Are psychosocial factors and menopause symptoms related to hair cortisol concentrations among midlife Latinas? Results from the Menopausia, Salud, Corazón pilot study

Abstract

Objective

To explore the relationship between psychosocial factors, menopause symptoms, and hair cortisol concentrations (HCC) among midlife Latinas in a community-engaged pilot study.

Methods

A cross-sectional analysis was conducted using baseline data of 45 midlife Latinas enrolled in Menopausia, Salud, Corazón, a community-based cardiovascular disease risk reduction pilot study. Participants completed self-report measures to assess psychosocial factors (e.g., perceived stress, everyday discrimination, resiliency) and menopause symptoms. Cortisol was extracted from hair samples using Salimetrics kit 1-3002 and ran in duplicates (pg/mg). HCC were log-transformed and linear regression models adjusting for age and education were performed to explore the effects of psychosocial factors and menopause symptoms on HCC.

Results

Participants were on average aged 47.4 ± 4.5 years, 48.9% had a high school education or higher, and 98% were born outside the continental United States. The median HCC was 4.3 pg/mg (2.7- 8.2 pg/mg). After adjusting for age and education, experiencing any vasomotor symptoms (VMS) in the previous two weeks was associated with higher HCC (β[SE]: 0.59[0.26], p = 0.03). When additionally adjusting for perceived stress, there was a suggestive association between VMS and HCC (β[SE]: 0.49[0.27], p = 0.08).

Conclusions

Hair cortisol may capture aspects of chronic stress exposure, although its interpretation remains complex. Few studies have assessed HCC in midlife women, particularly Latinas. Though exploratory, our findings suggest a possible association between VMS and HCC that should be studied further. Future research with larger sample sizes and multiple time points is necessary to confirm these results and explore the use of hair cortisol in menopause research.

Clinical trial registration

Prospectively registered, NCT04313751 (03/19/2020), Protocol version 1.0.

Highlights

• •Study contributes data on HCC among midlife Latinas, an understudied population.
• •VMS were associated with higher HCC, but not after accounting for perceived stress.
• •Findings underscore the need for larger, longitudinal studies.

1. Introduction

Cortisol, a biomarker of hypothalamic-pituitary-adrenal (HPA) axis activity that is commonly associated with physiological responses to stress [1], may increase during the menopause transition (MT) [[2], [3], [4]]. Rising cortisol during this window has been associated with reproductive hormone fluctuations [[3], [4], [5]], poor sleep quality [6], physical inactivity [7], depressive symptoms [8,9], obesity [[10], [11], [12]] and higher levels of psychosocial stress [4,10]. Prior research has linked cortisol dysregulation to psychosocial stress and menopause-related symptoms, including vasomotor symptoms (VMS) [[13], [14], [15], [16]].

Hair cortisol concentration (HCC) is an emerging measure of cumulative HPA axis activity that captures longer-term cortisol exposure and may provide insight into physiological stress processes over time [1,2,10,12,17,18]. Unlike salivary or serum cortisol, which reflect short-term fluctuations and require multiple time-sensitive samples [19], HCC offers a less burdensome approach to assessing chronic cortisol levels. In addition, serum cortisol may be invasive and pose additional logistical and financial concerns [20]. Since hair grows at a relatively constant rate at approximately 1 cm per month [21], the assessment of HCC may be a more consistent reflection of total HPA activity over the preceding months [1,2]. Thus, HCC may be a valuable tool to measure long-term concentrations of cortisol during the MT.

Midlife Latina women may be particularly vulnerable to stress-related physiological dysregulation, as they report higher levels of psychosocial stress compared to other populations [22], including exposure to discrimination and socioeconomic stressors [23,24]. These stressors may influence neuroendocrine pathways, including HPA axis activity, and in turn contribute to both psychological and physical health outcomes in adults [13,25,26]. Stress has been related to systemic inflammation, impaired sleep, and a worse diet [10,27]. In addition to perceived stress, other psychosocial factors such as discrimination and resilience may also play important roles in shaping stress-related physiological responses. Perceived discrimination represents a chronic sociocultural stressor that has been linked to dysregulation of the HPA axis, while resilience may buffer the physiological effects of stress [10]. However, these psychosocial factors have been less frequently examined in relation to HCC, particularly among midlife Latina women. Moreover, despite growing interest in cortisol as a biomarker of stress, few studies have examined HCC in Latina populations, and even fewer have explored how psychosocial factors and menopause symptoms jointly relate to HCC. Therefore, the purpose of this exploratory study was to examine the associations between psychosocial factors (perceived stress, discrimination, and resilience), menopause symptoms, and HCC among midlife Latina women. Findings from this study may be useful for investigators considering the use of HCC in their research. In addition, this study can provide a better understanding of biological measures of stress and their associated factors among Latinas during the MT.

2. Materials and methods

2.1. Study design

A cross-sectional analysis was conducted using baseline data from participants enrolled in Menopausia, Salud, Corazón– a pilot study designed to reduce cardiovascular disease (CVD) risk in midlife Latinas. The design and procedures of this pilot study have been previously published [28]. Briefly, this clinical trial was a two-group, randomized repeated measures pilot study to evaluate the feasibility and initial efficacy of a behavioral intervention consisting of 12 weekly sessions of nutrition education [[29], [30], [31]], physical activity [[32], [33], [34]], and coping skills training [35] followed by 3 monthly sessions of continued support. Data were collected at three time points (baseline, 6-month, 12-month) and included an interviewer-administered questionnaire as well as a clinical assessment.

2.2. Sample

Women were recruited between July 2020 and June 2022 from community settings including local businesses, churches, clinics, health fairs, and community-based organizations in North Carolina. Bilingual (English, Spanish) research assistants (RAs) and community health workers facilitated recruitment efforts. In-person and online strategies were also used such as posting flyers in public spaces, attending community events and meetings to share information about the research study, and social media posts. Additional recruitment strategies applied during the COVID-19 pandemic have been published [36].

Eligible participants were women aged 40-60 years who self-identified as Hispanic/Latina, able to understand spoken English or Spanish, were perimenopausal or early postmenopausal (change in menstrual cycle duration or inter-bleeding interval, or no menstrual bleeding in the past 3–24 months), and reported no menopause hormone therapy use in the preceding three months. Women were excluded if they were pregnant or had a prior history of CVD (i.e., heart attack, stroke, coronary heart disease, heart failure). Forty-nine women were enrolled in the pilot study of which 96% (n = 47) agreed to HCC. Two outliers (HCC ≥100 pg/mg) were excluded. Thus, the current study included 45 participants.

2.3. Ethical considerations and trial registration

The Institutional Review Board of The University of North Carolina at Chapel Hill approved the research, and written informed consent was obtained from all participants in their preferred language (English or Spanish). The study was registered at the www.ClinicalTrials.gov registry (NCT04313751).

2.4. Procedures and measures

Data collection took place in private locations such as the Biobehavioral Laboratory of The University of North Carolina at Chapel Hill School of Nursing, the participant's home, or in a private room within a community-based organization familiar to the participant. Bilingual research assistants (RAs) were responsible for collecting participant data, preferably in the morning following an overnight fast (8-12 h). Procedures were performed in the participant's preferred language and questionnaires were available in English and Spanish.

2.4.1. Hair cortisol concentrations

Hair cortisol was collected from the posterior vortex region of the scalp using standardized guidance [37]. Hair was parted horizontally, and a 3 cm sample (diameter about half a pencil) was cut from the deeper layers of hair as close to the scalp as possible. Additionally, it was noted if participants had received any hair treatment, including bleaching, dyeing, or permanent alterations such as a perm or straightening treatment. Samples were packaged in aluminum foil with a string tied towards the proximal end, and stored in a laboratory refrigerator at −20°C until analysis.

HCC were measured at the Biobehavioral Laboratory of the University of North Carolina at Chapel Hill School of Nursing following a standardized protocol. The hair was first cut into three 1 cm segments, weighed, and washed with isopropanol (1.0 mL per 1 cm) on a rotator for 5 min. After drying, the hair was ground to a fine powder in a mixer mill (Retsch MM400; Retscch GmbH, Germany) at 29-30 scale for 25 min. Cortisol was extracted from the ground hair using methanol and an enzyme-linked immunoassay (Salimetrics Assay kit 1-3002, Carlsbad, CA) was used to measure HCC [37, 38]. In brief, this assay has a sensitivity of <0.007 μg/dL, with a detection range of 0.012-3.000 μg/dL (#1-3002, 2024). Cortisol was run in duplicates and we report the average of the duplicates. The cortisol intra-assay % CV was <10% and inter-assay % CV was <15%. The results were expressed as pg/mg, with pg as the cortisol amount and mg as the hair amount. The hair amount is the hair weight after washing two times with isopropanol.

2.4.2. Psychosocial questionnaires

Interviewer-administered questionnaires were employed to collect information about psychosocial factors. These questionnaires included the Perceived Stress Scale-4 (PSS-4) [39], the Everyday Discrimination Scale [40], and the Brief Resilience Scale (BRS) [41]. The PSS-4 is a 4-item questionnaire that uses a 5-point Likert scale to assess the degree to which participants gauge certain situations as stressful in the past month [39,42]. The raw score was calculated by summing the responses, providing a range of 0 (low stress) to 16 (high stress) with good internal consistency (Cronbach's α = 0.82) [42]. The Everyday Discrimination Scale was used to assess the frequency of discriminatory treatment in the daily lives of participants (Cronbach's α > 0.74) [43]. Scores were summed, resulting in a possible overall score of 10 to 60, with higher scores indicating more frequently reported discriminatory encounters [40,43]. The BRS was used to evaluate the ability of each participant to bounce back or recover from stress [44]. Scores were summed and averaged, providing a range of 1 to 5, with higher scores indicating greater resilience [41]. The BRS has demonstrated sufficient reliability (Cronbach's α = 0.83; ICC = 0.69) and validity [41].

2.4.3. Menopause symptoms and sleep disturbances

Menopause symptoms were assessed using a questionnaire previously employed in the Study of Women's Health Across the Nation (SWAN) [45]. Women were asked how many days in the past two weeks they experienced vasomotor symptoms (hot flashes, night sweats), feeling down/depressed, irritability or grouchiness, feeling tense or nervous, forgetfulness, and frequent mood changes. For each symptom, women reported: not at all, 1–5 days, 6–8 days, 9–13 days, or every day. Due to the number of cell sizes with <5 participants, we dichotomized the frequency of each symptom as none versus any. Sleep disturbances in the past two weeks were self-reported using a separate 4-item questionnaire (trouble falling asleep, trouble staying asleep, early morning waking, restlessness) adapted from the Women's Health Initiative Insomnia Rating Scale [46]. Each item was scored using a 5-point Likert scale from 0 to 4, with higher scores indicating greater sleep disturbances.

2.4.4. Covariates

Covariates in this analysis included several sociodemographic factors. Participants self-reported age, Latino sub-group/nationality, and years living in the continental U.S. Participants reported their highest educational attainment (<high school; ≥high school) and financial strain (no difficulty paying for basic necessities; somewhat/very hard paying for basic necessities). Height was measured to the nearest 0.1 cm using a stadiometer, calibrated in 1/8-cm intervals. A Seca 874 scale was used to measure weight to the nearest 0.1 kg. Body mass index (BMI) was computed as weight divided by the square of height (kg/m²). RAs performed waist circumference measurements (at the umbilicus) with a Lufkin tape, following the protocol outlined in the Multi-Ethnic Study of Atherosclerosis [47]. The average of two consecutive waist measurements was used (cm).

2.5. Data analysis

Data were analyzed using SAS 9.4 (SAS Institute Cary, North Carolina). First, descriptive statistics (frequency, mean, standard deviations) were performed to summarize participant characteristics. Next, cortisol values were log-transformed for normality. Simple linear regression was performed to explore each independent variable in relation to HCC. Next, models were adjusted for covariates commonly associated with HCC (e.g., age, marital status, BMI, smoking). Age and education were selected because these two covariates yielded the highest adjusted R² when included in the models. Given that perceived stress and VMS were most strongly associated with HCC, to evaluate independent associations, the final model included perceived stress, VMS, age, and education. Model assumptions were evaluated using residual diagnostics to assess linearity, normality of residuals, and homoscedasticity; no major deviations were observed. All analyses were two-sided at a 5% level of significance.

3. Results

Table 1 presents participant characteristics. Participants were on average aged 47.4 ± 4.5 years, 48.9% had a high school education or higher, and 53.3% reported difficulty paying for basics. Nearly all participants (98%) were foreign-born, with an average of 18.3 ± 8.6 years residing in the U.S. In terms of psychosocial factors, mean scores were 5.1 ± 3.1 for perceived stress, 13.6 ± 4.5 for perceived everyday discrimination, and 20.6 ± 4.9 for the Brief Resiliency Scale. Over half reported experiencing VMS (66.7%), feeling tense or nervous (62.2%), forgetfulness (62.2%), and mood changes (64.4%) within the past two weeks. Participants were on average obese (BMI 30.4 ± 5.4) [48], with an average waist circumference of 94.8 ± 12.6 cm. The median HCC was 4.3 pg/mg, ranging from 2.7 pg/mg to 8.2 pg/mg. Scatterplots and boxplots illustrating associations are presented in the Supplementary Material.

Table 1.

Participant characteristics.

Variable	N = 45
Age, mean (SD)	47.4 (4.5)
Education, n (%)
Less than high school education	22 (48.9)
≥ high school education	23 (51.1)
Years in the U.S, mean (SD)	18.3 (8.6)
Difficulty paying for basics, n (%)
Not very hard at all	17 (37.8)
Somewhat/very hard	24 (53.3)
Refused	4 (8.9)
Body mass index kg/m2, mean (SD)	30.4 (5.4)
Waist circumference (cm), mean (SD)	94.8 (12.2)
Perceived stress scale (10-50), mean (SD)	5.1 (3.1)
Everyday discrimination scale (10-60), mean (SD)	13.6 (4.5)
Brief resilience scale (6-30), mean (SD)	20.6 (4.9)
Menopause symptoms in past 2 weeks
Vasomotor symptoms	30 (66.7)
Feeling blue, down, depressed	20 (44.4)
Irritability or grouchiness	22 (48.9)
Feeling tense or nervous	28 (62.2)
Forgetfulness	28 (62.2)
Frequent mood changes	29 (64.4)
Overall sleep (0-16), mean (SD)	6.8 (4.7)
Cortisol (pg/mg)	4.3 (2.7, 8.2)

Note. cm = centimeters; kg = kilograms; pg/mg = picogram per milligram; SD = standard deviation.

Table 2 presents findings from unadjusted and adjusted linear regression models. In unadjusted models, VMS was significantly associated with higher HCC (β [SE]: 0.63 [0.25], p = 0.01). There was inconclusive evidence on the association between perceived stress and HCC (β [SE]: 0.07 [0.04], p = 0.08). Similarly, there was a non-significant association between feeling tense or nervous and HCC (β [SE]: 0.45 [0.25], p=(0.08). After adjusting for age and education, reporting any vasomotor symptoms remained significantly associated with HCC (β [SE]: 0.59 [0.26], p = 0.03), and there was a non-significant association with perceived stress (β [SE]: 0.08 [0.04], p = 0.07). When we included perceived stress and VMS in the model, VMS was no longer significant (β [SE]: 0.49 [0.27], p = 0.08). No other variables were associated with HCC.

Table 2.

Factors associated with hair cortisol in midlife Latinas.

Variable	Model 1 β (SE)	P	Model 2 β (SE)	P	Model 3 β (SE)	P
Perceived stress scale	0.07 (0.04)	0.08	0.08 (0.04)	0.07	0.05 (0.04)	0.20
Everyday discrimination scale	0.03 (0.03)	0.27	0.03 (0.03)	0.28
Brief resilience scale	−0.03 (0.16)	0.84	−0.08 (0.17)	0.63
Menopause symptoms in the past 2 weeks (any versus none)
Vasomotor symptoms	0.63 (0.25)	0.01	0.59 (0.26)	0.03	0.49 (0.27)	0.08
Feeling blue, down, depressed	0.34 (0.24)	0.17	0.29 (0.25)	0.26
Irritability or grouchiness	0.04 (0.25)	0.88	0.006 (0.25)	0.98
Feeling tense or nervous	0.45 (0.25)	0.08	0.39 (0.27)	0.15
Forgetfulness	0.39 (0.25)	0.12	0.36 (0.26)	0.18
Frequent mood changes	0.19 (0.26)	0.46	0.20 (0.27)	0.47
Overall sleep	0.03 (0.03)	0.34	0.03 (0.03)	0.44
Covariates
Age	--	--	0.01 (0.02)	0.58	0.01 (0.02)	0.60
Education	--	--	−0.03 (0.08)	0.71	−0.04 (0.08)	0.65

Note. Factors were tested in separate models with log-transformed HCC as the outcome. Model 1 is unadjusted; Model 2 adjusts for age and education. Because vasomotor symptoms and perceived stress showed significant or inconclusive associations with HCC in Model 2, Model 3 additionally adjusts for perceived stress to assess whether the association between VMS and HCC is attenuated.

4. Discussion

This pilot study explored the association between psychosocial factors and menopause symptoms with HCC in Latinas during the MT. We found that worsening VMS over two weeks may be associated with a higher HCC. These findings could suggest an underlying aggregation of VMS pathophysiology that lead to compounded HCC over time. These findings are similar to those of other studies that have investigated VMS in relation to cortisol [16,49]. While most of these studies measured salivary or urine cortisol [3,[14], [15], [16]], one study reported that HCC was associated with an increased frequency of daily hot flashes using ambulatory and physiological data [3]. Our findings suggest a potential association between HPA axis activity and vasomotor symptoms, although the direction of this relationship cannot be determined. VMS are believed to result from altered thermoregulatory function in the hypothalamus, which is influenced by fluctuating estrogen levels during the MT [50]. There is evidence that increases in body temperature induce a typical stress response, with the secretion of catecholamines and cortisol, which induce thermoregulatory mechanisms to maintain internal temperature homeostasis [50]. Though findings were not significant when adjusting for perceived stress, they remained pronounced and serve to suggest a possible association that merits further investigation.

It is important to note that perceived stress was not significantly associated with HCC after adjusting for age and education; however, the direction of the association was positive. Perceived stress was higher in our sample than previous studies [10,51], which may be due to a combination of factors regarding our study sample: mostly immigrant Latinas; 53.3% experiencing financial strain; and our study was conducted during the COVID-19 pandemic when people were experiencing higher levels of social isolation and stress. It is also possible that because the PSS-4 measures stress over 4 weeks, and not a 12-week period, it reflects the average cortisol accumulation in hair samples at cross-sections; thus, longer self-report measures for chronic stress should be used to understand the relationship between stress and HCC better. Additionally, other studies have supported these associations between perceived stress and HCC among minoritized groups when controlled for financial strain [52]. Curiously, despite no association with HCC and psychosocial factors, participants in our cohort reported higher perceived everyday discrimination than previously reported literature in other minoritized populations [53].

Participants in our study all self-reported Latina ethnicity; however, participants were racially diverse with varied hair textures. Hair cortisol has been previously shown to vary by hair type, including hair color and texture [10,54]. For example, Moody et al. found that Latina participants with tightly curled hair had higher cortisol levels when compared to more wavy or straight hair [54]. Furthermore, hair type may impact individual preferences in frequency of hair washing, types of hair treatment, and the products utilized that may or may not impact cortisol binding protein at a molecular level [54]. Other studies have theorized that darker hair may have protective factors [55], although this has been contested in other cohort studies [56]. Future studies should consider the impact of hair type, particularly hair texture, including on cortisol concentration.

4.1. Strengths and limitations

There are several limitations and strengths to consider in this analysis. This is an exploratory study with a small sample size, largely conducted in the context of COVID-19 restrictions, which hindered recruitment for the study at large. Given the pilot nature of this study, we conducted a post hoc power analysis based on the observed association between vasomotor symptoms and hair cortisol concentrations. With an estimated small-to-moderate effect size (Cohen's d ≈ 0.34) and a sample size of 45, statistical power was approximately 60% at α = 0.05, indicating limited power to detect modest associations. However, to our knowledge, this is currently the largest ethnically-Latina cohort living in the U.S examining HCC using a Salimetrics assay and the only one examining midlife women [57]. Another limitation of this study is the lack of data on hair-washing practices, which may influence HCC [58]. Additionally, given the cross-sectional design of HCC collection, it is unknown if perceived stress results in worsening menopausal symptoms or vice versa; however, we included perceived stress and VMS in a model to evaluate independent associations. Another limitation is that we were exploring multiple independent variables, increasing the likelihood of Type I error. Nevertheless, our study was exploratory aiming to identify patterns to generate hypotheses for future research utilizing HCC methodology in this population.

The use of HCC was a strength in this investigation since it may reflect consistent cortisol concentration more accurately than urinary or salivary collection [1]. The latter measures may be impacted by sleep patterns, poor adherence to constant sampling, and biological limitations. While all participants (n = 45) denied hair treatment, a potential limitation of hair cortisol is the impact of hair treatments, frequency of hair washing, and chemical exposure [1]; however, the impact of these behaviors has been contested by other longitudinal studies [55,56,59]. As indicated, HCC was obtained using standardized protocol [37] with additional sources indicating no variation in HCC measurements based on where sampled is sourced [60]. As is best practice, samples were tested within a one-year window prior to any potential breakdown in analyte stability [60]. The immunoassay kit utilized in this study has popular use among other studies for hair cortisol measurement [52,61,62]; however, this kit is not validated for hair sampling (#1-3002, 2024). The distribution of HCC among this cohort was comparable to other women of heterogenous ethnicities in previous studies [51]. To our knowledge, this is the first study to examine hair cortisol in an exclusively midlife Latinas navigating the MT [13,25].

5. Conclusions

This study demonstrated the feasibility of measuring hair cortisol concentrations in midlife Latinas. We found a suggestive association between VMS and HCC that should be further evaluated with temporally aligned measures. Although not significant, this is an important finding supporting a potential biological link between stress and VMS. Future longitudinal studies in larger samples are necessary to confirm these results and examine associations between menopause-related factors and HCC.

Disclaimer

The content of this manuscript is solely the responsibility of the authors and does not necessarily represent the official views of the NIMHD, NIH or the Robert Wood Johnson Foundation.

Funding

This study was funded by the National Institutes of Health and the National Institute of Minority Health and Health Disparities (K23MD014767, PI: Cortés). At the time of submission, Latesha K. Harris, PhD, RN, was supported by the Robert Wood Johnson Foundation Health Policy Research Scholars Program and the National Institute on Nursing Research (T32NR007091; PI: Santacroce & Kneipp). A version of this manuscript was presented at the 2022 Menopause Society Annual Meeting.

CRediT authorship contribution statement

Christian A. Long: Conceptualization, Methodology, Writing – original draft, Writing – review & editing. Valentina Marginean: Methodology, Writing – original draft, Writing – review & editing. Latesha K. Harris: Data curation, Writing – original draft, Writing – review & editing. Chongben Zhang: Methodology, Writing – original draft, Writing – review & editing. Yamnia I. Cortés: Conceptualization, Data curation, Formal analysis, Funding acquisition, Methodology, Supervision, Writing – original draft, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article: