Comparison of salivary cortisol levels between women with functional hypothalamic amenorrhea and healthy women: a pilot study

Functional Hypothalamic Amenorrhea (FHA) has been associated with excessive-chronic stress, eating disorders, and weight loss. A common feature is the increased serum cortisol, but its measurement has many limitations. Currently, salivary cortisol (SC) has been proposed as a more sensitive and adequate index. To investigate the SC alterations through a 24-h period and the possible correlation with the severity of stress in women with FHA. Between July 2019 and March 2021, 12 FHA women and 12 healthy controls of comparable age were included. Psychological, eating, and physical abnormalities were evaluated by applying equivalent validated self-questionnaires. No significant differences were found between FHA women and healthy individuals with respect to morning (8:00 am) serum cortisol and ACTH (522.5 ± 162.2 vs 442 ± 138.5 nmol/l; p = 0.204 and 37 ± 44.8 vs 17.4 ± 10.8 pg/ml; p = 0.186, respectively). Women with FHA had statistically significant higher morning SC compared to control subjects (21.8 ± 5.9 nmol/l vs 12.8 ± 3.8 nmol/l; p < 0.001), while no significant differences were found regarding the afternoon (4:00 pm) and midnight (12:00 am) SC. A significant positive correlation was found between morning SC and morning serum cortisol (rho = 0.532; p = 0.007), EAT-26 (rho = 0.527; p = 0.008) and HADS-Anxiety score (rho = 0.471; p = 0.02). Additionally, a significant negative correlation between morning SC and BMI was observed (rho = −0.53; p = 0.009). Compared to serum cortisol, SC seems to express better the hypercortisolemic state of women with FHA and correlates well with the underlying contributing factors. Larger studies are needed in order to confirm these results and validate the optimal SC cut-off value associated with the development of FHA.


Introduction
Functional hypothalamic amenorrhea (FHA) is characterized by hypoestrogenism due to reduced gonadotropinreleasing hormone (GnRH) secretion [1].The diagnosis is based on the exclusion of other anatomical or organic causes of hypogonadotropic hypogonadism in adolescents and women whose menstrual cycle interval persistently exceeds 45 days and/or those who present with amenorrhea for 3 months or more [1][2][3].FHA is the third most common cause of secondary amenorrhea after pregnancy and polycystic ovary syndrome (PCOS).It can affect women of all ages but is most often diagnosed in young women and adolescents [4].It is associated with stress, weight loss, excessive exercise, undernutrition, or a combination of the above [1].
Serum cortisol is consistently found raised in women with FHA [5].Nevertheless, as serum cortisol levels vary through the day, in a circadian rhythm, a single measurement is usually inadequate or even deceptive for establishing the diagnosis of hypercortisolemia.Hence, there is a need for hospitalization to obtain a continuous monitoring of serum cortisol by taking several blood samples [6].However, this procedure may disrupt subject's daily routine, whereas it has a potential negative impact on serum cortisol levels due to stress related with hospitalization and repeated venipunctures [7].Furthermore, serum cortisol assays estimate the total cortisol, giving controversial results in cases of altered serum cortisol binding proteins, such as reduced cortisol-binding globulin (CBG) and albumin [6].Currently, the evaluation of cortisol through salivary measurements has been proposed [8].This method has the advantages of collecting samples without disturbing the normal daily routine of patients and avoiding iatrogenic stress-induced cortisol release.Moreover, it evaluates the free component with precision, instead of the total cortisol concentrations [9].Nonetheless, this method has not been evaluated in women with FHA.The aim of our study was to assess the aberrations of salivary cortisol (SC) through consecutive measures in a 24-h period (morning-afternoon-night) in women with FHA and to compare them with that of normally menstruating individuals.In addition, we aimed to investigate the possible correlation between the concentrations of SC and the underlying triggering factors of FHA.

Subjects
This comparative case-control study was performed in a gynecological outpatient clinic of a university hospital in Greece, between July 2019 and March 2021.Twelve women with FHA were identified and recruited through a Pediatric and Adolescent Gynecology and a Gynecological Endocrinology clinic, whereas 12 women with normal menstrual cycles were found through advertising within the hospital.Inclusion criteria: age between 15-35 years old and an ideal body weight between 90-110%.Exclusion criteria: time of menarche less than 3 years before the initiation of the study, state of pregnancy, Cushing syndrome, Diabetes Mellitus, hyperprolactinemia, thyroid dysfunction, non-classic congenital adrenal hyperplasia or other endocrine disorder, known or clinically apparent psychiatric or behavioral disease, known acute or chronic disease, treatment with metformin, use of corticosteroid or any hormonal medication including combined oral contraceptive pills in less than 6 months before study entry, treatment with antidepressants, anxiolytics or any other psychiatric medication, alcohol consumption greater than 20 gr/week and tobacco use.The study was performed according to the principles of the declaration of Helsinki as revised in 1983 and approval was obtained by the local Ethics Committee (reference number 507/16-6-2019).All subjects provided written consent for anonymous use of their data after a detailed explanation of the purpose and nature of the research protocol.At the initial visit, medical and reproductive history of FHA women and controls was noted.A clinical examination was performed by the same physician (E.S.).

Diagnosis of FHA
The diagnosis of FHA was established by the presence of secondary amenorrhea for 3 months or more, followed by a negative urinary pregnancy test, serum Follicular Stimulating Hormone (FSH) <10 mIU/ml, serum Luteinizing Hormone (LH) <10 mIU/ml, serum FSH/LH ratio: ~1, serum Estradiol (E2) <50 pg/ml and the exclusion of other diagnoses [1].The latter was verified by normal serum values of Prolactin (PRL), Thyroid Stimulating Hormone (TSH), Total Testosterone, Dehydroepiandrosterone Sulfate (DHEA-S), 17-Hydroxyprogesterone, and Sex Hormone Binding Globulin (SHBG), followed by a normal Magnetic Resonance Imaging (MRI) of the pituitary gland and a normal transabdominal or transvaginal ultrasound (uterus of normal size and volume according to age without anatomical malformations, and presence of two ovaries of normal volume without polycystic morphology indicative of PCOS as defined by the Rotterdam 2003 criteria [10]).

Anthropometrics and measurements
For all women, weight was measured to the nearest 0.1 kg using a beam balance (Beam Balance 710; Seca), while height was measured to the nearest 0.1 cm using a stadiometer (Stadiometer 208; Seca, Hanover, MD).The mean value of two measurements for each subject was used.Body mass index (BMI) was used as a marker of obesity.Women with values of 18 to 24.9 kg/m 2 were defined as normal, while those with 25-29.9kg/m 2 and ≥30 kg/m 2 , as overweight and obese, respectively.
Diurnal saliva was collected by using the Salivette Kit (Sarstedt company, Nunbrecht, Germany).Written and verbal instructions for use were given to patients prior to the procedure.The first sample was taken at 8:00 in the morning (at rest -at awakening), the second at 16:00 in the evening (at rest without employing any activity), and the third sample at 00:00 at night (at rest -at bed time) of the same day.Salivary cortisol was measured by using electrochemiluminescence immunoassay on a COBAS e411 analyzer (Roche Diagnostics, Mannheim, Germany) with inter-assay CV ≤ 5%, intra-assay CV ≤ 5% and sensitivity 1.5 nmol/l).

Questionnaires
We used the following validated self-questionnaires in order to assess the severity of anxiety, depression, eating disorders, sleeping disturbances and deterioration of patients' activities: a.The Eating Attitudes Test (EAT-26) which is a 26item self-report questionnaire, whereby participants indicate how frequently each statement "would be true for you" using a 3-point scale [range: 0 ("never or not applicable"), 1 ("always"), 2 ("very often"), 3 ("some time, not so often")].ΕΑΤ-26 consists of 3 sections relating to diet, bulimia and oral control.Scores ≥20 are indicative of an eating disorder (i.e., lower food intake, anorexia nervosa etc) [11].b.The Athens Insomnia Scale (AIS) which is intended to record persons' own assessment of any sleeping difficulty, they may have experienced over the last month.The scale consists of 8 items, scored from 0 to 3 [12].c.The International Physical Activity Questionnaire (IPAQ) that includes seven questions measuring the time of a subject's devotion to physical activity in the last 7 days.It estimates the degree of activities during work, on the go, at home or in the garden, as well as in the spare time [13].Data collected with the IPAQ long form can be converted to a continuous measurement and reported as metabolic equivalents (METs).Median values and interquartile ranges can be computed for walking (W), moderateintensity activities (M), and vigorous-intensity activities (V) within each domain using the formulas below: HADS-A: Anxiety scale and HADS-D: Depression scale) which estimates the level of anxiety and depression through fourteen questions regarding feelings during the previous week [15].
All four questionnaires have been validated for a Greek population.

Statistical analysis
Statistical analysis was performed by using SPSS (SPSS software; SPSS Inc, Chicago, IL, USA).Quantitative variables were compared with student's t-test or Mann-Whitney U test for normally and non-normally distribution, respectively.Categorical variables were compared with corrected Chi-squared test or two-sided Fisher's exact test, as appropriate.The relationship between parameters was estimated by using the nonparametric Spearman's correlation coefficient.All tests were two sided and p < 0.05 were considered to be significant.

Serum and Salivary cortisol
No significant differences were found with respect to morning serum cortisol and ACTH between women with FHA and the control group (Table 2).Considering SC, women with FHA had significantly higher morning levels compared to control subjects (21.8 ± 5.9 nmol/l vs 12.8 ± 3.8 nmol/l; p < 0.001), whereas the SC levels measured in the afternoon or at night did not significantly differ between the two groups (7.4 ± 3.9 vs 6.1 ± 4 nmol/l; p = 0.382 and 2.9 ± 1.9 vs 2.8 ± 1.1 nmol/l; p = 0.766, respectively) (Table 2 and Fig. 1).

Discussion
The suppression of the hypothalamic-pituitary-ovarian (HPO) axis plays a crucial role on the pathogenesis of FHA.A functional disruption of the pulsatile hypothalamic gonadotropin-releasing hormone (GnRH) secretion has been documented, leading to reduced gonadotropin secretion, absent midcycle surges of LH release, absence of normal follicular development, anovulation, and low serum E2 concentrations [16].Variable neuroendocrine patterns of LH secretion have been reported, including reduced frequency and/or amplitude of LH pulses.On the other hand, serum concentrations of FSH are usually low or normal, and often exceed those of LH, similar to the pattern observed in prepubertal girls [17].
The hyperactivity of the hypothalamic-pituitary-adrenal (HPA) axis seems to contribute to the development of FHA, as an increase in corticotropin-releasing hormone (CRH), in response to stress, has been proposed to be implicated in the GnRH suppression [18].Except from the psychological stress, eating disorders, weight loss, and excessive exercise, may also trigger CRH excretion [1].The increased levels of cortisol found in women with FHA, confirm this hyperactivity of HPA [19].Moreover, Lindahl et al. has shown an enlargement of the adrenal glands in women with FHA as a result of the increased adrenal responsiveness [20].In addition, a blunted cortisol response to CRH administration has also been reported [21].Interestingly, albeit hypercortisolemic, women with FHA seem to maintain the cortisol circadian rhythm.Previous studies demonstrated an increase in cortisol levels at night and early morning, but within the normal range [1].An increase in basal cortisol secretion or altered pulsatility of cortisol release has also been reported [7,8,22].
In our study women with FHA had significantly lower E2, as well as a trend towards lower LH, as expected.Furthermore, they had significantly lower TSH and a trend towards lower levels of FT4 compared to normally menstruating individuals, findings that are in agreement with those of previous studies, which showed alterations in the production of TSH and thus in the thyroid hormones concentrations in women with FHA [16,23].
Regarding morning serum cortisol, we found higher levels in the FHA compared to the control group, though the difference did not reach the statistical significance.However, there are several concerns about the sensitivity and specificity of serum cortisol as an indicator of the HPA hyperactivity [6][7][8].Alternatively, the measurement of SC has been currently promoted as more accurate, as it seems to express better the free component of the circulating cortisol [9,19].Nonetheless, to the best of our knowledge, there are no studies evaluating the SC in women with FHA up to now.According to our results, SC concentrations had a significant positive correlation with serum cortisol levels.More importantly, SC concentrations were significantly higher in FHA patients compared to control subjects but only in the morning, not in the evening or at night.In accordance with the studies already published, we confirmed the unaffected circadian cortisol rhythm in women with FHA, regardless of the increased concentrations of the morning SC.
Our findings confirmed the implication of stress, increased physical activity and eating disorders on the development of HPA hyperactivity and the overproduction of cortisol [24][25][26].Hence, women with FHA had significantly higher HADS-A score and higher EAT-26, though the latter did not reach the statistical significance.In addition, more FHA women exercised ≥3 times per week compared to healthy individuals, while a significantly higher proportion of them developed an EAT-26 score ≥20, which represents an increased possibility of eating disorders.Interestingly, higher HADS-A and EAT-26 scores and lower BMI, significantly correlated with higher morning SC concentrations.The connection between an abnormal autonomic response to stress and increased cortisol concentrations in FHA compared to eumenorrheic or PCOS women had been shown by Gallinelli et al. [27].Subsequently, Sanders et al. mentioned that moderate exercise leads to hypercortisolemia in FHA, supporting the hypothesis that these women demonstrate an altered response to moderate stress (stressful stimuli) [28].
We did not find a statistically significant difference between the two groups regarding the sleep disturbances, as was previously shown by our group.Tranoulis et al., demonstrated that women with FHA were twice as likely to have an abnormal score on the AIS questionnaire, and factors such as awakening during night time, final awakening and sleepiness during the day had been the most affected, in comparison to healthy individuals.In the current study, the lack of statistically significant difference between the two groups regarding the questionnaires related to sleep disturbances, is probably due to the smaller number of participants [29].
We believe that our study has many strengths.This is the first time that SC was evaluated in women with FHA.The procedure was easy to be performed and overcame the need of repeated blood measurements.Moreover, we used a sensitive and specific assay, not influenced by crossreactivity issues that commonly implicate the antibodybased direct quantification methods [30].The limitation of our study is the small sample size.However, the aim of this pilot study was to extract some primary results about the efficacy of SC to successfully differentiate women with FHA from healthy subjects.Interestingly, despite the small number of participants, SC was found to be superior to serum cortisol and showed a significant correlation with factors contributing to FHA.Obviously, after these encouraging primary findings, we are going to conduct a larger study in order to validate the accuracy of SC and determine the best cut-off value for the diagnosis of FHA or for the restoration of normal menstrual cycles.

7 FHAFig. 1
Fig. 1 Alterations of Salivary Cortisol in women with FHA and healthy subjects through a 24-h period of time

Table 1
Demographics and differences in hormonal levels between women with FHA and control subjects

Table 2
Serum and Salivary Cortisol between the two groups Control group Women with FHA p value

Table 3
Differences in psychological tests between women with FHA and control individuals