Abstract
Hair cortisol concentration (HCC) is a promising biomarker for chronic stress, yet its relationship with subjective distress and mental health remains unclear. Existing research has largely focused on group-level data over relatively short time frames. In the present longitudinal case study, we tracked HCC and subjective measures in a female PhD student across nine years (2014–2023), including depression severity, trait anxiety, resilience, coping strategies, and life events. Her data were compared with HCC, depression, and anxiety levels in a clinical outpatient sample (N = 33) and normative HCC data. Despite reporting numerous stressful life events, the PhD student’s HCC levels remained stable and comparable to the clinical sample, and the norm, with the exception of for a marked elevation during the first month of the PhD (HCC = 12.18 pg/mg; p < .001). She showed exceptionally high resilience (M = 4.7, p < .001) and use of problem-focused coping strategies (PR = 97.9), and low avoidant coping (PR = 28.2). In the clinical sample, HCC was not associated with diagnosis, depression severity, or trait anxiety. These findings suggest that HCC may not serve as a straightforward biomarker of psychological distress or reliably differentiate between clinical and non-clinical individuals. Instead, HCC appears to reflect a complex, context-dependent measure influenced by methodological and behavioral factors.
1. Introduction
In 2010, Insel and colleagues announced “[…] a new classification framework for research on mental disorders”, the Research Domain Criteria (RDoC) project, which included a strong emphasis on a neuroscientific approach to characterize dysfunctional manifestations of biopsychological processes underlying specific symptoms of mental disorders on different system levels of human behavior (Insel et al., 2010). To address the shortcomings of categorical diagnoses in favor of a descriptive approach to describe normal and abnormal functioning in different mental health domains, it is essential to understand the associated physiological mechanisms (Cuthbert & Insel, 2013; Insel et al., 2010). Nevertheless, the boundaries between adaptive and destructive processing are fluid and should be defined considering individual-specific and complex conditions.
A good example is the concept of psychosocial stress and its diverse effects on individual functioning, ranging from beneficial eustress to pathogenic distress (Selye, 1976). Thus, assessing stress over multiple domains, including biomarkers and self-report, might be promising not only to characterize individual responses and their adaptivity, but also to tailor further interventions in cases of high dysfunctionality (Steffen, 2023). Acute stress is typically short-lived, and arises from immediate and specific stressors. It activates the body’s stress response system, leading to the release of stress hormones, such as cortisol and norepinephrine. Acute stress can temporarily enhance cognitive performance and motivate attention (Schwabe et al., 2012; Weymar et al., 2012). However, when stress becomes overwhelming or occurs frequently, it can contribute to the development of mental and physical health problems (Cohen et al., 2007; McEwen, 2007). Chronic stress can result from ongoing stressors, such as chronic health conditions, financial difficulties, relationship problems, childbirth, or marriage, irrespective of event valence. It increases the risk of developing various mental disorders, including anxiety and depression (McEwen, 2007).
One elegant and non-invasive strategy for measuring chronic stress is the analysis of hair cortisol concentrations. Hair grows approximately .35 mm a day, which corresponds to approximately 1 cm per month, and free cortisol from the body is incorporated into the growing hair fibers (Greff et al., 2019; Webb et al., 2015). Thus, a 1 cm hair segment reflects approximately one month of cortisol activity (Stalder et al., 2017; Webb et al., 2015). In contrast to saliva, plasma, or urine cortisol (up to 24 h sampling), far longer time periods can be covered, and analyses are independent of diurnal hormone changes. Interestingly, cortisol analyses have been successfully used in archeological hair (Webb et al., 2015), indicating the possibility of longer storage times (but, see Abell et al., 2016; Huthsteiner et al., 2025; van den Heuvel et al., 2020). In Peruvian archeological hair samples (Nasca period; 1-1000 AD), mobility and dietary changes, as inferred from isotopic analyses, were associated with higher hair cortisol levels, suggesting that travel and food insecurity were the major life stressors in this sample. In addition, there was evidence of trauma resulting from interpersonal conflict (e.g., head injuries; Webb et al., 2015).
In current clinical samples, data show higher hair cortisol levels in depression and breast cancer survivors, whereas individuals with posttraumatic stress disorder and generalized anxiety had lower hair cortisol levels than healthy controls (Steudte et al., 2011, 2013; Wirkner et al., 2017; for reviews, see Staufenbiel et al., 2013; Koumantarou Malisiova et al., 2021). In contrast, other studies reported higher hair cortisol to be associated with more generalized anxiety disorder symptoms (Lewis et al., 2025). And, in another study, individuals with generalized anxiety disorder were found to have HCC levels comparable to healthy participants, but higher than in the depressed subsample, concluding that major depression might be rather associated with attenuated cortisol secretion over time (Steudte-Schmiedgen et al., 2017). Besides age, sex and sex hormones have been discussed as a possible confounder in HCC research (Dettenborn et al., 2012). However, there was no difference when controlling for participant’s gender (Lewis et al., 2025; Steudte et al., 2011) or hormone status (Steudte et al., 2013).
Given the mixed results from between-group comparisons, the question arises as to whether hair cortisol might be a biomarker for anxiety and depressive symptom severity in the individual, especially if the impact of high-stress experiences under conditions of functional processing is explicitly considered in the long-term. Therefore, our objectives were to follow an individual with an elevated risk of stress due to psychosocial living conditions but with good mental health, over an extended time period, and to examine hair cortisol in patients affected by mental disorders.
A PhD student was chosen because the PhD phase, especially the time before defense, was found to be a prolonged period of elevated stress (Van Laethem et al., 2017). For example, subjective stress ratings, but also overall pleasant anticipation increased until the PhD defense, followed by a fatigue after the ‘great day’ (Van Laethem et al., 2017). Research suggests that PhD students are at a 2.4 higher risk of developing mental disorders than the general population (Levecque et al., 2017). In addition, graduate students who completed their clinical year during psychotherapy licensing reported high stress levels (Engel et al., 2015), paralleling medical education (Dyrbye & Shanafelt, 2016; Meeks et al., 2019). Moreover, German psychotherapy trainees reported higher subjective stress than undergraduate controls and the general population (Engel et al., 2015). In contrast, doctoral psychology students in the U.S. did not report higher stress levels compared to the general population but reported higher emotional exhaustion in their third and fourth years of their seven-year education (Rico & Bunge, 2021). Usually, the PhD phase coincides with late emerging adulthood, a transition phase that can be challenging (Arnett, 2000; Arnett et al., 2014). Another more practical reason to follow a PhD student was the availability over such a long period of time and consent to provide enough hair samples that were kept constant for treatment (e.g., no dyeing).
Thus, in the present study, we (1) monitored a PhD student over a nine-year period and assessed self-report data (stressful life events, resilience, depression, and anxiety) and hair cortisol concentrations, (2) measured HCC and self-report in an outpatient sample with categorial diagnoses, and (3) related the individual PhD student’s HCC trajectory to these clinical data and additional laboratory reference values.
We expected the PhD student to exhibit phases of elevated stress, along with elevated HCC levels. Differences in HCC levels were expected in patients with primary depressive disorder compared with individuals with primary anxiety diagnoses. Under the premise of good mental health (i.e., no depressive and anxiety symptoms), the PhD student’s HCC was expected to be in the normative range of the reference sample and, given the mixed findings for clinical groups, eventually different from patients with anxiety disorders.
2. Material and Methods
2.1. Participants
Sociodemographic and Self-Report Data
*PR = percentage ranks in a healthy control sample (N = 316; Poulus et al., 2020); BDI = Beck Depression Inventory, STAI-T = State-Trait Anxiety Inventory (Trait version), BRS = Brief Resilience Scale.
2.1.1. PhD Student
The PhD student consented to providing sufficient material for HCC sampling and not undergoing hair treatment (e.g., dyeing, permanent waves) during the study period. Hair sampling was reduced to six measures to minimize load and respect individual demands. At the first hair sampling, the healthy PhD student was 24 years old. She reported no current or lifetime mental disorder and no current medical condition or medication intake (see below), which was also the case during follow-ups. Parallel to her PhD, she started working on psychotherapy licensing (postgraduate training in October 2014). This resulted in multiple changes to a different line of work during the PhD phase, mostly due to parallel contracts (research institution, psychiatric hospital, outpatient psychotherapy clinic).
2.1.2. Psychotherapy Outpatients
Of the 105 patients in the outpatient psychotherapy clinic of the University of Greifswald who were included in a larger psychotherapy study (Lang et al., 2018; M age = 37.2 years; 68% female), 34 hair samples were available for analyses (data not published before). One patient was excluded from the analyses because of notably high HCC levels outside the range of the values of the other patients (S1: 834.6 pg/ng; S2: 454.7 pg/ng, S3: 356,2 pg/ng; S4: 365.3 pg/ng), resulting in a final N of 33 (M age = 38.1 years, SD = 13.2, Range 19-68; 74.8% female). 1
Six patients (17.6%) had been diagnosed by trained clinicians (Standardized Diagnostic Interview for Mental Disorders; Margraf et al., 2017) with personality disorder, 8 (23.5%) with depression, 8 (23.5%) with anxiety disorder, and 12 (35.3%) with comorbid depression and anxiety. Fifteen (44%) individuals had a principal diagnosis of anxiety disorder and 13 (38%) had a principal diagnosis of depression.
Initially, from N = 105, only 49 (46.7%) individuals agreed to provide hair samples, of whom 15 had insufficient material, which was < 4.5 mg hair per 1 cm segment. Twenty-two patients (90.9% male) were excluded because their hair was too short for sampling. Missing consent was due to cosmetic reasons (N = 18, 77.8% female), and 16 patients provided no reason. Individuals included in the present study did not differ from the original psychotherapy study sample in terms of age, depression severity (BDI-II), or trait anxiety (all ps < .05).
The final sample was also representative of the whole cohort of patients treated at the outpatient clinic in the study year 2017: 440 individuals (M age = 34.1 years, SD = 13.1; 75.7% female) that were mainly diagnosed with affective (N = 220; 50.0%), anxiety (N = 358, 81.4%), and personality disorders (N = 87; 19.8%), with N = 128 (21.1%) having at least two comorbid diagnoses.
2.2. Hair Cortisol Sampling and Analysis
Hair cortisol samples were collected according to the instructions provided by the Technical University (TU) of Dresden. 2 Each individual provided two hair strands from the occipital scalp (resulting in the required minimum total diameter of 3mm), and samples were stored at room temperature until shipping.
In the clinical sample, the hair was cut only once before outpatient treatment was started (along with the initial study diagnostics). For longitudinal analyses, PhD student’s hair was collected six times over the entire doctorate period (T1: November 14, 2014 – 5 months after the PhD period started; T2: February 5, 2016; T3: January 9, 2020; T4: December 17, 2020; T5: June 30, 2022; T6: June 9, 2023 – successful PhD defense). PhD hair was not colored or dyed, and washing frequency remained stable with three washes per week (no excessive swimming, only shampoo was used). Hair samples were analyzed in June 2017, aiming at comparable storage times. Samples collected from the PhD student after this date were analyzed in June 2023. The lower limit of detection for the protocol was 0.3 pg/mg for cortisol and the two sample packages were analyzed in a single run each. The intra-assay coefficients of variation were 9% and 12%, respectively. From each sample, the four proximal segments (each segment was 1cm length of the hair strand) were analyzed by the Dresden LAB Service (https://dresden-labservice.com; Gao et al., 2013), resulting in four hair cortisol concentrations (HCC) ranging from one to four months before sampling (S1 = 1st month, S2 = 2nd month, S3 = 3rd month, S4 = 4th month). Wash-out effects were kept constant in the PhD student, but no information on hair treatment was available in the clinical group.
2.3. Questionnaires
To measure depressive symptom severity and trait anxiety, the Beck Depression Inventory (BDI-II, revised German version) and the State-Trait-Anxiety Inventory (Trait version; STAI-T; Laux et al., 1981) were used. The BDI-II and STAI-T were included in the initial study diagnostic for patients and administered to the PhD student at T6. In addition, the PhD student completed the Social Readjustment Rating Scale, commonly known as the Holmes and Rahe Stress Scale (HRSS (Holmes & Rahe, 1967; Noone, 2017); at all time points to assess critical life events, and the Brief Resilience Scale (BRS (Smith et al., 2008); and the Brief COPE (Carver, 1997) for resilience and coping strategies at T6. In 2018 (TX), the PhD student provided the HRSS, but hair sampling was omitted for cosmetic reasons (wedding hairstyle, honeymoon). The original instruction for the HRSS included a one-year time span; however, given the limitations resulting from reliable HCC measures (retrospective assessment is not recommended for one year) and sampling frequency, the time span for life events was reduced to six months. The HRSS conceptualizes stress primarily as the amount of life change requiring psychological readjustment, independent of whether the events are perceived as pleasant or unpleasant (Holmes & Rahe, 1967; Thoits, 2010).
2.4. Statistical Analyses
Statistical analyses were performed using RStudio: Integrated Development for R Version 2024.12.1 (RStudio, PBC, Boston, MA, USA).
First, following the characterization of the PhD student’s self-report (a one-sample t-test was used for BRS) and hair cortisol data, HCC values were related to HRSS scores. Second, hair cortisol concentrations in the clinical sample were analyzed. To account for the typically skewed distribution of hair cortisol concentrations (HCC), values were log-transformed prior to analysis and differences in HCC across hair segments were examined using a one-way analysis of variance. Given the small sample size, only the mean HCC over all segments was then analyzed in terms of sex (male vs. female) and diagnostic differences (anxiety vs. depressive vs. comorbid anxiety and depressive disorder vs. personality disorder; principal anxiety vs. principal depressive disorder). Kolmogorov-Smirnov tests were applied to test for normality. The BDI and STAI-T scores were normally distributed in the clinical sample, but the total mean HCC (mean of all four segments) was not. Therefore, non-parametric statistics (Spearman’s correlation, Kruskal-Wallis) were reported for HCC analyses. Correlational analyses included age, BDI-II, STAI-T, and HCC. Third, to classify the PhD student in relation to the clinical sample (BDI-II, STAI, and age), one-sample t-tests were conducted, and the PhD student’s HCC was also compared to the clinical sample. Moreover, a descriptive comparison to normative data provided by the Dresden LAB Service (Kirschbaum, personal communication) is presented. 3 The reference values are based on analyses from over 10,000 HCC samples using the same protocol.
3. Results
(1) Characterization of the PhD student. With a BDI score = 6 (Hautzinger et al., 2006) and a STAI-T score = 41 (T-value = 56; Laux et al., 1981), the PhD student showed no signs of depressive or anxiety symptoms. The resilience of 4.7 was above all samples in the original BRS publication (p < .001; Smith et al., 2008). PhD student’s use of problem-focused coping strategies (Brief COPE) was at the upper limit (PR = 97.9), whereas that of avoidant coping strategies was very low (PR = 28.2). The percentage rank for emotion-focused coping was 58.1, with only humor reaching the maximum value (“I’ve been doing this a lot.”). The significant life events and HRSS scores of the PhD student during the months prior to each hair sampling are listed in Table 2. The PhD student’s HRSS sum scores were > 300 at T1, TX, T4 and T5, (299 at T3). According to a statistical prediction model, sum scores > 300 are regarded as an 80% risk of health breakdown within the following two years (Rahe et al., 1970). No health breakdown occurred during study participation. Holmes and Rahe Stress Scale Scores and Selected Life Events in the PhD Student
*HRSS scores > 30; Date format is MM-DD-YYYY.
Hair cortisol concentrations for the PhD student are shown in Table 3. HRSS scores in the PhD student were not associated with HCC (p > .05). Also, exploratory analyses did not indicate systematic differences in HCC across time or hair segments. However, given the single-case design and, thus, lack of independence, findings should be interpreted only descriptively. On a descriptive level, the initial HCC value (November 2014: HCC = 12.18), which marked the beginning of the doctorate period, was three standard deviations above the individual mean. The second-highest value (October 2015: HCC = 7.34), corresponding to a major professional transition (i.e., responsibility for a large multicenter randomized controlled psychotherapy trial), and the third-highest value (May 2023: HCC = 7.03), occurring shortly before the PhD defense, were also elevated (i.e., approximately two standard deviations above the mean).
Hair Cortisol Concentrations
Mean (SD) hair cortisol concentrations [pg/mg] for psychotherapy outpatients and the PhD student for all segments and time points.
(3) Relating the PhD student to clinical and normative samples. The PhD student’s BDI-II (T (32) = 8.8, p < .001, d = 1.5) and STAI values (T (32) = 6.3, p <. 001, d = 1.1) were lower than those in the clinical sample, and she was younger than the mean clinical sample (T (33) = 6.2, p <. 001, d = 1.1). Hair cortisol concentrations are visualized in Figure 1. PhD student’s HCC did not differ from that of psychotherapy outpatients, except for the first segment at T1 (HCC = 12.18 pg/mg; T (33) = 7.02, p < .001, d = 1.2), when the PhD phase started. The following reference values for females aged between 21 and 30 years were provided by Dresden LAB Service (Kirschbaum, personal communication): 5th percentile = 1.64 pg/mg; 50th percentile = 4.95 pg/mg; 95th percentile = 19.49 pg/mg (see Figure 1). Although statistical comparisons were not feasible due to missing further information, the present HCCs apparently ranged around the 50th percentile of the reference. Hair cortisol concentrations (HCC) [pg/mg] for psychotherapy outpatients (left) and the PhD student (right) for hair segments S1, S2, S3, and S4 corresponding to 1, 2, 3, and 4 months prior to sampling, respectively. The sampling times for the PhD students were T1, November 14, 2014; T2, February 5, 2016; T3, January 9, 2020; T4, December 17, 2020; T5, June 30, 2022; and T6, June 9, 2023. Percentiles for reference values are indicated by an upward arrow (5th percentile), a square (50th percentile), and a downward arrow (95th percentile), the normative sample provided by Technical University Dresden
4. Discussion
As expected, the PhD student experienced a high number of stressful life events during the nine-year period. On the descriptive level, HCC was highest during the first month of the PhD qualification phase. In addition, two further intraindividual HCCs were elevated relative to the individual mean: the beginning of a large multicenter randomized controlled psychotherapy trial in October 2015 (with visual inspection of the data suggesting somewhat prolonged stress in the following two months), and the last month before the PhD thesis defense. Intriguingly, hair cortisol levels were not statistically related to the overall high amount of other stressful life events (HRSS), including pregnancy (Khoury et al., 2023; Kirschbaum et al., 2009). This suggests that objectively defined stressful life events do not reflect subjective stressful experiences. However, previous findings regarding the relationship between self-reported stress and HCC are inconsistent (Stalder et al., 2017). For example, a recent longitudinal study in adolescents found only limited associations between psychosocial stress and HCC, highlighting the influence of stressor type (more severe stressors), timing and duration of stress exposure (Finlay et al., 2026).
One possible explanation is that the PhD student’s high resilience and functional coping may have buffered acute stress responses, potentially reflecting adaptive functioning of stress-regulatory systems (e.g., noradrenergic and hypothalamic–pituitary–adrenal axis processes; De Kloet et al., 2005). High levels of problem-focused coping strategies and (social) support, along with low avoidance, may have contributed to overall well-being (Chao, 2011; Schmidt & Hansson, 2018). Despite four HRSS scores above the critical value of 300 points (Holmes & Rahe, 1967), no health breakdown was observed. The severe lockdown phase of the COVID-19 pandemic may have contributed to the high number of stressful life events in T4 (December 2020). However, no specific pandemic-related items were assessed in this study. Overall, there is growing evidence that the COVID-19 pandemic can be regarded as a multidimensional stressor (Wirkner et al., 2022).
Recent studies investigating pandemic-related stress and HCC have yielded mixed findings (Broadbent et al., 2024; Ding et al., 2024). For example, whereas Broadbent et al. observed a decrease in HCC levels in 2020 (Broadbent, 2024), increases were found in other samples, including adults, children, and adolescents (Fung et al., 2022; Jia et al., 2023). Moreover, associations between HCC, negative life events, perceived stress, depression, and anxiety were also inconsistent in these studies, suggesting that sample characteristics and contextual factors play a substantial role.
Consistent with these findings, previous research indicates that subjective stress, resilience and HCC might be only weakly related. For example, García-León et al. (2019) found no differences in HCC between individuals with high versus low resilience despite differences in perceived stress. Similarly, Planert et al. (2023) reported no associations between subjective measures and HCC in a very large sample (N = 1258). In summary, HCC measures alone might not be well suited to reflect the complex interplay between stressful life events, perceived stress, and resilience, especially at the single-subject level. Moreover, the high number of stressful life events experienced resulted in rather little variance (and thus ceiling effects) in the HRSS scores, suggesting that it might not be an ideal instrument for such individuals. To better understand this complex interplay, future longitudinal studies should complement life-event measures with assessments of subjective stress experience and resilience (Thoits, 2010).
Consistent with the present results, a longitudinal study investigating medical interns found increased HCC levels at the beginning and end of an internship, suggesting that novelty and unfamiliar professional challenges might drive strong cortisol responses and that subsequent cortisol decline reflects accommodation processes (Mayer et al., 2018). In the case of the PhD student, the final HCC rise one month before the PhD dissertation may similarly reflect anticipatory cognitive stress related to the social-evaluative threat situation (Allen et al., 2017; Gaab et al., 2005; Van Laethem et al., 2017). However, it remains unclear whether HCC levels indeed reflect these novel challenges, given the complex interplay between diurnal cortisol secretion, cortisol awakening response, sleep, and chronotype (Engert et al., 2018).
Across studies, there is a high variance in HCC concentrations, possibly resulting from different sampling and analysis methods, thus limiting comparisons to absolute values reported in the literature (Stalder et al., 2017). In the present study, all HCC samples were analyzed in the same laboratory and ranged around the 50th percentile for a normative sample reported by this laboratory. Thus, although analysis consistency was ensured, HCC did not discriminate between the PhD student, the clinical sample, and the lab norm.
In contrast to previous studies, HCC levels did not vary with patient diagnosis. Individuals with primary anxiety diagnoses had somewhat lower HCC scores than those with primary depression diagnoses, which is in line with a previous study (Stalder et al., 2017). However, this finding failed to reach statistical significance, perhaps because of the small sample size. Similarly, no sex differences were observed, despite previous findings suggest higher HCC levels in men (Stalder et al., 2017, but see Steudte et al., 2011; Lewis et al., 2025; Finlay et al., 2026). Likewise, such contrasting findings (Finlay et al., 2026; Lewis et al., 2025; Steudte et al., 2011) are limited by the small sample size and the predominance of female participants.
In addition, neither anxiety nor depression severity was associated with HCC levels. Although limited by the small sample size, this leads to the strong assumption that individual HCC levels alone may not be suitable as diagnostic biomarkers for depression or anxiety. This is also corroborated by the mixed findings in other clinical samples (e.g., Steudte-Schmiedgen et al., 2017; Lewis et al., 2025; for review, see Koumantarou Malisiova et al., 2021). Moreover, both depression and anxiety severity were highly correlated in the present sample, and this comorbidity is common (Kessler et al., 2015). Perhaps the concepts of depression and anxiety are still too specific and markers of chronic stress might be better associated with the broader domain of negative valence or specific sub-processes, such as sustained defensive responding (Cuthbert & Insel, 2013; Insel et al., 2010). This points to the Lang model, in that unpleasant events trigger defensive motivation observable in fight, flight (or freeze) responses (Lang et al., 1997).
Recent findings further illustrate the complex relationship between cortisol and behavioral responses. For example, a study on PTSD symptoms among traumatized refugees found that trauma and post-displacement stressors were not related to HCC; however, lower HCC levels were associated with high levels of avoidance, whereas higher HCC levels were related to mood symptoms (de Graaff et al., 2024). In contrast, higher long-term cortisol levels predicted avoidance behavior following a single traumatic event (Petrowski et al., 2020). It is very likely that cortisol and avoidance behaviors are intertwined in a complex manner, considering the timing of the measurement and the frequency and intensity of the stressor (for acute stress, see Vogel & Schwabe, 2019). Future research should therefore integrate measures of avoidance and related processes to better understand these dynamics.
A major limitation of the present study is the small sample size and the frequency of hair sampling in the PhD student. In the outpatient clinical sample, only 32 % provided hair strands (once) with sufficient material, and continuous longitudinal sampling was more limited for several reasons (e.g., hair availability and consent). In the present study, the main reasons for refusing hair sampling were somewhat specific for males and females with too short hair and cosmetic reasons, respectively. Missing consent for or feasibility of hair sampling might be a major obstacle for using HCC as a biomarker (for cultural reasons, see Broadbent, 2024).
In addition, HCC is influenced by various factors, such as dying or washing frequency (which was held constant in the PhD student but likely varied within the clinical sample) or sunlight exposure (Stalder et al., 2017; Wester et al., 2016). In addition, there is evidence that hair growth rates differ, for example, between sex or race identities, and that stress per se might affect hair growth rate, making it difficult to precisely trace back time intervals (Greff et al., 2019; Stalder et al., 2017; but see Dettmer et al., 2023). It would be fascinating to see if new, but still uncommon, approaches that analyze long-term cortisol concentrations in the cerumen could provide an alternative for HCC measures (Herane-Vives et al., 2020).
Although HCC has been widely acknowledged as a biomarker for long-term stress, it is still debatable whether it reflects basal or response-based cortisol secretion (Kalliokoski et al., 2019; Sugaya et al., 2020). For example, Sugaya et al. (2020) reported moderate correlations only between HCC in the first segment of hair and 30-day salivary cortisol and replicated findings of intra-individual stability of HCC (e.g., Zhang et al., 2017) but did not find correlations with the magnitude of the cortisol awakening response and the diurnal slope. Unfortunately, the authors did not assess the major life events during sample collection (Zhang et al., 2017). In line with this, Kalliokoski et al. (2019) argued that HCC levels can only be linked to ongoing stress at the time of hair sampling and that they might not reflect stressful experiences in the longer past (for a meta-analysis, see Stalder et al., 2017). Thus, corroborating the present findings, relating hair cortisol levels to retrospective (and specific) major life events may be challenging (Kalliokoski et al., 2019). Considering the multidimensionality and complex interplay of subjective and physiological measures of stress, network analyses might be a promising approach, which includes larger sample sizes (Engert et al., 2018) or collaborations in joint analyses (Boedhoe et al., 2019).
In concert with the above reasons, to date, hair cortisol might only be a limited biomarker for routine measures in mental health to characterize the individual and to individualize treatment. Because of the high number of possible confounders (Stalder et al., 2017), it would be favorable to develop a standardized instrument assessing such variables and to establish reporting guidelines in order to improve comparability between studies. Such approach also requires larger sample sizes. Thus, during study planning, a high number of missing samples (e.g., no consent, amount of hair available) should be taken into account, especially in clinical routine samples.
Footnotes
Ethical Considerations
All participants provided written informed consent for the study protocol approved by the Florida Institutional Review Board (Lang et al., 2018) in accordance with the World Medical Association’s Declaration of Helsinki.
Funding
The authors received no financial support for the research and authorship Support for the publication fee was provided by University of Greifswald's publication fund.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Declaration of Generative AI in Scientific Writing
During the preparation of this work the authors used chatGPT Version 25 September 2023 and Paperpal in order to to improve readability and language. After using these tools, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.