Chronic Psychosocial Stress and Levels of Ovarian Hormones in
Naturally Cycling Women
D Beem S2455331
Clinical Psychology Master
Supervisor: dr hab. A Ziomkiewicz Institute of Psychology/Social Sciences Universiteit Leiden
Table of Contents
1 Abstract ... 4
2 Introduction ... 4
2.1 Types of stressors ... 5
2.2 The Hypothalamic-Pituitary-Gonadal Axis ... 6
2.3 Effects of cortisol on reproduction during stress ... 6
2.4 Negative feedback loop of cortisol... 7
2.5 HPA dysregulation – hypercortisolism & hypocortisolism ... 7
2.6 Reproductive outcomes of hyper- & hypocortisolism ... 8
2.7 Acute stress & cross-talk between HPA & HPG ... 8
2.8 Chronic stress & cross-talk between HPA & HPG ... 8
2.9 Gonadotropin-Inhibitory Hormone ... 9
2.10 The under researched association between chronic stress and hormone levels ... 9
2.11 Hypothesis: Association between chronic stress, cortisol and hormone levels ... 10
2.12 Hypothesis: Moderation of repetition of chronic stressors on stress-hormone association ... 10
3 Method... 10 3.1 Research Design ... 10 3.2 Study Group ... 10 3.3 Procedure ... 11 3.4 Sample Collection ... 11 3.5 Hormone Analysis ... 11
3.6 Perceived Stress Assessment ... 12
3.7 Repeated Stressor Assessment ... 12
4 Data-analysis ... 13 4.1 Data-cleaning ... 13 4.2 Confounding variables ... 13 4.3 Statistical-analysis ... 14 4.4 Simple ANOVA’s ... 14 4.5 Two-way ANCOVA ... 14
4.6 Multiple Regression Models ... 14
4.7 Moderation analysis ... 15
5 Results ... 15
5.1 Sample Demographics ... 15
5.3 Little’s MCAR test ... 16
5.4 Group Comparisons ... 18
5.5 Chronic Stressors ... 18
5.6 Cycle Cortisol ... 18
5.7 Luteal Cortisol ... 18
5.8 Differences in estradiol between chronic stress and cortisol groups ... 19
5.9 Predictive value of chronic psychosocial stress and cortisol for cycle estradiol ... 20
5.10 Predictive value of chronic psychosocial stress and cortisol for luteal progesterone ... 21
5.11 Moderation of repeated stressors on chronic stress and estradiol ... 21
5.12 Moderation of repeated stressors on chronic stress and progesterone ... 21
5.13 Moderation of repeated stressors on cortisol and estradiol ... 21
5.14 Moderation of repeated stressors on luteal cortisol and luteal progesterone ... 21
6 Discussion ... 22
6.1 Chronic stressors and reproductive physiology ... 22
6.2 Possible mechanisms ... 23
6.3 Cortisol levels and ovarian hormone levels ... 25
6.4 Study limitations ... 26
6.5 Conclusions ... 27
Abstract
Chronic stress is defined as ongoing social andenvironmental events requiring an individual to adjust through physical and psychological reactions to maintain homeostasis. The HPA axis exerts inhibiting influences on the HPG axis, attenuating its secretion of hormones. Furthermore, prolonged HPA axis overactivation can lead to hypocortisolism, which is observed in chronic stress. The aim of this study is to find out if the repetitiveness of stressors strengthens the relationship between chronic stress, cortisol, and ovarian hormone levels in naturally cycling women. The Recent Life Changes Questionnaire measured stressors and competitive enzyme immunoassay tested estradiol as well as progesterone metabolites during entire menstrual cycle. Study participants were 106 healthy Polish women. A two-way ANCOVA, linear regression and hierarchical multiple regression analysis were employed for data-analysis. Perceived chronic stress was not associated with nor predicted ovarian hormone metabolite levels. An increase in cortisol levels predicted a rise in estradiol and progesterone metabolites during the menstrual cycle. Repeated stressors did not moderate the association between perceived chronic stress or cortisol and ovarian hormones. This could be due to the operationalization of chronic stressors as episodic stressors, which were demonstrated to have limited association with health outcomes. Alternatively, habituation or hormonal desensitization might have occurred, which can manifest in a lack or diminished physiological response to stressors. The positive predictive effect of cortisol on estradiol and progesterone could be explained by hypocortisolism and desensitization. It is recommended for future research to assess prospective effects of different stressors on cortisol and ovarian hormone levels.
Keywords: ovarian hormones, perceived psychosocial stress, chronic stress, natural cycle
Stress is a common problem in today’s society. An example of a widely experienced stressor is chronic psychosocial stress, which is defined by ongoing social, environmental events that require an individual to adjust through physical and psychological reactions to maintain homeostasis (Gaillard, 2006). The Reproductive Suppression model was introduced by Wasser and Isenberg (1986) and states that reproduction might be inhibited during adverse environmental conditions, such as stress. Chronic stress is thus known to have adverse effects on reproductive functioning.
A clinical reproductive syndrome suggested to be influenced by stress is unexplained infertility. This describes couples who are failing to conceive despite the lack of an
established medical cause for infertility (Wallach & Moghissi, 1983). Research has found that couples experiencing unexplained infertility are more stressed than fertile couples and that the less stress women experience, the faster they succeed in conceiving (Harrison, O’Moore, & O’Moore, 1986; Hjollund et al., 1999). Furthermore, couples experiencing infertility who receive treatment for stress issues conceive faster than controls (De Liz & Strauss, 2005). Due to the fact that 15% of women during reproductive years are estimated to be affected by unexplained infertility, it is of high importance to find out more about the role stress has in reproductive outcomes (Morse & Van Hall, 1987).
Types of stressors
In the psychological literature numerous different types of stressors were described, of which chronic psychosocial stressors are the sole focus of this study. A chronic stressor can be differentiated from an acute stressor. While an acute stressor occurs suddenly and
terminates quickly, a long-term stressor is ongoing for an extended period. An example of a chronic stressor is an episodic stressor which can be described as a long-lasting stressor that terminates. Examples of this can be a divorce or moving house (Krohne, 2002). Pacák and Palkovits (2001) found that different types of stressors lead to different neuroendocrine responses. The precise physiological responses to stressor types, or neuroendocrine signatures, might exist to minimize the harmful effects of the stress response. It has been found that processive (emotional) stressors do not stimulate a direct hypothalamic-pituitary-adrenal axis (HPA axis) activation. Instead, they first travel through the limbic forebrain sites to be identified as a stressor. If identified as a stressor, the paraventricular nucleus is activated (Herman & Cullinan, 1997). Overall, it may be said that different types of stressors can evoke
different physical outcomes, and that an emotional stressor needs to be identified as a stressor to elicit HPA axis activation.
The Hypothalamic-Pituitary Axis
A physiological response to a stressor (psychological or physical) is generated in the hypothalamic-pituitary-adrenal axis, which consists of the centers located in the
hypothalamus, pituitary, and adrenal gland (Mastorakos, Pavlatou, & Mizamtsidi, 2006). In a stressful situation, the paraventricular nucleus of the hypothalamus starts producing
corticotropin-releasing hormone (CRH), which is released into the hypophysial portal vessels through the infundibulum. At the same time, arginine vasopressin (AVP) is secreted. CRH and AVP stimulate the anterior lobe of the pituitary gland to produce and release
adrenocorticotropic hormone (ACTH) (Stephens & Wand, 2012). ACTH reaches the adrenal cortex and stimulates the production and secretion of glucocorticoids (cortisol). The purpose of cortisol secretion is to mobilize the energy necessary to deal effectively with an
experienced environmental stressor. This process is adaptive and helps to maintain
homeostasis throughout changing ecological conditions (Oyola, & Handa, 2017). However, when chronic, this process may have harmful consequences on several bodily functions, including reproduction (Smith & Vale, 2006).
The Hypothalamic-Pituitary-Gonadal Axis
The hypothalamic-pituitary-gonadal axis (HPG) is responsible for controlling and regulating reproductive function in the body. The HPG works as follows: the hypothalamus produces gonadotropin-releasing hormone (GnRH), which is a neuropeptide that directly influences the reproductive axis by promoting the release of pituitary gonadotropins (Tsutsui, Ubuka, Bentley, & Kriegsfeld, 2012). GnRH travels to the pituitary glands, which produce luteinizing hormone (LH) and follicle-stimulating hormone (FSH). As FSH and LH reach the gonads, in case of women the ovaries, they stimulate growth and development of the ovarian follicles, which then produce estradiol. LH stimulates the ovulation in the developed follicle and the production of progesterone from the corpus luteum. As a result, FSH leads to a rise in estradiol during the follicular phase and LH to a rise in progesterone secretion beginning at ovulation (Reed & Carr, 2018).
Effects of cortisol on reproduction during stress
Cortisol has been found to inhibit reproductive function in laboratory animals and in humans (Sapolsky, Romero & Allan, 2000). Cortisol inhibits the release of GnRH from the hypothalamus in castrated male rats (Belhadj, De Besi, Bardin, & Thau, 1989) and enhances
the release of gonadotropin inhibiting hormone (Kirby, Geraghty, Ubuka, Bentley & Kaufer, 2009). Further, it inhibits the basal and GnRH-stimulated release of LH from the pituitary and enhances the release of FSH (Suter & Schwartz, 1985). Cortisol has been found to suppress the responsivity of the pituitary to GnRH, which leads to a lowered secretion of LH (Hagino, Watanabe, & Goldzieher, 1969). Last, in terms of human reproductive function a study has showed that increased cortisol levels promote amenorrhea (Loucks, Mortola, Girton, & Yen, 1989). To conclude, cortisol has a prominent role in the inhibition of reproductive function during a stress response.
Negative feedback loop of cortisol
Cortisol receptors are widely spread throughout the body to enable a broad variety of hormone related responses. Although, a release of cortisol during a stress response is adaptive and promotes survival, the long-term effects of cortisol on the body are adverse. Therefore, there are feedback mechanisms in place, for example the negative feedback loop of cortisol in the central nervous system. Part of the negative feedback loop is the binding of cortisol to glucocorticoid receptors in the amygdala, hippocampus, hypothalamus, pituitary, and adrenals (Conrad, Hubold, Fischer, & Peters, 2009). Through the binding to the receptors in those regions, their glucocorticoid receptors are down-regulated, less ACTH is released and less cortisol is secreted (Marin, Martin, Blackwell, Stetler, & Miller, 2007). It can consequently be stated that a precise deactivation of the stress response is just as important as a precise
activation in survival (Pacak & Palkovits, 2001).
HPA dysregulation – hypercortisolism & hypocortisolism
During the experience of long-term chronic stress, the HPA is activated excessively and as time progresses the negative feedback loop of cortisol becomes impaired. The
consequence can be a dysregulation of the HPA axis (Loucks, Mortola, Girton & Yen, 1989). An indicator of HPA dysregulation is a lowered or heightened cortisol awakening response (Björntorp & Rosmond, 2000). The cortisol awakening response is the peak in cortisol immediately after waking in the morning (Pruessner et al., 2003). During normal regulatory functioning of the HPA the cortisol production should be attenuated in response to
glucocorticoids (Rodriguez et al., 2015). During dysregulation the HPA can become desensitized to the glucocorticoid feedback and thus cortisol release is not attenuated. As a result of that cortisol levels become elevated, which is called hypercortisolism. Over a
timeframe, which varies individually, the HPA axis adapts to the elevated cortisol levels. As a result, cortisol release is reduced, which leads to low cortisol levels, called hypocortisolism. During hypocortisolism the HPA axis is not triggered by awaking and thus the CAR is
blunted and cortisol remains low throughout the day (Guilliams, & Edwards, 2010). In conclusion, HPA dysregulation can lead to hyper- and hypocortisolism.
Reproductive outcomes of hyper- & hypocortisolism
In their review, Heim, Ehlert and Hellhammer (2000) have concluded that
hypocortisolism occurs in healthy individuals experiencing chronic, intermittent stress. An example of this is the study of Caplan et al. (1979), which has found decreased cortisol levels in people experiencing prolonged increased workload. Chronic stress can lead to various adverse effects on the female reproductive system. For example, failure of ovulation was observed in macaca fascicularis (Adams, Kaplan, Koritnik, 1985) as well as in college students (Bachman & Kemmann, 1982). To illustrate, Owens and Nemeroff (1991) have concluded in their review that increased cortisol levels, as seen in hypercortisolism are connected to attenuated reproductive behavior in animals. Therefore, it is known
observationally and experimentally that long-term stress exposure is associated with hyper- and hypocortisolism and adverse reproductive outcomes.
Acute stress & cross-talk between HPA & HPG
Reproductive function in females depends on the availability of environmental energy resources since reproduction is highly energy consuming. Physical and psychological
stressors might signal a decrease in the availability of these resources. As a result, the existing cross-talk between the HPA axis and the hypothalamic-pituitary-gonadal axis (HPG) allows for adequate responding to environmental stressors. An example of this cross-talk is the protective effect of estradiol during acute stress responses. Estradiol attenuates the activation of the HPA axis and thus the autonomic response to stress (Kajantie & Phillips, 2006). This is supported by the fact that the HPA axis has estradiol receptors (Mastorakos, Pavlatou, & Mizamtsidi, 2006). In accordance to this, a study by Albert, Pruessner, and Newhouse (2015) showed that women with higher estradiol levels have lower levels of perceived stress, and vice versa. In conclusion, there is a bidirectional relationship between hormone levels and perceived stress.
Chronic stress & cross-talk between HPA & HPG
The effect of hormones on the HPA during stress change with the type of stressor experienced. During incidents of chronic stress, the HPG is inhibited by the HPA to limit higher energy investment in reproductive function (pregnancy and breastfeeding). The purpose of this is to save energy used for reproduction as well as immune functions, while they are needed for survival (Toufexis, Rivarola, Lara & Viau, 2014). In line with the
inhibition of the HPG axis, it has been found that animals experiencing prolonged stress, have lower levels of progesterone (O'Connor et al., 2011). It has been found that in infertile
patients there is an association between low progesterone levels and decreased ovulation (Sanchez, Giviziez, Sanchez, Agostinho, Barros, & Approbato, 2015). From this it can be concluded that changes to the HPG axis under stress are dependent on type, duration and frequency of the stressor.
Gonadotropin-Inhibitory Hormone
The increased production of cortisol that accompanies a stress reaction leads to the secretion of gonadotropin-inhibitory hormone (GnIH), which is a hypothalamic neuropeptide synthesized in the brain as well as the gonads. GnIH is the hypothalamic antagonist of GnRH. While GnRH is responsible for the activation of the reproductive axis through the promotion of the release of pituitary gonadotropins, GnIH deactivates it by inhibiting the synthesis and release of GnRH (Tsutsui K, et al., 2000).This process leads to a temporary inhibition of reproduction in response to unpredictable environmental cues, such as stress. Another way that GnIH regulates reproductive function is through receptors in the gonadal cells. This is a direct causal pathway to the inhibition of steroidogenesis (Tsutsui, Ubuka, Bentley, & Kriegsfeld, 2012). Results in animal studies have found that during metabolic stress, GnIH directly increases in the ovaries of European starlings (McGuire, Koh & Bentley, 2013). The inhibitory action of GnIH takes place in a twofold sequence: firstly, the release of
gonadotropins is reduced, followed by the synthesis of LH and FSH being inhibited. It has been found in male rats, that during periods of chronic stress, there is an upregulation of GnIH. The possible mechanism behind that is glucocorticoid receptors in GnIH cells. The upregulation could lead to lower levels of reproductive hormones after prolonged stress exposure (Kirby, Geraghty, Ubuka, Bentley & Kaufer, 2009).
The under researched association between chronic stress and hormone levels In conclusion, there is a bidirectional association between the stress response and hormone levels. Acute stress is associated with a short-term increase in LH and estradiol production and chronic stress with inhibited secretion of LH and sex steroids. Moreover, it can be concluded that healthy individuals experiencing prolonged stressors can suffer from hypocortisolism after prolonged exaggerated cortisol levels. While a significant number of studies explored the effects of acute stress on ovarian hormone levels, research on the effect of chronic stress remains limited. The aim of the planned study is to fill this knowledge gap. To achieve this, the present study aims to answer the following questions 1) what is the association between female ovarian reproductive hormone levels and chronic stress? And 2)
does the repetition of the stressors moderate the relationship between female ovarian reproductive hormone levels and perceived psychosocial stress?
Hypothesis: Association between chronic stress, cortisol and hormone levels
When women experience more chronic stressors, they are expected to have lower cortisol levels due to possible dysregulated HPA axis function. Additionally, it is
hypothesized that women who experience more chronic stress have lower levels of estradiol during the menstrual cycle. That is based on the inhibitory effects of the HPA axis on the HPG axis during stressful conditions and the inhibiting effects of GnIH on steroidogenesis. It is also hypothesized that women who experience more chronic stress have lower levels of luteal progesterone. Therefore, it is expected that the more chronic psychological stress the women have experienced, the lower their level of cortisol, estradiol and luteal progesterone. Hypothesis: Moderation of repetition of chronic stressors on stress-hormone association When considering the fact that prolonged stress has been associated with lower levels of progesterone and that changes of the HPG axis under stress are dependent on type, duration and frequency of the stressor (Roney & Simmons, 2015), a hypothesis emerges: The
repetition of chronic stressors moderates the relationship between perceived chronic
psychosocial stress, cortisol and levels of estradiol and luteal progesterone. It is expected that the more often stressors have been experienced, the stronger the negative association between chronic stress, cortisol and ovarian hormone levels is.
Method Research Design
The present study is of observational nature, as no manipulation was performed. Due to the hormone samples being collected daily, the study is a repeated measures design. The study and analysis are carried out using a preexisting data set collected during the project STAR (Ziomkiewicz, Wichary, Bochenek, Pawlowski & Jasienska, 2012).
Study Group
The study sample consisted of 106 healthy women from the city of Wroclaw in the South-West of Poland. View table 1 for demographics of the sample. Participants were recruited to the study based on several inclusion criteria: 1) having a regular menstrual cycle (between 24 and 36 days), 2) not pregnant or lactating at least three months prior to the study, 3) not taking hormonal contraceptives for at least three months prior to the study, 4) being of good health with no reproductive or endocrinological disorders. Study participants were recruited using radio broadcastings and newspaper advertisements that were made, available
to the general population. The recruitment was carried out from the early spring of 2008 to the end of 2009. Participation was voluntary and the participants did not receive a monetary reimbursement, however they were provided with their hormone analysis results. There were no drop-outs during the study (Ziomkiewicz at al., 2012).
Procedure
During the initial meeting the purpose and procedure of the study was explained to the participants. Women were also given the material necessary to collect saliva samples and to record information about perceived stress together with the instructions for the execution of the study. The material consisted of a set of questionnaires and 35plastic vials to cover each day of menstrual cycle. Each participant was asked to collect first morning urine samples into plastic vials throughout the entire menstrual cycle, in order to assess the level of ovarian hormones and cortisol. After collecting the sample, the participants were instructed to label the vial with their initials, date and hour of sample collection. Additionally, they were told to keep the vials on the coldest temperature setting of their home freezer. The women were asked to fill out the set of questionnaires until the collection of the urine samples. During the first meeting anthropometric measurements were taken from every woman. These
measurements included: height, weight and body composition. After the last urine sample was collected, the women were instructed to contact the researchers. During that contact a second meeting was planned with the women in order to collect the samples and repeat the
measurements. The second meeting took place at least two weeks after the end of sampling. Sample Collection
The sample collection was to be started on the first day of the menstrual cycle, namely the first day of the menstruation and stopped on the starting day of the next menstruation. The samples were collected by a research assistant and transferred to the lab on ice and stored at -80 degrees Celsius until the hormone analysis was carried out.
Hormone Analysis
The levels of the metabolites of progesterone (PdG) and estradiol (EiG) were assessed via urinary concentration of their glucuronides using competitive enzyme immunoassay kits from Immunometrics Ltd. according to the published manufacturer procedure. In the analysis the glucuronide binds to the antibody and is made visible using alkaline phosphates, which are receptor enzymes. For estradiol the antibody used was Anti-Estrone-3-Glucuronide rabbit and the alkaline phosphate Estrone-3-Glucuronide. Progesterone was assessed using the Anti-Glucuronide rabbit antibody and the alkaline phosphate
Pregnanediol-3a-Glucuronide. For every woman 32 daily samples were analyzed, namely 18 samples for estradiol during the follicular and luteal phase and 14 for progesterone during the luteal phase (Ziomkiewicz at al., 2012).
The urinary concentration of cortisol was assessed using competitive
immunoenzymatic colorimetric method kits from DRG international, Inc. according to the published manufacturer manual. This kit is made up of a microplate (the solid phase) coated with antibodies targeting cortisol (anti-cortisol antibodies) and a modified cortisol which is covalently linked to a horseradish peroxidase (HRP-cortisol). During incubation, urinary cortisol competes with HRP-cortisol for antibody binding. During the washing, unbound urinary and HRP-cortisol are removed, while bound molecules remain attached to the solid phase. As urinary and HRP-cortisol have the same affinity for the solid phase, an increase in urinary cortisol will lead to a lower amount of HRP-cortisol binding to the microplate. After washing, the microplate is treated with H2O2 and TMB substrate, turning blue in the presence of HRP. Thus, the more urinary cortisol, the fewer HRP-cortisol bound, the lower the signal. Finally, a stop solution (H2SO4) is added to the sample, which turns the blue color into yellow. The color intensity is inversely related to the concentration of free cortisol in the sample. The exact cortisol concentration is then calculated using a calibration curve. Perceived Stress Assessment
Information about the perceived psychosocial stress was collected with the Recent Life Changes Questionnaire (Rahe, 1975). In alignment with the study population, the Polish version of the questionnaire was used. The polish version of the questionnaire is internally consistent with a Cronbach’s alpha of .85 (Sobolewski, Strelau, & Zawadzki, 1999). The measure consists of 55 items, which are grouped together in five domains: health, work, home and family, personal and social, and financial. For each domain, the subject has to indicate, which of the stressors they have experienced over the past six, 12, 18 or 24 months. Examples of items are: “Changes in residence”; or “Death of a Spouse”. Chronic psychosocial stress was operationalized by calculating the sum of scores for each stressor over the four timeframes. The more life stressors indicated, the higher the chronic psychosocial stress score.
Repeated Stressor Assessment
Over the four timeframes, the occurrence of a stressor on each timeframe was summed up. A sum of score higher than two on a stressor item indicated its repeated occurrence, which in this study translates to a repeated stressor. The sum of scores were then added up into one
variable per woman. Thus, if a woman experienced job loss in two timeframes, for example in the last six months and in the last seven to twelve months, she received a repeated stressor score of one. If this happened for two different stressors, she received a score of two. The sum of scores were then added up into one variable per woman.
Data-analysis Data-cleaning
The data was analyzed using the statistical software SPSS version 23. Before the main statistical analysis, an outlier check was performed using boxplots. Flagged values in the boxplots were analyzed by calculating z-scores of the variables with outliers. The z-score variables were arranged from lowest to highest value. It was then checked if there were z-values present above 3.5. Observations with a z-value above +3.5 or below -3.5 were considered outliers (Kannan, Manoj, & Arumugam, 2015).
Furthermore, two data sets were restructured and merged. The RLCQ data set was restructured from a long to a wide format. The hormone measures data set was merged with the data set containing the values of the RLCQ. Then, missing values were detected. This was achieved by asking frequencies for all variables. The variables with missing data were then further analyzed to find out if the data was missing at random. For that purpose, three Little’s MCAR tests were performed for the 18 Estradiol day variables; the 14 Progesterone day variables; and for the independent variable Chronic Stress and Cycle Cortisol, Luteal Cortisol and covariates BMI and Temperamental Emotional Reactivity separately (Little, 1988). Following this, the missing data was replaced using the multiple imputation technique, because the data was missing at random. Afterwards, frequencies were performed again to check if there was no missing data.
Additionally, assumptions were checked. Normality was visually tested using Q-Q plots. This resulted in the finding that the estradiol and progesterone variables were not normally distributed. Thus, logarithm variables were calculated in SPSS to be used in the analysis. Q-Q plots were made for the logarithm variables. The logarithm variables were visually normally distributed.
Confounding variables
The variable BMI (Body Mass Index) was included in the analysis, as body fat may affect level of hormones. Several studies demonstrated an association between BMI and ovarian hormone levels in naturally cycling women (Bohlke, Cramer, & Barbieri, 1998; Ecochard, Marret, Barbato, & Boehringer, 2000).
In addition, temperamental trait of emotional reactivity was included in the models as a confounder for the reason that the RLCQ does not offer information on the emotional values of a stressor. In addition, studies have found that temperamental traits of endurance, activity and emotional reactivity were correlated with the levels of estradiol and progesterone during the menstrual cycle (Ziomkiewicz, Wichary, Bochenek, Pawlowski, & Jasienska, 2012). Thus, by including emotional reactivity, the analysis was corrected for the differences in the strength of emotional reaction to a stressor.
Statistical-analysis
The association between chronic psychosocial stress and the levels of ovarian
hormones was tested using : 1) two-way ANCOVA with levels of Cycle Estradiol and Luteal Progesterone as dependent variables and groups based on RLCQ score and Cycle Cortisol as a between-subject factor, as well as Temperamental Emotional Reactivity and BMI as the covariates; 2) multiple regression models with levels of Cycle Estradiol and Luteal
Progesterone as dependent variables and RLCQ score, BMI and Cycle or Luteal Cortisol as independent predictors; 3) a moderation analysis with Repetition Stressor as moderator, RLCQ score and Cycle or Luteal Cortisol as independent variables and Cycle Estradiol and Luteal Progesterone as the dependent variables. The analysis was conducted using moderation model one of the PROCESS macro for SPSS (Montoya & Hayes, 2017). An alpha level of .05 was established for all analyses.
Simple ANOVA’s
Simple ANOVA’s were carried out to compare the groups for differences in age, BMI, education, Cycle and Luteal Cortisol, Cycle Estradiol and Luteal Progesterone.
Two-way ANCOVA
To perform the two-way ANCOVA with levels of Cycle Estradiol and Luteal
Progesterone as dependent variables and groups based on the Chronic Stress score and Cycle Cortisol levels as between-subject factors, and BMI and Temperamental Emotional Reactivity as covariates, participants were assigned to the groups based on Luteal or Cycle Cortisol level (high/low) and Chronic Stress score (high/low). The RLCQ and cortisol groups were formed by performing a median split. This resulted in two groups for each variable. The grouping variables were the between subject factors in the two-way ANCOVA.
Multiple Regression Models
The multiple regression models with levels of Cycle Estradiol and Luteal Progesterone as dependent variables and Chronic Stress, BMI and Luteal/Cycle Cortisol as independent
predictors were performed. This was done to examine if the amount of experienced psychological stressors or cortisol level can predict the level of cycle estradiol or luteal progesterone in women. The sum score for the RLCQ stressors (Chronic Stress) per woman was used as an independent predictor in the model for Cycle Estradiol and Luteal
Progesterone. Further, the cycle average of cortisol (Cycle Cortisol) per woman was used as an additional independent predictor. For the multiple regression model performed on
estradiol, the cycle average of estradiol (Cycle Estradiol) per woman was used as the
dependent variable. For the multiple regression model performed on progesterone, the Luteal Progesterone variable was used as a dependent variable. As an independent predictor the average cortisol during the luteal phase, namely Luteal Cortisol was used.
Moderation analysis
This analysis was performed to find out if the repeated stressors have an additional effect on the relationship between the experienced chronic psychosocial stressors or cycle/luteal cortisol and ovarian hormone levels in women. This analysis was carried out separately for the level of estradiol and progesterone and for cycle cortisol, luteal cortisol and chronic stressors. In the moderation analysis for estradiol, Cycle Estradiol was used as the dependent variable. Chronic Stress was used as an independent predictor, along with BMI and Temperamental Emotional Reactivity as covariates. The Repeated Stressor variable was used as the moderator. In the analysis with Luteal Progesterone as the independent variable the same variables were added, except for Cycle Cortisol, which was replaced by Luteal Cortisol. The PROCESS macro, model number one, automatically centers the variables around the means and adds the interaction term to the model.
Consequently, to test the hypothesis that the repetition of the stressor moderates the relationship between perceived chronic psychosocial stress and levels of hormones, the following variables were included into the model: Chronic Stressors, Temperamental Emotional Reactivity, BMI, Luteal Cortisol or Cycle Cortisol, and the interaction between Chronic Stressors and Repeated Stressors.
Results Sample Demographics
All analyses were carried out on the study sample consisting of 106 healthy, polish women of reproductive age with an average of 29.5 years (SD=3.58). Most of the women (80.2%) hold a master’s degree, 12.3% a bachelor’s degree, 6.6% a secondary degree and 0.9% a vocational degree. View table 1 for further demographics of the sample.
Table 1 Sample Demographics Demographics M/ % (SD) N 106 Age 29.6 (3.58) Education Master 80.2% Bachelor 12.3% Secondary 6.6% Vocational 0.9% BMI 22.85 (3.62) Cycle Cortisol 122.78 (46.91) Luteal Cortisol 129.06 (54.33) Cycle Estradiol 86.92 (77.07) Luteal Progesterone 5.51 (2.68) Chronic Stressors 16.13 (8.40) Repeated Stressors 2.10 (2.14) Temperamental Emotional Reactivity 11.93 (4.24) Outliers
The outlier check through visual inspection of boxplots identified no outliers on the variables Cycle Estradiol (E1G cycle) or Luteal Progesterone (Luteal PDG). Thus, all 106 women were included in the analyses.
Little’s MCAR test
The frequencies analysis to search for missing values resulted in the finding of 29 missing values. The missing values were found in the following variables: five on the variable E1Gm9, one on the variable E1Gm8, two on the variable E1Gp6, four on E1Gp7, seven on EiGp8, one on PdGd3 and PdGd2, two on PdGd1, one on Cycle Cortisol and Luteal Cortisol and four on the Repeated Stressors variable.
Little’s MCAR test showed an insignificant result for estradiol days +1 to 8, day 0, and 1 to -9, E1Gcycle/follicular/luteal/midcycle, which means that the values were missing at random (χ² (101, N = 106) = 69.152, p = .994). Little’s MCAR test for the independent variable Chronic Stressors and the covariates BMI, Temperamental Emotional Reactivity, and Cycle Cortisol was insignificant (χ² (101, N = 106) = 69.152, p = .994), thus the values were missing at random. In contrast Little’s MCAR test for progesterone (days -14 to -1, LutealPdG and Midluteal PdG) was significant (χ² (28, N = 106) = 51.134, p = .005). As a result, the patterns of the missing values were analyzed. The multiple imputation pattern analysis showed that the four missing values were distributed over the variables PdGd1, 2 and 3. Two of the values were missing for PdG day 1, and one for PdG day 2 and 3. The percentage of the missing data was lower than five percent, with 1.9% on PdGd1 and 0.9% on PdGd2 and 3 and hence unproblematic for the analysis. Missing values were therefore imputed with multiple imputation.
Descriptive statistics for Cortisol and Stress Groups
Table 2
Means (Standard deviation) of age, BMI, Cycle Estradiol, Luteal Progesterone, Cycle and Luteal Cortisol and Temperamental Emotional Reactivity and percentage of Education levels for the four groups.
Cortisol and Stress Groups
Cortisol High Cortisol Low
Stress Low Stress High Stress Low Stress High
Age 29.07 (3.63) 28.68 (2.56) 31.03 (3.56) 29.72 (3.95) Eduction Master Vocational 75.76% 0% 90% 0% 73.91% 4.35% 83.33% 0% BMI 22.89 (3.59) 21.09 (3.34) 24.41 (3.83) 22.81 (3.28) Cycle Estradiol 4.094 (1.03) 4.274 (0.95) 3.824 (0.93) 4.064 (0.85) Luteal Progesterone 1.59 (0.53) 1.76 (0.48) 1.46 (0.62) 1.46 (0.62) Cycle Cortisol 161.00 (33.17) 159.66 (37.04) 85.54 (17.13) 84.73 (20.85)
Luteal Cortisol 168.59 (44.07) 173.94 (42.29) 85.31 (22.19) 89.23 (25.93) Temperamental Emotional Reactivity 12.30 (4.81) 13.35 (4.43) 11.26 (3.61) 11.10 (3.81) Group Comparisons Chronic Stressors
Chronic Stress groups did not differ significantly with Age (F(1, 104) = .676, p > .05), Temperamental Emotional Reactivity (F(1, 104) = .023, p > .05), Cycle Estradiol (F(1, 104) = .807, p > .05) and Luteal Progesterone (F(1, 104) = .310, p > .05), Luteal Cortisol (F(1, 104) = 1.141, p > .05) and Cycle Cortisol (F(1, 104) = 3.295, p > .05). However when BMI is considered the Chronic Stress groups differ significantly (F(1, 104) = 3.968, p < .05), with the low Chronic Stress group having a significantly higher BMI (M = 23.511; SD = 3.735) than the high Chronic Stress group (M = 22.126; SD = 3.378).
Cycle Cortisol
Cycle Cortisol groups did not differ significantly with BMI (F(1, 104) = 3.453, p > .05), Temperamental Emotional Reactivity (F(1, 104) = 3.513, p > .05), Cycle Estradiol (F(1, 104) = 1.223, p > .05), Luteal Progesterone (F(1, 104) = 1.893, p > .05) and Age (F(1, 104) = 3.939, p = .05). As expected, the Cycle Cortisol groups differed significantly with Cycle Cortisol (F(1, 104) = 207.571, p <.001), with the low Cycle Cortisol group having
significantly lower Cycle Cortisol (M = 4.415; SD = .25) than the high Cycle Cortisol group (M = 5.056; SD = .206). Lastly, the Luteal Cortisol groups differed significantly on Luteal Cortisol, (F(1, 104) = 148.729, p <.001), with the low Luteal Cortisol group having
significantly lower Luteal Cortisol (M = 4.428; SD = .315) than the high Luteal Cortisol group (M = 5.108; SD = .256).
Luteal Cortisol
Luteal Cortisol groups did not differ significantly with BMI (F(1, 104) = 3.078, p > .05), Age (F(1, 104) = 2.842, p > .05), Temperamental Emotional Reactivity (F(1, 104) = 1.475, p > .05), Cycle Estradiol (F(1, 104) = 2.225, p > .05) and Luteal Progesterone (F(1, 104) = 3.209, p > .05. The Luteal Cortisol groups significantly differ with Cycle Cortisol (F(1, 104) = 162.280, p <.001), with the low Luteal Cortisol group having significantly lower Cycle Cortisol (M = 4.429; SD = .269) than the high Luteal Cortisol group (M = 5.04; SD = .224). The Luteal Cortisol groups did further differ significantly on Luteal Cortisol, (F(1, 104)
= 196.939, p <.001), with the low Luteal Cortisol group having significantly lower Luteal Cortisol (M = 4.409; SD = .296) than the high Luteal Cortisol group (M = 5.126; SD = .225). Differences in estradiol between chronic stress and cortisol groups
A two-way ANCOVA was carried out to test if women who perceived more life stressors and had higher cortisol level had lower levels of estradiol during the menstrual cycle. View Table 3 for descriptive statistics and estimates. The effect of Chronic Stress group on Cycle Estradiol was not significant, F(1, 100) = 0.249, p = .619. Thus, the levels of Cycle Estradiol in the group with high Chronic Stress (M = 4.10, SE = 0.13) and the group with low Chronic Stress (M = 4.01, SE = 0.12) did not differ significantly, when controlled for BMI and Temperamental Emotional Reactivity. Further, the effect of Cycle Cortisol group on the Cycle Estradiol was not significant, F(1, 100) = 0.932, p = .337. Therefore, the levels of Cycle Estradiol in the group with high Cycle Cortisol (M = 4.15, SE = 0.13) and the group with low Cycle Cortisol (M = 3.97, SD = 0.13) did not differ significantly, when controlling for BMI and Temperamental Emotional Reactivity. Lastly, there was no statistically
significant interaction effect of Cycle Cortisol Group X Chronic Stress Group on Cycle Estradiol levels, whilst controlling for BMI and Temperamental Emotional Reactivity, F(1, 100) = 0.005, p = .944.
Table 3
Means, Adjusted Means, Standard Deviations and Standard Error for Cycle Estradiol for the four groups.
Cycle Cortisol and Stress Groups
Cycle Cortisol High Cycle Cortisol Low Stress Low Stress High Stress Low Stress High Estradiol
M 4.09 4.27 3.82 4.06
(SD) (1.03) (0.94) (0.93) (0.84)
Madj 4.11 4.19 3.92 4.02
(SE) (0.15) (0.20) (0.19) (0.16)
Another two-way ANCOVA was conducted to test if women who perceived more life stressors had lower levels of luteal progesterone. View table 4 for descriptive statistics and estimates. The effect of Chronic Stress group on the Luteal Progesterone levels was not significant, F(1, 100) = 0.31, p = .578. Thus, the levels of Luteal Progesterone in the high Chronic Stress group (M = 1.60, SD = 0.08) and the low Chronic Stress group (M = 1.54, SD = 0.08) did not differ significantly, whilst controlling for BMI and Temperamental Emotional Reactivity. Also, the effect of Luteal Cortisol group on the Luteal Progesterone level was non-significant, F(1, 100) = 2.66, p = .106. Therefore, the levels of Luteal Progesterone in the group with high Luteal Cortisol (M = 1.67, SD = 0.08) and the group low Luteal Cortisol (M = 1.48, SD = 0.08) did not differ significantly, whilst controlling for BMI and
Temperamental Emotional Reactivity. Lastly, there was no statistically significant interaction between Chronic Stress and Luteal Cortisol groups on Luteal Progesterone levels, whilst controlling for BMI and Temperamental Emotional Reactivity, F(1, 100) = 0.58, p = .449. Table 4
Means, Adjusted Means, Standard Deviations and Standard Errors for Luteal Progesterone in the four groups.
Luteal Cortisol High Luteal Cortisol Low Stress Low Stress High Stress Low Stress High Progesterone
M 1.59 1.76 1.46 1.46
(SD) (0.53) (0.48) (0.62) (0.62)
Madj 1.59 1.74 1.48 1.46
(SE) (0.10) (0.12) (0.12) (0.11)
Predictive value of chronic psychosocial stress and cortisol for cycle estradiol
A multiple linear regression model was run to predict the level of Cycle Estradiol from Cycle Cortisol, Chronic Stress, BMI and Temperamental Emotional Reactivity. These
variables significantly predicted Cycle Estradiol, F(4, 101) = 6.117, p < .001, R2 = .195. In
total BMI, Temperamental Emotional Reactivity and the level of Cycle Cortisol explained nearly 20% of the variance in the level of Cycle Estradiol. Women’s estradiol decreased -.263 for each point of BMI and -.197 with a one-point increase in Temperamental Emotional Reactivity score. Women’s estradiol increased .215 per nmol/L of Cycle Cortisol.
Predictive value of chronic psychosocial stress and cortisol for luteal progesterone A multiple linear regression model run to predict the level of Luteal Progesterone from Luteal Cortisol, Chronic Stress, BMI and Temperamental Emotional Reactivity was not significant (F(4, 101) = 2.078, p = .089), with an R² of .076. Luteal Cortisol level (β= .246, p < .05) was the only significant predictor of Luteal Progesterone level. Neither Chronic stress (β= .05, p > .05), BMI (β= -.056, p > .05) nor Temperamental Emotional Reactivity (β= .031, p > .05) predicted the level of Luteal Progesterone.
Moderation of repeated stressors on chronic stress and estradiol
The variables Chronic Stress, Temperamental Emotional Reactivity, BMI, Repeated Stressors and ChronicStressXRepeatedStressors in the regression model accounted for around 18% of variance in Cycle Estradiol, R² = .180, F(5,100) = 6.840, p < .001. The variables that significantly accounted for the variance were Temperamental Emotional Reactivity (p < .05), Cycle Cortisol (p < .05) and BMI (p < .05). There was no significant effect of the interaction between ChronicStressXRepeatedStressors on the level of Cycle Estradiol (p > .05).
Moderation of repeated stressors on chronic stress and progesterone
None of the variables Chronic Stress, Temperamental Emotional Reactivity, BMI, Repeated Stressors or ChronicStressXRepeatedStressors in the regression model accounted for a significant amount of variance in the level of Luteal Progesterone, R² = .036, F(5,100) = .511, p > .05.
Moderation of repeated stressors on cortisol and estradiol
The variables Temperamental Emotional Reactivity, BMI, Cycle Cortisol, Repeated Stressors and interaction of CycleCortisolXRepeatedStressors in the regression model accounted for around 16% of variance in Cycle Estradiol (R² = .163, F(5,100) = 5.199, p < .001). The variables that significantly predicted Cycle Estradiol were Temperamental Emotional Reactivity (p = .056), Cycle Cortisol (p < .05) and BMI (p < .05). The interaction between CycleCortisolXChronicStress did not significantly account for the variance in the level of Cycle Estradiol (p > .05).
Moderation of repeated stressors on luteal cortisol and luteal progesterone
Temperamental Emotional Reactivity, BMI, Luteal Cortisol, Repeated Stressors and interaction LutealCortisolXRepeatedStressors in the regression model were not associated with the level of Luteal Progesterone (R² = .089, F(5,100) = 1.510, p > .05.
Discussion
This research was conducted to find out if there is an association between the levels of female ovarian sex steroids produced during the menstrual cycle and perceived chronic psychosocial stress. The research also tried to answer the question if the repetition of the experienced chronic psychosocial stressors moderates the relationship between female ovarian hormone levels and chronic psychosocial stress.
Based on the results of conducted analyses these questions should be answered negatively. No differences between groups of women with high and low number of experienced chronic psychosocial stressors was found in the levels of estradiol and progesterone. Also, no difference in the levels of estradiol and progesterone was found between groups of women with high and low levels of cortisol during the whole cycle and luteal phase. These results may suggest that the number of perceived psychosocial stressors did not affect the amount of physiological stress affecting women during the period of the observation.
In contrast, the analyses demonstrated a significant effect of cortisol on the level of estradiol during women’s cycle. A higher level of cortisol during the entire cycle predicted a higher level of estradiol level in the women during the same period. Similarly, a higher level of cortisol during the luteal phase predicted a higher level of progesterone during this
menstrual cycle phase. The observed associations were independent from the effect of known confounders such as body mass index and temperamental trait of emotional reactivity, that were demonstrated to affect levels of ovarian hormones in a previous research (Ecochard, Marret, Barbato, & Boehringer, 2000; Ziomkiewicz, Wichary, Bochenek, Pawlowski, & Jasienska, 2012).
Finally, the analysis did not confirm the expected moderating effect of repetition of the perceived psychosocial stressors over a longer period of time on the association between chronic stressors and sex hormone levels. This finding may further support the notion that perceived stressors alone have limited effect on women’s physiology.
Chronic stressors and reproductive physiology
Against the expectation, no effect of chronic stress on ovarian hormone levels was found. This is consistent with a study that administered chronic stress in the form of a foot shock paradigm to female rats over fourteen days, which led to a significant increase in ACTH, but not corticosterone concentrations. The researchers did not find an effect of
Bauman, Chu, Ghosh, & Kant, 1996). While the study population consisted of rodents and a stressor threatening physical integrity was employed the findings are still an indication of what could be expected to be found in a human sample. Also, a study using an anxiety provoking exam as a psychological stressor did not find an effect on ovarian steroid levels one or six months after the exam (Ellison, Lipson, Jasienska, & Ellison, 2007). While the study assessed a shorter stressor timeframe, it studied a comparable population and psychosocial stressor. In addition, the researchers Roney & Simmons (2015) yielded a
comparable result in a sample of healthy, fertile women, finding that throughout the menstrual cycle lower daily salivary estradiol, but not progesterone was predicted by daily perceived stress. Even though the study suggested that longer periods of psychosocial stress might reveal an effect of perceived stress on progesterone levels, that notion was not supported by the present study. The studies differed on the length of stressor assessment, with the Roney and Simmons measuring for one menstrual cycle, which might not be comparable to a two-year assessment. Lastly, a large American cohort study that partially studied the association of experienced stressors during women’s pregnancies with their estradiol and other sex steroid levels found a comparable result. The researchers have found no association between stressors such as illness or injury of family members, relationship difficulties, financial strains or unemployment and sex steroid levels. The sample was pregnant women, which does not generalize to normally menstruating women, because of elevated cortisol levels and lower cortisol increase in response to stressors during pregnancy. Despite this, the study is
comparable in the type and chronicity of stressors to the present study (Barrett et al., 2019). Possible mechanisms
The finding of the present study that the number of perceived psychosocial stressors did not affect the amount of physiological stress affecting women during the period of the observation could be explained by the research of Marin et al. (2007), from which it has been concluded that episodic stressors do not affect biological outcomes in the absence of daily hassles. As previously mentioned, there are different types of stressors. Examples of episodic stressors can be losing one’s job, having a baby or moving houses, more chronic examples of episodic stressors would be an abusive relationship or bullying at the workplace. Episodic stressors do not usually occur frequently, but are often major life events, as measured in the Recent Life Changes Questionnaire. The stressors typically associated with negative long-term consequences are rather unconscious and small but happen frequently and repeatedly. Those types of stressors are called daily hassles (Guilliams, & Edwards, 2010). This is in accordance with the fact stated previously that different types of stressors elicit different types
of biological reactions (Kajantie & Phillips, 2006; O'Connor et al., 2011). The present study has solely measured chronic episodic stressors and it is possible, that the women have not simultaneously experienced daily hassles, which could explain the absence of an effect.
Additional support for no effect of perceived psychosocial stressors on the amount of physiological stress experienced by the women, comes from a study that concluded that episodic stressors are not associated to changes in the cardiovascular response (Lepore, Miles, & Levy, 1997). Thus, it is possible that the assessed stressors do not promote a strong enough physical response to bring about observable change. Furthermore, it is possible that the reaction of the HPA axis to this type of stressors is not sufficient to account for long-term adverse effects (Marin, Martin, Blackwell, Stetler, & Miller, 2007). Finally, Lazarus (1990) has summarized, that studies using questionnaires of life-events have typically found no or small correlations with health outcomes. This could further explain the absence of an
observable effect of chronic perceived psychosocial stressors on the reproductive hormones in the present study.
A possible explanation for the observed lack of a moderation effect of repetition on the association between the stressors and ovarian hormone levels is the process of habituation or hormonal habituation. Habituation can be described as a weakened or lack of physiological response to a stressor due to repeated experience of it and familiarization to it (Fioravante, Antzoulatos, & Byrne, 2009). During familiarization a person learns that a stressor is harmless and thus does not feel stressed by it anymore. This mechanism is further strengthened when the stressor occurs in the same context (Grissom, Iyer, Vining, &
Bhatnagar, 2007). As a result of these mechanisms, the reaction to the stressor is the same as in the baseline condition without stressors (Cyr, & Romero, 2009). In this study a moderation was tested for the repetition of the experience of chronic psychosocial stressors. If an
individual is exposed to a stressor repeatedly, familiarization can take place. Thus, it can be concluded that women might have been habituated to the stressors they experienced.
Furthermore, this could explain not only the absence of a moderation effect of the repeated stressors, but also a general lack in the effect of chronic psychosocial stress on reproductive hormone levels. A number of the chronic stressors were experienced repeatedly, which could have led to habituation.
As explained previously, when one undergoes prolonged stress reactions, the activity of the HPA axis is heightened at first, but over time it can be attenuated (Rodriguez et al., 2015). This process possibly takes place to protect the body from adverse effects of prolonged exposure to increased cortisol levels, which may compromise overall health and biological
fitness. The consequence of such prolonged exposure is usually a dysregulation of the HPA axis. Accordingly, Cyr and Romero (2009) have found that HPA axis dysregulation can lead to desensitization of the response to any physical stressor, which usually takes place after a period of hypercortisolism. Meaning that chronically stressed people, might show no sign of physiological stress or show a significantly weakened response to the perceived stressor. In this case scenario, there is no habituation to the stressors, but hypocortisolism is developed instead (Guilliams, & Edwards, 2010). Possible mechanisms of desensitization include downregulation of glucocorticoid receptors (Hauger, Millan, Lorang, Harwood, & Aguilera, 1988) or a downregulation of stress hormone production (Heim, Ehlert, Hanker, &
Hellhammer, 1998). Therefore, it is possible that women in the sample may have felt chronically stressed, but did no longer experience a physical response to the stressors.
Desensitization and hypocortisolism could explain why no association was found between the number of experienced chronic stressors and the ovarian hormone levels of women.
Finally, what is experienced as a stressor is highly subjective. A few variables that account for the perception of an event as a stressor include coping style (Harkness, &
Monroe, 2016), but also the previous experiences of women (Rabkin, & Struening, 1976) and temperamental trait characteristics (Ziomkiewicz, Wichary, Bochenek, Pawlowski, &
Jasienska, 2012). The association between levels of estradiol and emotional reactivity found in this study suggest that the effect of perceived chronic stressors could be moderated by the emotional reactivity. However, such hypothesis requires further testing, which is beyond the scope of current analysis.
Cortisol levels and ovarian hormone levels
The finding that an increase in cortisol predicts an increase in estradiol and progesterone was the opposite of what was expected in accordance to the hypothesis that ovarian hormone levels decrease with increasing cortisol levels. It is extensively stated in existing literature, that cortisol has a suppressive effect on reproductive behavior (Sapolsky, Romero & Allan, 2000). As mentioned previously, possible mechanisms behind the
suppressive effect are a decrease of GnRH from the hypothalamus, elicited by a rise in cortisol (Belhadj, De Besi, Bardin, & Thau, 1989) and enhanced release of gonadotropin inhibiting hormone (Kirby, Geraghty, Ubuka, Bentley & Kaufer, 2009). The absence of the expected effect in contrast with the described literature could be explained by the proposed mechanism of desensitization, which can be caused by HPA axis dysregulation. When desensitization takes place, there is an absence in the rise of cortisol in response to a stressor, thus low levels of cortisol can be an indicator of desensitization. In that case, a rise of
estradiol with cortisol can be expected, as higher cortisol levels signal an absence of HPA axis dysregulation and desensitization in the women. In line with this result, it has been stated in literature that at a homeostatic level, cortisol plays a role in gonadotropin release and thus normal ovarian hormone levels (Breen & Mellon, 2014). Another study supporting this notion concluded that HPA axis activity is positively associated with FSH and LH levels in young female cancer survivors (Hardy, Garnier-Villarreal, Ohlendorf, & McCarthy, 2019).
Another possible explanation can be found in inconclusive findings on the effect of various stressor types and characteristics on the release of cortisol and dysregulation of the HPA axis. A meta-analysis by Chide and Steptoe (2009) showed that there are various effects of different types of stressors, like occupation stress or burnout, on the cortisol awakening response. Occupational stress was associated with an increased cortisol awakening response, while burnout or fatigue are associated with a decrease in cortisol awakening response. As existing literature on the effects of stressors on cortisol has been inconclusive, researchers have carried out an extensive review. Overall, it was found that chronic stress is associated with lower concentrations of morning cortisol, but higher levels of cortisol throughout the day, which is in line with proposed hypocortisolism. Further, various characteristics of the stressors are associated with different cortisol reactions. A stressor that is still present elicits higher morning cortisol, while a stressor that is terminated elicits lower morning cortisol. A stressor that feels controllable to the experiencer elicits higher morning cortisol and an uncontrollable stressor lower (Miller, Chen & Zhou, 2007). In accordance to the present research, it cannot be concluded from the RLCQ if stressors were still present and if women felt in control over the perceived stressor. Thus, it is unclear which effect the various assessed stressors may have had on cortisol. Further research is needed to gain a clear understanding of the effects of stressors on cortisol and vice versa.
Study limitations
A limitation of the present study is the retrospective nature of the psychosocial stressor assessment. Since the Recent Life Changes Questionnaire addresses the past two years, there might be an erroneous recollection of the experienced stressors. It is also possible that women do not remember if they actually experienced the event as stressful. The risk of recollection bias increases with time. Therefore, it is possible that information about repetitiveness of the stressors was not precise and as such affected the results of the current analysis. If that is the case, the stressor score of the women might not have been a reliable measurement of the stress level of the studied women. Future studies should consider prospective methods for monitoring of repeated chronic stressors.
Another limitation of the present study is the difference of timeframes of the variables. While the stress variable covers a time-span of two years in retrospect, the time-frame of the ovarian hormone levels and cortisol only covered the length of one menstrual cycle. Looking for long-term effects of stressors might better be carried out on a data-set with purely
longitudinal data. Conclusions
The aim of the presented study was to enrich the knowledge on the effects of chronic stress on reproductive physiology. Certainly, the results of this study should stimulate further research, which is necessary to assemble conclusive evidence of the effects of different types of stressors, but especially chronic stressors. Based on the results of the current study it can be concluded that neither number nor repetition of stressors affected ovarian hormone levels in naturally cycling women. Simultaneously, results of the analysis considering the level of cortisol and its association with the level of reproductive hormones bring possible new insights into the physiological path of the impact of chronic stress. Although, the observed effect of cortisol was not expected it could be partially explained based on evoked hypothesis of desensitization.
Longitudinal studies in human samples are a promising route for further knowledge enrichment. Therefore, it is recommended to carry out further research to find out more about the effects of chronic or episodic stress. In order to have a better match on the timeframes of the variables, a longitudinal study with a continuous measurement of stressors and ovarian hormone levels could be carried out in the future. A stressor might better be defined as ongoing in order to be chronic in future research. In this research stress was viewed as chronic, when it was repeatedly experienced by the woman. For that operationalization a cohort study design is suggested, because it is impossible to know if stressors will become chronic by being ongoing without following a sample over time. In addition, it is suggested to use a different stress measurement, or add it to the RLCQ. It is recommended to make use of the Hassles Scale developed by Lazarus in 1970. The Hassles Scale measures minor daily hassles, which are perceived as stressful. That will allow to find out if there are indeed different effects of major life event stressors and daily hassle stressors on ovarian hormone levels in women.
Adams, M. R., Kaplan, J. R., & Koritnik, D. R. (1985). Psychosocial influences on ovarian endocrine and ovulatory function in Macaca fascicularis. Physiology &
behavior, 35(6), 935-940.
Albert, K., Pruessner, J., & Newhouse, P. (2015). Estradiol levels modulate brain activity and negative responses to psychosocial stress across the menstrual cycle.
Psychoneuroendocrinology, 59, 14-24.
Anderson, S. M., Saviolakis, G. A., Bauman, R. A., Chu, K. Y., Ghosh, S., & Kant, G. J. (1996). Effects of chronic stress on food acquisition, plasma hormones, and the estrous cycle of female rats. Physiology & behavior, 60(1), 325-329.
Bachman, G. A.; Kemmann, E. (1982). Prevalence of oligomenorrhea and amenorrhea in a college population. Am. J. Obstet. Gynecol. 1(44), 98-102.
Barrett, E., Mbowe, O., Thurston, S.W., Butts, S., Wang, S., Nguyen, R., Bush, N., Redmon, B., Sheshu, S., Swan, S.H., & Sathyanarayana, S. (2019). Predictors of Steroid Hormone Concentrations in Early Pregnancy: Results from a Multi-Center Cohort. Maternal and Child Health Journal, 23, 397-407.
Belhadj, H., De Besi, L., Bardin, C. W., & Thau, R. B. (1989). The implication of opiates in the glucocorticoid-mediated inhibition of LH secretion in rats. Journal of
endocrinology, 122(2), 451-456.
Björntorp, P., & Rosmond, R., (2000). Obesity and cortisol. Nutrition, 16, 924-936.
Bohlke, K., Cramer, D. W., & Barbieri, R. L. (1998). Relation of luteinizing hormone levels to body mass index in premenopausal women. Fertility and sterility, 69(3), 500-504. Breen, K.M., & Mellon, P.L. (2014). Influence of stress-induced intermediates on
gonadotropin gene expression in gonadotrope cells. Molecular and Cellular Endocrinology, 385, 71–77.
Caplan, R. D., Cobb, S., & French Jr, J. R. (1979). White collar work load and cortisol: Disruption of a circadian rhythm by job stress?. Journal of Psychosomatic Research, 23(3), 181-192.
Chida, Y., & Steptoe, A. (2009). Cortisol awakening response and psychosocial factors: a systematic review and meta-analysis. Biological psychology, 80(3), 265-278.
Conrad, M., Hubold, C., Fischer, B., & Peters, A. (2009). Modeling the hypothalamus– pituitary–adrenal system: homeostasis by interacting positive and negative feedback. Journal of biological physics, 35(2), 149-162.
Cyr, N. E., & Romero, L. M. (2009). Identifying hormonal habituation in field studies of stress. General and comparative endocrinology, 161(3), 295-303.
De Liz, T. M., & Strauss, B. (2005). Differential efficacy of group and individual/couple psychotherapy with infertile patients. Human Reproduction, 20(5), 1324-1332. Ecochard, R., Marret, H., Barbato, M., & Boehringer, H. (2000). Gonadotropin and body
mass index: high FSH levels in lean, normally cycling women. Obstetrics & Gynecology, 96(1), 8-12.
Ellison, P. T., Lipson, S. F., Jasienska, G., & Ellison, P. L. (2007). Moderate anxiety, whether acute or chronic, is not associated with ovarian suppression in healthy, well‐
nourished, Western women. American journal of physical anthropology, 134(4), 513-519.
Fioravante, D., Antzoulatos, E. G., & Byrne, J. H. (2009). 18 Sensitization and Habituation: Invertebrate. Concise Learning and Memory: The Editor's Selection, 373.
Gaillard, A.W.K. (2006). Persoon en Omgeving. Stress, Productiviteit en Gezondheid. Amsterdam: Uitgeverij Nieuwezijds.
Grissom, N., Iyer, V., Vining, C., & Bhatnagar, S. (2007). The physical context of previous stress exposure modifies hypothalamic–pituitary–adrenal responses to a subsequent homotypic stress. Hormones and behavior, 51(1), 95-103.
Hagino, N., Watanabe, M., & Goldzieher, J. W. (1969). Inhibition by adrenocorticotrophin of gonadotrophin-induced ovulation in immature female rats. Endocrinology, 84(2), 308-314.
Hardy, T. M., Garnier-Villarreal, M., Ohlendorf, J. M., & McCarthy, D. O. (2019). Chronic Stress and Ovarian Function in Female Childhood Cancer Survivors. Oncology nursing forum, 46(3), 75-85.
Harrison, R. F., O'Moore, R. R., & O'Moore, A. M. (1986). Stress and fertility: some modalities of investigation and treatment in couples with unexplained infertility in Dublin. International Journal of Fertility, 31(2), 153-159.
Harkness, K. L., & Monroe, S. M. (2016). The assessment and measurement of adult life stress: Basic premises, operational principles, and design requirements. Journal of Abnormal Psychology, 125(5), 727.
Hauger, R. L., Millan, M. A., Lorang, M., Harwood, J. P., & Aguilera, G. (1988). Corticotropin-releasing factor receptors and pituitary adrenal responses during immobilization stress. Endocrinology, 123(1), 396-405.
Heim, C., Ehlert, U., Hanker, J. P., & Hellhammer, D. H. (1998). Abuse-related posttraumatic stress disorder and alterations of the hypothalamic-pituitary-adrenal axis in women with chronic pelvic pain. Psychosomatic medicine, 60(3), 309-318.
Heim, C., Ehlert, U., & Hellhammer, D. H. (2000). The potential role of hypocortisolism in the pathophysiology of stress-related bodily disorders. Psychoneuroendocrinology, 25(1), 1-35. disorders.
Herman, J. P., & Cullinan, W. E. (1997). Neurocircuitry of stress: central control of the hypothalamo–pituitary–adrenocortical axis. Trends in neurosciences, 20(2), 78-84. Kajantie, E., & Phillips, D. I. (2006). The effects of sex and hormonal status on the
physiological response to acute psychosocial stress. Psychoneuroendocrinology, 31(2), 151-178.
Kirby, E. D., Geraghty, A. C., Ubuka, T., Bentley, G. E., & Kaufer, D. (2009). Stress increases putative gonadotropin inhibitory hormone and decreases luteinizing hormone in male rats. Proceedings of the National Academy of Sciences, 106(27), 11324-11329.
Kannan, K. S., Manoj, K., & Arumugam, S. (2015). Labeling methods for identifying outliers. International Journal of Statistics and Systems, 10(2), 231-238.
Krohne, H. W. (2002). Stress and coping theories. International Encyclopedia of the Social Behavioral Sciences, 22, 15163-15170.
Lazarus, R. S. (1990). Theory-based stress measurement. Psychological inquiry, 1(1), 3-13. Lepore, S. J., Miles, H. J., & Levy, J. S. (1997). Relation of chronic and episodic stressors to
psychological distress, reactivity, and health problems. International Journal of Behavioral Medicine, 4(1), 39-59.
Little, R. J. A. (1988). A test of missing completely at random for multivariate data with missing values. Journal of the American Statistical Association, 83(404), 1198–1202. Loucks, A. B., Mortola, J. F., Girton, L., & Yen, S. S. C. (1989). Alterations in the
hypothalamic-pituitary-ovarian and the hypothalamic-pituitary-adrenal axes in athletic women. The Journal of Clinical Endocrinology & Metabolism, 68(2), 402-411. Marin, T. J., Martin, T. M., Blackwell, E., Stetler, C., & Miller, G. E. (2007). Differentiating
the impact of episodic and chronic stressors on hypothalamic-pituitary-adrenocortical axis regulation in young women. Health Psychology, 26(4), 447.
Mario G. Oyola & Robert J. Handa (2017) Hypothalamic–pituitary–adrenal and hypothalamic–pituitary–gonadal axes: sex differences in regulation of stress responsivity, Stress, 20:5, 476-494.
Mastorakos, G., Pavlatou, M. G., & Mizamtsidi, M. (2006). The hypothalamic-pituitary-adrenal and the hypothalamic-pituitary-gonadal axes interplay. Pediatric endocrinology reviews: PER, 3, 172-181.
McGuire, N. L., Koh, A., & Bentley, G. E. (2013). The direct response of the gonads to cues of stress in a temperate songbird species is season-dependent. PeerJ, 1, e139. Mikolajczak, M., Roy, E., Luminet, O., Fillée, C., & De Timary, P. (2007). The moderating
impact of emotional intelligence on free cortisol responses to stress. Psychoneuroendocrinology, 32(8-10), 1000-1012.
Montoya, A. K., & Hayes, A. F. (2017). Two condition within-participant statistical mediation analysis: A path-analytic framework. Psychological Methods, 22, 6-27. Morse, C. A., & Van Hall, E. V. (1987). Psychosocial aspects of infertility: a review of
current concepts. Journal of Psychosomatic Obstetrics & Gynecology, 6(3), 157-164. O'Connor, K. A., Brindle, E., Shofer, J., Trumble, B. C., Aranda, J. D., Rice, K., & Tatar, M.
(2011). The effects of a long‐term psychosocial stress on reproductive indicators in the baboon. American journal of physical anthropology, 145(4), 629-638.
Owens, M. J., & Nemeroff, C. B. (1991). Physiology and pharmacology of corticotropin-releasing factor. Pharmacological reviews, 43(4), 425-473.
