nonlinear mixed-effects model was developed describing maternal dexamethasone pharmacokinetics. In stochastic simulations (n = 1000), a typical pregnant population (n = 124) was split into two dosing arms receiving either the traditional 20 µg/kg/day dexamethasone dose or reduced doses between 5 and 10 µg/kg/day. Target maternal dexamethasone concentrations, identified from the literature, served as a threshold to be exceeded by 90% of mothers at a steady state to ensure foetal hypothalamic-pituitary-adrenal axis suppression.
Results
A two-compartment dexamethasone pharmacokinetic model was developed and subsequently evaluated to be fit for purpose. The simulations, including a sensitivity analysis regarding the assumed foetal:maternal dexamethasone concentration ratio, resulted in 7.5 µg/kg/day to be the minimum effective dose and thus our suggested dose.
Conclusions
We conclude that the traditional dexamethasone dose is three-fold higher than needed, possibly causing harm in treated foetuses and mothers. The clinical relevance and appropriateness of our recommended dose should be tested in a prospective clinical trial.
Introduction
Congenital adrenal hyperplasia (CAH) due to 21-hydroxylase deficiency is a recessively inherited disorder of adrenal steroidogenesis resulting in cortisol deficiency and, due to the activation of the hypothalamic-pituitary-adrenal (HPA) axis, in adrenal androgen excess (1). The latter is already present during foetal development resulting in virilisation of female foetuses and the current practice of urogenital reconstructive surgery in severe cases (1). Prenatal therapy with dexamethasone (Dex) has been used since its first description by David and Forest in 1984 (2, 3) at risk pregnancies for classic CAH due to 21-hydroxylase deficiency. Dex is a potent glucocorticoid (GC) with a long half-life and the ability to traverse the placenta and suppress the foetal hypothalamic-pituitary-adrenal axis and thus can prevent virilisation of the external genitalia caused by foetal adrenal androgen excess. If started before the window of sexual differentiation (7–12 week post conception), there is evidence that prenatal Dex can significantly reduce virilisation in female foetuses with CAH (4, 5).
Dex therapy during pregnancy, however, is still considered an experimental therapy (6) and current guidelines do not recommend a specific treatment protocol (1). Its use is highly controversial for several reasons (6, 7, 8). One main reason is its unclear safety for treated mothers and foetuses. High doses of Dex administered to pregnant animals have teratogenic effects (9), and prenatal GC overexposure in animals has resulted in postnatal hypertension, programming effects on glucose homeostasis, decreased glomerular filtration, fatty liver disease and multiple negative effects on brain development (6, 9). Clinical outcome studies of foetuses treated with prenatal Dex at risk pregnancies for CAH show conflicting results with potentially negative neuropsychological and behavioural outcomes in Dex-treated children for CAH (10, 11, 12, 13, 14, 15, 16, 17, 18, 19). All studies, however, included very low numbers of patients and thus low power. Possible adverse effects described in treated mothers are more striae, oedema, increased mean weight gain, mood swings, and a slightly increased occurrence of hypertension, pre-eclampsia and gestational diabetes (4, 5, 6).
Since its first description, the same experimental (traditional) Dex dose of 20 µg/kg/day (pre-pregnancy maternal body weight), to a maximum of 1.5 mg/day given in 2–3 daily doses, is used (2). As Dex is 50–80 times more potent than cortisol, the traditional prenatal Dex treatment provides already the mother with about six-fold her physiologic GC, that is, cortisol, need (6). Additionally, Dex is minimal if at all inactivated by placental 11β-hydroxysteroid dehydrogenase (20). Investigations in non-CAH affected and untreated pregnancies showed that physiological foetal cortisol concentrations are low in early gestation and reach only a small fraction (approximately 10%) of maternal cortisol concentrations in midgestation (21). Consequently, the currently used Dex treatment during pregnancy bears the risk of a massively increased foetal exposure with Dex (6). As potential adverse effects are dose-related, we asked ourselves if the currently used Dex dose can be reduced to minimise potential adverse events but still be efficacious with regard to the prevention of genital virilisation in female foetuses.
To investigate a rational dose recommendation, we developed a pharmacokinetic (PK)-based modelling and simulation framework. A key component is the availability of a PK model for Dex in prenatal CAH treatment. For its development, the nonlinear mixed-effects (NLME) modelling approach was applied. NLME modelling is a widely used and well-established gold standard approach for population PK analysis (22) and allows for simultaneous analysis of PK on the population as well as individual level. Characterising the PK of all individuals enables ,for example, the determination of dosages that are adequate for the large majority of the population. Additionally, patient factors that can explain interindividual variability (IIV) in the PK can be identified. NLME modelling has been previously applied to assess CAH treatment with hydrocortisone in paediatric patients (23, 24, 25). PK models are structured in so-called compartments, representing kinetically homogenous regions of the body, such as well-perfused and poorly perfused organs (26). Finally, we applied the developed PK model to generate a rationale for the lowest possible but still effective Dex dose for successful HPA axis suppression in the foetus, with the ultimate aim to suggest a reduced dosing regimen for prenatal CAH therapy.
Methods
A graphical overview of the stepwise pharmacokinetics-based modelling and simulation framework is depicted in Fig. 1.
Pharmacokinetics-based modelling and simulation framework. PK, pharmacokinetic; Dex, dexamethasone.
Citation: European Journal of Endocrinology 185, 3; 10.1530/EJE-21-0395
Pharmacokinetics-based modelling and simulation framework. PK, pharmacokinetic; Dex, dexamethasone.
Citation: European Journal of Endocrinology 185, 3; 10.1530/EJE-21-0395
Pharmacokinetics-based modelling and simulation framework. PK, pharmacokinetic; Dex, dexamethasone.
Citation: European Journal of Endocrinology 185, 3; 10.1530/EJE-21-0395
Selection of pharmacokinetic dexamethasone data base
The development of a PK model requires an underlying dataset including a population that is representative of the target population to be treated with DEX and thus is 'fit for purpose'. Therefore, literature research was conducted to identify clinical studies investigating Dex PK that could serve as a suitable database for the development of a Dex PK model. Criteria to be met were a study population that well-matched the pregnant target population with respect to health status, age, weight and sex. Further criteria were a well-matched Dex dose, administration of an oral-immediate release formulation and access to individual dosing and sampling data as well as PK and demographic data. The database preparation, the graphical analysis of the identified dataset and the plotting of the PK modelling and simulation results were performed using R (3.6.0) (27) and R Studio 1.3.1056 (28).
Population pharmacokinetic dexamethasone model development
This data base retrieved from literature was used for the development of a comprehensive PK model by applying NLME modelling. Models were developed in NONMEM (7.4.3, ICON, Dublin, Ireland) Development Solutions, Ellicott City, MD, USA) (29) using Perl speaks Nonmem (3.4.2, Uppsala University, Uppsala, Sweden) (30) embedded in the workbench Pirana (version 2.9.6) (31). One-, two- and three-compartment models were tested on the PK dataset and compared using the likelihood ratio test and the Akaike Information Criterion (AIC). Subsequently, interindividual variabilities (IIV) and covariate-parameter relations were investigated using the forward inclusion (P = 0.05, df = 1)/backwards deletion (P = 0.01, df = 1) approach (32). The reliability of the models was evaluated using objective function value (OFV), a quality criterion defined as minus twice the natural logarithm of the likelihood of observing the data given the model indicating how well the model can predict the data (26), and standard model evaluation techniques such as goodness-of-fit (GOF) plots (33) and visual predictive checks (VPC, n(simulations) = 1000) (34).
Identification of target maternal Dex concentration suppressing foetal HPA axis
A literature search was performed to determine the minimum maternal Dex threshold concentration that leads to successful suppression of the foetal HPA axis as a decision criterion for an appropriate Dex dose. To this end and since no foetal Dex target concentrations for HPA axis suppression were found for prenatal CAH therapy, as a proxy, physiological foetal cortisol concentrations were identified first. Taking into account the higher potency of Dex compared to cortisol (hydrocortisone), the required foetal Dex target concentration was calculated. Finally, to obtain a maternal Dex target concentration, literature was searched for maternal-foetal Dex blood concentration ratios.
Simulation of maternal dexamethasone dosing regimens reaching target
The developed PK model was then applied for identifying rational dosing regimens that suppress the foetal HPA axis by conducting stochastic simulations (n = 1000) in NONMEM 7.4.3, randomly sampling individual PK parameters based on the estimated interindividual variability of the developed PK model. With these stochastic simulations, not only the Dex concentration-time profiles for typical patients were obtained but also variability in these profiles between patients was taken into account. First, a virtual but typical target population was generated with uniformly distributed body weights ranging from 50 to 95 kg. This virtual target population (n = 124) was divided into a 'light' (50–72 kg) and 'heavy' (73–95 kg) group and the individuals were dosed according to the group’s median body weight (61 and 84 kg, respectively). These two weight groups enabled the simulation of a realistic scenario for a clinical study in which pharmacies would only need to compound one dose strength per study arm. Two dosing arms were simulated with half of the patients receiving the traditional dose of 20 μg/kg/day and the other half receiving reduced doses of 5, 6, 7.5, 9 or 10 μg/kg/day, divided into three single doses given every 8 h. Regimens with reduced doses where the 10th percentiles of the maternal concentration-time profiles, that is, 90% of the patients, exceeded the minimum maternal target threshold at steady state (after the 5th Dex administration), were judged as rational and effective Dex dosings. We assume that for the remaining 10% of the patients, having the minimum below the assumed threshold for only several minutes, a successful suppression of the foetal HPA axis might be anticipated.
Furthermore, in order to assess the impact of the assumed foetal-maternal Dex concentration ratio, a sensitivity analysis was conducted by simulating the administration of the reduced doses with lower than reported foetal-maternal Dex concentration ratios as ‘worst-case scenarios’.
Results
Selection of pharmacokinetic dexamethasone data base
The retrieved clinical study best matching the pre-set criteria was a bioavailability study with 24 healthy volunteers receiving a dose of 2 mg Dex as an orally administered immediate-release tablet (EudraCT number: 2008-001389-10) (35). The study population had a median age of 32 years and 15 of 24 volunteers (63%) of the healthy volunteers were female. Body weights ranged from 60 to 90 kg and were overlapping between sexes (Table 1). Plasma was sampled up to 24 h post-dose, resulting in a total of 432 plasma samples and 18 samples per individual. Total Dex concentrations were determined in plasma using a HPLC-UV assay. Method validation was performed in accordance with current International Conference on Harmonisation Guideline recommendations (Center for Drug Evaluation and Research, Guidance for Industry: Bioanalytical Method Validation, 2001) and the conference report of Shah et al. (36). The lower limit of quantification (LLOQ) was 0.7 ng/mL (1.78 nmol/L). Dex concentrations reported below the LLOQ were discarded for the graphical analysis and the subsequent model development; thus, the Dex PK database comprised 349 Dex concentrations.
Study characteristics (median (range)).
| Characteristics | Female (n = 15) | Male (n = 9) | Total (n = 24) |
|---|---|---|---|
| Age, years | 30.0 (22.0–54.0) | 38.0 (25.0–54.0) | 32.0 (22.0–54.0) |
| Body weight, kg | 65.0 (60.0–77.5) | 79.0 (73.0–90.0) | 70.0 (60.0–90.0) |
| Height, m | 1.66 (1.63–1.76) | 1.84 (1.72–1.88) | 1.71 (1.63–1.88) |
| BMI, kg/m2 | 23.2 (21.0–26.9) | 25.5 (20.7–27.0) | 23.7 (20.7–27.0) |
| Dexa*, nmol/L | 15.4 (1.12–39.0) | 10.1 (1.9–29.3) | 12.61 (1.12–39.0) |
*After peroral administration of a 2 mg dexamethasone immediate release tablet.
Dexa, dexamethasone concentration.
The individual Dex concentration-time profiles revealed a bi-phasic decline, that is, an initial phase with a steep concentration decline over time, followed by a phase with a less steep slope, in a semilogarithmic plot (Fig. 2), suggesting a two-compartment PK model structure (32). Moreover, the individual profiles suggested high interindividual variability between healthy volunteers.
Individual profiles (n = 24) of dexamethasone concentrations (nmol/L) over time (h) after administration of a 2 mg immediate-release tablet. Dashed line: Lower limit of quantification. Top: Percentage of concentrations below the lower limit of quantification at 0.25, 12, 16 and 24 h post-dose.
Citation: European Journal of Endocrinology 185, 3; 10.1530/EJE-21-0395
Individual profiles (n = 24) of dexamethasone concentrations (nmol/L) over time (h) after administration of a 2 mg immediate-release tablet. Dashed line: Lower limit of quantification. Top: Percentage of concentrations below the lower limit of quantification at 0.25, 12, 16 and 24 h post-dose.
Citation: European Journal of Endocrinology 185, 3; 10.1530/EJE-21-0395
Individual profiles (n = 24) of dexamethasone concentrations (nmol/L) over time (h) after administration of a 2 mg immediate-release tablet. Dashed line: Lower limit of quantification. Top: Percentage of concentrations below the lower limit of quantification at 0.25, 12, 16 and 24 h post-dose.
Citation: European Journal of Endocrinology 185, 3; 10.1530/EJE-21-0395
Population pharmacokinetic dexamethasone model development
PK models with two compartments were superior to one and three-compartmental models according to all model diagnostics, for example, the OFV was 55.1 points better than for the one-compartment models, whereas a third compartment did not result in any significant model improvement (OFV non-significantly better (2.20 points); AICs: 1558.489, 1507.399, 1509.197 with one, two, three compartments, respectively). The final PK model contained the first-order absorption of Dex with an estimated lag-time of 12 min and a first-order elimination process (schematic depiction: Fig. 3). Residual variability was small (14.6% CV; proportional model). All PK parameter estimates were plausible and in line with the literature. IIV (assuming a log-normal distribution of individual parameters) was moderate to high: on clearance (33.1% CV), central volume of distribution (36.0% CV), intercompartmental clearance (98.7% CV), peripheral volume of distribution (42.9% CV), and absorption rate constant (116% CV). The identified effect of body weight on clearance and volume of distribution parameters was implemented using theory-based allometric scaling (37). Standard model evaluation techniques, such as GOF plot and VPCs, showed that the PK model adequately predicted the measured Dex concentrations and that standard criteria for a suitable and reliable model were fulfilled. More details about the model development and evaluation are presented in the Supplementary material (see section on supplementary materials given at the end of this article).
Left: structure of the developed two-compartment dexamethasone (Dex) pharmacokinetic (PK) model describing maternal Dex PK after intake of a 2 mg Dex immediate-release tablet. Solid line boxes: central and peripheral PK model compartments. Dashed line box: Dex dosing compartment. Right: calculated foetal Dex and cortisol concentrations (dashed line boxes), based on foetal-to-maternal Dex plasma concentration ratio and Dex-cortisol potency ratio retrieved from literature. Pharmacokinetic parameters: bioavailability (F, fixed to 1), absorption rate constant (ka), absorption lag time (tlag), clearance (CL), volume of distribution of central maternal compartment (Vc), intercompartmental clearance (Q), volume of distribution of central maternal compartment (Vp).
Citation: European Journal of Endocrinology 185, 3; 10.1530/EJE-21-0395
Left: structure of the developed two-compartment dexamethasone (Dex) pharmacokinetic (PK) model describing maternal Dex PK after intake of a 2 mg Dex immediate-release tablet. Solid line boxes: central and peripheral PK model compartments. Dashed line box: Dex dosing compartment. Right: calculated foetal Dex and cortisol concentrations (dashed line boxes), based on foetal-to-maternal Dex plasma concentration ratio and Dex-cortisol potency ratio retrieved from literature. Pharmacokinetic parameters: bioavailability (F, fixed to 1), absorption rate constant (ka), absorption lag time (tlag), clearance (CL), volume of distribution of central maternal compartment (Vc), intercompartmental clearance (Q), volume of distribution of central maternal compartment (Vp).
Citation: European Journal of Endocrinology 185, 3; 10.1530/EJE-21-0395
Left: structure of the developed two-compartment dexamethasone (Dex) pharmacokinetic (PK) model describing maternal Dex PK after intake of a 2 mg Dex immediate-release tablet. Solid line boxes: central and peripheral PK model compartments. Dashed line box: Dex dosing compartment. Right: calculated foetal Dex and cortisol concentrations (dashed line boxes), based on foetal-to-maternal Dex plasma concentration ratio and Dex-cortisol potency ratio retrieved from literature. Pharmacokinetic parameters: bioavailability (F, fixed to 1), absorption rate constant (ka), absorption lag time (tlag), clearance (CL), volume of distribution of central maternal compartment (Vc), intercompartmental clearance (Q), volume of distribution of central maternal compartment (Vp).
Citation: European Journal of Endocrinology 185, 3; 10.1530/EJE-21-0395
Identification of target maternal Dex concentration suppressing foetal HPA axis
The minimum maternal Dex threshold concentration was derived as a target concentration to suppress the foetal HPA axis based on the following knowledge obtained in the literature research: in Goto et al. (38), a physiological foetal cortisol concentration of around 4 pmol/mg was determined in adrenal tissue at 8 weeks post conception. Assuming a tissue density of 1 g/cm3, the foetal target cortisol concentration was 4 pmol/mL or 4 nmol/L. Since Dex is reported to be between 50 and 80 times more potent than cortisol (39), the corresponding foetal target Dex concentration was derived to be 0.05–0.08 nmol/L. Considering the reported ratio of 0.45 for the foetal:maternal Dex concentration (40), the identified maternal plasma Dex threshold concentration was thereby calculated to be between 0.11 and 0.18 nmol/L.
Simulation of maternal dexamethasone dosing regimens
The developed PK Dex model combined with the derived target maternal Dex concentration was employed to evaluate maternal Dex concentration-time profiles of the traditional compared to reduced maternal Dex doses (62 individuals per dose group). All reduced doses of 5, 6, 7.5, 9 and 10 µg/kg/day met the pre-set decision criterion of being above the target 0.05–0.08 nmol/L maternal Dex threshold concentration to successfully suppress the foetal HPA axis at steady state (Supplementary Fig. 3). As shown in Fig. 4A, with maternal Dex concentrations for the reduced dose of 7.5 µg/kg/day, in 90% of the patients (i.e. 10th percentile), all Dex concentrations, including the minimum Dex concentration at a steady state (see grey box), were above the maternal Dex thresholds. Due to the dosing according to the group’s median body weight, almost no difference was visible in the profiles of the light and heavy body weight group, with the concentrations of the lightweight group being slightly lower which was also the case for further analyses. Compared to the traditional dose of 20 µg/kg/day, the Dex exposure of the 7.5 µg/kg/day dose was considerably reduced over the entire time period with 63.1 % lower maximum (dashed box) and 62.7% lower minimum (grey box) Dex concentrations at steady state (Fig. 4B: 10th percentile concentration-time profiles).
Dexamethasone (Dex) concentration–time profiles after administration of (A) 7.5 µg/kg/day as reduced dose for light (50–72 kg, dashed lines) and heavy (73–95 kg, solid lines) body weight groups (n = 62 patients per group with n = 1000 simulations each) with 10th percentile (black lines), median (dark grey lines) and 90th percentile (light grey lines), (B) 7.5 µg/kg/day as reduced dose (black solid line of A) and traditional dose (20 µg/kg/day, grey line) of 10 percentiles and light body weight group only. Dashed horizontal lines: Dex threshold concentration if Dex is 50- (upper line) or 80-fold (lower line) more potent than cortisol, light grey filled boxes: minimum Dex concentration at steady state, dashed non-filled box: maximum Dex concentration at steady state. Arrows: Dex dose administrations before (dashed arrows) and after (solid line arrows) steady state.
Citation: European Journal of Endocrinology 185, 3; 10.1530/EJE-21-0395
Dexamethasone (Dex) concentration–time profiles after administration of (A) 7.5 µg/kg/day as reduced dose for light (50–72 kg, dashed lines) and heavy (73–95 kg, solid lines) body weight groups (n = 62 patients per group with n = 1000 simulations each) with 10th percentile (black lines), median (dark grey lines) and 90th percentile (light grey lines), (B) 7.5 µg/kg/day as reduced dose (black solid line of A) and traditional dose (20 µg/kg/day, grey line) of 10 percentiles and light body weight group only. Dashed horizontal lines: Dex threshold concentration if Dex is 50- (upper line) or 80-fold (lower line) more potent than cortisol, light grey filled boxes: minimum Dex concentration at steady state, dashed non-filled box: maximum Dex concentration at steady state. Arrows: Dex dose administrations before (dashed arrows) and after (solid line arrows) steady state.
Citation: European Journal of Endocrinology 185, 3; 10.1530/EJE-21-0395
Dexamethasone (Dex) concentration–time profiles after administration of (A) 7.5 µg/kg/day as reduced dose for light (50–72 kg, dashed lines) and heavy (73–95 kg, solid lines) body weight groups (n = 62 patients per group with n = 1000 simulations each) with 10th percentile (black lines), median (dark grey lines) and 90th percentile (light grey lines), (B) 7.5 µg/kg/day as reduced dose (black solid line of A) and traditional dose (20 µg/kg/day, grey line) of 10 percentiles and light body weight group only. Dashed horizontal lines: Dex threshold concentration if Dex is 50- (upper line) or 80-fold (lower line) more potent than cortisol, light grey filled boxes: minimum Dex concentration at steady state, dashed non-filled box: maximum Dex concentration at steady state. Arrows: Dex dose administrations before (dashed arrows) and after (solid line arrows) steady state.
Citation: European Journal of Endocrinology 185, 3; 10.1530/EJE-21-0395
Sensitivity analysis
To test the sensitivity of our results to the assumed value of the foetal-to-maternal Dex plasma concentration ratio of 0.45 reported in the literature (40), the ratio was decreased to 0.3 and 0.25; For these decreased Dex transplacental transfer values, when 9 or 10 µg/kg/day was administered as a reduced dose respectively, for 90% of patients (10th percentile) the Dex concentration-time profiles entirely exceeded the two maternal threshold Dex concentrations (Supplementary Fig. 4). With reduced doses of 5 or 6 µg/kg/day, and with both lower ratios and Dex being 50 times as potent as cortisol, the minimum concentrations of the 10th percentile (90% of patients) Dex profiles were not exceeding the upper maternal target threshold. Thus, these doses were considered too low for successful suppression of the foetal HPA axis (Fig. 5). After administration of 7.5 µg/kg/day with a ratio of 0.3, the 10th percentile Dex profiles were exceeding the upper maternal target threshold entirely. However, with a ratio of 0.25, the minimum concentrations of the 10th percentile Dex profiles were slightly below the upper maternal target threshold (Fig. 5C). Taking into account that in this ‘worst-case-scenario’ only 0.47% or 1.47% (with Dex being 80 or 50 times as potent as cortisol) of the total simulated Dex concentrations were below the target thresholds at steady state, we nevertheless selected 7.5 µg/kg/day as the lowest maternal dosage resulting in sufficient HPA axis suppression in the foetus.
Sensitivity analysis for foetal:maternal dexamethasone concentration ratio. Simulated 10th percentile dexamethasone concentration-time profiles of light (50–72 kg) body weight group (n = 62 patients) after administration of reduced doses with 5 (A), 6 (B) and 7.5 (C) µg/kg/day, if foetal:maternal dexamethasone concentration ratio = 0.3 (top) or 0.25 (bottom) instead of 0.45 (ratio retrieved from literature). Dashed horizontal lines: dexamethasone threshold concentration if dexamethasone is 50- (upper line) or 80-fold (lower line) more potent than cortisol, light grey boxes: minimum dexamethasone concentration at steady state. Arrows: Dex dose administrations before (dashed arrows) and after (solid line arrows) steady state.
Citation: European Journal of Endocrinology 185, 3; 10.1530/EJE-21-0395
Sensitivity analysis for foetal:maternal dexamethasone concentration ratio. Simulated 10th percentile dexamethasone concentration-time profiles of light (50–72 kg) body weight group (n = 62 patients) after administration of reduced doses with 5 (A), 6 (B) and 7.5 (C) µg/kg/day, if foetal:maternal dexamethasone concentration ratio = 0.3 (top) or 0.25 (bottom) instead of 0.45 (ratio retrieved from literature). Dashed horizontal lines: dexamethasone threshold concentration if dexamethasone is 50- (upper line) or 80-fold (lower line) more potent than cortisol, light grey boxes: minimum dexamethasone concentration at steady state. Arrows: Dex dose administrations before (dashed arrows) and after (solid line arrows) steady state.
Citation: European Journal of Endocrinology 185, 3; 10.1530/EJE-21-0395
Sensitivity analysis for foetal:maternal dexamethasone concentration ratio. Simulated 10th percentile dexamethasone concentration-time profiles of light (50–72 kg) body weight group (n = 62 patients) after administration of reduced doses with 5 (A), 6 (B) and 7.5 (C) µg/kg/day, if foetal:maternal dexamethasone concentration ratio = 0.3 (top) or 0.25 (bottom) instead of 0.45 (ratio retrieved from literature). Dashed horizontal lines: dexamethasone threshold concentration if dexamethasone is 50- (upper line) or 80-fold (lower line) more potent than cortisol, light grey boxes: minimum dexamethasone concentration at steady state. Arrows: Dex dose administrations before (dashed arrows) and after (solid line arrows) steady state.
Citation: European Journal of Endocrinology 185, 3; 10.1530/EJE-21-0395
Discussion
This study presents a scientific rationale for Dex dosing in prenatal CAH therapy. We successfully developed a pharmacokinetic NLME modelling and simulation framework in order to investigate different reduced Dex dosing regimens for prenatal therapy in CAH. Using simulations and sensitivity analysis, a suggested dose of 7.5 µg/kg/day was determined, which represents about a third of the traditional dose and judged as effective according to the identified target threshold for successful suppression of foetal HPA axis and thus prevention of foetal female virilisation.
The chosen PK database for the PK model development well-matched with the target population. As opposed to typical healthy volunteers, the majority (63%) of the subjects enrolled in the clinical study were female. The body weight range (60–90 kg) well-matched the expected weight of women in their early pregnancy and could be incorporated as a covariate on clearance and volume of distribution parameters. Moreover, with a median age of 32 (interquartile range: 27–46), the subjects had a well-matched age for representing pregnant women. The administered Dex dose of 2 mg in the model development dataset is approximately four times higher than the traditional dose in prenatal CAH therapy for a pregnant patient weighing 70 kg. Cummings et al. (41) observed a linear binding within a Dex concentration range of around 0.02–2 µg/L which is corresponding to the Dex concentrations in our analysis. Another publication described a nonlinear binding of Dex to albumin (42). However, the binding was investigated for substantially higher Dex concentrations up to 2 mg/L. Moreover, Loew et al. estimated very similar PK parameters after oral administration of 0.5, 0.75 and 1.5 mg Dex in healthy females (43), which are well in line with the doses investigated in our simulations. Given these findings, the PK of Dex was assumed to be linear within the range of the simulated doses and the dose administered in the retrieved clinical study.
One limitation of the database is the absence of pregnant individuals. In near-term pregnant women, Tsuei et al. stated that the Dex clearance is higher than in non-pregnant women whereas pregnancy did not affect the terminal half-life or plasma protein binding of Dex (40). Similar findings were described by Ke et al. in a physiologically based pharmacokinetic model (PBPK) for Dex in late pregnancy (44). Since sexual differentiation takes place at 7 to 12 weeks post conception (38), efficacy of the prenatal Dex administration is most important in this time period. Dex PK in early stage of pregnancy has not been characterised in a PK model yet as, in general, only few PK models exist for the first trimester of pregnancy (45). In pregnancy PBPK models, organ volumes are assumed to steadily increase with relatively little changes within the first trimester (46). It was, therefore, assumed that in this early period of pregnancy, being the sensitive phase of genital development, changes of Dex PK were negligible for the analysis. Most important, from the retrieved clinical study, all relevant individual-level data were available.
The resulting parameter estimates from the final structural PK model were well in line with the knowledge on the PK of the well-studied drug Dex. The estimated typical clearance of 40.4 L/h was in accordance with the area under the curve (AUC) which was calculated in the publication presenting the PK model database (geometric mean of 48.1 ng/mL/h, corresponding to clearance of 41.6 L/h) (35). Moreover, the literature reports high variability in the PK of Dex (43) similar to the estimated interindividual variability parameters, as well as Dex PK models with a two-compartment structure and rapid first-order absorption (47, 48). In addition, the PK model adequately predicted the measured Dex concentrations (Supplementary Figs 1 and 2). The developed PK model was, therefore, judged well suitable for application in the subsequent stochastic simulations. These simulations showed that all reduced doses were effective according to the determined maternal Dex thresholds and underlying assumptions.
As the retrieved foetal-to-maternal Dex plasma concentration ratio was determined in near-term pregnant women (40), this ratio was identified as the most critical assumption. Thus, a sensitivity analysis was conducted by altering the ratio of 0.45 to 0.3 and, as a ‘worst-case-scenario’, to 0.25. In these investigated scenarios, 7.5 µg/kg/day met the pre-defined criterion to be the lowest effective dose, with around 90% of pregnant women exceeding the pre-defined upper maternal Dex concentration threshold at a steady state. For the assumed physiological foetal cortisol concentration, the cortisol concentration which was measured in adrenal tissue of human foetuses 8 weeks post-conception (38) was chosen as the only and, therefore, best information on human foetal cortisol concentrations during early pregnancy that could be found in the literature. Regarding the difference in potency between cortisol and Dex, the two factors of 50 and 80 (39) were used throughout the simulation-based evaluations, thereby assuming that the difference in potency between cortisol and Dex translates from adults and children to the foetus. By using both factors, the uncertainty on the target concentration was taken into account in the simulation study, as a difference in potency directly leads to required differences in target concentration. The recommended dose of 7.5 µg/kg/day corresponds to less than 1 mg/day (for body weights up to 130 kg), which is commonly regarded as a threshold Dex dose for iatrogenic Cushing’s syndrome (49) and would thus substantially decrease the mother’s risk for GC-related adverse events. In conclusion, a considerably, that is, about three-fold reduced recommended Dex dose for prenatal CAH therapy was identified based on NLME modelling and simulation and thereby based on a scientific rationale. These calculations can be the basis of a prospective research protocol which has been recommended for years (1, 4, 50, 51) to investigate the success of prenatal Dex therapy in CAH with regard to the prevention of virilisation in female foetuses and to prospectively monitor the adverse effects for the treated mother and child.
Supplementary materials
This is linked to the online version of the paper at https://doi.org/10.1530/EJE-21-0395.
Declaration of interest
V S and R M have nothing to declare. C K and W H report grants from an industry consortium (AbbVie Deutschland GmbH & Co. KG, AstraZeneca, Boehringer Ingelheim Pharma GmbH & Co. KG, Grünenthal GmbH, F. Hoffmann-La Roche Ltd, Merck KGaA and SANOFI) for the PharMetrX PhD program. C K, U N and O B report a grant from Diurnal Ltd. C K reports an additional grant from the Innovative Medicines Initiative-Joint Undertaking (‘DDMoRe’), the Federal Ministry of Education and Research within the Joint Programming Initiative on Antimicrobial Resistance Initiative (JPIAMR) and from the European Commission within in the Horizon 2020 framework programme (‘FAIR’). U F report grants from InfectoPharm Arzneimittel und Consilium GmbH. All funding was outside the submitted work.
Funding
This work was supported by the Deutsche Forschungsgemeinschaft (Heisenberg Professorship 325768017 to N R and Projektnummer: 314061271-TRR205 to N R).
Acknowledgement
The authors acknowledge the High-Performance Computing Service of ZEDAT at Freie Universitaet Berlin (https://www.zedat.fu-berlin.de/HPC/Home) for computing time.
References
- 1↑
Speiser PW, Arlt W, Auchus RJ, Baskin LS, Conway GS, Merke DP, Meyer-Bahlburg HFL, Miller WL, Hassan Murad MH & Oberfield SE et al. Congenital adrenal hyperplasia due to steroid 21-hydroxylase deficiency: an Endocrine Society clinical practice guideline. Journal of Clinical Endocrinology and Metabolism 2018 103 4043–4088. (https://doi.org/10.1210/jc.2018-01865)
- 2↑
David M, Forest MG. Prenatal treatment of congenital adrenal hyperplasia resulting from 21-hydroxylase deficiency. Journal of Pediatrics 1984 105 799–803. (https://doi.org/10.1016/s0022-3476(8480310-8)
- 3↑
Forest MG Prenatal diagnosis, treatment, and outcome in infants with congenital adrenal hyperplasia. Current Opinion in Endocrinology and Diabetes 1997 4 209–217. (https://doi.org/10.1097/00060793-199706000-00005)
- 4↑
Forest MG Recent advances in the diagnosis and management of congenital adrenal hyperplasia due to 21-hydroxylase deficiency. Human Reproduction Update 2004 10 469–485. (https://doi.org/10.1093/humupd/dmh047)
- 5↑
New MI, Carlson A, Obeid J, Marshall I, Cabrera MS, Goseco A, Lin-Su K, Putnam AS, Wei JQ, Wilson RC. Prenatal diagnosis for congenital adrenal hyperplasia in 532 pregnancies. Journal of Clinical Endocrinology and Metabolism 2001 86 5651–5657. (https://doi.org/10.1210/jcem.86.12.8072)
- 6↑
Miller WL Fetal endocrine therapy for congenital adrenal hyperplasia should not be done. Best Practice and Research: Clinical Endocrinology and Metabolism 2015 29 469–483. (https://doi.org/10.1016/j.beem.2015.01.005)
- 7↑
Hirvikoski T, Nordenstrm A, Wedell A, Ritzn M, Lajic S. Prenatal dexamethasone treatment of children at risk for congenital adrenal hyperplasia: the Swedish experience and standpoint. Journal of Clinical Endocrinology and Metabolism 2012 97 1881–1883. (https://doi.org/10.1210/jc.2012-1222)
- 8↑
New MI, Abraham M, Yuen T, Lekarev O. An update on prenatal diagnosis and treatment of congenital adrenal hyperplasia. Seminars in Reproductive Medicine 2012 30 396–399. (https://doi.org/10.1055/s-0032-1324723)
- 9↑
Khulan B, Drake AJ. Glucocorticoids as mediators of developmental programming effects. Best Practice and Research: Clinical Endocrinology and Metabolism 2012 26 689–700. (https://doi.org/10.1016/j.beem.2012.03.007)
- 10↑
Hirvikoski T, Lindholm T, Lajic S, Nordenström A. Gender role behaviour in prenatally dexamethasone-treated children at risk for congenital adrenal hyperplasia – a pilot study. Acta Paediatrica 2011 100 e112–e119. (https://doi.org/10.1111/j.1651-2227.2011.02260.x)
- 11↑
Hirvikoski T, Nordenström A, Lindholm T, Lindblad F, Ritzén EM, Lajic S. Long-term follow-up of prenatally treated children at risk for congenital adrenal hyperplasia: does dexamethasone cause behavioural problems? European Journal of Endocrinology 2008 159 309–316. (https://doi.org/10.1530/EJE-08-0280)
- 12↑
Hirvikoski T, Nordenström A, Lindholm T, Lindblad F, Ritzén EM, Wedell A, Lajic S. Cognitive functions in children at risk for congenital adrenal hyperplasia treated prenatally with dexamethasone. Journal of Clinical Endocrinology and Metabolism 2007 92 542–548. (https://doi.org/10.1210/jc.2006-1340)
- 13↑
Maryniak A, Ginalska-Malinowska M, Bielawska A, Ondruch A. Cognitive and social function in girls with congenital adrenal hyperplasia – influence of prenatally administered dexamethasone. Child Neuropsychology 2014 20 60–70. (https://doi.org/10.1080/09297049.2012.745495)
- 14↑
Meyer-Bahlburg HFL, Dolezal C, Baker SW, Carlson AD, Obeid JS, New MI. Cognitive and motor development of children with and without congenital adrenal hyperplasia after early-prenatal dexamethasone. Journal of Clinical Endocrinology and Metabolism 2004 89 610–614. (https://doi.org/10.1210/jc.2002-021129)
- 15↑
Meyer-Bahlburg HFL, Dolezal C, Haggerty R, Silverman M, New MI. Cognitive outcome of offspring from dexamethasone-treated pregnancies at risk for congenital adrenal hyperplasia due to 21-hydroxylase deficiency. European Journal of Endocrinology 2012 167 103–110. (https://doi.org/10.1530/EJE-11-0789)
- 16↑
Trautman PD, Meyer-Bahlburg HFL, Postelnek J, New MI. Effects of early prenatal dexamethasone on the cognitive and behavioral development of young children: results of a pilot study. Psychoneuroendocrinology 1995 20 439–449. (https://doi.org/10.1016/0306-4530(9400070-0)
- 17↑
Van’t Westeinde A, Karlsson L, Nordenström A, Padilla N, Lajic S. First-trimester prenatal dexamethasone treatment is associated with alterations in brain structure at adult age. Journal of Clinical Endocrinology and Metabolism 2020 105 dgaa340. (https://doi.org/10.1210/clinem/dgaa340)
- 18↑
Van’t Westeinde A, Zimmermann M, Messina V, Karlsson L, Padilla N, Lajic S. First trimester DEX treatment is not associated with altered brain activity during working memory performance in adults. Journal of Clinical Endocrinology and Metabolism 2020 105 e4074–e4082. (https://doi.org/10.1210/clinem/dgaa611)
- 19↑
Karlsson L, Barbaro M, Ewing E, Gomez-Cabrero D, Lajic S. Epigenetic alterations associated with early prenatal dexamethasone treatment. Journal of the Endocrine Society 2019 3 250–263. (https://doi.org/10.1210/js.2018-00377)
- 20↑
Seckl JR Glucocorticoids, feto-placental 11β-hydroxysteroid dehydrogenase type 2, and the early life origins of adult disease. Steroids 1997 62 89–94. (https://doi.org/10.1016/S0039-128X(9600165-1)
- 21↑
Kari MA, Raivio KO, Stenman UH, Voutilainen R. Serum cortisol, dehydroepiandrosterone sulfate, and steroid-binding globulins in preterm neonates: effect of gestational age and dexamethasone therapy. Pediatric Research 1996 40 319–324. (https://doi.org/10.1203/00006450-199608000-00021)
- 22↑
Pillai GC, Mentré F, Steimer JL. Non-linear mixed effects modeling – from methodology and software development to driving implementation in drug development science. Journal of Pharmacokinetics and Pharmacodynamics 2005 32 161–183. (https://doi.org/10.1007/s10928-005-0062-y)
- 23↑
Michelet R, Melin J, Parra-Guillen ZP, Neumann U, Whitaker JM, Stachanow V, Huisinga W, Porter J, Blankenstein O & Ross RJ et al. Paediatric population pharmacokinetic modelling to assess hydrocortisone replacement dosing regimens in young children. European Journal of Endocrinology 2020 183 357–368. (https://doi.org/10.1530/EJE-20-0231)
- 24↑
Melin J, Parra-Guillen ZP, Michelet R, Truong T, Huisinga W, Hartung N, Hindmarsh P, Kloft C. Pharmacokinetic/pharmacodynamic evaluation of hydrocortisone therapy in pediatric patients with congenital adrenal hyperplasia. Journal of Clinical Endocrinology and Metabolism 2020 105 E1729–E1740. (https://doi.org/10.1210/clinem/dgaa071)
- 25↑
Melin J, Parra-Guillen ZP, Hartung N, Huisinga W, Ross RJ, Whitaker MJ, Kloft C. Predicting cortisol exposure from paediatric hydrocortisone formulation using a semi-mechanistic pharmacokinetic model established in healthy adults. Clinical Pharmacokinetics 2018 57 515–527. (https://doi.org/10.1007/s40262-017-0575-8)
- 26↑
Mould DR, Upton RN. Basic concepts in population modeling, simulation, and model-based drug development. CPT: Pharmacometrics and Systems Pharmacology 2012 1 e6. (https://doi.org/10.1038/psp.2012.4)
- 27↑
R Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing, 2019. (available at: https://www.R-project.org/)
- 28↑
RStudio Team. RStudio: Integrated Development Environment for R. Boston, MA: Rstudio, PBC. 2020. (available at: http://www.rstudio.com/)
- 29↑
Beal S, Sheiner LB, Boeckmann A, Bauer RJ. NONMEM User’s Guides 1989–2009. Ellicott City, MD, USA: Icon Development Solutions, 2009.
- 30↑
Lindbom L, Pihlgren P, Jonsson EN. PsN-Toolkit – a collection of computer intensive statistical methods for non-linear mixed effect modeling using NONMEM. Computer Methods and Programs in Biomedicine 2005 79 241–257. (https://doi.org/10.1016/j.cmpb.2005.04.005)
- 31↑
Keizer RJ, Benten M, Beijnen JH, Schellens JH, Huitema AD. Piraña and PCluster: a modeling environment and cluster infrastructure for NONMEM. Computer Methods and Programs in Biomedicine 2011 101 72–79. (https://doi.org/10.1016/j.cmpb.2010.04.018)
- 32↑
Mould DR, Upton RN. Basic concepts in population modeling, simulation, and model-based drug development-part 2: introduction to pharmacokinetic modeling methods. CPT: Pharmacometrics and Systems Pharmacology 2013 2 e38. (https://doi.org/10.1038/psp.2013.14)
- 33↑
Owen JS, Fielder-Kelly J. Introduction to Population Pharmacokinetic/Pharmacodynamic Analysis with Nonlinear Mixed Effect Models. Hoboken, New Jersey: John Wiley & Sons, Inc., 2014.
- 34↑
Bergstrand M, Hooker AC, Wallin JE, Karlsson MO. Prediction-corrected visual predictive checks for diagnosing nonlinear mixed-effects models. AAPS Journal 2011 13 143–151. (https://doi.org/10.1208/s12248-011-9255-z)
- 35↑
Queckenberg C, Wachall B, Erlinghagen V, Di Gion P, Tomalik-Scharte D, Tawab M, Gerbeth K, Fuhr U. Pharmacokinetics, pharmacodynamics, and comparative bioavailability of single, oral 2-mg doses of dexamethasone liquid and tablet formulations: a randomized, controlled, crossover study in healthy adult volunteers. Clinical Therapeutics 2011 33 1831–1841. (https://doi.org/10.1016/j.clinthera.2011.10.006)
- 36↑
Shah VP, Midha KK, Dighe S, McGilveray IJ, Skelly JP, Yacobi A, Layloff T, Viswanathan CT, Cook CE, McDowall RD. Analytical methods validation: bioavailability, bioequivalence and pharmacokinetic studies. Conference report. European Journal of Drug Metabolism and Pharmacokinetics 1991 16 249–255. (https://doi.org/10.1007/BF03189968)
- 37↑
Holford NHG A size standard for pharmacokinetics. Clinical Pharmacokinetics 1996 30 329–332. (https://doi.org/10.2165/00003088-199630050-00001)
- 38↑
Goto M, Hanley KP, Marcos J, Wood PJ, Wright S, Postle AD, Cameron IT, Mason JI, Wilson DI, Hanley NA. In humans, early cortisol biosynthesis provides a mechanism to safeguard female sexual development. Journal of Clinical Investigation 2006 116 953–960. (https://doi.org/10.1172/JCI25091)
- 39↑
Rivkees SA, Crawford JD. Dexamethasone treatment of virilizing congenital adrenal hyperplasia: the ability to achieve normal growth. Pediatrics 2000 106 767–773. (https://doi.org/10.1542/peds.106.4.767)
- 40↑
Tsuei SE, Petersen MC, Ashley JJ, McBride WG, Moore RG. Disposition of synthetic glucocorticoids. Clinical Pharmacology and Therapeutics 1980 28 88–98. (https://doi.org/10.1038/clpt.1980.136)
- 41↑
Cummings DM, Larijani GE, Conner DP, Ferguson RK, Rocci ML. Characterization of dexamethasone binding in normal and uremic human serum. DICP: the Annals of Pharmacotherapy 1990 24 229–231. (https://doi.org/10.1177/106002809002400301)
- 42↑
Gonciarz A, Kus K, Szafarz M, Walczak M, Zakrzewska A, Szymura-Oleksiak J. Capillary electrophoresis/frontal analysis versus equilibrium dialysis in dexamethasone sodium phosphate-serum albumin binding studies. Electrophoresis 2012 33 3323–3330. (https://doi.org/10.1002/elps.201200166)
- 43↑
Loew D, Schuster O, Graul EH. Dose-dependent pharmacokinetics of dexamethasone. European Journal of Clinical Pharmacology 1986 30 225–230. (https://doi.org/10.1007/BF00614309)
- 44↑
Ke AB, Milad MA. Evaluation of maternal drug exposure following the administration of antenatal corticosteroids during late pregnancy using physiologically-based pharmacokinetic modeling. Clinical Pharmacology and Therapeutics 2019 106 164–173. (https://doi.org/10.1002/cpt.1438)
- 45↑
Dallmann A, Mian P, Anker Van den J, Allegaert K. Clinical pharmacokinetic studies in pregnant women and the relevance of pharmacometric tools. Current Pharmaceutical Design 2019 25 483–495. (https://doi.org/10.2174/1381612825666190320135137)
- 46↑
Dallmann A, Pfister M, van den Anker J, Eissing T. Physiologically based pharmacokinetic modeling in pregnancy: a systematic review of published models. Clinical Pharmacology and Therapeutics 2018 104 1110–1124. (https://doi.org/10.1002/cpt.1084)
- 47↑
Li J, Chen R, Yao QY, Liu SJ, Tian XY, Hao CY, Lu W, Zhou TY. Time-dependent pharmacokinetics of dexamethasone and its efficacy in human breast cancer xenograft mice: a semi-mechanism-based pharmacokinetic/pharmacodynamic model. Acta Pharmacologica Sinica 2018 39 472–481. (https://doi.org/10.1038/aps.2017.153)
- 48↑
Hochhaus G, Barth J, Al-Fayoumi S, Suarez S, Derendorf H, Hochhaus R, Möllmann H. Pharmacokinetics and pharmacodynamics of dexamethasone sodium-m-sulfobenzoate (DS) after intravenous and intramuscular administration: a comparison with dexamethasone phosphate (DP). Journal of Clinical Pharmacology 2001 41 425–434. (https://doi.org/10.1177/00912700122010285)
- 49↑
Ceccato F, Artusi C, Barbot M, Lizzul L, Pinelli S, Costantini G, Niero S, Antonelli G, Plebani M, Scaroni C. Dexamethasone measurement during low-dose suppression test for suspected hypercortisolism: threshold development with and validation. Journal of Endocrinological Investigation 2020 43 1105–1113. (https://doi.org/10.1007/s40618-020-01197-6)
- 50↑
Clayton PE, Oberfield SE, Martin Ritzén E, Sippell WG, Speiser PW, Hintz RL, Savage MO. Consensus: Consensus statement on 21-hydroxylase deficiency from the Lawson Wilkins Pediatric Endocrine Society and the European Society for Pediatric Endocrinology. Journal of Clinical Endocrinology and Metabolism 2002 87 4048–4053. (https://doi.org/10.1210/jc.2002-020611)
- 51↑
Technical Report: congenital adrenal hyperplasia. Section on Endocrinology and Committee on Genetics. Pediatrics 2000 106 1511–1518. (https://doi.org/10.1542/peds.106.6.1511)
