Biomarkers of stress as mind–body intervention outcomes for chronic pain: an evaluation of constructs and accepted measurement

1. Introduction

Physiological mechanisms involved in the stress response may underlie chronic pain maintenance in youth and adults.^1,3,14,55,57 Research investigating these mechanisms has included measurement of the hypothalamic–pituitary–adrenal (HPA) axis^6,27 and allostatic load (AL),^8,49,55,58 including telomere attrition,^69–71 sympathoadrenal and sympathetic function,^7,83 and alterations in stress-based networks in the brain (eg, aspects of the fear network and limbic system).⁷⁵ Oxidative stress also has been identified as a salient physiological stress indicator.² Both in the context of chronic pain and more broadly, it has been suggested that these mechanisms may be modifiable in response to mind–body interventions, such as mindfulness training.^36,47,56 Specifically, evidence suggests that in response to these interventions, the normalization of an altered stress response and, in turn, sensory, emotional, and cognitive circuitry changes, would address central sensitization, leading to positive intervention effect.⁵⁶ However, direct investigation of these mechanisms as mind–body intervention outcomes for chronic pain has been minimal and, thus, remains poorly understood. Accordingly, the aims of this topical review are to provide an overview of potentially modifiable mechanisms of stress and pain and summarize currently accepted practices on their measurement, including evidence for timeframes that show malleability. We end with a call for research that focuses on using these mechanisms to objectively measure intervention responses in clinical trials for chronic pain.

2. Methods

Based on a previously published conceptual model,⁵⁶ we performed a review of the literature in 3 key areas: (1) AL, including telomere attrition and HPA-axis functioning; (2) oxidative stress; and (3) functional and structural brain changes in response to stress, with a focus on the fear-network and relevant areas of the limbic system. In reviewing the literature, we used the main search engines (ie, PubMed, Google Scholar) and focused on operationalization of each primary concept and accepted measurement in the context of behavioral health/mind–body clinical trials (eg, “allostatic load” + “mindfulness”). In addition to using the main constructs in the literature search described above, terms such as (but not limited to) “measurement,” “clinical trial,” “assay,” “f/MRI,” “biomarkers,” “intervention,” “psychosocial,” and “mind-body” were used in combinations with each of the above identified focus areas. Results were synthesized by content area and are summarized below.

3. Results

3.1. Mind–body intervention for chronic pain: objective domains vs the “mind”

Historically, the majority of nonpharmacological (eg, mind–body) interventions for chronic pain have used subjective outcomes to measure responsivity, such as self-report of reductions in pain intensity or functional disability, or reductions in anxiety or depressive symptoms associated with a chronic pain state.⁸² These self-report measures tend to capture the “mind” state (ie, a subjective experience) rather than objective outcomes for how the body changes in relation to the experience of pain, stress, anxiety, and depression. More recently, efforts have been made to identify objective measurements of mind–body interventions for chronic pain, with a recent review highlighting functional neuroimaging as a primary approach.¹⁹ However, outside of imaging functional and structural brain changes, objective outcomes of mind–body intervention that (1) are feasible to measure in research and clinical contexts and (2) capture responsivity to change (to evidence appropriateness as an intervention outcome) remain sparse. Previous conceptual work has proposed the physiological stress–response as an appropriate target for mind–body intervention,⁵⁶ which opens several valuable avenues of measurement. Modifiable mechanisms that can be measured by objective methods are summarized below by substantive area.

3.2. Biological indicators of an altered stress response

3.2.1. Allostatic load

Burgeoning research has identified AL, defined as nervous system wear and tear in response to repeated or prolonged stress, as a potential underlying mechanism of chronic pain.^8,49,55,58 Outside of pain, research has indicated that youth with a higher number of AL risk factors (measurement described below) evidence poorer health and higher rate of healthcare utilization.^65,79 More broadly, research indicates that health outcomes are poorer in youth with a history of maltreatment,^35,60 indicating a strong association between stress exposure and physical health in youth.

3.2.2. Telomere attrition

Telomere attrition or the shortening of telomere length (ie, the ends of chromosomes that protect against chromosomal damage) has been examined both in relation to pain (eg, data from adults link shortened telomeres to several chronic pain conditions)^30,71,74 and in the context of AL as an indicator of cellular aging and disease susceptibility after exposure to stress or maltreatment.^26,61,62 Research also has indicated that physiological sequelae relating to stress exposure, such as cortisol reactivity, play significant roles in adaptation to stressors long-term and may contribute to telomere length dynamics, including shorter telomere length and more rapid telomere attrition over time.^26,37,53,64 Frequency of exposure to stressors (eg, violence) is also predictive of greater telomere attrition.⁴⁵

3.2.3. Oxidative stress

Cellular damage resulting from an imbalance of antioxidants and free radicals, known as oxidative stress, has been linked repeatedly to severe psychological stress, inflammation, and biological aging, including telomere shortening.² In particular, a large body of research has indicated an association between early life stress and activation of pathophysiological oxidative pathways.^33,38 Research also indicates that oxidative stress is associated with mental health pathology, both independently and in association with early life stress.^33,52 In the context of chronic pain, a growing body of literature has connected maintenance of chronic pain to biological factors related to oxidative stress (eg, free radicals, inflammation) as well as to associated risk for oxidative stress, such as poor diet or exercise and exposure to psychological stress.² To date, these relations have yet to be investigated in pediatric samples. Nevertheless, oxidative stress has been proposed to be related to telomere attrition, AL, and pain maintenance among pediatric as well as adult populations.^34,40,48,76

3.2.4. Functional and structural brain changes

A large body of research has identified substantial changes in several aspects of brain function or structure in response to early life stress and ongoing posttraumatic stress.^12,13,22 It has been noted that healthy brain circuitry in multiple brain regions in addition to various networks (eg, default mode network, sensorimotor) is remodeled in response to stressful experiences to enable protective responses to these threats,^13,17,73 potentially leading to short- or long-term remodeling or volumetric changes in a variety of loci.^23,32 Notably, research has captured overlap between areas of the brain involved in stress and in pain processes.⁵⁷ For example, early life stress has been associated with impact on the development of circuits throughout the limbic and fear network systems (eg, prefrontal cortex, hippocampus, and amygdala). Reduced hippocampal volume, in turn, has been studied in relation to its observed impact on pain, memory, and HPA axis function.^12,31 Evidence also suggests that cumulative stress (eg, severe, chronic neglect) is associated with heightened reactivity to provoked stressors (eg, emotional images) in the amygdala,^16,28 a brain area involved in nociception responsivity.⁷²

3.3. Evidence of modifiability by mind–body intervention?

A large body of research has indicated that mind–body interventions are capable of modifying aspects of the physiological stress response, both in the moment (eg, heart rate, respiration) and in the long term (eg, AL, oxidative stress). However, evidence for the responsivity to change of these constructs is inconsistent and difficult to capture in the context of pain management. Specifically, there has yet to be a standardized guide on the magnitude of change in relation to treatment response (ie, what is indicative of a “therapeutic dose”), in addition to specific evidence regarding which variables are expected to change under what conditions or in what timeframe or dose (eg, number of sessions). In order for greater adoption of objective measurement of intervention responses, research needs to delineate the responsivity of these mechanisms (eg, dose–response, therapeutic benefit for pain intensity, physical functioning, and stress reduction) for broader adoption by researchers and clinicians alike. Research on biological responsivity by substantive area is included below.

Within the area of AL, minimal research has used the full construct (ie, multifactorial measurement of AL risk) to assess intervention response. However, aspects of AL, such as glucocorticoid release and HPA axis functioning, have shown responsivity to mindfulness-based strategies.⁹ Glucocorticoids (eg, cortisol), a primary construct of AL and HPA axis functioning, have been studied most extensively compared with other aspects of AL in the context of intervention response, with studies showing cortisol alterations after psychosocial intervention in as little as 3 to 4 months' time.^44,59,84 Other mind–body interventions, such as art therapy and yoga, have shown promise in alleviating AL,^20,66 but these studies have primarily been pilot investigations with small sample sizes. Further research is needed to understand the usefulness and feasibility of measuring AL in intervention studies, particularly those involving chronic pain, given the hypothesized role of AL in pain maintenance in children and adults.^8,49,55

Telomere attrition, a marker of cellular aging and an aspect of AL,⁵ has shown responsivity to mind–body intervention through slowing of the rate of attrition.²⁵ A recent meta-analysis examining the association between mindfulness meditation and telomere length suggested that individuals who underwent meditation training exhibited longer telomeres when compared with individuals who did not meditate.⁶⁸ Recent research examining the impact of an intensive meditation retreat suggested that measurable changes in telomere regulation can be captured in a 3-week time frame.¹⁸ However, the measurement and interpretation of telomere length or attrition has been controversial because many researchers argue that there is not a plausible biological mechanism by which telomeres can be lengthened; at best, the rate of telomere attrition may be slowed by protective health behaviors.⁸⁰ It is also notable that telomere measurement demands specificity as either length or attrition. The former measurement remains most meaningful in cross-sectional studies (ie, when one's relative telomere length is compared with others in the sample in relation to variable(s) of interest with the latter requiring multiple measurement over a defined period [eg, preintervention to postintervention]).^41,45 Future research involving the use of telomere length dynamics as an intervention outcome in chronic pain (and broadly) should take these considerations into account.

It also has been proposed that oxidative stress may be responsive to mind–body intervention through targeting the subjective experiences of stress, which then leads to alterations in the expression of stress hormones and in turn, oxidative stress levels in the body.²⁵ Oxidative stress has been associated with psychiatric illness (eg, depression, posttraumatic stress disorder²⁹), and evidence suggests that psychotherapy intervention that directly targets these symptoms can address oxidative stress as well.^10,67 Irrespective of mental health or emotional functioning, psychotherapy has been found to decrease biomarkers of oxidative stress in the body in immediate postintervention (after 7 sessions) and follow-up (6 months later) time points.³⁹ These findings suggest that outcome measurements associated with oxidative stress are responsive in periods amenable to intervention research. Furthermore, evidence suggests that managing oxidative stress and inflammation may also address telomere dynamics because antioxidant and anti-inflammatory processes are associated with longer telomeres.³⁴ Future research should examine these processes in relation to the known psychological and physical correlates of chronic pain conditions and treatment responses.

3.4. An overview of integrative measurement strategies

As mentioned above, recent efforts have been made to more definitively identify objective measurements of mind–body interventions for chronic pain, with a recent review highlighting functional neuroimaging as a primary approach.¹⁹ In the stress literature, there is also an abundance of evidence that indicates biological samples, such as saliva, blood, or urine, may be valuable, but the application of these methods to pain intervention work remains limited. A summary of research on these measurement strategies in each major methodology category is provided below (Fig. 1).

Figure 1.

Overview of regulatory systems and integrative measurement methods of the physiological stress response.

3.4.1. Neuroimaging

Functional neuroimaging (eg, functional MRI [fMRI]) is one of the most common methods to objectively measure intervention outcomes in mind–body clinical trials. For example, a number of pain intervention trials have successfully captured alterations in several areas of the brain relevant to both pain and stress physiology, such as the prefrontal cortex, fear-network, default mode network, limbic system, and somatosensory cortex.^43,77,85 Moreover, changes in some of these areas have been captured after as little as 80 minutes of mindfulness practice,^85–87 which indicates that fMRI can be an appropriate and effective method to capture initial intervention response. However, engaging patients/participants in fMRI can be burdensome and is costly.^24,54 As such, research that investigates alternative methods to measure mind–body treatment responses is needed. Biological sampling may be a useful and cost-effective avenue to examine treatment responses, particular in the areas of stress physiology as detailed in prior sections.

3.4.2. Biological samples

The construct of AL is most effectively examined by a multifactorial composite of measures across multiple regulatory systems.⁸³ Common markers included in the AL construct represent dysregulation across metabolic (eg, waist-to-hip ratio, cholesterol, low-density lipoprotein, high-density lipoprotein, hemoglobin A1c, body mass index); cardiovascular (eg, systolic blood pressure, diastolic blood pressure); neuroendocrine (cortisol, dehydroepiandrosterone sulfate); and autonomic nervous system (epinephrine/norepinephrine) functioning.^36,46,47 More recently, markers of immune system functioning (eg, C-reactive protein) and telomere attrition have also been included.⁵ As a composite, measuring AL in the context of pain treatment has been shown to be feasible in pilot investigations in both children and adults.^50,51,55 Typical methods to measure aspects of AL include saliva sampling (cortisol, dehydroepiandrosterone, C-reactive protein^4,11,15); blood sampling (high-density lipoprotein/low-density lipoprotein, hemoglobin A1c^4,83); urine sampling (epinephrine/norepinephrine⁷); and biometric measurement (blood pressure, body mass index, and waist-to-hip ratio¹⁵). Telomere measures can be assessed from DNA extracted from saliva or venous blood draws, or less commonly, dried blood spots.^26,42,63 Interpretation of these factors can vary, but a common method to examine AL is by a “risk” ratio, where indicators of dysfunction across area are coded dichotomously in reference to the larger sample means or to accepted clinical metrics (eg, high vs normal blood pressure).^15,83 Research indicates that 2 or more indicators of AL indicate “high risk.”⁷⁸ There is no specific guidance on the number and type of AL indicators to include, which underscores the need for further research in the area. However, as mentioned above, AL measurement has shown good reliability when including multiple indicators of physiological dysregulation in complementary aspects of functioning.⁸³ AL as a construct may be particularly amenable to prospective mind–body clinical trials with multiple follow-up timepoints, as measurements included in the AL composite are cost-effective to store and assay, when compared with imaging, which, as mentioned above, is more expensive, invasive, and time-intensive. Future research should also investigate the most appropriate AL indicators for specific conditions (eg, feasible and responsive metrics for pain populations).

Accepted methodology for oxidative stress measurement includes examining reactive oxygen species/reactive nitrogen species metabolites, antioxidant levels, and activities related to antioxidant enzyme and related gene expression.^33,52 Research indicates that the F2-isoprostane/creatinine ratio is the “gold standard” for measuring oxidative stress,^52,76 with elevation of F2-isoprostanes interpreted as maladaptive oxidative stress.³⁸ The primary biological sample required to measure these metabolites is urine sampling, a noninvasive and cost-effective sampling procedure. Indirect ways for measuring oxidative stress may include telomere length or attrition and DNA methylation age.³⁸

4. Conclusions/future directions

The lack of objective markers of mind–body intervention response is a glaring issue within the chronic pain field, with evidence of effectiveness predominantly being guided by self-reported subjective measures. By using objective, as well as self-reported measures, we could better understand the effectiveness of treatments that are currently being provided and improve future treatment efforts. Objective markers of mind–body intervention response may also aide in the elucidation of the placebo effect. For example, previous research has indicated placebo to be effective on physiological stress but not pain.⁸¹ Evidence also suggests that the impact of mindfulness on pain may be placebo analgesia rather than mindfulness-based analgesia.²¹ However, we are still lacking a great amount of knowledge around the functionality, reliability, and validity of these biological measurements. More research must be done to understand the connection of biomarkers such as AL, oxidative stress, telomere attrition, and brain changes to chronic pain and stress, as well as the effects of mind–body treatment on these biological features. Such efforts will allow for the standardization of the specific changes and “doses” appropriate for short-term and long-term health changes.

Conflict of interest statement

The authors have no conflicts of interest to declare.