Waves of change: Sex hormones, Depression and Sleep Margot W. L. Morssinkhof
Copyright © Margot W. L. Morssinkhof, 2024 All rights reserved. No part of this thesis may be reproduced, stored, transmitted in any way or by any means without the prior permission of the author. Design: van Kira | www.vankira.nl/phd Lay-out: Tiny Wouters Printing: Ridderprint | www.ridderprint.nl DOI: http://doi.org/10.5463/thesis.676 ISBN: 978-94-6506-061-3 The work presented in this thesis was partly funded by a grant from ZonMW (nr. 91619085) and by the Amsterdam UMC Young Talent Fund. Philips supported the work in this thesis by providing the measurement equipment used in the research on loan.
VRIJE UNIVERSITEIT Waves of change: Sex hormones, Depression and Sleep ACADEMISCH PROEFSCHRIFT ter verkrijging van de graad Doctor of Philosophy aan de Vrije Universiteit Amsterdam, op gezag van de rector magnificus prof.dr. J.J.G. Geurts, in het openbaar te verdedigen ten overstaan van de promotiecommissie van de Faculteit der Geneeskunde op vrijdag 28 juni 2024 om 9.45 uur in een bijeenkomst van de universiteit, De Boelelaan 1105 door Margaretha Wilhelmina Laurence Morssinkhof geboren te Enschede
Promotoren prof.dr. B.F.P. Broekman prof.dr. M. den Heijer Copromotoren prof.dr. O.A. van den Heuvel prof.dr. Y.D. van der Werf Promotiecommissie prof.dr. C.H. Vinkers prof.dr. M. Lancel prof.dr. P.H.L.T. Bisschop prof.dr. E. De Bruijn dr. M.A. Bremmer
“When a subject is highly controversial – and any question about sex is that – one cannot hope to tell the truth. One can only show how one came to hold whatever opinion one does hold.” – Virginia Woolf
Table of contents Chapter 1 General Introduction 9 Chapter 2 Oral contraceptives, depressive and insomnia symptoms 41 in adult women with and without depression Chapter 3 Changes in depression symptom profile with gender- 69 affirming hormone use in transgender persons Chapter 4 Cortisol dynamics and sleep quality: the role of sex and 103 oral contraceptive use Chapter 5 Sex hormones, insomnia, and sleep quality: subjective sleep 133 in the first year of hormone use in transgender persons Chapter 6 Influence of sex hormone use on sleep architecture in 165 a transgender cohort: findings from the prospective RESTED study Chapter 7 Chronotype changes after sex hormone use: a prospective 195 cohort study in transgender users of gender- affirming hormones Chapter 8 Associations between sex hormones, sleep problems and 215 depression: a systematic review Chapter 9 General discussion 255 Appendix 295 English summary 297 Nederlandse samenvatting 305 Author contribution statement 313 Publications 317 Portfolio 319 Dankwoord 323 About the author 329
General introduction 11 1. General introduction 1.1. Sex differences in mental health in history Long before modern psychiatric classifications, healthcare providers believed that their patients’ sex could affect their mental health. Physicians from the ancient Egyptian and Greek societies believed that in females, the uterus could make “spontaneous, wandering movements” through the body. This could result in anxiety and nervousness, insomnia, low mood, irritability and a wide range of somatic symptoms. The Greek physician Hippocrates called this syndrome “hysteria”, named after the Greek word for uterus (Trimble & Reynolds, 2016), which was possibly the first sex-specific diagnosis in psychiatry. Although its meaning changed over the course of history, diagnoses of hysteria were most prevalent in women, and they were a commonplace occurrence up until the first half of the 20th century. The use of “hysteria” as a diagnostic term declined, and in 1980 references to hysteria were removed from psychiatric handbooks (Tasca et al., 2012). However, sex differences still exist in the prevalence of psychiatric illnesses: women are two times more likely to experience depression and 1.5 times more likely to experience insomnia compared to men (Kessler, 2003; Zhang & Wing, 2006). Although these prevalence differences have been attributed to social and psychological factors, researchers hypothesize that biological factors also contribute to higher prevalence of depression and insomnia (Altemus et al., 2014; Suh et al., 2018). One of the biological factors that gained more research attention is the role of sex hormones (SchweizerSchubert et al., 2021). There is an increased risk of mood disorders and insomnia in the female lifespan during times when sex hormones fluctuate. This is seen around pregnancy and in the peri-menopause: 10 to 15% of pregnant people experience peripartum depression, and risk of depression doubles in perimenopause (de Kruif et al., 2016; Shorey et al., 2018). Furthermore, 3 to 18% of women report depressive moods associated with their menstrual cycle (Halbreich et al., 2003). The presence of significant sex differences in the prevalence of depression and insomnia, and the increased risk of both disorders in life phases when hormones strongly change have raised questions on the possible role of sex hormones: could sex hormones actually be a contributing factor for these conditions?
Chapter 1 12 1.2. Central theme of this thesis To better understand the impact of exogenous sex hormones and to elucidate the sex disparities in depression and insomnia, in thesis we aimed to study the effects of exogenous hormone use, specifically oral contraceptives and gender-affirming hormones, on depression and sleep. 2. Background: Sex hormones, depression and sleep 2.1. Sex hormones Estrogen, progesterone and testosterone are the body’s main sex hormones. Estrogen and progesterone are generally regarded as “female sex hormones” and testosterone as the “male sex hormone”, although all three sex hormones are present in both sexes. Generally, externally administered hormones forms are called exogenous sex hormones, to distinguish from hormones which are naturally produced by the body, which are called endogenous sex hormones. Endogenous sex hormones Endogenous hormone production is a well-orchestrated process that is coordinated by the Hypothalamic-Pituitary-Gonadal (HPG) axis. The hypothalamus produces gonadotropin-releasing hormone (GnRH), which triggers the anterior pituitary to produce luteinizing hormone (LH) and follicle-stimulating hormone (FSH), which in turn stimulate the gonads to regulate reproductive processes and produce sex hormones (Dwyer & Quinton, 2019). The female gonads produce progesterone and estrogen and the male gonads produce testosterone. The production of sex hormones is regulated by a negative feedback loop: in both males and females, estrogen regulates the release of LH, FSH and GnRH (Pitteloud et al., 2008a; Shaw et al., 2010), and testosterone also regulates GnRH release (Marques et al., 2000; Pitteloud et al., 2008b). This process is also illustrated in Figure 1.1. Estrogen and progesterone The HPG axis in female babies is shortly activated during prenatal development, and again in the postnatal period: this activation in the postnatal period is also called the neonatal minipuberty. The postnatal
General introduction 13 activation of the HPG-axis results in increases in GnRH, LH, FSH and estrogen, which peak two months after birth in female babies and then decline until the age of two (Dwyer & Quinton, 2019). In the female minipuberty, LH and FSH show steady declines in serum levels, but estrogen already shows a cyclic decline in serum levels in this phase (Kuiri-Hänninen et al., 2014). Although the function of this phase is not yet clear, researchers hypothesize that it affects both gonadal and neural development (Hines et al., 2016). After the minipuberty, estrogen and progesterone levels are low during childhood and increase again with the start of puberty, starting the development of secondary sex characteristics such as breast growth and redistribution of body fat. During the reproductive age, most women show predictable and regular changes in estrogen and progesterone levels in a roughly 28-day cycle. Estrogen levels peak around 14 days after the first day of menstrual bleeding, marking ovulation, and progesterone levels peak around 21 days after the first day of menstrual bleeding, after which levels of estrogen and progesterone decline and trigger menstruation. This predictable rhythmicity in hormone levels ends during the menopausal transition, when menstrual cycles first becomes irregular and then stop altogether. During the menopausal transition, levels of estrogen and progesterone decline, but day-to-day levels of estradiol can still strongly fluctuate within this downward slope (Gordon et al., 2016; Joffe et al., 2020). Estrogen and progesterone also play a small part in the male HPG axis. In men, testosterone can be aromatized into estradiol (17β-estradiol, or E2), and this process plays important role in the regulation of gonadal hormones. GnRH can inhibit testosterone production, but the regulation of FSH is mainly regulated via estradiol (Hayes et al., 2001). In men, progesterone serves as a precursor for testosterone, and progesterone levels are similar compared to postmenopausal women (Muneyyirci-Delale et al., 1999). These studies indicate that although estrogen is often overlooked in male physiology, it serves an important function in male sex hormone regulation. Testosterone As seen in female babies, male babies also undergo two phases of hormone exposure during early development. Firstly, exposure to testosterone during fetal development triggers the sexual differentiation of the gonadal system in male babies. Secondly, in the postnatal period male babies also go
Chapter 1 14 through the so-called neonatal minipuberty, showing increases in LH, FSH and testosterone in the first six months of life (Kuiri-Hänninen et al., 2014). Studies also show that this postnatal minipuberty can affect gonadal and neuronal development in male babies (Hines et al., 2016). Following this phase, testosterone levels are also low during childhood and increase during puberty in boys, stimulating the development of masculine secondary sex characteristics such as beard growth, low voice and male musculature. Although reproductive-age males do not show the pronounced hormone fluctuations that is seen in females, testosterone levels in adult males do fluctuate in a 24-hour rhythm: they are highest early in the morning, and decline during the daytime. During sleep at night, especially deep sleep, testosterone production is highest, resulting in high testosterone levels in the morning again. There is no marked point at which testosterone production strongly declines in men: some studies show that testosterone levels decrease with age (Feldman et al., 2002; Harman et al., 2001), but it is not yet clear whether this directly caused by ageing or indirectly via increased comorbidity and lifestyle factors, especially obesity (Dean et al., 2015; Tajar et al., 2012). Some men show very low testosterone levels, which is also called hypogonadism, which can be treated with testosterone supplementation. Testosterone levels in women are lower than in men. Testosterone in women also shows small circadian fluctuations, with higher levels in testosterone in the morning (Bungum et al., 2013), as well as fluctuations throughout the menstrual cycle, when it peaks around ovulation and remains higher in the luteal phase (Rothman et al., 2011). Throughout the lifespan, testosterone levels also decline in women from the age of 45, independently from the menopausal transition (Burger et al., 2000; Davison et al., 2005). Although excessively high testosterone levels in women are be associated with hormonal disorders, most commonly with polycystic ovarian syndrome (Rodriguez Paris & Bertoldo, 2019), the clinical significance of testosterone in women is not so clear as it is in men, and testosterone supplementation in women is not part of regular clinical care.
General introduction 15 Figure 1.1. Image of the effect of endogenous sex hormones and exogenous sex hormones on the Hypothalamic-Pituitary-Gonadal axis. GnRH=Gonadatropin-Releasing Hormone, FSH= Follicle-Stimulating Hormone and LH=Luteinizing Hormone, AR=Androgen Receptor, ER= Estrogen Receptor, PR=Progesterone Receptor. Created using Biorender.com. Exogenous sex hormones The molecule structures of oestrogen, progesterone and testosterone were discovered in the late 1920s and early 1930s (Tata, 2005). These discoveries gave rise to research into the production of synthetic versions of these hormones, and in the 1940s, researchers developed methods to synthesize exogenous sex hormones to be used as therapies (Sturgis & Albright, 1940). In the decades following these experiments, scientists discovered ways to modify or optimize the molecule structures of exogenous sex hormones in order to optimize absorption (Strecker et al., 1979), adapt effects on the menstrual cycle (Pincus et al., 1958) and to reduce the risk of adverse effects, such as risk for thrombosis (Inman et al., 1970). These discoveries led to the release of the first forms of hormone therapy for menopause
Chapter 1 16 symptoms in the 1940s (Stefanick, 2005) and the development and approval of the first oral contraceptive pill in 1960 (Junod & Marks, 2002). Nowadays, exogenous hormones are used for a wide range of indications, ranging from fertility (e.g. contraception or fertility treatments) to symptom relief when endogenous sex hormone levels are outside of the eugonadal physiological range (e.g. hormone therapy for hypogonadism or menopausal symptoms) or to change physical characteristics in transgender persons (e.g. use of gender-affirming hormones). Exogenous hormones have similar effects to endogenous hormones, since they can bind to and activate the same hormone receptors. Due to the negative feedback mechanism in the HPG-axis, exogenous hormones can also suppress endogenous sex hormone production, as shown in Figure 1.1. However, exogenous hormones can differ from endogenous hormones in terms of potency, selectivity and systemic effects, which depends on the chemical composition and administration route of the hormones. Firstly, the potency of hormones can differ between hormone formulations. For example, ethinylestradiol, which is a non-bioidentical form of estradiol, is 100 times more potent than bioidentical estradiol (Jeyakumar et al., 2011). Secondly, differences in selectivity of hormone forms can change the effects of hormone formulations: the progestin levonorgestrel, which is present in many forms of hormonal contraceptives, not only binds to progesterone receptors but also to testosterone and cortisol receptors (Kuhl, 2005). These effects also seem to affect cortisol dynamics, since levonorgestrel-containing hormonal contraceptives were found to affect cortisol responsivity (Aleknaviciute et al., 2017; Herrera et al., 2019). Thirdly, differences in the hormone administration form can also influence the systemic effects of hormones: use of oral estradiol is associated with an increase in sex hormone-binding globulin (SHBG), which can bind to sex hormones and reduce the amount of free estradiol and testosterone, whereas the use of transdermal estradiol was not found to affect SHBG concentrations (Campagnoli et al., 2002; Ropponen et al., 2005). Lastly, effects of exogenous hormones could also be dependent on the timeframe in which they are used. Some exogenous hormones are used daily, which results in a relatively stable hormone level with small day-to-day fluctuations, while other exogenous hormones can be used weekly, monthly or every 3 months, resulting in stronger fluctuations in hormone levels. Differences in day-to-day sex hormone dynamics are during use of exogenous hormones also illustrated in Figure 1.2.
General introduction 17 Oral contraceptives Oral contraceptives (OCs) are the most commonly used form of hormonal contraceptives: worldwide, an estimated 151 million women use OCs (Haakenstad et al., 2022), with the majority using combined oral contraceptives, which consist of a combination of progestins and estrogens. Although intended use of OCs is for contraceptive purposes, it is also commonly used for non-contraceptive purposes, to regulate menstruation timing or to manage symptoms of gynaecological conditions. Although use of OCs has local effects on the uterus and ovaries, they also have systemic effects. The most well-studied risk resulting from these systemic effects are a slight increase in risk of breast cancer and thrombosis, but other possible side effects include weight gain, headaches, breast tenderness and possible mood changes (Burkman, 2001). Other possible systemic effects of OCs include changes in cortisol dynamics, with studies indicating that OC users show higher cortisol levels at rest (Boisseau et al., 2013), reduced cortisol responses to after awakening or stress (Høgsted et al., 2021; Nielsen et al., 2013) and chronic increased glucocorticoid signalling (Hertel et al., 2017). Despite increasing insights in these effects of OCs, it is yet unknown how some systemic effects, including changes in cortisol dynamics, relate to OC-associated side effects. It is estimated that more than 50% of OC starters discontinue OC use within 6 to 12 months, with 27% of OC starters reporting mood changes and 38% of starters reporting side effects as one of the most frequent reasons for discontinuation (Hall et al., 2012, 2014; Westhoff et al., 2007). Recent statistics show that the use of OCs is decreasing in youth, with user rates declining from 76% in 2012 to 46% in 2023 in youth aged 13 to 24 (De Graaf et al., 2023). In this group, reported side effects were one of the three most common reasons not to use OCs. Gender-affirming hormone therapy Gender-affirming hormone therapy (GAHT) is used by transgender persons, whose gender identity is not in line with their sex assigned at birth. GAHT can change their physical characteristics to be more in line with their gender identity. Transgender people commonly experiences gender dysphoria, which is the experience that one’s body is not in line with the gender that they identify with.
Chapter 1 18 There are, generally speaking, two forms of GAHT: masculinizing and feminizing hormones. Masculinizing GAHT is used by transmasculine (TM) persons who were assigned female at birth, and it consists of the use of testosterone. Feminizing GAHT is used by transfeminine (TF) persons who were assigned male at birth, and it generally consists of the use of estradiol and anti-androgens. In persons using masculinizing GAHT, who still have a uterus, ovulation is not always suppressed (Taub et al., 2020), and it is also possible to use medication to suppress ovulations or suppress menstrual bleeding using progestins or hormonal contraceptives. The effects of GAHT on the HPG-axis and daily hormone fluctuations are also illustrated in Figure 1.1 and 1.2. Figure 1.2. Image of hormone levels in a 28-day cycle without use of exogenous hormones compared to hormone levels during use of Gender-Affirming Hormones and Oral Contraceptives. The left panel (yellow) displays the endogenous hormone levels in females (top panel) and males (bottom panel). The middle panel displays hormone levels in users of oral contraceptives, with stable daily fluctuations of progesterone and estradiol for 21 days and a steep reduction in estradiol and progesterone in the placebo week. The right panel (blue) displays possible hormone fluctuations in GAHT users: the top panel displays hormone levels after masculinizing hormone use, with low progesterone levels and either small daily fluctuations (in users of testosterone gel) or a fluctuation over the course of 3 weeks (in users of testosterone injections), as well as partial but not full suppression of progesterone. The possible 12-week fluctuation in testosterone levels in users of long-acting testosterone is not displayed. The bottom panel displays hormone levels after feminizing hormone use, with low testosterone levels and either daily fluctuations in estrogen (in users of estradiol forms used daily) or multiday fluctuations (in the case of estradiol plasters, which are used biweekly). Created using Biorender.com.
General introduction 19 2.2. Depression Major depressive disorder (MDD), commonly called depression, is one of the most prevalent mental disorders worldwide, with a high impact on quality of life. Although presence of key symptoms (e.g. depressed mood or anhedonia) is required for a clinical diagnosis of depression, the symptoms that depressed people report vary widely: one can experience increases or decreases in appetite, increased propensity to sleep or total inability to sleep, and physical restlessness or a total slowing down of physical activity (American Psychiatric Association, 2013). This diversity in depression symptoms is also called heterogeneity, which refers to the idea that depression is not a single condition, but it is a set of symptoms which can manifest differently in different individuals. To gain further insight into heterogeneity in depression, clinicians and researchers have also studied subtypes of depression, such as melancholic or atypical depression, or specific clusters of symptoms. In many people, depression can show a recurrent or chronic course: longitudinal data show that in the 9 years after a depressive episode, an estimated 68% of patients reported at least one recurrent depressive episode (Solis et al., 2021). Sex differences in depression Depression is more common in women compared to men across cultures and countries (Hopcroft & Bradley, 2007). In the Netherlands, the lifetime prevalence of depression in women is 24%, whereas it is 13% in men (de Graaf et al., 2010). As the sex differences are evident, almost all researchers studying depression correct for sex. Most studies do not yet conduct sexstratified analyses, meaning they do not examine whether their research results are generalizable to both sexes (Rechlin et al., 2022). Sex differences are also found show in the type of depressive symptoms and in comorbidity: atypical depression, which is characterized by the absence of anhedonia and the presence of hypersomnia, excessive eating, leaden paralysis and sensitivity to rejection, is up to four times more prevalent in depressed women than in depressed men (Lamers et al., 2010; Łojko & Rybakowski, 2017). Depressed women are also more likely to report somatic and cognitive-affective symptoms than men and they are more likely to present with comorbid anxiety (Altemus et al., 2014). However, studies find
Chapter 1 20 no sex differences in the age of onset or rates of recurrence or chronicity (Rubinow & Schmidt, 2019). Sex hormones and depression during the lifespan Changes in female sex hormones could increase the risk of depression (Schiller et al., 2016). Epidemiological studies show an increased risk of depression during and after pregnancy, and it has been suggested that this is associated with the strong increase and drop in sex hormone levels during pregnancy and postpartum (Eisenlohr-Moul et al., 2023). Furthermore, sex hormones also show strong fluctuations in the perimenopause, which could also contribute to depressed mood (Joffe et al., 2020). Experimental hormone manipulation studies have further explored this hypothesis. Suppression of endogenous estrogen and progesterone production in healthy reproductive-age females resulted in increased depressive symptoms (Ben Dor et al., 2013; Frokjaer et al., 2015). Frokjær et al. (2015) also found an association between the magnitude of endogenous hormone changes and depression: participants showing stronger reductions in estradiol after hormone suppression also reported stronger increases in depressive symptoms. Other studies assessed effects of hormone suppression and add-back of exogenous hormones in supraphysiological doses: they find that in women with a history of postpartum depression, both hormone suppression and add-back of exogenous hormones had adverse effects on mood (Eisenlohr-Moul et al., 2023). Others have hypothesized that testosterone in men could be protective against depression, although studies on this topic are scarce. Hypogonadal men, who have clinically low testosterone levels, are more likely to report depressive symptoms (Hintikka et al., 2009). Men aged 60 or older are more likely to show clinically low testosterone (Kaufman & Vermeulen, 2005), but it is not clear whether this is also associated with clinically relevant symptoms, nor whether this requires hormonal treatment (Yeap et al., 2018). Research on the association between depression and testosterone levels in nonhypogonadal men is mostly inconclusive, with most studies finding no association between testosterone and depression (de Wit, Giltay, de Boer, Nolen, et al., 2021). Studies on the effect of testosterone treatment in depressed men show a possible favourable effect of testosterone treatment on depressive symptoms compared to placebo (Walther et al., 2019),
General introduction 21 although important methodological concerns have been raised in these studies (Bhasin & Seidman, 2019). The relationship between endogenous sex hormones and depression is difficult to study, since there seems to be a bidirectional relationship between depression and the HPG axis. In men, some studies find that depression is associated with lower testosterone levels (Hintikka et al., 2009; Westley et al., 2015), although others do not find this effect (de Wit, Giltay, de Boer, Nolen, et al., 2021). This bidirectional association could be due to changes in health and lifestyle during a depressive episode, including weight gain, lack of exercise and poor sleep, all of which could affect testosterone levels (D’Andrea et al., 2020; Su et al., 2021). In women, cross-sectional studies show higher levels of testosterone in those with remitted or current depressive episodes, but testosterone levels were not predictive of future depressive episodes (de Wit, Giltay, de Boer, Bosker, et al., 2021). 2.3. Sleep Sleep is important for general health and well-being, and disturbances of sleep can be caused by or exacerbate mental health problems. There are numerous aspects of sleep that are used to define sleep health, including subjectively reported sleep quality, sleep timing and objectively measured sleep, using polysomnography (PSG), including electroencephalography (EEG). Subjectively reported sleep concerns one’s perception of sleep quality. This can include subjectively reported sleep duration and disruptions, but also perceived trouble falling asleep, staying asleep or not feeling rested in the morning. If these symptoms are frequently present and cause burden on daily life, these could also be classified as insomnia. Sleep can also be measured using objective measures, such as actigraphy or PSG measurements, including EEG. Actigraphy measurements are conducted using a wrist-worn device which estimates sleep length and stages based on patterns of movement. Sleep EEG measurements, which are often considered the golden standard for sleep measurements, are conducted using electrodes that are directly placed on the scalp, recording electrical activity during the night. Sleep timing can be measured by assessing one’s chronotype, which is the natural preference of the body for the timing of wakefulness and sleep, which is assessed by asking about preferred bed-
Chapter 1 22 and wake-up times. Having a late chronotype can affect sleep health, especially when sleep-wake timings are not synchronous with social obligations (Antypa et al., 2016). People’s perception of their sleep can misalign with objectively measured sleep. People with insomnia commonly show underestimations of sleep durations and longer wakefulness at night (Bianchi et al., 2013). The phenomenon of reporting poor sleep or insomnia without objective sleep disturbances is also called sleep state misperception (Moon et al., 2015). Sex differences in sleep In subjective assessments of sleep, women show a higher likelihood of reporting insomnia and poorer sleep quality, including lower sleep efficiency, more sleep disturbances and more use of sleeping medication (Li et al., 2019). However, when assessing sleep using objective sleep measurements, women on average actually show better sleep than men, including longer sleep duration, less wakefulness during the night, longer slow wave sleep (SWS) and longer rapid eye movement (REM) sleep latency (Bixler et al., 2009; Roehrs et al., 2006). Thus far, there is no explanation yet for the paradoxical higher prevalence of poor sleep and insomnia and concurrent better objective sleep in women. Women also generally report an earlier chronotype than men, meaning they prefer an earlier bedtime and earlier wake-up time (Randler & Engelke, 2019). Sex hormones and sleep during the lifespan The sex difference in the prevalence of insomnia specifically increases from the time of the first menarche in girls (Johnson et al., 2006), suggesting that sex hormone changes could contribute to insomnia risk. Incidence of insomnia also increases during pregnancy and perimenopause (Ciano et al., 2017; Hashmi et al., 2016). Experimental studies have provided evidence indicating sex hormones could contribute to this risk of sleep problems: endogenous hormone suppression in reproductive-age women resulted in increased rates of disturbed sleep (Ben Dor et al., 2013). Vice versa, supplementation of estrogen in women going through perimenopause, was associated with improved sleep quality compared to the placebo group (Silva et al., 2011), and Bixler et al. (2009) found that postmenopausal women using hormone therapy showed shorter sleep onset latency and longer slow wave sleep compared to postmenopausal women not using hormone
General introduction 23 therapy. In men, Leibenluft et al. (1997) found that testosterone suppression resulted in shorter stage N4 sleep and a shorter REM sleep latency compared to testosterone suppression combined with add-back testosterone. However, further experimental studies assessing effects of sex hormones on objective sleep are scarce. 2.3. Sleep in relation to depression Sleep and depression are closely linked to each other: insomnia is a risk factor for depression, and it is also a common symptom during and after depressive episodes (Baglioni et al., 2011), although depressed persons also tend to underestimate their total sleep duration (Difrancesco et al., 2019). Objective sleep measures using sleep EEG also show alterations in depressed persons: SWS is often shorter and more fragmented, and REM sleep is longer and takes place earlier in the night (Steiger & Kimura, 2010). Persons in a depressive episode are more likely to have a later chronotype than persons who are not depressed (Antypa et al., 2016). 3. Exogenous sex hormone use, depression and sleep 3.1. Oral contraceptives and depression The use of OCs could contribute to depressive symptoms, although there might be individual differences in sensitivity for effects of OC (Oinonen & Mazmanian, 2002). Studies found indications of increased risk of depression in younger OC users (De Wit et al., 2020; Skovlund et al., 2016), people who previously had mood-related side effects using OC (Gingnell et al., 2013) and in new users (Johansson et al., 2023). However, there is no conclusive evidence indicating whether depression risk changes after starting OC use, and empirical studies show mixed results. These mixed results could be explained by differences in study samples and setups. One of the likely biases in the study samples in OC studies is a healthy user bias, also sometimes called the survivor bias. The healthy user bias is based on the assumption that side effects of OCs could lead to discontinuation of OC use, resulting in an underestimation of mood effects of OCs in long-time users (Westhoff et al., 2007). Therefore, it is important in
Chapter 1 24 OC research to distinguish prospective from cross-sectional studies and to assess whether studies include new users or long-term users. It is not clear whether women who are known with previous depressions are more at risk for mood-related side effects during OC use, since most studies on OCs and mood have excluded participants with a history of psychiatric illnesses. In reproductive-age women, 24% to 32% have or have had a depressive disorder (Ten Have et al., 2023), but it is still largely unknown whether they are at increased risk of experiencing mood-related side effects during OC use. 3.2. Gender-affirming hormone therapy and depression Transgender persons show a high prevalence of depression, especially transgender persons who desire to use GAHT but who have not yet started GAHT use: estimates show between 42% and 48% of transgender persons reports clinically significant depressive symptoms at the start of GAHT use (Aldridge et al., 2021; Colizzi et al., 2014). Studies on GAHT and depression generally show improvements in depressive symptoms after GAHT use, although findings also differ between studies (Doyle et al., 2023). Despite significant work in this topic, knowledge of changes in specific depression symptomatology (e.g. prevalence of specific symptoms or groups of symptoms) after the start of GAHT use is still lacking. 3.3. Oral contraceptives and sleep Studies on OC use and sleep are scarce, and findings on the effects of oral contraceptives and sleep are mixed. In the domain of subjective sleep, Bezerra et al. (2020) found that OC users reported more insomnia symptoms and a prospective pre-post study reported that participants reported poorer sleep after starting OC use (Albuquerque et al., 2015), but a meta-analysis assessing sleep quality, summarizing four studies, did not find significant differences between OC users and non-users (Bezerra et al., 2023). Individual studies also found effects of oral contraceptive use on sleep architecture, but the aforementioned meta-analysis found no consistently significant differences in seven studies, most of which were cross-sectional (Bezerra et al., 2023). Only one study examined OCs and chronotype and found no associations between OCs and chronotype (Toffol et al., 2013).
General introduction 25 3.4. Gender-affirming hormones and sleep Studies on GAHT and sleep are very scarce. Although the prevalence of poor sleep in transgender persons was estimated to be 80% (Auer et al., 2017), the effect of GAHT on insomnia, sleep quality or chronotype has not yet been studied. One study investigated the effects of three months of GAHT on sleep EEG outcomes, but they only examined effects of feminizing GAHT in seven transgender women (Kunzel et al., 2011). The authors found an increase in light sleep, and no other changes in sleep EEG in participants. The small study sample size and lack of participants using masculinizing GAHT largely limited the generalizability of this study. 4. This thesis Aims This thesis adds to the current literature by focusing on research gaps (as described in Box 1.1) in the relation between sex hormones, sleep and depression. We focused on three specific research aims. First, we aimed to examine the effects of OC use and GAHT use on depression, including depression subtypes and symptomatology. Second, we aimed to examine the effects of OC use and GAHT use have on subjective sleep, sleep architecture and chronotype. Third, we aimed to examine the effects of sex hormones on the association between depression and sleep. Hypotheses We hypothesized that the use of OCs would result in changes in depression and sleep in line with findings in sex hormone intervention studies, and that the use of GAHT would results in changes in depression or sleep in line with the cisgender sex differences in the population. We firstly hypothesized that use of female hormones, including OC and feminizing GAHT, would be associated with increases in depressive symptoms and specifically with increased atypical depression symptoms. Vice versa, we hypothesized that masculinizing hormone use would be associated with reduced depressive symptoms and specifically with reduced symptoms of atypical depression.
Chapter 1 26 Box 1.1 Summary of research gaps in the current literature on depression and sleep. Oral contraceptives (OCs) Most studies on OC use and depression have used cross-sectional data or naturalistic prospective data, which increases the likelihood of healthy user bias in the sample. Furthermore, the majority of studies have excluded participants with a lifetime history of psychiatric diagnoses, meaning these current studies cannot be generalized to those with a previous or current psychiatric disorder. There are numerous studies on OC use and sleep, but most studies used retrospective setups and were limited by small sample sizes. Furthermore, the relationship between OC-associated cortisol changes and sleep have not yet been addressed. Gender-affirming hormone therapy (GAHT) The topic of depression in transgender hormone users has been studied, but most studies used clinical diagnoses or total depressive symptoms to address depressive symptoms, and further knowledge on symptom changes after GAHT use is still lacking, as is knowledge on changes in depressive symptoms after masculinizing compared to feminizing GAHT use. Numerous previous studies also used data obtained from medical files, which could be biased by participants providing socially desirable answers to their healthcare providers in order to gain or keep access to gender-affirming care. Other studies compared depressive symptom severity when participants were still waitlisted to depressive symptom severity after years of GAHT, which means that the outcomes could be confounded by a multitude of psychosocial changes. Studies on sleep and GAHT thus far are still very scarce. Secondly, we hypothesized that the use of OCs and feminizing GAHT would be associated with increases in poor sleep quality and insomnia, whereas use of masculinizing GAHT would be associated with decreases in poor sleep quality and insomnia. However, based on the existing cisgender paradox in sleep, we also hypothesized that use of feminizing GAHT would be associated with better objective sleep, including longer TST, shorter SOL, shorter WASO, longer SWS and longer REM sleep latency, and with an earlier chronotype. We also hypothesized that use of masculinizing hormones would be associated with poorer objective sleep, including shorter TST, longer SOL, longer WASO, shorter SWS and shorter REM sleep latency, as well as a later chronotype.
General introduction 27 Box 1.2: Study cohorts in this thesis. To investigate the research questions, we used data from different cohorts and study populations. The distribution of the cohorts in the different chapters of this thesis are also described in Table 1.1. Oral contraceptives OC use was studied in data from the Netherlands Study of Depression and Anxiety (NESDA) and in data from the Center for Integrated Molecular Brain Imaging (Cimbi) database. NESDA is a multi-centre longitudinal naturalistic cohort study with 2981 participants, who were followed up for 9 years in five follow-up waves. It aims to describe the longterm course and consequences of depression and anxiety and to study possible biological and psychosocial predictors of depression and anxiety. The NESDA study conducted interviews, questionnaires, medical examinations and collected demographic, psychosocial, clinical and biological data (Penninx et al., 2008). The Cimbi database is a database containing study data from over 2000 healthy volunteers and from patients with psychiatric or somatic illnesses participating in studies conducted within the Lundbeck Foundation Center for Integrated Molecular Brain Imaging (Cimbi) in Copenhagen, Denmark. Study data from neuroimaging and psychiatry studies from 2008 onwards were saved in a structured and standardized manner, in order to enable researchers to conduct secondary analyses on previously collected data. The database currently holds data from over 2000 participants, both healthy volunteers and patients, including neuroimaging data, genetic and biochemical data, questionnaire data and data from neuropsychological tests (Knudsen et al., 2016). Gender-affirming hormone therapy Effects of GAHT were studied in the European Network for the Investigation of Gender Incongruence (ENIGI) study and the Relationship between Emotions and Sleep in Transgender persons: Endocrinology and Depression (RESTED) study. The ENIGI study is a multi-centre observational prospective study, with participating centres from Ghent (Belgium), Oslo (Norway), Florence (Italy), Tel Aviv (Israel) and Amsterdam (the Netherlands). Within the ENIGI study, questionnaires on sleep quality, insomnia and depression were collected at the start of GAHT and after three, six, nine and twelve months of GAHT, depending on the study outcome, and data on comorbidity, medication use, hormone dosage and forms and serum hormone levels were collected from medical patient files (Dekker et al., 2016). The RESTED study is a multi-centre prospective cohort study with the aim of studying the effects of GAHT on sleep and depression, with participating centres from Amsterdam (the Netherlands) and Groningen (the Netherlands). The RESTED study assessed depressive symptom severity, sleep quality and insomnia, sleep EEG and chronotype before starting GAHT, after three months of GAHT and after twelve months of GAHT. These data were combined with data on comorbidity, medication use, hormone dosage and forms and serum hormone levels from medical patient files.
Chapter 1 28 Thirdly, based on the hypothesis that poorer sleep quality is associated with more depressive symptoms, we hypothesized that in those undergoing hormonal changes, unfavourable changes in sleep (i.e., more insomnia symptoms, poorer sleep quality, less SWS, more WASO and later chronotype) would be associated with an increase in depressive symptoms.
General introduction 29 Thesis outline Section 1: Are OC use and GAHT use associated with changes in depressive symptoms? In this first section, we studied the relationship between OC use and GAHT use and depressive symptoms. In Chapter 2, we investigated depressive symptoms and mood disorders in a subgroup of participants from the NESDA study. We grouped all measurements from reproductive-age women who were either naturally cycling or using OCs, and assessed differences in depressive symptoms, atypical depression symptoms and mood disorder diagnoses between the groups. Due to the prospective data collection setup, we also assessed within-participant and between-participant associations with OC use. The assessment of within- and between-participant effects provided new insights into population effects and the healthy user bias in studies on OC use. In Chapter 3, we assessed depressive symptoms before and after 3 and 12 months of GAHT in the RESTED and ENIGI studies. We conducted an exploratory factor analysis to assess which depressive symptoms formed symptom clusters. Then, we assessed changes in depressive symptom clusters after 3 and 12 months of GAHT use, to determine the effect of GAHT on depressive symptomatology. Section 2: Are OC use and GAHT use associated with changes in sleep? In Chapter 2, we investigated insomnia symptoms in the aforementioned subgroup of participants from the NESDA study who were either naturally cycling or using OCs, and we assessed differences in insomnia symptoms between the groups. In Chapter 4, we used data from the Cimbi database to assess subjective sleep disturbances, sleep quality and their association with cortisol dynamics in naturally cycling women, oral contraceptive users and men. We tested whether oral contraceptive users and men were less or more likely to report sleep disturbances compared to naturally cycling women, and we assessed differences in sleep quality. Additionally, we tested whether the associations between the cortisol awakening response and sleep disturbances or sleep quality differed between the groups. In Chapter 5, we used data from the ENIGI study to assess changes in subjective sleep, including sleep quality and insomnia symptoms, in transgender persons during the first year of GAHT use. In Chapter 6, we used sleep EEG measurements from the RESTED study to assess whether sleep EEG, including total sleep duration, sleep onset latency, wake after sleep onset,
Chapter 1 30 SWS duration, REM sleep latency and REM sleep duration changed after 3 months of GAHT. In Chapter 7, we used data from the RESTED study to assess whether participants’ chronotype changed after 3 months of masculinizing or feminizing GAHT. Section 3: What are the effects of sex hormones on the association between depression and sleep? In Chapter 2, we also examined whether having a previous or current mood disorder moderated the group differences in insomnia between participants who were naturally cycling or who were using OCs. In Chapter 3, we examined correlations between depressive symptoms and insomnia in an exploratory factor analysis, and we prospectively examined whether the resulting symptom clusters changed after starting GAHT. In Chapter 8, we conducted a systematic review on studies assessing the relationship between sex hormones, depression and sleep. We selected studies that measured endogenous hormones or studied effects of using exogenous hormones and that measured both depression and sleep, to assess whether sex hormones could change the association between depression and sleep. Table 1.1. Description of study cohorts per chapter. Cohort Participant population Hormone intervention Comparison group(s) Used in chapter NESDA Women with previous or current depression and healthy controls Oral contraceptive users Naturally cycling women 2: Effects of oral contraceptives on depression and sleep Cimbi Healthy control participants Oral contraceptive users Naturally cycling women and men 4: Effects of oral contraceptives on sleep and cortisol dynamics ENIGI Transgender participants Gender-affirming hormones Prospective setup: baseline vs. 3 months and 12 months of GAHT 3: Effects of GAHT on depressive symptom profile 5: Effects of GAHT on insomnia and sleep quality RESTED Transgender participants Gender-affirming hormones Prospective setup: baseline vs. 3 months and 12 months of GAHT 3: Effects of GAHT on depressive symptom profile 6: Effects of GAHT on sleep architecture 7: Effects of GAHT on chronotype
General introduction 31 References Albuquerque, R. G., da Rocha, M. A. D., Hirotsu, C., Hachul, H., Bagatin, E., Tufik, S., & Andersen, M. L. (2015). A randomized comparative trial of a combined oral contraceptive and azelaic acid to assess their effect on sleep quality in adult female acne patients. Archives of Dermatological Research, 307(10), 905-915. https://doi.org/10.1007/s00403-015-1600-0 Aldridge, Z., Patel, S., Guo, B., Nixon, E., Pierre Bouman, W., Witcomb, G. L., & Arcelus, J. (2021). Long-term effect of gender-affirming hormone treatment on depression and anxiety symptoms in transgender people: A prospective cohort study. Andrology, 9(6), 1808-1816. https://doi.org/10.1111/andr.12884 Aleknaviciute, J., Tulen, J. H. M., De Rijke, Y. B., Bouwkamp, C. G., van der Kroeg, M., Timmermans, M., Wester, V. L., Bergink, V., Hoogendijk, W. J. G., Tiemeier, H., van Rossum, E. F. C., Kooiman, C. G., & Kushner, S. A. (2017). The levonorgestrel-releasing intrauterine device potentiates stress reactivity. Psychoneuroendocrinology, 80, 39-45. https://doi.org/10.1016/j.psyneuen.2017.02.025 Altemus, M., Sarvaiya, N., & Neill Epperson, C. (2014). Sex differences in anxiety and depression clinical perspectives. Frontiers in Neuroendocrinology, 35(3), 320-330. https://doi.org/ 10.1016/j.yfrne.2014.05.004 American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Antypa, N., Vogelzangs, N., Meesters, Y., Schoevers, R., & Penninx, B. W. J. H. (2016). Chronotype Associations with Depression and Anxiety Disorders in a Large Cohort Study. Depression and Anxiety, 33(1), 75-83. https://doi.org/10.1002/da.22422 Auer, M. K., Liedl, A., Fuss, J., Nieder, T., Briken, P., Stalla, G. K., Hildebrandt, T., Biedermann, S. V., & Sievers, C. (2017). High impact of sleeping problems on quality of life in transgender individuals: A cross-sectional multicenter study. PloS One, 12(2), e0171640. https://doi.org/10.1371/journal.pone.0171640 Baglioni, C., Battagliese, G., Feige, B., Spiegelhalder, K., Nissen, C., Voderholzer, U., Lombardo, C., & Riemann, D. (2011). Insomnia as a predictor of depression: A meta-analytic evaluation of longitudinal epidemiological studies. Journal of Affective Disorders, 135(1-3), 10-19. https://doi.org/10.1016/j.jad.2011.01.011 Ben Dor, R., Harsh, V. L., Fortinsky, P., Koziol, D. E., Rubinow, D. R., & Schmidt, P. J. (2013). Effects of Pharmacologically Induced Hypogonadism on Mood and Behavior in Healthy Young Women. American Journal of Psychiatry, 170(4), 426-433. https://doi.org/10.1176/ appi.ajp.2012.12010117 Bezerra, A. G., Andersen, M. L., Pires, G. N., Banzoli, C. V., Polesel, D. N., Tufik, S., & Hachul, H. (2020). Hormonal contraceptive use and subjective sleep reports in women: An online survey. Journal of Sleep Research, 29(6), e12983. https://doi.org/10.1111/jsr.12983 Bezerra, A. G., Pires, G. N., Andersen, M. L., Tufik, S., & Hachul, H. (2023). The Effects of Hormonal Contraceptives on the Sleep of Women of Reproductive Age. Sleep Medicine Clinics, 0(0). https://doi.org/10.1016/j.jsmc.2023.06.005 Bhasin, S., & Seidman, S. (2019). Testosterone Treatment of Depressive Disorders in Men: Too Much Smoke, Not Enough High-Quality Evidence. JAMA Psychiatry, 76(1), 9-10. https://doi.org/10.1001/jamapsychiatry.2018.2661 Bianchi, M. T., Williams, K. L., Mckinney, S., & Ellenbogen, J. M. (2013). The subjective–objective mismatch in sleep perception among those with insomnia and sleep apnea. Journal of Sleep Research, 22(5), 557-568. https://doi.org/10.1111/jsr.12046 Bixler, E. O., Papaliaga, M. N., Vgontzas, A. N., Lin, H.-M., Pejovic, S., Karataraki, M., Vela-Bueno, A., & Chrousos, G. P. (2009). Women sleep objectively better than men and the sleep of young women is more resilient to external stressors: Effects of age and menopause. Journal of Sleep Research, 18(2), 221-228. https://doi.org/10.1111/j.1365-2869.2008.00713.x
Chapter 1 32 Boisseau, N., Enea, C., Diaz, V., Dugué, B., Corcuff, J. B., & Duclos, M. (2013). Oral contraception but not menstrual cycle phase is associated with increased free cortisol levels and low hypothalamo-pituitary-adrenal axis reactivity. Journal of Endocrinological Investigation, 36(11), 955-964. https://doi.org/10.3275/8971 Bungum, L., Franssohn, F., Bungum, M., Humaidan, P., & Giwercman, A. (2013). The Circadian Variation in Anti-Müllerian Hormone in Patients with Polycystic Ovary Syndrome Differs Significantly from Normally Ovulating Women. PLOS ONE, 8(9), e68223. https://doi.org/10.1371/journal.pone.0068223 Burger, H. G., Dudley, E. C., Cui, J., Dennerstein, L., & Hopper, J. L. (2000). A prospective longitudinal study of serum testosterone, dehydroepiandrosterone sulfate, and sex hormone-binding globulin levels through the menopause transition. The Journal of Clinical Endocrinology and Metabolism, 85(8), 2832-2838. https://doi.org/10.1210/jcem.85.8.6740 Burkman, R. T. (2001). Oral Contraceptives: Current Status. Clinical Obstetrics and Gynecology, 44(1), 62. Campagnoli, C., Colombo, P., De Aloysio, D., Gambacciani, M., Grazioli, I., Nappi, C., Serra, G. B., & Genazzani, A. R. (2002). Positive effects on cardiovascular and breast metabolic markers of oral estradiol and dydrogesterone in comparison with transdermal estradiol and norethisterone acetate. Maturitas, 41(4), 299-311. https://doi.org/10.1016/S03785122(01)00300-0 Ciano, C., King, T. S., Wright, R. R., Perlis, M., & Sawyer, A. M. (2017). Longitudinal Study of Insomnia Symptoms Among Women During Perimenopause. Journal of Obstetric, Gynecologic, and Neonatal Nursing: JOGNN, 46(6), 804-813. https://doi.org/10.1016/ j.jogn.2017.07.011 Colizzi, M., Costa, R., & Todarello, O. (2014). Transsexual patients’ psychiatric comorbidity and positive effect of cross-sex hormonal treatment on mental health: Results from a longitudinal study. Psychoneuroendocrinology, 39, 65-73. https://doi.org/10.1016/ j.psyneuen.2013.09.029 D’Andrea, S., Spaggiari, G., Barbonetti, A., & Santi, D. (2020). Endogenous transient doping: Physical exercise acutely increases testosterone levels—results from a meta-analysis. Journal of Endocrinological Investigation, 43(10), 1349-1371. https://doi.org/10.1007/s40618020-01251-3 Davison, S. L., Bell, R., Donath, S., Montalto, J. G., & Davis, S. R. (2005). Androgen levels in adult females: Changes with age, menopause, and oophorectomy. The Journal of Clinical Endocrinology and Metabolism, 90(7), 3847-3853. https://doi.org/10.1210/jc.2005-0212 De Graaf, H., Oldenhof, A., Kraan, Y., Beek, T., Kuipers, L., & Vermey, K. (2023). Seks onder je 25ste. https://rutgers.nl/onderzoeken/seks-onder-je-25e-2023/ de Graaf, R., ten Have, M., & van Dorsselaer, S. (2010). The Netherlands Mental Health Survey and Incidence Study-2 (NEMESIS-2): Design and methods. International Journal of Methods in Psychiatric Research, 19(3), 125-141. https://doi.org/10.1002/mpr.317 de Kruif, M., Spijker, A. T., & Molendijk, M. L. (2016). Depression during the perimenopause: A meta-analysis. Journal of Affective Disorders, 206, 174-180. https://doi.org/10.1016/ j.jad.2016.07.040 De Wit, A. E., Booij, S. H., Giltay, E. J., Joffe, H., Schoevers, R. A., & Oldehinkel, A. J. (2020). Association of Use of Oral Contraceptives with Depressive Symptoms among Adolescents and Young Women. JAMA Psychiatry, 77(1), 52-59. https://doi.org/10.1001/ jamapsychiatry.2019.2838 de Wit, A. E., Giltay, E. J., de Boer, M. K., Bosker, F. J., Cohn, A. Y., Nolen, W. A., Kaiser, U. B., Joffe, H., Penninx, B. W. J. H., & Schoevers, R. A. (2021). Plasma androgens and the presence and course of depression in a large cohort of women. Translational Psychiatry, 11(1), Article 1. https://doi.org/10.1038/s41398-021-01249-2 de Wit, A. E., Giltay, E. J., de Boer, M. K., Nolen, W. A., Bosker, F. J., Penninx, B. W. J. H., & Schoevers, R. A. (2021). Plasma androgens and the presence and course of depression in a large cohort of men. Psychoneuroendocrinology, 130, 105278. https://doi.org/10.1016/ j.psyneuen.2021.105278
www.ridderprint.nlRkJQdWJsaXNoZXIy MTk4NDMw