Complete Thesis

University of Groningen

Hydrocortisone dose in adrenal insufficiency Werumeus Buning, Jorien

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record

Publication date: 2017

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA): Werumeus Buning, J. (2017). Hydrocortisone dose in adrenal insufficiency: Balancing harms and benefits. Rijksuniversiteit Groningen.

Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

Download date: 26-09-2021 HYDROCORTISONE DOSE IN ADRENAL INSUFFICIENCY Balancing harms and benefits

Jorien Werumeus Buning Hydrocortisone dose in adrenal insufficiency Balancing harms and benefits

Thesis, University of Groningen, the Netherlands

Cover-design: Machteld Hardick – www.mhardick.nl Lay-out: Ridderprint BV – www.ridderprint.nl Printing: Ridderprint BV – www.ridderprint.nl

ISBN: 978-90-367-9699-6 (printed) 978-90-367-9698-9 (eBook)

Copyright © 2017 Jorien Werumeus Buning All rights reserved. No parts of this thesis may be reproduced, stored in retrieval system, or transmitted in any form or by any means without prior written permission of the author.

Financial support for the printing of this thesis is kindly provided by The Endocrinol- ogy Fund (as part of the Ubbo Emmius Fund), Goodlife Healthcare BV, Pfizer BV, Novartis Pharma BV, Graduate School of Medical Sciences/University Medical Center Groningen and University of Groningen. Hydrocortisone dose in adrenal insufficiency

Balancing harms and benefits

Proefschrift

ter verkrijging van de graad van doctor aan de Rijksuniversiteit Groningen op gezag van de rector magnificus prof. dr. E. Sterken en volgens besluit van het College voor Promoties.

De openbare verdediging zal plaatsvinden op

maandag 8 mei 2017 om 16.15 uur

door

Jorien Werumeus Buning

geboren op 29 december 1987 te Steenwijk, Nederland Promotores Prof. dr. B.H.R. Wolffenbuttel Prof. dr. O.M. Tucha

Copromotores Dr. A.P. van Beek Dr. J. Koerts

Beoordelingscommissie Prof. dr. A.R.M.M. Hermus Prof. dr. O.C. Meijer Prof. dr. B.R. Walker Paranimfen H.E. Koning R. Werumeus Buning

Contents

Chapter 1 Introduction and aims of the thesis 9

Part I The effect of hydrocortisone treatment on psychological outcome measures in patients with secondary adrenal insufficiency Chapter 2 The effect of two different doses of hydrocortisone on cognition 29 in patients with secondary adrenal insufficiency – results from a randomized controlled trial. Psychoneuroendocrinology 2015;55:36–47 Chapter 3 Hydrocortisone dose influences pain, depressive symptoms and 63 perceived health in adrenal insufficiency: a randomized controlled trial. Neuroendocrinology 2016;103(6):771–8 Chapter 4 The role of cortisol in the association between perceived stress and 87 pain: a short report on secondary adrenal insufficient patients. Journal for Person-Oriented Research 2016;2(3):135–141

Part II The effect of hydrocortisone treatment on somatic outcome measures in patients with secondary adrenal insufficiency Chapter 5 Somatosensory functioning in patients with secondary adrenal 101 insufficiency treated with two different doses of hydrocortisone – results from a randomized controlled trial. Submitted Chapter 6 Effects of hydrocortisone on the regulation of blood pressure – 115 results from an RCT. Journal of Clinical Endocrinology & Metabolism 2016;101(10):3691–3699 Chapter 7 Pharmacokinetics of hydrocortisone: results and implications from 141 a randomized controlled trial. Accepted in Metabolism Chapter 8 General discussion and future perspectives 165 Chapter 9 Summary 185 Samenvatting 193 Dankwoord 201 About the author 207 List of publications 211

Chapter 1

Introduction and aims of the thesis

Introduction 11

The hypothalamic-pituitary-adrenal axis 1 The hypothalamic-pituitary-adrenal (HPA) axis functions through a complex set of interactions between the hypothalamus, the pituitary and the adrenal glands. It is our central stress response system which is activated during physically or mentally challenging situations. Activation of the system causes the hypothalamus to secrete corticotrophin releasing hormone (CRH). When CRH binds to CRH receptors on the anterior pituitary gland, adrenocorticotropic hormone (ACTH) is released by the pituitary. ACTH stimulates the secretion of cortisol by the adrenal cortex. Cortisol, in turn, exerts inhibitory effects to the hypothalamus and the pituitary via a negative feedback loop (Figure 1). Cortisol secretion shows strong diurnal variation with the lowest levels at around midnight, an initiation of the rise at approximately 02.00–03.00 h and peak values in the morning directly after waking. Thereafter, levels slowly decrease throughout the day until the nadir at midnight. Besides being activated during stressful situations, the HPA axis is also vital for sup- porting normal physiological functioning. Cortisol serves several functions in the body like the regulation of blood glucose, suppression of the immune system and assistance in fat, protein and carbohydrate metabolism.

Adrenal insufficiency

Patients with adrenal insufficiency (AI) are characterized by the loss of endogenous cortisol production. This lack of cortisol production can be caused by loss of function of the adrenal gland itself, in which case it is called primary AI (PAI) (Figure 1). This form of AI is most frequently caused by autoimmune adrenalitis (Addison’s disease) or by a disturbed function of one of the enzymes involved in cortisol synthesis (congenital adrenal hyperplasia).1,2 AI can also be caused by impairment of the pituitary (secondary AI (SAI)) or hypothalamus (tertiary AI), resulting in a deficiency of ACTH or CRH, respectively, and subsequently a lack of stimulation of the adrenal cortex to produce cortisol. Pituitary tumors or treatment of these pituitary tumors, by means of pituitary surgery or radiotherapy, are the most frequent causes of SAI.3 Pituitary tumors can be classified as functioning or non-functioning, depending on their hormonal activity. Non-functioning adenomas have no clinical or biochemical features of excessive hormonal secretion, whereas functional pituitary adenomas are characterized by excessive pituitary hormonal secretion. Other pituitary region tumors, such as craniopharyngiomas, meningeomas, germinomas, intrasellar or suprasellar metastases as well as traumatic brain injury can also cause SAI.4 The first choice of treatment of pituitary (region) tumors is surgical resection of the tumor, usually via 12 Chapter 1

Figure 1. The hypothalamus-pituitary-adrenal axis during A) the normal physiological situation, B) with primary adrenal insufficiency, C) with secondary adrenal insufficiency and D) with tertiary adrenal insufficiency. Introduction 13 the transsphenoidal route. It is a minimally invasive technique in which the tumor is approached through the nose and sphenoid sinus. Less commonly, when the tumor 1 is large or is not accessible through the transsphenoidal route, transcranial surgery is performed, in which a section of the skull is removed in order to access the brain. In some cases additional radiotherapy is applied, for instance when the tumor is not fully resected, when the tumor is recurring or in case of secreting adenomas where hormonal control cannot be achieved after surgery and medical therapy. SAI is a rare disease, it has an estimated prevalence of 150–280 per million.3,5–7 Patients with untreated SAI present with weight loss, anorexia or loss of appetite, generalized fatigue, loss of energy, reduced muscle strength, and increased irritability.4 However, presenting signs and symptoms are often subtle and unspecific, impeding the diagnosis which often leads to a delay in diagnosis.

Diagnosis The diagnosis of SAI is based on the finding of low early morning cortisol levels. Furthermore, in case of indeterminate cortisol values, various stimulation tests are available to assess the integrity of the HPA axis. For instance, the ACTH stimulation test, also known as the short synacthen test, has been advocated by several investigators in cases with a suspected disease duration of at least one month. This test can be done at any time of the day. Serum cortisol levels are measured at baseline and 30 or 60 minutes after the administration of 250 µg ACTH. Cortisol responses below 500–550 nmol/L (specific cut-off values can vary, depending on the assay used for the measurement of cortisol) are indicative of AI. However, when performed within 4 weeks after pituitary surgery, the adrenal glands are not yet atrophied and thus may still respond to ACTH, wrongly indicating intact HPA axis. Therefore, the insulin tolerance test is the gold standard for the diagnosis of SAI. In this test, a hypoglycemic state is provoked by the administration of 0.1 units of insulin/ kg body weight. Hypoglycemia is a powerful stressor that stimulates the HPA axis.8 Cortisol concentrations in response to the insulin tolerance test below 500 nmol/L are considered to indicate AI. However, the test is cumbersome, expensive and contraindi- cated in patients with a history of seizures or cardiovascular diseases.2,9,10

Treatment of SAI

Glucocorticoid substitution treatment Until corticosteroid replacement became available, AI was a deadly disease. Edward Kendall was the first to isolate cortisosterone in 1936 and two years later Leonard Simpson first used synthetic deoxycorticosterone acetate in the treatment of Addison’s 14 Chapter 1

disease with success.11 However, it was only after the synthesis of cortisone that glucocorticoids (GCs) became widely available,12 for which Edward Kendall, Philip Hench and Tadeus Reichstein received the Nobel Prize in Physiology or Medicine in 1950. Patients with AI are treated with GCs, of which oral hydrocortisone (HC) is the most commonly used preparation. Traditionally it was recommended to administer two-thirds (20 mg) of the substitution dose in the morning and one-third (10 mg) in the evening. This dose was based on cortisol production rate estimates in healthy individuals of 12–15 mg/m²/day. However, using stable isotope tracers the daily cortisol production rate is currently estimated to be approximately 6–10 mg/m²/day,13,14 corresponding to total daily doses of 15–20 mg/day. Body weight was found to be an important factor in the clearance of HC and therefore a weight-adjusted morning dose of 0.12 mg/kg body weight is recommended by some investigators.15 Furthermore, thrice-daily dosing resulted in more stable and physiological cortisol levels throughout the day compared to twice-daily dosing.16 However, uniform guidelines are lacking, resulting in a wide variety of substitution regimens with different doses and number of daily doses used in clinical practice.17 Next to the lack of consensus about the dose and dose frequency, agreement about how to objectively monitor the adequacy of current GC substitution therapy is also absent. Blood sampling is informative only when knowing the time of HC administration and time of blood sampling. Some clinicians use cortisol day curves,18 but they are inconvenient for the patient and time-consuming. Moreover, day curves turned out to be unable to discriminate between well-substituted and under- or over-substituted patients.19 Urinary free cortisol levels have been used as an overall indicator of the adequacy of cortisol substitution therapy.18 However, urinary free cortisol levels are influenced by corticosteroid-binding-globulin (CBG) binding capacity. CBG saturates rapid after oral HC ingestion, at approximately total cortisol concentrations of 450–550 nmol/L,20,21 which leads to increases of cortisol in urine. As a result, normal ranges for healthy individuals cannot be used in the assessment of urinary cortisol excretion during GC substitution. Furthermore, daily fluctuations in cortisol are missed due to the nature of this measure.22 Salivary cortisol has been used as an alternative, non-invasive method for monitoring substitution therapy. It has several advantages compared to serum cortisol day curves, as it is inexpensive, easy to perform, and can be collected at home. However, correlations between salivary cortisol and serum cortisol vary.22–25 Furthermore, in remains unclear to what extext salivary cortisol reflects tissue cortisol levels.25–27 In practice, clinical assessment of symptoms potentially suggestive of over- or under-treatment is often used. Muscle and joint pain, reduced strength, nausea, fatigue and lack of energy are symptoms suggestive of under-replacement, whereas weight Introduction 15 gain, new onset abdominal obesity, sleeplessness, hypertension and diabetes may indicate over-replacement.19 Under-treatment bears the risk of insufficient cortisol 1 supply in the case of severe stress risking an adrenal crisis,4 whereas chronic exposure to high cortisol levels is associated with increased mortality,28,29 increased risk for cardiovascular diseases,30,31 osteoporosis,32 reduced quality of life (QoL) 33 and cognitive impairment.34

Hydrocortisone treatment and cognition The brain is a major target area for GCs.35 GCs can easily pass the blood-brain barrier and they exert their effect via corticosteroid receptors. There are two types of corticosteroid receptors: the mineralocorticoid receptor (MR) and the glucocorticoid receptor (GR). GRs are widely expressed throughout the brain with highest densities in regions such as the prefrontal cortex and limbic areas including the hippocampus, amygdala, thalamus and hypothalamus.36 Data on mapping of MRs in the brain is less consistent, but high MR density is found in the hippocampus.37 GCs bind to MRs with a 6–10 fold greater affinity than to GRs.38 This differential affinity results in differences in occupation of the two receptors under different conditions. At basal cortisol levels, predominantly MRs are occupied. GRs will be activated additionally to MRs only when cortisol levels are high, e.g. at the circadian peak or during stress.39 An association between GCs and cognitive functioning is well described. Both in animals and in humans, there is convincing evidence of an inverted “U”-shape relation between plasma cortisol levels and cognitive functioning.36,40 This means that very low and very high cortisol levels impair information processing and cognition.40,41 Particu- larly the hippocampus is a major target area for GCs, due to its high density of GRs and MRs. Hippocampal functioning is therefore likely to be affected by GC levels. It is generally accepted that the hippocampus plays an important role in reactivity to novel situations and is essential for learning and memory.42 In healthy individuals, several studies reported decreased memory performance in association with higher cortisol levels.34,43–51 The impairing effects of GC excess seem to be most pronounced in declarative memory as studies have shown that increases in endogenous cortisol levels due to stressful conditions induce impairment in declarative memory, but not in non-declarative memory.34,44 Another study showed that the administration of 25 mg of cortisone to healthy subjects before retention of a previously learned list of nouns impaired delayed recall of those words.43 Immediate recall and recognition were not impaired, suggesting that cortisol impaired retrieval specifically. In addition, declarative verbal memory performance was decreased after administration of HC doses resembling cortisol levels during major stress in healthy subject.52 The negative side effects of cortisol excess on cognition have also been shown in several patient groups. Patients that were exposed to excessive levels of cortisol, for 16 Chapter 1

instance due to Cushing’s syndrome, showed deficits in memory, visual and spatial information processing, reasoning, language performance, verbal learning and attention.53–56 On the other hand, patients with hypocortisolism also reported impairments in cognitive functioning.57–60 Compared to healthy matched controls, patients with Addison’s disease performed worse on memory and executive functioning tasks,57 selective attention,58 episodic memory and speed of processing,59 and verbal learning.60 Besides being present in the hippocampus, GRs are also widely expressed in the prefrontal cortex, a brain structure known to be involved in executive functioning. Executive functions are a set of cognitive processes, including attentional control, inhibitory control, working memory, cognitive flexibility/set-shifting, reasoning and planning. So far, only a few studies focused on the influence of cortisol on prefrontal cortex-dependent cognitive functioning and results are inconclusive. McCormick and colleagues61 demonstrated a sex-dependent relationship in healthy subjects between salivary cortisol levels and the number of perseverative errors, as a measure of cognitive flexibility and set-shifting. In this task, participants had to sort cards according to a certain classification rule (for instance color or shape) and feedback was provided after each trial whether the classification was right or wrong. After several trials, the rules of classification changed and participants had to shift to a new method of classification. Perseverative errors were the trials in which the participants gave the same response as in the previous trial even though they received feedback that the response was not correct. In women, higher cortisol levels were associated with more perseverative errors, and thus less cognitive flexibility and reduced set-shifting, whereas in men higher cortisol levels were associated with more cognitive flexibility. In contrast, Newcomer and colleagues52 were not able to demonstrate an effect of HC administration on the performance on another test of executive functioning (verbal fluency test) in healthy subjects. Besides being a key area for executive functioning, the prefrontal cortex is also known to be involved in social cognition and attention.62,63 Social cognition refers to how people process, store and apply information about other people and social situations. To our knowledge there are no studies examining the effect of cortisol on social cognition. With regard to the effects of cortisol on attention, results are also inconclusive. A study showed that, compared to a placebo, the administration of 80 mg prednisone/day for 5 days to healthy subjects, leading to increased levels of cortisol, led to less salient imprinting of meaningful stimuli and may impair selective attention.64 In contrast, another study did not find impairments in attentional functioning in patients with Addison’s disease,59 neither in healthy subjects after dexamethasone treatment48 or after HC treatment.52 In conclusion, these results suggest that cognitive functions that are mediated by the hippocampus and the prefrontal cortex are affected by cortisol levels. However, most Introduction 17 of the studies performed are cross-sectional studies and included healthy individuals. It is therefore likely that the intact negative feedback mechanism in healthy individuals 1 influenced the results. No studies have been performed so far assessing the effect of the administration of different doses of HC treatment on memory, executive functioning, attention and social cognition for a substantial period of time in patients treated for SAI.

Hydrocortisone treatment and quality of life Patients who receive GC substitution often report an impaired QoL compared to healthy controls,33,65,66 which results in a high number of patients being unable to work and receiving disablement pension.66 The reason for this impairment is likely to be multifactorial and, among others, inappropriate GC treatment might play a role. The pharmacokinetic properties of currently available GC treatment results in over- or under-substitution during certain periods of the day. Indeed, retrospective analyses revealed that there is a relationship between the daily GC dose and QoL.33,65,66 Hahner and colleagues66 showed a reduced subjective health status in patients with chronic AI. Patients with SAI exhibited a slightly more impaired subjective health status, particularly in the domains physical functioning and bodily pain, compared to patients with PAI, probably due to common concomitant endocrine disorders that accompany SAI. Higher doses were associated with a more profound impairment of subjective health status. Ragnarsson and colleagues33 confirmed these findings by showing that higher HC equivalent doses were associated with more severely impaired QoL in hypopituitary patients. However, there was no difference in QoL between ACTH sufficient and ACTH insufficient patients, indicating that other factors also influence QoL. Due to the cross-sectional nature of the above mentioned studies, it remains unclear whether the impaired QoL was a result of the higher GC dose taken, or that a higher dose was prescribed in order to improve the comprised QoL. To disentangle cause and effect, QoL in relation to GC dose was assessed within a few randomized controlled trials. Wichers and colleagues67 performed a small randomized double-blind crossover study and found no differences in QoL between the three doses (15, 20 or 30 mg HC/day) administered. Furthermore, in another randomized open label crossover study no difference in QoL scores between the three doses (15, 20 and 30 mg HC/day) was found; irrespective of the dose, patients reported significantly lower levels of energy compared to healthy controls.68 In another randomized double- blind crossover study patients showed improvements in physical functioning on the lowest HC dose administered (15 mg HC/day) compared to the other regimens (20 mg HC/day or 5 mg prednisone/day).69 However, even though QoL improved on the lowest dose regimen, it remained impaired compared to healthy controls. Furthermore, in 18 Chapter 1

accordance with the study by Ragnarsson,33 QoL did not differ from ACTH-sufficient patients. It can thus be concluded that GC substitution treatment appears to have an effect on QoL. However, in the abovementioned controlled trials overall sample sizes were small, treatment periods were short and in some studies timing of the dosing was changed together with the total daily dose, introducing an extra potential influencing factor. Furthermore, results are conflicting and the exact relationship between GC dose and QoL remains inconclusive.

Hydrocortisone treatment and cardiovascular risk factors Patients with SAI show an increased mortality rate, predominantly caused by cardiovascular and cerebrovascular diseases.7,28 Inadequate substitution therapy with GCs might contribute to this increased risk for cardiovascular diseases. In order to prevent low levels of cortisol as a consequence of the short half-life of currently available immediate release tablets, patients are administered doses resulting in supraphysiological cortisol concentrations after intake. Higher doses of GC substitution treatment are associated with increased severity of cardiovascular risk factors. A study in the Scot- tish population showed that patients receiving exogenous GCs of doses equivalent to more than 30 mg HC/day had increased rates of all cardiovascular diseases, including myocardial infarction, heart failure and cerebrovascular disease.30 Cardiovascular risk in patients exposed to low dose GCs was similar to the risk in patients not receiving exogenous GCs.30 Furthermore, a large Scandinavian study in 2424 patients with hypopituitarism also demonstrated that higher doses of HC are associated with increased cardiovascular risk.17 Hypopituitary patients requiring GC replacement therapy had higher waist circumference, hemoglobin A1c, total cholesterol and triglycerides than ACTH sufficient patients (i.e. those not requiring GC treatment). Only those receiving doses of 20 mg/day or more showed an unfavorable metabolic profile. Moreover, all new cases of diabetes, stroke and myocardial infarction occurred in the GC treated group.17 High blood pressure is known to be an independent risk factor for cardiovascular diseases and hypertension is responsible for approximately 50 % of deaths from stroke or cardiovascular disease.70 A few controlled studies have investigated the effect of GC dose on blood pressure. Dunne and colleagues71 studied whether a reduction of the GC replacement dose would improve cardiovascular parameters in 13 hypopituitary patients. A reduction of the dose from 30 mg HC/day to 15 mg HC/day did not alter blood pressure values. Furthermore, several other cardiovascular function parameters were comparable between patients and a matched control group. Another group studied whether an increase in GC dose would alter blood pressure in 17 SAI patients.72 After Introduction 19 increasing the HC equivalent dose from < 20 mg/day to 30 mg/day for 7 days, blood pressure did not change. 1 Thus, even though there is a clear relationship between GCs and cardiovascular risk factors, the exact effect of GC dose on risk factors and the mechanisms underlying this relationship have not yet been fully defined.

Pharmacokinetics of hydrocortisone The pharmacokinetic properties of oral HC make it difficult to adequately mimic the physiological circadian rhythm of cortisol secretion. Oral HC is well absorbed, bioavailability is reported to be 96 ± 20 %, indicating complete oral absorption.73 Peak levels in plasma are reached at approximately 1.2 h after ingestion and the terminal half-life is around 1.8 h.22,73 As mentioned before, a fixed dose of 20 mg in the morning and 10 mg in the evening was initially suggested as the standard substitution therapy. However, Mah and colleagues15 showed that body weight was an important factor in the clearance of HC. Therefore, a weight-adjusted morning dose of 0.12 mg HC/kg body weight was recommended as this reduced variability in the maximum cortisol concentration, reduced variability in area under the curve and reduced overexposure to less than 5 %.15 More than 90 % of circulating cortisol is bound, predominantly to CBG (approximately 70 %) with high affinity and low capacity and to a lesser extent to albumin (approximately 20 %) with low affinity and high capacity.74 This leaves approximately 2–12 % of free cortisol in the circulation, dependent on the total cortisol concentration. Free cortisol is considered the biologically active part of cortisol and can bind to MRs and GRs. CBG acts as a reservoir for circulating cortisol. Several drugs influence cortisol metabolism through altered activity of cytochrome P450 3A4 (CYP3A4). Drugs that induce CYP3A4 activity, such as barbiturates, carbamazepine, phenytoin, rifampicin, lead to increased metabolism and hence decreased cortisol levels, whereas CYP3A4 inhibitors such as arepitant, ketoconazole, ritonavir, and diltiazem reduce metabolism and thus increase cortisol levels.75,76 A high inter-individual variability in pharmacokinetic parameters is reported in patients receiving HC substitution therapy.22,77 Due to the pharmacokinetic properties of currently available immediate release tablets, patients are over- or under-treated during certain periods of the day.77 Insight into the effect of dose adjustments on pharmacokinetic parameters of HC could provide an indication for evaluation and adjustment of HC substitution therapy. 20 Chapter 1

Considerations and aims for this thesis

Due to the lack of high-quality data underlying current recommendations about treatment of SAI, a randomized controlled trial investigating the effects of HC substitution dose in patients with SAI was desirable. This study was designed in 2011–2012 and conducted in 2012–2013. Two different doses of HC within the physiological range and used in clinical practice were compared. Thrice daily, weight-adjusted dosing before food intake, with a morning dose of 0.12 mg HC/kg body weight (with corresponding total daily doses of 0.2–0.3 mg HC/kg body weight), was found by other investigators to reduce inter-patient variability in maximum cortisol concentrations and this scheme was chosen as base for HC substitution in our study.15 However, much higher doses are also used in clinical practice,17 and therefore we compared this dose of 0.2–0.3 mg HC/ kg body weight/day to the double amount of it (0.4–0.6 mg HC/kg body weight/day). To reduce inter-subject variability, a crossover design was applied in which patients were their own controls. In this study we focussed on the clinical outcome measures cognitive functioning, QoL, somatosensory functioning, blood pressure and regulating hormones, and pharmacokinetic parameters. As described above, all these variables are known to be influenced by cortisol levels and are important parameters in the quality of the substitution therapy. The aim of this thesis is to assess the pathophysiologic effects of these two doses of cortisol, and to add evidence for recommendations regarding GC substitution therapy in SAI. As described above, we initiated a randomized double-blind crossover study (the Supreme Cort study, Clinicaltrials.gov identifier: NCT01546922). This thesis describes the results of our study on the effects of two different physiological doses of HC with regard to psychological outcome measures (cognition and QoL) as well as several somatic outcome measures (somatosensory functioning, blood pressure and regulating hormones, and pharmacokinetic parameters).

Outline of this thesis

This thesis is divided into two parts. The first part focuses on psychological outcome measures, the second part describes somatic outcome measures. The first part consists of the Chapters 2, 3, and 4 and describes the effect of HC substitution dose on psychological measures. In Chapter 2 we aimed to evaluate whether a lower dose of HC would be beneficial for cognitive functioning. The cognitive domains memory, attention, executive functioning and social cognition were Introduction 21 studied. These domains rely on the integrity of brain structures known to be influenced by cortisol due to the high density of GRs in these areas. 1 In Chapter 3 we aimed to evaluate the effect of the two different doses of HC on several aspects of health-related QoL. To this end we used validated questionnaires assessing QoL at the end of each treatment period as well as the daily assessment of somatic complaints, depression, and anxiety by means of diaries. In Chapter 4 we aimed to assess the mediating role of cortisol levels in the relationship between stress and pain. An individual approach was used in the analysis of a relationship between perceived stress, as measured with an anxiety questionnaire, and pain, and the mediating role of low cortisol concentrations therein. The second part consists of Chapters 5, 6, and 7 and describes the effect of HC substitution dose on somatic outcomes. In Chapter 5 we aimed to investigate whether somatosensory functioning would be affected by treatment with two different doses of HC. Detection and pain thresholds were established using mechanical stimuli. In Chapter 6 we aimed at assessing the effect of HC dose on blood pressure and regulating hormones. High doses of GCs are associated with increased cardiovascular risk factors including blood pressure. However, the mechanisms underlying this relationship remain inconclusive. Elaborate laboratory measurements enabled us to also explore the underlying mechanisms. In Chapter 7 we aimed at parameterizing a pharmacokinetic population model of HC in patients with SAI. Furthermore, we compared pharmacokinetic properties of HC on the two doses of HC for plasma total cortisol, plasma free cortisol and salivary cortisol. Chapter 8 provides the general discussion of the main findings in this thesis and addresses future perspectives. In Chapter 9 a summary is given. 22 Chapter 1

References

1. Johannsson G, Falorni A, Skrtic S, et al. Adrenal insufficiency: review of clinical outcomes with current glucocorticoid replacement therapy. Clin Endocrinol (Oxf) 2015;82(1):2-11. 2. Bancos I, Hahner S, Tomlinson J, Arlt W. Diagnosis and management of adrenal insufficiency. Lancet Diabetes Endocrinol 2015;3(3):216-226. 3. Regal M, Paramo C, Sierra SM, Garcia-Mayor RV. Prevalence and incidence of hypopituitarism in an adult Caucasian population in northwestern Spain. Clin Endocrinol (Oxf) 2001;55(6):735-740. 4. Arlt W, Allolio B. Adrenal insufficiency. Lancet 2003;361(9372):1881-1893. 5. Bates AS, Van’t Hoff W, Jones PJ, Clayton RN. The effect of hypopituitarism on life expectancy. J Clin Endocrinol Metab 1996;81(3):1169-1172. 6. Nilsson B, Gustavasson-Kadaka E, Bengtsson BA, Jonsson B. Pituitary adenomas in Sweden between 1958 and 1991: incidence, survival, and mortality. J Clin Endocrinol Metab 2000;85(4):1420- 1425. 7. Tomlinson JW, Holden N, Hills RK, et al. Association between premature mortality and hypopituitarism. West Midlands Prospective Hypopituitary Study Group. Lancet 2001;357(9254):425-431. 8. Grinspoon SK, Biller BM. Clinical review 62: Laboratory assessment of adrenal insufficiency.J Clin Endocrinol Metab 1994;79(4):923-931. 9. Vance ML. Hypopituitarism. N Engl J Med 1994;330(23):1651-1662. 10. Oelkers W. Adrenal insufficiency. N Engl J Med 1996;335(16):1206-1212. 11. Simpson SL. The use of synthetic desoxycorticosterone acetate in Addison’s disease. Lancet 1938;232:557-558. 12. Benedek TG. History of the development of corticosteroid therapy. Clin Exp Rheumatol 2011;29(5 Suppl 68):S-5-12. 13. Esteban NV, Loughlin T, Yergey AL, et al. Daily cortisol production rate in man determined by stable isotope dilution/mass spectrometry. J Clin Endocrinol Metab 1991;72(1):39-45. 14. Kraan GP, Dullaart RP, Pratt JJ, Wolthers BG, Drayer NM, De Bruin R. The daily cortisol production reinvestigated in healthy men. The serum and urinary cortisol production rates are not significantly different. J Clin Endocrinol Metab 1998;83(4):1247-1252. 15. Mah PM, Jenkins RC, Rostami-Hodjegan A, et al. Weight-related dosing, timing and monitoring hydrocortisone replacement therapy in patients with adrenal insufficiency. Clin Endocrinol (Oxf) 2004;61(3):367-375. 16. Groves RW, Toms GC, Houghton BJ, Monson JP. Corticosteroid replacement therapy: twice or thrice daily? J R Soc Med 1988;81(9):514-516. 17. Filipsson H, Monson JP, Koltowska-Haggstrom M, Mattsson A, Johannsson G. The impact of glucocorticoid replacement regimens on metabolic outcome and comorbidity in hypopituitary patients. J Clin Endocrinol Metab 2006;91(10):3954-3961. 18. Howlett TA. An assessment of optimal hydrocortisone replacement therapy. Clin Endocrinol (Oxf) 1997;46(3):263-268. 19. Arlt W, Rosenthal C, Hahner S, Allolio B. Quality of glucocorticoid replacement in adrenal insufficiency: clinical assessment vs. timed serum cortisol measurements. Clin Endocrinol (Oxf) 2006;64(4):384-389. 20. Tunn S, Mollmann H, Barth J, Derendorf H, Krieg M. Simultaneous measurement of cortisol in serum and saliva after different forms of cortisol administration. Clin Chem 1992;38(8 Pt 1):1491- 1494. Introduction 23

21. Aardal E, Holm AC. Cortisol in saliva--reference ranges and relation to cortisol in serum. Eur J Clin Chem Clin Biochem 1995;33(12):927-932. 1 22. Thomson AH, Devers MC, Wallace AM, et al. Variability in hydrocortisone plasma and saliva pharmacokinetics following intravenous and oral administration to patients with adrenal insufficiency. Clin Endocrinol (Oxf) 2007;66(6):789-796. 23. Wong V, Yan T, Donald A, McLean M. Saliva and bloodspot cortisol: novel sampling methods to assess hydrocortisone replacement therapy in hypoadrenal patients. Clin Endocrinol (Oxf) 2004;61(1):131-137. 24. Jung C, Greco S, Nguyen HH, et al. Plasma, salivary and urinary cortisol levels following physiological and stress doses of hydrocortisone in normal volunteers. BMC Endocr Disord 2014;14:91- 6823-14-91. 25. Lovas K, Thorsen TE, Husebye ES. Saliva cortisol measurement: simple and reliable assessment of the glucocorticoid replacement therapy in Addison’s disease. J Endocrinol Invest 2006;29(8):727- 731. 26. Debono M, Harrison RF, Whitaker MJ, et al. Salivary Cortisone Reflects Cortisol Exposure Under Physiological Conditions and After Hydrocortisone. J Clin Endocrinol Metab 2016;101(4):1469- 1477. 27. Ross IL, Lacerda M, Pillay TS, et al. Salivary Cortisol and Cortisone do not Appear to be Useful Biomarkers for Monitoring Hydrocortisone Replacement in Addison’s Disease. Horm Metab Res 2016;48(12):814-821. 28. Rosen T, Bengtsson BA. Premature mortality due to cardiovascular disease in hypopituitarism. Lancet 1990;336(8710):285-288. 29. Sherlock M, Reulen RC, Alonso AA, et al. ACTH deficiency, higher doses of hydrocortisone replacement, and radiotherapy are independent predictors of mortality in patients with acromegaly. J Clin Endocrinol Metab 2009;94(11):4216-4223. 30. Wei L, MacDonald TM, Walker BR. Taking glucocorticoids by prescription is associated with subsequent cardiovascular disease. Ann Intern Med 2004;141(10):764-770. 31. Walker BR. Glucocorticoids and cardiovascular disease. Eur J Endocrinol 2007;157(5):545-559. 32. Zelissen PM. Effect of glucocorticoid replacement therapy on bone mineral density in patients with Addison disease. Ann Intern Med 1994;120(3):207-210. 33. Ragnarsson O, Mattsson AF, Monson JP, et al. The relationship between glucocorticoid replacement and quality of life in 2737 hypopituitary patients. Eur J Endocrinol 2014;171(5):571-579. 34. Lupien SJ, McEwen BS. The acute effects of corticosteroids on cognition: integration of animal and human model studies. Brain Res Brain Res Rev 1997;24(1):1-27. 35. McEwen BS, Weiss JM, Schwartz LS. Selective retention of corticosterone by limbic structures in rat brain. Nature 1968;220(5170):911-912. 36. Gallagher P, Reid KS, Ferrier IN. Neuropsychological functioning in health and mood disorder: Modulation by glucocorticoids and their receptors. Psychoneuroendocrinology 2009;34 (suppl 1):S196-207. 37. Ito T, Morita N, Nishi M, Kawata M. In vitro and in vivo immunocytochemistry for the distribution of mineralocorticoid receptor with the use of specific antibody. Neurosci Res 2000;37(3):173-182. 38. de Kloet ER, Oitzl MS, Joels M. Stress and cognition: are corticosteroids good or bad guys? Trends Neurosci 1999;22(10):422-426. 39. Reul JM, de Kloet ER. Two receptor systems for corticosterone in rat brain: microdistribution and differential occupation. Endocrinology 1985;117(6):2505-2511. 24 Chapter 1

40. Lupien SJ, Gaudreau S, Tchiteya BM, et al. Stress-induced declarative memory impairment in healthy elderly subjects: relationship to cortisol reactivity. J Clin Endocrinol Metab 1997;82(7):2070-2075. 41. Belanoff JK, Gross K, Yager A, Schatzberg AF. Corticosteroids and cognition. J Psychiatr Res 2001;35(3):127-145. 42. De Kloet ER, Vreugdenhil E, Oitzl MS, Joels M. Brain corticosteroid receptor balance in health and disease. Endocr Rev 1998;19(3):269-301. 43. de Quervain DJ, Roozendaal B, Nitsch RM, McGaugh JL, Hock C. Acute cortisone administration impairs retrieval of long-term declarative memory in humans. Nat Neurosci 2000;3(4):313-314. 44. Kirschbaum C, Wolf OT, May M, Wippich W, Hellhammer DH. Stress- and treatment-induced elevations of cortisol levels associated with impaired declarative memory in healthy adults. Life Sci 1996;58(17):1475-1483. 45. Lupien S, Lecours AR, Lussier I, Schwartz G, Nair NP, Meaney MJ. Basal cortisol levels and cognitive deficits in human aging. J Neurosci 1994;14(5 Pt 1):2893-2903. 46. Lupien SJ, Wilkinson CW, Briere S, Ng Ying Kin NM, Meaney MJ, Nair NP. Acute modulation of aged human memory by pharmacological manipulation of glucocorticoids. J Clin Endocrinol Metab 2002;87(8):3798-3807. 47. Lupien SJ, Fiocco A, Wan N, et al. Stress hormones and human memory function across the lifespan. Psychoneuroendocrinology 2005;30(3):225-242. 48. Newcomer JW, Craft S, Hershey T, Askins K, Bardgett ME. Glucocorticoid-induced impairment in declarative memory performance in adult humans. J Neurosci 1994;14(4):2047-2053. 49. Young AH, Sahakian BJ, Robbins TW, Cowen PJ. The effects of chronic administration of hydrocortisone on cognitive function in normal male volunteers. Psychopharmacology (Berl) 1999;145(3):260-266. 50. Domes G, Rothfischer J, Reichwald U, Hautzinger M. Inverted-U function between salivary cortisol and retrieval of verbal memory after hydrocortisone treatment. Behav Neurosci 2005;119(2):512- 517. 51. Maheu FS, Collicutt P, Kornik R, Moszkowski R, Lupien SJ. The perfect time to be stressed: a differential modulation of human memory by stress applied in the morning or in the afternoon. Prog Neuropsychopharmacol Biol Psychiatry 2005;29(8):1281-1288. 52. Newcomer JW, Selke G, Melson AK, et al. Decreased memory performance in healthy humans induced by stress-level cortisol treatment. Arch Gen Psychiatry 1999;56(6):527-533. 53. Mauri M, Sinforiani E, Bono G, et al. Memory impairment in Cushing’s disease. Acta Neurol Scand 1993;87(1):52-55. 54. Forget H, Lacroix A, Somma M, Cohen H. Cognitive decline in patients with Cushing’s syndrome. J Int Neuropsychol Soc 2000;6(1):20-29. 55. Starkman MN, Giordani B, Berent S, Schork MA, Schteingart DE. Elevated cortisol levels in Cush- ing’s disease are associated with cognitive decrements. Psychosom Med 2001;63(6):985-993. 56. Michaud K, Forget H, Cohen H. Chronic glucocorticoid hypersecretion in Cushing’s syndrome exacerbates cognitive aging. Brain Cogn 2009;71(1):1-8. 57. Tiemensma J, Andela CD, Biermasz NR, Romijn JA, Pereira AM. Mild cognitive deficits in patients with primary adrenal insufficiency. Psychoneuroendocrinology 2016;63:170-177. 58. Klement J, Hubold C, Hallschmid M, et al. Effects of glucose infusion on neuroendocrine and cognitive parameters in Addison disease. Metabolism 2009;58(12):1825-1831. 59. Henry M, Thomas KG, Ross IL. Episodic memory impairment in Addison’s disease: results from a telephonic cognitive assessment. Metab Brain Dis 2014;29(2):421-430. Introduction 25

60. Schultebraucks K, Wingenfeld K, Heimes J, Quinkler M, Otte C. Cognitive function in patients with primary adrenal insufficiency (Addison’s disease). Psychoneuroendocrinology 2015;55:1-7. 1 61. McCormick CM, Lewis E, Somley B, Kahan TA. Individual differences in cortisol levels and performance on a test of executive function in men and women. Physiol Behav 2007;91(1):87-94. 62. Van Overwalle F. Social cognition and the brain: a meta-analysis. Hum Brain Mapp 2009;30(3):829- 858. 63. Li C, Chen K, Han H, Chui D, Wu J. An FMRI study of the neural systems involved in visually cued auditory top-down spatial and temporal attention. PLoS One 2012;7(11):e49948. 64. Wolkowitz OM, Reus VI, Weingartner H, et al. Cognitive effects of corticosteroids. Am J Psychiatry 1990;147(10):1297-1303. 65. Bleicken B, Hahner S, Loeffler M, et al. Influence of hydrocortisone dosage scheme on health-related quality of life in patients with adrenal insufficiency. Clin Endocrinol (Oxf) 2010;72(3):297-304. 66. Hahner S, Loeffler M, Fassnacht M, et al. Impaired subjective health status in 256 patients with adrenal insufficiency on standard therapy based on cross-sectional analysis.J Clin Endocrinol Metab 2007;92(10):3912-3922. 67. Wichers M, Springer W, Bidlingmaier F, Klingmuller D. The influence of hydrocortisone substitution on the quality of life and parameters of bone metabolism in patients with secondary hypocortisolism. Clin Endocrinol (Oxf) 1999;50(6):759-765. 68. Behan LA, Rogers B, Hannon MJ, et al. Optimizing glucocorticoid replacement therapy in severely adrenocorticotropin-deficient hypopituitary male patients. Clin Endocrinol (Oxf) 2011;75(4):505- 513. 69. Benson S, Neumann P, Unger N, et al. Effects of standard glucocorticoid replacement therapies on subjective well-being: a randomized, double-blind, crossover study in patients with secondary adrenal insufficiency. Eur J Endocrinol 2012;167(5):679-685. 70. Laslett LJ, Alagona P,Jr, Clark BA,3rd, et al. The worldwide environment of cardiovascular disease: prevalence, diagnosis, therapy, and policy issues: a report from the American College of Cardiology. J Am Coll Cardiol 2012;60(25 Suppl):S1-49. 71. Dunne FP, Elliot P, Gammage MD, et al. Cardiovascular function and glucocorticoid replacement in patients with hypopituitarism. Clin Endocrinol (Oxf) 1995;43(5):623-629. 72. Petersons CJ, Mangelsdorf BL, Thompson CH, Burt MG. Acute effect of increasing glucocorticoid replacement dose on cardiovascular risk and insulin sensitivity in patients with adrenocorticotrophin deficiency. J Clin Endocrinol Metab 2014;99(6):2269-2276. 73. Derendorf H, Mollmann H, Barth J, Mollmann C, Tunn S, Krieg M. Pharmacokinetics and oral bioavailability of hydrocortisone. J Clin Pharmacol 1991;31(5):473-476. 74. Lewis JG, Bagley CJ, Elder PA, Bachmann AW, Torpy DJ. Plasma free cortisol fraction reflects levels of functioning corticosteroid-binding globulin. Clin Chim Acta 2005;359(1-2):189-194. 75. Czock D, Keller F, Rasche FM, Haussler U. Pharmacokinetics and pharmacodynamics of systemi- cally administered glucocorticoids. Clin Pharmacokinet 2005;44(1):61-98. 76. Flockhart DA. Drug Interactions: Cytochrome P450 Drug Interaction Table. Indiana University School of Medicine (2007) . 2007; Available at: http://medicine.iupui.edu/clinpharm/ddis/. Accessed 08/05, 2016. 77. Simon N, Castinetti F, Ouliac F, Lesavre N, Brue T, Oliver C. Pharmacokinetic evidence for suboptimal treatment of adrenal insufficiency with currently available hydrocortisone tablets. Clin Pharmacokinet 2010;49(7):455-463.

Part 1

The effect of hydrocortisone treatment on psychological outcome measures in patients with secondary adrenal insufficiency

Chapter 2

The eff ect of two diff erent doses of hydrocortisone on cognition in patients with secondary adrenal insuffi ciency – results from a randomized controlled trial

J Werumeus Buning* P Brummelman* J Koerts RPF Dullaart G van den Berg MM van der Klauw O Tucha BHR Wolff enbuttel AP van Beek

* Both authors contributed equally

Psychoneuroendocrinology 2015;55:36-47 30 Chapter 2

Abstract

Context A wide variety in hydrocortisone (HC) substitution dose-regimens are considered physiological for patients with secondary adrenal insufficiency (SAI). However, it is likely that cognition is negatively influenced by higher cortisol exposure to the brain.

Objective To examine the effects of a high physiological HC dose in comparison to a low physiological HC dose on cognition.

Design and Setting This study was a randomized double blind cross-over study at the University Medical Cen- ter Groningen. This study is registered with ClinicalTrials.gov, number NCT01546922.

Patients Forty-seven patients (29 males, 18 females; mean [SD] age, 51 [14] years, range 19–73) with SAI participated.

Intervention Patients randomly received first a low dose of HC (0.2–0.3 mg/kg body weight/day) during 10 weeks followed by a high dose (0.4–0.6 mg/kg body weight/day) for another 10 weeks, or vice versa. HC substitution was given in three divided doses with the highest dose in the morning.

Main outcome measures Cognitive performance (memory, attention, executive functioning and social cognition) of patients was measured at baseline and after each treatment period using a battery of 12 standardized cognitive tests.

Results The higher dose of HC resulted in significantly higher systemic cortisol exposure for example measured at one hour after first dose ingestion (mean [SD], low dose: 653 [281] nmol/L; high dose: 930 [148] nmol/L; P < 0.001). No differences in cognitive performance were found between the two dose regimens.

Conclusions No negative influence on memory, attention, executive functioning and social cognition was observed after 10 weeks of treatment with a higher physiological dose of HC in patients with SAI when compared to a lower dose. Hydrocortisone Treatment and Cognition 31

Introduction

Patients with adrenal insufficiency are treated with glucocorticoids (GCs) to compensate for the loss of endogenous cortisol production. Usually this is done by oral administration of hydrocortisone (HC) or cortisonacetate (CA) with the aim to mimic 2 the endogenous cortisol rhythm, with peak values in the early morning before waking and a nadir at bedtime. Body weight is an important determinator of the exposure of a dose measured using repeated pharmacokinetic sampling over a period of time.1 Experts therefore recommend a weight-adjusted HC dose of 0.12 mg/kg body weight for the morning dose.2 Furthermore, thrice-daily weight-adjusted administration mimics the day-curve of cortisol seen in healthy volunteers and is necessary because of the short half-life of HC.1 GC substitution therapy, although referred to as physiological, has its imperfections. Current GC dose-regimens inevitably result in over- or under-replacement during certain periods of the day which may result in poor quality of life.2 Furthermore, Filipsson and colleagues found that GC substitution, especially in higher physiological doses (> 20 mg/day), was associated with an unfavorable metabolic profile when compared with patients with normal adrenal function.3 Besides the above mentioned side effects, cognitive side effects are reported in healthy individuals and in patients treated with pharmacological doses of GC. There is convincing evidence of an inverted “U”-shape relation between plasma cortisol levels and cognitive performance both in animals and humans.4,5 In healthy volunteers, an association has been found between higher cortisol levels and decreased memory performance,5–12 and decreased executive functioning.13 Fur- thermore, these associations are supplemented by findings from randomized controlled trials with healthy subjects further emphasizing the negative effect of elevated cortisol levels on memory14,15 and executive functioning.16 These impairments could be explained by the fact that memory and executive functioning rely on brain structures that contain a high concentration of GC receptors, such as the hippocampus17 and the prefrontal cortex.18 Indeed, Lupien and colleagues showed that chronically elevated cortisol levels are associated with reduced hippocampal volume and impairments in learning and memory tasks which depend upon the integrity of the hippocampus.9 Furthermore, the prefrontal cortex is not only associated with executive functioning, but also with social cognition19 and attention.20 A study by Wolkowitz and colleagues showed that increased levels of cortisol may lead to less salient encoding of meaningful stimuli and may impair selective attention.21 However, later studies suggest that cortisol does not affect attention.14,15 Because attention is of great importance for almost all cognitive functions, it is important to further elucidate the effects of GCs on attention. 32 Chapter 2

There seems to be compelling evidence of an association between levels of cortisol and several cognitive functions. However, most of the studies examining the effect of cortisol on cognition have been performed in healthy subjects, in which intact feedback regulation is likely to affect results. Furthermore, none of the studies has focused on the influence of administration for a substantial period of time of HC substitution on cognitive performance. Therefore, the aim of this study was to examine whether a low physiological HC substitution dose is better for cognition compared to a high physiological HC substitution dose, both administered for a substantial period of time, in patients treated for secondary adrenal insufficiency (SAI) who are characterized by absence of feedback regulation of the HPA-axis.

Materials and Methods

Patients In this randomized double blind cross-over study, patients were recruited for participation at the endocrine outpatient clinic of the University Medical Center Groningen (UMCG), a tertiary referral center for pituitary surgery in the Netherlands. A total of 63 patients were included in this study, of whom 60 patients completed the run-in phase and the baseline measurement (mean (SD) age, 52 (13), range 19–73, 35 males, 25 females). Eligibility, inclusion and follow-up is shown in Figure 1. All patients had SAI for which they received GC substitution therapy. To avoid effects of switching to a different type of GCs, all patients on CA were converted to treatment with HC in a bioequivalent dose, (i.e. CA dose (in mg) times 0.8 when compared to HC dose (mg)) during a four week run-in phase. The diagnosis of SAI was based on internationally accepted biochemical criteria, principally early morning (0800 – 0900 h) serum cortisol measurements and, if necessary, an insulin tolerance test. Early morning cut-off cortisol levels for adrenal insufficiency in our center were validated for patients with hypothalamic-pituitary disorders as previously published.22 Thyroid hormone deficiency was based on a low serum free thyroxine concentration (< 11.0 pmol/l). Growth hormone (GH) deficiency was based on a low insulin-like growth factor 1 (IGF-1) Z-score (less than 2 SD) and/or an insufficient peak GH concentration (< 10 mU/l) in response to insulin-induced hypoglycaemia or a peak growth hormone < 18 mU/l during an arginine-GHRH test. Insufficiency of the pituitary-gonadal axis was defined in men as a testosterone concentration below 10 nmol/l, in premenopausal women (aged < 50 years) as loss of menses and in postmenopausal women (aged > 50 years) as LH and FSH concentrations below 15 mU/l. Diabetes insipidus was defined as the incapac- ity to properly concentrate urine (increased urine volume with low urine osmolality) in the face of a high plasma osmolality (and sodium) and/or current treatment with Hydrocortisone Treatment and Cognition 33

Figure 1. Eligibility, inclusion and follow-up of the patients. 34 Chapter 2

desmopressin. Biochemical control of adequacy of hormonal substitution treatment was judged by the physicians that were responsible for the care of the participating patients using free thyroxine, IGF-1 and testosterone levels where necessary. Other inclusion criteria were age 18–75 years, body weight of 50–100 kg at screening, time interval of at least one year between study entry and tumor treatment with surgery and/ or radiotherapy, and adequate replacement of all other pituitary hormone deficiencies for at least six months prior to entry of the study. Main exclusion criteria were inability of legal consent, documented major cognitive impairment (MMSE < 24),23 drug abuse or dependence, current psychiatric disorders, treatment for a malignancy, shift work, previous Cushing’s disease, hospital admission during the study, diabetes mellitus with medication known to be able to induce hypoglycemia (e.g. Sulfonylurea derivatives and insulin) and a history of frequent episodes of clinical hypocortisolism. The concomitant use of other corticosteroids and drugs known to interfere with HC metabolism, e.g. anti-epileptics, was not allowed either.2 All patients were tested in the period between May 2012 and June 2013. The baseline characteristics of the entire cohort of patients with SAI (n = 135) known in our Univer- sity Medical Center did not differ from the presented study population, confirming the representativeness of our study population (data not shown).

Ethics Statement The study protocol was approved by the local ethics committee at the UMCG, The Netherlands. Patients gave written informed consent before entering the study. This study is registered with ClinicalTrials.gov, number NCT01546922.

Intervention Patients were randomly assigned to either group 1 or group 2 by the research pharmacy with a block size of 4. Group 1 first received a physiological low dose of HC for 10 weeks, followed by a physiological high dose for another 10 weeks. Group 2 first received a physiological high dose for 10 weeks, followed by a physiological low dose (supplemental Fig. 1). Patients were treated with oral tablets containing either 5 mg HC (low dose) or 10 mg HC (high dose). Only the research pharmacy knew which dose was administered in each period. In the low dose condition, patients received a cumulative daily dose of 0.2–0.3 mg HC per kg body weight in three divided doses (before breakfast, before lunch, before diner). In the high dose condition, patients received the double amount, 0.4–0.6 mg HC per kg body weight. For the exact dosing schemes see Table 1. In cases of intercurrent illness or fever, patients were allowed to double or triple their HC dose. Because the study aimed to investigate two different dosing schemes, increasing the dose of HC was allowed for a maximum of 7 days (i.e. 10 % of the study time and of the cumulative HC dose) but not in the week preceding Hydrocortisone Treatment and Cognition 35

Table 1. Weight-adjusted dosages Low dose Weight (kg) Before breakfast Before lunch Before diner Total 50–74 7.5 5.0 2.5 15 75–84 10.0 5.0 2.5 17.5 2 85–100 10.0 7.5 2.5 20 High dose Weight (kg) Before breakfast Before lunch Before diner Total 50–74 15.0 10.0 5.0 30 75–84 20.0 10.0 5.0 35 85–100 20.0 15.0 5.0 40 Dose is given in mg; Total: cumulative daily dose.

the second and the third visit. Compliance with the study medication was assessed in several ways. Firstly, by patient reports in daily diaries: patients were instructed to daily report if they had forgotten and/or doubled their medication and if so, how many doses they had forgotten or doubled. Secondly, the tablets returned by the patients after each study period were counted. Lastly, cortisol concentrations in plasma between the two study periods were compared.

Laboratory measurements Serum cortisol was measured by a commercially available electrochemiluminescence immunoassay (ECLIA, Roche Modular Systems). Intra-assay coefficient of variations of serum cortisol were 1.5 %, 1.1 % and 0.9 % at cortisol concentrations of 69.5, 348 and 952 nmol/l. The functional sensitivity was < 8.5 nmol/l with this assay. Cortisol in urine was analyzed by automated online solid phase extraction in combination with liquid chromatography tandem mass spectrometry (XLC-MS/MS) essentially as described by Jones et al..24 The intra-assay coefficient of variation of urinary free cortisol was < 2.4 % at a concentration of 57.2 nmol/l, and the inter-assay coefficient of variation was < 7.8 % at a concentration of 54.3 nmol/l.

Cognitive tests A battery of 12 standardized cognitive tests covering 4 cognitive domains (memory, attention, executive functioning and social cognition) was used. To assess memory function, the following tests were used: the Rivermead Behavioural Memory Test (RBMT), the 15 Words Test, the Digit Span Test (forward), the Rey Complex Figure Test and the 15 Figures Test. To assess attention, the subtests Vigilance, Divided Attention, Visual Scanning, and Alertness of the Test of Attentional Performance (TAP) were used. The 36 Chapter 2

Verbal Fluency Tests, the Trail Making Test and the Digit Span Test (backwards) were used to evaluate executive functioning. Lastly, in order to assess social cognition the Reading the Mind in the Eyes Test was used. See supplemental materials and methods for a detailed description of the tests and reference data.

Questionnaire A common questionnaire on demographic and health-related data was used to assess educational level, social status, full-time/part-time employment, social security benefit and use of medication. Educational level was classified using a Dutch education system, comparable to the International Standard Classification of Education (ISCED).25 This scale ranges from 1 (elementary school not finished) to 7 (university level).

Procedure At the three visit days (baseline [M1], 10 weeks later after the first treatment period [M2] and 10 weeks later after de second treatment period [M3]), patients were instructed to take their morning dose of HC at 07.00 h. At 08.00 h blood samples were drawn in a fasted state after which patients received breakfast. Breakfast was followed by a physical examination that included measuring body weight, height, and blood pressure. Between 08.30 h and 09.00 h the cognitive test battery started and took approximately 3 h including a 15 min break. During the break caffeine and nicotine intake was allowed. After finishing the cognitive test battery, the second blood samples were drawn (approximately 5 h after the morning HC intake). In addition, at the end of visit M1 and M2, study medication for the following treatment period and urine collection material was dispensed. During the day preceding M2 and M3, patients were instructed to collect 24 hour urine. All testing and scoring of tests was performed by trained personnel under the supervision of two neuropsychologists (PB and JK).

Safety A total of 34 adverse events (AE) were reported in 29 of the 63 patients (46.0 %). Two serious adverse events (SAE) were reported. One occurred on the low dose (hospitalization for influenza A infection) and one on the high dose (hospitalization for minor stroke in the left cerebral hemisphere). Both patients prematurely terminated the study due to this SAE as hospitalization was a criterion for withdrawal. No deaths occurred during the study. A total of three patients withdrew from the study during the run-in phase, none of the withdrawals was related to HC (two withdrew their informed consent due to personal reasons, one stopped because of the inability to complete the test battery). A total of eight patients withdrew from the study while on a low dose, of which three withdrawals were possibly related to the HC dose (influenza A infection [n = 1], an inability to Hydrocortisone Treatment and Cognition 37 tolerate the dose [n = 1] and a Herpes Zoster Ophthalmicus infection [n = 1]). Other reasons for withdrawal on the low dose were protocol violation (n = 3, incompliance with study medication [patient took an incorrect amount of tablets], initiation of anti epileptics and the use of prednisolone), broken arm (n = 1) and one patient withdrew the informed consent due to personal reasons. Five patients withdrew while on a high 2 dose, of which one withdrawal was possibly related to the HC dose (unpleasant feelings). The other reasons for withdrawal were protocol violation (n = 3, incompliance with study medication (compliance was questioned by the investigator and suspicion was confirmed by the patient self and a relative), kenacort injection, changed job to shift worker) and a minor stroke in the left cerebral hemisphere (n = 1). The patients completing the study did not differ from those not completing the study with regard to age, sex, educational level, age at diagnosis, childhood or adult onset and body weight (data not shown). With regard to the use of stress related dose adjustments, a doubling of the dose was reported 150 times in patients on the low dose and 146 times in patients on the high dose.

Statistical analyses No relevant data can be inferred from literature to estimate a reasonable treatment effect of HC dose on cognition in secondary adrenal insufficiency. Because of the absence of relevant data from literature, we performed a power analysis. A study with 2 arms, each with 25 patients (total number of patients of 50) is able to detect an effect size of 0.4 (two-sided α = 0.05 and β = 0.80) in test results even when between test correlations are poor (0.50). An effect size of 0.4 was chosen because it was considered a relevant change with a small to medium size effect. To allow for a drop-out rate of ±20 % a total number of ±60 patients were needed. All statistical analyses were performed using Statistical Package for the Social Sciences (SPSS, Inc., Armonk, NY, USA), version 20. Demographic data are presented as median, interquartile ranges [IQR], frequencies or percentages. Cognitive performance data were presented as mean Z-score (SD). Higher Z-scores represent better cognitive performance compared to healthy subject of the same age, sex and educational level. Normality of data was analyzed using Q-Q plots. Since not all data were normally distributed, non-parametric tests for paired samples were used. To compare the cognitive performances at baseline of group 1 and group 2, the Mann-Whitney U-Test was used. The cognitive performance which was obtained while on a low dose of HC was compared to the performance on cognitive tests while on a high dose of HC by using the Wilcoxon Signed Rank Test. In addition, Cohen’s d effect sizes were calculated to give a measure of the magnitude of the difference. An effect size of d = 0.2 is considered a small effect, d = 0.5 a moderate effect and d = 0.8 a large effect.26 Furthermore, the performances of patients were compared to published normative datasets from healthy populations (see 38 Chapter 2

supplemental materials and methods for a more detailed description), and standard scores were derived for each measure. A cognitive impairment on a test was defined as a performance equivalent to the performance of the lowest 10 % of the reference samples.23 To determine whether patients were more often impaired while on a high dose compared to a low dose, cognitive tests measuring similar cognitive processes were combined and the number of impaired patients on the high dose were compared to the number of impaired patients on the low dose using the McNemar Test. The two-tailed alpha level of < 0.05 was considered statistically significant. In case of statistical differences between the cognitive test performances on both doses, a Bonferroni correction was performed to correct for multiple comparisons. Given 42 test scores were compared, a P-value of < 0.001 (0.05/42) was regarded significant.

Results

Study population A total of 63 patients with SAI were included in this study. Forty-seven patients completed the study (29 men and 18 women, mean [SD] age 51 [14] years, range 19–73). Patients’ characteristics are given in Table 2. Twenty-four patients received treatment for a pituitary adenoma (14 patients had a non-functioning macroadenoma, five patients had a macroprolactinoma and five patients had acromegaly). Three patients were treated for a cyst with pituitary localization, four were treated for a craniopharyn- gioma, five had a tumor distant from the pituitary gland (including one germinoma, one dysgerminoma, two meningeomas and one ENT tumor) and four had other acquired forms of ACTH deficiency (including traumatic brain injury). Furthermore, seven had a congenital form of ACTH deficiency (six combined, one isolated). Of the 32 patients who underwent surgery, 72 % patients had undergone transsphenoidal surgery and 28 % had undergone a craniotomy. The median [IQR] cumulative daily HC dose before the study was 25 [20; 30] mg. When corrected for kg body weight, the median [IQR] prestudy HC dose was 0.32 [0.25; 0.35] mg/kg body weight. Most patients (70 %) received twice daily dosing prior to randomization. Group 1 and group 2 did not differ on clinical characteristics, except for the average time since radiotherapy in years (median [IQR], group 1, 20 [17; 31]; group 2, 10 [6; 18]; P = 0.045) and the number of daily doses of HC ([1/2/3], n; group 1, 2/11/9; group 2, 1/22/2; P = 0.016) (Table 2).

Serum and urinary cortisol The administration of the morning dose of HC in the high dose condition (0.24 mg HC/ kg body weight) resulted in significantly higher serum cortisol levels when compared to in the low dose condition (0.12 mg HC/kg body weight), both 1 hour and 5 hours Hydrocortisone Treatment and Cognition 39 0.502 0.074 0.474 0.370 0.130 0.215 0.186 0.014 0.575 0.730 0.045 0.647 0.966 0.085 0.171 0.316 0.297 0.052 0.213 0.447 P-value* 2 25)

5 = 19 13

16/9 2/23 74/26 1/22/2 85/15/0 (N Group 2 Group 10 [4; 12] 10 [3; 14] 10 [6; 18] 44 [20; 51] 30 [20; 30] 54 [40; 62] 49 [24; 57] 34 [21; 48] 0/1/1/1/13/8/1 0.34 [0.25; 0.37] 25.9 [23.2; 30.3] 85.4 [73.1; 93.2] 47)

N 22)

6 0 = 13

13/9 4/18 69/31 2/11/9 83/0/17 (N Group 1 Group 18 [8; 29] 14 [7; 25] 38 [30; 42] 23 [20; 30] 20 [17; 31] 55 [47; 61] 32 [25; 40] 31 [19; 42] 0/1/0/5/8/7/1 0.29 [0.25; 0.34] 26.8 [24.7; 29.1] 80.8 [71.2; 93.1] 47)

5 = 32 19

6/41 ary adrenal insufficiency ( ary adrenal 72/28 29/18 3/33/11 84/11/5 (N 11 [6; 20] 11 12 [5; 22] 12 [9; 22] 39 [28; 50] 25 [20; 30] 55 [43; 61] 43 [25; 52] 31 [20; 46] 0/2/1/6/21/15/2 0.32 [0.25; 0.35] 26.6 [24.5; 29.4] 82.5 [72.2; 93.0] Total patient group Total n n a n n surgery, surgery, nd Clinical characteristics of pituitary patients treated for second

Average time since surgery (y), median [IQR] time since surgery Average Patients with 2 Age at surgery (y), median [IQR] Age at surgery Number of daily dosings (1/2/3), Transsphenoïdal surgery/Craniotomy (%) surgery/Craniotomy Transsphenoïdal Total dose/kg body weight/day (mg/kg/day), median Total [IQR] Duration of hydrocortisone treatment (y), median [IQR] Total daily dose (mg/day), median [IQR] Total Pituitary radiotherapy/cranial irradiation/radiotherapy for extracranial tumors (%) time since radiotherapy (y), median [IQR] Average Age at radiotherapy (y), median [IQR] Surgery Educational level (1/2/3/4/5/6/7) Hydrocortisone treatment prior to randomization BMI (kg/m²), median [IQR] Sex (males/females), Childhood onset/Adult onset of SAI, Body weight (kg), median [IQR] Age (y), median [IQR] Radiotherapy Age at diagnosis (y), median [IQR] Table 2. Table 40 Chapter 2 This 25 0.119 0.368 1.000 0.302 1.000 0.920 1.000 0.492 P-value* 25)

= 24 60 56 68 35 44 92

NA (N Group 2 Group 1/5/11/5/3 47) (continued)

N 22)

= 14 33 41 46 28 46 91

NA (N Group 1 Group 2/4/10/6/0 47)

= 19 50 79 57 21 45 92

NA ary adrenal insufficiency ( (N 3/9/21/11/3 Total patient group Total 10: estrogens

8: estrogens (% of patients

n n Clinical characteristics of pituitary patients treated for second

Postmenopausal women, Premenopausal women, substituted) Men: testosterone (% of patients substituted) Growth hormone (% of patients unsubstituted) -value of group 1 versus 2. Desmopressin (% of patients substituted) Sex hormone (% of patients substituted) Growth hormone (% of patients substituted) Thyroid hormone (% of patients substituted) P No. of hormonal replacements (1/2/3/4/5) Educational level was classified using a Dutch education system, comparable to the International Standard Classification of Education (ISCED). scale ranges from 1 (elementary school not finished) to 7 (university level). IQR: Interquartile range. SAI: Secondary NA: adrenal Not insufficiency. applicable. * Table 2. Table a Hydrocortisone Treatment and Cognition 41

Table 3. Serum cortisol and 24 h urinary free cortisol levels Low dose High dose (N = 47) (N = 47) P-value* Serum cortisol levels 1 hour after ingestion, nmol/L 653 (281) 930 (148) < 0.001 2 5 hours after ingestion, nmol/L 208 (157) 312 (130) < 0.001 Urinary free cortisol levels (N = 46) (N = 46) 24 hour urinary cortisol, nmol/24 hr 88 (63) 339 (240)a < 0.001 Data is given as mean (SD); N = 46 for urinary cortisol because of incontinence of one patient; aN = 45 due to missing data; * Low dose versus High dose by Wilcoxon. after ingestion (Table 3). Furthermore, when comparing 24 hour urinary free cortisol levels, a high dose of HC resulted in significantly higher free cortisol levels compared to the low dose of HC (Table 3).

Cognitive tests At baseline, group 1 (first receiving the low dose followed by the high dose) and group 2 (first receiving the high dose followed by the low dose) did not differ with regard to their performances on tests of cognition (supplemental Table 1). Cognitive performances of the 47 patients who completed the study are given in Table 4. The number of patients who showed an impaired performance on the cognitive tests are given in Table 5.

Memory performance No significant differences between the doses were found for all memory measures (Table 4). Furthermore, there were no differences between both dose regimen in the number of patients scoring in the impaired range (Table 5).

Attention performance No differences were found between the doses for neither the measures of attention (Table 4) nor the number of impaired patients (Table 5).

Executive functioning Patients on a high dose did not differ from patients on a low dose on any of the measures of executive functioning (Table 4 and 5). 42 Chapter 2 0.21 0.14 0.07 0.01 0.02 0.03 0.07 0.02 0.26 0.20 0.14 0.24 -0.10 -0.38 -0.07 -0.19 Effect size c c b b Low dose High dose – 0.00 [0.20; 0.20] 0.00 [0.20; 0.40] 0.00 [0.60; 0.70] 0.00 [0.00; 0.36] 0.00 [0.00; 0.53] 0.20 [0.00; 0.40] 0.00 [0.00; 0.00] 0.04 [-0.69; 0.77] 0.00 [-0.30; 0.40] 0.22 [-0.44; 0.87] 0.10 [-0.10; 0.50] 0.10 [-0.10; 0.40] -0.64 [-1.14; 0.00] -0.98 [-3.38; 1.42] -0.30 [-0.50; 0.20] -7.80 [-35.85; 20.24] a 0.111 0.451 0.385 0.263 0.606 0.861 0.668 0.713 0.614 0.501 0.227 0.130 0.106 0.350 0.093 0.028 P-value* c b b 47)

(N High dose 0.10 (1.33) 0.89 (1.34) 0.56 (0.97) 0.25 (0.92) 0.38 (1.19) 0.23 (0.97) 0.08 (0.71) 1.28 (1.29) 1.40 (1.29) -0.03 (1.03) -0.42 (1.33) -0.16 (0.99) 10.23 (3.80) 28.43 (2.14) 39.83 (14.18) 612.54 (117.88) c 47)

(N Low dose 1.11 (1.22) 1.11 0.07 (0.98) 0.85 (1.09) 0.36 (1.01) 0.49 (0.97) 1.14 (1.26) 0.43 (1.02) 0.08 (1.05) 0.02 (1.04) -0.11 (0.88) -0.11 -0.41 (0.95) -0.30 (0.99) 28.36 (2.52) 10.02 (3.47) 40.72 (14.64) 623.21 (85.65) d,f d d,f d d Cognitive performances of pituitary patients treated for secondary adrenal insufficiency in the low and high dose condition adrenal Cognitive performances of pituitary patients treated for secondary

d Delayed recall Recognition Immediate memory Delayed memory Delayed corrected for immediate memory Short-term memory Short-term memory Immediate memory memory Total Reaction time of correct responses Total immediate memory Total Delayed memory Learning score Delayed memory Delayed corrected for total memory Recognition 15 Figures Test (raw data) Test 15 Figures 15 Words test Words 15 Rey Complex Figure Memory RBMT Attention (raw data) Vigilance Digit Span forward Table 4. Table Hydrocortisone Treatment and Cognition 43 0.11 0.13 0.04 0.18 0.03 0.01 0.06 0.09 0.23 0.10 0.08 0.14 0.05 -0.11 -0.35 -0.08 -0.13 -0.09 -0.08 Effect size 2 c c c Low dose High dose – 0.00 [0.00; 0.00] 0.00 [0.00; 0.40] 0.10 [-0.11; 0.31] 0.10 [-0.11; 0.00 [-0.28; 0.30] 0.00 [-0.27; 0.31] 0.00 [-0.30; 0.60] 0.00 [-0.20; 0.58] 0.00 [-0.40; 0.37] 0.08 [-0.17; 0.31] 0.00 [-0.51; 0.00] 0.28 [-0.18; 0.75] 0.15 [-0.39; 0.69] -0.11 [-0.30; 0.20] -0.11 -0.38 [-0.51; 0.00] -0.09 [-0.50; 0.60] -0.08 [-0.27; 0.00] -0.10 [-0.20; 0.40] -0.17 [-0.28; 0.00] 8.54 [-13.08; 30.17] a 0.331 0.732 0.291 0.905 0.857 0.713 0.192 0.536 0.502 0.536 0.748 0.572 0.277 0.229 0.164 0.531 0.329 0.681 0.015 P-value* c c c 47)

(N High dose 0.09 (1.18) 0.03 (0.94) 0.21 (1.31) 0.07 (1.00) 0.08 (1.08) 0.34 (0.76) 0.98 (2.47) 1.91 (1.56) -0.89 (0.82) -0.06 (1.00) -0.36 (1.03) -0.99 (0.73) -0.57 (0.92) -0.23 (1.04) -1.10 (0.66) -0.42 (0.79) -0.14 (0.93) -0.70 (0.28) 130.39 (63.57) 47)

(N Low dose 0.72 (2.41) 1.81 (1.28) 0.02 (1.24) 0.03 (1.03) 0.32 (1.48) 0.30 (1.01) -0.87 (0.81) -0.38 (1.04) -0.90 (0.73) -0.06 (1.00) -0.48 (0.95) -0.33 (0.92) -0.14 (0.88) -0.03 (1.06) -1.03 (0.67) -0.14 (0.82) -0.02 (0.86) -0.72 (0.26) 122.02 (62.02) d e d e Cognitive performances of pituitary patients treated for secondary adrenal insufficiency in the low and high dose condition (continued) adrenal Cognitive performances of pituitary patients treated for secondary

Number of omission errors Number of commission errors Reaction time auditory responses stimuli Reaction time for target alertness reaction time Tonic Semantic fluency Variability of reaction time Variability Variability of reaction time auditory responses Variability Reaction time visual responses stimuli of reaction time for target Variability alertness variability in reaction time Tonic Phonemic fluency Variability of reaction time visual responses Variability Number of omission errors Phasic alertness reaction time Phasic alertness variability in reaction time Number of omission errors Number of commission errors Number of commission errors Executive functioning Fluency test Divided attention Alertness Visual scanning Visual Table 4. Table 44 Chapter 2 0.07 0.04 0.05 -0.02 -0.01 Effect size Low dose High dose – 0.00 [-0.31; 0.44] 0.00 [-0.29; 0.29] 0.10 [-0.20; 0.40] -0.30 [-1.30; 1.10] -0.60 [-1.40; 1.30] : Rivermead Behavioral Memory Test. : Rivermead Behavioral Memory 0.632 0.819 0.512 0.861 0.958 1.000 0.403 P-value* percentile) (See supplemental materials and methods). th 10

47)

43/4 (N 30/9/6/1/1 High dose 0.18 (1.31) 0.36 (1.48) -0.01 (1.02) -0.71 (1.15) -0.26 (1.06) ) due to lacking reference data. d copy drawing, < 47)

43/4 (N Low dose 34/7/4/1/1 0.05 (0.86) 0.16 (1.46) 0.37 (1.42) -0.67 (1.20) -0.32 (1.14) d,f d g g e Cognitive performances of pituitary patients treated for secondary adrenal insufficiency in the low and high dose condition (continued) adrenal Cognitive performances of pituitary patients treated for secondary

45 (2 patients were excluded from analysis because of an impaire 46 due to a missing value. 0.001 for period effect.

Working memory Working A, time Condition Reading the Mind in the Eyes Test Reading the Mind in Eyes Condition A, no. of corrections (0/1) Condition Condition B, no. of corrections (0/1/2/3) Condition B, time Condition B/A = = <

-value not statistically significant after Bonferonni correction. P Trail Making Test Making Trail Social cognition Digit Span backwards N test. Verbal P N Non-verbal test. Cognitive performances on these tests are given as raw data mean (SD a b c d e f g Table 4. Table Wilcoxon. * LD versus HD by Cognitive performance data are given as mean Z-scores (SD) or [95 % confidence interval]; RBMT Hydrocortisone Treatment and Cognition 45

Table 5. Comparison of impaired scores between the low dose and the high dose of hydrocortisone Low dose High dose (N = 47) (N = 47) P-value* Memory 2 Immediate memory 13 15 0.727 Short-term memory 3 4 1.000 Delayed memory 8 8 1.000 Recognition 6 3 0.453 Attention Divided attention errors 6 7 1.000 Visual scanning errors 5 5 1.000 Executive functioning Fluency 10 10 1.000 Working memory 3 4 1.000 Cognitive flexibility 6 6 1.000 Social cognition 18 11 0.065 Psychomotor speed 17 24 0.092 Data is the number of patients showing an impaired score. Immediate memory: RBMT immediate memory + 15 Words Test total immediate memory + Rey’s Complex Figure immediate memory; short-term memory: 15 Words Test short-term memory + Digit Span forward; delayed memory: RBMT delayed memory + 15 Words Test delayed memory + Rey’s Complex Figure delayed memory; recognition: 15 Words Test recognition; divided attention errors: omission errors + commission errors; visual scanning errors: omission errors + commission errors; fluency: semantic fluency + phonemic fluency; working memory: Digit Span backwards; cognitive flexibility: Trail Making Test Condition B/A; social cognition: RMET; psychomotor speed: tonic alertness reaction time + phasic alertness reaction time + Trail Making Test Condition A. * LD versus HD by McNemar Test.

Social cognition With regard to social cognition no differences were found between the high and the low dose conditions (Table 4). Furthermore, there were no differences between the doses in the number of patients showing an impaired performance (Table 5).

Discussion

This is the first randomized controlled double blind cross-over study demonstrating that there are no differences in the cognitive domains of memory, attention, executive functions and social cognition in patients with SAI treated with a physiological high dose of HC when compared to a physiological low dose of HC. 46 Chapter 2

Although current substitution therapies for adrenal insufficiency are life-saving, optimization of treatment remains a challenge. Objective biochemical parameters to control dosage are lacking and the individual total dose is adjusted based on subjective parameters such as fatigue or quality of life.27 A low dose of HC is recommended3, but the clinical experience is that 40 % of the patients receive a higher dose (> 30 mg HC equivalents/day).3 The integrity of cognition is important for everyday functioning. Several cognitive functions are a basic requirement for participation in society, for example by maintaining a job and having a social life. Cognitive impairments can have an impact on these daily activities. The relationship between GC levels and cognitive performance has been studied in healthy subjects. However, no studies in patients with adrenal insufficiency were performed examining cognitive functioning in relation to HC dose. The weight related dosing resulted in serum cortisol levels comparable with cortisol levels found by Mah et al..1 The administration of a low dose of HC (0.12 mg HC/kg body weight for the morning dose) resulted in average serum cortisol levels of 646 ± 243 nmol/L 1 h after ingestion. In comparison, Mah found serum cortisol concentrations of 600–700 nmol/L 1 h after a single dose of 0.12 mg/kg body weight.1 No significant differences were observed between the weight groups (50–74 kg, 75–84 kg, 85–100 kg) on the low dose and the high dose, both 1 h after ingestion and 5 h after ingestion (data not shown). The results of the current study thus confirm the reproducibility of weight related dosing. Furthermore, the higher dose of HC resulted in approximately one and a half times higher cortisol concentrations than the lower dose of HC. The current study shows that there is no major negative effect of a higher dose of HC on memory performance. This finding is in accordance with a study by Newcomer and colleagues (1999) who did not find an effect of 4 days of treatment of 40 mg HC/day on memory performance in healthy subjects.14 Furthermore, our results are in agreement with previous studies indicating that cortisol does not affect memory28 and attention,14,15 and executive functioning.14 However, our findings are in contrast to studies in healthy volunteers showing an effect of higher levels of cortisol on memory performance5–12,14,15 and executive functioning.12 Considering social cognition, our findings are in contrast to a study showing a strong negative association between cortisol and social cognition in women.29 Clear differences in study design and population may underlie this observed discrepancy as different doses and routes of administration were used. In addition, healthy volunteers have a circadian rhythm which is obviously lacking in SAI patients. Without a Bonferroni correction for multiple testing, subtle differences were found between the low and high dose of HC regarding short-term memory and the variability of reaction time during the phasic alertness task, with patients on the high dose performing worse compared to patients on the low dose. This is in accordance with a Hydrocortisone Treatment and Cognition 47 study performed by Harbeck et al., who found a negative correlation between cortisol levels and short-term memory in patients with primary adrenal insufficiency or SAI.30 The number of AE’s, SAE’s, and withdrawals were comparable on both dose regimens and between group 1 and group 2, indicating that there was no pattern in AE’s and SAE’s favoring one group or dose over the other. The AE’s and SAE’s reported 2 can be considered a reflection of the events seen in daily practice in patients with SAI including more hospital admissions. Thus, in this view our patient sample can also be seen representative of the average population of patients with SAI. Although this randomized controlled trial was performed according to the highest standards, some limitations need to be addressed. First, ten patients (21 %) had IGF-1 Z-scores <-2 but did not receive GH substitution, indicating unsubstituted severe GH deficiency. This is usually the consequence of an absence of GH deficiency related symptoms or lack of patient benefit of previous GH treatment. Untreated GH deficiency may have influenced our results31. Even though all patients in this study received adequate replacement for all other pituitary hormone deficiencies prior to entry of the study and these substitution therapies remained unchanged during the study, other influences or interactions are possible and these might theoretically have also influenced our findings. Furthermore, some patients (40 %) were previously treated with radiotherapy and some had a history of GH excess (11 %). Previously, we have shown that neither relates to impaired cognitive functioning.32,33 It was our goal to perform a study in a population representative of all patients with SAI visiting the endocrine outpatient clinic. In such a population, multiple covariates are present. Due to the cross-over study design, the influence of confounding covariates was reduced because each cross-over patient served as his or her own control. Secondly, learning effects were found for 2 tests (Trail Making Test condition B and Trail Making Test condition B/A). Thus a remarkably low period (learning) effect was found in our study design in which we used parallel version of most of the tests. A possible explanation for this learning effect of the Trail Making Test condition B and condition B/A was that patients developed a strategy after multiple testing, which may lead to the ameliorated performance. We also need to consider the possible influence of carry-over effects. For pharmacokinetic reasons, one would not expect this as it is found that blood concentrations of GCs reach peak levels after 15–30 min after the stressor, and decline slowly to pre-stress levels 60–90 min later.34 Accordingly, New- comer and colleagues showed that a difference on verbal memory performance between a higher dose of HC compared to a lower dose of HC and placebo were restored after a 6-day washout period.14 However, cortisol is also known to induce genomic and structural changes in the limbic network.35 Because of the vital function of cortisol, it is not possible to implement a wash-out period. Thus, even though we have corrected the best way we could for possible influences by our cross-over design, we cannot fully exclude learning and/or carry-over effects. Reassuring is that we did not find statistical evidence 48 Chapter 2

for a carry-over effect. Thirdly, a heterogeneous patient group with regard to diagnosis and treatment was studied. However, we do believe that this is a realistic representation of the diagnoses seen in clinical practice. Therefore, our patient sample is representative of the patient group seen in practice. Lastly, one might argue that ten weeks is too short to find subtle differences. We believe however, that ten weeks is a reasonable time period to examine the effects of different doses of HC on cognition because some studies have found an effect of cortisol on cognition as early as after only 1 dose of GCs. Furthermore, other studies used similar time frames to assess clinically relevant endpoints.36–38 This study showed no differences in cognitive functioning between both doses in resting conditions. However, it has been found that glucocorticoids enhance memory consolidation of emotionally arousing experiences, but impair memory retrieval and working memory during emotionally arousing test situations.39 It is therefore possible that emotional stress induction yields different effects on cognition in our patients. In conclusion, we found no differences in cognition in patients with SAI on a physiological high dose of HC for 10 weeks when compared to a physiological low dose of HC. Therefore, both doses can be considered safe to administer with regard to cognitive functioning. However, opposite effects of higher doses of GCs on, for example, metabolic profile,3 bone density,40 and quality of life37,41 need to be taken into consideration.

Acknowledgement

We would like to thank Dr. W. J. Sluiter for his help with the statistical analyses. Hydrocortisone Treatment and Cognition 49

References

1. Mah PM, Jenkins RC, Rostami-Hodjegan A, et al. Weight-related dosing, timing and monitoring hydrocortisone replacement therapy in patients with adrenal insufficiency. Clin Endocrinol (Oxf) 2004;61(3):367-375. 2. Debono M, Price JN, Ross RJ. Novel strategies for hydrocortisone replacement. Best Pract Res Clin 2 Endocrinol Metab 2009;23(2):221-232. 3. Filipsson H, Monson JP, Koltowska-Haggstrom M, Mattsson A, Johannsson G. The impact of glucocorticoid replacement regimens on metabolic outcome and comorbidity in hypopituitary patients. J Clin Endocrinol Metab 2006;91(10):3954-3961. 4. Gallagher P, Reid KS, Ferrier IN. Neuropsychological functioning in health and mood disorder: Modulation by glucocorticoids and their receptors. Psychoneuroendocrinology 2009;34 (suppl 1):S196-207. 5. Lupien SJ, McEwen BS. The acute effects of corticosteroids on cognition: integration of animal and human model studies. Brain Res Brain Res Rev 1997;24(1):1-27. 6. Kirschbaum C, Wolf OT, May M, Wippich W, Hellhammer DH. Stress- and treatment-induced elevations of cortisol levels associated with impaired declarative memory in healthy adults. Life Sci 1996;58(17):1475-1483. 7. Lupien S, Lecours AR, Lussier I, Schwartz G, Nair NP, Meaney MJ. Basal cortisol levels and cognitive deficits in human aging. J Neurosci 1994;14(5 Pt 1):2893-2903. 8. Lupien SJ, Wilkinson CW, Briere S, Ng Ying Kin NM, Meaney MJ, Nair NP. Acute modulation of aged human memory by pharmacological manipulation of glucocorticoids. J Clin Endocrinol Metab 2002;87(8):3798-3807. 9. Lupien SJ, Fiocco A, Wan N, et al. Stress hormones and human memory function across the lifespan. Psychoneuroendocrinology 2005;30(3):225-242. 10. Maheu FS, Collicutt P, Kornik R, Moszkowski R, Lupien SJ. The perfect time to be stressed: a differential modulation of human memory by stress applied in the morning or in the afternoon. Prog Neuropsychopharmacol Biol Psychiatry 2005;29(8):1281-1288. 11. Newcomer JW, Craft S, Hershey T, Askins K, Bardgett ME. Glucocorticoid-induced impairment in declarative memory performance in adult humans. J Neurosci 1994;14(4):2047-2053. 12. Young AH, Sahakian BJ, Robbins TW, Cowen PJ. The effects of chronic administration of hydrocortisone on cognitive function in normal male volunteers. Psychopharmacology (Berl) 1999;145(3):260-266. 13. McCormick CM, Lewis E, Somley B, Kahan TA. Individual differences in cortisol levels and performance on a test of executive function in men and women. Physiol Behav 2007;91(1):87-94. 14. Newcomer JW, Selke G, Melson AK, et al. Decreased memory performance in healthy humans induced by stress-level cortisol treatment. Arch Gen Psychiatry 1999;56(6):527-533. 15. de Quervain DJ, Roozendaal B, Nitsch RM, McGaugh JL, Hock C. Acute cortisone administration impairs retrieval of long-term declarative memory in humans. Nat Neurosci 2000;3(4):313-314. 16. Lupien SJ, Gillin CJ, Hauger RL. Working memory is more sensitive than declarative memory to the acute effects of corticosteroids: a dose-response study in humans. Behav Neurosci 1999;113(3):420- 430. 17. Joels M. Functional actions of corticosteroids in the hippocampus. Eur J Pharmacol 2008;583(2- 3):312-321. 50 Chapter 2

18. Sanchez MM, Young LJ, Plotsky PM, Insel TR. Distribution of corticosteroid receptors in the rhesus brain: relative absence of glucocorticoid receptors in the hippocampal formation. J Neurosci 2000;20(12):4657-4668. 19. Van Overwalle F. Social cognition and the brain: a meta-analysis. Hum Brain Mapp 2009;30(3):829- 858. 20. Li C, Chen K, Han H, Chui D, Wu J. An FMRI study of the neural systems involved in visually cued auditory top-down spatial and temporal attention. PLoS One 2012;7(11):e49948. 21. Wolkowitz OM, Reus VI, Weingartner H, et al. Cognitive effects of corticosteroids. Am J Psychiatry 1990;147(10):1297-1303. 22. Dullaart RP, Pasterkamp SH, Beentjes JA, Sluiter WJ. Evaluation of adrenal function in patients with hypothalamic and pituitary disorders: comparison of serum cortisol, urinary free cortisol and the human-corticotrophin releasing hormone test with the insulin tolerance test. Clin Endocrinol (Oxf) 1999;50(4):465-471. 23. Lezak MD, Howieson D, Loring DW. Neuropsychological Assessment. 4th ed. New York: Oxford University Press; 2004. 24. Jones RL, Owen LJ, Adaway JE, Keevil BG. Simultaneous analysis of cortisol and cortisone in saliva using XLC-MS/MS for fully automated online solid phase extraction. J Chromatogr B Analyt Technol Biomed Life Sci 2012;881-882:42-48. 25. United Nations Educational, Scientific and Cultural Organization. International Standard Classifica- tion of Education (ISCED). 2006. 26. Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates; 1988. 27. Arlt W, Rosenthal C, Hahner S, Allolio B. Quality of glucocorticoid replacement in adrenal insufficiency: clinical assessment vs. timed serum cortisol measurements. Clin Endocrinol (Oxf) 2006;64(4):384-389. 28. Alhaj HA, Massey AE, McAllister-Williams RH. Effects of cortisol on the laterality of the neural correlates of episodic memory. J Psychiatr Res 2008;42(12):971-981. 29. Smeets T, Dziobek I, Wolf OT. Social cognition under stress: differential effects of stress-induced cortisol elevations in healthy young men and women. Horm Behav 2009;55(4):507-513. 30. Harbeck B, Kropp P, Monig H. Effects of short-term nocturnal cortisol replacement on cognitive function and quality of life in patients with primary or secondary adrenal insufficiency: a pilot study. Appl Psychophysiol Biofeedback 2009;34(2):113-119. 31. Oertel H, Schneider HJ, Stalla GK, Holsboer F, Zihl J. The effect of growth hormone substitution on cognitive performance in adult patients with hypopituitarism. Psychoneuroendocrinology 2004;29(7):839-850. 32. Brummelman P, Elderson MF, Dullaart RP, et al. Cognitive functioning in patients treated for non- functioning pituitary macroadenoma and the effects of pituitary radiotherapy. Clin Endocrinol (Oxf) 2011;74(4):481-487. 33. Brummelman P, Koerts J, Dullaart RP, et al. Effects of previous growth hormone excess and current medical treatment for acromegaly on cognition. Eur J Clin Invest 2012;42(12):1317-1324. 34. de Kloet ER, Joels M, Holsboer F. Stress and the brain: from adaptation to disease. Nat Rev Neurosci 2005;6(6):463-475. 35. Joels M, Karst H, Krugers HJ, Lucassen PJ. Chronic stress: implications for neuronal morphology, function and neurogenesis. Front Neuroendocrinol 2007;28(2-3):72-96. Hydrocortisone Treatment and Cognition 51

36. Johannsson G, Nilsson AG, Bergthorsdottir R, et al. Improved cortisol exposure-time profile and outcome in patients with adrenal insufficiency: a prospective randomized trial of a novel hydrocortisone dual-release formulation. J Clin Endocrinol Metab 2012;97(2):473-481. 37. Benson S, Neumann P, Unger N, et al. Effects of standard glucocorticoid replacement therapies on subjective well-being: a randomized, double-blind, crossover study in patients with secondary adrenal insufficiency. Eur J Endocrinol 2012;167(5):679-685. 2 38. Ekman B, Bachrach-Lindstrom M, Lindstrom T, Wahlberg J, Blomgren J, Arnqvist HJ. A randomized, double-blind, crossover study comparing two- and four-dose hydrocortisone regimen with regard to quality of life, cortisol and ACTH profiles in patients with primary adrenal insufficiency. Clin Endocrinol (Oxf) 2012;77(1):18-25. 39. de Quervain DJ, Aerni A, Schelling G, Roozendaal B. Glucocorticoids and the regulation of memory in health and disease. Front Neuroendocrinol 2009;30(3):358-370. 40. Lovas K, Gjesdal CG, Christensen M, et al. Glucocorticoid replacement therapy and pharmacoge- netics in Addison’s disease: effects on bone. Eur J Endocrinol 2009;160(6):993-1002. 41. Ragnarsson O, Mattsson AF, Monson JP, et al. The relationship between glucocorticoid replacement and quality of life in 2737 hypopituitary patients. Eur J Endocrinol 2014;171(5):571-579. 52 Chapter 2

Supplemental data

Supplemental Materials and Methods. Full description of cognitive tests and reference data The following cognitive domains were assessed using several standardized and validated cognitive tests: memory, attention, executive functioning and social cognition.

Memory During the Rivermead Behavioral Memory Test (RBMT),1 a short story was read out loud to patients who were then asked to immediately recall as many details as possible from the story. This allowed the calculation of an immediate memory score for coherent verbal information. A delayed memory score for coherent verbal memory was determined based on the number of details of the story recalled by patients after a period of about 30 minutes. Different stories were used during the second and third visit. Verbal memory was also assessed with the 15 Words Test, which is a Dutch equivalent of the Rey Auditory Verbal Learning Test.2 During this test, 15 words were presented five times. After each trial, patients were asked to name immediately the words they remembered. This allowed the calculation of three different scores describing immediate memory. The short-term memory score is based on the number of words patients were able to name after the first presentation of the word list. The total immediate memory score represents the total number of words the patients remembered over the five trials. The learning score describes the difference between the number of words remembered in the third trial in comparison with the first trial. Besides immediate memory, delayed memory was measured with the 15 words test. The delayed memory score is based on the number of words a patient could recall after a period of about 30 minutes. To correct for the total immediate memory, there was also a delayed corrected for total memory score. The last part of the test was a recognition task. A total of 30 words were presented to the patient, of which 15 words were presented during the previous five trials and of which 15 words were new to the patient. The patients had to decide whether the words were new or familiar, which allowed the calculation of a recognition memory score. During the second and third visit, different lists of words were applied. To assess short term memory, the Digit Span Forward test was used.3 Patients were presented with a sequence of digits and immediately had to repeat the digits in the same order as they were presented. The test started with sequences of two digits which expanded until the maximum of nine digits. When a patient could not correctly repeat two sequences of the same length, the test was aborted. The total number of correctly repeated sequences was registered. During the second and third visit, different versions of the test were performed. Visual memory was assessed with the Rey Complex Figure Hydrocortisone Treatment and Cognition 53

Test.4 During this test, a complex geometrical figure was placed in front of patients, who were instructed to copy the figure as accurately as possible. After a 3-minute delay, patients were asked to reproduce the figure from memory. Approximately 15 minutes later, patients were again asked to reproduce the figure from memory. The number of details correctly recalled after 3 and 15 minutes were used to calculate 2 an immediate and delayed memory score. If patients had an impaired copy drawing (< 10th percentile) measures of immediate and delayed memory were excluded for analysis. An impaired copy drawing might indicate problems with visuoconstructive functions causing unreliable scores for immediate and delayed memory. During the second and third visit, different complex figures were used. Aspects of visual memory were also assessed with the 15 Figures Test. During this test, 15 simple, geometrical figures were presented five times. After each trial, patients were asked to draw immediately the figures they remembered. This provides a score for immediate visual memory. The delayed visual memory score was based on the number of figures a patients could recall after a period of 15 minutes. Finally, a recognition task was administered, which consisted of 30 figures, of which 15 were presented during the five trials and of which 15 figures were new. Patients had to decide whether the figure was new of familiar, which allowed the calculations of a recognition score for visual memory. To prevent learning effects, different figures were used during the assessment after each treatment period.

Attention The Test of Attentional Performance is a computerized test consisting of several subtests.5 In the present study, the subtests vigilance, divided attention, visual scanning and alertness were used. Test instructions were presented on a computer screen and read out loud by the test examiner. In order to familiarize the participants with the tasks, a brief sequence of practice trials preceded each test. During the vigilance task, two vertically arranged squares were shown in the middle of the screen and a pattern jumped from one square to the other. However, every now and then the pattern appeared twice in succession in the same square. When this hap- pened, patients had to respond as quickly as possible by pressing a key. Reaction time of correct responses, variability in reaction time and the number of omission (lack of response to target stimuli) and commission (response to non-target stimuli) errors were calculated. Divided attention is needed when two (or more) tasks have to be performed simul- taneously. In the visual task, a series of matrices was presented in the center of the computer screen. Each matrix (4x4) consisted of a regular array of sixteen dots and crosses. When four of these crosses formed a square, patients had to press a key as quickly as possible. The acoustic task consisted of a high (2000 Hz) and low (1000 54 Chapter 2

Hz) tone which were presented alternately according to the synchronous rhythm of the changing positions of the crosses. When the same tone occurred twice in a row, patients had to press a key as quickly as possible. A total of 100 visual and 200 acoustic stimuli was presented including 17 visual and 16 acoustic targets. Reaction time of correct responses, variability in reaction time and the number of omission and commission errors were calculated and represented measures of divided attention. During the visual scanning task a series of 5 by 5 matrices were presented. Each matrix consisted of a regular array of 25 squares which each had an opening on one side (top, bottom, left or right side). A square with an opening on the top was defined as the target stimulus. The target stimulus occurred in 50 % of matrices and was randomly distributed across the matrix. Patients were asked to press the left response button as quickly as possible when a matrix contained the target stimulus and to press the right response button if the target stimulus was not present. A total of 50 trials was presented (25 target trials and 25 non-target trials). Reaction times of correct responses, variability of reaction time and the number of commission and omission errors were calculated and represented measures of selective attention, i.e. attention for the target stimulus and not for non-target stimulus. During the alertness task, reaction times were measured during two conditions. The first condition concerned a simple reaction time measurement, during which a cross appeared on the screen at random varying intervals. Patients were instructed to respond as quickly as possible by pressing a key when the cross appeared (tonic alertness task). In the second condition, a warning tone preceded the appearance of the stimulus (phasic alertness task). A total of 40 trials were presented, 20 trials with warning tone and 20 trials without warning tone. The time span between the warning tone and the appearance of the stimulus was random (between 300 and 700 ms). Measures of tonic and phasic alertness were calculated on the basis of reaction times of patients. In addition, the variability of reaction time and the number of omission errors were measured.

Executive functioning Divergent thinking is the ability to approach a task or situation in different ways. So- called fluency tests rely on divergent thinking and evaluate the spontaneous production of words under restricted search conditions.6 During the first fluency test applied in the present study, the semantic subtest, patients were asked to name as many animals as possible within one minute. For the second and third measurement patients were asked to name as many professions and grocery articles respectively. Participants were not allowed to name the same word twice. In a second fluency test, the phonemic subtest, patients were asked to name as many words as possible in one minute, starting with a specific letter. All existing nouns were allowed. However, personal names and names of places were not allowed. This part was administered three times, each time using a Hydrocortisone Treatment and Cognition 55 different letter, namely D, A and T at the first measurement, K, O and M at the second measurement and P, G and R at the third measurement.7 For both subtests, the total number of correctly produced words was counted. To assess working memory, the Digit Span Backward test was used.3 Patients were presented with sequences of digits, which they had to repeat backwards. The test started 2 with sequences of two digits which expanded until the maximum of eight digits. When a patient failed to correctly repeat two sequences of the same length, the test was aborted. The total number of sequences that was repeated correctly was registered. Different strings of digits were used during each measurement. The Trail Making Test assesses visual scanning, speed of processing (condition A) and cognitive flexibility (condition B).8 During condition A, patients were asked to connect 25 circles containing numbers as quickly as possible in ascending order. Part B consisted of 25 circles, containing both numbers and letters. Patients were again required to connect the circles as fast as possible in ascending order, this time alternating between numbers and letters i.e. 1-A-2-B-3-C, etc.. The target measure for cognitive flexibility was the performance on condition B corrected for condition A (Trail Making Test B/A). Parallel versions of the test were performed during each measurement.

Social cognition Social cognition can be defined as the mental operations underlying social interactions, including perceiving, interpreting, and generating responses to the intentions, disposi- tions, and behaviors of others.9 In this study, social cognition was assessed with the Reading the Mind in the Eyes Test (RMET).10 During this test, patients were presented with 36 photographs, one after each other, of the eye-region of the face. Along with each pair of eyes, four words describing an emotion were presented. Patients were asked to choose which word described best what the person in the picture is feeling or thinking. To exclude any influence of difficulties with word comprehension, a glossary with the presented words was made available.

Reference data: Healthy control subjects The performances of patients were compared to published normative datasets from healthy populations.

Memory The results of the RBMT were compared to a healthy reference group (n = 344) consisting of participants aged 17–89.11 The scores of patients were corrected for age, sex and education. In the study by Schmand and colleagues participants were presented with two stories of comparable length. However, in the present study, patients were presented with only one story. To compensate for this difference, the score obtained 56 Chapter 2

in the present study was doubled. Reference data for the 15 words test were derived from control subjects (n = 847) consisting of participants aged 14–87 for the measures total immediate memory, delayed memory and delayed corrected for total memory.11 Reference data for the measures short term memory, learning score and recognition of the 15 words test were derived from control subjects of the Maastricht Aging Study.2 The scores were corrected for age, sex and education. The results of the Digit Span test were compared to the reference group as described in the Wechsler Memory Scale Manual (n = 316). These scores were corrected for age.3 Patients’ scores on the Rey Complex Figure were compared to reference data as provided in the manual of the Rey Complex Figure Test and Recognition Trial and were also corrected for age.12 No normative data was available for the 15 Figures Test.

Attention Norm data for the divided attention, visual scanning and alertness subtests were provided by the test authors within the Test of Attentional Performance Version 2.2.5 A correction for age and sex was applied. No normative data was available for the vigilance subtest.

Executive functioning The results of the semantic subtest of the Verbal Fluency Test were compared to a reference group (n = 464) consisting of healthy participants aged 17–90.11 The scores were corrected for age and education. The scores of the phonemic subtest of the Verbal Fluency Test were compared to a reference group (n = 570) consisting of participants aged 17–90 and corrected for education.11 The results of the TMT were compared to a reference group (n = 478), consisting of healthy participants aged 17–90.11 The scores were corrected for age, sex and education.

Social cognition The performance on the RMET was compared to a healthy reference group (n = 200) and corrected for sex.13 Hydrocortisone Treatment and Cognition 57

Supplemental Figure 1. Schematic representation of the study design. M1: Baseline measurement; M2: Measurement after first treatment period; M3: Measurement after second treatment period. 58 Chapter 2

Supplemental Table 1. Cognitive performances of pituitary patients treated for adrenal insufficiency at baseline Group 1 (first low Group 2 (first high dose followed by dose followed by high dose) low dose) (N = 22) (N = 25) P-value* Memory RBMT † Immediate memory -0.80 (1.05) -0.83 (0.87) 0.949 Delayed memory -0.69 (0.76) -0.86 (0.77) 0.535 Delayed corrected for immediate 0.08 (1.12) -0.41 (1.41) 0.311 memory 15 Words test † Short-term memory -0.01 (0.94) -0.59 (0.79) 0.030a Total immediate memory -0.24 (1.22) -0.78 (0.75) 0.060 Learning score 0.17 (1.18) 0.09 (0.85) 0.831 Delayed memory -0.31 (1.26) -0.78 (0.80) 0.281 Delayed corrected for total memory -0.28 (0.94) -0.30 (0.73) 0.915 Recognition -0.13 (1.11) -0.22 (0.65) 0.126 Digit Span forward † Short-term memory 0.15 (0.95) 0.47 (1.00) 0.330 Rey Complex Figure ‡ Immediate memory 0.70 (1.32)b 0.64 (1.51) 0.921 Delayed recall 0.64 (1.24)b 0.74 (1.43) 0.691 15 Figures Test (raw data) ‡Δ Total memory 42.91 (13.27) 44.04 (12.46) 0.915 Delayed memory 10.55 (3.31) 10.92 (3.00) 0.780 Recognition 26.86 (6.39) 28.68 (1.55) 0.372 Attention Vigilance (raw data) ‡ Δ Reaction time of correct responses 606.24 (84.27)b 643.16 (131.02) 0.402 Variability of reaction time 128.67 (135.43)b 131.44 (62.85) 0.120 Number of omission errors 0.62 (0.87)b 1.44 (1.36) 0.463 Number of commission errors 9.52 (37.96)b 2.04 (2.65) 0.026a Divided attention ‡ Reaction time auditory responses -1.13 (0.88) -0.92 (0.82) 0.288 Variability of reaction time auditory 0.16 (1.06) 0.10 (1.15) 0.473 responses Reaction time visual responses -0.48 (0.82) 0.05 (0.97) 0.035 Variability of reaction time visual -0.46 (1.21) -0.05 (0.96) 0.175 responses Number of omission errors -0.37 (0.93) -0.39 (0.93) 0.974 Number of commission errors 0.13 (1.05) 0.23 (0.97) 0.842 Hydrocortisone Treatment and Cognition 59

Supplemental Table 1. Cognitive performances of pituitary patients treated for adrenal insufficiency at baseline (continued) Group 1 (first low Group 2 (first high dose followed by dose followed by high dose) low dose) 2 (N = 22) (N = 25) P-value* Visual scanning ‡ Reaction time for target stimuli -0.87 (0.73) -0.75 (0.90) 0.886 Variability of reaction time for target -0.78 (0.74) -0.72 (0.88) 0.912 stimuli Number of omission errors -0.56 (1.03) 0.34 (1.16) 0.015 Number of commission errors -0.78 (0.32) -0.72 (0.36) 0.346 Alertness ‡ Tonic alertness reaction time -0.75 (1.29) -0.80 (0.76) 0.583 Tonic alertness variability in reaction -0.08 (0.85) -0.35 (0.98) 0.410 time Phasic alertness reaction time -1.17 (0.56) -0.92 (0.77) 0.366 Phasic alertness variability in reaction -0.31 (0.65) -0.19 (0.91) 0.492 time Executive functioning Fluency test † Semantic fluency -0.21 (1.00) -0.04 (1.32) 0.789 Phonemic fluency 0.12 (1.39) -0.34 (0.82) 0.709 Digit Span backwards † Working memory 0.29 (0.83) 0.18 (1.12) 0.398 Trail Making Test ‡ Condition A, time -0.80 (1.01) -0.52 (1.21) 0.522 Condition B, time -0.43 (0.94) -0.38 (0.97) 0.773 Condition B/A -0.04 (0.99) -0.15 (0.84) 0.631 Social cognition Reading the Mind in the Eyes Test †‡ -0.85 (1.05) -0.80 (0.90) 0.923 Cognitive performance data are given as mean Z-scores (SD); RBMT: Rivermead Behavioral Memory Test; * Group 1 versus Group 2 by Mann-Whitney U test; †: verbal test; ‡: non-verbal test; a P-value not statistically significant after Bonferroni correction; b N = 21; Δ Performances on these cognitive tests are given as raw data mean (SD) due to lacking reference data 60 Chapter 2

References

1. Wilson BA, Cockburn J, Baddeley AD. The Rivermead Behavioural Memory Test. Bury St Ed- munds: Thames Valley Test Company; 1985. 2. Van der Elst W, van Boxtel MP, van Breukelen GJ, Jolles J. Rey’s verbal learning test: Normative data for 1855 healthy participants aged 24-81 years and the influence of age, sex, education, and mode of presentation. J Int Neuropsychol Soc 2005;11:290-302. 3. Wechsler D. Wechsler Memory Scale Revised. The Psychological Corporation; 1987. 4. Rey A. L’examen psychologique dans les cas d’encéphalopathie traumatique. Archives de Psycholo- gie 1941;28:286-340. 5. Zimmermann P, Fimm B. Test of Attentional Performance Version 2.2. Psychologische Testsysteme; 2007. 6. Straus E, Sherman E, Spreen O. A compendium of neuropsychological test: administration norms and commentary. New York: University Oxford Press; 2006. 7. Schmand B, Groenink SC, van den Dungen M. Letter fluency: Psychometric properties and dutch normative data. Tijdschr Gerontol Geriatr 2008;39:64-76. 8. Reitan RM. Validity of the Trail Making Test as an indicator of organic brain damage. Percept. Mot Skills 1958;8:271-276. 9. Davis MC, Lee J, Horan WP, Clarke AD, McGee MR, Green MF, Marder SR. Effects of single dose intranasal oxytocin on social cognition in schizophrenia. Schizophr Res 2013;147:393-7. 10. Baron-Cohen S, Wheelwright S, Hill J, Raste Y, Plumb I. The “reading the mind in the eyes” test revised version: A study with normal adults, and adults with asperger syndrome or high-functioning autism. J Child Psychol Psychiatry 2001;42:241-251. 11. Schmand B, Houx P, de Koning I. Normen neuropsychologische tests 2012. Nederlands Instituut van Psychologen; 2012. 12. Meyers JE, Meyers KR. RCFT Rey Complex Figure Test and Recognition Trial: Professional Manual. Ltz, FL. Psychological Assessment Recources; 1995. 13. Vellante M, Baron-Cohen S, Melis M, Marrone M, Petretto DR, Masala C, Preti A. The “Reading the Mind in the Eyes” Test: systematic review of psychometric properties and a validation study in Italy. Cognitive Neuropsychiatry 2013;18:326-354.

Chapter 3

Hydrocortisone dose infl uences pain, depressive symptoms and perceived health in adrenal insuffi ciency: a randomized controlled trial

J Werumeus Buning P Brummelman J Koerts RPF Dullaart G van den Berg MM van der Klauw WJ Sluiter O Tucha BHR Wolff enbuttel AP van Beek

Neuroendocrinology 2015;103(6):771-8 64 Chapter 3

Abstract

Background There is a major lack of randomized controlled trials (RCTs) evaluating the effects of hydrocortisone (HC) substitution therapy in patients with secondary adrenal insufficiency. Therefore, we evaluated the effects of two different replacement doses of HC on health-related quality of life (HRQoL) in a RCT.

Methods This RCT with double-blind cross-over design was performed at the University Medi- cal Center Groningen. Forty-seven patients (29 men, age 51 ± 14 years, range 19–73) with secondary adrenal insufficiency participated. Patients received both a lower and a higher dose of HC (0.2–0.3 and 0.4–0.6 mg/kg body weight/day) for 10 weeks in random order. HRQoL was assessed with a daily mood and symptom checklist (Pa- tient Health Questionnaire-15 [PHQ-15], Generalized Anxiety Disorder-7 [GAD-7], Patient Health Questionnaire-9 [PHQ-9]) and with questionnaires assessing general well-being (RAND 36-Item Health Survey [RAND-36]), mood (Hospital Anxiety and Depression Scale [HADS]) and fatigue (Multidimensional Fatigue Inventory-20 [MFI- 20]). ClinicalTrials.gov Identifier: NCT01546922.

Results Patients receiving the higher dose of HC reported significantly fewer symptoms of depression (p = 0.016 and p = 0.045 for HADS and PHQ-9, respectively), less general and mental fatigue (p = 0.004 and p = 0.003, respectively, both MFI-20), increased motivation (p = 0.021, MFI-20), better physical functioning (p = 0.041), better general health (p = 0.013) and more vitality (p = 0.025) (all RAND-36). In addition, while on the higher dose, fewer somatic symptoms (p = 0.022) and less pain (p < 0.001) (both PHQ-15) were experienced.

Conclusions On the higher dose of HC, patients reported a better HRQoL on various domains as compared to the lower dose of HC. The fact that a higher dose of HC may improve patients well-being should be taken into consideration when individualizing the HC substitution dose. Hydrocortisone Treatment and Quality of Life 65

Introduction

Adrenal insufficiency (AI) requires life-long, daily medical treatment with glucocorticoids (GCs). The standard replacement therapy consists of the oral administration of GCs, usually hydrocortisone (HC), with the aim of mimicking the daily rhythm in cortisol concentrations seen in healthy people. However, there are no criteria to objectively monitor and evaluate the quality of substitution therapy.1,2 Consequently, 3 current practice varies widely with regard to the administered dose of GCs. This was recently highlighted in a large study that reported that 20 % of patients received low daily doses (< 20 mg HC equivalent dose), 40 % received intermediate doses and 40 % received high doses (≥ 30 mg HC equivalent dose).3 Current HC dosing schemes are the result of a complex balancing of factors including the endogenous cortisol production as documented in healthy individuals, variation in plasma cortisol in relation to the HC substitution dose, and the risks and benefits of applying (long-term) higher or lower dosing schemes.2,4–6 Health-related quality of life (HRQoL) is another important issue in the individualization of the dosing scheme. HRQoL refers to subjective and multidimensional domains encompassing, for instance, physical functioning, psychological state, and social interaction. As such, it is often a subject of discussion between physician and patient. Cross-sectional studies suggest that there is a relationship between the mean daily dose of GCs and HRQoL, with higher GC doses being significantly related to more severely impaired subjective health status.7–10 However, in view of the cross-sectional nature of these studies, it is impossible to distinguish whether diminished QoL is a result of a higher dose or, conversely, whether diminished QoL results in the prescription of a higher dose. A few controlled studies have assessed HRQoL in relation to HC dose. All studies were small and applied different HC regimens, often with changes to both timing and dosing.6 These studies produced variable results, with increases,11 no change,12,13 or decreases14 in QoL being reported with higher doses of GCs. Thus the exact relation between HC dose and HRQoL remains unknown. This randomized, double-blind cross-over study was initiated to determine the effect of the total daily dose of HC on several indices of HRQoL by comparing a lower replacement dose of HC to a higher replacement dose of HC in patients with secondary AI. 66 Chapter 3

Patients and Methods

Patients For this randomized, double blind cross-over study patients were recruited from the endocrine outpatient clinic of the University Medical Center Groningen. A total of 63 patients were included, of whom 60 completed the run-in phase and the baseline measurement. Eligibility, inclusion and follow-up are shown in online supplemental figure 1 (for all online suppl. material, see www.karger.com/doi/10.1159/000442985). All patients had secondary AI and were on stable GC substitution therapy at least 6 months prior to study entry. The diagnosis of secondary AI was based on internationally accepted biochemical criteria, principally early morning (08.00 – 09.00 h) serum cortisol measurements and, if necessary, an insulin tolerance test. Early morning cut- off cortisol levels for AI in our center were validated for patients with hypothalamic- pituitary disorders as published previously.15 Other pituitary hormone deficiencies were adequately replaced when necessary for at least 6 months prior to entry of the study and treatment remained stable during the study. Other inclusion and exclusion criteria have been described earlier.16 All patients were tested in the period between May 2012 and June 2013.

Intervention Group 1 first received a lower dose of 0.2–0.3 mg HC/kg body weight/day (i.e. a total daily dose of 15–20 mg HC/day) divided in three doses (before breakfast, before lunch, before dinner) for 10 weeks, followed by a higher dose of 0.4–0.6 mg HC/kg body weight/day (i.e. a total daily dose of 30–40 mg HC/day) for a further 10 weeks. Group 2 received the two doses in reversed order. For the exact dosing scheme see online supplemental table 1. Randomization to one of the two treatment groups was performed by Tiofarma Inc, The Netherlands, with a block size of four. In cases of intercurrent illness or fever, patients were advised to double or triple their HC dose according to a fixed protocol. In the last week before testing this was not allowed. After each treatment period patients returned to the hospital to hand in questionnaires and diaries. In addition, blood samples were drawn, both at one hour and five hours after ingestion of the morning dose of HC.

Withdrawals A total of 63 patients were included in the study. During the run-in phase, 3 patients withdrew from the study; none of these withdrawals was suspected to be related to the dose of HC (2 patients withdrew their informed consent, 1 patient was withdrawn by the investigator because of the presence of a chronic-pain syndrome). Eight patients withdrew from the study during the lower dose condition; three of those withdrawals Hydrocortisone Treatment and Quality of Life 67 were suspected to be related to the dose of HC (influenza A infection [n = 1], inability to tolerate the dose [n = 1], and a Herpes zoster ophthalmicus infection [n = 1]). Five patients withdrew from the study during the higher-dose condition; one of these withdrawals was suspected to be related to the dose of HC (unpleasant feelings). All other withdrawals were not suspected to be related to the dose of HC. The withdrawals were evenly distributed in the two study arms and study periods.

Laboratory measurements 3 Serum cortisol levels were measured by a commercially available electrochemiluminescence immunoassay (ECLIA, RocheModular Systems) as described earlier.16

HRQoL Measures At the end of each treatment period, patients were asked to complete the Hospital Anxiety and Depression Scale (HADS),17 the Multidimensional Fatigue Inventory-20 (MFI-20),18 the RAND 36-Item Health Survey (RAND-36),19 and the Cognitive Failures Questionnaire (CFQ),20,21 assessing multiple aspects of HRQoL. In addition to these questionnaires, patients were instructed to complete a daily mood and symptom diary, consisting of items of the Patient Health Questionnaire-15 (PHQ-15),22 the Generalized Anxiety Disorder-7 (GAD-7)23 questionnaire, and the Patient Health Questionnaire-9 (PHQ-9).24 A detailed description of the questionnaires and the reference population is given in the online supplemental materials and methods.

Statistics A power analysis performed before the study indicated that two arms of 25 patients each were required to detect an effect size of 0.4 with a two-sided alpha of 0.05 and a beta of 0.80, even when between test correlations are poor (0.50). An effect size of 0.4 was chosen because it was considered a relevant change with a small to moderate effect. To allow for a dropout rate of about 20 %, a total of 60 patients were needed. In this study, descriptive data are presented as mean ± SD, median with interquartile ranges (IQR) in brackets, frequencies or percentages. Data on HRQoL are presented as median with IQR in brackets. Data for the PHQ-15, GAD-7, and PHQ-9 were pooled and averaged over the last four weeks of each treatment period to give a stable measure of the severity of symptoms during the treatment period. Normality of data was analyzed using Q-Q plots. Because of the non-normal distribution of the HRQoL data, non-parametric tests were used. In cross-over studies a treatment effect can be distinguished from a period effect as well as from a treatment-by-period interaction effect. When a treatment effect is present, the effect can be ascribed to the dose administered. A period effect can be a result of time, for example familiarization with the study situation. A treatment-by-period interaction is often referred to as a carryover 68 Chapter 3

effect. To test for period effects and carryover effects, the procedure developed by Altman was used.25 Differences in HRQoL between the lower and the higher doses were assessed with the Wilcoxon signed-rank rest for paired data. Cohen’s d effect sizes were calculated to give a measure of the magnitude of the difference. An effect size of d = 0.2 was considered a small effect; d = 0.5 a moderate effect; and d = 0.8 a large effect.26 The level of significance was set at two-sided P < 0.05. All statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS, Inc., Armonk, NY, USA), Version 22.

Results

Study Population The clinical characteristics of the study population have been published previously.16 In short, 47 patients (29 men, age 51 ± 14, range 19–73 years) completed both study periods and the data of these patients were used for further analysis. The patients’ characteristics are given in online supplemental table 2. The two groups were similar with respect to demographic and clinical characteristics, as described previously.16

Cortisol Levels Administration of the higher dose of HC resulted in significantly higher serum cortisol levels than the lower dose of HC, both 1 h after ingestion of the morning dose (925 [820–1,045] nmol/L and 620 [490–730] nmol/L for the higher and lower dose, respectively, p < 0.001) and 5 h after ingestion of the morning dose (305 [235–524] nmol/L and 155 [95–270] nmol/L for the higher and lower dose respectively, p = 0.001).

Stress-Related Dose Adjustments The frequency of dose adjustments was equal during treatment with the lower dose (159 adjustments, 1.6 % of the total dose administrations) and the higher dose (146 adjustments, 1.5 %). Furthermore, the number of temporary dose increments did not differ between the study groups or study periods (data not shown).

Somatic Complaints, Pain and HRQoL On treatment with the higher dose of HC, patients reported fewer symptoms of depression than when on treatment with the lower dose of HC (p = 0.016 for HADS, fig. 1, online suppl. table 3; p = 0.041 for PHQ-9, table 1). Similarly, they reported a better motivation (p = 0.021), less general fatigue (p = 0.004), less mental fatigue (p = 0.003, MFI-20, fig. 1, online suppl. table 3), better physical functioning (p = 0.041), more vitality (p = 0.025), and better general health perception (p = 0.013, RAND-36, fig. 1, Hydrocortisone Treatment and Quality of Life 69

Table 1. Daily assessment of somatic complaints, anxiety, and depressive symptoms in pituitary patients with secondary AI in the lower-dose and the higher-dose condition (n = 45 due to missing data) Effect Lower-dose HC Higher-dose HC p value1 size PHQ-15 12.89 [11.13; 12.32 [11.28; 0.023 0.2 16.04] 14.76] Back pain 1.00 [1.00; 1.84] 1.00 [1.00; 1.47] 0.011 0.3 3 Pain in legs, arms or joints 1.44 [1.00; 2.25] 1.14 [1.00; 1.98] 0.001 0.4 Headache 1.04 [1.00; 1.80] 1.04 [1.00; 1.54] 0.037 0.1 Shortness of breath 1.00 [1.00; 1.95] 1.00 [1.00; 1.57] 0.046 0.2 PHQ-9 10.63 [9.07; 14.52] 9.13 [10.00; 12.87] 0.041 0.2 Little interest or pleasure in 1.07 [1.00; 1.81] 1.00 [1.00; 1.69] 0.015 0.1 doing things Feeling tired or having little 1.29 [1.00; 2.56] 1.14 [1.00; 2.01] 0.007 0.2 energy Poor appetite or overeating 1.00 [1.00; 1.18] 1.00 [1.00; 1.09] 0.014 0.2 Data were pooled and averaged over the last 4 weeks of each treatment period. Higher scores indicate more complaints. The scores on the two doses are given as median [IQR]. Only significant subitems of the PHQ-15and PHQ-9 are shown. 1 p value lower-dose versus higher-dose by Wilcoxon signed-rank test for paired observations. online suppl. table 3). Furthermore, while receiving the higher dose of HC, patients reported fewer somatic complaints (p = 0.023, PHQ-15, table 1), particularly less pain (p < 0.001, fig. 2) and less shortness of breath (p = 0.046, table 1). In addition, while receiving the higher dose of HC patients reported more interest or pleasure in doing things (p = 0.015), more energy (p = 0.007), and less often a disturbed eating pattern (p = 0.014, PHQ-9, table 1) compared to the lower dose. No differences in anxiety (HADS, fig 1, online suppl. table 3; GAD-7, data not shown) and self-reported cognitive failures (CFQ, online suppl. table 3) were found between the doses. Effect sizes for the differences found ranged between 0.1 and 0.4. A significant period effect was found for depressive symptoms (p = 0.018), general fatigue (p = 0.008) and mental fatigue (p = 0.044) (online suppl. fig. 2). Graphical analysis of this effect showed that the group first receiving the lower dose of HC showed no change in reported symptoms after switching to the higher dose, whereas the group first receiving the higher dose of HC showed a significant increase in complaints after switching to the lower dose of HC. No significant carryover effects were found for the HRQoL measures, except for the blunders subscale of the CFQ (p = 0.024). 70 Chapter 3

HADS Anxiety --> worse QoL * # Depression * # General Fatigue Physical Fatigue MFI-20 # Mental Fatigue --> worse QoL * Reduced Activity * Reduced Motivation * Physical Functioning Social Functioning Role Limitations due to Physical Problems RAND-36 Role Limitations due to Emotional Problems worse QoL <-- Mental Health * Vitality Pain * General Health Perception -2 -1 0 1 2 Z-Score

Figure 1. Z-scores for HADS, MFI-20 and RAND-36 for patients treated for secondary AI (n = 47). Higher HADS scores indicate more anxiety and/or depression, higher MFI-20 scores indicate more fatigue, and higher RAND-36 scores indicate better quality of life. Gray circles represent scores on the higher dose; black squares represent scores on the lower dose. Data represent mean ± SEM. * p < 0.05. # A significant period effect was found in addition to a significant treatment effect.

10 Treatment period 1 Treatment period 2 Group 1 9 Lower dose Lower dose Group 2

6 Higher dose Higher dose 5 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Week

Figure 2. Pain scores plotted for each week For the composite pain score, the items ‘stomach pain’, ‘back pain’, ‘pain in legs, arms or joints’, ‘headache’, and ‘chest pain’ of the PHQ-15 were summed up, with scores ranging from 5 to 35. Scores represent mean ± SEM. Figures on the x-axis are weeks. Hydrocortisone Treatment and Quality of Life 71

Discussion

This randomized, double-blind cross-over trial showed that patients experience benefits on various aspects of HRQoL while on a higher dose of HC substitution when compared to a lower dose of HC. These effects included, with a striking congruence, improved sense of general and mental health, improved physical functioning, fewer symptoms of depression, fewer somatic complaints, less pain and less fatigue. Our 3 results emphasize that HRQoL is a clinically relevant aspect of HC treatment that needs to be taken into account when individualizing HC substitution therapy. Overall, most of the findings appear be related to energy and vitality. Physical health seems to be more affected than mental health. For example, with regard to depression, a close examination of the specific symptoms of the PHQ-9 showed that the somatic complaints (e.g., lack of energy, disturbed eating patterns) are affected by HC dose to a greater degree than mood (e.g. feeling down, depressed or hopeless or feeling bad about oneself). Furthermore, pain and fatigue were strongly influenced by HC dose. On the other hand, HC dose does not seem to impact symptoms of anxiety (HADS anxiety scale, GAD-7). However, the depression subscale of the HADS assesses mostly mental aspects (e.g. I enjoy the things I used to enjoy, I can laugh and see the funny side of things), so mental health is not unaffected. Nevertheless, the effects of HC dose on energy and vitality related aspects are most pronounced. Secondary AI is often accompanied by other pituitary hormone deficiencies. Thyroid hormone deficiency, growth hormone deficiency or sex hormone deficiency could also have an effect on quality of life.27 However, in our study other pituitary hormone deficiencies were adequately replaced when necessary for at least 6 months prior to study entry and replacement therapy was held constant during the study periods. Furthermore, due to the crossover design of the study, any remaining effect of these deficiencies would have influenced patients during both treatment periods. Therefore the effect of the other pituitary hormone deficiencies is unlikely to have influenced the results. One might argue that patients with primary AI lack these possible influential factors. However, primary AI is often accompanied by other comorbidities, e.g. thyroid disorders, gonadal disorders, diabetes, and treatments (e.g. fludrocortisone and dehydroepiandrosterone) which in turn have an effect on QoL. Previous studies on HC substitution dose and HRQoL were inconclusive. In a recently published large cross-sectional study, Ragnarsson et al. showed that an increasing dose of HC was associated with impaired HRQoL.9 However, due to the cross-sectional design, no conclusions could be drawn with regard to the causality of this association. In addition, previous randomized controlled trials (RCTs) showed conflicting results, mostly in contrast to our findings. A study by Alonso et al. showed better general health perception on a dose of 30 mg HC/day compared to a dose of 20 mg HC/day.11 In 72 Chapter 3

contrast, two studies comparing doses of 15, 20 or 30 mg HC/day found no differences in QoL.12,13 However, patients samples were small (n = 9 for Wichers et al.12 and n = 10 for Behan et al.13) which might be a reason for the lack of differences. In another RCT, a decline in physical QoL and current well-being was reported after treatment with doses of 20 mg HC/day or 5 mg prednisone/day as compared to 15 mg HC/day.14 However, besides the dose, the timing of dose administration was also altered in this study. This precludes proper interpretation of the results because not only dose but also rhythm was changed. The latter is an inherent feature of the intact hypothalamic- pituitary-adrenal axis and may be responsible for the observed differences. In our study, HRQoL was causally related to HC dose and the ensuing serum cortisol concentrations. It has previously been described that high doses of steroids may result in a state of euphoria (potentially mediated by mineralocorticoid receptors in the brain).28 This effect, however, was observed at supraphysiological doses of GCs. Our results extend the previous data, showing that subjective well-being is influenced at much lower doses of GCs. Although the effect sizes indicate that the observed differences were small to moderate (ranging between 0.1 and 0.4), they can be considered meaningful. For example, a difference of 3 points is regarded to be the minimally clinically important difference for the RAND-36.29 We found this clinically important difference in six of the eight domains. Considering that substitution therapy is life long, even small contributions to perceived health are important. One of the most striking results of our study is the effect of HC dose on pain. The relationship between cortisol and pain is not univocally reported. A study in healthy subjects showed that higher cortisol levels prior to a cold pain pressor test predicted lower subsequent pain sensitivity in men, but not in women.30 Similarly, medically induced hypocortisolism resulted in enhanced subjective pain sensitivity.31 On the other hand, some studies showed an inverse relationship, with higher levels being associated to greater pain sensitivity32,33 and decreases in pain thresholds.34 Furthermore, Wingenfeld et al.35 were unable to show an effect of the administration of HC or of dexamethasone-induced hypocortisolism on pain sensitivity. The results of the present study support the knowledge that HC dose plays a role in pain sensitivity and that even small changes of cortisol levels within the physiological range attenuate pain perception. This increased pain perception was not specifically located in one bodily area, suggesting a central effect of HC. In our study, bodily pain on the RAND-36 was not found to be different between the two HC doses. Although this may seem contradic- tory, it must be noted that the RAND-36 has only two non-specific questions regarding pain, while the PHQ-15 measures pain on a daily basis based on five different bodily areas. This discrepancy is therefore likely to be a methodological issue. Another potentially relevant finding of this study was that in the group receiving the higher dose of HC followed by the lower dose, a lowering of the dose resulted Hydrocortisone Treatment and Quality of Life 73 in a significant increase in depressive symptoms, mental fatigue and general fatigue. In comparison to the previous period, they experienced more symptoms of fatigue and depression. It might be that these patients realized ‘what they were missing’ once they switched to the lower dose. A decrease in reported symptoms was, however, not observed in the other group switching from the lower dose to the higher dose. Due to the design of the study, this group could not experience the ‘realizing what you are missing once it is gone’ effect. However, it could be expected that the same effect 3 would be present in this group when the higher dose of HC was followed by a lower dose. Therefore, from a practical point of view, these findings suggest that it may be better to start with a lower dose and subsequently increase it when needed, because the reverse will possibly increase depressive symptoms, general fatigue and mental fatigue. That this period effect was not found in for example the pain scores, could be due to methodological differences. Pain was assessed every day by diaries and then combined to create a total score per treatment period, while depression and fatigue were measured at the end of each treatment period only. It might be possible that these different approaches led to this discrepancy. Lower plasma cortisol levels are found in several stress-associated neuropsychiatric disorders, e.g., posttraumatic stress disorder, panic disorders and chronic pain and fatigue syndromes.36–38 This resembles the findings in our study, with more complaints of pain and fatigue being reported by patients receiving the lower dose of HC, an intriguing finding which points to a direct role for HC rather than to other hormones of the hypothalamic-pituitary-adrenal axis because feedback regulation is absent in our patients due to their pituitary disease. It would be of interest to study whether receptor sensitivity is of importance for those who appear particularly responsive to HC dose lowering.39 Unfavorable effects of higher doses of GCs are known. Higher doses of GCs are associated with increased cardiovascular risk.3 The effects of GC dose on bone mineral density remain inconclusive, with some studies reporting a linear decrease in bone mineral density with increasing HC doses in men40 while other studies show no difference in bone mineral density between three different doses of HC.41 Since there are no objective parameters to assess the quality of GC replacement therapy, these aspects also need to be taken into account when determining the appropriate dose for a patient. However, it should be noted that improving cardiovascular risk profile and bone strength can be achieved by a number of medical interventions, while improving HRQoL is not otherwise easily managed. Future studies should study the effects of different substitution doses on possible side-effects of GC substitution therapy in RCTs. The strength of our study is its cross-over design, making the study sufficiently powered to detect small to moderate effect sizes. In addition, we measured HRQoL not only at the end of each treatment period but also during the entire study while at the same time documenting temporary dose adjustments. Measuring HRQoL, both 74 Chapter 3

with diaries and retrospectively with questionnaires, resulted in a consistent pattern of complaints. Potential weaknesses include the possibility of a carryover effect inherent to the study design and the number of patients withdrawn from the study. Considering carryover effects, we found no statistical evidence for these effects, except for one subscale of self-reported cognitive functioning. Given that a wash-out period is not feasible in patients with secondary AI because they cannot be left untreated without endangering their safety, carryover effects cannot be fully excluded. However, since there was very little statistical evidence for a carryover effect, we believe this has not influenced our results to any relevant extent. Accordingly, we found that several health-related aspects monitored with diaries during the treatment periods, including the effects on pain, did not change after the first week of treatment. Furthermore, even though the number of withdrawals was slightly higher than suspected, few of these withdrawals were suspected to be related to HC dose. A comparable number of study withdrawals took place in both treatment groups and periods, with the most common reason for withdrawal being protocol violations.16 In conclusion, it is generally recommended to use the lowest dose of HC that relieves symptoms of GC deficiency. This advice in current guidelines remains true only when HRQoL is adequately considered in determining the appropriate dose for the individual patient. Our results show that a higher dose of HC substitution dose will improve patient well-being.

Acknowledgments

The authors would like to thank Judith Rosmalen for her expert advice with regard to the use of daily diary questionnaires. Hydrocortisone Treatment and Quality of Life 75

References

1. Arlt W, Rosenthal C, Hahner S, Allolio B. Quality of glucocorticoid replacement in adrenal insufficiency: clinical assessment vs. timed serum cortisol measurements. Clin Endocrinol (Oxf) 2006;64(4):384-389. 2. Grossman A, Johannsson G, Quinkler M, Zelissen P. Therapy of endocrine disease: Perspectives on the management of adrenal insufficiency: clinical insights from across Europe. Eur J Endocrinol 2013;169(6):R165-75. 3. Filipsson H, Monson JP, Koltowska-Haggstrom M, Mattsson A, Johannsson G. The impact of glu- 3 cocorticoid replacement regimens on metabolic outcome and comorbidity in hypopituitary patients. J Clin Endocrinol Metab 2006;91(10):3954-3961. 4. Mah PM, Jenkins RC, Rostami-Hodjegan A, et al. Weight-related dosing, timing and monitoring hydrocortisone replacement therapy in patients with adrenal insufficiency. Clin Endocrinol (Oxf) 2004;61(3):367-375. 5. Bancos I, Hahner S, Tomlinson J, Arlt W. Diagnosis and management of adrenal insufficiency. Lancet Diabetes Endocrinol 2015;3(3):216-226. 6. Johannsson G, Falorni A, Skrtic S, et al. Adrenal insufficiency: review of clinical outcomes with current glucocorticoid replacement therapy. Clin Endocrinol (Oxf) 2015;82(1):2-11. 7. Tiemensma J, Andela CD, Kaptein AA, et al. Psychological morbidity and impaired quality of life in patients with stable treatment for primary adrenal insufficiency: cross-sectional study and review of the literature. Eur J Endocrinol 2014;171(2):171-182. 8. Hahner S, Loeffler M, Fassnacht M, et al. Impaired subjective health status in 256 patients with adrenal insufficiency on standard therapy based on cross-sectional analysis.J Clin Endocrinol Metab 2007;92(10):3912-3922. 9. Ragnarsson O, Mattsson AF, Monson JP, et al. The relationship between glucocorticoid replacement and quality of life in 2737 hypopituitary patients. Eur J Endocrinol 2014;171(5):571-579. 10. Bleicken B, Hahner S, Loeffler M, et al. Influence of hydrocortisone dosage scheme on health-related quality of life in patients with adrenal insufficiency. Clin Endocrinol (Oxf) 2010;72(3):297-304. 11. Alonso N, Granada ML, Lucas A, et al. Evaluation of two replacement regimens in primary adrenal insufficiency patients. effect on clinical symptoms, health-related quality of life and biochemical parameters. J Endocrinol Invest 2004;27(5):449-454. 12. Wichers M, Springer W, Bidlingmaier F, Klingmuller D. The influence of hydrocortisone substitution on the quality of life and parameters of bone metabolism in patients with secondary hypocortisolism. Clin Endocrinol (Oxf) 1999;50(6):759-765. 13. Behan LA, Rogers B, Hannon MJ, et al. Optimizing glucocorticoid replacement therapy in severely adrenocorticotropin-deficient hypopituitary male patients. Clin Endocrinol (Oxf) 2011;75(4):505- 513. 14. Benson S, Neumann P, Unger N, et al. Effects of standard glucocorticoid replacement therapies on subjective well-being: a randomized, double-blind, crossover study in patients with secondary adrenal insufficiency. Eur J Endocrinol 2012;167(5):679-685. 15. Dullaart RP, Pasterkamp SH, Beentjes JA, Sluiter WJ. Evaluation of adrenal function in patients with hypothalamic and pituitary disorders: comparison of serum cortisol, urinary free cortisol and the human-corticotrophin releasing hormone test with the insulin tolerance test. Clin Endocrinol (Oxf) 1999;50(4):465-471. 76 Chapter 3

16. Werumeus Buning J, Brummelman P, Koerts J, et al. The effects of two different doses of hydrocortisone on cognition in patients with secondary adrenal insufficiency - Results from a randomized controlled trial. Psychoneuroendocrinology 2015;55:36-47. 17. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand 1983;67(6):361-370. 18. Smets EM, Garssen B, Bonke B, De Haes JC. The Multidimensional Fatigue Inventory (MFI) psychometric qualities of an instrument to assess fatigue. J Psychosom Res 1995;39(3):315-325. 19. van der Zee, K. I., Sanderman R. The Measurement of the General Health with the RAND-36, A Handbook. Rijks Universiteit Groningen, Noordelijk Centrum voor Gezondsheidsvraagstukken, Groningen, The Netherlands 2012. 20. Wallace JC, Kass SJ, Stanny CJ. The cognitive failures questionnaire revisited: dimensions and correlates. J Gen Psychol 2002;129(3):238-256. 21. Broadbent DE, Cooper PF, FitzGerald P, Parkes KR. The Cognitive Failures Questionnaire (CFQ) and its correlates. Br J Clin Psychol 1982;21:1-16. 22. Kroenke K, Spitzer RL, Williams JB. The PHQ-15: validity of a new measure for evaluating the severity of somatic symptoms. Psychosom Med 2002;64(2):258-266. 23. Spitzer RL, Kroenke K, Williams JB, Lowe B. A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch Intern Med 2006;166(10):1092-1097. 24. Lowe B, Kroenke K, Herzog W, Grafe K. Measuring depression outcome with a brief self-report instrument: sensitivity to change of the Patient Health Questionnaire (PHQ-9). J Affect Disord 2004;81(1):61-66. 25. Altman D. Practical statistics for medical research. London: Chapman and Hall; 1991. 26. Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates; 1988. 27. Dekkers OM, van der Klaauw AA, Pereira AM, et al. Quality of life is decreased after treatment for nonfunctioning pituitary macroadenoma. J Clin Endocrinol Metab 2006;91(9):3364-3369. 28. Plihal W, Krug R, Pietrowsky R, Fehm HL, Born J. Corticosteroid receptor mediated effects on mood in humans. Psychoneuroendocrinology 1996;21(6):515-523. 29. Hays RD, Morales LS. The RAND-36 measure of health-related quality of life. Ann Med 2001;33(5):350-357. 30. al’Absi M, Petersen KL, Wittmers LE. Adrenocortical and hemodynamic predictors of pain perception in men and women. Pain 2002;96(1-2):197-204. 31. Kuehl LK, Michaux GP, Richter S, Schachinger H, Anton F. Increased basal mechanical pain sensitivity but decreased perceptual wind-up in a human model of relative hypocortisolism. Pain 2010;149(3):539-546. 32. Goodin BR, Smith MT, Quinn NB, King CD, McGuire L. Poor sleep quality and exaggerated salivary cortisol reactivity to the cold pressor task predict greater acute pain severity in a non-clinical sample. Biol Psychol 2012;91(1):36-41. 33. Crettaz B, Marziniak M, Willeke P, et al. Stress-induced allodynia--evidence of increased pain sensitivity in healthy humans and patients with chronic pain after experimentally induced psychosocial stress. PLoS One 2013;8(8):e69460. 34. Choi JC, Chung MI, Lee YD. Modulation of pain sensation by stress-related testosterone and cortisol. Anaesthesia 2012;67(10):1146-1151. 35. Wingenfeld K, Wolf S, Kunz M, Krieg JC, Lautenbacher S. No effects of hydrocortisone and dexamethasone on pain sensitivity in healthy individuals. Eur J Pain 2015;19(6):834-841. Hydrocortisone Treatment and Quality of Life 77

36. Yehuda R, Seckl J. Minireview: Stress-related psychiatric disorders with low cortisol levels: a metabolic hypothesis. Endocrinology 2011;152(12):4496-4503. 37. Jezova D, Vigas M, Hlavacova N, Kukumberg P. Attenuated neuroendocrine response to hypoglycemic stress in patients with panic disorder. Neuroendocrinology 2010;92(2):112-119. 38. Tak LM, Cleare AJ, Ormel J, et al. Meta-analysis and meta-regression of hypothalamic-pituitary- adrenal axis activity in functional somatic disorders. Biol Psychol 2011;87(2):183-194. 39. Manenschijn L, van den Akker EL, Lamberts SW, van Rossum EF. Clinical features associated with glucocorticoid receptor polymorphisms. An overview. Ann N Y Acad Sci 2009;1179:179-198. 40. Zelissen PM. Effect of glucocorticoid replacement therapy on bone mineral density in patients with 3 Addison disease. Ann Intern Med 1994;120(3):207-210. 41. Ragnarsson O, Nystrom HF, Johannsson G. Glucocorticoid replacement therapy is independently associated with reduced bone mineral density in women with hypopituitarism. Clin Endocrinol (Oxf) 2012;76(2):246-252. 78 Chapter 3

Supplemental data

Supplemental materials and methods: Quality of life questionnaires The Hospital Anxiety and Depression Scale (HADS) measures anxiety and depression on a scale ranging from 0 (no symptoms) to 21 (severe symptoms).1 Dutch normative data based on the general population were derived from Spinhoven et al.2 The Multidimensional Fatigue Inventory-20 (MFI-20) records fatigue on five sub- dimensions (general fatigue, physical fatigue, reduced activity, reduced motivation, and mental fatigue), with scores ranging from 4 to 20 on each subscale.3 Higher scores indicate higher levels of fatigue or impairment. Dutch normative data based on the general population were derived from Smets et al.4 The RAND-36, which is identical to the 36-item short-form health survey (SF-36), records general well-being over the preceding four weeks. The SF-36 and RAND- 36 include the same set of items; however, scoring for the general health and bodily pain subscales is slightly different. The questions are organized into eight domains: physical functioning, role limitation due to physical health problems, bodily pain, general health perception, vitality, social functioning, role limitations due to emotional problems, and mental health. The domain scores range from 0–100, with higher scores indicating better QoL.5 Scores on the RAND-36 were standardized using normative data by age for the Dutch population.6 The Cognitive Failures Questionnaire (CFQ) is a 25-item questionnaire measuring self-reported failures in perception, memory, and action in everyday life. The total scores range 0–100, with higher scores reflecting more cognitive problems.7,8 Somatic symptoms were assessed with items of the Patient Health Questionnaire-15 (PHQ-15).9 The items include the most prevalent DSM-IV symptoms of somatization disorder.10 Patients were asked to rate the severity of symptoms over the preceding 24 hours from 1 (“not bothered at all”) to 7 (“bothered a lot”). The symptoms ‘menstrual cramps or other problems with your periods (women only)’ and ‘pain or problems during sexual intercourse’ were deleted because they were not considered relevant for the current study. Two other physical symptoms – ‘feeling tired or having little energy’ and ‘trouble sleeping’ – are also part of the PHQ-9 depression module (see below) and were therefore deleted from the PHQ-15 questionnaire. Consequently, patients were asked to rate the severity of 11 symptoms and total (daily) scores ranged from 11 to 77, with higher scores indicating more severe somatic symptoms. When one of the items was missing, no total daily score could be calculated. An average weekly score was calculated by adding the total daily scores of that week and dividing by the number of days a total daily score was available. The weekly scores of the preceding four weeks were averaged to give a stable measure of severity of symptoms over that treatment period. In addition, the scores on the items ‘stomach pain’, ‘back pain’, ‘joint pain’, Hydrocortisone Treatment and Quality of Life 79

‘headache’, and ‘chest pain’ were summed up in a composite ‘pain’ score, with scores ranging from 5 to 35. The Generalized Anxiety Disorder-7 questionnaire (GAD-7) was used to assess generalized anxiety disorder and symptom severity.11 Patients were asked to rate how much they had been bothered by seven items in the preceding 24 hours. Response options ranged from 1 (“not bothered at all”) to 7 (“bothered a lot”). The total scores ranged from 7 to 49, with higher scores indicating more severe symptoms of anxiety. 3 With regard to missing data and the calculation of a score for symptoms severity of the treatment period, the same procedure as for the PHQ-15 was used. Depression was measured using the PHQ-9 depression module (PHQ-9).12 The nine PHQ-9 depression items correspond to the DSM-IV diagnostic criteria for major depressive disorder.10 Patients were asked to rate the severity of their symptoms of depression for the preceding 24 hours from 1 (“not bothered at all”) to 7 (“bothered a lot”). Total daily scores ranged from 9 to 63, with higher scores indicating more severe depressive symptoms. With regard to missing data and the calculation of a score for symptoms severity of the treatment period, the same procedure as for the PHQ-15 was used. 80 Chapter 3

Supplemental Figure 1. Eligibility, inclusion and follow-up of the patients Hydrocortisone Treatment and Quality of Life 81

               

         3 

                  

      



                    

    

 Supplemental Figure 2. Period effect Mean z-scores per period per group for depressive symptoms, general fatigue and mental fatigue. Group 1 received the low dose in treatment period 1 and the high dose in treatment period 2. Group 2 received the high dose in treatment period 1 and the low dose in treatment period 2. P-value for period effect by Mann-Whitney U-test. 82 Chapter 3

Supplemental Table 1. Weight adjusted doses Low dose condition Weight (kg) BB BL BD Total 50–74 7.5 5.0 2.5 15 75–84 10.0 5.0 2.5 17.5 85–100 10.0 7.5 2.5 20 High dose condition Weight (kg) BB BL BD Total 50–74 15.0 10.0 5.0 30 75–84 20.0 10.0 5.0 35 85–100 20.0 15.0 5.0 40 Dose in mg; BB: before breakfast; BL: before lunch; BD: before diner; Total: cumulative daily dose. Hydrocortisone Treatment and Quality of Life 83

Supplemental Table 2. Clinical characteristics of pituitary patients with adrenal insufficiency (N = 47) Sex (men/women), n 29/18 Age (y), median [IQR] 55 [43; 61] Educational level (1/2/3/4/5/6/7) a 0/2/1/6/21/15/2 Age at diagnosis (y), median [IQR] 31 [20; 46] Childhood onset/Adult onset of SAI, n 6/41 Body weight (kg), median [IQR] 82.5 [72.2; 93.0] 3 BMI (kg/m²), median [IQR] 26.6 [24.5; 29.4]

Surgery (n = 32) Transsphenoidal surgery/Craniotomy (%) 72/28 Age at first surgery (y), median [IQR] 39 [28; 50] Average time since first surgery (y), median [IQR] 11 [6; 20] Patients with 2nd surgery (%) 16 Radiotherapy (n = 19) Pituitary radiotherapy/cranial irradiation/radiotherapy for extracranial 84/11/5 tumors (%) Age at radiotherapy (y), median [IQR] 43 [25; 52] Average time since radiotherapy (y), median [IQR] 12 [9; 22] Hydrocortisone treatment prior to randomization Total daily dose (mg/day), median [IQR] 25 [20; 30] Dose/kg body weight (mg/kg), median [IQR] 0.32 [0.25; 0.35] Number of daily dosages (1/2/3), n 3/33/11 Duration of glucocorticoid treatment (y), median [IQR] 12 [5; 22] No. of hormonal replacements (1/2/3/4/5) 3/9/21/11/3 Thyroid hormone deficiency (% of patients) 92 Growth hormone deficiency (% of patients) 66 Growth hormone deficiency (% of patients receiving substitution) 68 Sex hormone deficiency (% of patients) 57 Men: testosterone (% of patients receiving substitution) 83 Premenopausal women, n = 8: estrogens (% of patients receiving 50 substitution) Postmenopausal women, n = 10: estrogens NA Desmopressin (% of patients) 19 Abbreviations: IQR: Interquartile range, SAI: Secondary adrenal insufficiency, NA: Not applicable. a Educational level was classified using a Dutch education system, comparable to the International Standard Classification of Education (ISCED). This scale ranges from 1 (elementary school not finished) to 7 (university level). 84 Chapter 3

Supplemental Table 3. Quality of life in pituitary patients with secondary adrenal insufficiency in the lower dose and in the higher dose condition (N = 47) Lower dose Higher dose HCa HC P-value* Effect size HADS Anxiety 3.0 [1.0; 5.3] 3.5 [1.0; 5.0]a 0.724 0.0 Depression 3.5 [1.0; 6.0] 2.0 [0.0; 5.0] 0.016b 0.3 MFI-20 General fatigue 11.0 [8.0; 16.0] 10.0 [6.0; 15.0] 0.004b 0.3 Physical fatigue 10.0 [6.0; 13.0] 9.0 [6.0; 12.0] 0.056 0.3 Mental fatigue 10.5 [5.8; 16.0] 8.0 [5.0; 13.0] 0.003b 0.3 Reduced activity 9.5 [7.0; 13.0] 8.0 [6.0; 12.0] 0.170 0.2 Reduced motivation 9.5 [6.0; 12.3] 8.0 [5.0; 12.0] 0.021 0.3 RAND-36 Physical functioning 90 [80; 95] 95 [85; 100] 0.041 0.1 Social functioning 88 [75; 100] 88 [75; 100] 0.599 0.1 Role limitations physical 75 [50; 100] 100 [50; 100]a 0.079 0.3 problems Role limitations emotional 100 [67; 100] 100 [67; 100]a 0.886 0.1 problems Mental health 76 [68; 89] 80 [68; 88] 0.662 0.1 Vitality 65 [45; 71] 70 [50; 80] 0.025 0.3 Bodily pain 80 [67; 100] 90 [67; 100] 0.687 0.1 General health 60 [40; 75] 65 [55; 80] 0.013 0.3 CFQ Memory 8.5 [5.3; 11.0] 8.0 [5.8; 11.3] 0.850 0.0 Distractibility 12.5 [10.0; 17.0] 12.0 [9.0; 16.0] 0.725 0.1 Blunders 9.0 [6.0; 11.0] 7.0 [4.8; 10.3] 0.117 0.2 (Memory for) Names 4.0 [3.0; 6.0] 4.0 [3.0; 5.0] 0.281 0.2 Total CFQ 33.5 [27.3; 42.0] 31.0 [23.0; 40.0] 0.544 0.1 Abbreviations: HC: hydrocortisone, HADS: Hospital Anxiety and Depression Scale, MFI-20: Multidimensional Fatigue Inventory-20, CFQ: Cognitive Failures Questionnaire. Higher HADS scores indicate more anxiety and/or depression, higher MFI-20 scores indicate more fatigue, higher RAND-36 scores indicate better quality of life, and higher CFQ scores indicate more subjective cognitive failures. Data were given as median [interquartile range]. a N = 46 due to missing data; b A significant period effect was found in addition to a significant treatment effect. *P-value lower dose versus higher dose by Wilcoxon Signed Rank Test for paired observations. References

1. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand 1983;67(6):361-370. 2. Spinhoven P, Ormel J, Sloekers PP, Kempen GI, Speckens AE, Van Hemert AM. A validation study of the Hospital Anxiety and Depression Scale (HADS) in different groups of Dutch subjects. Psychol Med 1997;27(2):363-370. 3. Smets EM, Garssen B, Bonke B, De Haes JC. The Multidimensional Fatigue Inventory (MFI) psychometric qualities of an instrument to assess fatigue. J Psychosom Res 1995;39(3):315-325. 4. Smets EM, Visser MR, Willems-Groot AF, Garssen B, Schuster-Uitterhoeve AL, de Haes JC. Fatigue and radiotherapy: (B) experience in patients 9 months following treatment. Br J Cancer 1998;78(7):907-912. 5. Brazier JE, Harper R, Jones NM, et al. Validating the SF-36 health survey questionnaire: new outcome measure for primary care. BMJ 1992;305(6846):160-164. 6. van der Zee, K. I., Sanderman R. The Measurement of the General Health with the RAND-36, A Handbook. Rijks Universiteit Groningen, Noordelijk Centrum voor Gezondsheidsvraagstukken, Groningen, The Netherlands 2012. 7. Broadbent DE, Cooper PF, FitzGerald P, Parkes KR. The Cognitive Failures Questionnaire (CFQ) and its correlates. Br J Clin Psychol 1982;21:1-16. 8. Wallace JC, Kass SJ, Stanny CJ. The cognitive failures questionnaire revisited: dimensions and correlates. J Gen Psychol 2002;129(3):238-256. 9. Kroenke K, Spitzer RL, Williams JB. The PHQ-15: validity of a new measure for evaluating the severity of somatic symptoms. Psychosom Med 2002;64(2):258-266. 10. APA. Diagnostic and Statistical Manual of Mental Disorders DSM-IV-TR. 4th edition ed. Washington DC: American Psychiatric Press; 2000. 11. Spitzer RL, Kroenke K, Williams JB, Lowe B. A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch Intern Med 2006;166(10):1092-1097. 12. Lowe B, Kroenke K, Herzog W, Grafe K. Measuring depression outcome with a brief self-report instrument: sensitivity to change of the Patient Health Questionnaire (PHQ-9). J Affect Disord 2004;81(1):61-66.

Chapter 4

The role of cortisol in the association between perceived stress and pain: a short report on secondary adrenal insuffi cient patients

KAM Janssens J Werumeus Buning AP van Beek JGM Rosmalen

Journal for Person-Oriented Research 2016;2(3):135-141 88 Chapter 4

Abstract

Background Given the pain-dampening effect of cortisol, low cortisol levels have often been proposed to mediate the association between perceived stress and pain. However, studying causal effects is difficult given the complex interplay between perceived stress, cortisol release and pain levels in people with an intact hypothalamic-pituitary-adrenal axis . In the current study, we approached this problem by examining the association between perceived stress and pain levels in patients with secondary adrenal insufficiency (SAI) who were successively supplemented by lower and higher doses of hydrocortisone in a double-blinded cross-over design.

Methods Forty-seven patients with SAI (29 males, 18 females; mean [SD] age, 51 [14] years, range 19–73) participated in an RCT. Patients randomly received low doses of hydrocortisone (0.2–0.3 mg/kg body weight/day) during 10 weeks followed by high doses (0.4–0.6 mg/kg body weight/day) for another 10 weeks, or vice versa. Patients filled out a daily diary on perceived stress (GAD-7) and pain levels (PHQ-15) throughout the RCT. Non-seasonal autoregressive moving average (ARMA) modeling was performed to test associations between daily perceived stress levels and daily pain levels during low and high hydrocortisone dose.

Results Out of 47 study patients, twelve patients showed high enough fluctuations in perceived stress and pain levels (MSSD> 0.01) to study their associations during intake of two different hydrocortisone doses. Six patients showed associations between perceived stress and pain during both hydrocortisone doses, one showed this association only under the low dose, and two only under the high dose.

Conclusion This study does not suggest a role for lower cortisol levels in the association between perceived stress and pain. The Role of Cortisol in the Association between Stress and Pain 89

Introduction

A wide variety of stressors, such as sexual abuse, bullying, and daily hassles have been identified to predict pain levels.1–3 The exact physiological mechanisms behind this association are unknown. Stressors are known to increase perceived stress levels that in turn assert physiological effects, most importantly by increasing levels of the stress hormone cortisol.4 Increases in cortisol levels have been suggested to be in the short run beneficial for pain reduction, since cortisol release is known to suppress pain levels.4 Long term stress-exposure has been suggested to result in down-regulation of the hypothalamic-pituitary-adrenal (HPA) axis and thus in low cortisol levels.5 In turn, 4 low cortisol levels have been found to be related to high pain levels.6 However, it has recently been questioned whether these low cortisol levels are consequences rather than causes of high pain levels.7 Examining cause-effect relationships between perceived stress, cortisol levels and pain in healthy individuals is difficult, since perceived stress, cortisol levels, and pain are highly intertwined and several feedback loops play a role, particularly that of the HPA-axis. Patients with secondary adrenal insufficiency (SAI) are characterized by loss of endogenous adrenocorticotropic hormone (ACTH) and cortisol production and consequently lack the feedback loop of the HPA axis. These patients receive hydrocortisone substitution that mimics normal endogenous cortisol rhythms and compensates for low endogenous cortisol levels.8 Thus, cortisol levels can be externally controlled in these patients. Therefore, patients with SAI serve as a unique model to examine the association between perceived stress levels and pain under different cortisol conditions. This study presents data from a double-blind cross-over RCT in patients with SAI who successively received a lower and higher hydrocortisone dose, in a random order, both administered for a substantial time period. Patients completed a daily diary about perceived stress and pain levels during the entire study. Given the pain suppressing effect of cortisol, we hypothesized that the association between perceived stress and pain would become particularly evident under low hydrocortisone doses.

Methods

Patients This study is part of a double-blinded cross-over RCT.8 A total of 63 patients with SAI was included in the study, of whom 60 completed the run-in phase and the baseline assessment (mean [SD] age 52 [13] years, minimum and maximum age 19–73 years, 35 males, 25 females). All patients fulfilled internationally recognized criteria for SAI, were 18–75 years old, had a body weight of 50–100 kg at screening, had a time interval 90 Chapter 4

between study entry and tumor treatment with surgery and/or radiotherapy of at least one year, and were on adequate and stable replacement of all other pituitary hormone deficiencies for at least six months prior to entry of the study. Main exclusion criteria were inability of legal consent, documented major cognitive impairment, drug abuse or dependence, current psychiatric disorders, treatment for a malignancy, shift work, previous Cushing’s disease, hospital admission during the study, diabetes mellitus with medication known to be able to induce hypoglycemia, and a history of frequent episodes of clinical hypocortisolism. The concomitant use of other corticosteroids and drugs known to interfere with hydrocortisone metabolism, e.g. anti-epileptics, was not allowed either. Sixteen patients did not complete the study, since they violated the study protocol (n = 8), withdrew consent (n = 6), or were withdrawn by the investigator (n = 2). A total 47 patients remained for the analyses (29 men and 18 women, mean [SD] age 51 [14] years, minimum-maximum 19–73 years). Details about the patient group, the inclusion criteria, the exclusion criteria, and the study protocol can be found elsewhere.8

Intervention Patients were randomly divided into two groups. Group 1 first received a lower hydrocortisone dose during 10 weeks (i.e. a cumulative dose of 0.2–0.3 mg/kg body weight, in three divided doses: before breakfast, before lunch, before dinner) followed by a higher hydrocortisone dose during 10 weeks (i.e. a cumulative dose of 0.4–0.6 mg/kg body weight, in three divided doses). Group 2 first received the higher and then the lower hydrocortisone dose. Patients were allowed to double or triple their hydrocortisone intake in case of intercurrent illness or fever for a period of maximum 7 days during the research periods. A doubling of the dose was reported 150 times in patients on low dose and 146 times in patients on high dose.

Daily diaries Patients filled out a daily paper-and-pencil diary at the end of each day during the study period in which they reported among other things about perceived stress and pain levels. Perceived stress levels were assessed with seven items of the GAD-7 (“Feel- ing nervous, anxious or on the edge”, “Not being able to stop or control worrying”, “Worrying too much about different things”, “Trouble relaxing”, “Being so restless that it is hard to sit still”, “Becoming easily annoyed or irritable”, and “Feeling afraid as if something awful might happen”). Pain levels were assessed with five items of the Physical Health Questionnaire (PHQ-15), namely “Stomach pain”, “Back pain”, “Pain in arms, legs or joints”, “Headaches” and “Chest pain”. Patients could endorse whether they experienced these perceived stress or pain symptoms during the past day on a seven-point scale ranging from 1 (‘not at all’) to 7 (‘extremely’). Daily mean The Role of Cortisol in the Association between Stress and Pain 91 item scores for perceived stress and pain were calculated for each patient. Patients completed these diaries during both study periods, for a period of 20 weeks in total, resulting in a maximum of 140 daily observations per patient. Cronbach’s alpha of the GAD-7 in our sample was: 0.92; Cronbach’s alpha of the pain items was 0.72.

Statistical analysis Since previous research indicated that the association between perceived stress and somatic symptoms differs for individuals, analyses were performed at the individual.3 Associations between perceived stress levels and pain levels could only be studied in those individual patients showing high enough daily fluctuations.9 Insight in fluctuation 4 level was obtained by calculating the mean squared successive differences (MSSD). Given the seven-point likert scales used, the MSSD could range from 0 to 36. A score of 36 would be obtained if a participant alternatingly scored 1 and 7 during the entire diary period. Patients that showed a time series with an MSSD of < 0.01, indicating minimal fluctuations for both pain and stress, were excluded. Time series of participants with an MSSD < 0.05 were visually inspected. After the selection of suitable participants, se- rial correlations, also called autocorrelations, within the time series of perceived stress and pain were identified for each individual using non-seasonal autoregressive moving average (ARMA) modeling. To adjust for an increasing or decreasing trend over time and for weekly rhythm, time (date) and day of the week were included in these models if these variables were significant. If time-series showed significant autocorrelation, as indicated by the plots of the autocorrelation function (ACF) and the partial autocorrelation function (PACF) and a significant Ljung-Box test, autoregressive effects as suggested by the ACF and PACF were included in the ARMA model. The best model without autocorrelation was chosen based on the ACF and PACF, a non-significant Ljung-Box test, and BIC values. After running the best identified model, the residuals of this time series, also called “white noise” were saved and this autocorrelation free time series was used for the second step of cross-correlation calculation. If the ACF, PACF and Ljung-Box test did not indicate any form of autocorrelation, the original time series were used. During the second step, the cross-correlation function (CCF) was used to examine cross-correlations between (autocorrelation free) time series of perceived stress and (autocorrelation free) time series of pain. With these analyses, the lag numbers of the cross-correlation and the sign of the association (positive or negative) between perceived stress and pain could be identified for each individual. A lag number is the number of past time points that contains meaningful information on current time point, for example a lag number of 1 means that today’s perceived stress level predict tomorrow’s pain level. All analyses were performed using SPSS Version 22. Associations were considered statistically significant if thep -value was lower than 0.05. 92 Chapter 4

Results

Descriptive statistics Reported daily perceived stress levels and pain levels during the study period were generally low. They were generally higher during low hydrocortisone dose (range stress levels: 1.09 to 3.87, range pain levels: 1.14 to 3.64) than during high hydrocortisone dose (range stress levels: 1.06 to 2.91, range pain levels: 1.09 to 2.66).

Table 1. Differences between patients who show sufficient (n = 12) and insufficient (n = 35) fluc- tation in perceived stress and pain levels Patients showing Patients showing hardly or no fluctuation on fluctuation on pain and stress pain and stress scores (n = 12) scores (n=35) P-value* Sex (male/female), n 7/5 22/13 1.0 Age (years) 41 (14) 54 (13) 0.88 Plasma cortisol LD 1 hr after intake HC 519 [445; 645] 482 [359; 599] 0.34 (nmol/L) Plasma cortisol HD 1 hr after intake HC 819 [676; 888] 735 [667; 872] 0.39 (nmol/L) Free cortisol LD CL (L/h/70kg) 184 [127; 298] 204 [146; 268] 0.75 Free cortisol HD CL (L/h/70kg) 169 [156; 240] 206 [153; 265] 0.88 Free cortisol LD AUC 24hr (h*nmol/L) 223 [149; 325] 202 [144; 259] 0.58 Free cortisol HD AUC 24hr (h*nmol/L) 478 [340; 513] 382 [308; 518] 0.43 Total cortisol LD CL (L/h/70kg) 11 [8; 16] 10 [7; 14] 0.53 Total cortisol HD CL (L/h/70kg) 13 [10; 17] 12 [10; 15] 0.66 Total cortisol LD AUC 24hr (h*nmol/L) 3795 [2776; 5122] 3907 [2843; 5343] 0.86 Total cortisol HD AUC 24hr (h*nmol/L) 6235 [5260; 8095] 6559 [5299; 7553] 0.97 Pain levels LD 1.99 (0.79) 1.24 (0.24) 0.001 Pain levels HD 1.78 (0.59) 1.16 (0.18) < 0.001 Stress levels LD 1.76 (0.76) 1.11 (0.25) < 0.001 Stress levels HD 1.66 (0.53) 1.10 (0.21) < 0.001 Abbreviations: CL: clearance, AUC: area under the curve, HC: hydrocortisone, LD: Lower dose, HD: higher dose. Data are frequency, mean (SD) or median [interquartile range]. * P-values were based on Χ2-tests for dichotomous variables and Mann-Whitney U tests for continuous variables. The Role of Cortisol in the Association between Stress and Pain 93

Association between perceived stress and pain levels Thirty-two participants were excluded, since at least one of their time series showed insufficient fluctuations, as indicated by a MSSD of < 0.01. Time series of six participants with a MSSD of < 0.05 were visually inspected to determine their suitability. Three out of these six patients showed patterns that were not suitable for analyses

Table 2a. Association between perceived stress and pain during low hydrocortisone dose Cross- Autocor- correlation Mean (SD) MSSD relation Autocor- perceived 4 Patient perceived perceived perceived Mean (SD) MSSD relation stress and letters stress stress stress pain pain pain pain A 2.17 (0.47) 0.27 AR(1)a 3.05 (0.84) 0.65 AR(1)a 1: r = 0.30 sd = 0.12; 4: r = 0.27 sd = 0.13 B 1.35 (0.32) 0.11 AR(1) 1.37 (0.33) 0.14 AR(1,4) 0: r = 0.49 sd = 0.12 C 1.53 (0.19) 0.04 AR(1)a 2.09 (0.47) 0.25 AR(1,5) 0: r = 0.30 sd = 0.12 -2: r = 0.37, sd = 0.12 D 1.59 (0.36) 0.15 AR(1) a,b 2.36 (0.53) 0.22 AR(1,7) a NS E 1.15 (0.21) 0.06 NSa 1.32 (0.31) 0.15 AR(2) b NS F 1.88 (0.28) 0.11 AR(1) 1.83 (0.35) 0.07 AR(1,4,6) 0: r = 0.43, sd = 0.12 G 3.87 (0.74) 0.31 MA(1) a 3.64 (0.58) 0.41 AR(1,2,3) NS H 1.74 (0.68) 0.69 AR(1) 2.64 (0.60) 0.49 AR(1) 2: r = 0.25, sd = 0.12 I 2.30 (0.31) 0.18 NS 2.02 (0.31) 0.13 AR(7) a NS J 1.16 (0.20) 0.04 AR 1.19 (0.23) 0.06 AR(1) -5: r = 0.30, (1,5,9) a sd = 0.13; 2: r = 0.25, sd = 0.13 K 1.09 (0.17) 0.06 NS a,b 1.14 (0.16) 0.05 NS a,b NS L 1.25 (0.28) 0.08 AR(1) 1.25 (0.27) 0.13 NS a -4: r = 0.26, sd = 0.13 Abbreviations: MSSD: mean squared successive differences, AR: autoregressive effect, e.g. AR(1,4) means that today’s perceived stress level predict tomorrow’s perceived stress level and the perceived stress level 4 days later, MA: moving average, e.g. MA(1) means that the sd of today’s stress level predict tomorrow’s stress level; aARMA model adjusted for time (date); bARMA model adjusted for weekly rhythm (day of week); NS: non-significant p( < 0.05). 94 Chapter 4

Table 2b. Association between perceived stress and pain during high hydrocortisone dose Cross- Autocor- correlation Mean (SD) MSSD relation Autocor- perceived Patient perceived Perceived perceived Mean (SD) MSSD relation stress and letters stress stress stress pain Pain pain pain A 2.18 (0.31) 0.14 NS a 2.53 (0.47) 0.17 AR(1) -1: r = 0.24, sd = 0.12 B 1.44 (0.38) 0.16 AR(1,3) 1.37 (0.32) 0.13 AR(1) a 0: r = 0.37, sd = 0.12 C 1.37 (0.15) 0.03 AR(1,2) 1.75 0.29 AR(1) 0: r = 0.56, (0.45) sd = 012 2: r = 0.27, sd = 0.12 D 1.64 (0.27) 0.10 AR(7) a 2.33 (0.43) 0.16 AR(1) 0: r = 0.28, sd = 0.12 -1: r = 0.25, sd = 12 E 1.32 (0.26) 0.12 AR(4) 1.24 (0.28) 0.10 NS NS F 1.57 (0.29) 0.11 AR(1) 1.72 (0.42) 0.12 AR(1) 0: r = 0.33, sd = 0.12; 3: r = 0.27, sd = 0.12 G 2.91 (0.61) 0.30 AR(1) a 2.66 (0.43) 0.22 AR(1) a 0: r = 0.37, sd = 0.12 2: r = 0.28, sd = 0.14 H 2.01 (0.43) 0.25 AR(1) 2.53 (0.41) 0.27 AR(1) NS I 2.12 (0.47) 0.24 AR(1) 1.68 (0.42) 0.29 AR(2)b 0: r = 0.26, sd = 0.13 J 1.06 (0.12) 0.03 NS 1.09 (0.21) 0.07 AR(2) -1: r = 0.32, sd = 0.12 2: r = 0.33, sd = 0.13 K 1.07 (0.10) 0.02 NS b 1.11 (0.15) 0.04 NS b NS L 1.38 (0.41) 0.30 NS a,b 1.26 (0.37) 0.26 NS -7: r = 0.26, sd = 0.13 Abbreviations: MSSD: mean squared successive differences, AR: autoregressive effect, e.g. AR(1,4) means that today’s perceived stress level predicts tomorrow’s perceived stress level and perceived stress level 4 days later, MA: moving average, e.g. MA(1) means that the sd of today’s stress level predict tomorrow’s stress level; a ARMA model adjusted for time (date); b ARMA model adjusted for weekly rhythm (day of week); NS: non-significant p( < 0.05). The Role of Cortisol in the Association between Stress and Pain 95

(i.e., more than 20 of the same observations in a row). So, ultimately twelve subjects showed enough fluctuation in the time series to allow us to study the association between perceived stress and pain during both hydrocortisone conditions. Characteristics of these 12 patients included in present analysis and the 35 patients that were excluded based on insufficient fluctuation, are shown inTable 1. Patients with sufficient fluctuation in pain and stress level did not differ from persons without sufficient fluctuation with regard to age, sex, and hydrocortisone levels. However, patients with sufficient fluctuation levels showed higher pain and stress levels during both low and high hydrocortisone dose than patients without sufficient fluctuation, which probably contributed to their higher fluctuation patterns. Mean pain levels, perceived stress levels, MSSD, 4 information about autocorrelation and cross-correlation is given in Table 2a for low hydrocortisone dose and in Table 2b for high hydrocortisone dose. Six patients showed associations between perceived stress and pain during both hydrocortisone doses, one showed this association only under low dose, and two only under high doses. Six positive contemporaneous associations (i.e. lag number is 0) were found, which means that higher perceived stress levels were associated with higher pain levels on the same day. In seven cases, higher pain levels predicted higher perceived stress levels on subsequent days (negative lags), and in eight cases higher perceived stress predicted higher pain levels on subsequent days (positive lags).

Discussion

This double-blinded cross-over trial in patients with SAI found that the association between perceived stress levels and pain levels was comparable during supplementation with lower and higher hydrocortisone doses. Therefore, this study does not provide evidence for a prominent role of cortisol in the association between perceived stress and pain. One strong point of this study is the unique patient group in which we were able to study the role of cortisol in the association between perceived stress and pain without influence of HPA-axis feedback loops. Moreover, the combined diary and RCT approach offered the unique opportunity to study these effects on an individual level during daily life. A weakness of this study is that fluctuations in stress and pain levels were on average low during the study periods, which allowed us to study the association between perceived stress and pain only in a subset of patients. The small sample size might have impaired the generalizability of our findings. It is good to note that SAI patients that showed enough fluctuation to be included were comparable with regard to age, sex and cortisol levels to SAI patients that did not show enough fluctuation. Further, SAI patients of the RCT were believed to be a representative sample of the total population of patients with SAI.8 96 Chapter 4

Nevertheless, findings could have been different for other SAI samples. Moreover, inher- ently to the design, SAI patients were studied and one has to be careful with generalizing the result, i.e. a lack of moderation effect of cortisol in the relation between stress and pain, to healthy persons. Further, the levels of perceived stress and pain experienced during the research period were in general also low. Another limitation was that the GAD-7 was used as a proxy to assess perceived stress. The GAD-7 is not specifically validated for measurement of perceived stress, but has a high face validity for the assessment of stress given the content of the items, especially when administered daily. A final potential limitation was that participants were allowed to incidentally double their hydrocortisone dose in case of acute stress or illness. Such a doubling was allowed during a maximum of 10 per cent of the study period and patients were strictly informed not to double their hydrocortisone intake the week prior to their hospital visit (so in the week before they switched to the other hydrocortisone dose).8 A hydrocortisone dose was doubled seventy-six times during low hydrocortisone intake and fifteen times during high hydrocortisone intake. The two patients that were mainly responsible (65 %) for the doubling of hydrocortisone intake during low hydrocortisone dose, showed significant associations between perceived stress and pain during both the period in which they were on the lower dose and the period in which they were on the higher dose. Therefore, we do not expect that the incidental doubling of hydrocortisone (substantially) affected our findings. This is to the best of our knowledge the first study that examined the effect of cortisol on the association between perceived stress and pain in a diary design during a longer treatment period. We did not find a moderation effect of hydrocortisone in the association between perceived stress and pain. Although a negative finding can have the trivial explanation that the power to detect differences is too low, we do not believe that this is the case in our study. In idiographic research the power to detect an association is mainly determined by the number of data points for each individual and not by the number of participants.9 The number of data points in this study was sufficient to show significant associations between stress and cortisol, as significant associations were found in 10 out of 12 participants. Therefore, this study was sufficiently powered to show differences in the association between perceived stress and pain during low and high hydrocortisone dose in individual patients. Our findings are in line with a recent general population cohort study that did also not found that a dysregulated HPA-axis could predict widespread pain onset in relation to stressful life events.7 It should be kept in mind that our study tested the effect of hydrocortisone dose on the association between perceived stress and pain during normal daily life, and not during a stressful period, while the effect of hydrocortisone dose on the association between perceived stress and pain is likely to be particularly evident during stressful life periods. Therefore, repetition of the current study design with introduction of a mild stress experiment during both hydrocortisone doses is suggested. The Role of Cortisol in the Association between Stress and Pain 97

References

1. Gini G, Pozzoli T. Association between bullying and psychosomatic problems: a meta-analysis. Pediatrics 2009;123(3):1059-1065. 2. Paras ML, Murad MH, Chen LP, et al. Sexual abuse and lifetime diagnosis of somatic disorders: a systematic review and meta-analysis. JAMA 2009;302(5):550-561. 3. van Gils A, Burton C, Bos EH, Janssens KA, Schoevers RA, Rosmalen JG. Individual variation in temporal relationships between stress and functional somatic symptoms. J Psychosom Res 2014;77(1):34-39. 4. Hannibal KE, Bishop MD. Chronic stress, cortisol dysfunction, and pain: a psychoneuroendocrine rationale for stress management in pain rehabilitation. Phys Ther 2014;94(12):1816-1825. 5. Miller GE, Chen E, Zhou ES. If it goes up, must it come down? Chronic stress and the hypothalamic- pituitary-adrenocortical axis in humans. Psychol Bull 2007;133(1):25-45. 4 6. Tak LM, Cleare AJ, Ormel J, et al. Meta-analysis and meta-regression of hypothalamic-pituitary- adrenal axis activity in functional somatic disorders. Biol Psychol 2011;87(2):183-194. 7. Generaal E, Vogelzangs N, Macfarlane GJ, et al. Biological stress systems, adverse life events and the onset of chronic multisite musculoskeletal pain: a 6-year cohort study. Ann Rheum Dis 2015. 8. Werumeus Buning J, Brummelman P, Koerts J, et al. The effects of two different doses of hydrocortisone on cognition in patients with secondary adrenal insufficiency - Results from a randomized controlled trial. Psychoneuroendocrinology 2015;55:36-47. 9. van Ockenburg SL, Booij SH, Riese H, Rosmalen JG, Janssens KA. How to assess stress biomarkers for idiographic research? Psychoneuroendocrinology 2015;62:189-199.

Part 2

The effect of hydrocortisone treatment on somatic outcome measures in patients with secondary adrenal insufficiency

Chapter 5

Somatosensory function in patients with secondary adrenal insuffi ciency treated with two diff erent doses of hydrocortisone – Results from a randomized controlled trial

J Werumeus Buning KH Konopka P Brummelman J Koerts RPF Dullaart G van den Berg MM van der Klauw O Tucha BHR Wolff enbuttel AP van Beek

Submitted 102 Chapter 5

Abstract

Background Low cortisol levels are associated with several functional pain syndromes. In patients with secondary adrenal insufficiency (SAI), the lack in endogenous cortisol production is substituted by the administration of oral hydrocortisone (HC). Our previous study showed that a lower dose of HC led to an increase in reported subjective pain symptoms. Whether different doses of HC substitution alter somatosensory functioning in SAI patients has not been established yet.

Methods In this randomized double blind cross-over trial, forty-six patients with SAI participated. Patients randomly received either first a lower dose (0.2–0.3 mg HC/kg body weight/day) for 10 weeks followed by a higher dose (0.4–0.6 mg HC/kg body weight/ day) for another 10 weeks, or vice versa. After each treatment period, blood samples were drawn and somatosensory functioning was assessed by determining the mechanical detection threshold (MDT), mechanical pain threshold (MPT), mechanical pain sensitivity (MPS) and the pain pressure threshold (PPT), according to the Quantitative Sensory Testing (QST) battery by the German Network on Neuropathic Pain.

Results The administration of the higher dose of HC resulted in significantly higher levels of cortisol (mean [SD] 748 [245] nmol/L) than the lower dose (537 [250] nmol/L, P< 0.001). No differences were found in MDT, MPT, MPS and PPT z-scores between the two doses of HC. Furthermore, the number of patients showing sensory abnormalities did not differ between the two different doses.

Conclusions The results suggest that the dose of HC has no impact on somatosensory functioning in response to mechanical stimuli in patients with SAI, despite previously found altered subjective pain reports. Hydrocortisone Dose and Somatosensory Functioning 103

Introduction

Stress hormones are known to play an important role in stress-related bodily disorders, such as functional pain syndromes. Stress, i.e. internal or external threats to the maintenance of homeostasis, activates the hypothalamic-pituitary-adrenal (HPA) axis in order to reestablish homeostasis. Activation of the HPA axis results among others in the secretion of cortisol. There is some evidence that cortisol increases the tolerance of pain.1,2 Several functional pain syndromes are linked to hypocortisolism.1,3 For instance, patients with fibromyalgia showed reduced 24 h-urinary cortisol excretion compared to healthy controls.4–6 On the other hand, it was found that higher cortisol levels are associated with decreased pain thresholds7 and increased pain sensitivity in response to thermal stimuli.8 Other studies were unable to find an effect of cortisol levels and pain sensitivity.9 Thus, the relationship between cortisol levels and pain 5 sensitivity remains unclear. The relationship between cortisol levels and sensory functioning using mechanical stimuli has been studied in healthy volunteers.10 However, in healthy individuals the negative feedback mechanisms of the HPA axis makes it difficult to study the relationship between cortisol levels and sensory functioning. Patients with secondary adrenal insufficiency (SAI) are characterized by a loss of endogenous cortisol production due to impaired hypothalamus or pituitary functioning, thereby lacking this negative feedback mechanism. These patients receive substitution therapy, usually oral tablets of hydrocortisone (HC). Because the negative feedback mechanisms is absent in patient with SAI, cortisol levels can be controlled externally, making them a unique group of patients to study the direct relationship between cortisol concentrations and somatosensory function. As shown previously by our group, patients with SAI reported more pain symptoms as assessed by daily diaries during treatment with a lower dose of HC (total daily doses ranging from 15–20 mg HC) when compared to treatment with a higher dose of HC (total daily doses ranging from 30–40 mg HC).11 However, it is unknown whether these two substitution doses also influence somatosensory function. This study is part of a recently conducted randomized double blind cross-over study which assessed the effect of two different doses of HC on cognition,12 health related quality of life,11 blood pressure,13 pharmacokinetics and somatosensory functioning. Sensory function was assessed using parts of the German Research Network on Neuropathic Pain (DFNS) Quantitative Sensory Testing (QST) battery which assesses the perceptual response to systematically applied and quantifiable sensory stimuli.14,15 QST has been shown to be valid, reliable and sensitive in quantifying sensory abnormalities.16–18 It is hypothesized that patients show increased sensory functioning as assessed by mechanical stimuli after receiving a low dose of HC in comparison to after receiving a high dose of HC. 104 Chapter 5

Materials and Methods

Patients This study is part of a randomized double blind cross-over study.12 Patients with SAI were recruited from the endocrine outpatient clinic at the University Medical Center Groningen (UMCG) between May 2012 and June 2013. Details about eligibility, recruit- ment and follow-up are shown in Fig 1. All patients were diagnosed with SAI according to internationally accepted criteria. Other inclusion criteria were age between 18 and 75 years, body weight between 50 and 100 kg, at least one year between treatment for pituitary disorder (surgery and/or radiotherapy) and study entry and stable replacement of other pituitary hormonal deficiencies for at least 6 months prior to study entry and during the study periods (thyroid hormone deficiency, growth hormone deficiency, testosterone/ estradiol deficiency, diabetes insipidus).12 Main exclusion criteria were major cognitive impairment, drug abuse or dependence, current psychiatric disorders, malignancy, shift work, previous Cushing’s disease, hospital admission, medically treated diabetes mellitus potentially leading to hypoglyceamia (insulin, sulfonylurea derivatives), a history of frequent episodes of hypocortisolism and anti-epileptic drugs.12 Forty-seven patients completed all testing days, QST data was available for 46 patients. Patients were allowed to use paracetamol or NSAIDs during the study. In total six patients reported the incidental use of these drugs but not on the test day and one participant reported the daily use of paracetamol/codeine. One patient used transdermal fentanyl and was excluded from the present analysis, resulting in a total patient group of 45 patients for the present analysis. The study protocol was approved by the local ethic committee at the UMCG, The Netherlands. Patients gave written informed consent before entering the study. This study is registered with Clinicaltrials.gov, number NCT01546922.

Intervention and protocol Patients were randomly assigned to either group 1 or group 2. Group 1 received first a low dose of HC (total daily dose of 0.2–0.3 mg/kg body weight) for 10 weeks, followed by a high dose of HC (total daily dose of 0.4–0.6 mg/kg body weight) for another 10 weeks. Group 2 received the two doses in reversed order. Randomization to one of the two treatment groups was performed by Tiofarma Inc, The Netherlands, with a block size of 4. The total daily dose was divided into three doses, administered before breakfast, before lunch and before dinner, with the highest dose given in the morning. After each treatment period patients returned to the hospital for evaluation. On testing days, patients were instructed to take their morning dose at 0700 hour. At 0800 hour the first blood samples were drawn in a fasting state in sitting position after a short period of rest, followed by physical, neuropsychological and sensory functional examination. Approximately five hours after the morning intake of HC a second blood sample was drawn. Hydrocortisone Dose and Somatosensory Functioning 105

Fig 1. Patient flow diagram 106 Chapter 5

Laboratory measurements Total cortisol concentrations in plasma was measured by isotope dilution liquid chromatography tandem mass spectrometry (LC-MS/MS) and was performed essentially 19 as described by Hawley et al., using cortisol-13C3 as internal standard. Intra- and inter-assay CVs were < 2.6 % and < 5.3 %, respectively.

Quantitative sensory testing The QST battery utilized in the present study consisted of four tests and was applied on the hand (dorsum) according to the standardized protocol of Rolke et al.14 The QST battery was performed as part of an extensive neuropsychologic assessment battery. Not every QST parameter was thought to be relevant for the present patient group, for example dynamic mechanical allodynia and wind up ratio are used to assess central components of sensory function in neuropathic pain which was not expected to be present in the present patient population. In addition, in order not to exceed the time constraint of four hours of testing, only the following subset of mechanical tests of the QST battery were used: The mechanical detection threshold (MDT) was measured with a standardized set

of modified von Frey hairs (Optihair2-Set, Marstock Nervtest, Germany). Using the ‘method of limits’, five ascending and five descending series of stimulus intensities were applied, resulting in ten thresholds. The final threshold is the geometric mean of these ten thresholds. The mechanical pain threshold (MPT) was measured using a set of seven pinprick devices (flat contact area of 0.2 mm in diameter) with fixed stimulus intensities that exerted forces of 8, 16, 32, 64, 128, 256 and 512 mN. Using the ‘method of limits’, patients were instructed to indicate whether or not they perceived a sharp sensation. The final threshold was determined in equivalence to the procedure described for MDT, i.e. the geometric mean of ten thresholds. Mechanical pain sensitivity (MPS) was assessed using the same set of seven pinprick stimuli to obtain a stimulus-response function for pinprick-evoked pain. Five runs were performed and in each run the pinpricks were applied in balanced and pseudo-randomized order. Patients were asked to give a pain rating for each stimulus on a numerical rating scale ranging from ‘0’ (“no pain”) to ‘100’ (“most intense pain imaginable”). MPS was calculated as the geometric mean of all numerical ratings for pinprick stimuli. The pressure pain threshold (PPT) was determined over muscle with a pressure gauge device (FDN200, Wagner Instruments, USA) with three series of ascending stimulus intensities, each applied as a slowly increasing ramp of 50 kPa/s (0.5 kg/cm2 s). The final threshold was the geometric mean of three stimulations.

Statistics Details about sample size calculation have been described elsewhere.12 QST data were transformed into z-scores as described by Rolke et al.,14 using age- and sex-specific Hydrocortisone Dose and Somatosensory Functioning 107 reference values obtained in healthy volunteers.20 A z-score below -1.96 or above 1.96 was considered a sensory abnormality. Normality of data was analyzed using Q-Q plots. Since the data were not normally distributed, differences in z-scores on the QST parameters between the two doses of HC were assessed using the non-parametric Wilcoxon Signed Rank Test for paired observations. In addition, Cohen’s d effect sizes were calculated to give a measure of the magnitude of the difference. An effect size of d = 0.2 is considered to be a small effect, d = 0.5 a moderate effect, and d = 0.8 a large effect.21 Furthermore, the number of patients showing a sensory abnormality on each dose of HC were compared using the McNemar test. The two-tailed significance level of < 0.05 was used. All statistical analyses were performed using Statistical Package for the Social Sciences (SPSS, Inc., Armonk, NY, USA), version 22.

Table 1. Clinical characteristics at baseline of pituitary patients treated for adrenal insufficiency 5 (N = 45) Sex (men/women), n 27/18 Age (y), median [IQR] 55 [43; 60] Age at diagnosis (y), median [IQR] 31 [20; 45] Body weight (kg), median [IQR] 81.5 [71.9; 93.1] Hydrocortisone treatment prior to randomization Total daily dose (mg/day), median [IQR] 25 [20; 30] Dose/kg body weight (mg/kg), median [IQR] 0.31 [0.25; 0.35] Number of daily dosages (1/2/3), n 2/33/10 Duration of glucocorticoid treatment (y), median [IQR] 12 [6; 24] No. of hormonal replacements (1/2/3/4/5), n 3/9/20/11/2 Thyroid hormone deficiency (% of patients) 91 Growth hormone deficiency (% of patients) 67 Growth hormone deficiency (% of patients receiving substitution) 67 Sex hormone deficiency (% of patients) 56 Men: testosterone (% of patients receiving substitution) 78 Premenopausal women, n = 8: estrogens (% of patients receiving 50 substitution) Postmenopausal women, n = 10: estrogens NA Desmopressin (% of patients) 18 Abbreviations: IQR: Interquartile range, NA: Not applicable. 108 Chapter 5

Results

Patients Reporting of the study conforms to the CONSORT 2010 statement.22 Baseline characteristics are given in Table 1. A detailed description of the patient group and reasons for withdrawal can be found elsewhere.12

Cortisol levels One hour after the administration of the morning dose, the higher dose of HC resulted in significantly increased plasma cortisol levels compared to the lower dose of HC (median [IQR], 741 [669; 870] nmol/L and 499 [383; 605] nmol/L for the higher and the lower dose respectively, P < 0.001). A similar difference was found five hours after administration of the morning dose (235 [170; 314] nmol/L and 112 [75; 199] nmol/L for the higher and the lower dose respectively, P < 0.001).

Quantitative sensory testing There were no differences in z-scores on the MDT, MPT, MPS and the PPT between the two doses of HC (Fig 2). Effect sizes ranged between 0.01 and 0.11. Furthermore, the number of patients showing at least one sensory abnormality did not differ between the lower and the higher dose of HC (29 % of patients on the lower dose versus 27 % of patients on the higher dose).

Fig 2. Z-scores for the mechanical detection threshold, mechanical pain threshold, mechanical pain sensitivity and pressure pain threshold on the lower dose and the higher dose (N = 45). Abbreviations: HC: hydrocortisone; MDT: mechanical detection threshold; MPT: mechanical pain threshold; MPS: mechanical pain sensitivity; PPT: pressure pain threshold. Data are mean ± SEM. Hydrocortisone Dose and Somatosensory Functioning 109

Discussion

In this randomized double blind cross-over study, the effect of dose of HC on mechanical pain perception were studied. The administration of two different doses of HC for a period of 10 weeks did not result in a difference in mechanical QST parameters. This suggests that dose of HC does not alter mechanical pain perception in patients with SAI. Other studies using drug administration to investigate the relationship between cortisol and pain sensitivity have revealed conflicting results. Some studies reported an association between cortisol administration and reduced pain,23 whereas others did not find altered pain sensitivity and pain thresholds after administration of exogenous cortisol.9,24 Several studies used drugs to suppress the endogenous cortisol production to establish a state of hypocortisolism. Dexamethasone administration led to 5 lowered pain thresholds.25 However, next to decreased cortisol levels, dexamethasone administration leads to decreased ACTH or CRH, which could also result in reduced pain thresholds. Furthermore, reduced mechanical pain detection thresholds and enhanced temporal summation of pressure pain were reported in a hypocortisolemic state after oral administration of metyrapone.10 However, the decrease in cortisol was accompanied by an increase in adrenocorticotropic hormone (ACTH) and possibly by other opioid proopiomelanocortin-derived peptides (including β-endorphin).26 Hypo- cortisolism and concomitant endorphin hyper-secretion can lead to hyperalgesia due to both insufficient anti-inflammatory glucocorticoid action and chronic opioid-induced activation of pro-inflammatory neuro-immune responses,27 therefore these peptides and mechanisms also might have played a role in the relationship between hypocortisolism and increased sensitivity of pain as found by Kuehl et al., (2010). Furthermore, Kuehl et al., (2010) assessed somatosensory function in healthy individuals, making a direct comparison with the results of the patient population of the present study difficult. Despite the previously observed difference in subjective pain reports as assessed by daily diaries comparing both doses of HC,11 sensory changes supportive for HC induced alteration of mechanical function could not be established in this study. This methodological difference might underlie the discrepancy in findings. Previously, a poor relationship between clinical pain and pain thresholds measured with QST was reported.28 A possible explanation for this poor association is that both methods measure different constructs of pain and may not be directly related. QST is a psychometric measure of evoked pain. However, mechanically induced pain is not necessarily the same as clinical pain in which patients refer to ongoing pain without an evocative stimulus. For example, in people with osteoarthritis of the knee it was shown that during spontaneous pain, i.e. pain experienced without stimulation, other brain areas were activated than brain areas that were active during evoked pain.29 This implies that 110 Chapter 5

by measuring evoked pain, such as with mechanical stimuli, a differently constructed pain is experienced by patients without such a stimulation. Furthermore, even though the administration of the two different doses of HC in the present study resulted in significant and clinically relevant differences in plasma cortisol levels, the overall cortisol concentrations remained within the physiological range. In contrast, the state of hypocortisolism aimed for in the study by Kuehl et al. can be considered subphysi- ological.10 This conceptual difference in study design could be an explanation for the lack of differences in QST parameters in our patient group. A strength of the present study is the unique patient population enabling a comparison of the direct relation between cortisol levels and mechanical pain perception by excluding potential interference of the negative feedback mechanism of the HPA axis. A limitation is the small number of patients, which should be considered when interpreting the results of this study. However, given the negligible effect sizes, we do not expect to find different results in a larger sample. Furthermore, 10 weeks of treatment might be too short to find dose-dependent effects of HC on mechanical pain perception. However, the study by Kuehl et al. (2010) showed significant differences in sensory perception as soon as six hours after induced hypocortisolism, so a lack of differences due to a too short treatment period is in our opinion unlikely. To the best of our knowledge, this is the first study examining the effect of two different doses of HC on mechanical parameters in patients with SAI. Our results indicate that the dose of HC has no impact on mechanical pain perception in patients with SAI, even though a difference in subjectively experienced pain symptoms was reported previously by the same patient group.11 Hydrocortisone Dose and Somatosensory Functioning 111

References

1. Heim C, Ehlert U, Hellhammer DH. The potential role of hypocortisolism in the pathophysiology of stress-related bodily disorders. Psychoneuroendocrinology 2000;25(1):1-35. 2. Chrousos GP. The hypothalamic-pituitary-adrenal axis and immune-mediated inflammation. N Engl J Med 1995;332(20):1351-1362. 3. Tak LM, Cleare AJ, Ormel J, et al. Meta-analysis and meta-regression of hypothalamic-pituitary- adrenal axis activity in functional somatic disorders. Biol Psychol 2011;87(2):183-194. 4. McCain GA, Tilbe KS. Diurnal hormone variation in fibromyalgia syndrome: a comparison with rheumatoid arthritis. J Rheumatol Suppl 1989;19:154-157. 5. Crofford LJ, Pillemer SR, Kalogeras KT, et al. Hypothalamic-pituitary-adrenal axis perturbations in patients with fibromyalgia. Arthritis Rheum 1994;37(11):1583-1592. 6. Griep EN, Boersma JW, Lentjes EG, Prins AP, van der Korst JK, de Kloet ER. Function of the hypothalamic-pituitary-adrenal axis in patients with fibromyalgia and low back pain. J Rheumatol 1998;25(7):1374-1381. 7. Choi JC, Chung MI, Lee YD. Modulation of pain sensation by stress-related testosterone and corti- 5 sol. Anaesthesia 2012;67(10):1146-1151. 8. Crettaz B, Marziniak M, Willeke P, et al. Stress-induced allodynia--evidence of increased pain sensitivity in healthy humans and patients with chronic pain after experimentally induced psychosocial stress. PLoS One 2013;8(8):e69460. 9. Wingenfeld K, Wolf S, Kunz M, Krieg JC, Lautenbacher S. No effects of hydrocortisone and dexamethasone on pain sensitivity in healthy individuals. Eur J Pain 2015;19(6):834-841. 10. Kuehl LK, Michaux GP, Richter S, Schachinger H, Anton F. Increased basal mechanical pain sensitivity but decreased perceptual wind-up in a human model of relative hypocortisolism. Pain 2010;149(3):539-546. 11. Werumeus Buning J, Brummelman P, Koerts J, et al. Hydrocortisone Dose Influences Pain, Depres- sive Symptoms and Perceived Health in Adrenal Insufficiency - A Randomized Controlled Trial. Neuroendocrinology 2016;103(6):771-8. 12. Werumeus Buning J, Brummelman P, Koerts J, et al. The effects of two different doses of hydrocortisone on cognition in patients with secondary adrenal insufficiency - Results from a randomized controlled trial. Psychoneuroendocrinology 2015;55:36-47. 13. Werumeus Buning J, van Faassen M, Brummelman P, et al. Effects of Hydrocortisone on the Regula- tion of Blood Pressure: Results From a Randomized Controlled Trial. J Clin Endocrinol Metab 2016;101(10):3691-3699. 14. Rolke R, Baron R, Maier C, et al. Quantitative sensory testing in the German Research Network on Neuropathic Pain (DFNS): standardized protocol and reference values. Pain 2006;123(3):231-243. 15. Rolke R, Magerl W, Campbell KA, et al. Quantitative sensory testing: a comprehensive protocol for clinical trials. Eur J Pain 2006;10(1):77-88. 16. Felix ER, Widerstrom-Noga EG. Reliability and validity of quantitative sensory testing in persons with spinal cord injury and neuropathic pain. J Rehabil Res Dev 2009;46(1):69-83. 17. Freynhagen R, Rolke R, Baron R, et al. Pseudoradicular and radicular low-back pain--a disease con- tinuum rather than different entities? Answers from quantitative sensory testing. Pain 2008;135(1- 2):65-74. 18. Geber C, Klein T, Azad S, et al. Test-retest and interobserver reliability of quantitative sensory testing according to the protocol of the German Research Network on Neuropathic Pain (DFNS): a multi-centre study. Pain 2011;152(3):548-556. 112 Chapter 5

19. Hawley JM, Owen LJ, MacKenzie F, Mussell C, Cowen S, Keevil BG. Candidate Reference Mea- surement Procedure for the Quantification of Total Serum Cortisol with LC-MS/MS. Clin Chem 2016;62(1):262-269. 20. Konopka KH, Harbers M, Houghton A, et al. Bilateral sensory abnormalities in patients with unilat- eral neuropathic pain; a quantitative sensory testing (QST) study. PLoS One 2012;7(5):e37524. 21. Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates; 1988. 22. Schulz KF, Altman DG, Moher D, CONSORT Group. CONSORT 2010 statement: updated guidelines for reporting parallel group randomised trials. BMJ 2010;340:c332. 23. Michaux GP, Magerl W, Anton F, Treede RD. Experimental characterization of the effects of acute stresslike doses of hydrocortisone in human neurogenic hyperalgesia models. Pain 2012;153(2):420- 428. 24. Schwegler K, Ettlin D, Buser I, et al. Cortisol reduces recall of explicit contextual pain memory in healthy young men. Psychoneuroendocrinology 2010;35(8):1270-1273. 25. Kemppainen P, Paalasmaa P, Pertovaara A, Alila A, Johansson G. Dexamethasone attenuates exercise-induced dental analgesia in man. Brain Res 1990;519(1-2):329-332. 26. Papadimitriou A, Priftis KN. Regulation of the hypothalamic-pituitary-adrenal axis. Neuroimmuno- modulation 2009;16(5):265-271. 27. DeLeo JA, Tanga FY, Tawfik VL. Neuroimmune activation and neuroinflammation in chronic pain and opioid tolerance/hyperalgesia. Neuroscientist 2004;10(1):40-52. 28. Hubscher M, Moloney N, Leaver A, Rebbeck T, McAuley JH, Refshauge KM. Relationship between quantitative sensory testing and pain or disability in people with spinal pain-a systematic review and meta-analysis. Pain 2013;154(9):1497-1504. 29. Parks EL, Geha PY, Baliki MN, Katz J, Schnitzer TJ, Apkarian AV. Brain activity for chronic knee osteoarthritis: dissociating evoked pain from spontaneous pain. Eur J Pain 2011;15(8):843.e1- 843.14.

Chapter 6

Eff ects of hydrocortisone on the regulation of blood pressure: results from a randomized controlled trial

J Werumeus Buning M van Faassen P Brummelman RPF Dullaart G van den Berg MM van der Klauw MN Kerstens CA Stegeman AC Muller Kobold IP Kema BHR Wolff enbuttel AP van Beek

Journal of Clinical Endocrinology and Metabolism 2016;101(10):3691-3699 116 Chapter 6

Abstract

Context Cardiovascular risk is increased in patients with secondary adrenal insufficiency, which may be ascribed to an unfavorable metabolic profile consequent to a relatively high hydrocortisone replacement dose.

Objective We determined the effects of a higher versus a lower glucocorticoid replacement dose on blood pressure (BP), the renin-angiotensin-aldosterone system, 11β-hydroxysteroid dehydrogenase enzyme activity and circulating (nor)metanephrines.

Design, setting and patients Forty-seven patients with secondary adrenal insufficiency from the University Medical Center Groningen participated in this randomized double-blind crossover study.

Interventions Patients randomly received 0.2–0.3 mg hydrocortisone/kg body weight followed by 0.4–0.6 mg hydrocortisone/kg body weight, or vice versa, each during 10 weeks.

Main outcome measure(s) BP and regulating hormones were measured.

Results The higher hydrocortisone dose resulted in an increase in systolic BP of 5 (12) mm Hg (P = .011), diastolic BP of 2 (9) mm Hg (P = .050), and a median [interquartile range] drop in plasma potassium of -0.1 [-0.3; 0.1] nmol/liter (P = .048). The higher hydrocortisone dose led to decreases in serum aldosterone of -28 [-101; 9] pmol/liter (P = .020) and plasma renin of -1.3 [-4.5; 1.2] pg/ml (P = .051), and increased the ratio of plasma and urinary cortisol to cortisone (including their metabolites) (P < .001 for all). Furthermore, on the higher dose, plasma and urinary normetanephrine decreased by -0.101 [-0.242; 0.029] nmol/liter (P < .001) and -1.48 [-4.06; 0.29] µmol/mol creatinine (P < .001) respectively.

Conclusions A higher dose of hydrocortisone increased systolic and diastolic BP and was accompanied by changes in the renin-angiotensin-aldosterone system, 11β-hydroxysteroid dehydrogenase enzyme activity, and circulating normetanephrine. This demonstrates that hydrocortisone dose even within the physiological range affects several pathways involved in BP regulation. Effects of Hydrocortisone on Blood Pressure 117

Introduction

Patients with adrenal insufficiency (AI) require life-long glucocorticoid (GC) replacement therapy, most commonly administered as oral hydrocortisone (HC). Despite replacement therapy, these patients have an increased overall mortality rate compared to the general population; this is predominantly attributable to an increased risk of cardiovascular and cerebrovascular diseases.1–4 Inappropriate GC substitution therapy may be an explanation for the higher risk of mortality and cardiovascular disease found in patients with AI. GC substitution therapy aims to mimic the endogenous circadian rhythm of cortisol secretion, with peak values in the morning, slowly declining levels throughout the day, and a nadir at midnight. However, conventional oral GC substitution therapy is unable to accurately reproduce this physiological variation,5 inevitably resulting in over- and under-replacement during certain periods of the day. The association between the level of GCs and a number of cardiovascular risk factors, of which hypertension is the most prominent, is well-recognized.6–8 Globally, 6 elevated blood pressure (BP) is the cause of approximately 50 % of deaths from stroke or cardiovascular disease.9 Both in animal models and in human studies, exogenous GCs have been shown to raise BP,10 and patients with AI treated with GCs show increased cardiovascular risk.11 However, to the best of our knowledge, the relationship between BP and different GC replacement doses has only been examined in two small studies, which did not reveal a significant change in BP after a change in GC dose.12,13 Nevertheless, the GC dose-response relationship with BP is still poorly documented in humans, as are the pathogenic mechanisms that conceivably underlie a possible BP response. Patients with secondary AI (SAI) are characterized by a loss of endogenous cortisol production, making them a unique patient cohort because this group lacks the negative feedback mechanisms of the hypothalamic-pituitary-adrenal axis. As a result, cortisol levels can be controlled externally, allowing study of the direct effects of cortisol concentrations on BP and regulating systems. The present randomized double-blind crossover study was initiated to compare the effect of two different doses of HC on cognition (primary endpoint),14 quality of life (QOL),15 metabolic parameters, somatosensory functioning, and pharmacokinetic parameters (secondary endpoints). Here we report the results of the ancillary analysis of the effect of HC dose on BP and pathways related to BP regulation: the renin-angiotensin-aldosterone-system (RAAS), the 11β-hydroxysteroid dehydrogenase (11β-HSD) enzymes catalyzing the cortisol and cortisone interconversion, the sympathetic nervous system, and vasopressin. 118 Chapter 6

Materials and methods

Patients Reporting of the study conforms to the Consolidated Standards of Reporting Trials 2010 statement.16 This study is part of a randomized double-blind crossover study, of which the effects on cognition,14 pain, depressive symptoms, and QOL have been reported previously.15 Patients with SAI for which they received GC substitution therapy were eligible for this study. They were recruited from the endocrine outpatient clinic of the University Medical Center Groningen, a tertiary referral and expertise center for pituitary disease in the Netherlands. A total of 63 patients were included in the study, 60 of whom completed the run-in phase and the baseline measurement. The diagnosis of SAI was based on internationally accepted biochemical criteria. Early morning cutoff cortisol levels for AI in our center were validated for patients with hypothalamic-pituitary disorders, as published previously.17 Other inclusion and exclusion criteria and reasons for withdrawal have been described earlier.14 The study protocol was approved by the local medical ethics review committee at the University Medical Center Groningen. Patients provided informed written consent before entering the study.

Intervention Patients were randomly assigned to either group 1 or group 2. Group 1 first received a physiological lower dose of HC for 10 weeks, followed by a physiological higher dose for another 10 weeks. Group 2 received the two doses in reverse order. Patients were treated in a double blind fashion with oral tablets containing either 5 mg HC (lower dose) or 10 mg HC (higher dose). In the lower dose condition, patients received a cumulative daily dose of 0.2–0.3 mg HC/kg body weight, divided in three doses administered before breakfast, before lunch and before dinner, resulting in total daily doses between 15 and 20 mg HC depending on body weight. In the higher dose condition this was 0.4–0.6 mg HC/kg body weight, resulting in total daily doses between 30 and 40 mg HC dependent on body weight. For the exact dosing scheme, see Werumeus Buning et al.14 Compliance with study medication was assessed using daily diaries in which patients had to report every day whether they had forgotten or doubled their medication, and if so, how often. In addition, HC tablets returned by the patient at the end of each treatment period were counted. In cases of intercurrent illness or fever, patients were allowed to double or triple their HC dose according to a fixed protocol for a maximum of seven days (ie, 10 % of the study time and of the cumulative HC dose). Adjustments of the dose of HC were not allowed in the week preceding testing. Randomization to one of the two treatment groups was performed by Tiofarma Inc., with a block size of four. Effects of Hydrocortisone on Blood Pressure 119

Protocol After each 10-week treatment period, patients returned to the hospital for evaluation. On testing days, they were instructed to take their morning dose of HC at 07:00 am. At 08:00 am, the first blood samples were drawn in a fasting state in sitting position after a short period of rest for the measurement of sodium, potassium, creatinine, cortisol, cortisone, renin concentration, aldosterone, osmolarity, copeptin, and metanephrines (metanephrine, normetanephrine and 3-methoxytyramine). In addition, 24-hour urinary samples were collected for determination of sodium, potassium, osmolarity, cortisol, cortisone, GC metabolites, and metanephrines. Breakfast was provided for the patients, followed by a physical examination including the measurement of body weight, height and BP. Body mass index was calculated as the ratio between the weight (in kilograms) and height (in meters) squared (kg/m²). Blood pressure was measured three times with an automated device (Dinamap XL Model 9300; Johnson & Johnson Medical), with the patient in a sitting position, and reported as the mean of the three measurements. Approximately 5 hours after intake of the morning dose of HC, a second blood sample was drawn for the measurement of cortisol and cortisone. 6

Biochemical measurements Total and free cortisol and cortisone in plasma, as well as free cortisol and free cortisone in urine, were all measured by isotope dilution liquid chromatography tandem mass spectrometry (LC-MS/MS). Total and free cortisol and cortisone in plasma were performed essentially as described by Hawley et al, using cortisol-13C3 and cortisone- 18 D7 as internal standards. For free cortisol and cortisone, intra- and interassay coefficients of variation (CVs) were less than 6 % and less than 5 %, respectively. For total cortisol and cortisone, intra- and interassay CVs were less than 2.6 % and less than 5.3 %, respectively. Plasma equilibrium dialysis for free cortisol and cortisone was performed as described by Fiers et al,19 the only difference being that we used 10-kD cellulose membranes (Harvard Apparatus). Free cortisol and cortisone in urine were performed essentially as described by Jones et al.20 Plasma renin concentration was measured with an immunoradiometric renin assay (Renin III Generation; Cisbio). The intra-assay CVs were 5.5, 4.7, and 1.6 % at 6.0 ng/liter, 19.9 ng/liter, and 52.2 ng/ liter, respectively. The interassay CVs were 14, 9.2, and 4.0 % at 5.1 ng/liter, 19.8 ng/ liter, and 54.7 ng/liter, respectively.21 Aldosterone in serum was measured by LC-MS/ MS, essentially as described by Van der Gugten et al,22 but using additional online solid-phase extraction in combination with LC-MS/MS analysis. Intra- and interassay CVs were less than 6 %. Osmolarity in serum and urine was measured by freezing- point depression on an osmometer (Menarini Diagnostics). Plasma sodium, potassium, creatinine, and urinary sodium and potassium were all measured routinely using a Roche Modular ISE/P system (Roche Diagnostics). Serum corticosteroid-binding 120 Chapter 6

globulin concentrations were determined in duplicate using a RIA (IBL International GmbH). GC metabolites (tetrahydrocortisol [THF], allo-tetrahydrocortisol [alloTHF], tetrahydrocortison [THE]) in urine were measured using gas chromatography in combination with mass spectrometry, as previously described.23 Free metanephrines in plasma were analyzed by LC-MS/MS.24 Urinary free metanephrine concentrations were also determined by LC-MS/MS,25 and concentrations were normalized to the urinary excretion of creatinine, measured using an enzymatic method (Roche Diagnos- tics), and expressed in units of µmol/mol creatinine. Plasma copeptin was measured using a sandwich immunoassay on a Kryptor immunoassay analyzer (Thermo Fischer) based on the assay described by Morgenthaler et al.26 Copeptin can be considered a surrogate marker of vasopressin and has the same rapid in vivo response to osmotic changes as vasopressin but has higher ex vivo preanalytical stability than vasopressin and measurement is faster and easier.27

11β-HSD enzyme activity The urinary cortisol to cortisone metabolite ratio (THF+alloTHF)/THE and plasma cortisol to cortisone ratio are regarded as an overall measure of 11β-HSD activity 28,29 and determined by the combined enzymatic activities of both 11β-HSD types 1 and 2. The ratio of urinary free cortisol (UFF) to urinary free cortisone (UFE) is considered specifically to reflect renal 11β-HSD type 2 activity.30,31 In the liver 11β-HSD type 1 predominantly converts the inactive cortisone to the biologically active cortisol, whereas 11β-HSD type 2, which is highly expressed in mineralocorticoid target tissues, converts active cortisol to inactive cortisone to protect mineralocorticoid receptors (MR) from stimulation by GC.

Statistics As described previously,14 a prestudy power analysis revealed that a study design with two arms of 25 patients each was required to detect a change in primary or secondary endpoints with an effect size of 0.4 (two-sided α = 0.05 and β = 0.80), even when between-test correlations are poor (0.50). An effect size of 0.4 was chosen because it is considered a relevant change with a small to moderate effect size. To allow for a dropout rate of about 20 %, a total of 60 patients needed to be included. Normally distributed data were presented as mean (SD), non-normally distributed data were presented as median [interquartile range], and categorical data were presented as number or percentage. Normality of data was analyzed by visual inspection of Q-Q plots and histograms. To test for period and carryover effects, the procedure developed by Altman was used.32 In this procedure, to test for a carry-over effect, the average response to both treatments (ie, of the low dose and high dose combined) was compared between the Effects of Hydrocortisone on Blood Pressure 121 two treatment groups. If these average responses were not different between the treatment groups, the effect of the treatment was considered the same irrespective of the order in which the treatments were administered.32 All variables were compared using the Wilcoxon signed rank test for paired observations. Statistical significance was set at a P < .05. All statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS, Inc.), version 22.

Results

Study population A total of 63 patients were included in this trial. Forty-seven of them completed both study periods (29 men, age 51 (14) years [range, 19–73]). The patient flow through the study is shown in Supplemental Figure 1. Thirteen patients withdrew from the study, with various reasons for withdrawal that have been described previously.14 The number of patients that did not complete the study was comparable between the two doses and 6 the two treatment groups (data not shown). Furthermore, subjects completing the study did not differ from those who did not with regard to age, sex, educational level, age at diagnosis, childhood onset or adult onset, body weight, or number of patients taking antihypertensive medication (data not shown). One patient was excluded from the analysis because of starting BP-lowering medication during the study; therefore a total of 46 patients were included in the present analysis. The patients’ characteristics are given in Supplemental Table 1. Fourteen of the patients used BP-lowering medication before the study. Of those, four received angiotensin-converting enzyme inhibitors, one received thiazides, nine received beta blockers, six received calcium antagonists, two received angiotensin II receptor antagonists, and one received alpha blockers (some patients used multiple BP-lowering drugs). All other medication apart from the study medication remained stable during the study periods.

Effects of HC dose on cortisol and cortisone concentrations in blood and urine During the low dose condition, patient received on average 17.9 (2.2) mg HC/day, during the high dose condition they received on average 35.8 (4.5) mg HC/day. Administration of the higher HC dose resulted in significantly higher plasma total cortisol levels compared to the administration of the lower HC dose, both 1 hour after ingestion (median difference, 233 [136; 367] nmol/liter) and approximately 5 hours after ingestion (median difference, 100 [10; 156] nmol/liter, both P < .001) (Table 1). Plasma free cortisol levels were significantly increased after the higher dose of HC, both 1 hour after administration (median difference, 31 [14; 43] nmol/liter) and approximately 5 hours after administration (median difference, 4 [1; 7] nmol/liter, both 122 Chapter 6

Table 1. Cortisol concentrations in plasma and urine (n = 46) Lower dose Higher dose P value Plasma total cortisol levels 1 h after ingestion (nmol/liter) 500 [389; 602] 745 [676; 880]a <.001 5 h after ingestion (nmol/liter) 117 [76; 211] 232 [171; 311] <.001 Plasma total cortisone levels 1 h after ingestion (nmol/liter) 65 [53; 70] 63 [55; 71]a .852 5 h after ingestion (nmol/liter) 30 [21; 40] 44 [36; 57] <.001 Plasma free cortisol levels 1 h after ingestion (nmol/liter) 32 [24; 44]a 71 [50; 81]b <.001 5 h after ingestion (nmol/liter) 4 [2; 8]c 9 [6; 15]c .005 Plasma free cortisone levels 1 h after ingestion (nmol/liter) 11 [9; 12]a 11 [10; 13]b .018 5 h after ingestion (nmol/liter) 4 [3; 5]c 7 [5; 8]c .007 Urinary free cortisol (nmol/24 h) 78 [47; 112]a 281 [200; 411]d <.001 Urinary free cortisone (nmol/24 h) 229 [147; 374]a 445 [326; 663]d <.001 Data are median [interquartile range]. a n = 45, b n = 42, c n = 37, d n = 44.

P < .01) (Table 1). UFF median excretion levels were increased after receiving the higher dose of HC by 209 [134; 320] nmol/24 hours compared to the lower dose (P < .001) (Table 1). Plasma total cortisone levels did not differ between the two HC doses 1 hour after ingestion of the morning dose, whereas 5 hours after the morning dose of HC a significant increase was observed after the higher dose (median difference, 17 [-1; 27] nmol/liter, P < .001) (Table 1). For plasma free cortisone levels, a significant increase was found both 1 hour and 5 hours after intake of the higher HC dose (median difference, 1 [-1; 2] nmol/liter and 3 [-1; 4] nmol/liter, respectively, both P < .05) (Table 1). Furthermore, UFE median excretion levels were increased with 230 [133; 370] nmol/24 hours (P < .001) after receiving the higher dose of HC (Table 1).

Effects of HC dose on BP and body weight An increase in systolic BP of 5 (12) mm Hg (P = .011) and a borderline significant increase in diastolic BP of 2 (9) mm Hg (P = .050) were observed in patients using the higher compared with the lower HC dose (Table 2). The effect of HC dose on BP was not different between patients treated with BP-lowering medication and those who were not (Supplemental Table 2). Body weight and body mass index tended to increase by 0.5 (1.7) kg (P = .060) and 0.2 (0.5) kg/m2 points (P = .045), respectively, in response to the higher HC dose (Table 2). Effects of Hydrocortisone on Blood Pressure 123

Table 2. Anthropometric measures and biochemical and hormonal analysis (n = 46) Lower dose Higher dose P value Systolic BP (mm Hg) 133 (14)a 138 (16)a .011 Diastolic BP (mm Hg) 76 (10)a 78 (9)a .050 Weight (kg) 82.8 (14.0) 83.3 (14.3) .060 BMI (kg/m²) 26.9 (4.0) 27.1 (4.0) .045 Plasma sodium (mmol/liter) 142 [141; 143] 142 [141; 143] .099 Plasma potassium (mmol/liter) 3.9 [3.7; 4.0] 3.8 [3.6; 4.0] .048 Plasma creatinine (µmol/liter) 82 [66; 88] 80 [68; 89] .109 Serum CBG (µg/ml) 53.2 [49.1; 63.0]a 56.5 [49.0; 62.5]a .102 Plasma renin concentration (pg/ml) 11.6 [6.7; 17.3]b 8.6 [5.9; 14.9]b .051 Serum aldosterone (pmol/liter) 150 [77; 256]a 107 [43; 235]a .020 ARR (pmol/ng) 13.8 [7.3; 21.3]b 11.0 [6.1; 19.8]b .044 TTKG 7.42 [6.12; 9.48]c 7.47 [6.00; 9.18]b .352 Plasma copeptin (pmol/liter) 3.7 [2.5; 5.0]b 3.4 [2.5; 4.9]b .819 Urine sodium (mmol/24 h) 142 [119; 206]a 161 [112; 200]c .534 Urine potassium (mmol/24 h) 77 [64; 96]a 83 [69; 103]c .032 6 Creatinine clearance calculated (ml/min) 117 (37)a 117 (29)c .700 Abbreviations: ARR, aldosterone to renin ratio; CBG, corticosteroid-binding globulin; TTKG,: transtubular potassium gradient. Data are mean (SD) or median [interquartile range]. Fourteen patients received blood pressure lowering medication. Diabetes mellitus treated with medication known to be able to induce hypoglycaemia was an exclusion criterion. a n = 45, b n = 43, c n = 44.

The RAAS The higher dose of HC led to a borderline significant median decrease in plasma renin of -1.3 [-4.5; 1.2] pg/ml (P = .051, Figure 1A), a median decrease in aldosterone of -28 [-101; 9] pmol/liter (P = .020, Figure 1B) and a median decrease in the aldosterone- to-renin ratio of -2.7 [-5.7; 1.5] pmol/ng (P = .044), compared to the lower dose of HC (Table 2). Treatment with the higher HC dose was followed by a small, albeit significant, reduction of plasma potassium concentrations compared to the lower HC dose (median difference, -0.1 mmol/liter [-0.3; 0.1] mmol/liter, P = .048). Plasma sodium and creatinine concentrations, as well as the transtubular potassium gradient, did not differ between the two doses. The higher dose of HC led to higher 24-hour urine potassium excretion (median difference, 10 [-9; 28] mmol/24 hours, P = .032), but sodium excretion remained unchanged. Of note, a significant treatment by period interaction effect was found for plasma renin concentration (P = .039). Thus, the effect of dose was different for the two treatment groups, depending on the order in which the doses were given (Supplemental Figure 2). We performed a sensitivity analysis 124 Chapter 6

 





 



       

  



 

   

  

Figure 1. Plasma renin concentration (A) and serum aldosterone (B) on the two doses of hydrocortisone. Box plots represent median and interquartile ranges, whiskers represent fifth and 95th percentiles. HD, high-dose hydrocortisone; LD, low-dose hydrocortisone. *** P < .05.

to further study the effects of HC dose on the RAAS. To this end, we excluded all patients with RAAS-influencing medication (n = 13). Results are shown in Supple- mental Table 3. The effects on BP, plasma renin, serum aldosterone, plasma potassium and the aldosterone-to-renin ratio remained significantly different and were even more pronounced than in the entire group.

11β-HSD enzyme activity A shift toward higher cortisol exposure on the higher HC dose was found as calculated by the ratios of plasma cortisol to cortisone (overall 11β-HSD enzyme activity) (median increase of 4 [2; 6] for total cortisol to cortisone; median increase of 3 [1; 4] for free cortisol to cortisone, Figure 2, A and B), urinary (THF+alloTHF)/THE ratio (overall 11β-HSD enzyme activity) (median increase, 0.41 [0.22; 0.69], Supplemental Figure 3A) and the ratio UFF/UFE (11β-HSD type 2 enzyme activity) (median increase, 0.30 [0.20; 0.47], Supplemental Figure 3B). Effects of Hydrocortisone on Blood Pressure 125

   



 

               

   



  6

               

Figure 2. Ratio of plasma total cortisol/total cortisone (A) and ratio of plasma free cortisol/free cortisone (B) on the two doses of hydrocortisone, 1 hour (hr) and 5 hr after ingestion of the morning dose. Box plots represent median and interquartile ranges, whiskers represent fifth and 95th percentiles. HD, high-dose hydrocortisone; LD, low-dose hydrocortisone. *** P < .001; * P < .05.

Sympathetic activity Plasma free normetanephrine concentrations showed a median decrease of -0.101 [-0.242; 0.029] nmol/liter (P < .01) on the higher dose of HC (Figure 3A). Plasma free metanephrine and 3-methoxytyramine levels remained unchanged (Figure 3, B and C). Free normetanephrine in 24-hour urinary samples decreased on the higher dose of HC (median difference, -1.48 [-4.06; 0.29] µmol/mol creatinine, P < .001, Figure 4A), as well as free 3-methoxytyramine (median difference, -2.20 [-3.99; -0.82] µmol/mol creatinine, P < .001, Figure 4C), whereas urinary free metanephrine concentrations remained unchanged (Figure 4B).

Copeptin Copeptin concentrations were not different on the two doses of HC (Table 2). 126 Chapter 6

Discussion

The present double-blind randomized controlled trial showed that higher levels of HC substitution doses lead to increased BP together with changes in several related

  





     

  

 



     

  

 





     

  

Figure 3. Plasma free normetanephrine (NMN) (A), metanephrine (MN) (B), and 3-methoxytyramine (3-MT) (C) concentrations on the two doses of hydrocortisone. Box plots represent median and interquartile ranges, whiskers represent fifth and 95th percentiles. HD, high-dose hydrocortisone; LD, low-dose hydrocortisone. ** P < .01. Effects of Hydrocortisone on Blood Pressure 127 systems such as the RAAS, 11β-HSD enzyme activity and sympathetic nerve activity. A major strength of this study is its crossover design. Patients were their own controls, thereby minimizing interindividual effects. Furthermore, patients with SAI have an intact RAAS, but lack the negative feedback mechanism of the hypothalamic- pituitary-adrenal axis that is normally present in healthy individuals. This enabled us to control the cortisol levels externally and study the effects of cortisol levels on

  



        

   6

 



         

  

  





   

     

  

Figure 4. Urinary free normetanephrine (NMN) (A), free metanephrine (MN) (B), and free 3-methoxytyramine (3-MT) (C) concentrations on the two doses of hydrocortisone. Box plots represent median and interquartile ranges, whiskers represent fifth and 95th percentiles. HD, high-dose hydrocortisone; LD, low-dose hydrocortisone. *** P < .001. 128 Chapter 6

BP and regulating mechanisms. Furthermore, elaborate biochemical measurements enabled us to explore the effect on mechanisms involved in BP regulation. A limitation of the study is that the BP value is based on three office BP measurements, whereas 24 hour ambulatory BP measurement is reported to be more accurate and is more related to cardiovascular events.33,34 The effects of HC on BP are likely to be caused by increased MR activation. The higher HC dose resulted in an increase in BP together with a decrease in plasma potassium, an increase in urinary potassium excretion, and a decrease in plasma renin and serum aldosterone concentrations that, all taken together, are compatible with activation of the MR.35 These effects were even more pronounced when studied in patients without RAAS-influencing medication. Our results are in contrast to those found by Dunne and colleagues.12 They found no change in BP in 13 patients after reduction of a cumulative dose of HC from 30 to 15 mg/d. It is possible that because of the small number of patients, this study may have lacked power to detect an effect on BP. Our results are also in contrast to the findings of Petersons and colleagues.13 In this study, among 17 subjects with SAI, BP was unaltered after increasing the HC daily replacement dose equivalent to less than 20 mg/d to 30 mg/d for 7 days. That study, however, was much smaller, had a nonrandomized design, used a different GC dosing protocol, and had a considerable shorter intervention period compared to our study. Further, in another open- label randomized trial, Johannsson and colleagues were able to show a beneficial effect of reduced cortisol exposure on BP.36 Using an oral dual-release HC formulation and compared to thrice daily dosing, a similar daily dose of the dual-release HC resulted in a more physiological cortisol profile and a 20 % reduction in 24-hour cortisol exposure. The investigators describe a reduction in BP of 6 mm Hg systolic and 2 mm Hg diastolic; however, mechanisms leading to this reduction in BP were not studied. Sodium is an important determinant of the RAAS and sodium retention results in decreased renin concentrations. In addition to MR, glucocorticoid receptors also mediate sodium retention37; thus the observed suppression of renin concentration could also be the result of activation of the glucocorticoid receptors. No information on sodium intake is available for the present patient group. However, plasma sodium concentrations as well as sodium excretion remained unchanged under the two doses of HC. Therefore, it is unlikely that sodium intake has influenced our results. Exogenous estrogen may interfere with cortisol level measurements because of estrogen induced elevation of serum corticosteroid-binding globulin levels.38 However, to correct for this effect, several measurements of cortisol were included that are not influenced by corticosteroid-binding globulin, most importantly plasma free cortisol. An additional analysis was performed in women with and without estrogen substitution and showed that both plasma total cortisol levels and plasma free cortisol levels did not differ, neither on the two different HC doses, nor at the different time points after HC intake (data not Effects of Hydrocortisone on Blood Pressure 129 shown). Furthermore, estrogen substitution therapy remained unchanged throughout the study. Even though we cannot fully exclude the effects of possible confounding covariates, because of the crossover design of the study, probable effects are reduced. It is known that cortisol and aldosterone have similar affinity for the MR in vitro, but 11β-HSD type 2 protects the MR in vivo from stimulation by cortisol through catalyzing the conversion of biologically active cortisol to inactive cortisone.39 The increased MR effects we observed may be caused by a change in set point of 11β-HSD enzyme activity, thereby favoring enhanced cortisol exposure. This seems attributable to a change in 11β-HSD type 2 enzyme activity as indicated by the increased UFF/UFE ratio, whereas changes in 11β-HSD type 1 activity cannot be excluded. Intriguingly, 11β-HSD activity may thus contribute to a feed-forward system enhancing systemic cortisol exposure. In accordance with these findings, Sherlock and colleagues showed that patients receiving exogenous HC therapy demonstrate a significantly altered corticosteroid metabolism with respect to urinary (THF+alloTHF)/THE ratios, with the highest ratios found in patients receiving the highest dose HC of 30 mg HC/day.40 Although the use of metabolite ratios is widely accepted as a tool to assess enzyme activity,35,39,40 newer study methodology, in 6 particular using isotope tracer infusions (ie deuterated cortisol and cortisone), may allow improved delineation of the activity of 11β-HSD type 1 and 2 in vivo.41,42 Plasma and urinary normetanephrine levels were decreased during treatment on the higher HC replacement dose. Normetanephrine is a metabolite of norepinephrine, which is a neurotransmitter released by sympathetic nerves. Normetanephrine concentrations are widely accepted to reflect changes in sympathetic nerve activity in normal physiology.43,44 The most likely explanation for the decrease in sympathetic activity is the baro- receptor reflex which is influenced by (a change in) BP. Elevated BP results in an increase in the sensory impulses transmitted from the baroreceptors to the vasomotor center in the brainstem, enhancing parasympathetic and decreasing sympathetic activity in order to reestablish normal BP values.44,45 In this context, it is also relevant that increased effector organ responsiveness towards norepinephrine under high GC doses is observed.46,47 Of note, supraphysiological doses of GCs were used in the described studies. We found no changes in copeptin as surrogate marker for vasopressin. Vasopressin is involved in homeostasis by regulating water absorption in the renal tubules and by vasoconstriction mediated by two different vasopressin receptors.48 Hypocortisolism changes the free water clearance by corticotropin-releasing hormone-stimulated vasopressin release.49 Potentially, we were unable to demonstrate an effect of HC dose on copeptin concentrations because our study subjects were deficient in ACTH and possibly also in part in corticotropin-releasing hormone. The observed response may thus be different in healthy individuals. A complex balancing of the harms and benefits needs to be made in the individualization of substitution therapy in patients with SAI. We suggest to use the lowest dose 130 Chapter 6

possible that relieves symptoms of GC deficiency to prevent the observed negative side effects of higher doses of GCs. However, in light of our previous paper, which showed improved QOL after higher substitution doses, an increase in dose should be considered when necessary or desirable. In conclusion, a higher dose of HC increased BP and resulted in changes in BP- regulating mechanisms, with most pronounced effects found in the RAAS, 11β-HSD enzyme activity, and sympathetic nerve activity.

Acknowledgement

The authors thank Dr. W. J. Sluiter for his statistical advice. Effects of Hydrocortisone on Blood Pressure 131

References

1. Bergthorsdottir R, Leonsson-Zachrisson M, Oden A, Johannsson G. Premature mortality in patients with Addison’s disease: a population-based study. J Clin Endocrinol Metab 2006;91(12):4849-4853. 2. Bülow B, Hagmar L, Mikoczy Z, Nordström CH, Erfurth EM. Increased cerebrovascular mortality in patients with hypopituitarism. Clin Endocrinol (Oxf) 1997;46(1):75-81. 3. Rosen T, Bengtsson BA. Premature mortality due to cardiovascular disease in hypopituitarism. Lancet 1990;336(8710):285-288. 4. Tomlinson JW, Holden N, Hills RK, et al. Association between premature mortality and hypopituitarism. West Midlands Prospective Hypopituitary Study Group. Lancet 2001;357(9254):425-431. 5. Howlett TA. An assessment of optimal hydrocortisone replacement therapy. Clin Endocrinol (Oxf) 1997;46(3):263-268. 6. Walker BR. Glucocorticoids and cardiovascular disease. Eur J Endocrinol 2007;157(5):545-559. 7. Wei L, MacDonald TM, Walker BR. Taking glucocorticoids by prescription is associated with subsequent cardiovascular disease. Ann Intern Med 2004;141(10):764-770. 8. Andrews RC, Walker BR. Glucocorticoids and insulin resistance: old hormones, new targets. Clin Sci (Lond) 1999;96(5):513-523. 9. Laslett LJ, Alagona P,Jr, Clark BA,3rd, et al. The worldwide environment of cardiovascular disease: prevalence, diagnosis, therapy, and policy issues: a report from the American College of Cardiology. 6 J Am Coll Cardiol 2012;60(25 Suppl):S1-49. 10. Whitworth JA. Mechanisms of glucocorticoid-induced hypertension. Kidney Int 1987;31(5):1213- 1224. 11. Filipsson H, Monson JP, Koltowska-Haggstrom M, Mattsson A, Johannsson G. The impact of glucocorticoid replacement regimens on metabolic outcome and comorbidity in hypopituitary patients. J Clin Endocrinol Metab 2006;91(10):3954-3961. 12. Dunne FP, Elliot P, Gammage MD, et al. Cardiovascular function and glucocorticoid replacement in patients with hypopituitarism. Clin Endocrinol (Oxf) 1995;43(5):623-629. 13. Petersons CJ, Mangelsdorf BL, Thompson CH, Burt MG. Acute effect of increasing glucocorticoid replacement dose on cardiovascular risk and insulin sensitivity in patients with adrenocorticotrophin deficiency. J Clin Endocrinol Metab 2014;99(6):2269-2276. 14. Werumeus Buning J, Brummelman P, Koerts J, et al. The effects of two different doses of hydrocortisone on cognition in patients with secondary adrenal insufficiency - Results from a randomized controlled trial. Psychoneuroendocrinology 2015;55:36-47. 15. Werumeus Buning J, Brummelman P, Koerts J, et al. Hydrocortisone Dose Influences Pain, Depres- sive Symptoms and Perceived Health in Adrenal Insufficiency - A Randomized Controlled Trial. Neuroendocrinology 2016;103(6):771-8. 16. Schulz KF, Altman DG, Moher D, CONSORT Group. CONSORT 2010 statement: updated guidelines for reporting parallel group randomised trials. BMJ 2010;340:c332. 17. Dullaart RP, Pasterkamp SH, Beentjes JA, Sluiter WJ. Evaluation of adrenal function in patients with hypothalamic and pituitary disorders: comparison of serum cortisol, urinary free cortisol and the human-corticotrophin releasing hormone test with the insulin tolerance test. Clin Endocrinol (Oxf) 1999;50(4):465-471. 18. Hawley JM, Owen LJ, MacKenzie F, Mussell C, Cowen S, Keevil BG. Candidate Reference Mea- surement Procedure for the Quantification of Total Serum Cortisol with LC-MS/MS. Clin Chem 2016;62(1):262-269. 132 Chapter 6

19. Fiers T, Delanghe J, T’Sjoen G, Van Caenegem E, Wierckx K, Kaufman JM. A critical evaluation of salivary testosterone as a method for the assessment of serum testosterone. Steroids 2014;86:5-9. 20. Jones RL, Owen LJ, Adaway JE, Keevil BG. Simultaneous analysis of cortisol and cortisone in saliva using XLC-MS/MS for fully automated online solid phase extraction. J Chromatogr B Analyt Technol Biomed Life Sci 2012;881-882:42-48. 21. Kerstens MN, Kobold AC, Volmer M, Koerts J, Sluiter WJ, Dullaart RP. Reference values for aldosterone-renin ratios in normotensive individuals and effect of changes in dietary sodium con- sumption. Clin Chem 2011;57(11):1607-1611. 22. Van Der Gugten JG, Crawford M, Grant RP, Holmes DT. Supported liquid extraction offers improved sample preparation for aldosterone analysis by liquid chromatography tandem mass spectrometry. J Clin Pathol 2012;65(11):1045-1048. 23. Kerkhofs TM, Kerstens MN, Kema IP, Willems TP, Haak HR. Diagnostic Value of Urinary Steroid Profiling in the Evaluation of Adrenal Tumors. Horm Cancer 2015;6(4):168-175. 24. de Jong WH, Graham KS, van der Molen JC, et al. Plasma free metanephrine measurement using automated online solid-phase extraction HPLC tandem mass spectrometry. Clin Chem 2007;53(9):1684-1693. 25. de Jong WH, Eisenhofer G, Post WJ, Muskiet FA, de Vries EG, Kema IP. Dietary influences on plasma and urinary metanephrines: implications for diagnosis of catecholamine-producing tumors. J Clin Endocrinol Metab 2009;94(8):2841-2849. 26. Morgenthaler NG, Struck J, Alonso C, Bergmann A. Assay for the measurement of copeptin, a stable peptide derived from the precursor of vasopressin. Clin Chem 2006;52(1):112-119. 27. Christ-Crain M, Fenske W. Copeptin in the diagnosis of vasopressin-dependent disorders of fluid homeostasis. Nat Rev Endocrinol 2016;12(3):168-176. 28. Ulick S, Levine LS, Gunczler P, et al. A syndrome of apparent mineralocorticoid excess associated with defects in the peripheral metabolism of cortisol. J Clin Endocrinol Metab 1979;49(5):757-764. 29. Kerstens MN, Riemens SC, Sluiter WJ, Pratt JJ, Wolthers BG, Dullaart RP. Lack of relationship between 11beta-hydroxysteroid dehydrogenase setpoint and insulin sensitivity in the basal state and after 24h of insulin infusion in healthy subjects and type 2 diabetic patients. Clin Endocrinol (Oxf) 2000;52(4):403-411. 30. Best R, Walker BR. Additional value of measurement of urinary cortisone and unconjugated cortisol metabolites in assessing the activity of 11 beta-hydroxysteroid dehydrogenase in vivo. Clin Endo- crinol (Oxf) 1997;47(2):231-236. 31. Palermo M, Shackleton CH, Mantero F, Stewart PM. Urinary free cortisone and the assessment of 11 beta-hydroxysteroid dehydrogenase activity in man. Clin Endocrinol (Oxf) 1996;45(5):605-611. 32. Altman D. Practical statistics for medical research. London: Chapman and Hall; 1991. 33. Clement DL, De Buyzere ML, De Bacquer DA, et al. Prognostic value of ambulatory blood-pressure recordings in patients with treated hypertension. N Engl J Med 2003;348(24):2407-2415. 34. Dolan E, Stanton A, Thijs L, et al. Superiority of ambulatory over clinic blood pressure measurement in predicting mortality: the Dublin outcome study. Hypertension 2005;46(1):156-161. 35. Ferrari P, Lovati E, Frey FJ. The role of the 11beta-hydroxysteroid dehydrogenase type 2 in human hypertension. J Hypertens 2000;18(3):241-248. 36. Johannsson G, Nilsson AG, Bergthorsdottir R, et al. Improved cortisol exposure-time profile and outcome in patients with adrenal insufficiency: a prospective randomized trial of a novel hydrocortisone dual-release formulation. J Clin Endocrinol Metab 2012;97(2):473-481. 37. Nishimoto M, Fujita T. Renal mechanisms of salt-sensitive hypertension: contribution of two steroid receptor-associated pathways. Am J Physiol Renal Physiol 2015;308(5):F377-87. Effects of Hydrocortisone on Blood Pressure 133

38. Musa BU, Seal US, Doe RP. Elevation of certain plasma proteins in man following estrogen administration: a dose-response relationship. J Clin Endocrinol Metab 1965;25(9):1163-1166. 39. van Uum SH, Hermus AR, Smits P, Thien T, Lenders JW. The role of 11 beta-hydroxysteroid dehydrogenase in the pathogenesis of hypertension. Cardiovasc Res 1998;38(1):16-24. 40. Sherlock M, Behan LA, Hannon MJ, et al. The modulation of corticosteroid metabolism by hydrocortisone therapy in patients with hypopituitarism increases tissue glucocorticoid exposure. Eur J Endocrinol 2015;173(5):583-593. 41. Kilgour AH, Semple S, Marshall I, Andrews P, Andrew R, Walker BR. 11beta-Hydroxysteroid dehydrogenase activity in the brain does not contribute to systemic interconversion of cortisol and cortisone in healthy men. J Clin Endocrinol Metab 2015;100(2):483-489. 42. Andrew R, Smith K, Jones GC, Walker BR. Distinguishing the activities of 11beta-hydroxysteroid dehydrogenases in vivo using isotopically labeled cortisol. J Clin Endocrinol Metab 2002;87(1):277- 285. 43. Eisenhofer G, Peitzsch M. Laboratory evaluation of pheochromocytoma and paraganglioma. Clin Chem 2014;60(12):1486-1499. 44. McCorry LK. Physiology of the autonomic nervous system. Am J Pharm Educ 2007;71(4):78. 45. Guyton AC, Hall JE. Textbook of medical physiology. 9th ed. Philadelphia: W.B. Saunders; 1996. 46. Sudhir K, Jennings GL, Esler MD, et al. Hydrocortisone-induced hypertension in humans: pressor responsiveness and sympathetic function. Hypertension 1989;13(5):416-421. 6 47. Heindl S, Vahlkamp K, Weitz G, Fehm HL, Dodt C. Differential effects of hydrocortisone on sympathetic and hemodynamic responses to sympathoexcitatory manoeuvres in men. Steroids 2006;71(3):206-213. 48. Oh YK. Vasopressin and vasopressin receptor antagonists. Electrolyte Blood Press 2008;6(1):51-55. 49. Rotondo F, Butz H, Syro LV, et al. Arginine vasopressin (AVP): a review of its historical perspectives, current research and multifunctional role in the hypothalamo-hypophysial system. Pituitary 2016;[Epub ahead of print; doi: 10.1007/s11102-015-0703-0]. 134 Chapter 6

Supplemental data

Supplemental Figure 1. Patient flow diagram Effects of Hydrocortisone on Blood Pressure 135

      



   



      

Supplemental Figure 2. Period by treatment interaction effect Group 1 received first the lower dose of hydrocortisone followed by the higher dose; Group 2 first received the higher dose followed by the lower dose. The mean values for renin concentration of the two groups differed significantly (P = 0.039), indicating a period by treatment interaction effect. 6 136 Chapter 6

  







      

  





   

  

Supplemental Figure 3. Ratio of (THF+alloTHF)/THE (A) and the ratio of UFF/UFE (B) on the two doses of hydrocortisone. Box plots represent median and interquartile ranges, whiskers represent 5th and 95th percentile. HD: high dose hydrocortisone, LD low dose hydrocortisone, THE: tetrahydrocortisone, THF: tetrahydrocortisol, allo-THF: allo-tetrahydrocortisol, UFE: urinary free cortisone, UFF: urinary free cortisol. *** P < 0.001. Effects of Hydrocortisone on Blood Pressure 137

Supplemental Table 1. Clinical characteristics at baseline of pituitary patients treated for adrenal insufficiency (N = 46) Sex (men/women), n 28/18 Age (y), median [IQR] 54 [42; 61] Age at diagnosis (y), median [IQR] 31 [20; 46] Body weight (kg), median [IQR] 82.0 [72.1; 93.1] Hydrocortisone treatment prior to randomization Total daily dose (mg/day), median [IQR] 25 [20; 30] Dose/kg body weight (mg/kg), median [IQR] 0.32 [0.25; 0.35] Number of daily dosages (1/2/3), n 3/33/10 Duration of glucocorticoid treatment (y), median [IQR] 12 [5; 23] No. of hormonal replacements (1/2/3/4/5) 3/9/21/10/3 Thyroid hormone deficiency (% of patients) 91 Growth hormone deficiency (% of patients) 66 Growth hormone deficiency (% of patients receiving substitution) 67 Sex hormone deficiency (% of patients) 57 6 Men: testosterone (% of patients receiving substitution) 79 Premenopausal women, n = 8: estrogens (% of patients receiving 50 substitution) Postmenopausal women, n = 10: estrogens NA Desmopressin (% of patients) 20 Abbreviations: IQR: Interquartile range, NA: Not applicable. 138 Chapter 6

Supplemental Table 2. Effect of dose of HC on blood pressure for patients with and without blood-pressure lowering medication Blood-pressure No blood- lowering medication pressure lowering (n = 13) medication (n = 33) P-value* Δ Systolic blood pressure (mm Hg) 5 (17) 5 (10) 0.542 Δ Diastolic blood pressure (mm Hg) 3 (10) 1 (8) 0.961 One patient was excluded from the analysis due to the introduction of a new blood-pressure lowering drug during the study. Data are mean (SD). * P-value blood-pressure lowering medication versus no blood-pressure lowering medication by Mann-Whitney U Test

Supplemental Table 3. Effect of HC dose on blood pressure and the RAAS for patients not using medication influencing the RAAS (N = 33) Lower dose Higher dose P-value * SBP (mmHg) 129 (12) 135 (14)a 0.007 DBP (mmHg) 76 (9) 77 (8)a 0.108 Plasma potassium (mmol/L) 3.9 [3.7; 4.1] 3.7 [3.6; 4.0] 0.006 Plasma renin concentration (pg/mL) 11.9 [7.6; 19.0] 8.6 [6.0; 5.0] 0.021 Serum aldosterone (pmol/L) 157 [68; 278] 91 [39; 220] 0.002 ARR (pmol/ng) 12.3 [7.3; 21.3] 9.5 [6.4; 18.6] 0.004 TTKG 7.9 [6.5; 9.8] 8.0 [6.2; 10.1] 0.570 Urine potassium (mmol/24hr) 82 [61; 100] 83 [68; 106]e 0.116 Patients using angiotensin-converting-enzyme inhibitors, angiotensin II receptor blockers, beta- blockers and diuretics were excluded. Abbreviations: ARR: aldosterone to renin ratio, DBP: diastolic blood pressure, SBP: systolic blood pressure, TTKG: transtubular potassium gradient. Data are mean (SD) or median [interquartile range]. * P-value lower dose versus higher dose by Wilcoxon Signed Rank Test for paired data.

Chapter 7

Pharmacokinetics of hydrocortisone: results and implications from a randomized controlled trial

J Werumeus Buning DJ Touw P Brummelman RPF Dullaart G van den Berg MM van der Klauw J Kamp BHR Wolff enbuttel AP van Beek

Accepted in Metabolism 142 Chapter 7

Abstract

Context and Objective This study aimed at comparing pharmacokinetics of two different doses of hydrocortisone (HC) in patients with secondary adrenal insufficiency (SAI).

Design, Setting and Patients Forty-six patients with SAI participated in this randomized double-blind crossover study.

Intervention Patients received two different doses of HC (0.2–0.3 mg HC/kg body weight/day and 0.4–0.6 mg HC/kg body weight/day).

Main Outcome Measures One- and two-compartment population models for plasma free cortisol, plasma total cortisol and salivary cortisol were parameterized. The individual pharmacokinetic

parameters clearance (CL), volume of distribution (Vd), elimination half-life (t1/2),

maximum concentration (Cmax), and area under the curve (AUC) were calculated.

Results The one-compartment models gave a better description of the data compared to the two-compartment models. Weight-adjusted dosing reduced variability in cortisol exposure with comparable AUCs between weight groups. However, there was large

inter-individual variation in CL and Vd of plasma free cortisol, plasma total cortisol

and salivary cortisol. As a consequence, AUC24h varied more than 10 fold. Cortisol exposure was increased with the higher dose, but this was dose proportional only for free cortisol concentrations and not for total cortisol.

Conclusions Cortisol concentrations after a doubling of the dose were only dose proportional for free cortisol. HC pharmacokinetics can differ up to 10-fold inter-individually and individual adjustment of treatment doses may be necessary. Doubling of the HC dose in fast metabolizers (patients that showed relative low AUC and thus high clearance compared to other patients), does not result in significantly enhanced exposure during large parts of the day and these patients may need other management strategies. Population Pharmacokinetics of Hydrocortisone 143

Introduction

Hydrocortisone (HC) substitution by immediate release tablets is the most widely used glucocorticoid treatment to substitute for the lack of endogenous cortisol production in adrenal insufficiency (AI). Since the introduction of HC in the 50’s, it is considered lifesaving, but it is also known for its imperfections with over- and undersubstitution during certain parts of the day.1,2 To overcome this problem, several pharmaceutical products modifying the release of HC have been developed.3,4 However, in spite of this progress, immediate release tablets remain the cornerstone of glucocorticoid substitution therapy today. Various methods to assess the adequacy of glucocorticoid treatment regimen are suggested, but none of them is fully satisfactory. Some have advocated that clinical judgment is best to avoid over- or undersubstitution.5,6 But considering that studies in patients with AI report increased mortality and morbidity rates in which undersubstitution as well as oversubstitution with glucocorticoids might play a role,7 additional guidance is needed. The use of biomarkers to assess the adequacy of HC replacement therapy has been promising, but it has not been successful so far.8,9 Others use plasma cortisol day curves, but they are considered time-consuming, expensive and inconvenient for the 7 patient,10 and they lack sensitivity to discriminate under- and over-substituted patients from well-substituted patients.5 Furthermore, total cortisol levels are usually measured, instead of free cortisol levels that is believed to reflect the biologically active form of cortisol. However, the measurement of plasma free cortisol is cumbersome and therefore not used in clinical practice. An alternative method to determine free cortisol concentrations is in 24-h urine samples, however, fluctuations in cortisol levels during the day are missed due to the nature of this method.11 Assessment of cortisol in saliva has been reported to be a good alternative for free cortisol measurement as it is easy to perform.12 However, there is large variation in the correlations between salivary cortisol and plasma total cortisol or plasma free cortisol, which may be induced by differences in glucocorticoids used or corticosteroid-binding globulin (CBG) concentrations.11,13–15 Alternatively, pharmacokinetic analyses have been performed.1,11,16,17 This has led to recommendations for HC substitution therapy including thrice daily, weight related dosing, administered before the meal.17 However, most of the studies performed either investigated healthy subjects,16 had relatively small samples of patients,11,17 or measured plasma total cortisol instead of plasma free cortisol.1,11,16,17 Studies assessing the pharmacokinetics of hydrocortisone on free and total cortisol, both in plasma and saliva, and studying the effect of dose adjustments in patients with SAI have not been performed. 144 Chapter 7

We have recently conducted a randomized double blind cross-over study aiming at assessing the effect of two different doses of HC on cognition (primary endpoint),18 health related quality of life,19 somatosensory functioning, and blood pressure.20 In this ancillary analysis based on various cortisol measures we studied the pharmacokinetic parameters of two different doses of HC in patients with secondary AI with the aim to better understand individual and population pharmacokinetics and the effects of HC dose adjustments.

Materials and Methods

Patients Sixty-three patients were recruited between May 2012 and June 2013 from the endocrine outpatient clinic at the University Medical Center Groningen, the Netherlands. Sixty of these patients completed the run-in phase and the baseline measurement (mean [SD] age, 52 [13], range 19–73, 35 males) 47 of them completed both study periods. One more patient was excluded due to violation of the study protocol, therefore 46 patients were included in the present analysis. All patients were diagnosed with secondary AI according to internationally accepted criteria and had been on stable glucocorticoid replacement therapy for at least six months. Other potential pituitary hormone deficiencies were adequately substituted and remained stable for at least six months prior to entry of the study and during the study. Renal and liver function tests were normal in all patients. Patients using medication known to alter cytochrome P450 (CYP3A4) activity (e.g. anti-epileptics) were excluded. Other inclusion and exclusion criteria have been described elsewhere.18 The study protocol was approved by the local ethics committee at the University Medical Center Groningen, The Netherlands. Patients gave written informed consent before entering the study. This study is registered with ClinicalTrial.gov, number NCT01546922.

Intervention Patients were randomly allocated to receive first a lower dose of 0.2–0.3 mg HC/kg body weight/day (total daily doses ranging from 15–20 mg HC depending on body weight) for ten weeks followed by a higher dose of 0.4–0.6 mg HC/kg body weight/ day (total daily doses ranging from 30–40 mg HC depending on body weight) for ten weeks (group 1) or in reversed order (group 2). HC substitution was given in three divided doses (before breakfast, before lunch, before dinner) with the highest dose given in the morning. The exact dosing scheme has been published previously.18 In case of intercurrent illness, patients were allowed to double or triple their dose, but not Population Pharmacokinetics of Hydrocortisone 145 for longer than 7 consecutive days. Patients were (strictly) instructed not to double or triple their dose in the week preceding the visit day.

Protocol After each 10 week treatment period patients attended the hospital outpatient clinic. On the visit days, patients were instructed to take their morning dose of HC at 0700 h. At 0800 h they attended to the hospital and blood samples were drawn in sitting position after a short period of rest for the measurement of plasma total cortisol and plasma free cortisol, which was repeated approximately five hours later. On the day preceding the hospital visit, patients collected 24-h urine for the measurement of 24-h urinary free cortisol and urinary free cortisone. In addition, seven patients collected saliva samples using a Salivette (Sarstedt, Nümbrecht, Germany) at 25 moments on one day, starting on the day preceding the scheduled hospital visit for the measurement of salivary cortisol.

Laboratory measurements Plasma free cortisol and plasma total cortisol, urinary free cortisol, urinary free cortisone as well as salivary cortisol were all measured by isotope dilution liquid chromatography tandem mass spectrometry (LC-MS/MS). Total and free cortisol in 7 plasma were performed essentially as described by Hawley et al., using cortisol-13C3 as internal standard.21 For plasma free cortisol, intra- and interassay coefficients of variability (CVs) were less than 6 %. For plasma total cortisol intra- and interassay CVs were less than 2.6 %. Plasma equilibrium dialysis for free cortisol was performed as described by Fiers et al.,22 with the only difference that we used 10 kD cellulose membranes (Harvard Apparatus; Holliston). Urinary free cortisol, urinary free cortisone and salivary cortisol assays were performed essentially as described by Jones et al.23 Intra-assay CV for salivary cortisol were less than 11.3 %. Serum CBG concentrations were determined in duplicate using a radioimmunoassay (IBL International GmbH, Hamburg, Germany). Albumin was measured using the brome cresol green method on a Roche Modular ISE/P (Roche Diagnostics, Mannheim, Germany).

Population pharmacokinetic analysis Using the Kinpop module of the pharmacokinetic software package MwPharm version 3.81 (Mediware, Zuidhorn, The Netherlands), a one-compartment and a two-compartment population model were parameterized for plasma free cortisol, plasma total cortisol and salivary cortisol taking the assay error into account. One- and two-compartment models were parameterized using an iterative two-stage Bayesian procedure and medians, means and standard deviations (SD) of the pharmacokinetic parameters total body clearance (CL), intercompartment clearance (CL12), apparent 146 Chapter 7

volume of distribution of the central compartment (V1) and apparent volume of dis-

tribution of the peripheral compartment (V2) were estimated. The elimination half-life

(t1/2) was calculated from 0.693/ke, where ke is the elimination rate constant. The ‘a priori’ models with fixed estimates for the absorption constant of 1.4 h-1 and bioavailability of 96 % based on existing literature were constructed.24 The ‘a priori’ model was fit to the data to construct the final models. The pharmacokinetic parameters for plasma free cortisol at the lower HC dose and the higher HC dose were compared. If there was no difference, the final one-and two-compartment models were constructed with combined data from both doses. Each model was validated with the Monte Carlo analysis using MwPharm. In this analysis, for all measures of cortisol, 100 patients were simulated per patient for the one-compartment and two-compartment models. Patient parameters were randomly calculated based on the estimated population parameter values. The relative root mean squared error (% RMSE), as a measure of precision of the one- and two-compartment models, was calculated. A RMSE below 15 % was considered to be sufficient.25 The model with the lowest RMSE was chosen for further pharmacokinetic analysis. Individual pharmacokinetic parameters were calculated by maximum a posteriori Bayesian estimation using the one-compartment final model for plasma free cortisol, plasma total cortisol and salivary cortisol for the two different doses of HC. CL,

volume of distribution (Vd), t1/2, maximum concentration (Cmax) and area under the

curve (AUC) after the first, second and third dose (Cmax1, AUC1, Cmax2, AUC2, Cmax3 and

AUC3 for the first, second and third dose respectively) were calculated. AUC24h was

determined as the sum of AUC1, AUC2 and AUC3. Three patients had cortisol levels that were higher 5 hours after intake of the hydrocortisone than after 1 hour. Therefore, these subjects, representing about 6 % of the study population, were not included in the population pharmacokinetic model. Other curves showing possible outliers after visual inspection were included, since population pharmacokinetic modeling and individual maximum a posteriori Bayesian estimation by iterative two-stage Bayesian analysis is robust for relatively small errors in registration.26 For the model for salivary cortisol no outliers were removed, as they did not influence the model parameter estimates.

Statistics The present analysis is a post-hoc pharmacokinetic analysis of a pharmacodynamic study.18 The power analysis performed for the study was based on the primary endpoint of the pharmacodynamic study: cognitive performance after the two different HC doses.18 A study arm with each 25 patients (50 patients in total) would be able to detect an effect size of 0.4 with α = 0.05 and β = 0.80. In order to allow a dropout rate of approximately 20 %, a total of 60 patients had to be included, as described previ- Population Pharmacokinetics of Hydrocortisone 147 ously.18 Data of all patients that completed the pharmacodynamic study were used for pharmacokinetic analysis. Data are presented as mean (SD) or median (interquartile range). Normality of the data was checked using QQ-plots and histograms. Because of skewness of the data, non-parametric tests were used. Individual pharmacokinetic parameters for plasma free cortisol, plasma total cortisol and salivary cortisol on the two different doses of HC were compared by using the Wilcoxon Signed Rank Test for paired observations. Individual pharmacokinetic parameters were compared between three weight groups by using the Kruskal-Wallis Test with post hoc pairwise comparisons. Correlations were calculated using Spearman’s correlation coefficient. All statistical analyses were performed using the Statistical Package for the Social Sciences version 22 (SPSS Inc., Armonk, N.Y., USA).

Table 1. Cortisol levels, CBG, albumin and glucocorticoid metabolites in patients with secondary adrenal insufficiency treated with two different doses of hydrocortisone (n = 46) Lower dose Higher dose P-value Plasma total cortisol levels 1 hr after ingestion (nmol/L) 495 [385; 598] 737 [671; 870]a < 0.001 5 hrs after ingestion (nmol/L) 112 [74; 195] 231 [170; 321] < 0.001 Plasma free cortisol levels 7 1 hr after ingestion (nmol/L) 30 [23; 44]a 70 [49; 80]c < 0.001 5 hrs after ingestion (nmol/L) 4 [2; 8]b 9 [6; 15]d 0.003 Serum CBG (µg/ml) 52.8 [49.1; 61.9]a 55.7 [48.1; 61.9]a 0.193 Serum albumin (g/L) 46 [45; 49] 48 [45; 50]a 0.010 Urinary free cortisol (nmol/24h) 78 [46; 112]a 281 [200; 411]e < 0.001 Urinary free cortisone (nmol/24h) 229 [147; 374]a 445 [326; 663]e < 0.001 Abbreviations: CBG: corticosteroid binding globulin. Data are median [interquartile range]. P-value lower dose versus higher dose by Wilcoxon Signed Rank Test for paired observations. a n = 45, b n = 38, c n = 43, d n = 37, e n = 44

Results

Patient sample Forty-six patients completed both study periods and were used in the present analysis (28 males, age 51 (14) years, range 19–73). The clinical characteristics of the study population have been published previously,18 a summary is added as Supplemental Table 1. 148 Chapter 7

Table 2. Population pharmacokinetic parameters and Monte Carlo Simulation of the lower and the higher dose combined Population estimates Monte Carlo Simulation Median Mean SD Mean 95 % CI RMSE (%) Plasma free cortisol (n = 89) CL (L/h) 244.13 235.78 108.22 236.22 217.83–254.61 2.5

Vd (L) 452.62 474.38 257.13 487.82 445.51–530.14 4.2

t1/2 (h) 1.28 1.39 1.65 1.43 1.42–1.44 Plasma total cortisol (n = 91) CL (L/h) 14.04 12.85 6.50 13.75 12.42–15.08 0.4

Vd (L) 41.75 39.82 14.85 39.97 37.24–42.70 0.5

t1/2 (h) 2.06 2.15 1.58 2.02 1.96–2.08 Salivary cortisol (n = 14) CL (L/h) 337.58 352.92 195.00 351.79 284.94–418.64 1.1

Vd (L) 486.25 348.37 445.02 447.35 229.21–665.50 10.4

t1/2 (h) 1.00 0.68 1.58 0.83 0.56–1.10 Abbreviations: CI: confidence interval; CL: total body clearance; n: number of curves included in the model; RMSE: Root Mean Squared Error; SD: standard deviation;

t1/2: elimination half-life; Vd: volume of distribution. The ‘a priori’ models were constructed with an absorption constant of 1.4 h-1 and bioavailability of 96 %

Cortisol, CBG and albumin The higher dose of HC resulted in significantly higher plasma total cortisol and plasma free cortisol concentrations (both 1 h and 5 h after ingestion of the morning dose), higher albumin, urinary free cortisol and urinary free cortisone concentrations compared to the lower dose of HC (Table 1). CBG concentrations did not change on the higher HC dose.

Population pharmacokinetics The population parameter estimates did not differ between the lower dose and the higher dose, therefore the data of both doses were combined to parameterize the final one- and two-compartment population models. Population pharmacokinetic estimates for plasma free cortisol, plasma total cortisol and salivary cortisol are given in Table 2. The one-compartment models gave a better description of the data than the two-compartment models as indicated by the lower RMSE (< 15 %) and thus were kept as the final models for the estimation of the individual parameters. There was

considerable inter-individual variation in CL and Vd for the models of plasma total and free cortisol as well as salivary cortisol (Table 2 and 3, Figure 1). This resulted in a high variability in cortisol concentration-time curves. In Figure 1 a simulation of the concentration-time curve (with 95 % confidence interval) is shown for plasma Population Pharmacokinetics of Hydrocortisone 149 total cortisol for a standard patient (weight of 70 kg, length of 1.75 m) based on the population pharmacokinetic parameters, for both the lower dose and the higher dose, illustrating the large inter-individual variation.

Figure 1. Concentration time profile of a standard patient (weight of 70 kg, length of 1.75 m). Median (thick solid and dashed line) with 95 % CI (thin solid and dashed lines). The 95 % CI is 7 based on the variability of the model parameters and the variability of the biochemical assays used. Arrows indicate time points of hydrocortisone administration.

Pharmacokinetic values on the two doses of hydrocortisone

CL and Vd were comparable between the two doses of HC for plasma free cortisol and salivary cortisol (each P > 0.05) (Table 3). CL and Vd for plasma total cortisol were significantly higher after receiving the higher dose of HC compared to the lower dose (P < 0.01). Furthermore, t1/2 was comparable between the lower and the higher dose for plasma free cortisol, and plasma total cortisol as well as for salivary cortisol. Cmax for plasma total cortisol and plasma free cortisol was significantly higher after the higher dose (eachP < 0.001), but this was only dose proportional for free cortisol concentrations and not for total cortisol. Thus, doubling of the

HC dose resulted in 2.1 ± 0.9 the Cmax1 for plasma free cortisol and 1.5 ± 0.5 for plasma total cortisol (P< 0.001). The AUC24h for plasma free cortisol was 2.2 ± 1.0 higher after doubling of the dose while this was 1.6 ± 0.6 for plasma total cortisol (P< 0.001). Large inter-individual variation was found, as AUC24hr for plasma free cortisol and plasma total cortisol on the same dose varied by a factor of approximately 10 (Figure 2A and B). To test whether age had an effect on our results, an additional correlation analysis was performed. For plasma free cortisol and plasma total cortisol, no pharmacokinetic parameter was associated with age, except for Vd of plasma total cortisol on the lower dose (r = 0.358, P = 0.016). 150 Chapter 7 : area under area : 1 0.89 96.71 339.62 247.81 152.80* 369.00* [0.56; 1.72] Higher dose Higher [52.75; 119.76] [155.25; 611.90] [123.37; 259.25] [289.32; 525.39] [177.33; 440.41] Salivary cortisol 0.91 65.05 28.53 133.45 619.07 389.61 [0.20; 1.79] Lower dose Lower [31.09; 81.57] [17.91; 47.79] [63.77; 165.88] [136.18; 674.11] [237.81; 643.26] 2.00 15.27** 50.33*** 754.94*** [1.78; 2.60] 2533.02*** 6433.07*** Higher dose Higher [11.93; 18.47] [11.93; [35.17; 58.76] [683.63; 903.35] [2173.09; 3206.01] [5260.17; 7786.22] : maximal concentration after the first dose of hydrocortisone, AUC hydrocortisone, of dose first the after concentration maximal : Plasma total cortisol max1 1.82 32.98 12.46 514.47 1743.93 3868.31 [1.40; 2.82] Lower dose Lower [8.80; 17.38] [26.01; 47.23] [390.28; 597.23] [1246.04; 2133.44] [2834.99; 5349.00] : 24-hour area under the curve. : 24-hour area under 24h 1.15 227.11 466.12 : volume of distribution, C distribution, of volume : 70.81*** 199.11*** 419.99*** d [0.98; 1.64] Higher dose Higher V [55.87; 80.78] [178.83; 311.93] [147.30; 247.25] [324.58; 515.76] [348.16; 607.55] 0.001 for lower dose vs higher dose by Wilcoxon Signed Rank Test for paired observations. Test Signed Rank Wilcoxon 0.001 for lower dose vs higher by

P Plasma free cortisol Plasma free 1.17 90.63 32.69 211.81 422.30 226.21 [0.99; 1.73] Lower dose Lower [25.13; 46.39] [68.34; 119.66] [144.69; 260.02] [315.88; 654.73] [168.20; 336.90] 0.01, ***

P Pharmacokinetic parameters for plasma free cortisol, plasma total cortisol and salivary cortisol for the two doses of hydrocortisone cortisol Pharmacokinetic parameters for plasma free cortisol, total

1 24h 0.05, ** (nmol/L)

< (h)

(L) max1 d P 1/2 AUC AUC t C (h*nmol/L) (h*nmol/L) CL (L/h) CL V Abbreviations: CL: total body clearance, body total CL: Abbreviations: Data are given as median [interquartile range]. * the curve after the first dose of hydrocortisone, AUC the curve after first dose of hydrocortisone, Table 3. Table Population Pharmacokinetics of Hydrocortisone 151

    

     

    

    



 

   

     7 Figure 2. AUC24h’s for plasma total cortisol (A) and plasma free cortisol (B) on the two doses of HC.

Weight-related dosing

AUCs and Cmax of plasma free cortisol and plasma total cortisol for the three daily dose administrations did not differ between the three weight groups on the higher dose (Supplemental Tables 2 and 3). The same was true for the lower dose, with the exception of a minor difference in cortisol exposure after the third gift in which the lowest weight group had higher Cmax3 and AUC3 compared to the highest weight group, both for plasma free cortisol and plasma total cortisol.

Associations between the various measures of cortisol

CL, Vd and AUC24h of free cortisol at the lower dose of HC correlated significantly with

CL, Vd and AUC24h of free cortisol at the higher dose (r = 0.476, r = 0.467, r = 0.449 for CL, Vd and AUC24h respectively, each P < 0.003). Similar but stronger correlations were found for total cortisol (r = 0.655, r = 0.540, and r = 0.634 for CL, Vd and AUC24h respectively, each P < 0.001).

AUC24h for plasma free cortisol correlated significantly with AUC24h for plasma total cortisol at the lower dose of HC (r = 0.872, P < 0.001) as well as at the higher dose (r = 0.752, P < 0.001, Figure 3A). Urinary free cortisol correlated significantly with AUC24h for plasma free cortisol (lower dose: r = 0.563, P < 0.001; higher dose: 152 Chapter 7

Table 4. Overview of literature assessing pharmacokinetic properties of hydrocortisone. Body Age weight Study Design N (year) (kg) Outcome measures Intervention Results

CL (L/h) Vd (L) t1/2 (h) Derendorf et al., Intervention 8 healthy 28 ± 9 74 ± 6 Pharmacokinetic Endogenous cortisol 18.2 ± 4.2 (L/h) 33.7 ± 4.6 (L) 1.82 ± 0.52 (h) 1991 study male parameters were estimated production was subjects for total cortisol using suppressed by the compartmental and non- administration of 4 mg compartmental analysis. dexamethasone at 2200 h the day preceding the study. 20 mg of HC was administered IV. Mah et al., 2004 Intervention 6 patients 46–68 56.6– Pharmacokinetic Patients received 10 mg 16.2 ± 1.62 (L/h) NA NA study with PAI (range) 112.4 parameters were estimated HC orally. 4 patients (range) for total cortisol with a with SAI user-defined model with P-Pharm software. Thomson et al., Intervention 9 47 ± 10 75 ± 14 Noncompartmental On day 1 patients Plasma IV: Plasma IV: Plasma IV: 2007 crossover patients with individual pharmacokinetic received 20 mg HC IV. 20.2 (16.3–23.1) (L/h) 61.9 (51.0–72.7) (L) 2.16 (1.94–2.40) (h) study PAI parameters were On day 2 patients Plasma oral: Plasma oral: 52.3 Plasma oral: 1.81 18 patients determined using received their own oral 19.1 (14.4–21.9) (L/h) (40.6–57.4) (L) (1.69–2.36) (h) with SAI WinNonlin. HC dose (10, 15 or 20 Saliva IV: Saliva IV: Saliva IV: mg HC). 256 (194–376) (L/h) 608 (338–1138) (L) 1.74 (1.01–2.37) (h) Saliva oral: Saliva oral: 457 Saliva oral: 1.47 215 (181–292) (L/h) (331–774) (L) (1.24–1.72) (h) Simon et al., Prospective, 20 patients 53 ± 17 74 ± 18 Population pharmacokinetic PAI patients received 28 12.1 (SE (%) 6.1) 38.7 (SE (%) 8.30) 1.7 (95 % CI 2010 open-label with PAI modelling for total cortisol mg (median 30 [range (L/h) (L) 1.5–1.9) (h) study 30 patients was done by NONMEN. 20–50] mg) HC (0.41 mg/ with SAI One and two-compartment kg body weight). models were calculated. SAI patients received 23 mg (median 20 [range 15–30] mg) HC (0.32 mg/ kg body weight). Present study Intervention 46 51 ± 15 83 ± 14 Population pharmacokinetic Patients received a lower Plasma total cortisol: Plasma total cortisol: Plasma total study (RCT patients with models were calculated for dose for 10 weeks (0.2– 12.85 ± 6.50 (L/h) 39.82 ± 14.85 (L) cortisol: double-blind SAI plasma free cortisol, plasma 0.3 mg/kg body weight) Plasma free cortisol: Plasma free cortisol: 2.15 ± 1.58 (h) crossover) total cortisol and salivary followed by a higher dose 235.78 ± 108.22 (L/h) 474.38 ± 257.13 (L) Plasma free cortisol: cortisol. for 10 weeks (0.4–0.6 Saliva cortisol: Saliva cortisol: 1.39 ± 1.65 (h) mg/kg/body weight). Or 352.92 ± 195.00 348.37 ± 445.02 (L) Saliva cortisol: in reversed order. (L/h) 0.68 ± 1.58 (h)

CL = clearance, Vd = volume of distribution, t1/2 = elimination half-life, HD = hydrocortisone, IV = intravenous, PAI = primary adrenal insufficiency, SAI = secondary adrenal insufficiency. RCT = randomized controlled trial. Data are presented as mean ± SD or median (interquartile range), unless otherwise stated. Population Pharmacokinetics of Hydrocortisone 153

Table 4. Overview of literature assessing pharmacokinetic properties of hydrocortisone. Body Age weight Study Design N (year) (kg) Outcome measures Intervention Results

CL (L/h) Vd (L) t1/2 (h) Derendorf et al., Intervention 8 healthy 28 ± 9 74 ± 6 Pharmacokinetic Endogenous cortisol 18.2 ± 4.2 (L/h) 33.7 ± 4.6 (L) 1.82 ± 0.52 (h) 1991 study male parameters were estimated production was subjects for total cortisol using suppressed by the compartmental and non- administration of 4 mg compartmental analysis. dexamethasone at 2200 h the day preceding the study. 20 mg of HC was administered IV. Mah et al., 2004 Intervention 6 patients 46–68 56.6– Pharmacokinetic Patients received 10 mg 16.2 ± 1.62 (L/h) NA NA study with PAI (range) 112.4 parameters were estimated HC orally. 4 patients (range) for total cortisol with a with SAI user-defined model with P-Pharm software. Thomson et al., Intervention 9 47 ± 10 75 ± 14 Noncompartmental On day 1 patients Plasma IV: Plasma IV: Plasma IV: 2007 crossover patients with individual pharmacokinetic received 20 mg HC IV. 20.2 (16.3–23.1) (L/h) 61.9 (51.0–72.7) (L) 2.16 (1.94–2.40) (h) 7 study PAI parameters were On day 2 patients Plasma oral: Plasma oral: 52.3 Plasma oral: 1.81 18 patients determined using received their own oral 19.1 (14.4–21.9) (L/h) (40.6–57.4) (L) (1.69–2.36) (h) with SAI WinNonlin. HC dose (10, 15 or 20 Saliva IV: Saliva IV: Saliva IV: mg HC). 256 (194–376) (L/h) 608 (338–1138) (L) 1.74 (1.01–2.37) (h) Saliva oral: Saliva oral: 457 Saliva oral: 1.47 215 (181–292) (L/h) (331–774) (L) (1.24–1.72) (h) Simon et al., Prospective, 20 patients 53 ± 17 74 ± 18 Population pharmacokinetic PAI patients received 28 12.1 (SE (%) 6.1) 38.7 (SE (%) 8.30) 1.7 (95 % CI 2010 open-label with PAI modelling for total cortisol mg (median 30 [range (L/h) (L) 1.5–1.9) (h) study 30 patients was done by NONMEN. 20–50] mg) HC (0.41 mg/ with SAI One and two-compartment kg body weight). models were calculated. SAI patients received 23 mg (median 20 [range 15–30] mg) HC (0.32 mg/ kg body weight). Present study Intervention 46 51 ± 15 83 ± 14 Population pharmacokinetic Patients received a lower Plasma total cortisol: Plasma total cortisol: Plasma total study (RCT patients with models were calculated for dose for 10 weeks (0.2– 12.85 ± 6.50 (L/h) 39.82 ± 14.85 (L) cortisol: double-blind SAI plasma free cortisol, plasma 0.3 mg/kg body weight) Plasma free cortisol: Plasma free cortisol: 2.15 ± 1.58 (h) crossover) total cortisol and salivary followed by a higher dose 235.78 ± 108.22 (L/h) 474.38 ± 257.13 (L) Plasma free cortisol: cortisol. for 10 weeks (0.4–0.6 Saliva cortisol: Saliva cortisol: 1.39 ± 1.65 (h) mg/kg/body weight). Or 352.92 ± 195.00 348.37 ± 445.02 (L) Saliva cortisol: in reversed order. (L/h) 0.68 ± 1.58 (h)

        

       

           

        

     

           

        

     

          

Figure 3. Correlation of plasma free cortisol AUC24h with plasma total cortisol AUC24h (A), plas-

ma free cortisol AUC24h with urinary free cortisol (B), and plasma total cortisol AUC24h with urinary free cortisol (C). * P < 0.05, ** P < 0001 Population Pharmacokinetics of Hydrocortisone 155

r = 0.359, P < 0.05, Figure 3B) and with AUC24h for plasma total cortisol (lower dose: r = 0.623, P < 0.001; higher dose: r = 0.381, P < 0.05, Figure 3C).

Discussion

Our results show that HC pharmacokinetics can differ up to 10-fold between individuals. A doubling of the HC dose results only in a doubling of the plasma free cortisol concentrations, whereas the increase in total cortisol concentrations was lower. Fur- thermore, in patients that showed relatively high CL and low AUC compared to other patients in the present study, designated as fast metabolizers, a doubling of the dose does not result in markedly enhanced cortisol exposure during large parts of the day. Based on pharmacokinetic characteristics, HC dosing should therefore be individualized in case of maintenance therapy. Attention should be paid to stress-dosing since doubling of the dose does not necessarily result in doubled exposure, and fast-metabolizers may not benefit from it in terms of marked increments of plasma cortisol concentrations. This is the largest study to date on pharmacokinetics of immediate release HC. Our extensive laboratory determinations together with pharmacokinetic analysis provide data on free and total cortisol in both plasma and saliva under two HC dose condi- 7 tions. In addition, other measures of HC metabolism like urinary free cortisol, CBG and albumin were taken into account. Another strength of the present study is that pharmacokinetic variables were calculated based on plasma cortisol concentrations measured with the LC-MS/MS, instead of previously used immunoassays, providing greater analytical specificity.27 A few limitations of the study need to be addressed. First, our population pharmacokinetic model was based on cortisol concentration measurements of two time-points, whereas most pharmacokinetic studies are based on elaborate regularly timed cortisol measurements. However, the Kinpop module of MwPharm using the Bayesian procedure is developed specifically for parameterizing population pharmacokinetic models based on scarce data from a relatively large patient sample, thereby circumventing this problem. In addition, the pharmacokinetic outcome variables were comparable to other studies (Table 4), and even related to urinary free cortisol concentrations. In agreement, a study by Rousseau and colleagues showed that single point cortisol concentrations 28 correlated significantly with cortisol AUC0800–1900hr in patients with primary AI. Thus, even without extensive individual sampling we could reliably perform a pharmacokinetic analysis. Second, although we tried to standardize the salivary cortisol collection by using protocols, sampling was done at home without supervision of the investigators. In addition, data on salivary cortisol were performed in a subset of patients and not in the entire study population, which could have led to missing a true effect due to 156 Chapter 7

the lack of statistical power. In future studies, it would be preferable to collect salivary cortisol in a larger patient sample along with blood sampling at the same time points. Third, CBG binding capacity varies dependent on the cortisol concentrations and this variation is known to influence pharmacokinetic parameters of cortisol.14 MwPharm is unable to take this variation in protein binding capacity into account and calculate its effect on CL. Consequently, linear models with a relatively higher variance are parameterized for total cortisol instead of non-linear models. Future studies should use a program incorporating this variability in protein binding. In agreement with other studies, we found large inter-individual differences in pharmacokinetic parameters.1,11 Several factors can influence the variability in pharmacokinetic values, for instance variation in CYP3A4 enzyme activity. CYP3A4 is an important enzyme in the metabolism of cortisol and its activity can vary 5 to 20- fold inter-individually.29 CYP3A4 activity can be induced by drugs like rifampicin and several anticonvulsant agents, and inhibited by drugs such as azole antifungal agents and macrolide antibiotics. We did not measure CYP3A4 activity nor determined CYP3A4 genotypes, however, we excluded patients using these enzyme inducers or inhibitors. We also corrected for body weight, which is known to influence cortisol clearance.17 Indeed, the administration of weight-adjusted doses resulted in highly comparable cortisol exposure between weight groups. However, on an individual basis variability in cortisol exposure was still substantial, indicating the need for individual dose adjustments. Approximately 6 % of our patients had higher cortisol levels 5 hours after administration compared to 1 hour after administration. These patients may have had very slow resorption of cortisol and/or very slow elimination. Residual cortisol production to explain this seems very unlikely since morning (0800 h) cortisol concentration are invariably higher than later during day time. Based on population pharmacokinetics, a substantial percentage of patients is predicted to have plasma total cortisol levels well below 100 nmol/L during large parts of the day (Figure 1). Although the functioning of the hypothalamic-pituitary-adrenal axis is not only represented by cortisol concentrations in plasma, in patients this plasma concentration is frequently used as a cut-off level below which adrenal insufficiency is diagnosed. Thus, when complaints compatible with undersubstitution are present and they are accompanied by low plasma cortisol concentrations, it appears reasonable to raise cortisol concentrations by dose adjustments. However, one must bear in mind that other determinants of the cortisol metabolism, like 11β-hydroxysteroid dehydrogenase activity, cortisone concentrations, intracellular concentrations, and the effects of cortisol efflux transporters, must be considered. Another point of attention is that doubling of the dose does not necessarily increase cortisol concentrations by a factor of 2. This is also found in previous studies, showing a non-proportional increase in cortisol exposure after increasing doses, both after Population Pharmacokinetics of Hydrocortisone 157 administration of immediate release tablets 30,31 and after administration of dual-release HC tablets.3,32 However, different results are found when looking at free cortisol. Our results show that for plasma free cortisol AUC24h values were dose proportional to the

HC dose administered. For AUC24h values for salivary cortisol this was even higher, as they increased by factor 3 on the higher dose. Whether this difference in behavior in response to dose increments is relevant remains elusive. The free fraction of cortisol is expected to represent the biologically active part of cortisol, but CBG, that binds approximately 80 % of cortisol, is not only known as an inactive carriers, but also as a protein that delivers cortisol when its binding capacity is decreased upon cleav- ing of the reactive center loop by proteases,33 assisting in tissue specific delivery of cortisol. It could be surmised that for patients with persistent complaints compatible with undersubstitution, not an increase in total daily dose but an increase in the number of daily doses or the administration of modified release HC would be beneficial if fast metabolism of cortisol is a contributing mechanism. The non-proportional total cortisol exposure after a doubling of the dose may be a consequence of increased clearance of cortisol due to saturation of CBG. Saturation of CBG occurs at total plasma cortisol concentrations of 450–550 nmol/L,34,35 and consequently, the higher dose of oral HC used readily saturates CBG, leading to increases in free plasma cortisol, salivary cortisol and urinary cortisol.13 This is accompanied 7 by a significant increase in CL of plasma total cortisol (Table 3) to restore the balance resulting in a relatively decreased total cortisol exposure. It is known that aging may have pharmacokinetic implications36. Aging is associated with reduced fist-pass metabolism, increasing bioavailability of a few drugs. With advancing age body fat increases and total body water and lean body mass decreases, leading to an increased volume of distribution and prolonged half-life for lipophilic drugs such as cortisol. Furthermore, a reduction in renal and hepatic clearance is found with advancing age. In our study with patients aged between 19–73 year we found minor evidence that this influenced our results. Only total cortisol Vd on the lower dose was positively associated with age, while other pharmacokinetic indices on both different doses remained non-significant. In conclusion, pharmacokinetics of HC can differ 10-fold between individuals. A doubling of the dose does not result in a doubling of total cortisol exposure. Further- more, fast-metabolizers may not benefit from dose increments with regard to improved concentrations-time curves, and other managing strategies should be considered. Due to the large inter-individual variation individually tailored treatment doses and regimens are required. 158 Chapter 7

References

1. Simon N, Castinetti F, Ouliac F, Lesavre N, Brue T, Oliver C. Pharmacokinetic evidence for suboptimal treatment of adrenal insufficiency with currently available hydrocortisone tablets. Clin Pharmacokinet 2010;49(7):455-463. 2. Ross IL, Levitt NS, Van der Walt JS, et al. Salivary cortisol day curves in Addison’s disease in patients on hydrocortisone replacement. Horm Metab Res 2013;45(1):62-68. 3. Johannsson G, Bergthorsdottir R, Nilsson AG, Lennernas H, Hedner T, Skrtic S. Improving glucocorticoid replacement therapy using a novel modified-release hydrocortisone tablet: a pharmacokinetic study. Eur J Endocrinol 2009;161(1):119-130. 4. Newell-Price J, Whiteman M, Rostami-Hodjegan A, et al. Modified-release hydrocortisone for circadian therapy: a proof-of-principle study in dexamethasone-suppressed normal volunteers. Clin Endocrinol (Oxf) 2008;68(1):130-135. 5. Arlt W, Rosenthal C, Hahner S, Allolio B. Quality of glucocorticoid replacement in adrenal insufficiency: clinical assessment vs. timed serum cortisol measurements. Clin Endocrinol (Oxf) 2006;64(4):384-389. 6. Debono M, Ross RJ. What is the best approach to tailoring hydrocortisone dose to meet patient needs in 2012? Clin Endocrinol (Oxf) 2013;78(5):659-664. 7. Mills JL, Schonberger LB, Wysowski DK, et al. Long-term mortality in the United States cohort of pituitary-derived growth hormone recipients. J Pediatr 2004;144(4):430-436. 8. Lovas K, Gjesdal CG, Christensen M, et al. Glucocorticoid replacement therapy and pharmacoge- netics in Addison’s disease: effects on bone. Eur J Endocrinol 2009;160(6):993-1002. 9. Perogamvros I, Keevil BG, Ray DW, Trainer PJ. Salivary cortisone is a potential biomarker for serum free cortisol. J Clin Endocrinol Metab 2010;95(11):4951-4958. 10. Howlett TA. An assessment of optimal hydrocortisone replacement therapy. Clin Endocrinol (Oxf) 1997;46(3):263-268. 11. Thomson AH, Devers MC, Wallace AM, et al. Variability in hydrocortisone plasma and saliva pharmacokinetics following intravenous and oral administration to patients with adrenal insufficiency. Clin Endocrinol (Oxf) 2007;66(6):789-796. 12. Restituto P, Galofre JC, Gil MJ, et al. Advantage of salivary cortisol measurements in the diagnosis of glucocorticoid related disorders. Clin Biochem 2008;41(9):688-692. 13. Jung C, Greco S, Nguyen HH, et al. Plasma, salivary and urinary cortisol levels following physiological and stress doses of hydrocortisone in normal volunteers. BMC Endocr Disord 2014;14:91- 6823-14-91. 14. Perogamvros I, Aarons L, Miller AG, Trainer PJ, Ray DW. Corticosteroid-binding globulin regulates cortisol pharmacokinetics. Clin Endocrinol (Oxf) 2011;74(1):30-36. 15. Oksnes M, Ross R, Lovas K. Optimal glucocorticoid replacement in adrenal insufficiency. Best Pract Res Clin Endocrinol Metab 2015;29(1):3-15. 16. Derendorf H, Mollmann H, Barth J, Mollmann C, Tunn S, Krieg M. Pharmacokinetics and oral bioavailability of hydrocortisone. J Clin Pharmacol 1991;31(5):473-476. 17. Mah PM, Jenkins RC, Rostami-Hodjegan A, et al. Weight-related dosing, timing and monitoring hydrocortisone replacement therapy in patients with adrenal insufficiency. Clin Endocrinol (Oxf) 2004;61(3):367-375. 18. Werumeus Buning J, Brummelman P, Koerts J, et al. The effects of two different doses of hydrocortisone on cognition in patients with secondary adrenal insufficiency - Results from a randomized controlled trial. Psychoneuroendocrinology 2015;55:36-47. Population Pharmacokinetics of Hydrocortisone 159

19. Werumeus Buning J, Brummelman P, Koerts J, et al. Hydrocortisone Dose Influences Pain, Depres- sive Symptoms and Perceived Health in Adrenal Insufficiency - A Randomized Controlled Trial. Neuroendocrinology 2016;103(6):771-8. 20. Werumeus Buning J, van Faassen M, Brummelman P, et al. Effects of Hydrocortisone on the Regula- tion of Blood Pressure: Results From a Randomized Controlled Trial. J Clin Endocrinol Metab 2016;101(10):3691-3699. 21. Hawley JM, Owen LJ, MacKenzie F, Mussell C, Cowen S, Keevil BG. Candidate Reference Mea- surement Procedure for the Quantification of Total Serum Cortisol with LC-MS/MS. Clin Chem 2016;62(1):262-269. 22. Fiers T, Delanghe J, T’Sjoen G, Van Caenegem E, Wierckx K, Kaufman JM. A critical evaluation of salivary testosterone as a method for the assessment of serum testosterone. Steroids 2014;86:5-9. 23. Jones RL, Owen LJ, Adaway JE, Keevil BG. Simultaneous analysis of cortisol and cortisone in saliva using XLC-MS/MS for fully automated online solid phase extraction. J Chromatogr B Analyt Technol Biomed Life Sci 2012;881-882:42-48. 24. Czock D, Keller F, Rasche FM, Haussler U. Pharmacokinetics and pharmacodynamics of systemi- cally administered glucocorticoids. Clin Pharmacokinet 2005;44(1):61-98. 25. Alffenaar JW, Kosterink JG, van Altena R, van der Werf TS, Uges DR, Proost JH. Limited sampling strategies for therapeutic drug monitoring of linezolid in patients with multidrug-resistant tubercu- losis. Ther Drug Monit 2010;32(1):97-101. 26. van der Meer AF, Touw DJ, Marcus MA, Neef C, Proost JH. Influence of erroneous patient records on population pharmacokinetic modeling and individual bayesian estimation. Ther Drug Monit 2012;34(5):526-534. 7 27. Kushnir MM, Rockwood AL, Bergquist J. Liquid chromatography-tandem mass spectrometry applications in endocrinology. Mass Spectrom Rev 2010;29(3):480-502. 28. Rousseau E, Joubert M, Trzepla G, et al. Usefulness of Time-Point Serum Cortisol and ACTH Measurements for the Adjustment of Glucocorticoid Replacement in Adrenal Insufficiency. PLoS One 2015;10(8):e0135975. 29. Wilkinson GR. Cytochrome P4503A (CYP3A) metabolism: prediction of in vivo activity in humans. J Pharmacokinet Biopharm 1996;24(5):475-490. 30. Toothaker RD, Craig WA, Welling PG. Effect of dose size on the pharmacokinetics of oral hydrocortisone suspension. J Pharm Sci 1982;71(10):1182-1185. 31. Toothaker RD, Sundaresan GM, Hunt JP, et al. Oral hydrocortisone pharmacokinetics: a comparison of fluorescence and ultraviolet high-pressure liquid chromatographic assays for hydrocortisone in plasma. J Pharm Sci 1982;71(5):573-576. 32. Johannsson G, Lennernas H, Marelli C, Rockich K, Skrtic S. Achieving a physiological cortisol profile with once-daily dual-release hydrocortisone: a pharmacokinetic study. Eur J Endocrinol 2016. 33. Zhou A, Wei Z, Stanley PL, Read RJ, Stein PE, Carrell RW. The S-to-R transition of corticosteroid- binding globulin and the mechanism of hormone release. J Mol Biol 2008;380(1):244-251. 34. Aardal E, Holm AC. Cortisol in saliva--reference ranges and relation to cortisol in serum. Eur J Clin Chem Clin Biochem 1995;33(12):927-932. 35. Tunn S, Mollmann H, Barth J, Derendorf H, Krieg M. Simultaneous measurement of cortisol in serum and saliva after different forms of cortisol administration. Clin Chem 1992;38(8 Pt 1):1491- 1494. 36. Mangoni AA, Jackson SH. Age-related changes in pharmacokinetics and pharmacodynamics: basic principles and practical applications. Br J Clin Pharmacol 2004;57(1):6-14. 160 Chapter 7

Supplemental data

Supplemental Table 1. Clinical characteristics at baseline of pituitary patients treated for secondary adrenal insufficiency (N = 46) Age (y), median [IQR] 54 [42; 61] Sex (males/females), n 28/18 Educational level (1/2/3/4/5/6/7) a 0/2/1/6/20/15/2 Body weight (kg), median [IQR] 84.0 [73.4; 93.1] Hydrocortisone treatment prior to randomization Total daily dose (mg/day), median [IQR] 25 [20; 30] Total dose/kg body weight/day (mg/kg/day), median [IQR] 0.32 [0.25; 0.35] Number of daily dosings (1/2/3), n 3/33/10 Duration of hydrocortisone treatment (y), median [IQR] 12 [5; 23] No. of hormonal replacements (1/2/3/4/5) 3/9/20/11/3 Thyroid hormone deficiency (% of patients) 91 Growth hormone deficiency (% of patients) 67 Growth hormone deficiency (% patients receiving substitution) 68 Sex hormone deficiency (% of patients) 59 Men: testosterone (% of patients receiving substitution) 69 Premenopausal women, n = 8: estrogens (% of patients receiving 50 substitution) Postmenopausal women, n = 10: estrogens NA Desmopressin (% of patients) 17 Abbreviations: IQR: interquartile range; NA: not applicable. Population Pharmacokinetics of Hydrocortisone 161

Supplemental Table 2. Pharmacokinetic parameters for the three weight groups for plasma free cortisol. 50–74 kg 75–84 kg 85–100 kg Lower dose n = 14 n = 9 n = 23 P-value

Cmax1 (nmol/L) 35.99 33.39 32.58 0.611 [25.73; 48.22] [28.15; 55.96] [21.67; 43.40]

AUC1 (h*nmol/L) 104.52 83.98 85.91 0.417 [75.70; 129.87] [70.90; 138.23] [61.00; 116.66]

Cmax2 (nmol/L) 25.95 20.64 24.51 0.683 [19.80; 34.43] [15.86; 32.31] [18.29; 35.67]

AUC2 (h*nmol/L) 79.56 64.16 74.02 0.244 [60.63; 111.01] [36.21; 78.17] [49.51; 93.87]

Cmax3 (nmol/L) 15.13 10.56 10.16 0.025* [10.70; 20.96] [8.08; 15.30] [7.43; 13.50]

AUC3 (h*nmol/L) 43.66 28.92 28.27 0.020* [37.40; 80.99] [18.90; 46.67] [23.72; 40.58]

AUC24h (h*nmol/L) 236.30 177.43 192.01 0.248 [173.18; 361.36] [126.18; 258.96] [128.15; 251.35] 50–74 kg 75–84 kg 85–100 kg Higher dose n = 14 n = 9 n = 23 P-value

Cmax1 (nmol/L) 71.51 81.43 64.72 0.175 7 [50.71; 77.79] [67.85; 113.29] [52.75; 77.95]

AUC1 (h*nmol/L) 200.96 225.51 187.37 0.441 [152.90; 222.44] [171.78; 269.68] [142.51; 247.64]

Cmax2 (nmol/L) 53.19 48.12 54.86 0.450 [38.29; 57.09] [36.33; 60.22] [44.10; 68.87]

AUC2 (h*nmol/L) 143.14 121.94 156.03 0.322 [113.08; 164.33] [91.88; 157.73] [123.60; 196.91]

Cmax3 (nmol/L) 28.52 24.66 21.68 0.243 [19.97; 30.46] [18.56; 29.96] [16.44; 28.03]

AUC3 (h*nmol/L) 83.87 64.65 67.33 0.204 [67.18; 97.12] [47.91; 92.19] [49.44; 91.36]

AUC24h (h*nmol/L) 421.63 411.77 416.46 0.999 [332.48; 478.43] [311.72; 515.64] [313.88; 538.62]

Abbreviations: Cmax1: maximum concentration after the first administration; AUC1: area under the curve after the first administration; Cmax2: maximum concentrations after the second administration; AUC2: area under the curve after the second administration; Cmax3: maximum concentrations after the third administration; AUC3: area under the curve after the third administration; AUC24h: 24 hour area under the curve. * Statistical significant difference between the lowest weight group (50–74 kg) and the highest weight group (85–100 kg). 162 Chapter 7

Supplemental Table 3. Pharmacokinetic parameters for the three weight groups for plasma total cortisol. 50–74 kg 75–84 kg 85–100 kg Lower dose n = 14 n = 9 n = 23 P-value

Cmax1 (nmol/L) 541.67 521.54 482.99 0.187 [438.42; 760.91] [464.69; 639.89] [353.08; 595.06]

AUC1 (h*nmol/L) 1984.34 1502.67 1504.10 0.144 [1381.15; 2728.00] [1297.27; 2070.43] [1092.05; 1872.48]

Cmax2 (nmol/L) 483.15 393.03 386.16 0.136 [391.14; 702.84] [264.56; 437.88] [282.45; 558.23]

AUC2 (h*nmol/L) 1634.98 1144.10 1183.09 0.107 [1141.14; 2686.29] [715.02; 1590.25] [954.31; 1720.02]

Cmax3 (nmol/L) 286.32 179.35 202.27 0.016* [187.56; 487.77] [135.59; 266.48] [147.55; 252.85]

AUC3 (h*nmol/L) 1223.90 535.60 662.24 0.054 [651.07; 2426.45] [394.78; 1321.44] [489.34; 1101.74]

AUC24h (h*nmol/L) 4862.89 3211.18 3759.60 0.128 [3075.09; 8201.59] [2441.57; 4933.21] [2850.41; 4745.90] 50–74 kg 75–84 kg 85–100 kg Higher dose n = 14 n = 9 n = 23 P-value

Cmax1 (nmol/L) 759.62 906.95 729.46 0.148 [731.81; 925.33] [673.64; 1156.61] [646.69; 783.03]

AUC1 (h*nmol/L) 2685.46 3013.60 2409.28 0.312 [2315.51; 3466.15] [2116.62; 3530.51] [2149.96; 2839.97]

Cmax2 (nmol/L) 677.53 628.76 687.50 0.656 [606.59; 824.15] [412.69; 757.10] [600.88; 760.96]

AUC2 (h*nmol/L) 227.37 1945.96 2332.25 0.402 [1932.39; 3200.90] [1279.56; 2554.02] [1873.00; 2671.04]

Cmax3 (nmol/L) 392.70 319.24 328.46 0.060 [336.35; 561.90] [219.71; 410.78] [255.88; 380.34]

AUC3 (h*nmol/L) 1598.97 1138.32 1489.74 0.064 [1282.61; 2916.63] [771.14; 1699.39] [918.88; 1859.88]

AUC24h (h*nmol/L) 6561.79 5974.13 5880.28 0.534 [5548.51; 9391.69] [4167.31; 7783.91] [5022.00; 7101.21]

Abbreviations: Cmax1: maximum concentration after the first administration; AUC1: area under the

curve after the first administration; Cmax2: maximum concentrations after the second administra-

tion; AUC2: area under the curve after the second administration; Cmax3: maximum concentrations

after the third administration; AUC3: area under the curve after the third administration; AUC24h: 24 hour area under the curve. * Statistical significant difference between the lowest weight group (50–74 kg) and the highest weight group (85–100 kg)

Chapter 8

General discussion, conclusions and future perspectives

General Discussion and Conclusions 167

A 67-year-old woman was seen at our endocrine outpatient clinic for her yearly check-up. Two years earlier a non-functioning pituitary adenoma was removed by transsphenoidal surgery due to mass effects on the optic chiasm. After surgery, her vision improved, but her anterior pituitary gland was damaged resulting in pituitary insufficiency, necessitating hormonal suppletion therapy. She received substitution therapy with thyroid hormone, growth hormone and hydrocortisone (HC). The dose of the latter medication was 10 + 5 + 2.5 mg divided over the day. Since her last visit she mentioned that she was not feeling well. She reported regular complaints of headache and muscle and joint pain. She felt tired, restless, emotionally unstable and a bit depressed from time to time. She had trouble falling and staying asleep. She described her overall well-being as “it’s just not right”. Her partner confirmed the reported complaints and insisted that something had to change. Her endocrinologist acknowledged her problems and evaluated her options. Laboratory measurements showed an fT4 level of 16 pmol/L and IGF-1 Z-score of +0.4. Cortisol level measured at 14.15 h was 125 nmol/L after a hydrocortisone dose of 5 mg taken at 12.00 h. She had been treated with hydrochlorothiazide 25 mg once daily for hypertension and this resulted in good control of her blood pressure. After surgery, she gained 5 kg while she already weighed 90 kg, and this increase in body weight was predominantly noticeable in the abdominal region. When reviewing her medical record it was also noticed that her younger sister was diagnosed with osteoporosis. In light of her physical complaints that are non-specific, but substantial enough to have an impact on her life and suggestive of under-substitution, it was investigated 8 whether the dose of HC underlay her reported complaints. Her high blood pressure, increase in body weight and the presence of osteoporosis in the family discouraged her specialist from increasing her HC dose thus far. However, her current quality of life was impaired and together with her endocrinologist they decided to experiment with increasing the HC dose.

This anecdote shows that the considerations for the optimal substitution dose can be difficult with several pros and cons for HC dose increments. Therefore, in this thesis we aimed at contributing to the discussion regarding the treatment of secondary adrenal insufficiency (SAI) by adding evidence for current guidelines about substitution doses. In order to do so, we have performed a randomized double-blind crossover trial in which patients with SAI were treated with two different doses of HC. 168 Chapter 8

Part I. The effect of hydrocortisone treatment on psychological outcome measures in patients with secondary adrenal insufficiency

Hydrocortisone and cognitive functioning High levels of cortisol have often been linked to impaired cognitive performance. An inverted “U”-shape relationship between cortisol level and cognitive performance is suggested, with moderate cortisol levels being necessary for proper cognitive functioning, and very low or very high levels being impairing.1,2 We assessed cognitive functioning after 10 weeks of treatment with two different doses of HC, and showed that a doubling of the HC dose did not affect cognitive functioning in the domains of memory, attention, executive functioning and social cognition (chapter 2). On the higher dose of HC, patients tended to score worse compared to the lower dose on short-term memory for verbal non-coherent information and tended to show increased variability in reaction time for phasic alertness, however, the overall pattern of cognitive performance showed no indication of cognitive deterioration on the higher dose of HC as compared to the lower dose. Even though these results are reassuring, there are several studies, both in animals and in humans, indicating that prolonged exposure to excessive levels of endogenous or exogenous cortisol is disadvantageous for cognitive functioning and, over time, leads to structural changes in the brain.3–5 For instance, patients with active Cushing’s disease report impairments in memory, visual and spatial information processing, reasoning, language performance, verbal learning, and attention,6–9 which is accompanied by structural changes in the brain such as cerebral cortical atrophy and hippocampal volume reduction.10,11 Moreover, these impairments and structural changes persist even after long-term remission of Cushing’s syndrome.12,13 Treatment with exogenous GCs can also lead to changes in brain structure, as white-matter abnormalities and temporal lobe atrophy were reported in patients with primary adrenal insufficiency (PAI) receiving GC substitution therapy.14 The temporary administration of a higher HC dose might thus not be detrimental for cognitive functioning, however, there is abundant evidence prompting us to be cautious about generalizing our results to longer time periods.

Two receptor types in the brain Cortisol can bind to glucocorticoid receptors (GRs) and mineralocorticoid receptors (MRs). Both receptors are found in the brain. Neurons in the hippocampus contain both receptor types, whereas other parts of the brain mainly contain GRs. The finding that stress and accompanying high levels of cortisol impair cognitive performance whereas moderate cortisol levels enhance cognition, can be explained by differential activation of these two receptors. General Discussion and Conclusions 169

Cortisol has a 6–10 fold higher affinity for the MR,15 resulting in high MRs oc- cupancy even under basal, non-stressful situations.16,17 GRs become activated when cortisol levels increase, i.e. after stress exposure or at the circadian peak. It was found that when cortisol levels are mildly elevated, so when most of the MRs and some of the GRs are activated, long-term potentiation (LTP) is enhanced. LTP is the reinforcement of synaptic contacts contributing to the storage of information.18 When cortisol levels increase further and GRs become extensively activated, LTP is reduced and instead, long-term depression (reducing synaptic contact) is enhanced, resulting in the opposite effect of LTP.18 Both receptor types are involved in different aspects of learning and memory. It was found that MRs play a role in the initial behavioral reaction in novel situations, whereas GRs are involved in the consolidation of learned information.18 In addition, the consequence of activation of MRs and GRs depends on the context, i.e. the envi- ronmental input, during the various stages of memory processing: acquisition, consolidation and retrieval.18 Both effects of GRs and MRs mediate the hippocampal circuits that underlie complex learning paradigms such as spatial learning. Some GR activation is necessary for the consolidation of learned information, but additional activation of GRs due to a stressor that is out of context with respect to the original learning task, disrupts consolidation and favors a more-opportune response. Thus, MRs and GRs mediate different effects, but they interact and proceed in a coordinated manner.18 Studies assessing the effects of GR and MR occupation were mostly performed in animals or healthy individuals. However, results of these studies cannot uniformly be 8 generalized to patients with AI, since regulation of the hypothalamic-pituitary-adrenal axis is disturbed in these patients. It would be interesting to study receptor activation in patients with AI, in which cortisol levels can be controlled externally. Coopera- tion between MRs and GRs is necessary to preserve cognitive performance, and even though we did not find differences on the current HC doses used, demands may be different when we expose patients to stress and thereby increase cortisol demand. One could speculate that the lower dose of HC as administered in the present study might not be sufficient to retain cognitive performance. In conclusion, a higher dose of HC is not harmful for cognitive functioning when administered for a substantial period of time. However, there is evidence that chronic exposure to high cortisol levels leads to impairments in cognitive functioning in the long run.12,13 Therefore, we need to be cautious about generalizing our results to longer time periods. Furthermore, it is known that GCs enhance memory consolidation of emotionally arousing experiences, whereas it impairs memory retrieval and working memory in stressful situations.19 Therefore, the introduction of an (emotional) stressor prior to cognitive assessment may result in different effects. 170 Chapter 8

Hydrocortisone and health-related quality of life Patients with SAI report a compromised quality of life (QoL). This relationship appears to be dose-dependent, with higher doses of HC being associated with a worse QoL.20–22 However, in the present thesis we showed that patients reported increased QoL after receiving the higher dose of HC (chapter 3). Particularly improvements in the domains fatigue and energy, pain, somatic complaints, and symptoms of depression were reported. Overall, physical aspects of QoL seemed more affected than mental aspects and most of the reported complaints are related to energy and vitality. Reduced QoL has a major impact on one’s functioning in daily life. In a cross- sectional study in the German AI population, more than 25 % of patients with SAI reported to be out of work and 70 % experienced that they were restricted in their leisure activities.22 Improving the QoL of patients with AI is a major short- and long- term goal of substitution therapy, but at the same time it was reported to be one of the main challenges associated with conventional GC substitution therapy.23 Surprisingly, QoL is still not routinely assessed by physicians.23 Awareness about the importance of QoL is rising, prompting clinicians in the UK to develop an Addison’s disease-specific questionnaire aimed at quantifying altered well-being and treatment effects.24 Initially a total of 36 disease-specific items were included in the questionnaire, based on literature research and in-depth interviews with 14 patients with Addison’s disease and seven of their partners. Validation of the questionnaire resulted in a revised 30-item AddiQoL questionnaire.25 This questionnaire could be an easy tool for the regular assessment of QoL during GC substitution therapy. In a study examining the factors affecting QoL in AI, longer time between onset of symptoms and diagnosis turned out to be the most important factor impairing QoL.26 In a German study in 216 AI patients it was found that 47 % of patients with AI were diagnosed within the first year of onset of symptoms, whereas for 20 % of patients it took more than 5 years before the correct diagnosis of AI was established.27 Furthermore, 85 % of patients consulted more than one physician before the diagnosis was made, emphasizing the need for increased awareness of this disease among physicians. Di- agnosis is often delayed due to the non-specific signs of AI. This association between delay in diagnosis and impairment in QoL was still present years after manifestation of the disease, indicating the relevance of a timely and adequate diagnosis. Therefore, more attention for this rare disease and an earlier diagnosis should be established as this may contribute to improvements in QoL. One of our most striking results from chapter 3 was the reported increase in pain symptoms after treatment with the lower dose of HC. In healthy individuals, stress leads to secretion of cortisol, which in turn reduces pain levels due to its pain-dampening effects.28 Correspondingly, low levels of cortisol are associated with high pain General Discussion and Conclusions 171 levels.29 However, it has recently been questioned whether low cortisol levels are the consequence rather than the cause of these high pain levels.30 To study this cause-effect relationship, we have investigated the possible mediating role of cortisol levels in the association between perceived stress and pain (chapter 4). We did not find evidence for a mediating role of cortisol in the association between perceived stress and pain, which was in accordance with a recently conducted large longitudinal study.30 It should be noted that overall reported stress levels in this study were relatively low. It is con- ceivable that the role of HC in the association between stress and pain is particularly evident in stressful situations and that the relationship between stress and pain is different when a mild stressor is applied. Therefore, the introduction of a mild stressor in addition to a change in HC substitution dose would be suggested for future research. In conclusion, QoL should be regularly assessed in GC substitution treatment since reduced QoL has a major impact on daily life functioning. Endocrinologist should be aware that patients might benefit from a higher dose of HC with regard to several aspects of QoL. We suggest to start with the lowest dose to relieve the symptoms of adrenal insufficiency (AI), but consider to increase the dose when QoL is impaired.

Part II. The effect of hydrocortisone treatment on somatic outcome measures in patients with secondary adrenal insufficiency 8 Hydrocortisone and somatosensory functioning As described earlier (chapter 3), after receiving the lower HC dose, patients reported more pain symptoms. Elaborating on this, we investigated whether the two HC doses also altered somatosensory functioning. To study this, the Quantitative Sensory Test- ing (QST) battery of the German Research Network on Neuropathic Pain was used to establish detection and pain threshold in response to mechanical stimuli.31,32 We applied several sensory stimuli, ranging from modified von Frey hairs to pinpricks, on the dorsal hand and patients had to indicate whether they felt the stimulus or not, or how much pain they experienced when the stimulus was applied. We did not find differences in sensitivity in response to mechanical stimuli between the two treatment doses (chapter 5). However, even though the QST has proven to be valid, reliable and sensitive in quantifying sensory abnormalities,33–35 it was mainly developed to study inter-individual differences in pain diagnoses and it might be questioned whether it was sensitive enough to assess potentially much smaller intra-individual differences. The QST can be used as a tool for discriminating between groups (pain versus no pain), however, its ability to predict pain intensity in patients based on pain thresholds seems limited.36 172 Chapter 8

Mechanisms regulating cortisol exposure It is well established that high levels of GCs increase risk factors for cardiovascular disease and can even increase cardiovascular mortality. In the present thesis it was found that 10 weeks of treatment with a higher dose of HC resulted in an increase in blood pressure, accompanied by suppression of the renin-angiotensin-aldosterone system (chapter 6), indicating MR activation. The mechanisms underlying these relationships are complex; exposure to cortisol is not just the equivalent of the dose of HC administered. This was illustrated in chapter 7, where we showed that administration of the same weight-adjusted dose of HC led to large inter-individual differences in cortisol exposure. Furthermore, tissue specific cortisol levels are not only determined by circulating plasma or serum cortisol levels. This system is much more complex and additional factors play a role. For instance plasma proteins that bind cortisol and steroid metabolism within target tissues both regulate tissue specific cortisol exposure.

CBG and albumin Most of the cortisol is bound to the plasma proteins corticosteroid-binding-globulin (CBG) and albumin. Approximately 80 % of cortisol is bound to CBG (with high affinity but low binding capacity) and 10 % to albumin (with low affinity but high binding capacity), leaving approximately 10 % of cortisol unbound, i.e. free cortisol. As cortisol levels increase, the relative amount of albumin-bound cortisol and free cortisol increases, while the relative CBG-bound cortisol fraction decreases.37 This is probably due to saturation of CBG at higher concentrations of cortisol. CBG is the principal transport protein of GCs. Cortisol is hydrophobic and it therefore requires a transport system in plasma to circulate and reach target tissues.38 In addition, CBG also acts as a reservoir of circulating cortisol. Upon interaction with target proteinases, the binding activity of CBG is lost and cortisol is released.39 This mechanism regulates local delivery of cortisol to tissues, for instance during inflammation.40 CBG concentrations and its binding capacity are affected by several factors such as sex (CBG levels are higher in women compared to men41), concomitant medications (for instance estrogen are known to increase CBG levels and consequently increase total cortisol levels42–44) and physiological changes in body temperature (in response to fever or external heat, CBG releases cortisol45). Changes in CBG have an impact on total serum cortisol levels, but concentrations of serum free cortisol, i.e. the biologically active part of cortisol, are believed to be unaffected. Nevertheless, total cortisol is usually measured in clinical practice as a reflection of free cortisol. However, free cortisol levels are not always dose proportional to total cortisol, as was shown in chapter 7. A doubling of the HC dose did result in a doubling of plasma free cortisol levels, whereas total cortisol levels increased less. This non-proportional increase in total cortisol could be attributed to saturation of CBG, leading to increased General Discussion and Conclusions 173 free cortisol levels and in order to compensate for this, resulting in increased total cortisol clearance and thus a relatively less increase in total cortisol levels.46 Serum free cortisol could potentially be a useful biomarker in the monitoring of GC replacement therapy, however, measurement of free cortisol is time-consuming and cumbersome as no commercial assays are available yet. Therefore, it is not used in daily clinical practice. It should be mentioned that even though free cortisol levels more closely reflect the biologically active part of cortisol, it still does not represent intracellular cortisol levels.47 Alternatively cortisol in saliva is suggested as a potential marker for monitoring HC substitution. Salivary cortisol is believed to reflect the free fraction of cortisol in plasma because of the absence of substantial amounts of CBG and albumin in saliva.48 Furthermore, the measurement of cortisol in saliva is non-invasive and convenient for the patient because it can be collected at home. However, it has some limitation as well. Salivary cortisol is undetectable when serum cortisol levels are low,49 samples are easily contaminated by oral HC intake resulting in very high salivary cortisol levels,50 and correlations with serum cortisol are variable.50–54 Therefore, its utility in the monitoring of HC substitution therapy remains uncertain. In contrast, salivary cortisone is not affected by contamination with oral HC. In addition, due to the activity of 11β-hydroxysteroid dehydrogenase type 2 (11β-HSD2) in the salivary gland, cortisone levels exceed cortisol levels in saliva, making them easier to measure even at low concentrations of serum cortisol. Furthermore, salivary cortisone is presumed to be generated during the production of saliva from free serum cortisol.49 With saliva 8 being considered a post-enzyme tissue, salivary cortisone levels could thus reflect tissue specific cortisol exposure. However, results thus far are inconclusive about whether salivary cortisone can be used as a biomarker for monitoring HC substitution therapy.51,55,56

11β-hydroxysteroid dehydrogenase Another mechanism involved in the targeted tissue exposure is the pre-receptor metabolism of cortisol by 11β-hydroxysteroid dehydrogenase type 1 (11β-HSD1) and type 2 (11β-HSD2). 11β-HSD2 is highly expressed in mineralocorticoid target tissues such as the salivary gland, kidney and colon.57 In vitro, cortisol has the same binding affinity to the MR as the mineralocorticoid aldosterone, but 11β-HSD2 protects the MR from unwanted activation by cortisol by converting cortisol to cortisone.58 In the absence of 11β-HSD2 or when its activity is reduced, the MR is activated by cortisol, leading to increased blood pressure, sodium retention, potassium loss with low renin and aldosterone concentrations. In chapter 6 we have shown that even GC substitution doses considered to be within the physiological range have an effect on blood pres- 174 Chapter 8

sure and decrease renin and aldosterone concentrations, suggesting MR activation by cortisol. The other enzyme, 11β-HSD1, is more widely expressed throughout the body, with highest expression detected in key metabolic tissues including adipose tissue, liver, and skeletal muscle.59 11β-HSD1 is bidirectional in vitro, it can activate GCs by converting cortisone to cortisol (reductase) and it can inactivate GCs by converting cortisol to cortisone. However, the reductase activity (activating GCs) predominates in vivo. In healthy men, the amount of cortisol regenerated through 11β-HSD1 is greater than endogenous secretion by the adrenal gland.60 Three pathways have been identified by which cortisol activates GRs and 11β-HSD1 plays a role in these mechanisms. Firstly, active cortisol is able to diffuse through the cell and bind to the GR. Secondly, excess active cortisol in the circulation is converted to inactive cortisone by 11β-HSD2, and once it has reached the target tissue, is transformed back into cortisol by 11β-HSD1 to enable activation of the GR. Thirdly, activation of the GR by GCs stimulates the expression and activity of 11β-HSD1 within the cell, resulting in a feed forward loop in which more cortisone is being reactivated to cortisol.58 It is suggested that 11β-HSD1 plays a role in the adverse metabolic features of GC excess. Two case reports describe patients with GC excess due to Cushing’s disease, but who lack the typical Cushingoid phenotype.61,62 In both these patients, a defect in 11β-HSD1 activity was identified. This led to the assumption that tissue regeneration of cortisol by 11β-HSD1, rather than circulating delivery, is the major determinant of the adverse metabolic manifestations of circulatory GC excess. This was replicated in a mouse model, in which 11β-HSD1 knock-out mice receiving exogenous GCs remained protected from the adverse metabolic side effects of GC excess such as glucose intolerance, hyperinsulinemia, hypertension, hepatic steatosis, myopathy and dermal atrophy.58 11β-HSDs are also expressed in the brain and are suggested to be involved in the cognitive consequences of GC overexposure. 11β-HSD2 can be found in the developing brain, where it protects the brain from the negative effects of GC excess, but its expression in the adult brain is very limited and restricted to the hippocampus, lateral septum, and some brain stem nuclei.17,63 In contrast, 11β-HSD1 is more widely expressed in the brain, particularly in the hippocampus, prefrontal cortex and cerebellum.64 Local regeneration of cortisol by 11β-HSD1 may cause neurotoxicity and this can result in age-related cognitive decline.60 In the mouse brain, 11β-HSD1 levels rise with age and correlate with impaired cognitive performance. Accordingly, 11β-HSD1 knockout mice are protected from the age-related impairment in cognitive performance.65 This was associated with reduced intrahippocampal glucocorticoid levels whereas circulating glucocorticoid levels remained stable or were modestly elevated.65 This shows that General Discussion and Conclusions 175

11β-HSD1 in the brain is more important for GC action in the brain than circulating cortisol levels. Inhibition of 11β-HSD1 has been suggested as a potential target for therapy in disorders in which local cortisol regeneration plays a key role, such as type 2 diabetes, obesity and age-related cognitive decline. A number of selective and non-selective 11β-HSD1 inhibitors have been developed and studied, but results are inconclusive. A few studies described the effects of a selective 11β-HSD1 inhibitor in patients with type 2 diabetes and those not responding to metformin treatment and reported mod- est improvements in blood glucose control, insulin sensitivity and lipid profile.66–68 However, due to the small magnitude of their glucose lowering effect, these inhibitors have not been further developed as antidiabetic agents, and are therefore not commercially available yet. With regard to cognition, the non-selective 11β-HSD inhibitor carbenoxolone improved verbal fluency and memory in healthy elderly and men with type 2 diabetes.64 However, a phase II clinical trial studying the selective 11β-HSD1 inhibitor ABT-384 in patients with mild to moderate Alzheimer’s disease was terminated prematurely due to lack of efficacy.69 One might speculate that 11β-HSD1 inhibition could also be used to counteract the negative side effects of GC excess in GC substitution therapy as treatment for SAI. However, none of these selective inhibitors have been tested in this application. Fur- thermore, the effects of long-term 11β-HSD1 inhibition on, for example, the immune system are unclear.58 Clinical studies are needed to further elucidate these long-term effects and to investigate the potential role in neutralizing the negative side effects of 8 potential GC excessive exposure in the substitution therapy of SAI.

Other mechanisms regulating cortisol exposure Binding of cortisol and the interconversion between cortisol and cortisone are not the only mechanisms influencing cortisol exposure. Cortisol is metabolized by cytochrome P450 3A4 (CYP3A4). CYP3A4 is the major enzyme metabolizing various substances made by the body (e.g. cholesterol and cortisol) and substances foreign to the body (e.g. several medications). The inhibition or induction of CYP3A4 by drugs often causes unfavorable and long-lasting drug-drug interactions. CYP3A4-inhibiting drugs (e.g. ketoconazole, clarithromycin, cimetidine) decrease the activity of the enzyme, resulting in decreased clearance and increased half-life (hence increased exposure) of cortisol. On the other hand, CYP3A4 inducing drugs (e.g. carbamazepine, barbiturates, rifampin) can increase enzyme activity and consequently increase clearance and decrease half-life of cortisol.70 When such interfering medications are used concomitantly with GCs, GC doses should be adjusted accordingly. Cortisol exerts its effect through binding with GRs within the cell. After binding with a GR, the GR-complex enters the nucleus of the cell where it binds to a GC response 176 Chapter 8

element to up-regulate and down-regulate certain genes. Several single nucleotide polymorphisms (SNPs) have been identified in the human GR gene which can influence receptor sensitivity for GCs.71 The main polymorphisms studied in the GR-gene are Tth111I (rs10052957), ER22/23EK (rs6189 + rs6190), N363S (originally rs6195, currently listed as rs56149945), BclI (rs41423247) and GR-9β (rs6198). The Tth111I was originally reported to be associated with elevated basal cortisol levels,72 but other studies suggested that this SNP coincides in haplotypes with either the 9β SNP, the 9β SNP and the ER22/23Ek SNP, or the BclI SNP.73 The BclI and the N363S SNP are associated with a phenotypes indicating increased GC sensitivity, whereas the ER22/23EK and GR-9β are associated with decreased GC sensitivity.71 Furthermore, the 11β-HSD1 SNP (rs11119328) was associated with increased salivary cortisol levels and increased susceptibility for depression.74 Most associations between SNPs and clinical outcome have been reported in body composition, metabolism, the cardiovascular system, the immune system and psychiatric illnesses.75 We know from patient experience and reported in this thesis that there is a large variation in the clinical response to GC treatment. Some patients reported to experience large differences in well-being between the two treatment doses, whereas others reported no change whatsoever. Genetic variation in the GR-gene may underlie this variation71 and it would be interesting to study whether genetic variation mediates the effect of the GC substitution dose.

Should we treat them all with dual-release hydrocortisone? One of the shortcomings of the currently available substitution treatments for SAI is that it does not mimic the normal diurnal cortisol rhythm seen in healthy people very well, which is a consequence of the pharmacokinetic properties of the oral immediate release tablets. New developments in timed-release HC preparations enable the mimicking of the daily rhythm in cortisol production. Plenadren® (dual-release HC) is one of those new formulations. It consists of a modified-release HC tablet with combined immediate and extended release characteristics, allowing only once daily administration.76 Initial studies showed that a more physiological cortisol profile could be established, leading to decreased overall cortisol exposure accompanied by improvements in blood pressure, body weight, body mass index, hemoglobin A1c, and QoL.77–79 Even though modified-release HC is able to create a more physiological cortisol profile during the day, patients still lack the nocturnal rise in cortisol and the peak upon awaking. Early morning fatigue as a consequence of low cortisol levels is a major problem with current replacement regimen.80–82 In order to overcome this limitation, another modified-release preparation of cortisol (also known as Chronocort®) was designed to provide nocturnal cortisol levels comparable to healthy individuals.83 Indeed, administration of 30 mg of Chronocort® at 2200 h with a delayed release of cortisol of 4 h, resulted in close to normal concentration-time curves, with time General Discussion and Conclusions 177 to reach maximum concentrations and area under the curve being similar to levels found in healthy volunteers with intact endogenous cortisol production.83–86 However, despite the more physiological profile during the night and the early morning, very low cortisol levels were observed at around midday after administration of Chronocort®, indicating the remaining need for a morning administration of GCs. Alternatively, continuous subcutaneous hydrocortisone infusion (CSHI) has been proposed as a treatment option, and it proved to allow accurate regulation of GC delivery restoring the biological daily rhythm.87 It was able to normalize morning cortisol levels and to restore GC metabolism close to normal.88 CSHI proved safe and might be a treatment option for selected patients who function poorly on regular oral HC treatment. Particularly patients with issues in first-pass hepatic metabolism or gastrointestinal diseases might benefit from CSHI. It has the potential to reduce risk of acute adrenal crisis in case of vomiting or diarrhea. If pump technology improves and costs will be reduced, CSHI has the potential to become an affordable treatment choice for patients who fail conventional treatment. In conclusion, results from initial studies with modified-release HC preparations and CSHI are promising. They provide a more physiological cortisol profile and metabolic parameters improve. Furthermore, adherence might be better compared to thrice daily dosing which is necessary with immediate release tablets.89 However, studies assessing long-term safety and effectiveness are still needed. In case of persistent compromised well-being or in patients that are poorly controlled by conventional therapy, dual- release HC or CSHI should be considered a treatment option in order to improve 8 disease control and outcomes.

Conclusion and future perspectives

In this thesis we looked into the effects of two different doses of HC on several psychological and somatic outcome measures. Treatment with a higher dose of HC was not harmful for cognitive functioning and it increased several components of QoL. However, it also increased blood pressure and affected several blood pressure regulating mechanisms. No differences were found for somatosensory functioning between the two HC doses and there was no evidence for a mediating role of cortisol in the relationship between perceived stress and pain. Large differences in cortisol exposure were found between individuals, which can be attributed to the many factors in the complex system that coordinate and mediate cortisol exposure. The data of this thesis support the use of a personalized approach in the treatment of SAI, taking into account the short- and long-term benefits and harms. A “one dose fits all” approach does not suffice. We suggest to start with the lowest dose that relieves symptoms of AI, but urge 178 Chapter 8

that QoL should be taken into account in determining the appropriate substitution dose because patients might benefit from a higher HC dose. In case of persistent impaired QoL or other factors resulting in failed treatment with immediate release tablets, dual- release HC preparations or CSHI might be a treatment option. During previous years, more insight has been gained into factors that have an impact on or are influenced by HC substitution therapy. Nevertheless, additional research is needed to evaluate long-term effects of HC treatment, and whether outcomes will be different in the presence of a stressor. Furthermore, other markers should be explored that could assist in the monitoring of currently used and new treatments. Therefore, future research should include the following: – The effects of long-term HC use on cognition. – The effects of the introduction of a mild stressor in the relationship between HC substitution and several clinical outcome measures. – The potency of selective 11β-HSD1 inhibitors in the prevention of negative side effects of GCs excess. – The number of daily doses (in contrast to a change in dose), in certain subgroups of patients (e.g. with high cortisol clearance). – The influence of cortisol exposure, or GC-sensitive biomarkers on clinical outcome measures. – Studies to identify better methods to estimate cortisol tissue exposure, for instance by salivary cortisol measurements or by dedicated sensor technology. General Discussion and Conclusions 179

References

1. Gallagher P, Reid KS, Ferrier IN. Neuropsychological functioning in health and mood disorder: Modulation by glucocorticoids and their receptors. Psychoneuroendocrinology 2009;34 (suppl 1):S196-207. 2. Lupien SJ, Gaudreau S, Tchiteya BM, et al. Stress-induced declarative memory impairment in healthy elderly subjects: relationship to cortisol reactivity. J Clin Endocrinol Metab 1997;82(7):2070-2075. 3. Bodnoff SR, Humphreys AG, Lehman JC, Diamond DM, Rose GM, Meaney MJ. Enduring effects of chronic corticosterone treatment on spatial learning, synaptic plasticity, and hippocampal neuro- pathology in young and mid-aged rats. J Neurosci 1995;15(1 Pt 1):61-69. 4. Bourdeau I, Bard C, Forget H, Boulanger Y, Cohen H, Lacroix A. Cognitive function and cerebral assessment in patients who have Cushing’s syndrome. Endocrinol Metab Clin North Am 2005;34(2):357-69, ix. 5. Fietta P, Fietta P, Delsante G. Central nervous system effects of natural and synthetic glucocorticoids. Psychiatry Clin Neurosci 2009;63(5):613-622. 6. Mauri M, Sinforiani E, Bono G, et al. Memory impairment in Cushing’s disease. Acta Neurol Scand 1993;87(1):52-55. 7. Forget H, Lacroix A, Somma M, Cohen H. Cognitive decline in patients with Cushing’s syndrome. J Int Neuropsychol Soc 2000;6(1):20-29. 8. Starkman MN, Giordani B, Berent S, Schork MA, Schteingart DE. Elevated cortisol levels in Cush- ing’s disease are associated with cognitive decrements. Psychosom Med 2001;63(6):985-993. 9. Michaud K, Forget H, Cohen H. Chronic glucocorticoid hypersecretion in Cushing’s syndrome exacerbates cognitive aging. Brain Cogn 2009;71(1):1-8. 10. Starkman MN, Gebarski SS, Berent S, Schteingart DE. Hippocampal formation volume, memory dysfunction, and cortisol levels in patients with Cushing’s syndrome. Biol Psychiatry 1992;32(9):756- 765. 8 11. Momose KJ, Kjellberg RN, Kliman B. High incidence of cortical atrophy of the cerebral and cer- ebellar hemispheres in Cushing’s disease. Radiology 1971;99(2):341-348. 12. Tiemensma J, Kokshoorn NE, Biermasz NR, et al. Subtle cognitive impairments in patients with long-term cure of Cushing’s disease. J Clin Endocrinol Metab 2010;95(6):2699-2714. 13. Andela CD, van Haalen FM, Ragnarsson O, et al. MECHANISMS IN ENDOCRINOLOGY: Cush- ing’s syndrome causes irreversible effects on the human brain: a systematic review of structural and functional magnetic resonance imaging studies. Eur J Endocrinol 2015;173(1):R1-14. 14. Nass R, Heier L, Moshang T, et al. Magnetic resonance imaging in the congenital adrenal hyperplasia population: increased frequency of white-matter abnormalities and temporal lobe atrophy. J Child Neurol 1997;12(3):181-186. 15. Grossmann C, Scholz T, Rochel M, et al. Transactivation via the human glucocorticoid and mineralocorticoid receptor by therapeutically used steroids in CV-1 cells: a comparison of their glucocorticoid and mineralocorticoid properties. Eur J Endocrinol 2004;151(3):397-406. 16. De Kloet ER, Reul JM. Feedback action and tonic influence of corticosteroids on brain function: a concept arising from the heterogeneity of brain receptor systems. Psychoneuroendocrinology 1987;12(2):83-105. 17. Reul JM, de Kloet ER. Two receptor systems for corticosterone in rat brain: microdistribution and differential occupation. Endocrinology 1985;117(6):2505-2511. 18. de Kloet ER, Oitzl MS, Joels M. Stress and cognition: are corticosteroids good or bad guys? Trends Neurosci 1999;22(10):422-426. 180 Chapter 8

19. de Quervain DJ, Aerni A, Schelling G, Roozendaal B. Glucocorticoids and the regulation of memory in health and disease. Front Neuroendocrinol 2009;30(3):358-370. 20. Ragnarsson O, Mattsson AF, Monson JP, et al. The relationship between glucocorticoid replacement and quality of life in 2737 hypopituitary patients. Eur J Endocrinol 2014;171(5):571-579. 21. Bleicken B, Hahner S, Loeffler M, Ventz M, Allolio B, Quinkler M. Impaired subjective health status in chronic adrenal insufficiency: impact of different glucocorticoid replacement regimens. Eur J Endocrinol 2008;159(6):811-817. 22. Hahner S, Loeffler M, Fassnacht M, et al. Impaired subjective health status in 256 patients with adrenal insufficiency on standard therapy based on cross-sectional analysis.J Clin Endocrinol Metab 2007;92(10):3912-3922. 23. Grossman A, Johannsson G, Quinkler M, Zelissen P. Therapy of endocrine disease: Perspectives on the management of adrenal insufficiency: clinical insights from across Europe. Eur J Endocrinol 2013;169(6):R165-75. 24. Lovas K, Curran S, Oksnes M, Husebye ES, Huppert FA, Chatterjee VK. Development of a disease- specific quality of life questionnaire in Addison’s disease.J Clin Endocrinol Metab 2010;95(2):545-551. 25. Oksnes M, Bensing S, Hulting AL, et al. Quality of life in European patients with Addison’s disease: validity of the disease-specific questionnaire AddiQoL.J Clin Endocrinol Metab 2012;97(2):568-576. 26. Meyer G, Hackemann A, Penna-Martinez M, Badenhoop K. What affects the quality of life in autoimmune Addison’s disease? Horm Metab Res 2013;45(2):92-95. 27. Bleicken B, Hahner S, Ventz M, Quinkler M. Delayed diagnosis of adrenal insufficiency is common: a cross-sectional study in 216 patients. Am J Med Sci 2010;339(6):525-531. 28. Hannibal KE, Bishop MD. Chronic stress, cortisol dysfunction, and pain: a psychoneuroendocrine rationale for stress management in pain rehabilitation. Phys Ther 2014;94(12):1816-1825. 29. Tak LM, Cleare AJ, Ormel J, et al. Meta-analysis and meta-regression of hypothalamic-pituitary- adrenal axis activity in functional somatic disorders. Biol Psychol 2011;87(2):183-194. 30. Generaal E, Vogelzangs N, Macfarlane GJ, et al. Biological stress systems, adverse life events and the onset of chronic multisite musculoskeletal pain: a 6-year cohort study. Ann Rheum Dis 2015. 31. Rolke R, Baron R, Maier C, et al. Quantitative sensory testing in the German Research Network on Neuropathic Pain (DFNS): standardized protocol and reference values. Pain 2006;123(3):231-243. 32. Rolke R, Magerl W, Campbell KA, et al. Quantitative sensory testing: a comprehensive protocol for clinical trials. Eur J Pain 2006;10(1):77-88. 33. Felix ER, Widerstrom-Noga EG. Reliability and validity of quantitative sensory testing in persons with spinal cord injury and neuropathic pain. J Rehabil Res Dev 2009;46(1):69-83. 34. Freynhagen R, Rolke R, Baron R, et al. Pseudoradicular and radicular low-back pain--a disease con- tinuum rather than different entities? Answers from quantitative sensory testing. Pain 2008;135(1- 2):65-74. 35. Geber C, Klein T, Azad S, et al. Test-retest and interobserver reliability of quantitative sensory testing according to the protocol of the German Research Network on Neuropathic Pain (DFNS): a multi-centre study. Pain 2011;152(3):548-556. 36. Hubscher M, Moloney N, Leaver A, Rebbeck T, McAuley JH, Refshauge KM. Relationship between quantitative sensory testing and pain or disability in people with spinal pain-a systematic review and meta-analysis. Pain 2013;154(9):1497-1504. 37. Dorin RI, Pai HK, Ho JT, et al. Validation of a simple method of estimating plasma free cortisol: role of cortisol binding to albumin. Clin Biochem 2009;42(1-2):64-71. 38. Perogamvros I, Ray DW, Trainer PJ. Regulation of cortisol bioavailability--effects on hormone measurement and action. Nat Rev Endocrinol 2012;8(12):717-727. General Discussion and Conclusions 181

39. Lin HY, Muller YA, Hammond GL. Molecular and structural basis of steroid hormone binding and release from corticosteroid-binding globulin. Mol Cell Endocrinol 2010;316(1):3-12. 40. Hammond GL, Smith CL, Paterson NA, Sibbald WJ. A role for corticosteroid-binding globulin in delivery of cortisol to activated neutrophils. J Clin Endocrinol Metab 1990;71(1):34-39. 41. Fernandez-Real JM, Pugeat M, Grasa M, et al. Serum corticosteroid-binding globulin concentration and insulin resistance syndrome: a population study. J Clin Endocrinol Metab 2002;87(10):4686- 4690. 42. Angeli A, Frajria R, Richiardi L, Agrimonti F, Gaidano G. Simultaneous measurement of circulating cortisol, corticosteroid binding globulin (CBG) binding capacity and “apparent free cortisol concentration” in human peripheral plasma using gel-exchange with Sephadex G-25. Clin Chim Acta 1977;77(1):1-12. 43. Doe RP, Fernandez R, Seal US. Measurement of Corticosteroid-Binding Globulin in Man. J Clin Endocrinol Metab 1964;24:1029-1039. 44. Wiegratz I, Kutschera E, Lee JH, et al. Effect of four different oral contraceptives on various sex hormones and serum-binding globulins. Contraception 2003;67(1):25-32. 45. Henley DE, Lightman SL. New insights into corticosteroid-binding globulin and glucocorticoid delivery. Neuroscience 2011;180:1-8. 46. Tunn S, Mollmann H, Barth J, Derendorf H, Krieg M. Simultaneous measurement of cortisol in serum and saliva after different forms of cortisol administration. Clin Chem 1992;38(8 Pt 1):1491- 1494. 47. Moisan MP. CBG: a cortisol reservoir rather than a transporter. Nat Rev Endocrinol 2013;9(2):78-c1. Epub 2012 Dec 18. 48. Hammond GL, Langley MS. Identification and measurement of sex hormone binding globulin (SHBG) and corticosteroid binding globulin (CBG) in human saliva. Acta Endocrinol (Copenh) 1986;112(4):603-608. 8 49. Perogamvros I, Owen LJ, Newell-Price J, Ray DW, Trainer PJ, Keevil BG. Simultaneous measurement of cortisol and cortisone in human saliva using liquid chromatography-tandem mass spectrometry: application in basal and stimulated conditions. J Chromatogr B Analyt Technol Biomed Life Sci 2009;877(29):3771-3775. 50. Perogamvros I, Aarons L, Miller AG, Trainer PJ, Ray DW. Corticosteroid-binding globulin regulates cortisol pharmacokinetics. Clin Endocrinol (Oxf) 2011;74(1):30-36. 51. Ross IL, Lacerda M, Pillay TS, et al. Salivary Cortisol and Cortisone do not Appear to be Useful Biomarkers for Monitoring Hydrocortisone Replacement in Addison’s Disease. Horm Metab Res 2016;48(12):814-821. 52. Lovas K, Thorsen TE, Husebye ES. Saliva cortisol measurement: simple and reliable assessment of the glucocorticoid replacement therapy in Addison’s disease. J Endocrinol Invest 2006;29(8):727-731. 53. Thomson AH, Devers MC, Wallace AM, et al. Variability in hydrocortisone plasma and saliva pharmacokinetics following intravenous and oral administration to patients with adrenal insufficiency. Clin Endocrinol (Oxf) 2007;66(6):789-796. 54. Jung C, Greco S, Nguyen HH, et al. Plasma, salivary and urinary cortisol levels following physiological and stress doses of hydrocortisone in normal volunteers. BMC Endocr Disord 2014;14:91- 6823-14-91. 55. Raff H. Measurement of Salivary Cortisone to Assess the Adequacy of Hydrocortisone Replacement. J Clin Endocrinol Metab 2016;101(4):1350-1352. 56. Debono M, Harrison RF, Whitaker MJ, et al. Salivary Cortisone Reflects Cortisol Exposure Under Physiological Conditions and After Hydrocortisone. J Clin Endocrinol Metab 2016;101(4):1469-1477. 182 Chapter 8

57. Whorwood CB, Ricketts ML, Stewart PM. Epithelial cell localization of type 2 11 beta-hydroxysteroid dehydrogenase in rat and human colon. Endocrinology 1994;135(6):2533-2541. 58. Morgan SA, Hassan-Smith ZK, Lavery GG. MECHANISMS IN ENDOCRINOLOGY: Tissue- specific activation of cortisol in Cushing’s syndrome. Eur J Endocrinol 2016;175(2):R83-9. 59. Ricketts ML, Verhaeg JM, Bujalska I, Howie AJ, Rainey WE, Stewart PM. Immunohistochemical localization of type 1 11beta-hydroxysteroid dehydrogenase in human tissues. J Clin Endocrinol Metab 1998;83(4):1325-1335. 60. Kilgour AH, Semple S, Marshall I, Andrews P, Andrew R, Walker BR. 11beta-Hydroxysteroid dehydrogenase activity in the brain does not contribute to systemic interconversion of cortisol and cortisone in healthy men. J Clin Endocrinol Metab 2015;100(2):483-489. 61. Tomlinson JW, Draper N, Mackie J, et al. Absence of Cushingoid phenotype in a patient with Cushing’s disease due to defective cortisone to cortisol conversion. J Clin Endocrinol Metab 2002;87(1):57-62. 62. Arai H, Kobayashi N, Nakatsuru Y, et al. A case of cortisol producing adrenal adenoma without phenotype of Cushing’s syndrome due to impaired 11beta-hydroxysteroid dehydrogenase 1 activity. Endocr J 2008;55(4):709-715. 63. Brown RW, Diaz R, Robson AC, et al. The ontogeny of 11 beta-hydroxysteroid dehydrogenase type 2 and mineralocorticoid receptor gene expression reveal intricate control of glucocorticoid action in development. Endocrinology 1996;137(2):794-797. 64. Sandeep TC, Yau JL, MacLullich AM, et al. 11Beta-hydroxysteroid dehydrogenase inhibition improves cognitive function in healthy elderly men and type 2 diabetics. Proc Natl Acad Sci U S A 2004;101(17):6734-6739. 65. Yau JL, Noble J, Kenyon CJ, et al. Lack of tissue glucocorticoid reactivation in 11beta -hydroxysteroid dehydrogenase type 1 knockout mice ameliorates age-related learning impairments. Proc Natl Acad Sci U S A 2001;98(8):4716-4721. 66. Rosenstock J, Banarer S, Fonseca VA, et al. The 11-beta-hydroxysteroid dehydrogenase type 1 inhibitor INCB13739 improves hyperglycemia in patients with type 2 diabetes inadequately controlled by metformin monotherapy. Diabetes Care 2010;33(7):1516-1522. 67. Feig PU, Shah S, Hermanowski-Vosatka A, et al. Effects of an 11beta-hydroxysteroid dehydrogenase type 1 inhibitor, MK-0916, in patients with type 2 diabetes mellitus and metabolic syndrome. Dia- betes Obes Metab 2011;13(6):498-504. 68. Shah S, Hermanowski-Vosatka A, Gibson K, et al. Efficacy and safety of the selective 11beta-HSD-1 inhibitors MK-0736 and MK-0916 in overweight and obese patients with hypertension. J Am Soc Hypertens 2011;5(3):166-176. 69. Marek GJ, Katz DA, Meier A, et al. Efficacy and safety evaluation of HSD-1 inhibitor ABT-384 in Alzheimer’s disease. Alzheimers Dement 2014;10(5 Suppl):S364-73. 70. Zhou SF. Drugs behave as substrates, inhibitors and inducers of human cytochrome P450 3A4. Curr Drug Metab 2008;9(4):310-322. 71. Koper JW, van Rossum EF, van den Akker EL. Glucocorticoid receptor polymorphisms and haplotypes and their expression in health and disease. Steroids 2014;92:62-73. 72. Rosmond R, Chagnon YC, Chagnon M, Perusse L, Bouchard C, Bjorntorp P. A polymorphism of the 5’-flanking region of the glucocorticoid receptor gene locus is associated with basal cortisol secretion in men. Metabolism 2000;49(9):1197-1199. 73. van Rossum EF, Roks PH, de Jong FH, et al. Characterization of a promoter polymorphism in the glucocorticoid receptor gene and its relationship to three other polymorphisms. Clin Endocrinol (Oxf) 2004;61(5):573-581. General Discussion and Conclusions 183

74. Dekker MJ, Tiemeier H, Luijendijk HJ, et al. The effect of common genetic variation in 11beta- hydroxysteroid dehydrogenase type 1 on hypothalamic-pituitary-adrenal axis activity and incident depression. J Clin Endocrinol Metab 2012;97(2):E233-7. 75. Manenschijn L, van den Akker EL, Lamberts SW, van Rossum EF. Clinical features associated with glucocorticoid receptor polymorphisms. An overview. Ann N Y Acad Sci 2009;1179:179-198. 76. Johannsson G, Bergthorsdottir R, Nilsson AG, Lennernas H, Hedner T, Skrtic S. Improving glucocorticoid replacement therapy using a novel modified-release hydrocortisone tablet: a pharmacokinetic study. Eur J Endocrinol 2009;161(1):119-130. 77. Johannsson G, Nilsson AG, Bergthorsdottir R, et al. Improved cortisol exposure-time profile and outcome in patients with adrenal insufficiency: a prospective randomized trial of a novel hydrocortisone dual-release formulation. J Clin Endocrinol Metab 2012;97(2):473-481. 78. Quinkler M, Miodini Nilsen R, Zopf K, Ventz M, Oksnes M. Modified-release hydrocortisone decreases BMI and HbA1c in patients with primary and secondary adrenal insufficiency. Eur J Endocrinol 2015;172(5):619-626. 79. Giordano R, Guaraldi F, Marinazzo E, et al. Improvement of anthropometric and metabolic parameters, and quality of life following treatment with dual-release hydrocortisone in patients with Addison’s disease. Endocrine 2016;51(2):360-368. 80. Lovas K, Husebye ES, Holsten F, Bjorvatn B. Sleep disturbances in patients with Addison’s disease. Eur J Endocrinol 2003;148(4):449-456. 81. Lovas K, Loge JH, Husebye ES. Subjective health status in Norwegian patients with Addison’s disease. Clin Endocrinol (Oxf) 2002;56(5):581-588. 82. Al-Shoumer KA, Ali K, Anyaoku V, Niththyananthan R, Johnston DG. Overnight metabolic fuel deficiency in patients treated conventionally for hypopituitarism. Clin Endocrinol (Oxf) 1996;45(2):171-178. 83. Newell-Price J, Whiteman M, Rostami-Hodjegan A, et al. Modified-release hydrocortisone for circadian therapy: a proof-of-principle study in dexamethasone-suppressed normal volunteers. Clin 8 Endocrinol (Oxf) 2008;68(1):130-135. 84. Verma S, Vanryzin C, Sinaii N, et al. A pharmacokinetic and pharmacodynamic study of delayed- and extended-release hydrocortisone (Chronocort) vs. conventional hydrocortisone (Cortef) in the treatment of congenital adrenal hyperplasia. Clin Endocrinol (Oxf) 2010;72(4):441-447. 85. Whitaker MJ, Debono M, Huatan H, Merke DP, Arlt W, Ross RJ. An oral multiparticulate, modified- release, hydrocortisone replacement therapy that provides physiological cortisol exposure. Clin Endocrinol (Oxf) 2014;80(4):554-561. 86. Debono M, Ghobadi C, Rostami-Hodjegan A, et al. Modified-release hydrocortisone to provide circadian cortisol profiles. J Clin Endocrinol Metab 2009;94(5):1548-1554. 87. Lovas K, Husebye ES. Continuous subcutaneous hydrocortisone infusion in Addison’s disease. Eur J Endocrinol 2007;157(1):109-112. 88. Oksnes M, Bjornsdottir S, Isaksson M, et al. Continuous subcutaneous hydrocortisone infusion versus oral hydrocortisone replacement for treatment of addison’s disease: a randomized clinical trial. J Clin Endocrinol Metab 2014;99(5):1665-1674. 89. Nilsson AG, Marelli C, Fitts D, et al. Prospective evaluation of long-term safety of dual-release hydrocortisone replacement administered once daily in patients with adrenal insufficiency. Eur J Endocrinol 2014;171(3):369-377.

Chapter 9

Summary

Summary 187

Activation of the hypothalamic-pituitary-adrenal (HPA) axis causes the hypothalamus to secrete corticotrophin releasing hormone (CRH). CRH stimulates the anterior pituitary gland to release adrenocorticotropic hormone (ACTH), which in turn leads to secretion of cortisol by the adrenal gland. Cortisol, in turn, exerts inhibitory effects to the hypothalamus and the pituitary via a negative feedback loop. Cortisol serves several functions in the body, like the regulation of blood glucose, suppression of the immune system and assistance in fat, protein and carbohydrate metabolism. Patients with adrenal insufficiency (AI) are characterized by the loss of endogenous cortisol production. This lack of cortisol production can either be caused by loss of function of the adrenal gland itself (primary AI), or by impairment of the pituitary or hypothalamus (secondary AI or tertiary AI, respectively). This thesis focuses on patients with secondary AI. The diagnosis of secondary AI is based on the finding of low early morning cortisol levels when pituitary pathology is present. Furthermore, in case of indeterminate cortisol values, various stimulation tests are available to assess the integrity of the HPA axis. Patients with AI are treated with glucocorticoids, most commonly by the oral administration of hydrocortisone (HC). Traditionally it was recommended to administer two-thirds (20 mg) of the substitution dose in the morning and one-third (10 mg) in the evening. This dose was based on cortisol production rate estimates in healthy individuals of 12–15 mg/m²/day. However, using stable isotope tracers the daily cortisol production rate is currently estimated to be approximately 6–10 mg/m²/day, corresponding to total daily doses of 15–20 mg/day. Parameters to objectively monitor the adequacy of the glucocorticoid substitution therapy are lacking. Therefore, in current practice, clinical assessment of symptoms potentially suggestive of over- or under-treatment is often used. Under-treatment bears the risk of 9 insufficient cortisol supply in the case of severe stress risking an adrenal crisis, whereas chronic exposure to high cortisol levels is associated with increased mortality and morbidity such as increased risk for cardiovascular diseases and osteoporosis. Reduced quality of life and cognitive impairment are also found in relation to high cortisol levels.

Chapter 1 provides an introduction in the HPA axis, the etiology, diagnosis and treatment of secondary AI, and several psychological and somatic outcome measures that are influenced by the treatment of secondary AI. Furthermore, the aims of the thesis are described: to assess the pathophysiological effects of two different doses of hydrocortisone, and to add evidence for recommendations regarding glucocorticoid substitution therapy in secondary AI. In order to do so, we performed a randomized, double-blind crossover study in which we compared treatment with a total daily dose of 0.2–0.3 mg HC/kg body weight with a total daily dose of 0.4–0.6 mg HC/kg body weight (both administered for 10 weeks), and its effect on cognitive functioning, quality of 188 Chapter 9

life (QoL), the association between stress and pain, somatosensory functioning, blood pressure and regulating hormones, and pharmacokinetic parameters.

Part I. The effect of hydrocortisone treatment on psychological outcome measures in patients with secondary adrenal insufficiency

In Chapter 2 we describe the effect of two different doses of HC and its effect on the cognitive domains memory, attention, executive functioning and social cognition. A total of 63 patients with secondary AI were included in the study, of which 47 completed both study periods and were used in the present analysis. Patients either first received a lower dose of HC for 10 weeks (0.2–0.3 mg HC/kg body weight/day) followed by a higher dose of HC for 10 weeks (0.4–0.6 mg HC/kg body weight/day), or in reversed order, divided into three daily administrations. Cognitive performance was assessed at the end of each treatment period using a battery of 12 standardized cognitive tests. Cortisol levels measured one hour after ingestion of the morning dose differed significantly between the two HC doses (mean [SD], low dose: 653 [281] nmol/L; high dose: 930 [148] nmol/L; P < 0.001). No differences in cognitive performance were found between the two dose regimens; neither did the number of patients showing a cognitive impairment differ between the two treatment doses. There appears to be no negative influence of 10 weeks of treatment with a higher dose of HC compared with a lower dose of HC on memory, attention, executive functioning and social cognition.

In Chapter 3 we evaluated health-related QoL (HRQoL) after treatment with the same two HC doses as described previously. Cross-sectional studies indicate a relationship between HC dose and QoL, with higher doses being associated with more severely impaired subjective health. However, randomized clinical trials are lacking. HRQoL was assessed with a daily mood and symptom checklist and at the end of each treatment period using several questionnaires. The 47 patients that completed both study period were used in the present analysis. Patients reported fewer symptoms of depression (P = 0.016 and P = 0.045 for the Hospital Anxiety and Depression Scale (measuring anxiety and depression) and the Patient Health Quesitonnaire-9 (measuring depression), respectively), less general and mental fatigue (P = 0.004 and P = 0.003, respectively, both for the Multidimensional Fatigue Inventory-20 (MFI-20), assessing fatigue), increased motivation (P = 0.021, MFI-20), better physical functioning (P = 0.041), better general health (P = 0.013) and more vitality (P = 0.025) (all RAND- 36, a generic QoL questionnaire), compared to the lower dose. Furthermore, fewer somatic symptoms (P = 0.022) and less pain (P < 0.001) (both for the Patient Health Questioannire-15, assessing somatic complaints) were experienced while treated with the higher HC dose, compared to the lower dose. We concluded that patients receiving Summary 189 the higher dose of HC, reported a better HRQoL on various domains and that this fact should be taken into consideration in tailoring individualized treatment.

In Chapter 4 we aimed to study the causal role of low cortisol levels in the relationship between perceived stress and pain. Cortisol is known to have pain-dampening effects, and low cortisol levels are suggested to mediate the relationship between stress and pain. Patients filled out a daily diary assess perceived stress and pain levels throughout the trial. Non-seasonal autoregressive moving average modelling was used to test for associations between daily perceived stress levels and daily reported pain levels during each of the HC doses administered. Out of 47 patient, 12 patients showed high enough fluctuations in stress and pain levels to study their associations. Six patients showed associations between perceived stress and pain during both HC doses, one showed this association only under the low dose, and two only under the high dose. We therefore concluded that this study did not provide evidence for a prominent role of cortisol levels in the association between perceived stress and pain.

Part II. The effect of hydrocortisone treatment on somatic outcome measures in patients with secondary adrenal insufficiency

In Chapter 5 we tested somatosensory functioning during treatment with the lower and higher HC dose as described in chapter 2. We previously showed increased subjective pain reports after treatment with the lower HC dose compared to the higher HC dose. Whether different doses of HC would also alter somatosensory functioning, has not been studied yet. We therefore determined the mechanical detection threshold 9 (MDT), mechanical pain threshold (MPT), mechanical pain sensitivity (MPS), and the pain pressure threshold (PPT) according to the Quantitative Sensory Testing battery of the German Network on Neuropathic Pain. QST data of 46 patients was available. No differences in MDT, MPT, MPS and PPT were found between the two treatment doses, nor did the number of patients showing a sensory abnormality. We therefore concluded that previously found altered subjective pain due to a lower HC substitution dose is not caused by different somatosensory functioning in response to mechanical stimuli.

Chapter 6 describes the effect of hydrocortisone on blood pressure (BP) and regulating hormones. Cardiovascular risk is increased in patients with secondary AI and this may be caused by an unfavourable metabolic profile, of which hypertension is an important factor, due to a relatively high HC replacement dose. However, the relationship between HC dose and BP and its underlying mechanisms is poorly documented. We therefore determined the effects of a higher versus a lower HC dose on BP, the renin-angiotensin-aldosterone system (RAAS), 11β-hydroxysteroid dehydrogenase 190 Chapter 9

(11β-HSD) enzyme activity and circulating (nor)metanephrines. Forty-six patients were included in the present analysis. After treatment with the higher dose of HC, we found an increase in systolic BP of 5 (12) mm Hg (P = 0.011), an increase in diastolic BP of 2 (9) mm Hg (P = 0.050), and a median [interquartile range] drop in plasma potassium of -0.1 [-0.3; 0.1] nmol/L (P = 0.048), compared to treatment with the lower dose. The higher hydrocortisone dose led to decreases in serum aldosterone of -28 [-101; 9] pmol/L (P = 0.020) and plasma renin of -1.3 [-4.5; 1.2] pg/mL (P = 0.051), and increased the ratio of plasma and urinary cortisol to cortisone (including their metabolites) (P < 0.001 for all), compared to the lower dose. Furthermore, on the higher dose plasma and urinary normetanephrine decreased by -0.101 [-0.242; 0.029] nmol/L (P < 0.001) and -1.48 [-4.06; 0.29] µmol/mol creatinine (P < 0.001) respectively, compared to the lower dose. Thus, treatment with higher HC doses leads to an increase in BP together with changes in several pathways involved in BP regulation such as the RAAS, 11β-HSD enzyme activity, and sympathetic nerve activity.

In Chapter 7 we evaluated the pharmacokinetic properties of two different doses of HC. Various methods are used to assess the adequacy of HC substitution therapy in patients with secondary AI, but none of them are fully satisfactory. Pharmacokinetic studies have been performed, but studies assessing pharmacokinetics of both total and free cortisol in plasma and saliva are lacking. Therefore, we performed this post-hoc pharmacokinetic analysis to better understand individual and population pharmacokinetics and the effect of dose adjustments. Pharmacokinetic data of 46 patients was available. One- and two-compartment population models for plasma free cortisol, plasma total cortisol and salivary cortisol were parameterized. The individual pharmacokinetic

parameters clearance (CL), volume of distribution (Vd), the elimination half-life (t1/2),

maximum concentration (Cmax), and area under the curve (AUC) were calculated. We

found large inter-individual variation in CL and Vd of plasma free cortisol, plasma

total cortisol, and salivary cortisol, with AUC24h varying more than 10 fold. Cortisol exposure was increased on the higher dose compared to the lower dose, but this was only dose proportional for free cortisol concentrations and not for total cortisol. This large variation in cortisol exposure warrants individually tailored treatment doses. For patients designated as ‘fast metabolizers’, a doubling of the dose does not result in double exposure and therefore other management strategies are required.

Chapter 8 is a general discussion of the role of the glucocorticoid receptors and mineralocorticoid receptors that mediate the relation between cortisol and cognitive functioning and other outcome measures studied. Furthermore we discuss the role of QoL assessment in the monitoring of HC substitution therapy. We additionally discuss mechanisms regulating cortisol exposure beyond circulating cortisol levels, and the Summary 191 prospect of dual-release HC as a treatment option. Finally, several recommendations for future research are given.

Samenvatting

Samenvatting 195

Door activatie van de hypothalamus-hypofyse-bijnier-as (HPA-as) wordt de hypothalamus gestimuleerd om corticotropin-releasing-hormoon (CRH) af te geven. Dit hormoon stimuleert de hypofyse tot het afgeven van adreno-corticotropic-hormoon (ACTH), als gevolg waarvan de bijnieren cortisol aanmaken. Via negatieve terugkop- peling remt cortisol de productie van CRH en ACTH door de hypothalamus en de hypofyse. Cortisol is betrokken bij verscheidene functies in het lichaam, zoals de regulering van bloeddruk, het immuunsysteem en het helpt bij het metabolisme van vet, proteïne en koolhydraten. Bij patiënten met secundaire bijnierschorsinsufficiëntie is de endogene cortisolproductie uitgevallen. Het ontbreken van de cortisolproductie kan veroorzaakt worden door disfunctioneren van de bijnieren zelf (primaire bijnierschorsinsufficiëntie), of door disfunctioneren van de hypofyse of hypothalamus (secundaire bijnierschorsinsuf- ficiëntie of tertiaire bijnierschorsinsufficiëntie, respectievelijk). Dit proefschrift focust op patiënten met secundaire bijnierschorsinsufficiëntie. De diagnose van secundaire bijnierschorsinsufficiëntie wordt gebaseerd op lage cortisolwaardes in de vroege ochtend. In het geval van cortisolwaardes die niet eenduidig wijzen op bijnierschor- sinsufficiëntie, zijn er verscheidene stimulatietests beschikbaar die gebruikt kunnen worden om de integriteit en het functioneren van de HPA-as te onderzoeken. Patiënten met secundaire bijnierschorsinsufficiëntie worden behandeld met glucocorticoïden, voornamelijk door de orale toediening van hydrocortison (HC). In eerste instantie werd aanbevolen om tweederde (20 mg) van de dosering in de ochtend te nemen, en éénderde (10 mg) in de avond. Deze dosering is gebaseerd op een geschatte cortisolproductie van 12–15 mg/m²/dag bij gezonde vrijwilligers. Echter, met behulp van stabiele isotopen wordt de dagelijkse cortisolproductie tegenwoordig geschat op 6–10 mg/m²/dag, wat overeenkomt met een totale dagelijkse dosering van 15–20 mg HC/ dag. Parameters om de kwaliteit van de substitutietherapie objectief te meten ontbreken. Daarom wordt in de huidige praktijk een klinische beoordeling van symptomen die mogelijk onder- of overbehandeling suggereren vaak gebruikt. Onderbehandeling heeft als risico dat de cortisolwaardes ontoereikend zijn in geval van ernstige stress, wat kan leiden tot een bijniercrisis. Chronische blootstelling aan te hoge cortisolwaardes is geassocieerd met een verhoogde kans op mortaliteit en morbiditeit zoals cardiovas- culaire ziekten en osteoporose. Daarnaast worden hoge cortisolwaardes geassocieerd met een verminderde kwaliteit van leven en beperkingen in het cognitief functioneren.

Hoofdstuk 1 geeft een introductie over de HPA-as, de etiologie, diagnose en behandeling van secundaire bijnierschorsinsufficiëntie en over verschillende psychologische en somatische uitkomsten die beïnvloed kunnen worden door de behandeling van secundaire bijnierschorsinsufficiëntie. Daarnaast worden de doelen van dit proefschrift 196 Samenvatting

beschreven: het evalueren van het pathofysiologische effect van twee verschillende doseringen HC en het toevoegen van evidentie voor de huidige aanbevelingen met betrekking tot glucocorticoïde substitutietherapie in secundaire bijnierschorsinsuf- ficiëntie. Om deze doelen te bereiken hebben we een gerandomiseerde, dubbelblinde crossover studie opgezet waarin we de behandeling met een totale dagelijkse dosering van 0.2–0.3 mg HC/kg lichaamsgewicht vergeleken met de behandeling met een totale dagelijkse dosering van 0.4–0.6 mg HC/kg lichaamsgewicht (allebei 10 weken toege- diend), en het effect van deze doseringen op het cognitief functioneren, kwaliteit van leven, de associatie tussen stress en pijn, somatosensorisch functioneren, bloeddruk en regulerende hormonen en farmacokinetische parameters.

Deel I. Het effect van behandeling met hydrocortison op psychologische uitkomstmaten bij patiënten met secundaire bijnierschorsinsufficiëntie

In Hoofdstuk 2 beschrijven we het effect van twee verschillende doseringen HC op de cognitieve domeinen geheugen, aandacht, executief functioneren en sociale cognitie. In totaal werden 63 patiënten met secundaire bijnierschorsinsufficiëntie geïncludeerd in deze studie waarvan 47 beide studieperiodes voltooiden en in de huidige analyse zijn gebruikt. Patiënten kregen of eerst tien weken de lagere dosering (0.2–0.3 mg HC/ kg lichaamsgewicht) gevolgd door tien weken de hogere dosering (0.4–0.6 mg HC/ kg lichaamsgewicht), of in omgekeerde volgorde. Aan het eind van iedere behandelperiode werd het cognitief functioneren gemeten door het gebruik van een testbatterij bestaande uit 12 gestandaardiseerde cognitieve testen. Cortisolwaardes gemeten één uur na inname van de ochtenddosering verschilden significant tussen de twee doseringen van HC (gemiddelde [SD], lage dosering: 653 [281] nmol/L; hoge dosering: 930 [148] nmol/L; P < 0.001). Er waren geen verschillen in het cognitief functioneren tussen de beide doseringen, noch in het aantal patiënten dat een cognitieve beperking liet zien. Er lijkt dus geen negatief effect te zijn van een behandeling van 10 weken met een hogere dosering HC, vergeleken met een lagere dosering HC, op geheugen, aandacht, executief functioneren en sociale cognitie.

In Hoofdstuk 3 hebben we kwaliteit van leven geëvalueerd na behandeling met dezelfde twee doseringen HC als eerder beschreven. Cross-sectionele studies wijzen op een relatie tussen de dosering van HC en kwaliteit van leven, waarbij hogere doseringen geassocieerd zijn met ernstigere beperkingen in het subjectieve welzijn. Echter, gerandomiseerde klinische studies naar deze relatie ontbreken. In deze studie hebben we gezondheidsgerelateerde kwaliteit van leven zowel dagelijks gemeten met een stemming en klachten vragenlijst, als aan het eind van iedere behandelperiode met enkele andere kwaliteit van leven vragenlijsten. De 47 patiënten die beide stud- Samenvatting 197 ieperiodes hebben volbracht zijn tevens gebruikt voor deze analyse. Patiënten rap- porteerden minder symptomen van depressie (P = 0.016 en P = 0.045 voor de Hospital Anxiety and Depression Scale (meet klachten van angst en depressie) en Patient Health Questionnaire-9 (meet klachten van depressie), respectievelijk), minder algemene en mentale vermoeidheid (P = 0.004 en P = 0.003, respectievelijk, allebei van de Multi- dimentional Fatigue Inventory-20 (MFI-20), een vragenlijst die vermoeidheid meet), verhoogde motivatie (P = 0.021, MFI-20), beter fysiek functioneren (P = 0.041 ), beter algemene gezondheid (P = 0.013) en meer vitaliteit (P = 0.025) (allen RAND-36, een algemene kwaliteit van leven vragenlijst) na behandeling met de hogere dosering HC in vergelijking met de behandeling met de lagere dosering. Daarnaast werden minder somatische klachten (P = 0.022) en minder pijn (P < 0.001) (allebei Patient Health Questionnaire-15, meet de mate van somatische klachten) gerapporteerd op de hogere dosering in vergelijking met de lagere dosering. We concluderen dat patiënten na behandeling met de hogere dosering een betere kwaliteit van leven ervaren op verscheidene gebieden en dat dit in overweging moet worden genomen bij het afstemmen van de individuele doseringen.

In Hoofdstuk 4 hebben we de causale rol van lage cortisolwaardes in de relatie tussen ervaren stress en pijn onderzocht. Cortisol heeft een pijn-dempende werking en er wordt een mediërende rol van lage cortisolwaardes gesuggereerd in de relatie tussen stress en pijn. Patiënten vulden dagelijks een dagboek in waarin ervaren stress en pijnklachten werden uitgevraagd. We hebben de associaties tussen dagelijks ervaren stress en dagelijks gerapporteerde pijnklachten tijdens de beide HC doseringen getest door middel van non-seasonal autoregressive moving average modelling. Van de 47 patiënten lieten 12 patiënten genoeg variatie zien in ervaren stress- en pijn-niveaus om de relatie tussen deze twee parameters te onderzoeken. Zes patiënten lieten een associatie tussen stress en pijn zien tijdens beide doseringen HC, één patiënt liet deze associatie alleen zien tijdens de lagere dosering en twee patiënten alleen tijdens de hogere dosering. De resultaten van dit onderzoek wijzen niet op een prominente rol van lage cortisolwaardes in de relatie tussen ervaren stress en pijn, zoals onderzocht binnen de gebruikte twee doseringen HC.

Deel II. Het effect van behandeling met hydrocortison op somatische uitkomstmaten bij patiënten met secundaire bijnierschorsinsufficiëntie

In Hoofdstuk 5 hebben we het somatosensorisch functioneren gedurende behandeling met de lagere en hogere dosering HC getest. We hebben eerder laten zien dat men meer pijn ervoer op de lagere dosering HC vergeleken met de hogere dosering. Of deze verschillende doseringen ook een effect hebben op het somatosensorisch functioneren 198 Samenvatting

is nog niet eerder onderzocht. We hebben de mechanische detectiedrempel (MDT), de mechanische pijndrempel (MPT), de mechanische pijngevoeligheid (MPS) en de ‘druk’ pijndrempel (PPT) bepaald met behulp van de ‘Quantitative Sensory Testting’ (QST) batterij van het Duits Netwerk van Neuropathische pijn. Van 46 patiënten was er QST data beschikbaar. We vonden geen verschil in MDT, MPT, MPS en PPT tussen de beide doseringen HC, noch in het aantal patiënten dat een abnormale response liet zien. We concluderen dat eerder gevonden verschillen in subjectieve pijn door een lagere dosering HC niet veroorzaakt worden door veranderingen in het somatosensorisch functioneren in reactie op mechanische stimuli.

In Hoofdstuk 6 beschrijven we het effect van HC op bloeddruk en regulerende hormonen. Mensen met secundaire bijnierschorsinsufficiëntie hebben een verhoogd risico op hart- en vaatziekten en dit wordt mogelijk veroorzaakt door een ongunstig metabool profiel, waarvan verhoogde bloeddruk een belangrijke factor is, als gevolg van een relatief hoge dosering HC als substitutietherapie. De relatie tussen HC dosering en bloeddruk en de onderliggende mechanismen is nog niet uitgebreid gedocumenteerd. Daarom hebben we gekeken naar het effect van de lagere dosering HC vergeleken met de hogere dosering HC op bloeddruk, het renine-agiotensine-aldosteronsysteem (RAAS), 11β-hydroxysteroid dehydrogenase (11β-HSD) enzymactiviteit en (nor) metanefrines. In totaal zijn 46 patiënten geïncludeerd in de analyse. Na behandeling met de hogere dosering HC vonden we een gemiddelde (SD) stijging in de systolische bloeddruk van 5 (12) mm Hg (P = 0.011), een stijging in de diastolische bloeddruk van 2 (9) mm Hg (P = 0.050), een mediane [interkwartielafstand] daling in plasma kalium van -0.1 [-0.3; 0.1] nmol/L (P = 0.048) in vergelijking met de lagere dosering. De hogere dosering HC leidde tot een daling in serum aldosteron van -28 [-101; 9] pmol/L (P = 0.020) en plasma renine van -1.3 [-4.5; 1.2] pg/mL (P = 0.051) en tot een stijging in de cortisol-cortison-ratio in plasma en urine (inclusief de metabolieten) (P < 0.001 voor allen) in vergelijking met de lagere dosering. Daarnaast zagen we na behandeling met de hogere dosering HC een daling in de normetanefrine concentratie in plasma en urine van -0.101 [-0.242; 0.029] nmol/L (P < 0.001) en -1.48 [-4.06; 0.29] µmol/ mol creatinine (P < 0.001) voor normetanefrine in plasma en urine, respectievelijk, in vergelijking met de lagere dosering. Concluderend kunnen we stellen dat de hogere dosering HC leidt tot een stijging in bloeddruk, gepaard gaande met veranderingen in verscheidene mechanismen die betrokken zijn bij de bloeddrukregulatie zoals het RAAS, 11β-hydroxysteroid dehydrogenase (11β-HSD) enzym activiteit en activiteit van het sympathische zenuwstelsel.

In Hoofdstuk 7 evalueren we de farmacokinetische eigenschappen van de twee verschillende doseringen HC. Verschillende methodes worden gebruikt om de adequaatheid Samenvatting 199 van HC substitutietherapie in patiënten met secundaire bijnierschorsinsufficiëntie te meten, maar geen van deze methodes is toereikend. Er zijn een aantal farmacokinetische studies gedaan, maar geen van deze studies heeft gekeken naar zowel totaal cortisol als vrij cortisol gemeten in plasma en vrij cortisol gemeten in speeksel. Om een beter begrip te krijgen van de individuele en populatie farmacokinetische eigenschappen van HC en het effect van aanpassingen in de dosering, hebben we deze post-hoc farmacokinetische analyse gedaan. Van 46 patiënten was farmacokinetische data beschikbaar. Er werden één- en twee-compartimentele populatie modellen voor vrij cortisol en totaal cortisol gemeten in plasma en vrij cortisol gemeten in speeksel gemaakt. De individuele farmacokinetische parameters klaring (CL), verdelingsvolume (Vd), de halfwaardetijd (t1/2), maximale concentratie (Cmax) en area under the curve (AUC) zijn berekend. We vonden grote interindividuele variatie in CL en Vd voor vrij cortisol en totaal cortisol gemeten in plasma en vrij cortisol uit speeksel, waarbij AUC24h tot een factor tien verschilde. Blootstelling aan cortisol was verhoogd tijdens de hogere dosering in vergelijking met de lagere dosering, maar dit was alleen proportioneel aan de dosis voor de vrije cortisol concentraties en niet voor totaal cortisol. De resultaten van de populatiemodellen laten zien dat er individueel afgestemde doseringen nodig zijn vanwege de grote interindividuele verschillen in cortisolblootstelling. Voor patiënten bij wie het metabolisme relatief snel gaat, leidde een verdubbeling van de dosering niet tot een verdubbeling in blootstelling aan cortisol, waardoor er voor deze patiënten andere strategieën nodig zijn om in hun cortisolbehoefte te voorzien.

Hoofdstuk 8 is een algemene discussie over de rol van glucocorticoïde receptoren en mineralocorticoïde receptoren die de relatie tussen cortisol en cognitieve functies en andere uitkomstmaten mediëren. Verder wordt de rol van kwaliteit van leven in het evalueren van de HC substitutietherapie besproken. Daarnaast bediscussiëren we de rol van mechanismen die, naast de cortisol concentraties in de circulatie, de blootstelling aan cortisol regelen en bespreken we de toepassing van HC tabletten met gereguleerde afgifte als optie voor de behandeling. Tenslotte worden enkele aanbevelingen voor vervolgonderzoek gedaan.

Dankwoord

Dankwoord 203

Daar ligt ‘ie dan! Dat ik dit dankwoord mag schrijven betekent dat het echt zo ver is: mijn proefschrift is klaar! Maar een proefschrift schrijf je niet alleen, dus een dankwoord voor een ieder die hier direct of indirect aan heeft bijgedragen is zeker op z’n plaats.

Allereerst wil ik de bereidwillige patiënten bedanken die aan dit onderzoek hebben deelgenomen. Zonder hen was dit proefschrift nooit tot stand gekomen.

Mijn promotoren prof. dr. Bruce H.R. Wolffenbuttel en prof. dr. Oliver M. Tucha:

Beste Bruce, ik wil je bedanken voor de kans die je me hebt gegeven om op de afdeling Endocrinologie mijn promotietraject te volgen. Je kritische inbreng de afgelopen jaren is van grote waarde geweest voor dit proefschrift. Ook wil ik je bedanken dat je me op mijn eerste grote internationale congres in Polen, waar ik waarschijnlijk ietwat ontredderd om me heen stond te kijken, op sleeptouw hebt genomen.

Beste Oliver, ik vind het bewonderenswaardig hoe snel jij de belangrijkste boodschap uit een brei van data en cijfertjes weet te filteren. Ik heb genoten van je positiviteit en de prettige en inspirerende manier waarop je me tijdens mijn onderzoek hebt begeleid.

Veel dank gaat uit naar mijn copromotoren dr. André P. van Beek en dr. Janneke Koerts:

Beste André, tijdens mijn scriptie onderzoek polste jij of ik ook geïnteresseerd was in onderzoek doen. Misschien dat ik toen niet helemaal wist waar ik aan begon, maar ik heb er zeker geen spijt van gekregen. Je enthousiasme voor dit vak en de wetenschap werkt aanstekelijk; zo liep ik na ieder overleg vrolijk en vol goede moed je kamer uit. Je geduld en manier van uitleggen heb ik als zeer prettig ervaren. Onze gesprekken over nieuwe hardlooprecords en afleveringen van Game of Thrones en Rundfunk, al dan niet onder het genot van een kop ‘lekkere koffie’ of een biertje op congressen, hebben daar ook zeker aan bijgedragen. André, bedankt voor het vertrouwen en de fijne begeleiding!

Beste Janneke, bedankt dat je altijd voor me klaar stond als ik even advies nodig had. Hoe vol je agenda soms ook was, er was altijd ruimte om even te sparren. Je afsluiting ‘niet te druk maken, het komt goed’, had ik af en toe zeker nodig. Janneke, bedankt voor de fijne samenwerking en ik hoop dat onze wegen elkaar in de toekomst weer zullen kruisen! 204 Dankwoord

I thank the members of the reading committee, prof. dr. A.R.M.M. Hermus, prof. dr. O.C. Meijer, prof. dr. B.R. Walker for their time to critically read and evaluate this thesis.

Veel dank gaat uit naar de afdeling Endocrinologie en aan alle co-auteurs betrokken bij de verschillende hoofdstukken. In het bijzonder wil ik prof. dr. Daan Touw bedanken. Beste Daan, met veel geduld heeft u me thuis gemaakt in een deel van de wereld van de farmacie; voor een alfa-mens als ik niet altijd even makkelijk. Als ik er even niet uit kwam had u altijd een gaatje vrij in uw drukke agenda om te overleggen. Dit heb ik erg gewaardeerd, bedankt daarvoor!

Machteld, ik wil je ontzettend bedanken voor het ontwerpen van de prachtige voorkant van dit proefschrift.

Verder wil ik iedereen bedanken van de afdeling Neuropsychologie. Het was fijn om tijdens mijn hele promotietraject een connectie te houden met de psychologie, mede door de inspirerende research meetings. Ik heb me er erg welkom gevoeld en beseft dat psychologie toch echt het leukste vak is.

Veel dank gaat uit naar de mensen van het lab en in het bijzonder naar Martijn van Faassen. Ook het lab was een compleet nieuwe wereld voor mij. Martijn, bedankt voor het me wegwijs maken in het pipetteren, dialyseren en opwerken van de monsters. Dank dat je mijn vraagbaak wilde zijn voor al mijn lab-gerelateerde vragen.

Natuurlijk wil ik ook mijn mede-onderzoekers Edward, Sandra, Sarah, Dineke, Robert, Karin, Marloes, Thamara en Mariëlle bedanken voor de leuke en leerzame tijd die we hebben gehad. Dames, heerlijk om samen in dit promotietraject op te trekken en onder toeziend oog van Manuel af toe te sparren over belangrijke en minder belangrijke dingen in het leven. Mannen, fijn om ook af en toe iemand te hebben die mij weer even liet relativeren en met beide benen op de grond zette.

Daarnaast wil ik Pauline Brummelman nog even extra benoemen. In 2012 mocht ik mijn scriptie onder jouw begeleiding schrijven. Jij stond aan de wieg van de trial beschreven in dit proefschrift, dus ook zonder jou was dit proefschrift er niet geweest. Mede door onze gezellige samenwerking is mijn interesse voor de wetenschap opgebloeid. Inmiddels ben je werkzaam als GZ-psycholoog (ik ben onderweg;)) en hoop ik je op professioneel en privégebied nog vaak tegen te komen! Dankwoord 205

Lieve teamgenootjes van Groningen Dames 1. Wanneer hockey zo’n grote rol speelt in iemands leven als in het onze, verdienen jullie zeker een plekje in dit proefschrift. Hoewel ik mezelf soms naar training toe moest slepen, sta ik nog steeds te genieten als we met z’n allen op het veld staan. We hebben lief en leed gedeeld, met als hoogtepunt natuurlijk onze promotie naar de Hoofdklasse en onze recent behaalde zilveren medai- lle op het NK zaalhockey. Dank voor de vele avondjes Cirkel en borreltjes op de club waardoor ik weer de energie kreeg om door te gaan met dit promotietraject!

Lieve JC Gula, lieve clubgenoten. Ik ben heel gelukkig met zoveel leuke mensen om me heen! Als enige nog over in Groningen zie ik jullie veel minder vaak dan ik zou willen. Dank voor de gezellige borrels, etentjes en weekendjes weg op z’n tijd, maar bovenal jullie begrip en interesse in wat ik nou allemaal uitspookte daar in het hoge noorden. Nu tijd voor een borrel in de stad waar het allemaal begon!

Lieve Hedi, 6 jaar geleden maakte ik de overstap naar Groningen Dames 1 en kwam ik bij jou in het team. Al vrij snel wist, ja, dit is een leukerd! Ik wil je bedanken dat je altijd voor me klaar staat met een knuffel, een goed glas wijn en niet te vergeten: kaas- jes! Onze etentjes en avondjes in de Soest hebben me de nodige afleiding en energie gegeven om dit traject af te maken. Ik vind het heel fijn om je als paranimf aan mijn zijde te hebben!

Lieve papa, Irene en Ellen. Lieve mama, Coen en Yannick. Ik kan de juiste woorden nog niet vinden om jullie te bedanken, maar ik zal een poging wagen. Jullie hebben me de ruimte gegeven om te ontdekken en te doen wat ik zelf wil en me daarin altijd gesteund. Bedankt dat ik bij jullie kon komen om af en toe weer even bij te tanken en voor jullie onvoorwaardelijke steun en liefde!

Lieve Rein en Nicky, lief broertje en zusje, wat is het fijn om jullie om me heen te hebben! Lieve Nick, wat hebben we een bijzonder gave reis gemaakt afgelopen zomer. Dat maakte de laatste loodjes net even wat makkelijker. Wanneer gaan we weer? Lieve Rein, naast dat ik je weet te vinden als mijn internet of tv weer eens niet naar mijn zin werkt, ben je ook gewoon m’n maatje en vind ik het super leuk dat je vandaag als paranimf naast me staat!

About the author

About the Author 209

Jorien Werumeus Buning was born on December 29th 1987 in Steenwijk, The Neth- erlands. In 2006 she graduated from Regionale Scholengemeenschap Steenwijk. The same year she started her study Psychology at the Faculty of Behavioural and Social Sciences at the University of Groningen. She received her Bachelor’s degree in 2011 and her Master’s degree in Clinical Neuropsychology in 2013. From 2013 till 2016 she worked as a PhD student under the supervision of prof. dr. B.H.R. Wolffenbuttel, prof. dr. O. M. Tucha, dr. A. P. van Beek and dr. J. Koerts at the department of Endo- crinology of the University Medical Center Groningen and the department of Clinical and Developmental Neuropsychology at the University of Groningen. During her PhD training she was able to present her research at several national and international conferences, for which she received several travel grants and an award for the best abstract in Clinical Endocrinology. In January 2017 Jorien started working as a psychologist at Kinnik, a child and youth mental health care organization.

List of publications

List of publications 213

Werumeus Buning J, Touw DJ, Brummelman P, Dullaart RPF, van den Berg G, van der Klauw MM, Kamp J, Wolffenbuttel BHR, van Beek AP. Pharmacokinetics of hydrocortisone: results and implications from a randomized controlled trial. Accepted for publication in Metabolism.

Janssens KAM, Werumeus Buning J, van Beek AP, Rosmalen JGM. The role of cortisol in the association between perceived stress and pain: a short report on secondary adrenal insufficiency patients. Journal for Person-Oriented Research 2016; 2(3):135–141.

Werumeus Buning J, van Faassen M, Brummelman P, Dullaart RPF, van den Berg G, van der Klauw MM, Kerstens MN, Stegeman CA, Muller Kobold AC, Kema IP, Wolffenbuttel BHR, van Beek AP. Effects of hydrocortisone on the regulation of blood pressure – results from an RCT. Journal of Clinical Endocrinology & Metabolism 2016; 101(10):3691–3699.

Werumeus Buning J, Kootstra-Ros JE, Brummelman P, van den Berg G, van der Klauw MM, Wolffenbuttel BHR, van Beek AP, Dullaart RPF. Higher hydrocortisone dose increases bilirubin in hypopituitary patients – Results from an RCT. European Journal of Clinical Investigation 2016; 46(5):475–80.

Werumeus Buning J, Brummelman P, Koerts J, Dullaart RPF, van den Berg G, van der Klauw MM, Sluiter WJ, Tucha O, Wolffenbuttel BHR, van Beek AP. Hydrocor- tisone dose influences pain, depressive symptoms and perceived health in adrenal insufficiency: a randomized controlled trial.Neuroendocrinology 2015; 103(6):771–8.

Werumeus Buning J*, Brummelman P*, Koerts J, Dullaart RPF, van den Berg G, van der Klauw MM, Tucha O, Wolffenbuttel BHR, van Beek AP. The effects of two different doses of hydrocortisone on cognition in patients with secondary adrenal insufficiency – results from a randomized controlled trial. Psychoneuroendocrinology 2015; 55:36–47.

* both authors contributed equally

Brummelman P, Sattler MGA, Meiners LC, van den Berg G, van der Klauw MM, Elderson MF, Dullaart RPF, Koerts J, Werumeus Buning J, Tucha O, Wolffenbuttel BHR, van den Bergh ACM, van Beek AP. Cognition and brain abnormalities on MRI in pituitary patients. European Journal of Radiology 2015; 84(2):295–300.