Brief Report: Trait Emotion Regulation Strategies and Diurnal Cortisol Profiles in Healthy Adults

Abstract

Objective

Experimental studies have shown that two emotion regulation strategies - suppression and reappraisal - are associated with differential profiles of physiological activation in response to a stress test. The present study aims to add to those findings by investigating whether individual differences in trait emotion regulation strategies are associated with diurnal cortisol patterns in a naturalistic context.

Method

A sample of 46 men and women from the Midlife in the United States II study completed the Emotion Regulation Questionnaire and provided four salivary cortisol samples per day over four consecutive days. Trait reappraisal and suppression were tested as predictors of three cortisol parameters averaged across days: cortisol awakening response (CAR), diurnal cortisol slope (DCS), and area under the curve with respect to ground (AUCg).

Results

Higher scores on the suppression scale were associated with more physiological activation, as indicated by steeper CAR and flatter DCS. Suppression was not associated with AUCg, and reappraisal was not predictive of any cortisol parameter.

Conclusions

Individual differences in suppression, but not reappraisal, were linked to greater cortisol activation in this naturalistic study. These preliminary results add to a growing body of findings that link suppression to adverse psychological and physiological profiles.

Emotions are whole-body phenomena that give rise to response tendencies that involve changes in subjective experience, behavior and physiology (Lazarus, 1991). Consequently, the strategies that individuals use to regulate their emotions can have implications for physical and mental health. Emotion regulation strategies can be classified into antecedent-focused strategies, such as reappraisal, where regulatory behaviors occur prior to emotion response tendencies, and response-focused strategies, such as suppression, which influence the experiential, behavioral and physiological aspects of the emotional response (Gross & John, 2003). According to the process model of emotion regulation (Gross, 1998), reappraisal acts early in the emotion-generative process and entails reevaluating a possibly emotion-eliciting situation, thereby changing its meaning and altering its emotional impact. Suppression acts later in the emotion-generative process, after an event has been deemed emotionally relevant and involves inhibiting or reducing outward expression of emotional arousal (Gross, 1998). Modulating emotional responses (suppression) rather than one’s appraisal of an emotional situation (reappraisal) is thought to be more effortful, and the regulation is achieved at a higher cost. Accordingly, suppression and reappraisal are associated with differential psychological and physiological profiles. In particular, previous research has linked reappraisal to greater experience and expression of positive emotion and decreased experience and expression of negative emotion, as well as fewer symptoms of depression. Suppression on the other hand, has been linked to decreased expression and experience of positive emotion and greater experience of negative emotion, a feeling of incongruence and elevated depressive symptoms (Gross & John, 2003). Physiologically, reappraisal has been linked to more adaptive patterns of physiological responding (e.g. Egloff, Schmukle, Burns, & Schwerdtfeger 2006; Mauss, Cook, Cheng, & Gross, 2007) as well as to beneficial health outcomes such as decreased levels of inflammation and decreased cardiovascular disease risk (Appleton, Loucks, Buka & Kubzansky 2014). Suppression has been associated with heightened sympathetic nervous system activation, higher levels of inflammation and increased cardiovascular disease risk (Appleton, Buka, Loucks, Gilman, & Kubzansky, 2013; Appleton et al., 2014; Gross, 1998; Gross & Levenson, 1997).

Among the physiological domains found to be affected by the regulation of emotion is the neuroendocrine secretion of cortisol. Diurnal cortisol rhythms of healthy individuals are characterized by a pronounced increase 30–45 minutes after waking (cortisol awakening response, CAR) and a steady decline over the remainder of the day (diurnal cortisol slope, DCS), reaching the nadir approximately 16 hours after waking (Miller et al., 2016). Cortisol exerts wide-ranging effects on the immune, metabolic and central nervous systems (Sapolsky, Romero, & Munck, 2000) and can thus be considered a biomarker of psychological and physiological health. In fact, adverse physical and mental health outcomes have been linked to altered cortisol patterns of flattened diurnal slopes and, both, enhanced and reduced cortisol awakening responses. General life stress, obesity, and depression have been linked to increased CAR, while posttraumatic stress disorder, fatigue syndrome, cardiovascular disorders, depression and, interestingly, positive psychological traits have been related to reduced CAR (for reviews see, Chida & Steptoe, 2009; Fries, Dettenborn, & Kirschbaum, 2008).

Previous research on emotion regulation and cortisol has mainly focused on neuroendocrine reactivity to a stressor in a laboratory setting. For example, trait suppression and reappraisal, as well as experimental induction of reappraisal, have been linked to higher cortisol reactivity in response to a psychosocial stressor (Denson, Creswell, Terides, & Blundell, 2014; Lam, Dickerson, Zoccola, & Zaldivar, 2009). In contrast, longer-term cognitive behavioral training (CBT) - of which reappraisal is a primary, however, not isolated aspect - has been shown to decrease cortisol reactivity to a psychosocial stress test, relative to controls without CBT (Gaab et al., 2003). To our knowledge, associations of trait emotion regulation strategies and diurnal cortisol patterns in healthy adults have not been studied in a naturalistic setting before. It is important to examine these associations outside the laboratory in daily life, as diurnal patterns in salivary cortisol are predictive of health outcomes and mortality (e.g., Kumari, Shipley, Stafford, & Kivimaki, 2011).

The present study investigates whether individual differences in suppression and reappraisal are linked to diurnal cortisol patterns in a naturalistic context. Based on the process model of emotion regulation (Gross, 1998) and previous research, we hypothesize that the higher cost of regulating emotions with suppression will be reflected in more activated patterns of diurnal cortisol, as indicated by steeper CAR, flatter DCS, and higher total cortisol output (area under the curve with respect to ground, AUCg). Additionally, we expect that reappraisal will be associated with healthier patterns of steeper DCS and lower AUCg, whereas the CAR could be either enhanced (suggesting an adaptive boost to face the demands of the day; Adam, Hawkley, Kudielka, & Cacioppo, 2006) or reduced (indicating less physiological activation; Chida & Steptoe, 2009).

Methods

Participants

This study uses cross-sectional data from the MIDUS II study, which was designed to examine how behavioral, psychological and social factors accounted for variations in health and well-being in a national sample of non-institutionalized, English-speaking adults. MIDUS II contains projects that include a daily diary substudy called the National Study of Daily Experiences (NSDE; Ryff & Almeida, 2009) and a neuroscience substudy (Ryff & Davidson, 2006). Assessments of salivary cortisol were obtained from the NSDE, and a questionnaire on emotion regulation was administered in the neuroscience substudy. The emotion regulation questionnaire was added to the study protocol late; of the 331 participants in the neuroscience substudy, 118 were administered the emotion regulation questionnaire. Of these 118, 46 participants also completed the saliva collection protocol in the NSDE. The final sample was 50% female and 63% Caucasian with a mean age of 54.04 years (SD = 10.24). Procedures were approved by Institutional Review Boards at the University of Wisconsin, the Pennsylvania State University, the University of California - Los Angeles, and Georgetown University and all participants provided informed consent.

Procedures and Measures

Emotion Regulation

Participants completed the Emotion Regulation Questionnaire (ERQ; Gross & John, 2003), which assessed the typical use of suppression and reappraisal. The suppression scale consisted of 4 items (e.g., “I control my emotions by not expressing them”), and the reappraisal scale consisted of 6 items (e.g., “I control my emotions by changing the way I think about the situation I’m in”). Ratings were made on a 7-point Likert scale (1 = “strongly disagree” to 7 = “strongly agree”) and averaged within each subscale. The subscales had good internal consistency reliability (Cronbach’s alpha = 0.80 for suppression and 0.74 for reappraisal).

Salivary cortisol

The saliva collection protocol for the NSDE has been described in detail elsewhere (Almeida, McGonagle, & King, 2009). Briefly, participants were instructed to collect four saliva samples per day (immediately upon waking, 30 min after waking, before lunch, and before bed) for four consecutive days using salivette collection devices (Sarstedt, Nümbrecht, Germany). As several substances can compromise salivary immunoassays by altering pH levels and impairing antibody binding reactions, participants were instructed to collect the samples before eating, drinking, brushing their teeth, or consuming any caffeinated products. The pH levels were checked and corrected (pH 4 – 9) prior to immunoassay for cortisol (Almeida, McGonagle, & King, 2009). Participants recorded the time of each saliva sample on a paper log, in addition to reporting the sampling times during nightly telephone interviews. Salivettes were processed and assayed at the MIDUS Biological Core at the University of Wisconsin, Madison.

Based on previous research on emotions and diurnal cortisol (e.g., Adam et al., 2006), we calculated three diurnal cortisol indices. CAR was calculated as 30-min post-waking cortisol level minus waking cortisol level, divided by time (in hours) between the two samples. DCS was calculated as bedtime cortisol minus waking cortisol, divided by hours between the two samples. AUCg was computed using the trapezoid formula from Pruessner, Kirschbaum, Meinlschmid, and Hellhammer (2003). The diurnal cortisol indices were calculated for each day, then averaged across days for each participant. Days were excluded from the calculation of the cortisol indices if (1) saliva collection time stamps were missing, (2) the participant woke up after 12 PM, (3) the participant was awake <12 hours or >20 hours, or (4) if there was an indication of non-compliance with the saliva collection protocol such that <15 or >60 minutes elapsed between the first two measurements (Stawski, Cichy, Piazza, & Almeida, 2013). The analytic sample sizes were 46 participants for DCS and 43 participants for CAR and AUCg.

Covariates

We chose to control for the following variables based on previous related research on cortisol (e.g., Adam & Kumari, 2009; Stawski et al., 2013) and on emotion regulation (Gross & John, 2003): age, gender, body mass index (BMI), smoking status, use of cortisol-altering medications, and self-rated health. As part of the parent MIDUS II study, participants rated their physical health on a 5-point scale (excellent, very good, good, fair, or poor). As scores increase, self-rated health decreases. Height and weight were measured at a clinic visit and used to calculate BMI. During daily diary telephone interviews across 8 evenings in the NSDE, participants reported daily cigarette smoking. A dummy-coded variable was created to indicate any smoking (1) versus no smoking (0). Another dummy-coded variable indicated the use of any cortisol-altering medication during saliva collection period (including cortisone, steroids, antidepressants, hormonal, and anti-anxiety medications). Data on daily smoking and medication use during saliva collection were missing for seven and five participants, respectively. For these participants, smoking status and use of corticosteroid and/or sex hormone medications were obtained from the biomarker substudy in MIDUS II. Daily stressors were reported in the daily diary study (Almeida, Wethington, & Kessler; 2002) but were not correlated with any of the cortisol parameters in the present study. To keep the models parsimonious, we did not include daily stressors in further analyses.

Statistical Analyses

Descriptive analyses and correlations were run to examine participant characteristics, emotion regulation, and cortisol variables. For the primary analyses, suppression and reappraisal were examined separately in linear regression models as predictors of CAR, DCS, and AUCg. Models were adjusted for age, gender, BMI, smoking (yes/no), medications (yes/no), and self-rated health (0–4 scale). Analyses were conducted using SAS 9.4.

Results

Descriptive analyses and correlations

Participants were 46 adults (50% female; 63% white) with a mean age of 54.04 years (SD = 10.24). Further descriptives are presented in the Online Supplementary Material. Cortisol values exhibited a diurnal pattern, in which the mean (SD) cortisol level at waking was 13.91 (6.13) nmol/L, increased to 19.90 (6.95) nmol/L 30-min after waking, declined to 6.63 (3.03) nmol/L before lunch, and further declined to 3.65 (2.83) nmol/L at bedtime.

Correlations between participant characteristics and emotion regulation showed that males had higher trait suppression (r = 0.49, p < 0.001), and with increasing age there was a marginal increase in trait reappraisal (r = 0.28, p = 0.06). Partialing for age, smoking and higher BMI were correlated with flatter DCS (r = 0.31 and r = 0.38, respectively, p’s < 0.05). In addition, as suppression increased, the slope of CAR increased (r = 0.32, p = 0.04) while the DCS got marginally flatter (r = 0.27, p = 0.08), partialing for age and gender. Suppression was not correlated with AUCg, and reappraisal was not correlated with CAR, DCS, or AUCg.

Emotion regulation strategies as predictors of cortisol

In linear regression models, trait reappraisal was not associated with any of the three cortisol parameters (Table 1). Suppression was positively correlated with the steepness of the slope of CAR, controlling for age and gender. This association remained after further adjustment for trait reappraisal, BMI, medications, smoking, and self-rated health. Suppression was also associated with flatter DCS but was not predictive of AUCg.

Table 1.

Emotion suppression and reappraisal as predictors of diurnal cortisol parameters

Variable	Cortisol awakening response (n = 43) Unstandardized B (SE)			Diurnal cortisol slope (n = 46) Unstandardized B (SE)			AUCg (n = 43) Unstandardized B (SE)
	Model 1a	Model 2b	Model 3c	Model 1a	Model 2b	Model 3c	Model 1a	Model 2b	Model 3c
Intercept	9.36 (1.66)***	9.53 (1.73)***	14.35 (2.19)***	−0.67 (0.06)***	−0.68 (0.06)***	−0.62 (0.08)***	131.32 (6.81)***	131.86 (6.94)***	136.21 (10.42)***
Suppression	3.16 (1.52)*	-----	3.01 (1.38)*	0.09 (0.05)†	-----	0.11 (0.05)*	8.05 (6.25)	-----	9.57 (6.54)
Reappraisal	-----	1.92 (2.10)	2.23 (1.83)	-----	−0.05 (0.07)	−0.05 (0.06)	-----	0.26 (8.44)	−2.04 (8.69)
Age	−0.16 (0.18)	−0.09 (0.18)	−0.23 (0.16)	0.01 (0.01)	0.01 (0.01)	0.01 (0.01)	0.22 (0.72)	0.47 (0.72)	−0.05 (0.77)
Genderd	−5.39 (2.00)**	−2.86 (1.80)	−6.46 (2.02)**	−0.14 (0.07)†	−0.08 (0.06)	−0.18 (0.07)*	0.32 (8.19)	5.97 (7.24)	1.71 (9.61)
BMI			0.70 (0.25)**			0.02 (0.01)**			0.44 (1.20)
Medication use			−11.25 (4.09)**			−0.25 (0.15)†			6.92 (19.42)
Current smoker			−7.30 (3.85)†			0.07 (0.14)			−28.91 (18.31)
Self-rated health			−3.91 (1.93)†			0.02 (0.07)			−0.09 (9.17)

^***p ≤ .001,

^**p ≤ .01,

^*p ≤ .05,

^†p ≤ .10

^aModel 1 tests suppression as a predictor of cortisol, adjusting for age and gender.

^bModel 2 tests reappraisal as a predictor of cortisol, adjusting for age and gender.

^cModel 3 includes additional covariates for BMI, use of any cortisol-altering medications (1=yes, 0 = no), smoking status (1=yes, 0 = no), and self-rated physical health (0–3 scale).

^dGender was effect coded (−1 = female, 1 = male), such that associations between emotion regulation and cortisol can be interpreted as the effects averaged across females and males.

Discussion

The present study adds to previous research on trait emotion regulation and cortisol by examining diurnal salivary cortisol in healthy adults in daily life. Results indicate that individual differences in suppression are linked to diurnal cortisol parameters. Specifically, as trait suppression increased, the steepness of the CAR curve increased, the DCS got flatter, and the AUCg remained unchanged. Trait reappraisal showed no associations with any of the three cortisol parameters.

The findings from this naturalistic study are consistent with previous laboratory research showing a link between trait suppression and physiological activation. On a broader level, these results are in line with the long-held notion that inhibition of emotional expression has negative effects on physical and mental health, especially when inhibition occurs chronically (for review, see Consedine, Magai, & Bonanno, 2002). Suppression is cognitively and physically effortful (e.g., Richards & Gross, 1999), which may place cumulative stress on the body and increase vulnerability to illness (e.g., Appleton et al., 2014).

The null results for reappraisal were unsurprising given the mixed findings in the literature on reappraisal and physiological activity (Denson et al., 2014; Egloff et al., 2006; Gaab et al., 2003; Lam et al., 2009). The equivocal literature suggests that there could be differences in cortisol patterns between experimentally-manipulated reappraisal—which might increase cortisol reactivity in the short-term—compared to trait reappraisal, which has been proposed to buffer against cortisol reactivity over the long-term as it is more automatic and less effortful (Denson et al., 2014; Gaab et al., 2003). Similarly, the timeframe of effects (such as immediate reactivity to a stressor vs. longer-term diurnal cortisol patterns) and contexts of emotion-eliciting situations (such as laboratory vs. naturalistic contexts) might be critical elements to consider when examining physiological correlates of emotion regulation strategies (Aldao, 2013). As each of those elements might be associated with different cortisol output, they may need to be considered separately to obtain consistent results. Lastly, the null findings between reappraisal and diurnal cortisol could be due to reporting bias, such that participants’ perceptions of how they typically regulate their emotions may differ from their actual emotion regulation in daily life.

Limitations & Future Directions

Our findings came from a small sample and therefore should be interpreted as preliminary. Caution should be taken in generalizing the results to other populations. As this was a cross-sectional study, inferences about causality cannot be made. Emotion regulation was assessed using a questionnaire administered at a single time point; we did not collect data on people’s actual use of emotion regulation strategies in real-life situations during the same days as the saliva collection. Future research could collect repeated assessments of state emotion regulation, emotion-eliciting situations, and diurnal cortisol in daily life to examine whether the use of suppression versus reappraisal strategies are linked to concurrent or subsequent changes in cortisol. Furthermore, future research is needed to examine diurnal cortisol rhythms as potential pathways linking trait emotion regulation to downstream health outcomes.