A systematic review of the effects of mindfulness interventions on cortisol

Abstract

Cortisol is increasingly included in examinations of mindfulness intervention effects as an indicator of efficacy; however, the association of cortisol and mindfulness has yet to be rigorously evaluated. A systematic review of six studies examining mindfulness intervention effects on cortisol was conducted. Inconsistent results were found for mindfulness effects on cortisol. Significant changes in cortisol levels were observed in within-participants studies but not observed in randomised controlled trial designs. Mindfulness may influence cortisol, but findings are inconclusive. Mindfulness pathways and methodological differences influence variations in mindfulness effects. Robust protocols are needed to adequately examine mindfulness effects on cortisol.

Keywords

cortisol intervention mindfulness well-being

Mindfulness is the intentional and non-judgemental awareness of experience in the present moment. It is consistently associated with health and well-being in diverse clinical and non-clinical populations (e.g. Fjorback et al., 2011; Keng et al., 2011). Mindfulness is associated with higher levels of quality of life (Grossman et al., 2007), life satisfaction, happiness and positive affect (Bogels et al., 2008; Brown and Ryan, 2003). It is associated with better sleep quality and sleep functioning (Howell et al., 2010) and with lower levels of stress (Carlson et al., 2001; Shapiro et al., 2005; Speca et al., 2000), depression and anxiety (Brown and Ryan, 2003; Monti et al., 2006; Shapiro et al., 2005). Examinations of the effects of mindfulness on biomarkers of well-being are in their infancy. A recent review highlighted the usefulness of cortisol as an outcome measure in mindfulness research (Matousek et al., 2010). However, it did not systematically evaluate the effects of mindfulness interventions on cortisol outcomes or the efficacy of current research approaches. Thus, studies examining the effects of mindfulness on cortisol have not yet been systematically evaluated.

Mindfulness incorporates an orientation to experience characterised by openness and acceptance of all internal and external stimuli in the present moment (Keng et al., 2011). It also involves regulation of awareness (Bishop et al., 2004; Keng et al., 2011), typically through directing attention to a point of focus such as the breath. Several mindfulness programmes and interventions have been developed, including Mindfulness-Based Stress Reduction (MBSR) and Mindfulness-Based Cognitive Therapy (MBCT). MBSR emphasises mindfulness meditation training to foster acceptance and non-judgemental observation of stimuli (Kabat-Zinn, 1982); it has consistently demonstrated benefits in clinical and non-clinical populations (Keng et al., 2011). MBCT was developed to help prevent recurrence of major depressive episodes by addressing maladaptive response patterns to distress and depressive episodes (Teasdale et al., 1995). MBSR and MBCT generally consist of 8-week-long courses, with weekly sessions of 2–2.5 hours and at-home mindfulness practice 6 days a week. A daylong intensive retreat is often included. MBSR and MBCT are two of the most widely used mindfulness interventions and have guided much of the research in the area.

Two reviews of randomised controlled trials (RCTs) examining the effects of MBSR and MBCT (Keng et al., 2011; Sedlmeier et al., 2012) reported improvements in psychosocial outcomes, such as quality of life, depression, anxiety and stress. Such interventions can also improve markers of health, such as chronic pain, fibromyalgia and psoriasis (Baer, 2003). To date, the reviews have not addressed evaluations of biomarkers of well-being or measures of biological outcomes. They instead focus on studies using self-report indices of health and well-being. Incorporating standardised markers of health and well-being allows for more rigorous investigations of the pathways between psychological and physical well-being and health outcomes.

An important mediator between psychological states and health-related outcomes is cortisol. Cortisol is the end product of hypothalamic pituitary adrenal (HPA) axis functioning, which helps regulate the stress response. Perception of a stressor induces the release of corticotropin releasing hormone (CRH) from the paraventricular nucleus in the hypothalamus. CRH then travels to the pituitary gland, and adrenocorticotropin hormone (ACTH) is secreted into the blood stream. ACTH binds to receptors in the adrenal cortex that stimulate secretion of cortisol. This stress response system most optimally regulates stress when a quick onset is required followed by a quick termination once a threat has passed (Kudielka and Kirschbaum, 2005). Dysregulations of the HPA axis are associated with chronic and/or prolonged stress and with adverse physical and psychological sequelae. HPA hyperactivity, for instance, is associated with major depression, susceptibility to disease and cardiovascular problems (Kudielka et al., 2012). Hyporeactivity is associated with autoimmune processes such as lupus, multiple sclerosis and chronic fatigue (Kudielka and Kirschbaum, 2005). Optimal functioning of the HPA axis is vital for physical and psychological well-being.

Salivary cortisol is a commonly used measure of physiological stress responses as it is a minimally invasive method of ambulatory assessment (Kudielka et al., 2012) and provides information about cortisol levels and changes over periods of time (Chida and Steptoe, 2009). Typically, cortisol follows a distinct diurnal pattern. Levels of cortisol begin to rise at waking and peak about 30–45 minutes later; this peak is followed by a steady decline throughout the day. The rise in cortisol levels from waking to peak is known as the cortisol awakening response (CAR). It was first established by Pruessner et al. (1997b) as a useful marker of change in cortisol levels in the first hour after waking. As a marker of HPA axis functioning, the CAR is particularly useful because it occurs naturally. Waking is a consistent, robust and sufficient stimulus that is necessary for the CAR to occur (Adam and Kumari, 2009; Clow et al., 2009). The CAR is not strongly associated with cortisol levels for the rest of the day but both the CAR and diurnal patterns are associated with psychological and physical well-being (Chida and Steptoe, 2009). Flatter diurnal slopes and higher diurnal cortisol output are associated with perceived stress in healthy adults (Lovell et al., 2011). Similarly, the CAR is positively associated with job stress and general life stress (Chida and Steptoe, 2009; Clow et al., 2004; Wust et al., 2000). It is also negatively associated with fatigue, burnout and depression (Chida and Steptoe, 2009). Examining the effects of mindfulness on the CAR and diurnal patterns of cortisol is important to contribute to models of mindfulness effects on health and well-being.

A number of studies have demonstrated reductions in cortisol subsequent to mindfulness interventions (Galantino et al., 2005; Kang and Oh, 2012; Lengacher et al., 2012; Lipschitz et al., 2013). Brand et al. (2012) found that participating in an MBSR programme decreased the CAR for both experienced and novice meditators. In clinical populations, changes in cortisol levels have been observed for individuals who have completed cancer treatment (Matousek et al., 2011), current cancer patients (Carlson et al., 2004; Lengacher et al., 2012) and their caregivers (Lengacher et al., 2012). For instance, in a group of breast and prostate cancer outpatient, Carlson et al. (2004) found significant decreases in morning and evening cortisol following an MBSR intervention. These decreases were specific to participants with higher levels of baseline mean daily cortisol. Participants with low or potentially blunted baseline cortisol demonstrated increases over time, suggesting a normalisation of cortisol levels in this group.

In a RCT of a 3-week mindfulness meditation intervention with cancer survivors, Lipschitz et al. (2013) failed to find significant changes in cortisol levels. Similarly, Bowden et al. (2012) did not find significant differences in cortisol levels between mindfulness and two other wellness interventions. However, both studies utilised active controls and did not include a no-treatment control against which the effects of the interventions could be compared. More robust examinations of MBSR (Matchim et al., 2011) and a low-dose MBSR programme (Klatt et al., 2009) also failed to demonstrate significant changes over time between the interventions and no-treatment control groups for working adults (Klatt et al., 2009) or early stage breast cancer survivors (Matchim et al., 2010). Matchim et al. (2011) did observe a significant decrease in cortisol following a mindfulness intervention for the intervention condition only, similar to effects observed in other within-participants designs (Carlson et al., 2004; Galantino et al., 2005; Lengacher et al., 2012). While these findings suggest that mindfulness interventions only demonstrate within-participants effects, Lynch et al. (2011) found no significant changes over time for university students following a mindfulness meditation intervention. In addition to not finding within-participants effects in a sample of college students Lynch et al. (2011) also failed to find significant differences between the intervention and control groups over time.

The inconsistent and contrary findings of mindfulness effects on cortisol, coupled with its increasing incorporation into studies of mindfulness, indicate a need to take stock of the value of cortisol measures in examinations of mindfulness interventions. As this is a rapidly growing field of enquiry, a review of the literature is essential to investigate whether mindfulness does demonstrate beneficial effects on cortisol. An evaluation of current research design, sampling protocols and cortisol measurement in mindfulness research is pressing; without it there is the potential for perpetuation of inappropriate research practices when combining two research areas which have been examined independently until recently. The primary aim of this review was, therefore, to systematically evaluate the effect of mindfulness interventions on salivary cortisol levels. A secondary aim was to evaluate current research approaches in the area of mindfulness effects on cortisol levels.

Methods

Study eligibility criteria

Eligibility criteria included examination of mindfulness interventions, 6–8 weeks duration, published in English. Interventions were required to involve training and teaching of mindfulness techniques, emphasising systematic practice. Studies were required to include salivary cortisol measures of the CAR and/or diurnal slope as an outcome variable. Both within-subject studies and RCTs were included if measurements were collected pre- and post-intervention. Studies including participants with psychotic illnesses or bipolar disorder were excluded due to disturbances in cortical regulations that may influence interpretation of intervention outcomes (Chida and Steptoe, 2009).

Search strategy

A systematic search of the literature was conducted using Medline, Academic Search Complete, PsycARTICLES, PyscINFO, CINAHL, Web of Science and Science Direct. The search terms ‘mindfulness’, ‘cortisol’, ‘cortisol awakening response’ and ‘awakening cortisol response’ were used, and the search was not limited by the date of publication.

Study and data collection process

Prior to study commencement, ethical approval was obtained from the appropriate University Ethics Committee. K.O.L. designed the search strategy, and both K.O.L. and S.O.N. executed it. The two assessors independently reviewed titles and abstracts of all identified articles and independently evaluated each full text article. Reference lists of retrieved articles were hand searched for further potential studies. This resulted in retrieval of 71 articles in total. Disagreements were resolved by consensus. The 71 potential articles were assessed using the inclusion and exclusion criteria; this resulted in 12 articles remaining for quality assessment.

Quality assessment

Study quality was assessed by evaluating six types of bias using Review Manager 5.2. Articles were considered to have high risk of bias if inadequate information about randomisation and blinding (if any) was provided or if groups were not randomly assigned. High risk of bias was also associated with unclear study design, unaccounted for attrition rates and inadequate measures of cortisol. Inadequate cortisol measures included baseline cortisol after intervention commencement or use of single cortisol samples. Following quality assessment, six articles remained, all with sample sizes ranging from 21 (Marcus et al., 2003) to 186 (Malarkey et al., 2013), total n = 378. When studies only provided descriptive statistics, effect sizes were calculated using the medians and ranges. A power analysis for meta-analysis was conducted to assess the feasibility of conducting a meta-analysis. The power analysis was conducted on five articles using means and standard deviations; where these were unavailable, effect sizes were calculated from median scores and ranges. The power analysis resulted in an estimated power of 0.08, indicating that the studies demonstrated lower than desirable power to examine significant effects. Due to the low power and the weak effect sizes of the included studies, a systematic review was deemed most appropriate.

Results

A summary of all studies examining the effects of mindfulness interventions on cortisol levels is shown in Table 1. Two studies utilise within-participants designs (Marcus et al., 2003; Matousek et al., 2011). These studies demonstrate high methodological quality, and their lack of control conditions can be considered ethically sound, as they include vulnerable populations. The remaining studies utilise RCT designs. As design is a potential contributor to differences in results, the findings will be presented by research design.

Table 1.

Characteristics of included studies.

Study	Year	Design	Population	N intervention	N controls	Saliva sampling protocol	Cortisol measurement	Intervention	Outcome
Marcus et al.	2003	Within subjects	Therapeutic community for substance abuse	21		1 day pre- and post-intervention, waking, +30, +45, +60	CAR	8-week MBSR	p = .03
Matousek et al.	2011	Within subjects	Breast cancer patients	33		3 days pre- and post-intervention, waking, +30, +45	CAR	8-week MBSR	p = .05
Gex-Fabry et al.	2012	RCT	Patients remitted with recurrent depression	30 (28)	30 (28)	Six collections: pre- and post-intervention and 3-, 6-, 9-, and 12-month follow-up. Waking, +15, +30, +45, +60, 3 p.m., 8 p.m.	CAR and diurnal slope	8-week MBCT	CAR p = .73; slope p = .36
Oken et al.	2010	Pilot RCT	Community dwelling caregivers	10 (8)	21 (17)	1 day pre- and post-intervention, +5, +35 and bedtime	CAR	7-week MBCT	+5 minutes p = .615; +35 minutes p = .175; bedtime p = .209
Daubenmier et al.	2011	RCT	Overweight and obese women	24	23	4 days pre- and post-intervention, waking, +30, bedtime	CAR and diurnal slope	9-week combination of MBSR, MBCT and MB-EAT	CAR p = .15; slope p = .58
Malarkey et al.	2013	RCT	University faculty and staff	93 (84)	93 (86)	3 days pre- and post-intervention, +20, 5 p.m., bedtime	Diurnal slope	8-week MBI-ld	p = .61

CAR: cortisol awakening response; RCT: randomised controlled trial; MBSR: Mindfulness-Based Stress Reduction; MBCT: Mindfulness-Based Cognitive Therapy; MB-EAT: Mindfulness-Based Eating Awareness Training; MB–ld: Mindfulness Based Intervention- low dose.

Within-participants examinations of mindfulness effects

The two within-participants examinations of mindfulness intervention effects included in this review (Marcus et al., 2003; Matousek et al., 2011) focus exclusively on the CAR. Notably, these are the only studies reviewed that demonstrate changes in cortisol levels over time.

Marcus et al. (2003) conducted a pilot study of a standard MBSR programme with 21 members of a therapeutic community for substance abuse. Salivary cortisol was measured using samples collected on the first morning pre-intervention and again at post-intervention at the following times: waking, +30, +45 and +60 minutes. No information is given regarding time of waking or whether variability in participant waking times was accounted for; this can have a significant impact on the CAR (Clow et al., 2004; Federenko et al., 2004; Kudielka et al., 2012; Kudielka and Kirschbaum, 2003). Information regarding the time frame between completing the intervention and collecting post-intervention saliva samples is also unclear. In addition, sample collection on a single-day pre- and post-intervention is problematic, as multiple sampling days are required to rigorously measure cortisol levels (Hellhammer et al., 2007; Powell and Schlotz, 2012). Marcus et al. (2003) calculated the CAR as area under the curve (AUC) using the trapezoid method with four time points. The analysis only included data from a small subsample of 12 participants who provided complete data; additional information about this subgroup, such as socio-demographic variables, is not provided. Despite this, Marcus et al. (2003) found a significant reduction in CAR from pre-test to post-test, even with a small sample size of 12 participants. One interpretation of this is that there is a strong effect of mindfulness on CAR; however, it is possible that the findings are attributable to a lack of rigour in timing of samples.

Matousek et al. (2011) also conducted a within-participants examination of the effects of MBSR on the CAR for 33 women who had completed medical treatment for cancer. Salivary cortisol measures were collected three times a day over three consecutive days pre- and post-intervention; samples were collected at waking and +30 and +45 minutes. Cortisol measures were collected within the 5 days preceding and following the intervention, and further specific information is not given. It is unclear whether participants collected samples on weekdays and/or weekends; the CAR is influenced by the day of the week and can differ between work and non-work days (Hellhammer et al., 2007; Kudielka et al., 2012; Kunz-Ebrecht et al., 2004; Schlotz et al., 2004). Matousek et al. (2011) examined the effect of sampling day in terms of stability of sampling within each sampling period and found no significant effect of day of sample collection on cortisol levels. The CAR was calculated AUC with respect to increase (AUC_i) (Pruessner et al., 2003). Cortisol levels were also transformed into a single value for correlational analyses by calculating the AUC for each day and then calculating the mean for the individual day values. While this approach has its merit, it also loses information regarding intra-individual differences across days by removing potential effects of individual days for individual participants. Matousek et al. (2011) found a significant difference in the CAR from pre- to post-test; the CAR increased following intervention completion, with prolonged increases at post-test measures. This indicates that the MBSR programme significantly increases the CAR and that observed increases are influenced by depressive symptomatology in women who have completed breast cancer treatment.

Summary of within-participants findings

The two within-participants examinations of mindfulness intervention effects included in the current review demonstrate significant changes in the CAR. Directionality of effects differed between studies, perhaps indicating differing mechanisms of mindfulness in different groups. For members of a therapeutic community treated for substance abuse, a reduction in cortisol levels was observed (Marcus et al., 2003). Individuals who had completed medical treatment for cancer demonstrated an increase in cortisol levels (Matousek et al., 2011). It is important to note that due to the lack of a control condition, our interpretation of these findings is limited.

RCT examinations of mindfulness effects

A number of RCTs have examined the effect of mindfulness interventions on cortisol levels, in comparison to control groups. Oken et al. (2010) conducted one such examination in a pilot RCT of 31 healthy adults providing care for a family member with dementia. Two control conditions were used. The first, powerful tools for caregivers, involved weekly sessions on topics including stress management and decision-making. The second was a respite only condition with respite care 3 hours weekly, for 7 weeks. Experimental conditions were not compared to caregivers maintaining usual routines. Inclusion criteria specified providing 12 hours minimum caregiving assistance weekly and having ‘high enough’ (p. 1032) baseline stress levels. The latter criterion potentially denies benefits to people below this cut-off and presupposes intervention utility in high-stress cases only. As this is a newly developed intervention based on MBSR and MBCT, such a presupposition may also artificially inflate intervention effects; high baseline levels may result in greater decreases. Single-day samples were collected within 3 weeks pre- and post-intervention, and information on standardisation of collection days is absent; this allows for variability between participants (Clow et al., 2004; Hellhammer et al., 2007; Powell and Schlotz, 2012; Stalder et al., 2010). Saliva samples were collected at three time points: within 5 minutes of waking, 30 minutes later and at bedtime (approximately 10–11 p.m.). No further information on sampling times is provided, such as variations in the waking or bedtime samples. This lack of specificity can result in inaccurate measurement of intervention efficacy. In addition, morning cortisol levels can peak between 30 and 45 minutes after waking, and several factors can influence this morning peak (Clow et al., 2004). For instance, the CAR peak can occur later in females, and most participants in this study were female. Thus, the sampling times used by Oken et al. (2010) may explain the lack of a significant effect in this study.

Gex-Fabry et al. (2012) also found no significant effect of a mindfulness intervention on the CAR or diurnal slope of 60 patients remitted from depression, following an MBCT intervention. Salivary cortisol samples were collected at baseline, intervention completion, and at 3-month intervals for 1 year. Single-day sample collection was used; samples were collected at waking, +15 minutes, +20 minutes, +45 minutes, +60 minutes, 3 p.m. and 8 p.m. Analysis of the CAR was conducted using AUC according to the trapezoid rule (Pruessner et al., 2003). No significant changes were observed in comparison to a treatment-as-usual (TAU) control group at post-intervention or follow-up. The use of single-day samples may contribute to the absence of observed effects. Similarly, no significant effect on diurnal slope was observed at post-intervention or follow-up. The diurnal slope was calculated using the difference between first and last samples divided by the time interval between samples. Absence of a bedtime sample limits analysis of the overall diurnal pattern. AUC for overall day levels was calculated using the trapezoid rule, with differences in sampling interval normalised to the mean 13-hour sampling period. The use of a median value of 13 does not take into account individual variations in sampling times, and it is unclear why differences in sampling interval are not allowed for (Pruessner et al., 2003). Despite this, the use of a TAU control group is the strength, indicating that mindfulness interventions may not be as beneficial when compared with TAU for such samples.

Daubenmier et al. (2011) examined the effects of mindfulness on cortisol for 47 overweight and obese women, body mass index (BMI) between 25 and 40. The intervention used, Mindfulness-Based Eating Awareness Training (MB-EAT), was based on MBSR and MBCT. It lasted 9 weeks and incorporated practices before and during meals. The wait-listed control group engaged in a 2-hour nutrition and exercise session halfway through the study. The rationale for this is unclear, as it provides neither a comparative level of engagement nor the relative neutrality of a TAU group. Samples were collected over four work days, pre- and post-test; they were collected at waking, +30 minutes and bedtime. The bedtime sample was reported as ‘just prior to bedtime’ (p. 4), without further clarification. The use of a +30-minute sample may also miss a CAR peak occurring later for women (Clow et al., 2004). Significant reductions in the CAR were observed for obese women in the treatment condition that were not observed for obese women in the control condition. Significant differences were not found for overweight women in either condition. This suggests potential benefits for the obese subgroup only. No changes in diurnal pattern were found for obese or overweight women in the study. MB-EAT does not, therefore, demonstrate an effect on the CAR or diurnal pattern in overweight and obese women in comparison to a control group; some within-participants benefits are apparent for obese women.

In a RCT of 186 university faculty and staff, Malarkey et al. (2013) examined the effect of mindfulness on biological measures of inflammation and chronic stress, including diurnal cortisol. An 8-week-long intervention was used; the intervention was characterised as low dose because the duration of weekly group sessions and at-home practice was reduced from standard MBSR and MBCT programmes. The daylong mindfulness retreat was also replaced by a 2-hour retreat. A lifestyle education control condition spent a comparable amount of time engaging with study exercises. Salivary cortisol was collected for three consecutive days at 2 weeks pre-intervention and 2 weeks post-intervention. Samples were collected at the following times: 20 minutes after waking, noon, 5 p.m. and bedtime. No further information is given for waking and bedtime sample collection, such as variance within and between participants. It is also unclear whether rising and bedtime occurred at pre-specified times or were determined by participants. The primary cortisol outcome, the average of three cortisol measurements (noon, 5 p.m. and bedtime) across 3 days, provides afternoon and evening cortisol information only. The morning sample provides little information about cortisol at waking or the CAR, however, as it is taken at a time unassociated with either. The difference between the +20-minute sample and the average of the 5 p.m. and bedtime levels was used in a sensitivity analysis. No significant differences in cortisol levels between the two groups following intervention completion were found.

Summary of RCT findings

No significant effects of mindfulness interventions were observed for the CAR or diurnal slope in comparison to control conditions. No significant effects were observed for dementia caregivers (Oken et al., 2010), individuals remitted from depression (Gex-Fabry et al., 2012), working adults (Malarkey et al., 2013) or obese and overweight women (Daubenmier et al., 2011). When a subgroup of obese women in the latter study was examined as a within-participants group, a reduction in the CAR was observed (Daubenmier et al., 2011). Issues of sampling strategies, intervention use and differential control groups may contribute to inconsistent findings.

Discussion

This systematic review evaluated the effect of mindfulness interventions on cortisol. Examinations of mindfulness effects on biomarkers of well-being are in their infancy and have not yet been subject to a rigorous review. The current article highlights some potential effects of mindfulness interventions on cortisol levels and the number of issues that constrain investigations and interpretations of mindfulness effects to date.

Significant changes in the CAR have been observed using within-participants designs in diverse groups, such as members of a substance abuse community (Marcus et al., 2003), obese women (Daubenmier et al., 2011) and individuals who have completed medical treatment for cancer (Matousek et al., 2011). In research designs utilising control conditions, mindfulness does not demonstrate significant effects on the CAR or diurnal slope for people remitted from depression (Gex-Fabry et al., 2011), dementia caregivers (Oken et al., 2010), obese and overweight women (Daubenmier et al., 2011) or working adults (Malarkey et al., 2013). The inconsistent findings of mindfulness effects on cortisol indicate the need for caution in interpreting the research findings. They also highlight the importance of robust research design, such as controlled trials, in studies of mindfulness and biology as such designs have yet to demonstrate significant effects of mindfulness.

Mindfulness optimises HPA axis functioning

A recurrent issue in the cortisol literature is that the direction of change observed for the CAR differs between studies (Fries et al., 2009). Inconsistent cortisol patterns have been found in association with a wide range of well-being variables across numerous studies (Mikolajczak et al., 2010). Mikolajczak et al. (2010) posit that examining cortisol flexibility, rather than increases or decreases in cortisol levels, may account for these inconsistencies and provide a more consistent account of cortisol functioning over time and across studies. In an examination of the influence of protective psychological factors, such as high happiness and low stress, Mikolajczak et al. (2010) found that people exhibiting these factors demonstrated a more flexible CAR. These participants had increased reactivity to challenge as a result of transitioning from weekend to weekday; individuals with dysregulated and less flexible CAR did not respond as well. In line with this flexibility hypothesis (Mikolajczak et al., 2010), mindfulness may function as a protective psychological factor to optimise cortisol functioning to induced or naturally occurring stressors. Mindfulness does not simply reduce stress in the moment or over time.

The flexibility hypothesis explains findings that mindfulness influences differential changes in cortisol based on baseline functioning (Carlson et al., 2004). It may also explain the findings of significant increases in the CAR from pre- to post-intervention for women who had completed breast cancer treatment (Matousek et al., 2011). Cancer patients may have had blunted CAR at baseline because the CAR is negatively associated with depression, fatigue and post-traumatic stress disorder (Chida and Steptoe, 2009); these outcomes can result from stressful cancer-related experiences and treatments. Matousek et al. (2011) provide support for blunted cortisol in their study, as their participants demonstrated lower cortisol levels than those found for healthy controls (Wust et al., 2000). Therefore, mindfulness functions to improve responding in this instance by increasing and thereby normalising cortisol levels.

Gex-Fabry et al. (2012) failed to find similar increases in cortisol levels following MBCT for patients remitted from depression. They expected a normalisation of blunted cortisol levels as the CAR is negatively associated with depression; this did not occur. The use of MBCT, designed to reduce depressive symptomatology rather than stress, may explain the lack of a significant effect. Similarities between MBSR and MBCT presuppose the utility of both interventions to influence stress, hence the inclusion of both in this review. It is possible that baseline effects on cortisol over time result from noise variables, such as health behaviours and socio-demographic factors (Adam and Kumari, 2009) that are present at baseline but may differ by follow-up. To fully investigate this, robust examinations of mindfulness interventions that incorporate such variables are needed to determine whether mindfulness truly exerts a meaningful effect on cortisol.

Direct effects of mindfulness on cortisol

Depression adversely influences cortisol levels (Chida and Steptoe, 2009), and mindfulness interventions demonstrate reductions in levels of depression (Brown and Ryan, 2003; Goyal et al., 2014; Monti et al., 2006). It could be assumed that a mindfulness-induced reduction in depression leads to lower cortisol levels, but there is little evidence to support this. It is equally possible that HPA axis functioning influences depression. This latter possibility is supported by the lack of a significant reduction in cortisol over time following MBCT (Gex-Fabry et al., 2012), despite observed reductions in depression in the original article (Bondolfi et al., 2010). Reductions in cortisol in people with depression, following a mindfulness intervention, do not, therefore, appear to result from improvements in depressive symptomatology. The findings of this review suggest that observed reductions in cortisol could occur through a direct mechanism that is not moderated by depression. Further support for this conclusion comes from the lack of association between cortisol and depression in a mindfulness trial in a sample of healthy adults (Malarkey et al., 2013). Additionally, Matousek et al. (2011) found that changes in the CAR following intervention use were more pronounced after controlling for depression. This indicates that depression does not moderate mindfulness effects and suggests potential direct effects of mindfulness on cortisol levels. While effects are observed in within-subjects designs, the extant literature does not provide sufficient support for direct effects of mindfulness based on the findings of controlled trials.

Mindfulness interventions in different subgroups

It is possible that mindfulness results in different effects on cortisol levels for different participant groups. For instance, within-participants mindfulness effects were observed for a group of obese women who were not observed in overweight participants in the same study (Daubenmier et al., 2011). Cortisol secretion differs between obese and overweight women (Kumari et al., 2010; Rutters et al., 2010), but these findings suggest that obese individuals receive more benefit from mindfulness interventions. This may extend to other groups demonstrating pervasive or clinically meaningful disease statuses. A heightened level of underlying stress and/or physical symptomatology may elicit a greater cortisol response to mindfulness. This would explain the lack of change in cortisol levels in the healthy, non-clinical populations included in this review: caregivers (Oken et al., 2010) and university faculty and staff (Malarkey et al., 2013). Previous research has also suggested that lack of effects are due to relatively normal baseline cortisol (Lipschitz et al., 2013) or a lack of room for further change due to significantly lowered cortisol towards the end of intervention use (Lengacher et al., 2012).

Methodological issues

A number of methodological issues may explain any observed intervention effects. The studies included in the current review use inconsistent approaches to cortisol sampling, research design and intervention adherence and attrition monitoring. The number of studies existing but excluded from this review highlight the pressing need for improvement in the methodological quality of examinations of mindfulness effects on cortisol.

Cortisol sampling

Implementation of appropriately structured sampling protocols is essential to accurately measure cortisol levels. This is because cortisol can be influenced by time of sampling and incorrect sampling (Adam and Kumari, 2009; Clow et al., 2004; Kudielka and Kirschbaum, 2003). Frequency and duration of sampling is often influenced by feasibility and cost, but sample collection protocols must still be adequate to capture variability in cortisol levels. In this article, incomplete information on timing of night-time samples (Daubenmier et al., 2011; Oken et al., 2010) limits our interpretation of the accuracy and meaningfulness of intervention effects on diurnal slope. It also reduces the standardisation of sampling times across participants within studies. Similarly, incomplete information for some waking times (Marcus et al., 2003) also allows for variability between participants. As time of waking and sampling during the early morning rise period can have a significant effect on the CAR (Clow et al., 2004; Kudielka and Kirschbaum, 2003), accurate CAR sampling is essential.

Additionally some time points included are ill suited to capture intended cortisol outcomes. For instance, some studies do not sample for the CAR beyond 30 minutes after waking (Daubenmier et al., 2011; Oken et al., 2010). It is well accepted that cortisol can peak between 30 and 45 minutes post waking, with this peak tending to occur later in women (Clow et al., 2004). Using a shorter sampling period for the CAR may miss this peak, thereby contributing to the absence of an effect. Restrictive sampling times for diurnal pattern were also noted, with some studies failing to adequately capture overall diurnal secretion (Malarkey et al., 2013). Restrictive duration of sampling periods, in terms of single-day samples, was also found (Gex-Fabry et al., 2012; Marcus et al., 2003; Oken et al., 2010). Multiple sampling days at each time point are necessary to adequately measure cortisol levels and minimise the influence of contextual and situational effects (Hellhammer et al., 2007; Powell and Schlotz, 2012). Not accounting for potential differences across days and within participants provides an incomplete and inaccurate measure of cortisol.

Conversely, multiple day measures must be standardised across all participants. Some studies in the current review do not indicate whether sampling occurred on weekdays and/or weekends (Matousek et al., 2010) allowing for variation across participants (Hellhammer et al., 2007; Kudielka et al., 2012; Mikolajczak et al., 2010; Schlotz et al., 2004). Potential variation may also occur when participants are instructed to collect samples within a certain time period pre- and post-intervention, without specific sampling days used (Marcus et al., 2013; Matousek et al., 2010; Oken et al., 2010). Future research on mindfulness effects must include standardised collection days and times to reduce potential variation between participants. Adequate sampling protocols that collect sufficient samples over a sufficient duration must also be implemented.

Population groups and research designs

The majority of studies included in this review have low sample sizes, raising the possibility that these studies may be underpowered to detect a significant effect. Additionally, the diverse populations examined present challenges for coherently interpreting the effects of mindfulness interventions. Goyal et al. (2014) recently highlighted this issue in a review and meta-analysis of meditation programmes for psychological well-being. Mikolajczak et al. (2010) have also highlighted that due to inconsistencies between studies, no clear guideline is available for what a healthy CAR should look like. As a result, interpretations of effects on cortisol levels are limited. Similarly, here we cannot gauge what an acceptable or healthy cortisol response to a mindfulness intervention is, due to the diversity of populations examined.

A further issue of note is the use of within- and between-participants research designs. When studying vulnerable or clinical patients, it is not often ethically sound to conduct RCTs due to potential denial of benefits. However, within-participants designs lack a useful control group against which intervention effects can be compared. Interestingly, the only significant effects observed in the current review are in within-participants designs. When control groups are utilised, effects are not observed, but this issue is less clear in the instances when control groups utilise alternative forms of intervention but do not include a non-intervention group. While this appears to imply within-participants effects of mindfulness only, it more clearly suggests that when more robustly examined using the RCT design, mindfulness interventions have not demonstrated any significant effects.

Conclusion

Mindfulness may have an influence on cortisol levels, although caution in interpreting the findings of extant results is warranted. Methodological differences are likely to have contributed to inconsistent findings across studies. Future research should adhere to robust methodological standards in terms of control groups, sampling protocols and intervention use. This is essential to facilitate a better understanding of effects and the development of efficacious mindfulness interventions.

Footnotes

Acknowledgements

The authors would like to acknowledge Dr Carol Linehan and Gillian Murphy of University College Cork for their assistance in refining and revising this article.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Adam

Kumari

(2009) Assessing salivary cortisol in large-scale, epidemiological research. Psychoneuroendocrinology 34(10): 1423–1436.

Baer

(2003) Mindfulness training as a clinical intervention: A conceptual and empirical review. Clinical Psychology: Science and Practice 10(2): 125–143.

Bishop

Lau

Shapiro

. (2004) Mindfulness: A proposed operational definition. Clinical Psychology: Science and Practice 11(3): 230–241.

Bogels

Hoogstad

van Dun

. (2008) Mindfulness training for adolescents with externalizing disorders and their parents. Behavioural and Cognitive Psychotherapy 36(2): 193–209.

Bondolfi

Jermann

Van der Linden

. (2010) Depression relapse prophylaxis with Mindfulness-Based Cognitive Therapy: Replication and extension in the Swiss health care system. Journal of Affective Disorders 122(3): 224–231.

Bowden

Gaudry

. (2012) A comparative randomised controlled trial of the effects of brain wave vibration training, iyengar yoga, and mindfulness on mood, well-being, and salivary cortisol. Evidence-Based Complementary and Alternative Medicine 1: 1–13.

Brand

Holsboer-Trachsler

Naranjo

. (2012) Influence of mindfulness practice on cortisol and sleep in long-term and short-term meditators. Neuropsychobiology 65(3): 109–118.

Brown

Ryan

(2003) The benefits of being present: Mindfulness and its role in psychological well-being. Journal of Personality and Social Psychology 84(4): 822–848.

Carlson

Speca

Patel

. (2004) Mindfulness-based stress reduction in relation to quality of life, mood, symptoms of stress and levels of cortisol, dehydroepiandrosterone sulfate (DHEAS) and melatonin in breast and prostate cancer outpatients. Psychoneuroendocrinology 29(4): 448–474.

10.

Carlson

Ursuliak

Goodey

. (2001) The effects of a mindfulness meditation-based stress reduction program on mood and symptoms of stress in cancer outpatients: 6-month follow-up. Supportive Care in Cancer 9(2): 112–123.

11.

Chida

Steptoe

(2009) Cortisol awakening response and psychosocial factors: A systematic review and meta-analysis. Biological Psychology 80(3): 265–278.

12.

Clow

Hucklebridge

Stadler

. (2009) The cortisol awakening response: More than a measure of HPA axis function. Neuroscience and Biobehavioural Reviews 35(1): 97–103.

13.

Clow

Thorn

Evans

. (2004) The awakening cortisol response: Methodological issues and significance. Stress 7(1): 29–37.

14.

Daubenmier

Kristeller

Hecht

. (2011) Mindfulness intervention for stress eating to reduce cortisol and abdominal fat among overweight and obese women: An exploratory randomized controlled study. Journal of Obesity 1: 1–13.

15.

Federenko

Wust

Hellhammer

. (2004) Free cortisol awakening responses are influenced by awakening time. Psychoneuroendocrinology 29(2): 174–184.

16.

Fjorback

Arendt

Ornbol

. (2011) Mindfulness-Based Stress Reduction and Mindfulness-Based Cognitive Therapy – A systematic review of randomized controlled trials. Acta Psychiatrica Scandinavica 124(2): 102–119.

17.

Fries

Dettenborn

Kirschbaum

(2009) The cortisol awakening response (CAR): Facts and future directions. International Journal of Psychophysiology 72(1): 67–73.

18.

Galantino

Baime

Maguire

. (2005) Association of psychological and physiological measures of stress in health-care professionals during an 8-week mindfulness meditation program: Mindfulness in practice. Stress and Health: Journal of the International Society for the Investigation of Stress 21(4): 255–261.

19.

Gex-Fabry

Jermann

Kosel

. (2012) Salivary cortisol profiles in patients remitted from recurrent depression: One-year follow-up of a mindfulness-based cognitive therapy trial. Journal of Psychiatric Research 46(1): 80–86.

20.

Goyal

Singh

Sibinga

EMS

. (2014) Meditation programs for psychological stress and well-being: A systematic review and meta-analysis. Journal of the American Medical Association Internal Medicine 174(3): 357–368.

21.

Grossman

Tiefenthaler-Gilmer

Rays

. (2007) Mindfulness training as an intervention for fibromyalgia: Evidence of postintervention and 3-year follow-up benefits in well-being. Psychotherapy and Psychosomatics 76(4): 226–233.

22.

Hellhammer

Fries

Schweistha

. (2007) Several daily measurements are necessary to reliably assess the cortisol rise after awakening: State- and trait components. Psychoneuroendocrinology 32(1): 80–86.

23.

Howell

Digdon

Buro

(2010) Mindfulness predicts sleep-related self-regulation and well-being. Personality and Individual Differences 48(4): 419–424.

24.

Kabat-Zinn

(1982) An outpatient program in behavioral medicine for chronic pain patients based on the practice of mindfulness meditation: Theoretical considerations and preliminary results. General Hospital Psychiatry 4(1): 33–47.

25.

Kang

(2012) Effects of Mindfulness Meditation program on perceived stress, ways of coping, and stress response in breast cancer patients. Journal of Korean Academy of Nursing 42(2): 161–170.

26.

Keng

Smoski

Robins

(2011) Effects of mindfulness on psychological health: A review of empirical studies. Clinical Psychology Review 31(6): 1041–1056.

27.

Klatt

Buckworth

Malarkey

(2009) Effects of low-dose mindfulness-based sress reduction (MBSR-ld) on working aduls. Health Education Behaviour 36(3): 601–614.

28.

Kudielka

Kirschbaum

(2003) Awakening cortisol responses are influenced by health status and awakening time but not by menstrual cycle phase. Psychoneuroendocrinology 28(1): 35–47.

29.

Kudielka

Kirschbaum

(2005) Sex differences in HPA axis responses to stress: A review. Biological Psychology 69(1): 113–132.

30.

Kudielka

Gierens

Hellhammer

. (2012) Salivary cortisol in ambulatory assessment – Some dos, some don’ts, and some open questions. Psychosomatic Medicine 74(4): 418–431.

31.

Kumari

Chandola

Brunner

. (2010) A nonlinear relationship of generalized and central obesity with diurnal cortisol secretion in the Whitehall II study. Journal of Clinical Endocrinology and Metabolism 95(9): 4415–4423.

32.

Kunz-Ebrecht

Kirschbaum

Marmot

. (2004) Differences in cortisol awakening response on work days and weekends in women and men from the Whitehall II cohort. Psychoneuroendocrinology 29(4): 516–528.

33.

Lengacher

Kip

Barta

. (2012) A pilot study evaluating the effect of mindfulness-based stress reduction on psychological status, physical status, salivary cortisol, and interleukin-6 among advanced-stage cancer patients and their caregivers. Journal of Holistic Nursing 30(3): 170–185.

34.

Lipschitz

Kuhn

Kinney

. (2013) Reduction in salivary alpha-amylase levels following a mind-body intervention in cancer survivors – An exploratory study. Psychoneuroendocrinology 38(9): 1521–1531.

35.

Lovell

Moss

Wetherell

(2011) Perceived stress, common health complaints and diurnal patterns of cortisol secretion in young, otherwise healthy individuals. Hormones and Behavior 60(3): 301–305.

36.

Lynch

Gander

Kohls

. (2011) Mindfulness-based coping with university life: A non-randomized wait-list-controlled pilot evaluation. Stress and Health 27(5): 365–375.

37.

Malarkey

Jarjoura

Klatt

(2013) Workplace based mindfulness practice and inflammation: A randomized trial. Brain Behavior and Immunity 27(1): 145–154.

38.

Marcus

Fine

Moeller

. (2003) Change in stress levels following mindfulness-based stress reduction in a therapeutic community. Addictive Disorders & Their Treatment 2(3): 63–68.

39.

Matchim

Armer

Stewart

(2010) Effects of mindfulness-based stress reduction (MBSR) on health among breast cancer survivors. Western Journal of Nursing Research 33(8): 996–1016.

40.

Matousek

Dobkin

Pruessner

(2010) Cortisol as a marker for improvement in mindfulness-based stress reduction. Complementary Therapies in Clinical Practice 16(1): 13–19.

41.

Matousek

Pruessner

Dobkin

(2011) Changes in the cortisol awakening response (CAR) following participation in mindfulness-based stress reduction in women who completed treatment for breast cancer. Complementary Therapies in Clinical Practice 17(2): 65–70.

42.

Mikolajczak

Quoidbach

Vanootighem

. (2010) Cortisol awakening response (CAR)’s flexibility leads to larger and more consistent associations with psychological factors than CAR magnitude. Psychoneuroendocrinology 35(5): 752–757.

43.

Monti

Peterson

Shakin Kunkel

. (2006) A randomized, controlled trial of mindfulness-based art therapy (MBAT) for women with cancer. Psycho-Oncology 15(5): 363–373.

44.

Oken

Fonareva

Haas

. (2010) Pilot controlled trial of mindfulness meditation and education for dementia caregivers. Journal of Complementary and Alternative Medicine 16(10): 1031–1038.

45.

Powell

Schlotz

(2012) Daily life stress and the cortisol awakening response: Testing the anticipation hypothesis. PLoS ONE 7(12): 1–10.

46.

Pruessner

Gaab

Hellhammer

. (1997a) Increasing correlations between personality traits and cortisol stress responses obtained by data aggregation. Psychoneuroendocrinology 22(8): 615–625.

47.

Pruessner

Kirschbaum

Meinlschmid

. (2003) Two formulas for computation of the area under the curve represent measures of total hormone concentration versus time-dependent change. Psychoneuroendocrinology 28(7): 916–931.

48.

Pruessner

Wolf

Hellhammer

. (1997b) Free cortisol levels after awakening: A reliable biological marker for the assessment of adrenocortical activity. Life Sciences 61(26): 2539–2549.

49.

Rutters

Nieuwenhuizen

Lemmens

SGT

. (2010) Hypothalamic-Pituitary-Adrenal (HPA) axis functioning in relation to body fat distribution. Clinical Endocrinology 72(6): 738–743.

50.

Schlotz

Hellhammer

Schulz

. (2004) Perceived work overload and chronic worrying predict weekend-weekday differences in the cortisol awakening response. Psychosomatic Medicine 66(2): 207–214.

51.

Sedlmeier

Eberth

Schwarz

. (2012) The psychological effects of meditation: A meta-analysis. Psychological Bulletin 138(6): 1139–1171.

52.

Shapiro

Astin

Bishop

. (2005) Mindfulness-based stress reduction and health care professionals. International Journal of Stress Management 12(2): 164–176.

53.

Speca

Carlson

Goodey

. (2000) A randomized, wait-list controlled clinical trial: The effect of a mindfulness meditation-based stress reduction program on mood and symptoms of stress in cancer outpatients. Psychosomatic Medicine 62(5): 613–622.

54.

Stalder

Evans

Hucklebridge

. (2010) Associations between psychosocial state variables and the cortisol awakening response in a single case study. Psychoneuroendocrinology 35(2): 209–214.

55.

Teasdale

Segal

Williams

JMG

(1995) How does cognitive therapy prevent depressive relapse and why should attentional control (mindfulness) training help. Behaviour Research and Therapy 33(1): 25–39.

56.

Wust

Wolf

Hellhammer

. (2000) The cortisol awakening response – Normal values and confounds. Noise and Health 2(7): 79–88.