Abstract

Introduction

Mind-body therapies, or MBTs, have been broadly defined as a group of therapies that emphasize use of the brain in conjunction with the body to assist the healing process (Spencer, 2003). These therapies, the majority of which are based on ancient practices and traditions, are believed to have beneficial effects on mental and physical health and are widely used to manage symptoms and improve well-being. Indeed, in a nationwide survey of community-dwelling adults in the US conducted in 2007, 19% reported that they had used at least one mind-body therapy in the past year, and rates are even higher among clinical populations (Barnes, Bloom, & Nahin, 2008). Over the past two decades, the efficacy of these approaches has been subjected to empirical scrutiny through randomized controlled trials conducted in clinical and non-clinical populations. Meta-analyses of these trials suggest that MBTs are effective in reducing symptoms and improving quality of life and certain functional outcomes (Bussing, Ostermann, Ludtke, & Michalsen, 2012; Goyal et al., 2014; Wang, Collet, & Lau, 2004).

Alterations in inflammatory processes are thought to play a role in many of the symptoms and conditions that are responsive to MBTs, including fatigue, depression, and pain (Irwin & Cole, 2011). Given the importance of inflammation on these patient-reported outcomes, a growing number of trials have evaluated effects of MBTs on markers of inflammation. The goal of this review is to qualitatively evaluate the evidence that MBTs lead to changes in these markers, and to discuss potential mechanisms and issues in this emerging area of research. We include results of randomized controlled trials that assessed inflammatory cytokine activity at multiple levels, including circulating, cellular, and genomic markers of inflammation.

We focus here on four types of MBTs that have received considerable research attention and are widely available to clinical and community populations: Tai Chi, Qigong, yoga, and meditation. Tai Chi and Qigong are practices from traditional Chinese medicine that combine specific movements or postures, coordinated breathing, and mental focus. Yoga has its origins in ancient Indian philosophy; as practiced in the West, it typically includes physical postures, breathing, and meditation or relaxation, though there is considerable variability across different schools of yoga and specific interventions. Meditation refers to a broad range of practices that involve training the mind, typically to focus attention. In particular, mindfulness meditation teaches individuals to bring attention to present moment experiences with openness, curiosity, and non-judgment.

To identify studies for inclusion in this qualitative review, we searched MEDLINE (from 1946), through November 1, 2014. Searches were limited to human studies and the English language. We searched using the following terms: mind-body therapies, tai chi, qigong, meditation, mindfulness, or yoga; and inflammation, cytokines, or proinflammatory. In addition, we screened the reference lists of selected reviews and primary articles for additional publications.

Effects of MBTs on circulating markers of inflammation

Effects of MBTs on cellular markers of inflammation

Systemic changes in circulating markers of inflammation might be due to effects of MBTs on the release of cytokines such as IL-6 from non-immune sources such as adipose tissue. To evaluate whether mind-body therapies have specific effects on immune cells, which might account for increases in circulating markers of inflammation, various cellular assays have been used to determine changes in stimulated production of IL-6, TNF, and IL-1 before and after administration of Tai Chi (n=3), yoga (n=1), and meditation (n=2). The total number of studies is small (n=6), and the findings are again mixed, which might be due to differences in assay methods as well as length of treatment administration and/or follow-up duration. For example, 4 studies evaluated the production of proinflammatory cytokines using ex vivo methods in whole blood or mixed peripheral blood mononuclear cells (Elsenbruch et al., 2005; Kiecolt-Glaser et al., 2014; McCain et al., 2008; Zautra et al., 2008), whereas 2 studies examined the intracellular production of IL-6 and TNF at the level of the monocyte, the primary source of proinflammatory cytokines (Irwin et al., 2014; Irwin et al., 2015). In addition, some studies have focused on changes immediately after 8–10 weeks of treatment (Elsenbruch et al., 2005; McCain et al., 2008), whereas other studies have examined changes at follow-ups lasting 3 months to one year post-treatment (Irwin et al., 2014; Irwin et al., 2015; Kiecolt-Glaser et al., 2014). Given the varying approaches and findings from these different studies, results from the individual trials are reviewed below, stratified by type of assay method. Studies are also summarized in Table 2.

Elsenbruch et al. (2005) evaluated the effects of mindfulness-based stress reduction (MBSR) on lipopolysaccharide (LPS) stimulated production of proinflammatory cytokines using whole blood. Despite significant improvements in patient-reported outcomes of health functioning and colitis symptoms, no change in the stimulated production of TNF was found. Moreover, the intervention did not have effects on other biologic markers such as cortisol that might regulate cellular production of inflammatory cytokines. Similar negative findings were found in two other studies that showed improvements in behavioral and/or patient-reported outcomes without showing changes in cellular production of proinflammatory cytokines (McCain et al., 2008; Zautra et al., 2008). However in the Zautra et al. study, a significant reduction in the production of IL-6 was found with the CBT group, but not in the mindfulness meditation group, as compared to the education control.

In contrast to the above findings that assessed proinflammatory cytokine production after 8 to 12 weeks of treatment, differences may emerge when assessment occurs later during the follow-up period. For example, Kiecolt-Glaser et al. (2014) found no differences LPS stimulated production of IL-6, TNF, and IL-1 in isolated peripheral blood mononuclear cells in 200 breast cancer survivors after a 12 week yoga intervention, but significant group differences were identified at 3 month follow-up. Furthermore, such differences were associated with the frequency of yoga practice, which predicted changes in these inflammatory markers. Additionally, sleep quality improved along with fatigue. Given that sleep disturbance can activate inflammatory signaling and increase markers of inflammation, it was suggested that improvements in sleep might have accounted for the decreases in cellular inflammation in this sample of breast cancer survivors.

To evaluate a cellular source of inflammation, Irwin et al. have examined the effects of Tai Chi on monocyte production of IL-6 and TNF in two independent samples of insomnia patients (Irwin et al., 2014; Irwin et al., 2015). Assessment of cellular inflammation focused on LPS or Toll-like receptor (TLR)-4 stimulated production of IL-6 and TNF in monocytic populations, as compared to prior studies that had used mixed mononuclear cell cultures or whole blood. Tai Chi administration over 12 or 16 weeks reversed the insomnia-related increases in the percentage of monocytes expressing IL-6 alone, expressing TNF alone, and co-expressing IL-6 and TNF, with significant decreases for each of these measures (Irwin et al. 2014; Irwin et al., 2015). Interestingly, Tai Chi administration induced decreases as early as 2 months, with effects maintained over the course of a one year follow-up.

Effects of MBTs on gene expression inflammatory pathways

As demonstrated in several studies in the growing field of social genomics, significant life adversity is associated with alterations in gene transcriptional programs expressed under basal conditions in circulating immune cells (Cole, 2014). Indeed, genome wide transcriptional profiling of leukocytes from individuals experiencing a range of life adversities (e.g., bereavement, loneliness, low socioeconomic status) has shown a common pattern of increased expression of proinflammatory genes, which is often accompanied by a focal suppression of genes involved in innate antiviral responses (Irwin & Cole, 2011). Bioinformatic analyses of these gene expression data have identified specific transcription factors (TFs) as potential mediators of this transcriptional shift, with evidence of activation of proinflammatory NF-κB/Relfamily TFs and GATA-family TFs, decreased activity of interferon response factors (IRFs), and decreased activity of the glucocorticoid receptor (Cole, 2014). The glucocorticoid receptor would otherwise antagonize NF-κB/Rel factors and reduce inflammation.

Building on this work, genomics-based methods have been recently used to assess inflammation and the ability of MBTs to reverse inflammatory gene expression. Seven RCTs have examined the effects of MBTs on inflammation-related gene expression, including trials of Tai Chi (n=3), meditation (n=3), and yoga (n=1), described in Table 3. These studies have included a variety of control and comparison groups, including CBT, health education, and wait list. Results are consistent across trials and approaches; all have been shown to reverse the pattern of leukocyte transcriptional alterations including activation of genes regulated by the proinflammatory NF-κB/Rel family.

The effects of three different types of meditation on gene expression profiles have now been examined in various samples – lonely older adults (Creswell et al., 2012), adult caregivers (Black et al., 2014), and women with early stage breast cancer (Bower et al., 2015) – yielding consistent findings across each type of meditation. Creswell et al. (2012) identified 143 genes showing greater than 25% differential change over time with mindfulness based stress reduction (MBSR), in which the MBSR group showed a down-regulation of 69 genes and an up-regulation of 74 genes relative to the wait list controls. Using a TELiS bioinformatics analytic strategy, there was evidence of reduced activity of NF-κB target genes in the MBSR subjects that were prominently found in the monocytes consistent with prior studies on the cellular production of inflammatory cytokines. Black et al. (2014) generated a similar set of results using a yogic meditation intervention, in which the yogic meditation group showed a down-regulation of 49 genes and an up-regulation of 19 genes relative to a control group that listened to relaxing music. Down-regulated transcripts included a number of proinflammatory cytokines (e.g., IL8) as well as activation-related immediate-early genes (e.g., JUN, FOSB). Given the prior findings by Creswell et al. (2012), 3 other prototypical proinflammatory cytokines were examined with small reductions in IL-6 and IL-1 transcripts and a small increase in TNF transcripts. Nevertheless, promoter-based bioinformatic analysis again showed reduced expression of genes bearing NF-κB-response elements in the yogic meditation group as compared to controls. Finally, Bower et al. (2015) interrogated differences in a 19-transcript composite of proinflammatory genes to reveal a significantly greater decline from baseline to post-intervention in a mindfulness training group versus wait list control. Reduced activity of the proinflammatory transcription factor NF-κB and increased activity of anti-inflammatory glucocorticoid receptor was again implicated using TELiS based bioinformatics analyses.

Consistent with these homogenous effects using various types of meditation, yoga appears to generate similar differences in inflammatory gene expression. In a study of yoga for breast cancer survivors with persistent fatigue, Bower et al. (2014) identified 435 gene transcripts showing a >15% differential change over time; 282 transcripts showed up-regulation, and 153 transcripts showed down-regulation in the yoga group relative to health education controls. TELiS promoter-based bioinformatic analyses of these differentially regulated genes again demonstrated reduced activity of NF-kB, and also showed increased activity of the glucocorticoid receptor in the yoga group vs. controls.

Finally, Irwin et al. (2014, 2015) have shown that Tai Chi produced decreases in inflammatory gene expression, similar to the effects of meditation and yoga. In a study of breast cancer survivors with insomnia, analyses of an a priori-selected set of 19 proinflammatory gene transcripts showed a 9.0% (± 2.5%) greater decline over time in average expression in TCC participants relative to CBT-I participants (Irwin et al., 2014). Genome wide analysis of this sample showed a ≥1.2-fold down-regulation of 68 gene transcripts in the TCC group relative to CBT-I, of which inflammatory transcripts were prominently identified. In a study of older adults with insomnia, Tai Chi similarly induced a down-regulation of genes transcripts involved in immunological activation and inflammation (e.g., IL-6, IFNGR1, CD69, FOSB, FOS, IFNG, JUNB, IL-8, IL-1B, PTGS2) (Irwin et al., 2015). TELiS analyses in both studies implicated reduced activity of NF-κB with the primary cellular context being the monocyte population. Only one small study has directly measured changes in activation of NF-κB (as opposed to the TELiS-based inferential analyses) and found that the practice of Tai Chi relative to a health education control attenuated the rise in NF-κB found in the controls(Black et al., 2014).

One issue relevant for interpretation of results from these trials that have evaluated gene expression profiling is the use of a single social genomics lab. Cole et al. have performed the gene expression assays and associated bioinformatics analyses for all 7 studies. However, there are corroborating data from another group that found decreased NF-kB activity following acute induction of the relaxation response (Bhasin et al., 2013). Because these results were not obtained as part of an RCT, the findings were not included in the overall review. Additional studies are needed to substantiate the positive findings seen in these initial trials.

Summary of results

Over the past decade investigators have become increasingly interested in the impact of MBTs on markers of inflammation. We reviewed a total of 26 trials that examined effects of Tai Chi, Qigong, meditation, and yoga interventions on inflammatory outcomes. The majority of studies (n = 19) focused on circulating markers, particularly CRP. We found mixed evidence that the MBTs evaluated in these trials led to alterations in circulating concentrations of CRP, with half of the studies showing decreases (or attenuated increases) of CRP in the intervention group and the other half of the trials showing no significant change of CRP. The evidence for effects of MBTs on IL-6 and other inflammatory markers was also mixed, with the majority of studies showing no change. Of note, the trials that yielded non-significant results for inflammatory markers did show beneficial effects on symptoms and other outcomes, indicating that the therapies were effective in improving some aspects of health.

Studies that evaluated cellular markers of inflammation, assessed by production of proinflammatory cytokines after ex vivo stimulation, were somewhat more promising. Among these studies, 50% (3 of 6) of the trials showed that production of inflammatory cytokines was reduced following Tai Chi or yoga administration. Two of the trials that showed a reduction of inflammatory cytokines were conducted with breast cancer survivors, though trials conducted with other populations at risk for elevated inflammatory activity (HIV, ulcerative colitis) did not show such effects.

Results for trials examining genomic inflammatory markers consistently indicated a decrease in inflammatory gene expression profiles; each of the 7 trials that assessed genomic markers of inflammation showed effects on inflammatory signaling pathways, specifically reductions in NF-kB activity. These effects were seen in trials of yoga, Tai Chi, and meditation, and occurred in diverse populations. Further, effects on genomic markers were seen even in trials that did not evidence decreases in circulating inflammatory markers, suggesting that alterations in molecular signaling pathways may be more sensitive to these interventions, at least in the short term. More sustained practice may be required to alter circulating markers, such as CRP, that provide a more integrated, comprehensive measure of systemic inflammatory activity. Indeed, in the exercise literature, year-long physical activity interventions have been required to reduce circulating levels of CRP (e.g., Nicklas et al., 2008).

The limitations of the current literature are several fold. First, many of the RCTs included only a small number of subjects, and did not rely on estimates of statistical power to determine sample size needed for outcomes related to inflammation. Hence, inability of several studies to identify reductions in markers of inflammation may be a consequence of inadequate statistical power, as opposed the lack of efficacy of the MBTs. Second, subjects included within studies, and across RCTs, showed considerable heterogeneity in terms of age as well as clinical characteristics (i.e., healthy adults vs. patient populations) which might contribute to mixed results. Third, studies did not routinely evaluate levels of inflammation as an inclusion criteria for study entry; inclusion of subjects who show no or little evidence of inflammation at baseline might limit the ability to determine whether MBTs or other interventions reduce inflammation given limits of assay sensitivity and potential “floor effects”. Fourth, many of the interventions were administered over weeks and such a short time frame may not be adequate to detect changes in circulating or systemic markers of inflammation. This possibility is particularly salient given evidence of more consistent effects of MBTs to reduce inflammatory gene expression profiles and cellular markers of inflammation; such upstream changes in the inflammatory cascade would likely be occurring prior to changes in circulating markers. Fifth, the latter issue is relevant to the vast majority of RCTs that did not conduct longer term follow-up following administration of the MBTs. Sixth, many of the RCTs relied on wait list or usual care controls, and it is possible that the effects of MBTs on the inflammatory markers might be a consequences of non-specific factors such as social support due to group participation or expectation for benefit. Finally, many of the RCTs did not systematically evaluate fidelity of intervention administration and/or effects of participant adherence on changes in the inflammatory outcomes.

Mechanisms

What are the mechanisms through which these diverse approaches might lead to changes in inflammation, particularly changes in cellular and genomic markers? A number of conceptual models have been proposed to explain the effects of MBTs on a broad range of outcomes (Creswell & Lindsay, 2014; Gard, Noggle, Park, Vago, & Wilson, 2014; Holzel et al., 2011; Taylor, Goehler, Galper, Innes, & Bourguignon, 2010), all of which emphasize the important of self-regulation as a central goal of MBTs. We draw from these models, and from the empirical literature, to identify pathways that are most relevant to inflammation, including neuroendocrine, neural, and psychological/behavioral mechanisms. These pathways are illustrated in Figure 1. Of note, although we focus here on “top down” pathways, MBTs may also have more direct physiological effects, as discussed in the section on active ingredients.

Open in a separate window

Figure 1

Potential pathways linking mind-body therapies and inflammatory biology, focusing on neuroendocrine mechanisms. Mind-body interventions may influence neural regions that regulate downstream stress response pathways, including the autonomic nervous system (ANS) and the hypothalamic-pituitary-adrenal (HPA) axis. The ANS and the HPA axis may, in turn, influence the production of proinflammatory cytokines via effects of their ligands on adrenergic, cholinergic, and glucocorticoid receptors on immune cells. Stimulation of adrenergic receptors by epinephrine and norepinephrine can activate proinflammatory transcription factors such as NF-kB, whereas stimulation of GC receptors can inhibit proinflammatory cytokine production. Randomized controlled trials of mind-body therapies have demonstrated reduced pro-inflammatory signaling through NF-kB and increased anti-inflammatory GR signaling, suggesting that these approaches may modulate gene expression to reduce inflammation. However, effects on stimulated production of proinflammatory cytokines and circulating cytokine concentrations are mixed.

Active ingredients

Mind-body treatments include multiple components, including breathing, focused attention, meditation, and physical movements. The different MBTs place different emphasis on these components. For example, mindfulness and other meditative practices focus on conscious and intentional mental activities. Physical movements, if they are included as part of the intervention, are typically performed in the service of cultivating awareness of one’s body and its interaction with the environment. On the other hand, Tai Chi is a more physical practice; indeed, the flowing physical movements that are performed during Tai Chi can achieve 50% to 74% of the maximal heart rate, depending on the age of the individual and the intensity of practice. Yoga is also a more physical practice that emphasizes the importance of postures (asanas) for treatment effects. Although our review did not identify systematic differences between these approaches for inflammatory outcomes, it is certainly plausible that different MBTs may work through different mechanisms, which may modulate effects on inflammatory biology. Certainly effects on other outcomes (functional, behavioral, psychological) may differ by approach.

Another important issue in the literature on MBTs is the degree to which intervention effects are due to what are thought to be the “active ingredients” of the intervention, or to non-specific factors, such as expectations, attention from an experienced instructor, and/or group context. Carefully designed control groups are required to test these alternative hypotheses. In conducting this review, it is encouraging to find that many of the trials assessing inflammatory outcomes did use active control groups, including health education, support/counseling, and exercise. The optimal control group will, of course, depend on the question under investigation. For example, to determine whether the effects of Tai Chi are primarily due to physical activity, an exercise control group would be optimal. Or, to determine whether the effects of yoga are primarily due to stretching and associated changes in muscle-derived inflammatory markers, a stretching control would be preferred. Of note, research comparing yoga to stretching has shown mixed effects on psychological and physiological measures (e.g., Chandwani et al., 2014; Corey et al., 2014), though differential effects on inflammatory processes have not been examined.

One potential complication in the selection of control groups is the possibility that they may control for what is considered to be an essential part of the intervention. For example, we have elected not to use relaxation as a control condition in our yoga research, as relaxation is an integral part of yoga. Given that many of these practices have been developed over thousands of years, it is unclear whether dismantling them to identify active ingredients will be particularly useful. An alternative approach is to use control groups that control for non-specific elements of the intervention, such as meeting in a group and learning something you consider useful; in this case, health education might be optimal. Indeed, health education has been recommended for use in yoga trials (Sherman, 2012). In general, decisions about control groups will depend on the question under investigation in a particular study.

Typically, investigators have examined MBTs separately from other intervention approaches. However, it is possible that combining MBTs with other approaches may have additive or synergistic effects on inflammation. In particular, combining MBTs with pharmacologic therapies that target inflammation may enhance the anti-inflammatory effects of each. For example, in regards to anti-viral immune response (as opposed to inflammation), Irwin et al. (2007) found that Tai Chi together with varicella zoster virus vaccination augmented the immune response in older adults to a greater extent than either Tai Chi or vaccination alone. MBTs might also be combined with dietary approaches (e.g., omega-3s) and other interventions that are thought to reduce inflammation for maximal benefit. As a related issue, MBTs are rarely compared with each other, which would provide information about specificity of effects. Further, few trials examine combined MBT approaches (e.g., yoga and TCC), although in practice many patients may use these therapies simultaneously.

Conclusions

There is a growing literature examining effects of MBTs on inflammatory processes, reflecting increasing use of these approaches and recognition of the importance of inflammation for physical and mental health. Results from this review suggest that effects of these approaches may initially be reflected in alterations in gene expression profiles and indicators of proinflammatory signaling; indeed, alterations in inflammatory gene expression were identified even after relatively short (6 week) interventions (Bower et al., 2015) and generally appeared to be driven by reduced proinflammatory gene expression in the monocyte population. Studies examining cellular markers of inflammation also suggest that monocytes may be particularly sensitive to effects of MBTs, with effects on monocyte production of proinflammatory cytokines observed as early as 8 weeks after intervention onset (Irwin et al., 2015). Effects on non-specific markers, including circulating inflammatory markers, may take longer to emerge, and may require more intensive practice of the treatment (e.g., Bower et al., 2014; Kiecolt-Glaser et al., 2014; Pace et al., 2009). One of the compelling features of MBTs is the ability to integrate these practices into one’s daily life, with the potential for long-term benefit. A critical next step in research on MBTs is to determine how to maintain these practices and to examine longer-term effects on inflammation and health.

Acknowledgments

Footnotes

References