Scientific Validation of Hair Stress Hormone Analysis

In this article, we explore the different levels of scientific validation used for the analysis of Cortisol, Cortisone, DHEA, Testosterone, Progesterone as well as experimentally Endocannabinoids.

The Sapiens Stress Diagnostics includes hair steroid hormone analysis in collaboration with Professor Clemens Kirschbaum and the laboratory "Dresden Lab Services". We use 3 cm hair samples where available representing a 3 months history of stress hormones.

Foundational Science & Analytical Reliability

Overall, the current state-of-the-art in hair steroid analytics which is used by Dresden Lab Services/Sapiens provides both the specificity and sensitivity necessary for reliable measurement of even low-abundance hormones like progesterone or testosterone in hair.

Analytical Specificity

Hair steroid measurements are most reliably performed using liquid chromatography-tandem mass spectrometry (LC-MS/MS), which offers high analytical specificity. This is also the analysis technique used as part of the Sapiens diagnostics. Unlike immunoassay-based methods (e.g. ELISA or RIA) that can cross-react with structurally similar compounds, LC-MS/MS separates and specifically quantifies each target hormone by its unique mass-to-charge signature. For example, an LC-MS/MS method has been developed to simultaneously measure cortisol, cortisone, testosterone, progesterone, corticosterone, DHEA, and androstenedione in human hair, with no cross-reactivity between these analytes (Gao et al., 2013).

In contrast, immunoassays for hair cortisol often overestimate levels due to antibody cross-reactivity – one study found ELISA values about 4-fold higher than LC-MS/MS in the same hair samples, partly because the ELISA was also detecting cortisone (Slominski et al., 2015). LC-MS/MS thus provides superior specificity and can distinguish cortisol from cortisone and other steroids that immunoassays might inadvertently detect. This specificity is crucial given that hair contains multiple glucocorticoids; notably, hair cortisone is often present at ~10–20% the level of cortisol and is concurrently measured via LC-MS/MS for a more complete assessment of HPA-axis output (Slominski et al., 2015).

The use of LC-MS/MS has become the gold standard for hair steroid analysis in research and clinical laboratories (Greff et al., 2019).

Analytical Sensitivity

Modern LC-MS/MS assays, like the one used by the Sapiens related laboratory Dresden Lab Services for hair steroids are highly sensitive. Detection limits in the sub-picogram per milligram range have been reported. For instance, Gao et al. (2013) achieved limits of quantification ≤0.1 pg/mg for cortisol, cortisone, and other steroids (0.9 pg/mg for DHEA) using 10 mg of hair per sample (Gao et al., 2013). Such sensitivity means that even the low endogenous concentrations of steroids in hair (which are on the order of a few pg/mg for many hormones) can be reliably detected in virtually all individuals. Immunoassays, while sometimes appearing very sensitive, can give inflated readings at the low end due to cross-reactivity rather than true signal. In practice, the typical hair cortisol concentration in healthy adults is well above assay detection limits – often around 5–20 pg/mg in the 3 cm of hair closest to scalp (reflecting 3 months) depending on the cohort and method, with observed ranges from <1 pg/mg up to dozens of pg/mg in non-pathological conditions. In children, a recent LC-MS/MS-based reference study found hair cortisol drops from high neonatal levels (160 pg/mg at birth) to much lower levels by childhood (median ~3 pg/mg by late adolescence) (de Kruijff et al., 2020). The assay sensitivity comfortably covers this range. In terms of precision and reliability, validated hair steroid LC-MS/MS assays show intra- and inter-assay coefficients of variation in the single-digit percentages (often ~5–10% CV) (Gao et al., 2013), indicating excellent analytical reproducibility. Furthermore, hair cortisol measurements show good test–retest consistency when the same individual’s hair is sampled at different times: studies report correlations around r = 0.7–0.8 for repeat hair cortisol measures, highlighting respectable reliability of the assay and the trait-like stability of long-term cortisol output (Stalder & Kirschbaum, 2012); (Kirschbaum et al., 2009). It should be noted that minor methodological differences (e.g. pulverizing vs. cutting hair, solvent extraction protocols) can affect absolute yields, but when validated, most protocols yield comparable and sensitive results (Slominski et al., 2015).

‍

‍

‍

Correlational & Associative Validity

Hair cortisol has emerged as a meaningful biomarker of chronic stress and exhibits both clinical specificity and sensitivity in certain contexts. Clinical specificity refers to how uniquely a biomarker is elevated in the condition of interest versus other conditions, and sensitivity refers to how well it detects those with the condition. Hair cortisol tends to reflect long-term hypothalamic–pituitary–adrenal (HPA) axis activity, and numerous studies show it is specifically elevated in states of sustained cortisol overproduction (or decreased in states of cortisol underproduction). For example, in Cushing’s syndrome – a disorder of chronic cortisol excess – hair cortisol is markedly higher than in healthy individuals, and it corresponds with disease activity. Manenschijn et al. (2012) reported that a hair cortisol cut-off of ~31 pg/mg distinguished Cushing’s patients from controls with 93% sensitivity and ~90% specificity (Manenschijn et al., 2012). Notably, hair cortisol in these patients correlated strongly with traditional clinical cortisol metrics: in one study, hair cortisol concentrations showed r ≈0.7 correlations with 24-hour urinary free cortisol and late-night salivary cortisol levels (indicating good agreement with established diagnostic tests) (Manenschijn et al., 2012). This demonstrates that hair cortisol is sensitive to the hypercortisolemic state and specific to disorders of cortisol excess.

Outside of overt endocrine disease, associative validity of hair steroids has been examined in relation to psychosocial stress and health outcomes. A meta-analysis of chronic stress studies found that groups with high chronic stress exposure had on average 22% higher hair cortisol levels than non-stressed groups (Stalder et al., 2017). This suggests hair cortisol is sensitive enough to pick up physiologically relevant differences in stress hormone output associated with chronic stress. Indeed, hair cortisol has been positively associated with various stress-related conditions: for instance, patients with ongoing depressive symptoms often show elevated hair cortisol compared to those in remission or healthy controls, consistent with HPA axis hyperactivation in chronic depression (Steudte-Schmiedgen et al., 2017). These associations underscore that hair cortisol is not only analytically valid but also correlates in expected ways with clinical and subclinical states of stress and disease.

It is important to note that the clinical sensitivity of hair cortisol to psychological stress, while present, is moderate – not every individual under stress will show a large hair cortisol elevation, as there is substantial inter-individual variability. Some studies find only weak correlations between hair cortisol and self-reported stress levels (possibly because subjective stress and physiological stress response can diverge). Nevertheless, in aggregate, the evidence supports that hair cortisol specifically reflects prolonged cortisol exposure and is sensitive enough to capture differences in chronic stress burden at a group level (Stalder et al., 2017). Moreover, hair cortisol changes appropriately with interventions: for example, individuals who underwent stress-reduction programs or recovered from depressive episodes have shown decreases in hair cortisol over time, whereas those with worsening stress or persistent illness maintain high levels (Steudte-Schmiedgen et al., 2015). Such findings give confidence in the biomarker’s validity: it behaves as expected under conditions of increasing or decreasing chronic cortisol exposure.

Beyond cortisol, other hair steroids like hair testosterone and hair DHEA have been studied for associative validity. These measures are less established, but early findings are intriguing. For instance, hair testosterone in men has been found to correlate with long-term average serum testosterone and has been explored in contexts like paternal behavior and polycystic ovary syndrome. Hair DHEA(S) tends to track parallel to hair cortisol to some extent – for example, chronic stress that elevates cortisol may also elevate adrenal DHEA output, which can be reflected in hair (though DHEA data are sparser). Some studies have looked at the hair cortisol:DHEA ratio as an index of HPA balance, but results are mixed $$(Gerber et al., 2013 – Psychoneuroendocrinology, 38: p.1990–2000). Overall, while cortisol is the flagship hair hormone with strong validity, the presence of other steroids in hair holds potential for a more complete picture of endocrinological changes; ongoing research is evaluating how specifically changes in hair androgens or other adrenal steroids align with clinical states.

‍

‍

Interpretive & Contextual Utility

Interpreting hair steroid results requires understanding what these measures represent. Hair grows approximately 1 cm per month, so the steroid content of a given 1 cm segment reflects the integrated hormone secretion over that month. This retrospective averaging is a key strength: hair cortisol (and other hormones) captures long-term endocrine activity that is not discernible from point-in-time blood or saliva tests. As a result, hair steroid levels are robust against short-term fluctuations due to circadian rhythm or acute stress events. For example, whereas cortisol in blood or saliva fluctuates widely over a day, cortisol incorporated into hair provides a smoothed, cumulative value. This makes it highly useful for assessing chronic stress load or long-term hormonal changes (Russell et al., 2012). Indeed, hair cortisol has been called an “integrated index” of HPA activity over weeks to months (Russell et al., 2012), offering a window into the past that complements acute measurements (Meyer & Novak, 2021).

However, several contextual factors must be considered for proper interpretation. Hair growth rate can influence hormone concentrations: slower-growing hair accumulates more hormone per unit length (since the hair spends longer time perfused by blood). An illustrative case is seen in infants – newborns have a much slower hair growth in early months, which likely contributes to their very high hair cortisol levels at birth that then decline as hair growth speeds up (de Kruijff et al., 2020). In one study, infants 0–6 months old had about one-third the hair growth rate of toddlers, coinciding with significantly higher cortisol per mg in infant hair; the authors suggest the low growth rate concentrates cortisol in the hair shaft (de Kruijff et al., 2020). Thus, age and hair growth dynamics are relevant: children and older adults (who may have slower hair growth or changes in hair physiology) can have different baselines than healthy adults.

Hair treatment and cosmetic factors are another contextual consideration. Fortunately, normal variations in natural hair color, thickness, or washing frequency do not appear to systematically affect hair cortisol levels in a meaningful way (Kirschbaum et al., 2009). Kirschbaum and colleagues found no significant difference in hair cortisol between individuals of different hair colors (e.g., dark vs. blond hair) and no correlation with how often hair was washed per week. This indicates the hormone is quite securely embedded in the hair matrix, and routine washing (shampooing) does not wash out the cortisol to any great extent after it has been incorporated. Chemical treatments, on the other hand, can have an effect. Bleaching or dyeing hair, especially with harsh chemicals, has the potential to reduce measurable steroid content by either partially destroying hormone molecules or by leaching them out. Some studies reported that in vitro hair dyeing can reduce cortisol readings (though others found minimal effect from commercial hair dyes) – due to inconsistent findings, it’s recommended to document any chemical treatments. Generally, if a patient heavily bleaches their hair, their hair cortisol might come out artificially low. In practice, many clinical protocols advise sampling an untreated segment of hair (or noting the treatment status) to interpret results properly (Stalder & Kirschbaum, 2012).

Environmental contamination is less of a worry with endogenous steroids than it is with external substances (like illicit drugs) in hair. Cortisol and other steroids are produced internally and incorporated from the follicle’s blood supply; although sweat and sebum contain steroids, the washing procedures prior to analysis (multiple solvent washes) remove surface contaminants. Studies have shown that external sweat cortisol does not substantially confound hair cortisol measurement under typical conditions (Stalder et al., 2015). Thus, an elevated hair cortisol is interpreted as reflecting true internal production (barring a scenario where someone’s hair was soaked in cortisol-containing fluids, which is unlikely in real life).

Another aspect of interpretive utility is the possibility of segmental analysis. By cutting hair into sequential segments (e.g., 3 cm segments), one can chart hormone levels over time. This has been used to map events like pregnancy or therapy response. For instance, in one of the first studies, hair cortisol of postpartum women showed a peak in the segment corresponding to the third trimester of pregnancy, then a drop in later segments – matching the known high cortisol of late pregnancy and normalization thereafter (Kirschbaum et al., 2009). Similarly, segmental hair analysis in Cushing’s syndrome has revealed cyclic disease activity: patients with cyclical Cushing’s show alternating high and normal cortisol segments, aligning with symptomatic and asymptomatic phases (Manenschijn et al., 2012). This retrospective calendar ability is a unique interpretive strength of hair analysis, enabling clinicians to look back at an endocrine timeline (e.g., identifying approximately when hypercortisolism began). The contextual utility of hair steroids is therefore high for conditions that develop or fluctuate over months.

Finally, it’s worth noting that each hormone in hair should be interpreted within its own physiological context. For example, hair testosterone in a female patient might be considered in light of clinical signs of hyperandrogenism or androgen therapy, whereas hair progesterone might be very low in postmenopausal women (as expected) or could reflect luteal phase output in cycling women. These measures are still exploratory, but as research grows, reference ranges and contextual norms for each hair hormone will become clearer. At present, cortisol and cortisone in hair have the most developed interpretive frameworks, while DHEA, testosterone, and progesterone are interpreted more cautiously (often relative to cortisol or as part of research indices). Endocannabinoids in hair are entirely new territory – initial findings suggest they could relate to chronic stress or neuropsychiatric conditions, but no clinical reference standard exists yet. Thus, context (such as whether a person is known to have high stress or certain metabolic conditions) is critical when interpreting these experimental markers.

‍

‍

‍

Validation of At-Home Sampling Method

At-Home Sampling Validity

One practical advantage of hair hormone analysis is that sample collection is non-invasive and can be done by individuals at home. Studies have tested whether self-collected hair samples are as valid as those collected by trained professionals. The evidence so far is reassuring: Enge et al. (2020) conducted a study with over 400 participants and found no significant difference in hair cortisol concentrations between self-collected samples (with a partner’s help at home) and samples collected by researchers in a lab setting (Enge et al., 2020). In a subset of participants who provided both types of samples, the hair cortisol levels were statistically equivalent. This suggests that as long as clear instructions are given, individuals can accurately cut a hair sample that yields reliable hormone data. Typically, instructions for at-home collection include: using clean scissors to cut a small lock of hair from the posterior vertex (crown of the head) as close to the scalp as possible, aligning the cut strands, and then securing the sample (usually in aluminum foil or a small bag) with the scalp end marked. The posterior vertex region is recommended because hair growth rate is most consistent there and not overly affected by acute patchy shedding. Participants are also instructed not to substitute shed hairs or hair from a brush (since those hairs may not have a definable segment of recent growth). When such protocols are followed, at-home collection is quite feasible and valid (Enge et al., 2020). There is typically a slight increase in sample loss or unusable samples from mail-ins (e.g. a few percent of kits might return with too little hair or hair that’s fallen out and not usable for analysis), but overall the success rate is high. This means large-scale studies and even clinical services can leverage mail-in hair sampling without significant loss of data quality.

Laboratory Transit

Hair samples are very stable under normal transit and storage conditions. Unlike blood or saliva, hair does not require special handling like refrigeration or immediate processing. Once the hair is cut and dry, the steroid content is essentially locked in place within the hair shaft’s keratin matrix. Research has shown that hair cortisol remains stable at room temperature for extended periods. Berger et al. (2024) demonstrated that hair cortisol concentrations were essentially unchanged even after 5 years of storage at room temp in an envelope, with a strong concordance (r > 0.8) between the values measured after 5 years and the initial values (Berger et al., 2024). Over shorter time frames, such as the days or 1–2 weeks it might take for a mailed sample to reach a lab, there is no evidence of any degradation. Cortisol and other steroid hormones are non-volatile, chemically stable molecules; they are not significantly affected by typical fluctuations in temperature or light during shipping (within normal environmental ranges). Even in extreme scenarios, like a sample being transported in very hot or cold weather, the hair’s physical protection and the robustness of steroid molecules mean the risk of hormone breakdown is minimal.

Additionally, hair samples are often mailed in simple packaging (a paper envelope or small container). As long as the package stays dry (no water exposure) – which is usually the case – the sample will arrive in good condition. There is also no need for preservatives or special liquids, simplifying logistics. For these reasons, laboratory transit of hair samples is considered reliable. Many large cohort studies have successfully employed mail-in hair collection, and analytes have remained consistent. One minor consideration is avoiding contamination or mix-ups: each sample is usually labeled carefully, and participants are instructed not to touch the hair too much or contaminate it with lotions, etc., before mailing (though even if they did, the lab’s wash steps would likely remove surface contaminants). Upon arrival at the lab, hair samples are typically stored dry at room temperature or cooler (some labs keep them in the dark or at -20°C for long-term storage, but this is not strictly required).

In summary, the transit and storage validation for hair steroids is excellent. Hair is a very convenient medium for remote collection – it does not spoil, it is lightweight and easy to ship, and hormone levels remain stable over time. This has been a boon for research during periods where clinic visits are challenging (e.g., during the COVID-19 pandemic, many studies pivoted to hair sample mailing for stress hormone assessment). The stability of hair samples in transit means that results from an at-home collected hair strand can be trusted to reflect the donor’s hormone levels at the time of sampling just as well as if the sample were collected and processed on-site immediately.

‍

‍

‍

Clinical Validation

Hair steroid analysis, especially hair cortisol measurement, has undergone extensive clinical validation in recent years. It has moved from an experimental technique to a tool that is being used in endocrinology and psychiatry research, with emerging clinical applications. Hair cortisol is the most validated marker: for diagnosing hypercortisolism (Cushing’s syndrome), hair cortisol testing has shown high diagnostic accuracy. As mentioned, Manenschijn et al. (2012) confirmed that patients with Cushing’s have dramatically higher hair cortisol, and that hair cortisol falls after curative treatment – indicating the test is not only diagnostic but also responsive to clinical improvement (Manenschijn et al., 2012). In cases of cyclic Cushing’s (intermittent cortisol excess), segmental hair analysis has been able to unmask the cycling, capturing historical peaks of cortisol that align with symptomatic periods. This is a game-changer for a condition that is otherwise hard to confirm with spot tests. Major endocrine societies are now considering how hair cortisol might integrate into the diagnostic workup for suspected Cushing’s or for monitoring remission. Similarly, in Addison’s disease (primary adrenal insufficiency), hair cortisol provides a novel way to verify chronically low cortisol production. Preliminary data show Addison’s patients have significantly lower hair cortisol than healthy controls, and importantly, hair cortisol levels rise appropriately in those patients who are on cortisol replacement therapy (hydrocortisone) – reflecting the hormone reaching the hair (Manenschijn et al., 2012). This suggests a potential role for hair cortisol in monitoring long-term adequacy of replacement in adrenal insufficiency (though in practice blood levels and clinical assessment are primary, the hair metric could add an objective cumulative measure of exposure).

Beyond adrenal disorders, clinical validation is expanding to other conditions. For example, chronic stress-related pathologies: research has validated that hair cortisol is elevated in conditions like chronic pain, burnout syndrome, and PTSD in many (though not all) studies. This has led to an interest in using hair cortisol as an objective adjunct marker in psychiatric evaluation. A notable finding in PTSD is that some patients (especially those with chronic PTSD) actually show lower hair cortisol, possibly due to HPA exhaustion or a different pathophysiology – such observations underscore that “validation” includes understanding in which clinical states hair cortisol goes up or down relative to normals. For instance, one study of PTSD patients found hair cortisol was lower on average than in non-traumatized controls, while another found no difference – so work is ongoing to validate hair cortisol’s role in psychiatric diagnostics. Nevertheless, hair cortisol has been successfully used as an outcome measure in clinical trials (e.g., to test if a psychotherapy or medication normalizes long-term cortisol output). A 2021 systematic review noted that in some psychotherapy trials, hair cortisol decreased in responders, highlighting its potential as a biomarker of treatment response (Botschek et al., 2023).

As for androgen and other steroid measurement in hair, clinical validation is in earlier stages. Hair testosterone has been evaluated in disorders of androgen excess: some small studies suggest that women with polycystic ovary syndrome (PCOS) have higher hair testosterone levels on average, correlating with their hirsutism scores and serum androgens, implying hair T could serve as a long-term index of androgen exposure $$(Allen et al., 2019 – Clin. Endocrinol., 90(3): p.383–390). Hair DHEA(S) has been preliminarily used in developmental studies (since DHEA rises in adrenarche during adolescence, hair DHEA might track that). There is also interest in using hair steroids for doping control or monitoring exogenous steroid use: for example, hair testosterone or hair progesterone could reveal long-term use of these hormones (since blood/urine tests only catch recent use). Validation studies in that domain have confirmed that exogenous steroid administration can be detected in hair (e.g., long-term glucocorticoid therapy leads to high hair cortisol that correlates with dose).

In summary, hair steroid analysis is transitioning into practice for select applications. Cortisol in hair is the most validated, with strong evidence supporting its use in diagnosing and monitoring Cushing’s syndrome (with sensitivity and specificity around 90% or above in experienced laboratories) and promising applications in chronic stress assessment. At-home collection validity and analytical robustness have been demonstrated, which paves the way for broader clinical deployment (Stalder et al., 2016). As always, establishing reference intervals is part of clinical validation: efforts like the Erasmus Rotterdam cohort and others have produced age-stratified reference ranges for hair cortisol (de Kruijff et al., 2020) and are working on similar data for other hair steroids. With each new study, the confidence in hair steroid measurements grows.

It’s important to communicate to clinicians that hair measurements complement rather than replace traditional tests in many cases. For example, in suspected Cushing’s, one would still do blood, urine, or saliva tests; hair cortisol adds a long-term perspective and can be especially useful if those tests are equivocal or if cyclic disease is suspected. In the context of mental health, hair cortisol provides biological validation of a patient’s chronic stress load, which can strengthen a holistic assessment. As for the endocannabinoids and other experimental hair biomarkers, these are in the research validation phase. A recent 2025 study measured anandamide (AEA), 2-AG, PEA, OEA in hair of PTSD patients, exploring correlations with symptom severity (Bergunde et al., 2025). Such studies are defining how these novel markers behave; early results suggest, for instance, that low hair AEA might be linked with high anxiety symptoms, but these are not yet validated for clinical use.

In conclusion, the clinical validation of hair steroids is most solid for cortisol (and to a degree cortisone), with growing support for its use in endocrinology and psychosomatic medicine. Other hair steroids and modulators (DHEA, testosterone, progesterone, endocannabinoids) are research-frontier markers – our current panel includes them on an exploratory basis, and we interpret them with caution. As validation studies continue, we anticipate these too may find specialized clinical roles. For now, they serve to enrich research data and generate hypotheses, while cortisol remains the cornerstone hair hormone for clinically actionable insights.

(Experimental/Research Markers Notice): The inclusion of hair endocannabinoids (AEA, 2-AG, PEA) and certain steroid hormones like DHEA and progesterone in our panel is primarily for research purposes at this stage. These analytes can be measured with advanced LC-MS/MS techniques (Behringer et al., 2021), but their clinical interpretation is not yet established. We consider these results as exploratory data points. Any findings related to these experimental markers will be interpreted with caution and in the context of supporting research literature, rather than as definitive clinical diagnostics.

‍