J Psyneuen 2017 08 008

Accepted Manuscript
Title: Yoga, mindfulness-based stress reduction and

stress-related physiological measures: A meta-analysis
Authors: Michaela C. Pascoe, David R. Thompson, Chantal F.

Ski
PII: S0306-4530(17)30040-9
DOI: http://dx.doi.org/10.1016/j.psyneuen.2017.08.008
Reference: PNEC 3694
To appear in:
Received date: 12-1-2017

Revised date: 7-8-2017
Accepted date: 8-8-2017
Please cite this article as: Pascoe, Michaela C., Thompson, David R.,
Ski, Chantal F., Yoga, mindfulness-based stress reduction and stress-
related physiological measures: A meta-analysis.Psychoneuroendocrinology
http://dx.doi.org/10.1016/j.psyneuen.2017.08.008
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
Yoga and Stress - Michaela Pascoe
Yoga, mindfulness-based stress reduction and stress-related physiological measures: A meta-analysis
Running Title: Yoga, MBSR and stress-related physiological measures
*Michaela C. Pascoe, PhD (Michaela.Pascoe@petermac.org)1
David R. Thompson, PhD (David.Thompson@unimelb.edu.au)2,3
Chantal F. Ski, PhD (Chantal.Ski@unimelb.edu.au)2
1
Peter MacCallum Cancer Centre, 305 Grattan St, Melbourne, VIC 3000, Australia
2
Department of Psychiatry, University of Melbourne, Melbourne, VIC 3010, Australia
3
Department of Epidemiology and Preventive Medicine, Monash University, Melbourne, VIC 3000,
Australia
Corresponding author
Michaela Pascoe Peter MacCallum Cancer Centre, 305 Grattan St, Melbourne, VIC 3000, Australia. E:
Michaela.Pascoe@petermac.org
Number of Tables: 2; Number of Figures: 6

Highlights
Meta-analysis of randomised control trails of involving yoga asana versus active control
Stress related physiological measures assessed
42 studies including all populations reviewed
Yoga reduced cortisol, systolic blood pressure, heart rate, heart rate variability
Yoga reduced fasting blood glucose, cholesterol and low density lipoprotein
Abstract
Background and Objectives
Practices that include yoga asanas and mindfulness-based stress reduction for the management of stress
are increasingly popular; however, the neurobiological effects of these practices on stress reactivity are
not well understood. Many studies investigating the effects of such practices fail to include an active
control group. Given the frequency with which people are selecting such interventions as a form of self-
management, it is important to determine their effectiveness. Thus, this review investigates the effects of
practices that include yoga asanas, with and without mindfulness-based stress reduction, compared to an
active control, on physiological markers of stress.
Materials and Methods
A systematic review and meta-analysis of randomised controlled trials published in English compared
practices that included yoga asanas, with and without mindfulness-based stress reduction, to an active
control, on stress-related physiological measures. The review focused on studies that measured
physiological parameters such as blood pressure, heart rate, cortisol and peripheral cytokine expression.
MEDLINE, AMED, CINAHL, PsycINFO, SocINDEX, PubMed, and Scopus were searched in May 2016
and updated in December 2016. Randomised controlled trials were included if they assessed at least one
of the following outcomes: heart rate, blood pressure, heart rate variability, mean arterial pressure, C-
reactive protein, interleukins or cortisol. Risk of bias assessments included sequence generation,
allocation concealment, blinding of assessors, incomplete outcome data, selective outcome reporting and
other sources of bias. Meta-analysis was undertaken using Comprehensive Meta-Analysis Software
Version 3. Sensitivity analyses were performed using ‘one-study-removed’ analysis. Subgroup analysis
was conducted for different yoga and control group types, including mindfulness-based stress reduction
versus non-mindfulness-based stress reduction based interventions, different populations, length of
intervention, and method of data analysis. A random-effects model was used in all analyses.
Results
Forty two studies were included in the meta-analysis. Interventions that included yoga asanas were
associated with reduced evening cortisol, waking cortisol, ambulatory systolic blood pressure, resting
heart rate, high frequency heart rate variability, fasting blood glucose, cholesterol and low density
lipoprotein, compared to active control. However, the reported interventions were heterogeneous.
Conclusions
Practices that include yoga asanas appear to be associated with improved regulation of the sympathetic
nervous system and hypothalamic-pituitary-adrenal system in various populations.
Key words: Yoga, Mindfulness-based stress reduction, Stress, Inflammation, Exercise

1.1 Introduction
The daily demands associated with modern life cause arousal and can lead to psychological stress and
activation of the stress response, or ‘fight-or-flight’ response (Nesse, 2016). The stress response, which
can be defined as real or perceived threats to homeostasis or safety/well-being (Herman et al., 2016) is
adaptive in situations of imminent threat. Persistent activation of the ‘fight-or-flight’ response can be
associated with the onset of psychiatric disorders such as anxiety and depression (Iwata et al., 2013;
Ventriglio et al., 2015).
The practice of yoga is aimed at achieving a union of mind, body and spirit and has become popular in
recent years as a form of stress management in Western cultures (Penman et al., 2012). Meta-analysis has
shown that yoga practice effectively decreases depressive and anxious symptomatology (Cramer et al.,
2013). While there is no definitive taxonomy of yoga as the many forms of the practice have unique
theoretical underpinnings and approaches (Ospina et al., 2007), common elements are shared by many
forms, such as controlled breathing, meditative techniques and physical postures (Farmer, 2012; Pflueger,
2011; Travis and Pearson, 2000). Approximately 10% of the US population were practising yoga in the
United States in 2012 (Clarke et al., 2015). In England, approximately 1.28% of the population were
practising it in 2006/08 (Ding and Stamatakis, 2014). Medical practitioners often prescribe yoga to their
patients (Nerurkar et al., 2011) with as many as 77% of surveyed Australian practitioners referring their
patients to a yoga therapist (Wardle et al., 2014).
Mindfulness-based stress reduction (MBSR), developed by Kabat-Zinn in the 1970s (Miller et al., 1995),
is a group program that uses a combination of mindfulness meditation, body awareness and yoga asanas
with the aim of increasing mindfulness (Praissman, 2008). Typically, MBSR is run across eight sessions
with at least one session plus a day-long retreat including yoga asanas, as well as home practice that may
or may not include further yoga asanas, thus the practice of yoga asanas is part of a wider program in
MBSR (Praissman, 2008). MBSR has been promoted to clinicians a safe and effective technique to
reduce stress and anxiety in diverse patient populations (Praissman, 2008).
In spite of the popularity of practices that include yoga asanas in the management of stress, research in
this area is in its infancy and the neurobiological effects of such practices are still not well elucidated.
Given the frequency with which people are choosing to engage in yoga asanas and MBSR for stress
management, it is important to validate and understand the neurobiological effects of these practices.
Some researchers have hypothesised that practices including yoga may decrease stress reactivity and thus
result in overall improved health and wellbeing (Riley and Park, 2015). We have shown previously in a
systematic review that yoga asanas appear to modulate the regulation of the sympathetic nervous system
(SNS) and hypothalamic-pituitary-adrenal (HPA) system in people experiencing depressive
symptomatology, indicated by decreases in blood pressure (BP), heart rate (HR), cortisol or cytokine
levels (Pascoe and Bauer, 2015). While this previous review provided preliminary evidence of the
beneficial effects of yoga asanas on stress activity, to date no meta-analysis has been conducted.
Furthermore, our previous review was limited to a particular population and highlighted that many studies
fail to include an active control (AC), a limitation in this developing field. Therefore, we aim to conduct a
meta-analysis investigating the effects of yoga asanas, including MBSR, on stress reactivity, in
randomised controlled trials (RCTs) that include an AC group, in all populations. We discuss the
relevance of these findings in the context of stress-related depression.

2.1 Methods
This study was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-
Analyses (PRISMA) guidelines (Moher et al., 2010). A prospective protocol for the systematic review
was not previously published.
2.2 Criteria
Eligible studies were RCTs published in English from any time. Eligible studies included any intervention
with yoga asanas (the physical postures practiced in yoga), including MBSR, compared to an AC group.
There was no restriction on participant eligibility and thus studies of all populations were included. All
RCTs with a yoga asanas intervention and AC control were included if they assessed at least one of the
following outcomes: measures related to autonomic parameters (heart rate [HR], blood pressure, heart
rate variability [HRV], mean arterial pressure [MAP]), inflammation (C-reactive protein [CRP]
Interleukin 6 [IL-6], Interleukin 8 [IL-8]) or HPA axis activation (including cortisol). Studies with lipid
outcomes were also eligible for inclusion as high cholesterol is associated with the accumulation of
cholesterol in macrophages and other immune cells, which promotes inflammation (Tall and Yvan-
Charvet, 2015), and individuals with chronic inflammatory diseases have shown a changed lipid profile
(Feingold and Grunfeld, 2000). Fasting blood glucose (FBG) was also included as inflammation
contributes the development of insulin resistance (Henriksen et al., 2011; Kalupahana et al., 2012;
Olefsky and Glass, 2010).
In order to meet the requirements of meta-analysis, eligible studies were required to report outcomes as
means with standard error (SE), standard deviation (SD) or confidence intervals (CI). We intended to
include only peer reviewed RCTs and thus dissertations were excluded. Conference abstracts and
technical reports were also excluded as these were not likely to include the detailed information required
for assessment of bias or meta-analysis inclusion.

2.3 Search strategy
The following electronic databases were searched: MEDLINE, AMED, CINAHL, PsycINFO,
SocINDEX, PubMed, and Scopus. Searches were undertaken in May 2016 and last updated on December
15, 2016. Exact search strategies are listed in the Appendix. Authors of eligible studies were contacted to
request unpublished data.
2.4 Study selection
Sourced studies were imported into Covidence Online Software (https://www.covidence.org). Two
independent reviewers screened studies for relevance based on titles/abstracts and later full texts (MCP,
MH) with disagreements resolved through discussion or by consulting a third reviewer (CFS).
2.5 Data extraction
Data were extracted using Covidence Online Software (https://www.covidence.org) and a predesigned
form that included study design, country undertaken, aims, ethical information, studied outcomes, sample
size, participant characteristics and intervention characteristics. Mean (M), standard deviation (SD) and
sample size (n) were extracted. The data were extracted by two independent reviewers (MCP, MH).
2.6 Risk of bias in individual studies and grades of recommendation, assessment, development
and evaluation
The methodological quality of the included studies was assessed independently by two reviewers (MCP,
MH) using the Cochrane Risk of Bias Tool (The Cochrane Collaboration, 2011) on Covidence Online
Software (https://www.covidence.org) and included assessments of sequence generation, allocation
concealment, blinding of assessors, incomplete outcome data, selective outcome reporting and other
sources of bias. To best capture the current state and quality of research in this field, studies were not
included or excluded based on quality assessment, and thus all eligible articles were included. Grades of
Recommendation, Assessment, Development and Evaluation (GRADE) were assessed using the GRADE
working group recommendations as published in the Cochrane Handbook. We considered five factors
when assessing the quality of evidence: 1) risk of bias; 2) heterogeneity; 3) population, intervention,
comparison, outcomes (PICO) and applicability; 4) precision; and 5) publication bias (The Cochrane
Collaboration, 2011).
2.7 Summary measures
For the meta-analysis we report the raw difference in means when the outcome is reported on the same
meaningful scale in all studies. (The SMD was used in place of raw difference in means when studies
used different outcome measures, unable to be converted to a common form, and thus the different scales
used are not comparable in raw form) To obtain the standardised mean difference (SMD) the raw
difference in means in each study is divided by the SD to create an index that is comparable across studies
(Borenstein, 2009). The Hedges’ G (g), form of the SMD was used. A small effect size was considered to
be Hedges’ G = 0.2, medium was = 0.5 and large was = 0.8. We report the confidence interval (CI) and p-
values. The proportion of the observed variance reflects differences in true effect-sizes rather than
sampling error as shown by the I2 statistic (Borenstein, 2009).
2.8 Data analysis
Meta-analysis was undertaken using Comprehensive Meta-Analysis Software Version 3 (CMA Version-
3). The primary analysis compared the effect of yoga on markers of stress and inflammation. A funnel
plot analysis which plots a study’s effect size against its standard error was conducted using CMA
Version 3 and was used to look for any publication bias. Sensitivity analyses were performed using ‘one-
study-removed’; results of this are only presented in text when removal of a study affected the outcome.
Planned a priori subgroup analysis were conducted for different yoga and control group types, including
MBSR-based inventions vs. non-MBSR-based interventions, different populations (healthy vs. clinical
populations, comparisons of different patient populations), length of yoga and control group intervention,
and method of data analysis (intention to treat [ITT] vs. completers only) and results of these subgroup
analyses are only presented in text when subgroup analysis significantly affected the outcome. A random-
effects model was used in all analyses, weighting the studies based on the sample size/standard error. In
the random effects analysis, each study is weighted by the inverse of its variance which includes the
original (within-studies) variance plus the between-studies variance (tau-squared). Results using ITT
analysis were used in the meta-analysis wherever possible. Results using completers only were used only
when ITT results were not reported. In cases when pre-post correlations were not reported in the
published papers, we used a 0.5 correlation for all analyses.
3.1 Results
3.2 Study selection
Initially 1329 articles were retrieved. Of these, 538 were duplicates and a further 40 were theses, leaving
789 for screening. One study collected cortisol outcomes but did not report these (Cohen et al., 2011).
Title/abstract screening excluded 706, 83 remained for full-text review and ultimately 42 RCTs were
included, consisting of 2944 participants. A PRISMA flow diagram shows the selection of papers for
inclusion and exclusion (Figure 1).
Insert Figure 1 here
3.3 Study characteristics
Study specifications are listed in Table 1. In cases where a study had a three-group design, with one
group being a no intervention control, we extracted data only from one of the two active intervention
groups. Five studies used a three-group, parallel-design, comparing a yoga asanas intervention with two
AC conditions (Carlson et al., 2015; Carlson et al., 2013; Jung et al., 2015; Long Parma et al., 2015; Ruby
et al., 2016; Saptharishi et al., 2009). We ran two separate analyses on any of the outcomes where more
than one control group was included, the first including one control group, and the second including the
other control group. In text, we chose to present results from the comparison between the intervention and
the AC group that was considered to be the most intensive and rigorous of the two control groups. The
meta-analysis result of the comparison of the intervention and the AC group that was considered to be
least intensive and rigorous is reported in the supplementary data. The results of these supplementary
meta-analyses are only mentioned in text when the overall result on a particular outcome differed between
the two analyses. Two studies reported outcomes on the same sample, and thus were combined into one
study (Carlson et al., 2015; Carlson et al., 2013). Two studies employed a crossover design and therefore
only outcomes before the crossover period were included in the meta-analysis (Blumenthal et al., 1991;
Bowman et al., 1997). Nine studies used ITT analysis (Carlson et al., 2013; Chacko et al., 2016; Cohen et
al., 2016; Creswell et al., 2016; Grossman et al., 2016; Hughes et al., 2013; Jedel et al., 2014; Kanaya et
al., 2014; Saptharishi et al., 2009), while all others reported outcomes on completers only. Four studies
reported outcomes as mean change scores (Chacko et al., 2016; Kanaya et al., 2014; Long Parma et al.,
2015; Sieverdes et al., 2014). All other studies reported outcomes as pre- and post-means and SD, SE or
CI within each group.
Intervention group sample sizes ranged from 7-118. In 15 studies participants were free from disease, i.e.
healthy individuals (Blumenthal et al., 1991; Bowman et al., 1997; Creswell et al., 2016; Cusumano,
1992; Gothe et al., 2016; Granath et al., 2006; Hagins et al., 2013; Harinath et al., 2004; Hayney et al.,
2014; Palta et al., 2012; Ray et al., 2001; Sawane and Gupta, 2015; Sieverdes et al., 2014) or were
pregnant (with depression in one study) (Babbar et al., 2016; Field et al., 2013). Participants had pre-
hypertension (Cohen et al., 2011; Cohen et al., 2016; Hagins et al., 2014; Hughes et al., 2013; Saptharishi
et al., 2009; Thiyagarajan et al., 2015) or hypertension (Ziv et al., 2013) in seven studies, breast cancer in
five studies (Bower et al., 2014; Carlson et al., 2015; Carlson et al., 2013; Long Parma et al., 2015; Rao et
al., 2008; Vadiraja et al., 2009) and metabolic syndrome in two studies (Corey et al., 2014; Kanaya et al.,
2014). The following populations were each included in only one study, people with osteoarthritis
(Ebnezar et al., 2012), type II diabetes (Jung et al., 2015), or at risk of type II diabetes (Yang et al., 2011),
ulcerative colitis (Jedel et al., 2014), impaired vision (Telles, 1998), increased pulse pressure (Patil et al.,
2015) polycystic ovary syndrome (Nidhi et al., 2012), vasomotor symptoms (Jones et al., 2016), bariatric
patients (Chacko et al., 2016), fibromyalgia patients (Grossman et al., 2016) or had restless leg syndrome
(Innes and Selfe, 2012) or were overweight or obese (Ruby et al., 2016).
The yoga asanas interventions in each study varied in their components, frequency and length as reported
in Supplementary Table 1 (template for intervention description and replication [TIDiER] table).
Ashtanga yoga was used in one study (Hagins et al., 2014). Hatha yoga was used in eight studies
(Bowman et al., 1997; Cohen et al., 2016; Cusumano, 1992; Gothe et al., 2016; Harinath et al., 2004;
Long Parma et al., 2015; Ray et al., 2001; Sieverdes et al., 2014). Integrated yoga (Cohen et al., 2011;
Patil et al., 2015; Rao et al., 2008; Vadiraja et al., 2009) and Iyengar yoga (Bower et al., 2014; Cohen et
al., 2011; Innes and Selfe, 2012; Sawane and Gupta, 2015) were each used in four studies. An MBSR
program or derivative was used in nine studies (Carlson et al., 2015; Carlson et al., 2013; Chacko et al.,
2016; Creswell et al., 2016; Grossman et al., 2016; Hayney et al., 2014; Hughes et al., 2013; Jedel et al.,
2014; Jung et al., 2015; Palta et al., 2012). Prenatal (Babbar et al., 2016; Field et al., 2013) and restorative
yoga (Corey et al., 2014; Kanaya et al., 2014) were each used in two studies. Vinyasa yoga was used in
one study (Yang et al., 2011). Ten studies did not specify the type of yoga used (Blumenthal et al., 1991;
Granath et al., 2006; Hagins et al., 2013; Jones et al., 2016; Nidhi et al., 2012; Ruby et al., 2016;
Saptharishi et al., 2009; Telles, 1998; Thiyagarajan et al., 2015; Ziv et al., 2013) but authors provided this
information upon request in two cases (Granath et al., 2006; Hagins et al., 2013)
Exercise or physical activity was used as an AC in 17 studies (Blumenthal et al., 1991; Bowman et al.,
1997; Ebnezar et al., 2012; Hagins et al., 2013; Hagins et al., 2014; Harinath et al., 2004; Hayney et al.,
2014; Jones et al., 2016; Jung et al., 2015; Long Parma et al., 2015; Nidhi et al., 2012; Patil et al., 2015;
Ray et al., 2001; Ruby et al., 2016; Saptharishi et al., 2009; Telles, 1998; Ziv et al., 2013), while health
education was used in six studies (Babbar et al., 2016; Bower et al., 2014; Cohen et al., 2016; Jung et al.,
2015; Yang et al., 2011). Social support (Field et al., 2013; Palta et al., 2012), stretching (Corey et al.,
2014; Kanaya et al., 2014) and progressive muscle relaxation (with physical therapy in Grossman et al.,
2016) (Cusumano, 1992; Grossman et al., 2016; Hughes et al., 2013) or other counselling/therapy
(Carlson et al., 2015; Carlson et al., 2013; Chacko et al., 2016; Rao et al., 2008) were each used in two
studies. Stretching was used in four studies (Corey et al., 2014; Gothe et al., 2016; Kanaya et al., 2014;
Patil et al., 2015). The following control interventions were each used in only one study: brief supportive
therapy (Vadiraja et al., 2009), cognitive behaviour therapy (CBT) (Granath et al., 2006), educational film
(Innes and Selfe, 2012), enhanced usual care (Cohen et al., 2011), lifestyle modification program
(Thiyagarajan et al., 2015), swimming (Sawane and Gupta, 2015), music and art classes (Sieverdes et al.,
2014), relaxation (Creswell et al., 2016) and diet (Ruby et al., 2016).
Insert Tables 1 and 2 here
3.4 Risk of bias within studies and grades of recommendation, assessment, development and
evaluation
As can be seen in Table 3, only one study had a high risk of bias for sequence generation, eight for
incomplete outcome data and selective outcome reporting, and 14 for other sources of bias. On each of
the individual domains the vast majority of the included RCTs were rated as having a low or unclear risk
of bias, which is insufficient to justify downgrading the level of evidence. In terms of heterogeneity, four
outcomes (cortisol slope, resting DBP, resting SBP, MAP) had an I2 value that showed that a high
proportion of the observed variance reflected differences in true effect-sizes, rather than sampling error.
The p value assessing heterogeneity in these outcomes was also under p=0.05. We suggest that this
observed heterogeneity can be explained by differences in study design such as intervention type,
duration, as well as differences in the studied populations. For resting DBP, resting SBP and cortisol
slope the confidence intervals for most studies overlapped. For MAP, the confidence intervals for most
studies did not overlap. For resting DBP, resting SBP and MAP there were also differences in the
estimate of effects. Therefore, for these above mentioned four outcomes, we suggest the level of evidence
be downgraded from high to moderate. In terms of PICO, we feel that the various populations included,
yoga based interventions, AC group comparisons, and outcomes assessed were appropriate to address the
question of if practices including yoga asanas, including MBSR, influence stress-related physiological
measures compared to AC groups in all populations. The following outcomes had small sample sizes:
CRP (n=119); IL-6 (n=154); IL-8 (n=70); mid-morning cortisol (n=166); and 60 mins post-waking
cortisol (n=163). Funnel plots and the classic fail safe N (see supplementary data) indicated possible
publication bias for the following outcomes: waking cortisol; evening cortisol; cholesterol; triglycerides;
LDL; HDL; and IL-6; thus we suggest that level of evidence for these outcomes should be downgraded to
moderate. For the following outcomes there were too few studies available to accurately assess
publication bias using funnel plots: 30 mins post-waking cortisol; 60 mins post-waking cortisol; mid-
morning and afternoon cortisol; cortisol diurnal slope; IL-8; CRP; 24hr DBP; 24hr SBP; and resting HRV
(LF and HF). Overall, we consider the level of evidence to be high for the following outcomes: resting
HR and FBG. We consider the GRADE of evidence to be moderate for the following outcomes: waking
cortisol evening cortisol; cortisol slope; resting DBP; resting SBP; MAP; cholesterol; triglycerides; LDL;
HDL; and IL-6. Due to the small number of included primary studies (n=≤4), we consider the level of
evidence to be low for the following outcomes: 30 mins post-waking cortisol; 60 mins post-waking
cortisol; mid-morning cortisol; afternoon cortisol; IL-8; CRP; 24hr DBP; 24hr SBP; and resting HRV (LF
and HF).
Of note, inappropriate statistical methods appear to have been used in a few primary studies: Mann-
Whitney when had baseline data (Yang et al., 2011) and T Tests (Bowman et al., 1997; Cohen et al.,
2011; Ray et al., 2001; Sawane and Gupta, 2015): however, the methods of statistical analysis used in
these primary analyses do not affect the outcome of the current meta-analysis, as only the raw means,
SD/SE/CI and sample sizes were required for conduct of the meta-analysis, and these are obtained from
the descriptive statistics.
Insert Table 3 here
3.5 Meta-Analysis
3.5.1 Cortisol
Cortisol was measured in 12 studies as can be seen in Table 1. Cortisol (except mid-morning cortisol) was
assessed using the raw difference in means.
Waking cortisol (n=386) was measured in five studies at post intervention (Bower et al., 2014; Carlson et
al., 2013; Corey et al., 2014; Sieverdes et al., 2014; Vadiraja et al., 2009). In one study, ug/dL was
converted to nmol/L (*27.59) so that analysis could be conducted on the raw difference in means
(Vadiraja et al., 2009). In one study, log-transformed values were reported in the text but raw scores were
provided by the author and included in the analysis (Carlson et al., 2013). Interventions including yoga
decreased waking cortisol by 1.51nmol/L compared to AC (p=0.03, I2=0%). Removal of the studies by
Corey and Raghavenda resulted in p-values of 0.11 and 0.07, indicating that these results are not
particularly robust.
Salivary cortisol at 30 mins post-waking was measured in three studies (Bower et al., 2014; Corey et al.,
2014; Sieverdes et al., 2014) (n=191) and no effect was found (p=0.30, I2=0%). Salivary cortisol at 60min
post-waking was measured in two studies (n=163) (Corey et al., 2014; Sieverdes et al., 2014). No effect
of the yoga intervention was found (p=0.20, I2=0%).
Salivary cortisol at mid-morning was measured in three studies using SMD (Field et al., 2013; Granath et
al., 2006; Vadiraja et al., 2009), as shown in Figure 2 (n=166). No effect of the yoga intervention was
found (p=0.08, I2=0%). Cortisol slope was measured in three studies (Bower et al., 2014; Carlson et al.,
2013; Corey et al., 2014) (n=301) and no effect was found (p=0.29, I2= 69.2%).
Afternoon salivary cortisol reported by three studies (n=285) (Bower et al., 2014; Carlson et al., 2013;
Gothe et al., 2016). Afternoon cortisol was 0.60nmol/L lower in the group interventions involving yoga,
compared to the AC (p=0.03, I2=0%). Removal of the study by Carlson resulted in p=0.42.
Evening cortisol (collected between 21:00 and 24:00) (n=385) was measured in five studies post-
intervention (Bower et al., 2014; Carlson et al., 2013; Corey et al., 2014; Sieverdes et al., 2014; Vadiraja
et al., 2009). Interventions including yoga asanas decreased salivary evening cortisol by -0.88nmol/L
compared to AC (p=.048, I2=14.95%). Removal of the studies by Bower, Corey, Raghavenda and
Sieverdes resulted in p-values of 0.08, 0.31, 0.09 and 0.09 respectively.
3.5.1.1 Comparison of MBSR and non-MBSR interventions on cortisol outcomes
Sensitivity analysis showed that removal of the studies using an MBSR-based intervention (MBCR)
(Carlson et al., 2013) did not alter the above cortisol outcome. This finding indicates that the observed
differences in salivary cortisol are common to both MBSR and non-MBSR interventions including yoga
asanas. There results of these subgroup analyses are presented in the legend of Figure 2.
3.5.2 Inflammatory changes
3.5.2.1 Interleukin-8 (IL-8)
The assessment of CRP and interleukins were conducted using SMD, as Long Parma et al., (2015)
measured these on a scale that was unable to be converted. Inflammatory changes are shown in Figure 3.
Only two studies measured IL-8 (n=70), one directly after (Long Parma et al., 2015) and one at 10 months
post-intervention (Jedel et al., 2014) and both reported on completers only. No significant effect of yoga
asanas was found using SMD (p=0.16, I2=0%).
3.5.2.2 Interleukin 6 (IL-6)
IL-6 was measured in five studies (n=154) with two studies collecting the outcome directly after (Bower
et al., 2014; Long Parma et al., 2015), one at 2 months (Chacko et al., 2016) one at 4 months (Creswell et
al., 2016) one at 10 months post intervention (Jedel et al., 2014). As with CRP, SMD was used. Two
studies used ITT analysis (Chacko et al., 2016; Creswell et al., 2016). No effect of yoga asanas compared
to AC was found (p=0.07, I2=0%). Removal of the study by Jedel et al., (2014) resulted in a significant p
value of 0.048. As the study by Long Parma et al., (Long Parma et al., 2015) had two control groups, we
ran a second meta-analysis including the second control group (comparison exercise). In this analysis,
interventions that included yoga were found to decrease IL-6 compared to an AC group (SMD (CI)=-
0.33(-0.65, -0.01), p=0.04, I2=0%) (see supplementary data).
3.5.2.2.1 Comparison of MBSR and non-MBSR interventions on IL-6 outcomes
Subgroup analysis showed that in the three studies that used an MBSR intervention (Creswell et al., 2016,
Jedel et al., 2014 and Chacko et al., 2016), a small effect of yoga asanas compared to AC was found
(p=0.03, I2=0%). Conversely, in the two studies using a non-MBSR-based yoga intervention (Bower et
al., 2014, Long Parma et al., 2015) no effect of yoga practice was found on IL-6 (p=0.65, I2=0%). This
finding indicates that the observed differences in IL-6 are associated with MBSR but not non-MBSR yoga
interventions. These conflicting results are depicted on Figure 3.
3.5.2.3 C-reactive protein (CRP)
The analysis of serum CRP was included in four studies (n=119), two collecting CRP directly after
intervention completion (Bower et al., 2014; Long Parma et al., 2015), one at 2 months (Chacko et al.,
2016) and one at 10 months follow up (Jedel et al., 2014). One study used ITT analysis (Chacko et al.,
2016). No effect of intervention including yoga asanas compared to AC was found (p=0.32, I2=0%).
Subgroup analysis including the two studies that collected CRP directly after intervention completion
similarly showed no effect of interventions including yoga asanas on CRP levels.
3.5.2.3.1 Comparison of MBSR and non-MBSR interventions on CRP outcomes
Subgroup analysis of studies that used an MBSR-based intervention and studies that did not, similarly
showed no effect of interventions on CRP levels, MBSR (p=0.44, I20%; no MBSR p=0.52, I2=0%), as
shown in Figure 3. These finding indicate that neither MBSR nor non-MBSR yoga interventions
influence CRP outcomes.
3.5.3 Autonomic measures
3.5.3.1 Diastolic blood pressure (DBP)
Blood pressure was assessed using raw difference in means. Resting BP outcomes for are shown in Figure
4. DBP was measured in 20 studies as seen in Table 1. Resting DBP was measured in 16 studies at post-
intervention (n=887) (Babbar et al., 2016; Cohen et al., 2011; Ebnezar et al., 2012; Granath et al., 2006;
Hagins et al., 2013; Harinath et al., 2004; Hughes et al., 2013; Innes and Selfe, 2012; Palta et al., 2012;
Patil et al., 2015; Ruby et al., 2016; Saptharishi et al., 2009; Sawane and Gupta, 2015; Sieverdes et al.,
2014; Thiyagarajan et al., 2015; Yang et al., 2011). Two studies used ITT analysis (Hughes et al., 2013;
Saptharishi et al., 2009). Interventions that included yoga asanas were found to decrease resting DBP by
3.66 mmHg compared to AC (p<.001, I2=81.51%). We ran sub-group analysis comparing different yoga
types/duration and control group type, however did not find any meaningful differences.
3.5.3.1.1 Comparison of MBSR and non-MBSR interventions on DBP outcomes

Subgroup analysis showed that both the studies using a MBSR intervention (Hughes et al., 2013, Palta et
al., 2012) and the studies using non-MBSR intervention decreased resting DBP: MBSR (p=0.03, I2=0%.;
non-MBSR, p<0.01, I2=83.89%).This finding indicates that the observed differences in DBP are due to
both MBSR and non-MBSR yoga interventions.
3.5.3.1.2 Ambulatory DBP
Ambulatory DBP (24hr) was measured in 3 studies at post intervention (n=242) (Cohen et al., 2016;
Hagins et al., 2014; Ziv et al., 2013), with only one study using ITT analysis (Cohen et al., 2016). No
effect of the intervention was found (p=0.36, I2=0%), as shown in Figure 5.
3.5.3.2 Systolic blood pressure (SBP)
Blood pressure was assessed using raw difference in means. SBP was measured in 22 studies as seen in
Table 1. Resting SBP was measured in 17 studies at post intervention (n=1058) (Babbar et al., 2016;
Cohen et al., 2011; Ebnezar et al., 2012; Granath et al., 2006; Hagins et al., 2013; Harinath et al., 2004;
Hughes et al., 2013; Innes and Selfe, 2012; Kanaya et al., 2014; Palta et al., 2012; Patil et al., 2015; Ruby
et al., 2016; Saptharishi et al., 2009; Sawane and Gupta, 2015; Sieverdes et al., 2014; Thiyagarajan et al.,
2015; Yang et al., 2011). Three studies used ITT analysis (Hughes et al., 2013; Kanaya et al., 2014;
Saptharishi et al., 2009). Interventions including yoga asanas were found to decrease resting SBP by 5.34
millimetres of mercury (mmHg) compared to AC (p<.001, I2=83.58%).
3.5.3.2.1 Comparison of MBSR and non-MBSR interventions on SBP outcomes
Subgroup analysis showed that in the two studies using a MBSR intervention (Hughes et al., 2013, Palta
et al., 2012) that no effect of yoga asanas were found on resting SPB, compared to AC (p=0.40, I2=0%).
In the studies using non-MBSR intervention, yoga asana interventions were found to decrease resting
SBP, compared by 3.23 mmHg compared to AC to AC (p<.001, I2=84.96%), as shown in Figure 4. This
finding indicates that the observed differences in SBP are associated with non-MBSR yoga interventions,
but not with MBSR interventions.
3.5.3.2.2 Ambulatory SBP
Ambulatory SBP (24hr) was measured in 3 studies at post intervention (n=272) (Cohen et al., 2016;
Hagins et al., 2014; Ziv et al., 2013), with only one study using ITT analysis (Cohen et al., 2016). No
effect of the interventions including yoga asanas was found (p=0.48, I2=0%).
3.5.3.3 Mean arterial pressure (MAP)
The analysis of resting MAP was done using the raw difference in means and included five studies with
outcomes collected at post-intervention (n=315). One study collected 24hr ambulatory MAP (Hagins,
2014) and four collected resting MAP (Cohen et al., 2011; Harinath et al., 2004; Patil et al., 2015;
Thiyagarajan et al., 2015). All studies reported outcomes on completers only data. Intervention including
yoga asanas decreased MAP by 6.82mmHg compared to AC (p<.001, I2=90.41%). (Figure 5)
3.5.3.4 Heart rate (HR)
HR was measured in 23 studies as seen in Table 1. Resting heart rate at post intervention was measured in
15 studies in completers only using the raw difference in means (n=879) (Babbar et al., 2016; Blumenthal
et al., 1991; Bowman et al., 1997; Cohen et al., 2011; Cusumano, 1992; Ebnezar et al., 2012; Granath et
al., 2006; Hagins et al., 2013; Harinath et al., 2004; Innes and Selfe, 2012; Ruby et al., 2016; Sawane and
Gupta, 2015; Sieverdes et al., 2014; Telles, 1998; Thiyagarajan et al., 2015). Participants in the
interventions including yoga asanas had a resting HR 3.20 beats per minute slower than those in the AC
group (p<.001, I2= 20.55%). (Figure 5)

3.5.3.5 Heart rate variability (HRV)
The analysis of resting HRV included four studies (supplementary Figure 1) (Bowman et al., 1997; Jones
et al., 2016; Patil et al., 2015; Sawane and Gupta, 2015) and was conducted using SMD as Bowman used
different scale which was unable to be converted with outcomes collected at post-intervention. All studies
reported outcomes on completers only data (n=367). The results for low frequency (LF) (0.04-0.15 Hz)
indicate a medium effect (g) (p<.001, I2=34.31%). High frequency (HF) (0.15-0.40 Hz) analysis similarly
showed a medium effect (p-value=0.01, I2=0%) and removal of any one study did not affect the outcome.
Grossman et al., (2016) also reported log transformed HRV (HF) however did not provide raw score
means and SD upon request and thus was unable to be included in this analysis (Grossman et al., 2016).
3.5.4 Fasting blood glucose (FBG)
FBG was assessed using the raw difference in means in seven studies (supplementary Figure 2) (Jung et
al., 2015; Kanaya et al., 2014; Nidhi et al., 2012; Ruby et al., 2016; Thiyagarajan et al., 2015; Yang et al.,
2011; Ziv et al., 2013) (n=534) with outcomes collected at post-intervention. Only one study used ITT
(Kanaya et al., 2014). Interventions including yoga asanas were seen to decrease FBG by 4.53mg/dL
compared to AC (p<.001, I2=0%).
3.5.5 Lipids
All lipids were assessed using SMD (supplementary Figure 3) as Blumenthal 1991 reported mg%, which
was unable to be converted to a common unit of measurement. Cholesterol at post intervention was
measured in six studies reporting data on completers only (n=389) (Blumenthal et al., 1991; Nidhi et al.,
2012; Ruby et al., 2016; Thiyagarajan et al., 2015; Yang et al., 2011; Ziv et al., 2013) and showed a small
effect of the intervention including yoga asanas (p-value=0.01, I2=0%).
Low-density lipoprotein (LDL) at post intervention was measured in six studies in completers only
(n=389) (Blumenthal et al., 1991; Jedel et al., 2014; Nidhi et al., 2012; Thiyagarajan et al., 2015; Yang et
al., 2011; Ziv et al., 2013). A small effect of the interventions including yoga asanas were found (p-<0.01,
I2=0%).
Triglycerides at post intervention were measured in seven studies (Blumenthal et al., 1991; Kanaya et al.,
2014; Nidhi et al., 2012; Ruby et al., 2016; Thiyagarajan et al., 2015; Yang et al., 2011; Ziv et al., 2013)
(n=560) with only one study using ITT (Kanaya et al., 2014). No effect of the interventions including
yoga asanas were found (p-value=0.16, I2=0%).
High-density lipoprotein (HDL) was assessed in seven studies at post intervention (n=560) (Blumenthal et
al., 1991; Kanaya et al., 2014; Nidhi et al., 2012; Ruby et al., 2016; Thiyagarajan et al., 2015; Yang et al.,
2011; Ziv et al., 2013) with only one study using ITT (Kanaya et al., 2014). No effect of interventions
including yoga asanas were found (p-value=0.12, I2=0).
4.1 Discussion
In this review we included 42 studies which examined the relationship between interventions including
yoga asanas and physiological measures of stress. Outcomes included cortisol, autonomic measures,
cytokines and lipid levels.

4.2 Cortisol
We found that practices involving yoga asanas reduced waking, afternoon and evening salivary cortisol.
Sensitivity analysis showed that for each of these outcomes, removal of some studies resulted in a non-
significant effect, indicating that these results are not particularly robust. Indeed, we considered that the
GRADE of evidence for waking and evening cortisol should be downgraded to moderate due to possible
publication bias and that the GRADE of evidence for afternoon cortisol should be considered low, as only
three primary studies were included in the analysis.
Stress is accepted to contribute to the onset of depressive symptomatology (Herman et al., 2016; Iwata et
al., 2013).and individuals with clinical depression and negative mood have elevated basal levels of
cortisol (Brown et al., 2004; Pruessner et al., 2003). Some authors argue that clinical depression
represents a dysregulated adaptive stress response (Gold, 2015), which is supported by evidence that that
elevated cortisol levels predict the onset of depressive disorder (Goodyer et al., 2000; LeMoult et al.,
2015). Therefore, practices involving yoga that decrease cortisol levels may protect against the
development of stress related mental illness such as depression.
It is unclear why we found decreases in waking cortisol but no effect of the interventions including yoga
asanas on cortisol at 30 mins, 60 mins, mid-morning or cortisol slope. This may be due to the fact that
only two and three studies were included in these analyses, and therefore the results should be considered
as preliminary. Indeed, we suggest that these outcomes be considered low level evidence. It is unknown if
values of salivary cortisol at post waking are meaningful independent of the change from waking cortisol
levels and we suggest that in future studies cortisol should be collected at multiple time points throughout
the day, in order to measure change from waking cortisol.
In summary, practices involving yoga asanas appear to reduce waking, afternoon and evening salivary
cortisol, which may protect against the development of stress-related mental illness; however, these
results should be considered preliminary.

4.3 Inflammatory measures
No effect of the intervention was found on CRP or IL-8. Interventions including yoga asanas were not
seen to decrease IL-6 compared to an AC in the primary analysis. Sub-group analysis for IL-6 showed
that in the three studies that used an MBSR intervention, there was a small effect of yoga asanas
compared to AC. Conversely, in the two studies using a non-MBSR yoga intervention there was no effect
of yoga practice on IL-6. MBSR includes structured practise of mindfulness meditation, body scanning
and cognitive awareness. Given that yoga asana practice in isolation did not appear to mediate IL-6 levels,
it is possible that the more meditative aspects of MBSR, such as the practise of mindfulness meditation
and body scanning underlie the observed change in IL-6, rather than the yoga asanas practice. However, it
is interesting to note that we conducted a secondary meta-analysis as the study by Long Parma et al.,
(Long Parma et al., 2015) included a second control group. In the secondary analysis, Hatha yoga was
compared to comparison exercise, (as opposed to comprehensive exercise, which consisted of an
individualized exercise program prescribed by a certified Clinical Exercise Specialist) where participants
engaged in exercise of their own choosing. In this secondary analysis, interventions including yoga
practice were seen to decrease IL-6. These results indicate that while MBSR and practices involving yoga
asanas appear to influence IL-6 levels to some degree that selecting an appropriate comparison group is
particularly important when assessing the impact of yoga asanas on IL-6 levels. Given the observed
heterogeneity in IL-6 outcomes, we suggest that the level of evidence for interventions including yoga
asanas on IL-6 levels should be considered moderate and interpreted with caution. The current results
indicate that IL-6 is an interesting marker of physiological stress in the context of practices involving
yoga and should be further studied to delineate what aspects of the practices are capable of mediating it,
compared to appropriately selected control groups. Decreasing IL-6 levels in people experiencing high
levels of stress may protect against the development of stress-related psychological illnesses such as
depression. Indeed, increased levels of IL-6 are seen in major depressive disorder (Alesci et al., 2005;
Maes, 2008), which is widely accepted to be precipitated by stress (Pariante, 2003) and meta-analysis
indicates that pharmacological antidepressants may decreases IL-6 levels (Hannestad et al., 2011).
In summary, practices involving yoga asanas do not seem to influence CRP or IL-8 levels and moderate
level evidence shows that only MBSR interventions decreased IL-6 levels. However non-MBSR
interventions were seen to decrease IL-6 when compared to a less time-intensive control group, indicating
that while practices of yoga asanas may influence IL-6 levels, further research needs to be done to better
understand potential effects.
4.4 Autonomic measures
Interventions that include yoga asanas reduced resting DBP, resting SBP, MAP, resting HR and increased
LF and HF-HRV. No effect of the interventions was found on ambulatory DBP and ambulatory SBP. For
a number of these outcomes (resting SBP, resting DPB, MAP, LF-HRV) there was heterogeneity in the
effect size between studies, which is not surprising given that the populations, interventions and duration
of interventions differed in each study. In the case of MAP, all study interventions were 12 weeks in
duration, and thus heterogeneity between studies likely results from different yoga techniques, control
groups or populations studied.
For resting SBP, subgroup analysis showed that there was no effect of MBSR interventions, compared to
AC. MBSR includes yoga asanas with low impact cardiovascular effects, which may not result in exercise
associated effects sufficient to result in changes in resting SBP (Cornelissen and Smart, 2013). However,
in seven of the 14 studies measuring SPB in non-MBSR yoga interventions compared to AC, the AC
group consisted of an exercise intervention. This indicates that yoga asanas practice decreased resting
SBP to a greater degree than non-yoga based exercise. Therefore, it would seem that the effects of yoga
on resting SBP may be related to a combination of the mindfulness practice and the cardiovascular effects
associated with the physical practice of yoga asanas. It is important to note that six studies did not specify
if multiple BP measures were taken and thus the reliability of these outcomes is unclear.
It is interesting that interventions including yoga asanas were associated with increased LF and HF-HRV.
Traditionally, increased HF-HRV has been accepted to reflect increased PNS. LF-HRV is more complex
and represents both branches of the PNS, but has been argued to have a dominant SNS
component (Billman, 2011; Task Force of the European Society of Cardiology, 1996; Thayer et al.,
2010). However, more recently various authors have highlighted that accumulating evidence
demonstrates that this interpretation oversimplifies the complex non-linear interactions between the SNS
and the PNS (Billman, 2011, 2013). Substantial evidence indicates that LF-HRV likely represents arterial
baroreflex control, a neural feedback mechanism controlling HR, rather than cardiac sympathetic tone
(Goldstein et al., 2011). Acute rises in BP cause baroreceptor activation, resulting in PNS activation,
decreased HR, vascular resistance, cardiac contractility and venous return to help maintain BP
homeostasis (Lanfranchi and Somers, 2002; Wesseling and Settels, 1985). Accordingly, some authors
argue that LF-HRV is determined primarily by the PSN rather than the SNS. Given the lack of clarity
regarding the precise role of LF-HRV it is difficult to interpret the significance of the present finding that
interventions including yoga asanas increased both HF-HRV and LF-HRV. Given the evidence
demonstrating the relationship between baroreflex sensitivity and LF-HRV however, it is possible that
interventions including yoga asanas increases baroreflex sensitivity. This is relevant as a high sympathetic
and/or a low cardiovagal activity is seen in patients with major depression (Agelink et al., 2002). It is
possible that interventions that decrease sympathetic and increase cardiovagal activity could protect
against the development of stress related depression.
Given that all the AC groups for HRV outcomes included an exercise-based intervention, the effects of
interventions including yoga asanas on these outcomes cannot likely be attributed to the exercise-related
effect of yoga asanas alone. Indeed, yoga practice incorporates philosophical teachings, mindful
awareness, controlled breathing, meditative techniques and physical asanas (Farmer, 2012; Pflueger,
2011; Travis and Pearson, 2000), which likely influence psychological processes. Many yoga teachings
emphasise learning to accept one’s reality, for example, and this way of thinking may aid in reducing the
stress associated with demanding goals or the rumination resulting from dissatisfaction about the
mismatch between reality and desires. Indeed, previous work has demonstrated significant psychological
effects of yoga practice, including decreased depressive and anxious symptoms, perceived stress (Cramer
et al., 2013; Li and Goldsmith, 2012) and improved emotional regulation (Gard et al., 2014). Thus, the
combination of philosophical teachings, mindful awareness, controlled breathing, meditative techniques
and asanas practiced in yoga appear to improve psychological outcomes, and these are likely related to
the effects of yoga on physiological measures of stress.
In summary, high level evidence indicates that interventions that include yoga asanas reduce resting HR.
Moderate level evidence indicates that interventions that include yoga asanas reduce MAP, resting DBP
and resting SBP and that the effects of yoga on resting SBP may be related to a combination of the
mindfulness practice and the cardiovascular effects associated with the physical practice of yoga asanas.
Low level evidence indicates that interventions that include yoga asanas increase LF and HF-HRV
possibly increasing baroreflex sensitivity, and that these interventions have no effect on ambulatory DBP
and ambulatory SBP.
4.5 Lipids
Interventions including yoga asanas reduced FBG, cholesterol and LDL. The effect size was not seen to
vary between studies, indicating that the effect of yoga asanas on these outcomes was robust to
intervention duration, yoga type and type of AC. No effect of the intervention was found on triglycerides
and HDL, suggesting that the cholesterol altering effects of yoga practice were due primarily to lowering
LDL. We consider the level of evidence to be high for FBG and moderate for cholesterol, triglycerides,
LDL, HDL.
The above findings are interesting in regards to inflammation-associated depressive disorder. As earlier
highlighted, high cholesterol promotes inflammation (Tall and Yvan-Charvet, 2015) and inflammation
contributes the development of insulin resistance (Henriksen et al., 2011; Kalupahana et al., 2012;
Olefsky and Glass, 2010). Chronic inflammation is widely accepted to contribute to the onset of
depressive disorder (Pariante, 2003) and previous research has demonstrated that there is a high
prevalence of lipid and glucose abnormalities in patients with depressive disorders (Wysokinski et al.,
2015). In clinical populations, melancholic features are independently associated with lower HDL
cholesterol, while atypical depression is independently associated with higher total and LDL cholesterol
(van Reedt Dortland et al., 2010). In pregnant women, HDL levels are inversely associated with changes
in depressive symptom scores after adjusting for socioeconomic, demographic, behavioural, nutritional,
biochemical and mental health factors (Teofilo et al., 2014). The altered lipid levels associated with yoga
practice suggest the practices involving yoga asanas may protect against illnesses such as clinical
depression.
In summary high level evidence indicates interventions including yoga asanas reduced FBG. Moderate
level evidence indicates that interventions including yoga asanas reduced cholesterol and LDL and did not
influence triglycerides and HDL.
4.6 The difference in physiological outcomes between MBSR and non-MBSR based interventions
varies according to the parameter
There were differences in outcomes between MBSR and non-MBSR yoga asana interventions for IL-6
and resting SBP, as discussed. For IL-6 and resting SBP it would seem that the mindfulness and stress
management skills taught in the MBSR program are required in order to have a physiological effect.
There was no difference in outcomes between MBSR and non-MBSR yoga asana interventions for
salivary cortisol, CRP and resting DBP. For salivary cortisol, CRP and resting DBP, the practise of yoga
asanas appears to have a physiological effect independent of the mindfulness and stress management
skills taught in the MBSR program. In summary, the above findings indicate that differences in
physiological outcomes between MBSR and non-MBSR interventions vary according to the parameter
measured.
4.7 Previous findings
The results of the present meta-analysis are consistent with our previous systematic review reporting that
yoga asanas practice appears to modulate the regulation of the SNS and HPA system in people
experiencing depressive symptomatology. Our previous review found that yoga practice was associated
with decreases in BP, HR, cortisol and cytokine levels, compared to usual care (Pascoe and Bauer, 2015).
The findings of the current meta-analysis are further consistent with another recent systematic review of
yoga asanas studies that measured stress as a primary dependent variable and assessed a mechanism of the
relationship (Riley and Park, 2015). This review identified four biological mechanisms associated with
the effects of yoga asanas on stress. The first was the posterior hypothalamus as indicated by decreased
BP after a single yoga asanas session (Bagga and Gandhi, 1983). Three inflammatory and endocrine
responses were also identified, IL-6, CRP and cortisol (Kiecolt-Glaser et al., 2010; Michalsen et al.,
2005). Yoga asanas were found to mitigate changes in IL-6 and CRP following a stressful laboratory test,
compared to a movement control condition and a video control condition (Kiecolt-Glaser et al., 2010).
Another study showed that salivary cortisol decreased after participation in 12 weeks of Iyengar yoga and
that these changes predicted decreases in perceived stress (Michalsen et al., 2005).
A more recent systematic review of mind–body therapies, including yoga asanas, tai chi, qigong
and meditation reported reduced signalling through the pro-inflammatory transcription factor, nuclear
factor kappa-light-chain-enhancer of activated B cells (NF-κB), and mixed effects of the mind-body
therapies on CRP and IL-6 (Bower and Irwin, 2016). The finding of mixed effects in terms of CRP and
IL-6 is not surprising given that they assessed yoga asanas, tai chi, qigong, and meditation altogether.
Two of the yoga asanas-based studies reviewed were also included in the current meta-analysis (Bower et
al., 2014; Rao et al., 2008). In those yoga asanas-based studies not included in the current meta-analysis,
MBSR was found to have no effect on IL-6 and a trend for decreasing CRP in healthy older adults,
compared to usual care (Creswell et al., 2012). In university staff, a low dose of MBSR compared to a
lifestyle education program was similarly seen to have no effect on IL-6 or salivary cortisol and a trend
for decreasing CRP was seen (Malarkey et al., 2013). These findings are consistent with those of the
present study, showing no effect of interventions including yoga asanas on CRP compared to an AC. The
results on IL-6 however are inconsistent with those of the present meta-analysis, where a small effect was
found in studies using MBSR. In two studies by Pullen et al., yoga asanas was found to decrease IL-6 and
hs-CRP in patients with chronic heart failure, (Pullen et al., 2008; Pullen et al., 2010) and in breast cancer
survivors yoga asanas was seen to decrease IL-6, TNF-α and IL-1β (Kiecolt-Glaser et al., 2014). The
findings of these three studies are inconsistent with those of the present meta-analysis, where no effect of
yoga asanas was seen on CRP or IL-6 (for subgroup analysis excluding MBSR based studies); however,
these three studies had no active control comparison conditions, while the studies included in the present
meta-analysis did.
4.8 Strengths and limitations.
A particular strength of this meta-analysis is that it only includes RCTs with an AC, making this the first
meta-analysis to investigate the immune modulating effects of practices including yoga asanas in well-
controlled studies. As yoga research is a still developing field and is characterised by an undefined
taxonomy (Ospina et al., 2007) studies in this field have historically, arguably been of poor
methodological rigour. As seen in the present meta-analysis, the conduct of RCTs with an AC are
becoming more common; however, the lingering perception of poor quality research may influence the
availability of research funding and accordingly negatively impact the future production of high quality
research trials. Thus, a meta-analysis assessing the impact of yoga interventions in only RCTs with an AC
group is both timely and important. The main limitation of the current meta-analysis is that many of the
reviewed studies do not include a follow up period, and thus the longevity of the observed effects is
unknown. Additionally, many of the included studies were classed as having a high risk of bias, however
we would suggest that this issue is not specific to research in the field, but rather a testimony of the
requirement of scientific research to increase transparency and rigour in publishing standards (Franco et
al., 2014; Macleod et al., 2015). For example, in one study included in this meta-analysis the authors did
not report statistical information for a non-significant differences in cortisol between the yoga asanas and
control group; this information was requested from the authors but was unable to be provided for
inclusion in the analysis (Cohen et al., 2016). A further limitation of the present meta-analysis is that only
studies published in English were eligible for inclusion, however as can be seen in Figure 1 no studies
were excluded due to being published in a language other than English, and thus it is unlikely that
relevant non English language studies were missed due to this inclusion criterion.
Given the high levels of stress experienced in modern society it is important to identify methods that can
protect against the persistent arousal. Our results suggest that interventions including yoga asanas can be
applied in various populations to decrease stress-related outcomes. It is possible that the effect of yoga
asanas on stress physiology contributes to the reported improved well-being associated with yoga asanas
practice (Cramer et al., 2013). Our findings indicate that interventions including yoga asanas improve
ANS and HPA axis regulation.
4.9 Conclusion
While yoga practice has become a popular method of stress management, its neurobiological
underpinnings are not well understood. This is the first meta-analysis of RCTs with ACs demonstrating
that yoga asanas practice appears to have inhibitory effects on physiological stress, as demonstrated by
decreased cortisol, decreased BP and cytokine levels. The current meta-analysis included a broad range of
populations, suggesting that the observed effects are not specific to a particular population. Importantly,
these effects are greater than those associated with an AC, most commonly exercise, therapy or education.
The authors declare no conflicts of interest
My co-authors have all contributed to this manuscript and approve of this submission. Neither this manuscript nor
substantial parts of it are under consideration for publication elsewhere, have been published nor made available
elsewhere in a manner that could be construed as a prior or duplicate publication of the same content. There is not a
manuscript of related content (e.g. from the same study with the same or very similar primary exposure and
outcome) under consideration for publication elsewhere, nor has one been published nor made available elsewhere. I
have communicated with all of my co-authors and obtained their full disclosures. My co-authors and I declare no
conflicts of interest. I have not submitted any images in colour.
Acknowledgments
Author contributions are as follows: MCP conceived the study including data sources and search strategy,
conducted the systematic search, performed study selection, extracted data, performed data synthesis and
wrote the manuscript. DRT conceived the study including data sources and search strategy and critically
appraised the manuscript. CFK conceived the study including data sources and search strategy and
critically appraised the manuscript. All authors take responsibility for the contents of this article. The
authors declare no conflict of interest. The authors thank Mr Maher Hana for his assistance with data
extraction.
Appendix: Search Strategy
Databases - CINAHL Complete; AMED - The Allied and Complementary Medicine
Database; MEDLINE Complete; PsycINFO; SocINDEX (Run in May and re-run in
December 2016)
ABSTRACT WORD neuroendocrine OR immune OR cortisol OR catecholamines OR
norepinephrine OR cytokine OR interleukin OR interferon OR blood pressure OR heart rate OR
inflammation OR proinflammatory AND ABSTRACT WORD biomarker OR blood OR saliva
OR urine OR telomere OR infection AND TITLE WORD yoga OR TI mindfulness based OR TI
mindfulness-based
Pubmed
MeSH or ABSTRACT WORD neuroendocrine OR immune OR cortisol OR catecholamines OR
inflammation OR proinflammatory AND MeSH or ABSTRACT WORD biomarker OR blood
OR saliva OR urine OR telomere OR infection AND MeSH or TITLE WORD yoga OR TI
mindfulness based OR TI mindfulness-based
Scopus
TI/ABS/KEYWORDS neuroendocrine OR immune OR cortisol OR catecholamines OR
inflammation OR proinflammatory AND TI/ABS/KEYWORDS biomarker OR blood OR saliva
OR urine OR telomere OR infection AND TI/ABS/KEYWORDS yoga OR TI mindfulness based
OR TI mindfulness-based
Cochrane-Central-Register-of-Controlled-Trials
MeSH terms neuroendocrine OR immune OR cortisol OR catecholamines OR norepinephrine
OR cytokine OR interleukin OR interferon OR blood pressure OR heart rate OR inflammation

OR proinflammatory AND MeSH terms biomarker OR blood OR saliva OR urine OR telomere
OR infection AND MeSH terms yoga OR TI mindfulness based OR TI mindfulness-based
References
Agelink, M.W., Boz, C., Ullrich, H., Andrich, J., 2002. Relationship between major depression
and heart rate variability. Clinical consequences and implications for antidepressive treatment.
Psychiatry Res 113, 139-149.
Alesci, S., Martinez, P.E., Kelkar, S., Ilias, I., Ronsaville, D.S., Listwak, S.J., Ayala, A.R.,
Licinio, J., Gold, H.K., Kling, M.A., Chrousos, G.P., Gold, P.W., 2005. Major depression is
associated with significant diurnal elevations in plasma interleukin-6 levels, a shift of its
circadian rhythm, and loss of physiological complexity in its secretion: clinical implications. J
Clin Endocrinol Metab 90, 2522-2530.
Babbar, S., Hill, J.B., Williams, K.B., Pinon, M., Chauhan, S.P., Maulik, D., 2016. Acute feTal
behavioral Response to prenatal Yoga: a single, blinded, randomized controlled trial (TRY
yoga). Am J Obstet Gynecol 214, 399 e391-398.
Bagga, O.P., Gandhi, A., 1983. A comparative study of the effect of Transcendental Meditation
(T.M.) and Shavasana practice on cardiovascular system. Indian Heart J 35, 39-45.
Billman, G.E., 2011. Heart rate variability - a historical perspective. Frontiers in physiology 2,
86.
Billman, G.E., 2013. The LF/HF ratio does not accurately measure cardiac sympatho-vagal
balance. Frontiers in physiology 4, 26.
Blumenthal, J.A., Emery, C.F., Madden, D.J., Coleman, R.E., Riddle, M.W., Schniebolk, S.,
Cobb, F.R., Sullivan, M.J., Higginbotham, M.B., 1991. Effects of exercise training on
cardiorespiratory function in men and women older than 60 years of age. Am J Cardiol 67, 633-
639.
Borenstein, M., Hedges, L. V., Higgins, J. P. T., Rothstein, H. R, 2009. Introduction to Meta-
Analysis. Wiley, Chichester.
Bower, J.E., Greendale, G., Crosswell, A.D., Garet, D., Sternlieb, B., Ganz, P.A., Irwin, M.R.,
Olmstead, R., Arevalo, J., Cole, S.W., 2014. Yoga reduces inflammatory signaling in fatigued
breast cancer survivors: a randomized controlled trial. Psychoneuroendocrinology 43, 20-29.
Bower, J.E., Irwin, M.R., 2016. Mind-body therapies and control of inflammatory biology: A
descriptive review. Brain, behavior, and immunity 51, 1-11.
Bowman, A.J., Clayton, R.H., Murray, A., Reed, J.W., Subhan, M.M., Ford, G.A., 1997. Effects
of aerobic exercise training and yoga on the baroreflex in healthy elderly persons. Eur J Clin
Invest 27, 443-449.
Brown, E.S., Varghese, F.P., McEwen, B.S., 2004. Association of depression with medical
illness: does cortisol play a role? Biological psychiatry 55, 1-9.
Carlson, L.E., Beattie, T.L., Giese-Davis, J., Faris, P., Tamagawa, R., Fick, L.J., Degelman, E.S.,
Speca, M., 2015. Mindfulness-based cancer recovery and supportive-expressive therapy maintain
telomere length relative to controls in distressed breast cancer survivors. Cancer 121, 476-484.
Carlson, L.E., Doll, R., Stephen, J., Faris, P., Tamagawa, R., Drysdale, E., Speca, M., 2013.
Randomized controlled trial of Mindfulness-based cancer recovery versus supportive expressive
group therapy for distressed survivors of breast cancer. Journal of clinical oncology : official
journal of the American Society of Clinical Oncology 31, 3119-3126.
Chacko, S.A., Yeh, G.Y., Davis, R.B., Wee, C.C., 2016. A mindfulness-based intervention to
control weight after bariatric surgery: Preliminary results from a randomized controlled pilot
trial. Complementary therapies in medicine 28, 13-21.
Clarke, T.C., Black, L.I., Stussman, B.J., Barnes, P.M., Nahin, R.L., 2015. Trends in the use of
complementary health approaches among adults: United States, 2002-2012. National health
statistics reports, 1-16.
Cohen, D.L., Bloedon, L.T., Rothman, R.L., Farrar, J.T., Galantino, M.L., Volger, S., Mayor, C.,
Szapary, P.O., Townsend, R.R., 2011. Iyengar Yoga versus Enhanced Usual Care on Blood
Pressure in Patients with Prehypertension to Stage I Hypertension: a Randomized Controlled
Trial. Evidence-based complementary and alternative medicine : eCAM 2011, 546428.
Cohen, D.L., Boudhar, S., Bowler, A., Townsend, R.R., 2016. Blood Pressure Effects of Yoga,
Alone or in Combination With Lifestyle Measures: Results of the Lifestyle Modification and
Blood Pressure Study (LIMBS). J Clin Hypertens (Greenwich) 18, 809-816.
Corey, S.M., Epel, E., Schembri, M., Pawlowsky, S.B., Cole, R.J., Araneta, M.R., Barrett-
Connor, E., Kanaya, A.M., 2014. Effect of restorative yoga vs. stretching on diurnal cortisol
dynamics and psychosocial outcomes in individuals with the metabolic syndrome: the PRYSMS
randomized controlled trial. Psychoneuroendocrinology 49, 260-271.
Cornelissen, V.A., Smart, N.A., 2013. Exercise training for blood pressure: a systematic review
and meta-analysis. J Am Heart Assoc 2, e004473.
Cramer, H., Lauche, R., Langhorst, J., Dobos, G., 2013. Yoga for depression: a systematic
review and meta-analysis. Depress Anxiety 30, 1068-1083.
Creswell, J.D., Irwin, M.R., Burklund, L.J., Lieberman, M.D., Arevalo, J.M., Ma, J., Breen, E.C.,
Cole, S.W., 2012. Mindfulness-Based Stress Reduction training reduces loneliness and pro-
inflammatory gene expression in older adults: a small randomized controlled trial. Brain,
behavior, and immunity 26, 1095-1101.
Creswell, J.D., Taren, A.A., Lindsay, E.K., Greco, C.M., Gianaros, P.J., Fairgrieve, A.,
Marsland, A.L., Brown, K.W., Way, B.M., Rosen, R.K., Ferris, J.L., 2016. Alterations in
Resting-State Functional Connectivity Link Mindfulness Meditation With Reduced Interleukin-
6: A Randomized Controlled Trial. Biol Psychiatry 80, 53-61.
Cusumano, J.A., Robinson, S. E., 1992. The short-term psychophysiological effects of hatha
yoga and progressive relaxation on female Japanese students. Applied Psychology: An
International Review 41, 77-90.
Ding, D., Stamatakis, E., 2014. Yoga practice in England 1997-2008: prevalence, temporal
trends, and correlates of participation. BMC research notes 7, 172.
Ebnezar, J., Nagarathna, R., Yogitha, B., Nagendra, H.R., 2012. Effect of integrated yoga
therapy on pain, morning stiffness and anxiety in osteoarthritis of the knee joint: A randomized
control study. International journal of yoga 5, 28-36.
Farmer, J., 2012. Yoga Body: The Origins of Modern Posture Practice. Rev Am Hist 40, 145-
158.
Feingold, K.R., Grunfeld, C., 2000. The Effect of Inflammation and Infection on Lipids and
Lipoproteins, in: De Groot, L.J., Chrousos, G., Dungan, K., Feingold, K.R., Grossman, A.,
Hershman, J.M., Koch, C., Korbonits, M., McLachlan, R., New, M., Purnell, J., Rebar, R.,
Singer, F., Vinik, A. (Eds.), Endotext, South Dartmouth (MA).
Field, T., Diego, M., Delgado, J., Medina, L., 2013. Yoga and social support reduce prenatal
depression, anxiety and cortisol. J Bodyw Mov Ther 17, 397-403.
Franco, A., Malhotra, N., Simonovits, G., 2014. Social science. Publication bias in the social
sciences: unlocking the file drawer. Science 345, 1502-1505.
Gard, T., Noggle, J.J., Park, C.L., Vago, D.R., Wilson, A., 2014. Potential self-regulatory
mechanisms of yoga for psychological health. Front Hum Neurosci 8, 770.
Gold, P.W., 2015. The organization of the stress system and its dysregulation in depressive
illness. Mol Psychiatry 20, 32-47.
Goldberg, S.B., Manley, A.R., Smith, S.S., Greeson, J.M., Russell, E., Van Uum, S., Koren, G.,
Davis, J.M., 2014. Hair Cortisol as a Biomarker of Stress in Mindfulness Training for Smokers. J
Altern Complem Med 20, 630-634.
Goldstein, D.S., Bentho, O., Park, M.Y., Sharabi, Y., 2011. Low-frequency power of heart rate
variability is not a measure of cardiac sympathetic tone but may be a measure of modulation of
cardiac autonomic outflows by baroreflexes. Exp Physiol 96, 1255-1261.
Goodyer, I.M., Herbert, J., Tamplin, A., Altham, P.M., 2000. Recent life events, cortisol,
dehydroepiandrosterone and the onset of major depression in high-risk adolescents. Br J
Psychiatry 177, 499-504.
Gothe, N.P., Keswani, R.K., McAuley, E., 2016. Yoga practice improves executive function by
attenuating stress levels. Biological psychology 121, 109-116.
Granath, J., Ingvarsson, S., von Thiele, U., Lundberg, U., 2006. Stress management: a
randomized study of cognitive behavioural therapy and yoga. Cogn Behav Ther 35, 3-10.
Grossman, P., Deuring, G., Walach, H., Schwarzer, B., Schmidt, S., 2016. Mindfulness-based
Intervention does not Influence Cardiac Autonomic Control or Pattern of Physical Activity in
Fibromyalgia During Daily Life: An Ambulatory, Multi-measure Randomized Controlled Trial.
Clin J Pain.
Hagins, M., Haden, S.C., Daly, L.A., 2013. A randomized controlled trial on the effects of yoga
on stress reactivity in 6th grade students. Evidence-based complementary and alternative
medicine : eCAM 2013, 607134.
Hagins, M., Rundle, A., Consedine, N.S., Khalsa, S.B., 2014. A randomized controlled trial
comparing the effects of yoga with an active control on ambulatory blood pressure in individuals
with prehypertension and stage 1 hypertension. J Clin Hypertens (Greenwich) 16, 54-62.
Hannestad, J., DellaGioia, N., Bloch, M., 2011. The effect of antidepressant medication
treatment on serum levels of inflammatory cytokines: a meta-analysis.
Neuropsychopharmacology 36, 2452-2459.
Harinath, K., Malhotra, A.S., Pal, K., Prasad, R., Kumar, R., Kain, T.C., Rai, L., Sawhney, R.C.,
2004. Effects of Hatha yoga and Omkar meditation on cardiorespiratory performance,
psychologic profile, and melatonin secretion. J Altern Complement Med 10, 261-268.
Hayney, M.S., Coe, C.L., Muller, D., Obasi, C.N., Backonja, U., Ewers, T., Barrett, B., 2014.
Age and psychological influences on immune responses to trivalent inactivated influenza vaccine
in the meditation or exercise for preventing acute respiratory infection (MEPARI) trial. Hum
Vaccin Immunother 10, 83-91.
Henriksen, E.J., Diamond-Stanic, M.K., Marchionne, E.M., 2011. Oxidative stress and the
etiology of insulin resistance and type 2 diabetes. Free radical biology & medicine 51, 993-999.
Herman, J.P., McKlveen, J.M., Ghosal, S., Kopp, B., Wulsin, A., Makinson, R., Scheimann, J.,
Myers, B., 2016. Regulation of the Hypothalamic-Pituitary-Adrenocortical Stress Response.
Compr Physiol 6, 603-621.
Hughes, J.W., Fresco, D.M., Myerscough, R., van Dulmen, M.H., Carlson, L.E., Josephson, R.,
2013. Randomized controlled trial of mindfulness-based stress reduction for prehypertension.
Psychosom Med 75, 721-728.
Innes, K.E., Selfe, T.K., 2012. The Effects of a Gentle Yoga Program on Sleep, Mood, and
Blood Pressure in Older Women with Restless Legs Syndrome (RLS): A Preliminary
Randomized Controlled Trial. Evidence-based complementary and alternative medicine : eCAM
2012, 294058.
Iwata, M., Ota, K.T., Duman, R.S., 2013. The inflammasome: pathways linking psychological
stress, depression, and systemic illnesses. Brain, behavior, and immunity 31, 105-114.
Jedel, S., Hoffman, A., Merriman, P., Swanson, B., Voigt, R., Rajan, K.B., Shaikh, M., Li, H.,
Keshavarzian, A., 2014. A randomized controlled trial of mindfulness-based stress reduction to
prevent flare-up in patients with inactive ulcerative colitis. Digestion 89, 142-155.
Jones, S.M., Guthrie, K.A., Reed, S.D., Landis, C.A., Sternfeld, B., LaCroix, A.Z., Dunn, A.,
Burr, R.L., Newton, K.M., 2016. A yoga & exercise randomized controlled trial for vasomotor
symptoms: Effects on heart rate variability. Complementary therapies in medicine 26, 66-71.
Jung, H.Y., Lee, H., Park, J., 2015. Comparison of the effects of Korean mindfulness-based
stress reduction, walking, and patient education in diabetes mellitus. Nurs Health Sci 17, 516-
525.
Kalupahana, N.S., Moustaid-Moussa, N., Claycombe, K.J., 2012. Immunity as a link between
obesity and insulin resistance. Mol Aspects Med 33, 26-34.
Kanaya, A.M., Araneta, M.R., Pawlowsky, S.B., Barrett-Connor, E., Grady, D., Vittinghoff, E.,
Schembri, M., Chang, A., Carrion-Petersen, M.L., Coggins, T., Tanori, D., Armas, J.M., Cole,
R.J., 2014. Restorative yoga and metabolic risk factors: the Practicing Restorative Yoga vs.
Stretching for the Metabolic Syndrome (PRYSMS) randomized trial. J Diabetes Complications
28, 406-412.
Kiecolt-Glaser, J.K., Bennett, J.M., Andridge, R., Peng, J., Shapiro, C.L., Malarkey, W.B.,
Emery, C.F., Layman, R., Mrozek, E.E., Glaser, R., 2014. Yoga's impact on inflammation,
mood, and fatigue in breast cancer survivors: a randomized controlled trial. Journal of clinical
oncology : official journal of the American Society of Clinical Oncology 32, 1040-1049.
Kiecolt-Glaser, J.K., Christian, L., Preston, H., Houts, C.R., Malarkey, W.B., Emery, C.F.,
Glaser, R., 2010. Stress, Inflammation, and Yoga Practice. Psychosom Med 72, 113-121.
Lanfranchi, P.A., Somers, V.K., 2002. Arterial baroreflex function and cardiovascular
variability: interactions and implications. American journal of physiology. Regulatory,
integrative and comparative physiology 283, R815-826.
LeMoult, J., Ordaz, S.J., Kircanski, K., Singh, M.K., Gotlib, I.H., 2015. Predicting first onset of
depression in young girls: Interaction of diurnal cortisol and negative life events. J Abnorm
Psychol 124, 850-859.
Li, A.W., Goldsmith, C.A., 2012. The effects of yoga on anxiety and stress. Alternative medicine
review : a journal of clinical therapeutic 17, 21-35.
Long Parma, D., Hughes, D.C., Ghosh, S., Li, R., Trevino-Whitaker, R.A., Ogden, S.M.,
Ramirez, A.G., 2015. Effects of six months of Yoga on inflammatory serum markers prognostic
of recurrence risk in breast cancer survivors. Springerplus 4, 143.
Macleod, M.R., Lawson McLean, A., Kyriakopoulou, A., Serghiou, S., de Wilde, A., Sherratt,
N., Hirst, T., Hemblade, R., Bahor, Z., Nunes-Fonseca, C., Potluru, A., Thomson, A.,
Baginskaite, J., Egan, K., Vesterinen, H., Currie, G.L., Churilov, L., Howells, D.W., Sena, E.S.,
2015. Risk of Bias in Reports of In Vivo Research: A Focus for Improvement. PLoS Biol 13,
e1002273.
Maes, M., 2008. The cytokine hypothesis of depression: inflammation, oxidative & nitrosative
stress (IO&NS) and leaky gut as new targets for adjunctive treatments in depression.
Neuroendocrinol Lett 29, 287-291.
Malarkey, W.B., Jarjoura, D., Klatt, M., 2013. Workplace based mindfulness practice and
inflammation: a randomized trial. Brain, behavior, and immunity 27, 145-154.
Michalsen, A., Grossman, P., Acil, A., Langhorst, J., Ludtke, R., Esch, T., Stefano, G.B., Dobos,
G.J., 2005. Rapid stress reduction and anxiolysis among distressed women as a consequence of a
three-month intensive yoga program. Medical science monitor : international medical journal of
experimental and clinical research 11, CR555-561.
Miller, J.J., Fletcher, K., Kabat-Zinn, J., 1995. Three-year follow-up and clinical implications of
a mindfulness meditation-based stress reduction intervention in the treatment of anxiety
disorders. Gen Hosp Psychiatry 17, 192-200.
Moher, D., Liberati, A., Tetzlaff, J., Altman, D.G., Group, P., 2010. Preferred reporting items for
systematic reviews and meta-analyses: the PRISMA statement. Int J Surg 8, 336-341.
Nerurkar, A., Yeh, G., Davis, R.B., Birdee, G., Phillips, R.S., 2011. When conventional medical
providers recommend unconventional medicine: results of a national study. Archives of internal
medicine 171, 862-864.
Nesse, R., M., Bhatnagar, S., Ellis, B., 2016. Evolutionary Origins and Functions of the Stress
Response System, Stress: Concepts, Cognition, Emotion, and Behavior Handbook of Stress
Series. Academic Press, pp. 95-101.
Nidhi, R., Padmalatha, V., Nagarathna, R., Ram, A., 2012. Effect of a yoga program on glucose
metabolism and blood lipid levels in adolescent girls with polycystic ovary syndrome. Int J
Gynaecol Obstet 118, 37-41.
Olefsky, J.M., Glass, C.K., 2010. Macrophages, inflammation, and insulin resistance. Annu Rev
Physiol 72, 219-246.
Ospina, M.B., Bond, K., Karkhaneh, M., Tjosvold, L., Vandermeer, B., Liang, Y., Bialy, L.,
Hooton, N., Buscemi, N., Dryden, D.M., Klassen, T.P., 2007. Meditation practices for health:
state of the research. Evid Rep Technol Assess (Full Rep), 1-263.
Palta, P., Page, G., Piferi, R.L., Gill, J.M., Hayat, M.J., Connolly, A.B., Szanton, S.L., 2012.
Evaluation of a mindfulness-based intervention program to decrease blood pressure in low-
income African-American older adults. J Urban Health 89, 308-316.
Pariante, C.M., 2003. Depression, stress and the adrenal axis. Journal of neuroendocrinology 15,
811-812.
Pascoe, M.C., Bauer, I.E., 2015. A systematic review of randomised control trials on the effects
of yoga on stress measures and mood. Journal of psychiatric research 68, 270-282.
Patil, S.G., Aithala, M.R., Das, K.K., 2015. Effect of yoga on arterial stiffness in elderly subjects
with increased pulse pressure: A randomized controlled study. Complementary therapies in
medicine 23, 562-569.
Penman, S., Cohen, M., Stevens, P., Jackson, S., 2012. Yoga in Australia: Results of a national
survey. International journal of yoga 5, 92-101.
Pflueger, L.W., 2011. Yoga Body: The Origins of Modern Posture Practice. Relig Stud Rev 37,
235-235.
Praissman, S., 2008. Mindfulness-based stress reduction: a literature review and clinician's
guide. Journal of the American Academy of Nurse Practitioners 20, 212-216.
Pruessner, M., Hellhammer, D.H., Pruessner, J.C., Lupien, S.J., 2003. Self-reported depressive
symptoms and stress levels in healthy young men: associations with the cortisol response to
awakening. Psychosom Med 65, 92-99.
Pullen, P.R., Nagamia, S.H., Mehta, P.K., Thompson, W.R., Benardot, D., Hammoud, R.,
Parrott, J.M., Sola, S., Khan, B.V., 2008. Effects of yoga on inflammation and exercise capacity
in patients with chronic heart failure. J Card Fail 14, 407-413.
Pullen, P.R., Thompson, W.R., Benardot, D., Brandon, L.J., Mehta, P.K., Rifai, L., Vadnais,
D.S., Parrott, J.M., Khan, B.V., 2010. Benefits of yoga for African American heart failure
patients. Med Sci Sports Exerc 42, 651-657.
Rao, R.M., Nagendra, H.R., Raghuram, N., Vinay, C., Chandrashekara, S., Gopinath, K.S.,
Srinath, B.S., 2008. Influence of yoga on mood states, distress, quality of life and immune
outcomes in early stage breast cancer patients undergoing surgery. International journal of yoga
1, 11-20.
Ray, U.S., Sinha, B., Tomer, O.S., Pathak, A., Dasgupta, T., Selvamurthy, W., 2001. Aerobic
capacity & perceived exertion after practice of Hatha yogic exercises. Indian J Med Res 114,
215-221.
Riley, K.E., Park, C.L., 2015. How does yoga reduce stress? A systematic review of mechanisms
of change and guide to future inquiry. Health psychology review, 1-30.
Ruby, M., Repka, C.P., Arciero, P.J., 2016. Comparison of Protein-Pacing Alone or With
Yoga/Stretching and Resistance Training on Glycemia, Total and Regional Body Composition,
and Aerobic Fitness in Overweight Women. J Phys Act Health 13, 754-764.
Saptharishi, L., Soudarssanane, M., Thiruselvakumar, D., Navasakthi, D., Mathanraj, S.,
Karthigeyan, M., Sahai, A., 2009. Community-based Randomized Controlled Trial of Non-
pharmacological Interventions in Prevention and Control of Hypertension among Young Adults.
Indian J Community Med 34, 329-334.
Sawane, M.V., Gupta, S.S., 2015. Resting heart rate variability after yogic training and
swimming: A prospective randomized comparative trial. International journal of yoga 8, 96-102.
Sieverdes, J.C., Mueller, M., Gregoski, M.J., Brunner-Jackson, B., McQuade, L., Matthews, C.,
Treiber, F.A., 2014. Effects of Hatha yoga on blood pressure, salivary alpha-amylase, and
cortisol function among normotensive and prehypertensive youth. J Altern Complement Med 20,
241-250.
Tall, A.R., Yvan-Charvet, L., 2015. Cholesterol, inflammation and innate immunity. Nat Rev
Immunol 15, 104-116.
Task Force of the European Society of Cardiology, t.N.A.S.o.P.a.E., 1996. Heart rate variability.
Standards of measurement, physiological interpretation, and clinical use. Task Force of the
European Society of Cardiology and the North American Society of Pacing and
Electrophysiology. Eur Heart J 17, 354-381.
Telles, S., Srinivas, R. B., 1998. Autonomic and respiratory measures in children with impaired
vision following yoga and physical activity programs. International Journal of Rehabilitation &
Health 4, 17-122.
Teofilo, M.M., Farias, D.R., Pinto Tde, J., Vilela, A.A., Vaz Jdos, S., Nardi, A.E., Kac, G., 2014.
HDL-cholesterol concentrations are inversely associated with Edinburgh Postnatal Depression
Scale scores during pregnancy: results from a Brazilian cohort study. J Psychiatr Res 58, 181-
188.
Thayer, J.F., Hansen, A.L., Johnsen, B.H., 2010. The non-invasive assessment of autonomic
influences on the heart using impedance cardiography and
heart rate variability, Handbook of Behavioral Medicine. Springer, New York, pp. 723-740.
The Cochrane Collaboration, 2011. Cochrane Handbook for Systematic Reviews of
Interventions, in: Higgins, J., Green, S. (Ed.).
Thiyagarajan, R., Pal, P., Pal, G.K., Subramanian, S.K., Trakroo, M., Bobby, Z., Das, A.K.,
2015. Additional benefit of yoga to standard lifestyle modification on blood pressure in
prehypertensive subjects: a randomized controlled study. Hypertens Res 38, 48-55.
Travis, F., Pearson, C., 2000. Pure consciousness: Distinct phenomenological and physiological
correlates of "consciousness itself". Int J Neurosci 100, 77-89.
Vadiraja, H.S., Raghavendra, R.M., Nagarathna, R., Nagendra, H.R., Rekha, M., Vanitha, N.,
Gopinath, K.S., Srinath, B.S., Vishweshwara, M.S., Madhavi, Y.S., Ajaikumar, B.S., Ramesh,
B.S., Nalini, R., Kumar, V., 2009. Effects of a yoga program on cortisol rhythm and mood states
in early breast cancer patients undergoing adjuvant radiotherapy: a randomized controlled trial.
Integrative cancer therapies 8, 37-46.
van Reedt Dortland, A.K., Giltay, E.J., van Veen, T., van Pelt, J., Zitman, F.G., Penninx, B.W.,
2010. Associations between serum lipids and major depressive disorder: results from the
Netherlands Study of Depression and Anxiety (NESDA). J Clin Psychiatry 71, 729-736.
Ventriglio, A., Gentile, A., Baldessarini, R.J., Bellomo, A., 2015. Early-life stress and
psychiatric disorders: epidemiology, neurobiology and innovative pharmacological targets. Curr
Pharm Des 21, 1379-1387.
Wardle, J., Adams, J., Sibbritt, D., 2014. Referral to yoga therapists in rural primary health care:
A survey of general practitioners in rural and regional New South Wales, Australia. International
journal of yoga 7, 9-16.
Wesseling, K.H., Settels, J.J., 1985. Baromodulation explains short-term blood pressure
variability, in: Orlebeke , J.F., Mulder, G., Van Doornen, L.P.J. (Eds.), The psychophysiology of
cardiovascular control. Plenum, New York, pp. 69–97.
Wysokinski, A., Strzelecki, D., Kloszewska, I., 2015. Levels of triglycerides, cholesterol, LDL,
HDL and glucose in patients with schizophrenia, unipolar depression and bipolar disorder.
Diabetes Metab Syndr 9, 168-176.
Yang, K., Bernardo, L.M., Sereika, S.M., Conroy, M.B., Balk, J., Burke, L.E., 2011. Utilization
of 3-month yoga program for adults at high risk for type 2 diabetes: a pilot study. Evidence-
based complementary and alternative medicine : eCAM 2011, 257891.
Ziv, A., Vogel, O., Keret, D., Pintov, S., Bodenstein, E., Wolkomir, K., Doenyas, K., Mirovski,
Y., Efrati, S., 2013. Comprehensive Approach to Lower Blood Pressure (CALM-BP): a
randomized controlled trial of a multifactorial lifestyle intervention. J Hum Hypertens 27, 594-
600.
Table 1. Characteristics of included studies
Reference Countr Setting Study Participants Intervention Control Time of Relevant Follow
y Design assessment Outcome Up
Measures
Babbar, USA Hospital Parallel Pregnant women, 18-45 Prenatal yoga Health Pre-post SBP, None
2016 group years education intervention; DBP, HR
presentati follow up (resting 1x
on at each
time
point)
Blumenthal USA Communi Crossover Healthy individuals, 60- Yoga Exercise Pre-post HR None
, 1991 ty based 83 years (unspecified) (aerobic) intervention (resting,
maximal
[continuou
s]); LDL,
HDL,
cholestero
l,
triglycerid
es (fasting
plasma)
Bower, USA Medical Parallel Women with breast Iyengar yoga Health Pre-post Cortisol 12 wks
2014 centre group cancer (stage 0-II) who education intervention; (salivary
had completed local follow up 4x daily/2
and/or adjuvant therapy days
(with the exception of [waking,
endocrine therapy) at 30m, 8h
least 6 months previously after
waking,
bedtime]),
CRP, IL-
6, TNF,
IL-1
receptor
antagonist
(plasma)
Bowman, UK Not Crossover Healthy, sedentary, Hatha yoga Exercise Pre-post HR, HRV, None
1997 specified elderly individuals (age (aerobic) intervention SBP,
ns) baroreflex
sensitivity
(resting
[mean of
continuou
s 20 min
recording]
)
Carlson, Canada Multi-site Parallel Women with breast MBCR Supportiv Pre-post T/S ratio, None
2013 cancer/ group cancer (stage I, II, or III) e- intervention cortisol
medical expressiv (salivary
centre e therapy 4x daily/3
OR days
didactic [waking,
stress 12:00,
managem 17:00,
ent bedtime])
seminar
Chacko USA Medical Parallel Patients who have MBSR, Weight Pre-post Hs-CRP, 16 wks
2016 Centre group undergone bariatric Mindfulness- managem intervention IL-6,
surgery 1-5 years prior to based eating ent HE TNF-
study start, weight loss awareness alpha
plateau (< 5lbs weight (MB-EAT)
loss in past month), aged
18-65
Cohen, USA Universit Parallel Individuals with Iyengar yoga Enhanced Pre-post SBP, None
2011 y and group - prehypertension/stage 1 usual care intervention DBP, HR,
communit open label hypertension, 22-69 MAP
y based years (ambulato
ry
continuou
s
monitorin
g) *non-
significant
cortisol
measures
not
reported in
paper*
Cohen, USA Communi Parallel Individuals with Hatha yoga Health Pre-post SBP, DBP None
2016 ty based group - prehypertension/stage 1 education intervention (ambulato
open label hypertension, 18-80 and ry
years walking continuou
s
monitorin
g)
Corey, USA Multi-site Parallel Individuals with Restorative Stretchin Pre-post Cortisol None
2014 Universit group - metabolic syndrome, 21- yoga g intervention (salivary
y open label 65 years 4x
daily/3day
s [waking,
30 min
and 60
min after
waking,
bedtime])
Creswell, USA Retreat Parallel Healthy, stressed, MBSR (HEM) Relaxatio Pre- IL-6 16 wks
2016 centre group unemployed individuals, Program n retreat intervention; (plasma
24-54 years follow up [10:00-
12:00])
Cusumano, Japan Universit Parallel Students, female Hatha yoga PMR Pre-post BP (SBP None
1992 y group- Japanese undergraduates, intervention and DBP
cluster 18-20 years combined)
randomisati , HR
on (averages
of
multiple
measures)
Ebnezar, India Medical Parallel Individuals with Integrated Exercise Pre-post SBP, None
2012 centre group osteoarthritic knees, 35- yoga and intervention DBP, HR
80 years physiothe (resting
rapy [ns])
Field, 2013 USA Universit Parallel Pregnant women with Prenatal yoga Social Pre-post Cortisol, None
y Medical group depression support intervention estriol,
Centre progestero
ne
(salivary
[mid-
morning])
Goldberg, USA Not Parallel Individuals who smoke, MBSR CBT Pre-post Cortisol None
2014 specified group ≥18 years (25-65 for hair intervention (hair)
donation)
Gothe, USA Universit Parallel Sedentary community Hatha yoga Stretchin Pre-post Cortisol None
2016 y group dwelling adults, 55-79 g and intervention (salivary
years strengthe 2x daily
ning [14:00,
14:40])
Granath, Sweden Work Parallel Healthy individuals with Yoga CBT Pre-post Adrenalin None
2006 place group self-reported stress (age (unspecified) intervention e,
ns) noradrenal
ine (urine
[ns]),
cortisol
(salivary
[ns]),
SBP,
DBP, HR
(resting
3x)
Grossman German Not Parallel Individuals with MBSR Relaxatio Pre-post Breathing 8 wks
2016 y specified group, fibromyalgia n intervention frequency
open-label (breaths
per min),
HR
(ambulato
ry)
Hagins, USA Primary Parallel Students (grade six) Yoga Physical Pre-post SBP, None
2013 school group (unspecified) education intervention DBP, HR
(resting
2x, during
stressor
2x)
Hagins, USA Universit Parallel Individuals with Ashtanga Exercise Pre-post SBP, None
2014 y group prehypertension/stage 1 yoga intervention DBP, HR,
hypertension, 21-70 MAP
years (ambulato
ry
continuou
s
monitorin
g)
Harinath, India Army unit Parallel Healthy male army Hatha yoga Physical Pre-post SBP, None
2004 group soldiers, 25–35 years and Omkar training intervention DBP,
meditation (army) MAP
(resting
[ns]) HR
(5 min
continuou
s)
Hayney, USA Communi Parallel Healthy adults, ≥50 years MBSR Exercise Pre-post Gene 10 wks
2014 ty based group intervention; regulation
follow up
Hughes, USA Medical Parallel Individuals with MBSR PMR Pre-post SBP, DBP None
2013 centre group prehypertension, intervention (resting
unmedicated, 30-60 3x)
years
Innes, 2012 USA Universit Parallel Women with restless legs Iyengar yoga Education Pre-post SBP, None
y group syndrome, 45–79 years al film intervention DBP, HR
(resting
3x)
Jedel, 2014 USA Universit Parallel Individuals with MBSR Health Pre-post Calprotect 18 and
y group ulcerative colitis, education intervention; in (stool), 44 wks
inactive at recruitment, follow up cortisol
18–70 years (urinary),
ACTH
(fasting),
CRP, IL-
10, IL-6,
IL-8
(serum)
Jones, 2016 USA Universit Parallel Women with vasomotor Yoga Exercise Pre-post HRV None
y and group symptoms, 40–62 years (unspecified) (aerobic) intervention
research
centre
Jung, 2015 Korea Multi-site Parallel Individuals with type 2 MBSR Patient Pre-post Cortisol None
communit group - diabetes (Korean) education intervention (plasma
y health cluster OR [fasting]),
centres randomisati walking FBG,
and on exercise PAI-1, t-
hospitals PA (serum
[fasting])
Kanaya, USA Universit Parallel Individuals with Restorative Stretchin Pre-post Insulin, None
2014 y group metabolic syndrome, 21- yoga g intervention FBG,
65 years triglycerid
es, HDL
(serum
[fasting]),
fasting 2-
hr
glucose,
HbA1c,
HOMA-
IR, SBP
(resting
3x)
Long USA Communi Parallel Women with breast Hatha yoga Exercise Pre-post IL-6, IL-8, None
Parma, ty based group cancer or ductal (aerobic, intervention TNFα,
2015 carcinoma in-situ, ≥18 resistance CRP
years and (serum
flexibility [not
) OR fasting])
DIVA
exercise
classes
Nidhi, 2012 India Residenti Parallel Students, girls with Yoga Exercise Pre-post FBG, None
al school group polycystic ovary (unspecified) and intervention HDL,
syndrome, 15-18 years supine LDL,
rest VLDL,
TC/HDL,
insulin,
cholestero
l,
triglycerid
es, FI,
HOMA-
IR (serum
[fasting
06:00-
08:00])
Palta, 2012 USA Senior Parallel Elderly individuals, MBSR Social Pre-post SBP, DBP None
housing group African-American, ≥62 (ELDERSHIN support intervention (resting
facility years E) 3x)
Patil, 2015 India Medical Parallel Men with increased pulse Integrated Yoga Pre-post DBP, None
college, group pressure, 60-75 years yoga (unspecifi intervention SBP,
hospital ed), MAP,
and qigong, pulse
research counselli pressure
centre ng and (resting 3x
diet daily/3day
s), NO
(serum)
Rao, 2008 India Hospital Parallel Women with breast Integrated Supportiv Pre-post Lymphocy None
group cancer, operable, 30-70 yoga e intervention tes
years counselli (CD4+,
ng and CD56+,
exercise CD8+),
rehabilitat IgA, IgG,
ion IgM
(serum
[08:00-
12:00])
Ray, 2001 India Army unit Parallel Healthy men, 19-23 Hatha yoga Physical Pre-post HR None
group years training intervention (maximal
(army) [continual
])
Ruby, 2016 USA Not Parallel Women who were Yoga Protein Pre-post HDL, None
specified group - overweight or obese, 25- (unspecified), diet OR intervention LDL,
open label 60 years stretching and resistance cholestero
protein diet exercise l,
and diet triglycerid
es, blood
glucose
(plasma
[fasting
06:00-
10:00]),
DBP,
SBP, HR
(resting
[30 min
continuou
s])
Saptharishi, India Communi Parallel Individuals with Yoga Physical Pre-post SBP, DBP None
2009 ty based group prehypertension/hyperten (unspecified) exercise intervention (resting
sion (age ns) OR salt [ns])
reduction
Sawane, India Communi Parallel Healthy adults with Iyengar yoga Swimmin Pre-post DBP, None
2015 ty based group sedentary occupations, g intervention SBP, HR,
18–40 years HRV,
LF/HF
ratio
(resting
[ns])
Sieverdes, USA Primary Parallel Students (grade seven) Hatha yoga Music Pre-post Cortisol None
2014 school group and art intervention and α-
classes Amylase
(salivary
5x
daily/1day
[waking,
before
leaving
bed, 30
min and
60 min
after
waking,
bedtime])
SBP,
DBP, HR
Telles, India Special Parallel Students with impaired Yoga Physical Pre-post HR, None
1998 needs group vision, 11-17 years (unspecified) activity intervention respiration
school rate, SR
(resting)
Thiyagaraja India Research Parallel Individuals with Yoga Lifestyle Pre-post SBP, None
n, 2015 centre group - prehypertension, 20–60 (unspecified) modificat intervention DBP, HR,
open label years and lifestyle ion MAP
modification (resting
2x), FBG,
HDL,
LDL,
cholestero
l,
triglycerid
es (plasma
[fasting
07:00-
09:00])
Vadiraja, India Hospital Parallel Women with breast Integrated Brief Pre-post Cortisol None
2009 group cancer, prescribed yoga supportiv intervention (salivary
adjuvant radiotherapy, e therapy 3x
30-70 years daily/3day
s [06:00,
09:00,
21:00])
Yang, 2011 USA Not Parallel Individuals, non-active, Vinyasa yoga Health Pre-post SBP, DBP None
specified group family history of type 2 education intervention (resting
diabetes with either [ns]),
impaired fasting glucose; insulin,
prehypertension; FBG,
overweight/obese; or HDL,
abnormal level of LDL,

cholesterol, 45-65 years cholestero
l,
triglycerid
es (plasma
[fasting])
Ziv, 2013 Israel Medical Parallel Individuals with Yoga Exercise Pre-post SBP, DBP None
centre group hypertension, 22–75 (unspecified), and diet intervention (3x
years qigong, daily/3x
counselling days/2x
and diet days - self
collected),
FBG,
HDL,
LDL,
cholestero
l,
triglycerid
es (plasma
[fasting])
ACTH = adrenocorticotropic hormone; CBT = cognitive behavioural therapy; CRP = C-reactive protein; DBP = diastolic blood pressure; DIVA
=deriving inspiration and vitality through activity; FBG = fasting blood glucose; FI = fasting insulin; HbA1c = haemoglobin A1c; HDL = high
density lipoprotein; HEM = health enhancement through mindfulness; HOMA-IR = homeostatic model assessment and insulin resistance; HR =
heart rate; HRV = heart rate variability; IgA = immunoglobulin A; IgG = immunoglobulin G; IgM = immunoglobulin M; IL-1=interleukin-1; IL-6
= interleukin-6; IL-8 = interleukin-8; IL-10 = interleukin-10; LDL = low density lipoprotein; LF/HF = low frequency/high frequency; MAP =
mean arterial pressure; MBCR = mindfulness-based cancer recovery; MBSR = mindfulness-based stress reduction program; NO = nitric oxide; ns
= not specified; PAI-1 = plasminogen activator inhibitor type 1; PMR = progressive muscle relaxation; SBP = systolic blood pressure; SR = skin
resistance; TC = total cholesterol; TNF = tumour necrosis factor; TNFα = tumour necrosis factor-alpha; t-PA = tissue plasminogen activator ; T/S
ratio = relative telomere to single copy gene; VLDL = very low density lipoprotein
Table 3: Risk of bias assessment for included studies
Study Sequence Allocation Blinding Blinding Incomplet Selective Other

Generatio concealm of of e outcome outcome sources of
n ent participa outcome data reporting bias
nts and assessor
personne
l
Babbar L L L L L L UC
2016
Blumenth UC UC L UC UC UC H
al 1991
Bower L L L UC L UC L
2014
Bowman UC UC L UC UC H H
1997
Carlson L L L L L L L
2013
Chacko L L L L L L L
2016
Cohen UC UC L L H H H
2011
Cohen L L L UC L H H
2016
Corey L UC L L UC L L
2014
Creswell L L L L L H UC
2016
Cusuman H UC L UC UC UC L
o 1993
Ebnezar L L L L L UC UC
2012
Field L UC L L UC UC H
2013
Goldberg L
2014 UC UC UC H UC L
Gothe UC UC L L L H L
2016
Granath UC UC L UC L UC H
2006
Grossman UC UC L UC L L H
2016
Hagins L L L L L UC L
2013
Hagins L L L L H L L
2014
Harinath L UC L UC UC UC UC
2004
Hayney L L L L L L UC
2014
Hughes L L L L L L H
2013
Innes L L L L L UC L
2012
Jedel L L L L L L L
2014
Jones L L L L L H UC
2016
Jung 2015 L L L L H UC L
Kanaya L UC L L L L H
2014
LongParm L UC L UC L UC L
a 2015
Nidhi UC L L L L H H
2012
Palta 2012 UC UC L UC L UC H
Patil 2015 L UC L UC L L L
Raghaven L L L UC L UC L
dra 2009
Rao 2008 L L L L L UC UC
Ray 2001 UC UC L UC H UC H
Ruby UC UC L L L L L
2016
Saptharish L UC L UC L UC H
i 2009
Sawane UC UC L UC UC UC UC
2015
Sieverdes UC UC L UC H UC L
2014
Telles UC UC L UC L H UC
1998
Thiyagara UC L L UC H UC L
jan 2015
Yang UC UC L L L UC UC
2011
Ziv 2013 UC UC L UC H UC H
Sequence Generation: UC = method of randomisation not stated; H = Block randomisation of class

with only two blocks and two classes in an unblended trial (Cusumano, 1992). Allocation: UC = does
not specify if allocation concealment maintained. Blinding of participants and personnel: All studies
were rated as having a low risk of bias on physiological outcomes, as even though participants and
personnel were not blind, knowledge of the assigned intervention is not considered likely to impact on
the physiological outcomes assessed in the current meta-analysis Blinding of outcome assessor: UC =
not specified if outcome assessor blind or not or if blinding likely to affect outcome result. Incomplete
outcome data: UC = drop out numbers and/or reasons in each group not stated; H = very high dropout
rates (Cohen et al., 2011; Cohen et al., 2016; Thiyagarajan et al., 2015), included only participants
who complied with the protocol (Ruby et al., 2016), reasons for drop out seem to differ between
groups (Jung et al., 2015; Ray et al., 2001; Ziv et al., 2013), participants who dropped out differed
significantly from completers (Hagins et al., 2014), excluded one participant due to non-adherence
(Sieverdes et al., 2014), study used a subset of participants from another study (Goldberg et al., 2014).
Selective outcome reporting: UC = as no protocol paper; H = an outcome in protocol paper does not
seem to be reported (Creswell et al., 2016; Gothe et al., 2016; Jones et al., 2016; Nidhi et al., 2012),
Diastolic blood pressure not reported in results (Bowman et al., 1997), BP taken at baseline but not
post intervention (Telles, 1998), results of biochemical measures not reported as differences were not
significant (Cohen et al., 2011). Other sources of bias: UC = authors did not collect measures of home
practice in days following intervention and leading up to MRI - within 2 weeks after intervention
(Creswell et al., 2016), seem to be baseline differences between groups (Babbar et al., 2016), unclear
if baseline differences between groups (Harinath et al., 2004; Hayney et al., 2014; Jones et al., 2016;
Rao et al., 2008; Yang et al., 2011), it is not stated if multiple measures of BP or pulse were taken
(Ebnezar et al., 2012; Sawane and Gupta, 2015; Telles, 1998; Yang et al., 2011), IgA not measured at
baseline in first cohort (Hayney et al., 2014); H = deviation from protocol (extended the duration of
the intervention and only some participants engaged in this (Blumenthal et al., 1991), changed
randomisation schedule (Cohen et al., 2011), baseline differences present in factors that are possibly
related to outcome measures (Cohen et al., 2016; Hughes et al., 2013; Kanaya et al., 2014; Nidhi et
al., 2012; Palta et al., 2012; Saptharishi et al., 2009; Ziv et al., 2013), Systolic and diastolic blood
pressure were combined into one measure and there was experimenter interaction with treatment
group confounding the results (Cusumano, 1992), single cortisol measure taken at each time point
(Field et al., 2013; Granath et al., 2006), Blood pressure measurements were taken by participants at
home and more side effects reported by people in the intervention group compared to control group
(Ziv et al., 2013), does not say if multiple measures of blood pressure taken (Blumenthal et al., 1991;
Saptharishi et al., 2009)
Fig 1
Fig 2
Fig 3
Fig 4
Fig 5

J Psyneuen 2017 08 008

Uploaded by

Copyright:

Available Formats

J Psyneuen 2017 08 008

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

J Psyneuen 2017 08 008

Uploaded by

Copyright:

Available Formats

Accepted Manuscript

Title: Yoga, mindfulness-based stress reduction and

Authors: Michaela C. Pascoe, David R. Thompson, Chantal F.

Received date: 12-1-2017

Yoga, mindfulness-based stress reduction and stress-related physiological measures: A meta-analysis

Running Title: Yoga, MBSR and stress-related physiological measures

*Michaela C. Pascoe, PhD (Michaela.Pascoe@petermac.org)1

David R. Thompson, PhD (David.Thompson@unimelb.edu.au)2,3

Chantal F. Ski, PhD (Chantal.Ski@unimelb.edu.au)2

Number of Tables: 2; Number of Figures: 6

Stress related physiological measures assessed

42 studies including all populations reviewed

Background and Objectives

active control, on physiological markers of stress.

Materials and Methods

versus non-mindfulness-based stress reduction based interventions, different populations, length of

nervous system and hypothalamic-pituitary-adrenal system in various populations.

Key words: Yoga, Mindfulness-based stress reduction, Stress, Inflammation, Exercise

Ventriglio et al., 2015).

patients to a yoga therapist (Wardle et al., 2014).

reduce stress and anxiety in diverse patient populations (Praissman, 2008).

(SNS) and hypothalamic-pituitary-adrenal (HPA) system in people experiencing depressive

relevance of these findings in the context of stress-related depression.

was not previously published.

Olefsky and Glass, 2010).

for assessment of bias or meta-analysis inclusion.

2.3 Search strategy

request unpublished data.

2.4 Study selection

2.5 Data extraction

Software (https://www.covidence.org) and included assessments of sequence generation, allocation

2.7 Summary measures

sampling error as shown by the I2 statistic (Borenstein, 2009).

2.8 Data analysis

published papers, we used a 0.5 correlation for all analyses.

3.2 Study selection

inclusion and exclusion (Figure 1).

Insert Figure 1 here

3.3 Study characteristics

CI within each group.

Insert Tables 1 and 2 here

the descriptive statistics.

Insert Table 3 here

assessed using the raw difference in means.

of the yoga intervention was found (p=0.20, I2=0%).

Sieverdes resulted in p-values of 0.08, 0.31, 0.09 and 0.09 respectively.

3.5.1.1 Comparison of MBSR and non-MBSR interventions on cortisol outcomes

Insert Figure 2 here

3.5.2 Inflammatory changes

3.5.2.1 Interleukin-8 (IL-8)

asanas was found using SMD (p=0.16, I2=0%).

3.5.2.2 Interleukin 6 (IL-6)

0.33(-0.65, -0.01), p=0.04, I2=0%) (see supplementary data).

3.5.2.2.1 Comparison of MBSR and non-MBSR interventions on IL-6 outcomes

interventions. These conflicting results are depicted on Figure 3.

3.5.2.3 C-reactive protein (CRP)

similarly showed no effect of interventions including yoga asanas on CRP levels.

3.5.2.3.1 Comparison of MBSR and non-MBSR interventions on CRP outcomes

influence CRP outcomes.