Residual endogenous corticosteroid production in patients with adrenal insufficiency

Abstract Objective This study aimed at comparing precursors of endogenous corticosteroid production in patients with primary adrenal insufficiency and in secondary adrenal insufficiency. Design Twenty patients with primary adrenal insufficiency and matched controls and 19 patients with secondary adrenal insufficiency participated in this ancillary analysis of two different studies. Patients and measurements Patients with primary adrenal insufficiency were on stable hydrocortisone and fludrocortisone therapy. Patients with secondary adrenal insufficiency received two different doses of hydrocortisone in a randomized crossover study. Main outcome measures were concentrations of precursors of cortisol and aldosterone measured by LC‐MS/MS Results Compared to controls, progressively lower concentrations of the glucocorticoid precursors 11‐deoxycortisol, 11‐deoxycorticosterone and corticosterone concentrations were found in patients with secondary adrenal insufficiency on lower hydrocortisone dose, secondary adrenal insufficiency on higher hydrocortisone dose and primary adrenal insufficiency, respectively. Half of the primary adrenal insufficient patients showed evidence of residual endogenous cortisol or aldosterone synthesis, as determined by quantifiable 11‐deoxycortisol, 11‐deoxycorticosterone and corticosterone conce ntrations. In secondary adrenal insufficient patients with higher endogenous cortisol production, as indicated by 11‐deoxycortisol concentrations above the median, no increased cortisol exposure was observed both by plasma pharmacokinetic parameters and 24‐hour free cortisol excretion in urine. Conclusions Adrenal corticosteroid production is likely to continue during treatment in a considerable percentage of patients with both primary and secondary adrenal insufficiency. In patients with secondary adrenal insufficiency, this synthesis appears to be sensitive to the dose of hydrocortisone. However, the residual corticosteroid concentrations were quantitatively low and its clinical significance remains therefore to be determined.


| INTRODUC TI ON
Patients with primary adrenal insufficiency (PAI) due to autoimmune adrenalitis are thought to have no adrenal cortical function after long-standing disease as a consequence of progressive destruction of the adrenal gland. This notion comes from autopsy reports in which the adrenal glands were found to be small after long-standing disease, with a capsule that was thickened and fibrotic and a cortex that was completely destroyed, although in some cases a few small clusters of adrenocortical cells were found surrounded by lymphocytes. 1 The previous research has shown that the development of adrenal insufficiency progresses through different stages whereby adrenocortical autoantibodies correlate with the degree of adrenal dysfunction. 2,3 This suggests that adrenal insufficiency is always the inevitable end result, but interestingly some cases of (partial) recovery of adrenal function in a patient with autoimmune Addison's disease have been described. [4][5][6][7] Further, in a cross-sectional study by Smans and Zelissen performed among 27 patients with long-standing autoimmune Addison's disease 10 patients were found to have detectable cortisol at the end of synacthen testing. 8 Although none of them showed (partial) recovery of adrenal function, their findings clearly indicate that in one third of patients with long-standing disease, some residual adrenal function was still present. Even though very rare, testing for residual function may have far reaching consequences because Gan and co-workers found that the treatment with tetracosactide improved adrenal function in one patient. 9 Patients with secondary adrenal insufficiency (SAI) have an insufficient production of ACTH due to pituitary injury. The adrenal glands therefore lack the stimulus for adequate cortisol production and become atrophic. At the time of the diagnosis of both PAI and SAI, the early morning cortisol or the stimulated cortisol is very low, but often still measurable. The clinical significance of this residual production is unknown. It can be speculated to reduce the requirements of hormone replacement, affect quality of life, or influence hospital admission rates for Addisonian crisis and even mortality.
Once the diagnosis of PAI or SAI is established, patients are rapidly replaced to compensate for the loss of hormone production. Patients suffering from PAI and SAI require cortisol replacement, commonly given as hydrocortisone (HC) tablets. In addition, patients with PAI need fludrocortisone replacement as substitute for aldosterone.
These therapies are lifelong. Interestingly, some investigators found that recovery of the adrenal function was not uncommon in secondary adrenal insufficiency even after a median of nearly 2 years. 10 Few studies have addressed the residual endogenous cortisol production in long-standing adrenal insufficiency during substitution therapy. This is difficult because the chemical composition of hydrocortisone tablets and endogenous cortisol is identical, and therefore, usually the strategy is chosen to temporarily withdraw hydrocortisone substitution. Alternatively, this problem can be circumvented by using stable isotope dilution/mass spectrometry, but these techniques are difficult and only a few research centres are equipped to do such studies. 11,12 A candidate approach to determine residual endogenous cortisol or aldosterone production may be the measurement of precursors of these corticosteroids. To explore this option, we recently developed a LC-MS/MS assay capable of measuring not only cortisol and cortisone in one run but also several precursor corticosteroids including 11-deoxycortisol (11S), 11-deoxycorticosterone (11DOC) and corticosterone (B) concentrations. This creates the possibility to routinely map the underlying function of the adrenal cortex in patients on hydrocortisone substitution and to gain further insight in its clinical consequence. The aim of this study was to assess the applicability of these techniques in estimating the residual function of the adrenal cortex in patients with adrenal insufficiency.

| Study design
In this study, we compared patients with primary adrenal insufficiency and their controls from a cross-sectional, single-centre, casecontrol study with patients with secondary adrenal insufficiency from a randomized double-blind crossover study.

| Ethical considerations
In both studies, ethical approval was obtained from the local ethics committee. The study was approved by the University Medical Center Groningen, the Netherlands and Sahlgrenska University Hospital, Gothenburg, Sweden. Informed consent was obtained from all participants before any elements of the study were performed.
The study was conducted according to the latest Declaration of Helsinki.

| Patients with primary adrenal insufficiency and matched controls
Twenty patients and their controls were derived from a larger sample participating in a cross-sectional, single-centre, case-control study conducted at the Centre for Endocrinology and Metabolism, Sahlgrenska University Hospital, Gothenburg, Sweden. The study design and recruitment have been published previously. 13 In short, adult patients from western Sweden, diagnosed with primary adrenal insufficiency K E Y W O R D S 11-deoxycortisol, adrenal insufficiency, aldosterone, corticosterone, cortisol, hydrocortisone, precursors at the age of 18 years or older, were invited to participate. Matched control subjects from the Gothenburg region were recruited from a random population sample obtained from the Swedish Population Registry. The patients took their usual HC morning dose at home before arriving to the hospital. Fasting blood samples were collected in the morning (between 8:00 and 10:00 AM) after an overnight fast.

| Patients with secondary adrenal insufficiency
This randomized double-blind crossover study was performed in a university hospital in Groningen, The Netherlands, between May 2012 and June 2013. The study has been registered with ClinicalTrials.
gov with number NCT01546922. The study design and rationale have been published in detail previously. 14 In short, patients with established secondary adrenal insufficiency (based on internationally accepted criteria) were recruited from the endocrine outpatient clinic at the University Medical Center Groningen, The Netherlands.
The following inclusion criteria were applied: treatment for pituitary disorder (surgery and/or radiotherapy) at least 1 year before study entry, stable replacement therapy for other pituitary hormone deficiencies for at least 6 months (thyroid hormone deficiency, growth hormone deficiency, testosterone/estradiol deficiency, diabetes insipidus), age between 18 and 75 years and body weight between 50 and 100 kg. 14 Main exclusion criteria were drug abuse or dependence, shift work, previous Cushing's disease, a history of frequent episodes of hypocortisolism and use of antiepileptic drugs. 14 Patients treated with cortisone acetate were converted to treatment with hydrocortisone in a bioequivalent dose during a 4-week run-in period. The hydrocortisone dose during the run-in period averaged 25 mg per day (0.31 mg/kg body weight). After the runin phase, patients were randomized to either first receive a lower hydrocortisone dose for 10 weeks followed by a higher hydrocortisone dose for 10 weeks, or vice versa. During the lower dose period, patients received 0.2-0.3 mg per kg body weight per day (total daily doses ranged from 15 to 20 mg per day), taken before breakfast, before lunch and before dinner. During the higher dose period, patients received the double amount, 0.4-0.6 mg per kg body per day (total daily doses ranged from 30 to 40 mg per day), taken at the same time points of the day. In case of intercurrent illness, doubling or tripling of the dose was allowed for a maximum of 7 successive days.
Compliance with study medication was assessed as described. 14 After each treatment period, blood sample drawing 1 hour after ingestion of the morning dose of hydrocortisone was done and 24hour urine collection was performed. The initial study population existed 47 patients that completed both study periods. 14 For present analysis, samples of 19 patients were available.

| Exposure to and pharmacokinetics of hydrocortisone
In patients with secondary adrenal insufficiency, estimations of the individual pharmacokinetic parameters were calculated as described in detail elsewhere. 15 In short, the Kinpop module of MwPharm version 3.81 was used to calculate a one-compartment and two-compartment population model for plasma total cortisol. The models were calculated using an iterative two-stage It is not possible to determine (without the use of stable isotopes) whether cortisol in plasma is solely present as the consequence of hydrocortisone substitution or as a consequence of both residual endogenous cortisol production and substitution therapy. We therefore studied cortisol exposure and pharmacokinetics to estimate whether higher producers of 11-deoxycortisol (suggestive of higher residual endogenous cortisol production) contributed to more cortisol exposure or a longer cortisol half-life. Urinary free cortisol was also measured using isotope dilution LC-MS/MS, as previously described. 17

| Statistics
Data are presented as mean ± standard deviation or median (interquartile range) when appropriate. Normality of the data was checked using QQ-plots and histograms. Parametric statistics were used between groups for most of the variables. However, differences in the pharmacokinetic parameters for plasma total cortisol and urinary free cortisol between groups separated by the median of 11-deoxycortisol were assessed using the Mann-Whitney U test.
This test was separately used for both the lower and the higher dose of HC.

| Subject characteristics
Twenty patients with PAI with matched controls and 19 with SAI were compared. Clinical characteristics are shown in Table 1. Age and age at diagnosis was not different between groups. The group of PAI consisted of more women than the SAI group (75% vs 53%). Weight, BMI and waist were all significantly higher in the SAI patients. Patients with PAI used a HC morning dose and total daily dose that were intermediate between the lower and higher HC dose of SAI patients. When corrected for body weight, the higher dose of hydrocortisone in SAI was similar to the dose of the PAI patients.

| 11-deoxycortisol, corticosterone and 11-deoxycorticosterone
Quantifiable amounts of 11-deoxycortisol or corticosterone or 11deoxycorticosterone were present in 100% of controls, in 94.7% of patients with SAI and in 50% of patients with PAI. The individual percentages for each of these compounds for the study groups are given in Table 2

| Exposure to and pharmacokinetics of hydrocortisone in SAI patients
Patients with SAI who had 11-deoxycortisol concentrations below the median on either the lower dose or the higher dose of HC showed no differences in pharmacokinetic parameters of HC, especially not elimination half-life (t ½ ) and 24-hour cortisol exposure in plasma (AUC 24h). In addition, urinary 24-hour cortisol excretion was also not different in these patient groups (Table 3).   and specific utility superior in this context to the cumbersome immuno-assays and without disturbing cross-reactivity. 21,23 We developed a LC-MS/MS that combines a low LOQ for all measured steroids (in the low picomolar range) and short retention time (Table   S1), making it possible to measure multiple samples in a short period of time. Another strength of the present study is the pharmacokinetic analysis under two HC dose conditions. In addition, urinary free cortisol was taken into account. Thus, several markers of cortisol exposure were available to assess if residual endogenous cortisol production adds to total cortisol exposure in these patients.

| D ISCUSS I ON
Evidence of production of cortisol precursors in adrenal insufficiency may have interesting implications. For example, low concentrations of 11-deoxycortisol or corticosterone may affect quality of life or cognition. In agreement with this, Nixon and co-workers have shown that corticosterone rather than cortisol may be the primary glucocorticoid for the brain, as a consequence of tissue-specific expression of transmembrane exporters of corticosteroids. 24 In addition, Schweitzer found that 11-deoxycortisol to cortisol ratios were different in depressed patients when compared to normal volunteers. 25 The use of 11-deoxycortisol may be of potential interest in sev- and osteoporosis, must be performed to address this. Thirdly, cortisol is produced directly by the adrenals but also by regeneration of cortisone by 11-HSD1 activity. Differentiation between these two pathways is only possible today by enzymatic inhibition or by use of stable isotopes. 26 The use of 11-deoxycortisol may be of help here to assess the relative contribution of direct cortisol production by the adrenal glands.
Several shortcomings need to be addressed. First, this is a pilot study addressing feasibility and potential implications in a relatively We conclude that cortisol synthesis is ongoing in SAI patients based on concentrations of 11-deoxycortisol, albeit at very low levels because no increase total cortisol exposure was observed in SAI patients with higher 11-deoxycortisol concentrations. This seems plausible, but additional data validated against other methods in healthy volunteers and patients are needed to gain further insight in this. However, when looking at the absolute concentrations of 11-deoxycortisol, which appears to be <1/1000th of the normal cortisol concentrations at 9 am, one comes to similar conclusions.
In conclusion, adrenal corticosteroid production is likely to continue during treatment in a considerable percentage of patients with PAI and in almost all patients with SAI. In patient with SAI, the higher doses of HC suppressed 11-deoxycortisol concentration, suggesting feedback inhibition of endogenous cortisol production. Whether residual cortisol production in adrenal insufficiency has clinical consequences remains to be determined.

CO N FLI C T O F I NTE R E S T
The authors declare no conflicts of interest. Johannsson and André P van Beek: agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.