Hair Cortisol Analysis: A Promising Biomarker of HPA Activation in Older Adults

Abstract

Prolonged stress is a potentially harmful and often undetected risk factor for chronic illness in older adults. Cortisol, one indicator of the body’s hormonal responses to stress, is regulated by the hypothalamic-pituitary-adrenal (HPA) axis and is commonly measured in saliva, urine, or blood samples. Cortisol possesses a diurnal pattern and thus collection timing is critical. Hair cortisol is a proxy measure to the total retrospective activity of the HPA axis over the preceding months, much like hemoglobin A1c is a proxy measure of glucose control over the past 3 months. The aim of this review is to examine a novel biomarker, hair cortisol, as a practical measure of long-term retrospective cortisol activity associated with chronic stress in older adults. Hair cortisol analysis advances the science of aging by better characterizing chronic stress as a risk factor for chronic illness progression and as a biomarker of the effectiveness of stress reduction interventions.

Chronic stress is linked to premature aging, early death, disability, chronic disease, depression, and poor quality of life (American Psychological Association, 2012; Epel & Lithgow, 2014; Klein et al., 2014). Among older adults, 63% report chronic illness as a primary source of stress (American Psychological Association, 2012). Chronic stress is a potentially harmful factor that worsens chronic illness and negatively impacts health outcomes of older adults (Buyck et al., 2013; Dawson, Powers, Krestar, Yarry, & Judge, 2013; Judge, Menne, & Whitlatch, 2010; McEwen & Gianaros, 2010; von Känel et al., 2006). Yet, chronic stress is often undetected in this vulnerable population. Cortisol, a significant component of the body’s response to chronic stress, is a steroid hormone regulated by the hypothalamic-pituitary-adrenal (HPA) axis (Hellhammer et al., 2007).

Most current measures of cortisol reflect acute stress, rather than chronic stress (Russell, Koren, Rieder, & Van Uum, 2012). Thus, we are challenged with the identification of a biomarker for measuring chronic stress in older adults. As the aging population increases worldwide, research on the biologic indicators of chronic stress holds significant promise to advance our understanding of interventions to improve the health of older adults exposed to this risk factor for chronic illness (Dawson et al., 2013; Judge et al., 2010). Therefore, we examined a novel biomarker, hair cortisol, for its potential as a proxy measure of long-term retrospective HPA activity associated with chronic stress.

Assessing Cortisol

Current approaches to measure cortisol include saliva, urine, or blood collection; these methods represent the acute status of the HPA axis and not chronic HPA activation (Russell et al., 2012). Although commonly used in cortisol and stress research, these measures share several disadvantages. Salivary cortisol varies diurnally, possesses a variable awakening response, as well as daily acute fluctuations (Clow, Hucklebridge, & Thorn, 2010; Hellhammer et al., 2007). The awakening response results in a 20%–50% increase in cortisol approximately 30–45min postawakening and declines thereafter (Chahal & Drake, 2007; Fries, Dettenborn, & Kirschbaum, 2009). The reduction in urinary output, commonly seen in older adults with chronic renal failure, can affect urinary cortisol results (Russell et al., 2012). Blood cortisol shares similar drawbacks as urinary and salivary cortisol. Plasma cortisol reflects both free and bound cortisol unlike salivary cortisol which reflects only the free or biological active component (Kirschbaum, Strasburger, Jammers, & Hellhammer, 1989).

Researchers have used repeated measures of saliva, urine, and blood as a proxy for long-term cortisol characterizations. However, providing repeated measures increases burden and discomfort for study participants. To obtain saliva samples, study participants must swab their mouths or spit into a container 2–8 times per day for 37 days, depending upon the research data being sought (Broderick, Arnold, Kudielka, & Kirschbaum, 2004; Kudielka, Hawkley, Adam, & Cacioppo, 2007). The results from a recent study on the number of samples needed to characterize salivary cortisol indicate that 3 days of samples are needed for mean results, 8 days for the area under the curve, and 21 days are required to measure diurnal decline (Segerstrom, Boggero, Smith, & Sephton, 2014). The disadvantage to multiple sampling in older adults is that following a complex regimen of multiple sampling reduces adherence to protocol (Kudielka et al., 2007). The changes in oral mucosa with age leading to dry mouth add to the difficulty in obtaining multiple saliva samples (Ghezzi & Ship, 2003).

The protocol to characterize long-term urinary cortisol exposure is no less tedious, as participants must collect urine for 12–24hr, which may not be feasible for older adults with memory impairment (Russell et al., 2012). This involves urinating in a collection pan placed on the toilet, then transferring the urine to a gallon container for 12–24hr. Although less effort is required from study participants, blood cortisol testing is the most invasive of the three approaches because of the pain and discomfort caused by venipuncture. This can lead to erroneous test results due to elevations in cortisol caused by the participant’s fear of venipuncture (Weckesser et al., 2014).

In addition to participant burden and discomfort, there are financial challenges for using current measures of cortisol. These challenges include increased time spent by the participant due to multiple sampling and added costs to the research budget for sample analysis and participant incentives. The average cost of salivary, urinary, and blood cortisol analysis is approximately $20.00U.S. dollars each. This could add up to $100.00 per day for multiple samples. An innovative alternative to the methodological and financial challenges of saliva, urine, and blood testing is hair cortisol analysis, which requires only one sample every 3 months for approximately $50.00 (USD) per sample (Sauvé, Koren, Walsh, Tokmakejian, & Van Uum, 2007).

Hair Cortisol as an Alternative Marker of HPA Activation

Scalp hair cortisol is an emerging measure of cumulative HPA activity without the limitations of other biological specimens (saliva, urine, blood) used to assess cortisol and HPA activity (O’Brien, Tronick, & Moore, 2012). Hair cortisol in the most proximal 3cm of hair is a proxy measure to the total retrospective activity of the HPA over the preceding 3 months, much like hemoglobin A1c is a proxy measure of glucose control over the past 3 months.

Cortisol obtained from scalp hair is a lipophilic substance that originates from the vascular supply which nourishes the hair shaft follicular cells. The source of cortisol within the hair shaft is from circulation to the medullary region in the core of the hair (Russell et al., 2014, p. 594). Sources of cortisol on the surface of the hair include both sweat as well as sebaceous glands which are most likely eliminated by the washing steps prior to grinding the hair for extraction. It is generally thought that the cortisol measured using the standard wash, grind, and extraction method most likely reflects free cortisol and not the bound steroid (Staufenbiel, Penninx, Spijker, Elzinga, & van Rossum, 2013).

It is important to view hair cortisol as a retrospective measure of a previous time period based on the length of hair collected and the specific segment analyzed. Hair grows approximately 1cm/month (LeBeau, Montgomery, & Brewer, 2011; Loussouarn, 2001). If one has collected a longer length of hair and does a segmental analysis of the hair beginning with the segment nearest the scalp, each centimeter is a proxy to the activity of the HPA axis during the month represented by that segment, beginning with the most recent month at the scalp, and prior months in subsequent segments. Therefore, 2cm of hair growth would represent total HPA activity of the previous 2 months, and so forth. Generally, a 3-cm sample is taken from the posterior vertex of the scalp because this area has been shown to have less variation in cortisol levels than other areas of the scalp (Sauvé et al., 2007).

The responsiveness of hair cortisol to changes in HPA regulation is illustrated in data collected from clinical populations in which the activity is abnormal. For example, in patients with Cushing’s syndrome who have alternating episodes of excessive and normal cortisol, hair cortisol concentration increases and decreases in association with the clinical course (Manenschijn et al., 2012; Thomson et al., 2010). Hair cortisol concentrations also decreased after surgical treatment in Cushing’s syndrome patients (Thomson et al., 2010). Hair cortisol is not an acute marker of HPA activity, but rather, a proxy to total HPA activity in the preceding months. Thus, this is a unique marker that provides the researcher with information about participant’s HPA activity prior to recruitment into a study from the historic information contained in their hair.

Although supporting research is still in its early stages, the available literature indicates that scalp hair cortisol is an emerging biomarker of chronic stress (O’Brien et al., 2012; Wosu, Valdimarsdóttir, Shields, Williams, & Williams, 2013). From 2004 to the present, 257 articles on hair cortisol analysis have been cataloged in PubMed, beginning with nine articles in 2004 and increasing to 54 articles in 2014 year-to-date (Figure 1). Of these articles, three report results using hair cortisol analysis in adults 50 and older, which is the focus of our review.

Figure 1.

The number of articles (N = 281) indexed in PubMed from January 2004 to October 17, 2014 featuring scalp hair cortisol analysis. *Year-to-date.

In a study of 57 healthy adults ages 56–77 (mean age: 64.75), the long-term effects of cortisol on cognitive performance were measured using scalp hair cortisol analysis (Pulopulos et al., 2014). Cognitive tests measured working memory, learning, short-term verbal memory, and long-term verbal memory. Lower hair cortisol concentrations correlated with poorer working memory, learning, short-term memory, and long-term memory. These results are counterintuitive in that high cortisol is associated with impaired cognition, as well as rising cortisol levels noted in aging (Feller et al., 2014). Likewise, in a larger study of 645 adults ages 47–82 (mean age: 65.8 years), a positive correlation was found between age and hair cortisol concentration: older subjects had higher cortisol levels than younger subjects (Feller et al., 2014). In most studies involving subjects in a wide range of ages, the concentration of hair cortisol is usually found to be higher with advanced age (Feller et al., 2014; Stalder & Kirschbaum, 2012). This difference may be attributed also to age-related changes in HPA axis function (Kudielka, Buske-Kirschbaum, Hellhammer, & Kirschbaum, 2004).

Hair cortisol concentration also correlates with body measurements, cardiovascular disease, and type 2 diabetes in adults 50 years of age and older (Feller et al, 2014; Manenschijn et al., 2013). The hair cortisol concentration is higher in those adults, whose waist-to-hip ratio and waist circumference is greater than the average (Feller et al., 2014). Furthermore, in a study of 283 community-dwelling older adults (median age: 75; range: 65–85), a high hair cortisol level was associated with an increased risk of cardiovascular disease and type 2 diabetes as compared with older adults with lower levels of cortisol (Manenschijn et al., 2013). We do know that in the three studies of adults over 50 years of age, the hair cortisol concentration means ranged from 21.0 to 40.5 picograms per milligram (Feller et al., 2014; Manenschijn et al., 2013; Pulopulos et al., 2014).

Advantages and Limitations of Scalp Hair Cortisol

Hair cortisol analysis has several advantages over other measures of cortisol. First, segmental analysis or the cutting of hair into segments from the scalp allows for the retrospective review of HPA axis activity reflected in cortisol (Gow, Thomson, Rieder, Van Uum, & Koren, 2010; Kirschbaum, Tietze, Skoluda, & Dettenborn, 2009). Hair cortisol can be measured reliably in hair up to 6cm in length, representing 6 months of HPA axis activity. However, a steady decline in cortisol concentration occurs past 4–5cm from the scalp due to a leaching effect (Gow et al., 2010; Kirschbaum et al., 2009). Thus, HPA axis activity in response to chronic stress can be measured up to 3–4 months previous to when a hair cortisol sample was taken. In addition, hair cortisol measures can be taken 3 months later from new growth, to measure change over time. Moreover, the hair collection process does not cause physical discomfort to the participant. Finally, unlike the additional biohazard and storage precautions that must be followed with saliva, urine, and blood, hair can be placed in foil inside labeled envelopes at room temperature for several years with minimal degradation (Gow et al., 2010; Russell et al., 2012; Webb et al., 2009). In fact, hair cortisol has even been measured in samples collected from mummified remains, suggesting considerable stability over time (Webb et al., 2009).

Although hair cortisol analysis provides an innovative approach to measuring a portion of the chronic HPA activity retrospectively over months, short-term acute stress responses and diurnal patterns cannot be inferred as with other measures (saliva, urine, or blood). In an older adult population, the prevalence of baldness limits participation, and those with thinning hair may be reluctant to provide a hair sample. Additionally, for balding subjects, one cannot use other sites for collection (axillary area, arms, legs), as the growth rates differ and cortisol collected from these sites does not mirror cortisol levels collected from the posterior vertex (M. L. Laudenslager, unpublished observation), the preferred site for collection due to stable growth rates. Variability in the growth rate of hair among different racial/ethnic groups and hair types is not well understood and could influence the amount of cortisol concentration in hair samples, with slower-growing hair potentially having higher levels of cortisol than faster-growing hair (Loussouarn, 2001). Hair cortisol is not affected by natural hair color (e.g., gray, black, blonde). However, hair cortisol levels are significantly affected by repeated shampooing (15–30 times in a laboratory experiment), demi-perms, and bleach, with bleach having the greatest influence on cortisol levels (Hoffman, Karban, Benitez, Goodteacher, & Laudenslager, 2014).

The lack of standardization in reporting the unit of measurement for hair cortisol concentration results presents an additional challenge to the interpretation of results. The majority of studies report hair cortisol concentration in picograms per milligram. However, cortisol concentration results are also reported in nanograms per gram or millimoles per gram not to mention the erroneous reporting in pictograms per milliliter, which makes it difficult to interpret factors among studies and establish norms for the older adult population. Although variation in assays can influence absolute levels, there is a relatively high interlaboratory correlation (Russell et al., 2014). At present, there are no normative value ranges established for hair cortisol concentration in older adults. Additional work needs to be done to determine norms for various ages especially older adults. This work is presently under way in several laboratories.

Conclusion

Chronic stress associated with persistent negative emotions is a risk factor for chronic illness in older adults. Cortisol, measured in saliva, urine, and blood is frequently used to measure HPA axis response to acute stress (stress only reflective of the past 24hr). Hair cortisol analysis offers an innovative approach to measuring HPA axis activity in response to chronic stressors with minimal burden to participants. Opportunities for future research include determining the longitudinal effects of chronic stress on hair cortisol levels and chronic illness progression in older adults. This would help researchers to better appreciate the HPA axis changes in response to chronic stress and target interventions to reduce the adverse effects of chronic stress on health outcomes. As the science advances, the retrospective and prospective data from hair cortisol has the potential to measure change across time in intervention studies designed to improve the health and well-being of older adults.

Funding

Thanks are due to Shirley Moore, PhD, RN, FAAN and The Edward J. and Louise Mellen Predoctoral Fellowship; Substance Abuse and Mental Health Services Administration Minority Fellowship Program at the American Nurses Association (grant #5T06SM060559-03); The John A. Hartford Foundation’s National Hartford Centers of Gerontological Nursing Excellence Award Program; and Dr. M. L. Laudenslager, Patient Centered Outcomes Research Institute (CE-1304–6208). This publication was made possible by the Clinical and Translational Science Collaborative of Cleveland, KL2TR000440 from the National Center for Advancing Translational Sciences (NCATS) component of the National Institutes of Health and NIH Roadmap for Medical Research.