Investigating relations among stress, sleep and nail cortisol and DHEA

Abstract

In the current study, we present data investigating the relationships among stress, sleep disturbance, self-control, and levels of cortisol (CORT) and dehydroepiandrosterone (DHEA) in fingernail clippings. Currently, hair CORT is the only routinely used noninvasive, validated, biomarker of chronic exposure to stress-related hormones. Nail clippings represent an important potential alternative sample matrix for assessing chronic hormone exposure, as it offers a different timeline of hormone incorporation than scalp hair, and may be obtainable from populations in which hair either is lacking or is unavailable for cultural reasons. Moreover, there is established precedent for using fingernail clippings to attain biomarker data. However, the value of nail hormone assessment for psychological research is currently unknown due to a paucity of information on the relations between nail hormone concentrations and environmental or psychological variables. In the present study, we collected data from a low income, minority population (N = 47; 97% African American) to demonstrate feasibility and acceptability of nail collection and analysis of the adrenal steroids CORT and DHEA. Participants reported on perceived stress, sleep and self-control abilities. Correlational analyses suggest that exposure to stressful events, disturbances in sleep and waking were associated with higher levels of nail DHEA, while self-control was associated with higher levels of nail CORT. We discuss the potential importance of this methodology for investigating biological, behavioral, and subjective indices of stress and well-being.

Keywords:

• Fingernails
• cortisol
• DHEA
• sleep
• self-regulation
• stress
• self-control

Lay summary

We assess levels of stress hormones cortisol (CORT) and dehydroepiandrosterone (DHEA) from nail samples. Our results suggest that stressful events and disturbances in sleep were associated with higher levels of nail DHEA. Interestingly, higher self-control was associated with higher nail CORT.

Levels of the hypothalamic–pituitary–adrenal axis (HPA) hormones is pivotal for the stress response, but also influence higher levels of cognitive and psychological functioning (Holsboer & Ising, 2010) as well as physical and mental health outcomes (de Kloet et al., 2006; Vogelzangs et al., 2010). However, while biomarkers of acute stress can be well-quantified (Goldstein, 1995), current assessments of chronic stress that involve the collection of serum, blood or urine are time intensive and costly. Hair analysis has been used to identify long-term exposure to hormones, particularly CORT as it is noninvasive and relatively stable. Hair CORT levels are associated with lower family income (Rippe et al., 2016), lifetime trauma (Simmons et al., 2016), body mass index and somatic complaints in children (Gerber et al., 2017), increased physical activity (Gerber et al., 2013), and chronic pain (Van Uum et al., 2008). It has also been demonstrated to be higher in long-term unemployed individuals as compared to controls (Dettenborn, Tietze, Bruckner, & Kirschbaum, 2010), dementia caregivers (Stalder et al., 2014) and higher in children during school entry (especially for fearful children) (Groeneveld et al., 2013). It is important to note, however, that higher levels of CORT does not automatically lead to negative outcomes, and the meaning of CORT needs to take into account context (Cichetti & Rogosh, 2007; Miller, Chen, & Zhou, 2007). Higher levels of CORT have been associated with positivity in preschool girls (Hatzinger et al., 2007), and individuals with higher levels of antisocial behavior have rather low levels of CORT (Flinn & England, 1995; Hawes, Brennan, & Dadds, 2009).

Unfortunately, hair collection excludes a wide range of individuals, including those with short or no hair, including many men, young infants and some of the elderly. Moreover, hair collection is also unacceptable to some individuals for cultural or religious reasons (Ferris, 1991; Kilcup, 2000; Mcleod, 1989). As an example, for individuals who practice Sikhism, cutting hair is forbidden. All of these factors limit the ability of individuals from specific racial/ethnic, age groups, and other demographic backgrounds to participate in research on stress, despite being over-represented among vulnerable populations. For example, African Americans are commonly underrepresented in clinical trials targeted to disorders from which they suffer disproportionately (Branson, Davis, & Butler, 2007). At the same time, the practice of keeping short hair, or shaving all hair among African American men, and perming, weaving, or adding extensions to hair among African American women make hair sampling difficult. An alternative method to assess chronic stress hormone exposure that is effective, valid, and involves samples that are easy to obtain from a wide range of individuals is thus vitally important.

The capability to obtain hormone levels through fingernails may be one method for addressing the limitations of current chronic stress assessments. Indeed, nail clippings have been used in epidemiological studies for toxins and drugs, and nutrients for over the past three decades (Jenkins, 1979). However, only recently have researchers consider the possibility of using fingernails for biomarking stress hormones. Both CORT and DHEA have been detected and quantified in nails by liquid chromatography/electrospray ionization-tandem mass spectrometry (Higashi et al., 2016; Khelil et al., 2011) and, in the case of CORT, by enzyme immunoassay (Izawa et al., 2015). Because nails consist of dead keratinous cells, once the compound is incorporated into the nail matrix, it does not become metabolized. On average, fingernails grow at about a rate of 1.0 mm/10 days; nails fully extend from the nail matrix over the course of several months (Gupta et al., 2005). Fingernail samples then capture retrospective exposure approximately several months before clipping (Warnock et al., 2010). While both hair and nail capture chronic rather than momentary exposure to hormone levels, unlike hair, nails are available from a vast majority of individuals, do not usually confer the same cultural and emotional restrictions. Nails are also resistant to decomposition and disintegration.

Nail CORT and DHEA: relations with stress, sleep and self-control

Nail CORT has similar concentration levels and variance with hair CORT (Hubmann et al., 2016), has been found to be moderately associated with both facial hair (Nejad, Ghaseminezhad, Sung, Hoseinzadeh, & Cabibi, 2016), and scalp hair CORT from the same participants, as well as saliva CORT across the day (Izawa et al., 2015). A few studies have investigated associations between nail CORT and DHEA and psychological outcomes. Warnock et al. (2010) found a significant increase in the CORT:DHEA ratio, due to an increase in DHEA during stressful times. Similarly, nail CORT was also found to be higher during exam times, as compared to a baseline for students (Nejad et al., 2016). Izawa demonstrated that the experience of stressful life events, but not perceived stress was associated with elevated nail CORT (Izawa, Matsudaira, Miki, Arisaka, & Tsuchiya, 2017). Higher levels of DHEA have been found in nails from infants of mothers who experienced stressful life events during pregnancy (Tegethoff et al., 2011). However, no association between nail CORT and neurocognitive impairment has also been reported (Herane-Vives et al., 2015).

No studies using nail has investigated the relations between stress hormones and sleep or self-control. But, sleep duration and disturbances have been found to be associated with flatter slope in diurnal CORT in a large cohort study (Kumari et al., 2009). Shift workers (those who work outside of regular daytime hours) (Manenschijn, van Kruysbergen, de Jong, Koper, & van Rossum, 2011), and daytime sleeping (Feller et al., 2014), are positively associated with higher levels of hair CORT. Sleep deprivation has also been associated with increased plasma CORT levels (Wright et al., 2015). DHEA has been much less researched, but has been implicated with sleep in several studies (Caufriez, Leproult, L’Hermite-Baleriaux, Kerkhofs & Copinshi, 2013; Opstad, 1992; Schule et al., 2003).

Finally, while nail CORT have not been investigated in relation to self-control, research using basal levels of CORT (e.g. as indexed by plasma), has found that low resting CORT is associated with low levels of self-control (Shoal, Giancola, & Kirillova, 2003), impulsivity (King, Jones, Scheuer, Curtis, & Zarcone, 1990) and other indicators of deficits in self-regulation such as aggression (Poustka et al., 2010). Additionally, research using a sample very similar to ours (low income, African American adolescents) have also found that self-regulation was associated with another measure of long-term physiological stress, namely allostatic load (Brody et al., 2013). There is also preliminary evidence that DHEA is associated with self-control (Sripada et al., 2013) and better regulation for maltreated children (Cicchetti & Rogosch, 2007).

The purposes of the current study

To date, no study has examined associations among nail CORT and DHEA and a broad range of psychosocial variables in a disadvantaged minority sample. In the current study, we explore associations among nail CORT and DHEA, with perceived stress, stressful events, sleep, and trait self-control, in a high risk, predominantly African American adolescent population. We chose to focus on African American adolescents because stressful events have been demonstrated to be more frequent in low socioeconomic status racial/ethnic minorities (Hatch & Dohrenwend, 2007). At the same time, hair collection to assess chronic CORT exposure is more difficult in this population due to specific hair practices discussed above. Moreover, the current data on nails has focused on adults, and we are interested in generalizing to adolescents.

Methods

Study sample

Participants were high school students enrolled in a public college preparatory charter school (N = 47, female = 24, male = 15.93). The majority (96.7%) of participants identified as Black or African American and two individuals identified as White. All participants were drawn from a largely low-income sample (89% of students in the school qualified for free and reduced lunch). Students and parents of minors provided consent and assent. The Institutional Review Board approved all measures and procedures.

Procedures

After obtaining consent, participants were told they would be participating in a study assessing stress among adolescents. Because, it takes a couple of months for nails to grow to the point where collection would be feasible, the questionnaire data was collected at the beginning of the spring semester (March), and nail samples were at the end of the Spring semester in May. To collect the nail samples, participants were asked to clip their fingernails 2 weeks before the beginning of the study. Two weeks later, participants once again clipped and provided their nail samples to the experimenter.

Measures

Perceived Stress Scale (PSS). The PSS is the most widely used psychological instrument for assessing subjective experience of stress (Cohen, Kamarck, & Mermelstein, 1983). The reliability for the scale in our sample was Cronbach’s α = 0.81.

Academic stress. In addition to PSS, we assessed the occurrence of stressful events that are pertinent to the current sample. Specifically, we inquired about exposure to stressful academic events in the past 6 months. Four items assessed whether the individual has failed a class (19 students reported failed a class), did poorly on a test or paper, received bad grades, or had a specific project that they worried a lot about. This resulted in a frequency score of stressful events. In addition, for each item that was checked “yes”, we asked the individual to rank on a scale of 1–5, how much the event bothered them. A mean then was computed for subjective reactions to the stressor. These scores were then summed to create the academic stress variable.

Pittsburgh Sleep Quality Index (PSQI). The PSQI consists of seven components: subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleeping medications, and daytime dysfunction over the last month. The components were combined to create three subscales (sleep efficiency, perceived sleep quality and daily disturbances) (Tomfohr, Schweizer, Dimsdale, & Loredo, 2013). Sleep efficiency taps into the proportion of time in bed spent sleeping, perceived sleep quality assesses subjective sleep quality as well as sleep latency (how long it takes the individual to fall asleep), and finally daily disturbances tap into both night time disturbances that keep the individual from sleeping (e.g. snoring, not being able to breathe, bad dreams), as well as day time disturbances including not being able to stay awake. The PSQI has demonstrated strong reliability and validity (Mollayeva et al., 2016). For the current sample, Cronbach’s α for items assessed on a scale was = 0.82.

Self-regulation. Trait self-regulation was assessed using the effortful control subscale of the Adult Temperament Questionnaire (Rothbart, Ahadi, & Evans, 2000). Items #72 and #76 were missing due to experimenter error. Overall reliability of the scale was Cronbach’s α = 0.54). Thus, we conducted a factor analysis and extracted items that seem to tap into “emotion regulation”. Items (N = 5, item 8, 40, 50, 60, and 63) include questions such as “When I am happy and excited about an upcoming event, I have hard time focusing”, “When I am excited about something, it’s usually hard for me to resist jumping right into it before I’ve consider the possible consequences”, and “I make plans but do not follow through.” The reliability for this scale was Cronbach’s α = 0.73. In our analyses, we find that the results were identical for both scales, thus we report findings for the original “effort control” scale.

CORT and DHEA in nails. Nails were analyzed in duplicate for CORT and DHEA content by enzyme immunoassay (EIA; Salimetrics, Carlsbad, CA; kit numbers 1-3002 and 1-1202, respectively) after sample processing using procedures optimized in preliminary studies in our laboratory. Each sample was weighed, placed into a 2.0-ml microcentrifuge tube, and washed twice for 1-min each with isopropanol to remove external contamination. Washes were carefully decanted and the samples were allowed to air dry. Following washing, nails were ground to a fine powder for 10 min at 25 Hz using a Retsch MM200 ball mill (Verder Scientific, Newtown, PA). Ground nails were weighed, placed into a clean 2.0-ml tube, and extracted overnight (18–24 h) with 1.5 ml HPLC-grade methanol. One milliliter of the methanol extract was evaporated under vacuum and reconstituted in 0.2 ml assay diluent (assay diluent is the same in both the CORT and DHEA EIA kits) for subsequent assay. Assay readouts were corrected for the powdered sample weight to yield pg steroid per mg dry nail weight. Because there was insufficient material left for DHEA analysis and CORT measurement was prioritized over DHEA, the CORT assays were conducted first and for those samples that had needed to be redone for technical reasons (e.g. high variability between duplicates). The intra- and inter-assay coefficients of variation (CVs) were 5.6 and 2.7%, respectively for CORT. For DHEA, the intra- and inter-assay CVs were 5.7 and 7.8%, respectively.

Results

One fingernail sample was dropped because the participant did not follow instructions and had utilized fingernail polish. Two other samples yielded insufficient powder weight for extraction and analysis. Of the remaining 44 samples (males = 23), powdered sample weights ranged from 3.1 to 75.8 mg with a mean of 22.8 [standard deviation (SD) = 17.3] mg. CORT samples were for 44 participants. In line with past research, concentrations of fingernail CORT levels ranged from 1.4 to 239.1 pg/mg (Izawa et al., 2015) with a mean of 21.1 (SD = 39.5). With regards to DHEA, three participants had samples that were unreadable because they were beyond the highest standard in the EIA kit and there was no extract remaining to dilute and re-analyze. For an additional 10 participants, there was insufficient reconstituted nail extract for DHEA assay after the CORT analysis had been completed. Thus DHEA quantifiable in 30 participants (males = 18), DHEA levels range from 16.6 to 1498.3 pg/mg, with a mean of 128.4 (SD = 256.7) pg/mg. One participant was eliminated from subsequent statistical analysis due to an extremely high level of DHEA (1498.3 pg/mg), which was five standard deviations above the mean. Both CORT and DHEA levels were non-normally distributed with a skewness of 4.42 and kurtosis of 22.48 for CORT, and 5.05 and 27.30 for DHEA; thus, the values were logarithmically transformed prior to conducting correlational analyses. There were no gender differences with regards to levels of nail CORT or DHEA, and age was not associated with either variable.

Both Pearson product-moment (r) and Spearman rank-order (r_s) correlations were conducted to investigate whether the nail CORT and DHEA levels were related to each other or were associated with subjective stress, sleep quality, and self-regulation. Table 1 reports the Pearson correlations for all study variables. For those samples in which both CORT and DHEA results were available, there was no significant Pearson or Spearman correlation between the concentrations of these hormones. Assessments of the occurrence of stressful academic events (r = 0.50, p = .006; r_s = 0.48, p = .008) was significantly correlated with DHEA. Perceived subjective stress was not associated with either nail CORT or DHEA. Among sleep efficiency, quality and disturbances, nail DHEA was positively associated with higher levels of sleep disturbances using both Pearson correlation between samples (r = 0.42, p = .038) and rank-order correlation (r_s = 0.35, p = .05). However, with regards to self-regulation, the effortful control scale was correlated with nail CORT (r = 0.42, p = .007; r_s = 0.386, p = .014). To look at the combined influence of CORT and DHEA, we computed a CORT/DHEA ratio. The ratio was positively correlated with self-control (r = 0.40, p = .031), as well as sleep disturbances (r = −0.37, p = .04). We also explored the combined effect of nail CORT and DHEA, by testing for moderation (Sollberger & Ehlert, 2016), using the PROCESS Macro (Hayes, 2013). However, the model was not significant. Table 1 provides the Pearson correlations for all variables. Next, we applied the Benjamini–Hochberg procedure to correct for multiple comparisons. The adjusted p values for the relation between CORT and effortful control, was p = .04. With regards to DHEA the corrected p value for sleep disturbances was p = .15, and academic stress, p = .04.

Table 1. Pearson correlations among all study variables

	1	2	3	4	5	6	7	8
1. Nail CORT^a	–	.28	−.06	.42**	.19	.08	−.12	.16
2. Nail DHEA^b		–	.12	.06	.16	.20	.42*	.50**
3. PSS^c			–	−.45**	.14	.30	.25	.08
4. Effortful control				–	−.16	−.04	−.20	−.14
5. Sleep efficiency					–	.52	.27	.23
6. Sleep quality						–	.45**	−.05
7. Daily dis.^d							–	.24
8. Academic Stress								–

^acortisol, ^bdehydroepiandrosterone, ^cperceived stress scale, ^ddaily disturbances.

*p < .05.

**p < .01 (all significance tests are two-tailed).

In our final set of analyses, to address missing data, we used Mplus 7.4 (Muthén & Muthén, 1998–2015) to re-run our analyses using multiple imputation. Our results were consistent, DHEA was still significantly and positively correlated with academic stress (p < .001) and sleep disturbances (p = .01). Self-control was positively associated with CORT (p = .005), and the CORT/DHEA ratio was positively associated with self-control (p = .01).

Discussion

The current study investigated feasibility of fingernail collection in a low-income African American sample, and examined the associations of nail CORT and DHEA with perceived subjective stress, self-control, and sleep. We found significant relationships between nail CORT and stressful events, self-control, as well as nail DHEA and sleep disturbances. However, consistent with research in hair, a more widely used marker of chronic exposure to CORT, there was no relationship among CORT, DHEA, and perceived subjective stress.

In our sample, higher levels of nail CORT was associated with better self-regulatory abilities. This is in contrast to past work (McLennan, Ihle, Streudte-Schmiedgen, Kirschbaum & Kliegel, 2016). However, our finding is consistent with another study utilizing African American adolescents and finding that better self-regulatory abilities were associated with higher levels of allostatic load, a multisystemic marker of chronic physiological stress (Brody et al., 2013). Moreover, an association between higher levels of hair CORT with better cognitive outcomes, including learning and memory have been reported (Pulopulos et al. 2014). Hair CORT was also found to be associated with lower levels of depressive symptoms (Gerber et al., 2013). This body of work suggests that in certain context, the ability to mount a stress response (rather than have a dampened one) may be adaptive.

Moreover, in our data, we found a relationship between nail DHEA, but not CORT in relation to sleep. In particular, nail DHEA was associated with night and daytime disturbances in sleep. DHEA has been implicated with sleep in several studies. In a study investigating the effects of partial sleep deprivation on individuals suffering from major depression, individuals for whom sleep deprivation had no effect on depression had significantly higher DHEA concentrations as compared to responders (Schule et al., 2003). DHEA administration has also been found to increase rapid eye movement during sleep (Friess et al., 1995). It is currently unclear whether these effects are exerted directly or through DHEA-induced changes in steroid metabolism.

Our data are consistent with past studies showing no relationship between chronic exposure to CORT (as indexed by hair) and subjective reports of stress. For example, hair CORT was not associated with either score on the perceived social stress scale (Milam, Slaugher, Verma & McConnell, 2014) or report of current stress levels using ecological momentary assessment (Gidlow, Randall, Gillman, Silk, & Jones, 2016). Moreover, no relationship was found between job stress and hair CORT (McLennan, Ihle, Steudte-Schmiedgen, Kirschbaum, & Kliegel, 2016). Additionally in children, parental reports of stress were not associated with children’s hair CORT levels (Gerber et al., 2017). Yet, our data also suggested that the presence of specific stressors and one’s subjective experience of that stressor was associated with nail DHEA, consistent with past work showing stressful life events, but not perceived stress was associated with nail (Izawa et al., 2017), and hair (Simmons et al., 2016) stress hormones. Together this body of work suggests that it is important to rethink the meaning of self-reported stress in relation to physiological indicators. It is likely, reports of stressful events may be better at capturing stress exposure as oppose to subject reports of general feelings of stress. Our data also extends past research by generalizing to adolescents, and demonstrates that academic stressors among adolescents can influence hormone levels. The extent that these academic stressors can affect biology, and potentially mental and physical health is an important consideration. We also provide preliminary data for the importance of nail DHEA. However, note that nail DHEA has not been validated with hair or saliva, despite having been found to be associated with stress (Tegethoff et al., 2011), thus future work should examine whether nail DHEA correlates with hair or saliva.

Our data is correlational, and it is not possible to test the direction of causality. Our sample is also small, and the majority of our sample is from an ethnic minority background, and thus caution should be taken concerning generalizability. Future studies should replicate the findings with a larger and more hetereogenous sample. Our current study was also exploratory in that as it was focusing on relations between nail hormone levels and psychosocial outcome. It would be important for future research to examine their independent contributions. For example, does academic stress predict hormone levels independent of sleep? Regardless of these limitations, these data are among the first to demonstrate that nail clipping may be a useful and informative approach to investigating the relationship between chronic exposure to CORT and DHEA to psychological outcomes in a sample of participants for whom traditional assessments of chronic stress would be more difficult to obtain.

Disclosure statement

No potential conflict of interest was reported by the authors.