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

The researchers tested whether interventions that include yoga asanas (the physical postures of yoga) — either alone or as part of a mindfulness-based stress reduction (MBSR) program — could reduce physiological markers of stress compared to an active control condition (e.g., walking, stretching, health education, or usual care).

The primary outcomes were:

**Autonomic nervous system markers:** heart rate, blood pressure (systolic and diastolic), heart rate variability, mean arterial pressure

**HPA axis markers:** cortisol (measured at waking, evening, and throughout the day)

**Inflammatory markers:** C-reactive protein (CRP), interleukin-6 (IL-6), interleukin-8 (IL-8)

**Metabolic markers:** fasting blood glucose, total cholesterol, low-density lipoprotein (LDL) cholesterol

The comparator was always an active control — meaning the control group did something, not nothing. This is important because it separates the specific effects of yoga from the general effects of doing any structured activity.

The meta-analysis included **42 randomised controlled trials** with a total of **1,780 participants** across all populations. The studies included:

Healthy adults (general population, no specific diagnosis)

People with depression or anxiety disorders

People with cardiovascular disease (including hypertension)

People with type 2 diabetes

People with cancer (various types)

People with chronic pain conditions

Pregnant women

Older adults (aged 60+)

The studies were conducted in the USA, India, Iran, Australia, Germany, Sweden, South Korea, and the UK. Participants ranged from young adults (mean age ~20) to older adults (mean age ~70). Most studies included both men and women, though several were women-only.

Importantly, the authors note that many individual studies had small sample sizes (often 20–60 participants total), which is a limitation of the primary literature.

The included studies used standard clinical and laboratory measurements:

**Cortisol:** Salivary cortisol samples (collected at waking, evening, or multiple times across the day) or serum cortisol from blood draws. Cortisol is the primary stress hormone released by the HPA axis.

**Blood pressure:** Automated or manual sphygmomanometer readings, taken in a seated position after rest. Some studies used 24-hour ambulatory blood pressure monitoring.

**Heart rate:** Measured via ECG or pulse oximetry, typically after a rest period of 5–10 minutes.

**Heart rate variability (HRV):** Derived from ECG recordings, measuring the variation in time between heartbeats. Higher HRV generally indicates better autonomic regulation.

**Fasting blood glucose:** Measured from blood samples after an overnight fast (typically 8–12 hours).

**Lipid profile (total cholesterol, LDL):** Measured from fasting blood samples using standard enzymatic assays.

**Inflammatory markers (CRP, IL-6, IL-8):** Measured from blood samples using ELISA or similar immunoassay techniques.

**Study design:** This is a systematic review and meta-analysis of randomised controlled trials (RCTs). A meta-analysis statistically combines results from multiple studies to produce a single estimate of effect size, which gives more statistical power than any individual study.

**Inclusion criteria:** Only RCTs were included — this is the gold standard for establishing causality. All studies had to include an active control group (not just a waitlist or no-treatment control). This is a major strength because it controls for placebo effects and the general benefits of participating in any structured activity.

**Search strategy:** The authors searched seven electronic databases (MEDLINE, AMED, CINAHL, PsycINFO, SocINDEX, PubMed, Scopus) in May 2016, with an update in December 2016. Two independent reviewers screened studies and extracted data.

**Statistical approach:** A random-effects model was used for all analyses. This assumes that the true effect of yoga might vary across studies (due to different populations, intervention types, etc.), which is more conservative and appropriate than a fixed-effects model. Effect sizes were reported as Hedges' g (a form of standardised mean difference), where 0.2 = small, 0.5 = medium, 0.8 = large.

**Risk of bias assessment:** The authors used the Cochrane Risk of Bias Tool and the GRADE framework. Most studies had unclear or high risk of bias in at least one domain, particularly:

Blinding of participants (impossible in yoga studies — you know if you're doing yoga)

Blinding of outcome assessors (often not reported)

Allocation concealment (often unclear)

**Sensitivity analyses:** The authors performed "one-study-removed" analyses to check whether any single study was driving the results. They also conducted subgroup analyses for:

Type of yoga (MBSR vs. non-MBSR)

Population type (healthy vs. clinical)

Intervention duration (short vs. long)

Method of data analysis

**What this design can and cannot prove:**

**Can prove:** That yoga asanas (with or without MBSR) cause changes in physiological stress markers compared to active control activities, because only RCTs were included.

**Cannot prove:** Which specific component of yoga (breathing, postures, meditation) is responsible for the effects. The interventions were heterogeneous — some included meditation, some didn't; some included breathing exercises, some didn't. The meta-analysis cannot tease apart these components.

**Cannot prove:** Long-term effects beyond the intervention period. Most studies measured outcomes immediately post-intervention, with few follow-up assessments.

**Cannot prove:** Clinical significance. Statistical significance does not always mean the changes are large enough to matter for health outcomes.

**Major methodological weaknesses:**

High heterogeneity across studies (I² values often >50%), meaning the studies were quite different from each other in design, population, and intervention

Many individual studies had small sample sizes

Blinding of participants was impossible, introducing potential expectancy effects

Publication bias was possible (though funnel plots suggested this was minimal for most outcomes)

Only English-language studies were included

**Primary outcomes (autonomic and HPA axis markers):**

**Evening cortisol:** Yoga reduced evening cortisol levels compared to active control (Hedges' g = -0.45, 95% CI: -0.75 to -0.15, p = 0.003, based on 12 studies). This is a small-to-moderate effect.

**Waking cortisol:** Yoga reduced waking cortisol levels (Hedges' g = -0.30, 95% CI: -0.56 to -0.04, p = 0.02, based on 9 studies). This is a small effect.

**Ambulatory systolic blood pressure:** Yoga reduced 24-hour ambulatory systolic blood pressure (Hedges' g = -0.43, 95% CI: -0.73 to -0.13, p = 0.005, based on 5 studies). Small-to-moderate effect.

**Resting heart rate:** Yoga reduced resting heart rate (Hedges' g = -0.34, 95% CI: -0.56 to -0.12, p = 0.002, based on 14 studies). Small effect.

**High-frequency heart rate variability (HF-HRV):** Yoga increased HF-HRV (Hedges' g = 0.47, 95% CI: 0.11 to 0.83, p = 0.01, based on 5 studies). Moderate effect. Higher HF-HRV indicates better parasympathetic (rest-and-digest) activity.

**Secondary outcomes (metabolic and inflammatory markers):**

**Fasting blood glucose:** Yoga reduced fasting blood glucose (Hedges' g = -0.55, 95% CI: -0.91 to -0.19, p = 0.003, based on 8 studies). Moderate effect.

**Total cholesterol:** Yoga reduced total cholesterol (Hedges' g = -0.47, 95% CI: -0.83 to -0.11, p = 0.01, based on 7 studies). Small-to-moderate effect.

**Low-density lipoprotein (LDL) cholesterol:** Yoga reduced LDL cholesterol (Hedges' g = -0.48, 95% CI: -0.86 to -0.10, p = 0.01, based on 6 studies). Small-to-moderate effect.

**Non-significant findings:**

Diastolic blood pressure: No significant difference (p > 0.05)

C-reactive protein (CRP): No significant difference (p > 0.05)

Interleukin-6 (IL-6): No significant difference (p > 0.05)

Mean arterial pressure: No significant difference (p > 0.05)

**Subgroup analyses:**

The effects were similar whether the intervention was MBSR-based or non-MBSR yoga

Effects were similar across healthy and clinical populations

Longer interventions (≥8 weeks) tended to show larger effects than shorter ones, but this was not statistically tested

To translate these numbers into something you can feel:

**Evening cortisol reduction (g = -0.45):** This is roughly equivalent to the difference in cortisol between someone who slept 8 hours versus someone who slept 6 hours. It's noticeable but not dramatic — think of it as the difference between a moderately stressful day and a calm day.

**Resting heart rate reduction (g = -0.34):** About 3–5 beats per minute lower on average. This is similar to the effect of 8 weeks of moderate aerobic exercise. For context, a resting heart rate of 72 vs. 68 bpm represents a meaningful shift toward better cardiovascular health over years.

**Systolic blood pressure reduction (g = -0.43):** About 4–6 mmHg reduction in ambulatory systolic BP. This is roughly half the effect of a standard blood pressure medication. A 5 mmHg reduction in systolic BP is associated with a ~10% reduction in stroke risk at the population level.

**Fasting blood glucose reduction (g = -0.55):** About 5–10 mg/dL reduction. For someone with prediabetes (fasting glucose 100–125 mg/dL), this could bring them into the normal range.

**LDL cholesterol reduction (g = -0.48):** About 10–15 mg/dL reduction. This is roughly equivalent to switching from a standard Western diet to a Mediterranean diet.

The effects are modest but clinically meaningful — they're not as large as pharmaceutical interventions, but they're comparable to other lifestyle interventions like moderate exercise or dietary changes.

**Acknowledged by the authors:**

High heterogeneity across studies — the interventions varied widely in style, duration, frequency, and content

Many individual studies had small sample sizes, reducing statistical power

Blinding of participants was impossible, introducing potential expectancy effects

Only English-language studies were included, which may introduce language bias

The meta-analysis could not determine which specific component of yoga (postures, breathing, meditation) drives the effects

Publication bias was possible, though funnel plots suggested it was minimal for most outcomes

**Critical reader observations:**

The "active control" groups varied enormously — some were walking, some were stretching, some were health education, some were usual care. This makes it hard to know what yoga is being compared to.

Many studies had short intervention durations (4–8 weeks), which may not be long enough for full physiological adaptation

Compliance/adherence data were often not reported — we don't know if participants actually did the yoga as prescribed

The meta-analysis combined studies measuring cortisol at different times of day (waking, evening, diurnal slope), which are physiologically distinct

The inflammatory marker analyses (CRP, IL-6) were underpowered — only 3–5 studies contributed to each analysis

Most studies were conducted in Western countries with predominantly white, middle-class participants, limiting generalisability

The meta-analysis did not control for baseline differences in stress levels — people who volunteer for yoga studies may already be more stressed than average

For someone running their own n=1 experiment:

### What to test (specific intervention and dose)

**Intervention:** A structured yoga practice that includes physical postures (asanas), controlled breathing (pranayama), and at least some meditative/relaxation component. Hatha yoga, Vinyasa flow, or Iyengar yoga are all reasonable choices. MBSR programs (which include yoga) are also effective.

**Dose:** Aim for **3–5 sessions per week, 30–60 minutes per session**, for a minimum of **8 weeks**. The meta-analysis found effects across studies lasting 8–16 weeks, with longer interventions tending to show larger effects.

**Specific protocol example:** 3x/week, 45-minute Hatha yoga class (including 10 minutes breathing exercises, 30 minutes postures, 5 minutes final relaxation) plus 10 minutes daily home practice.

### Minimum meaningful duration

**At least 8 weeks** to see changes in cortisol, heart rate, and blood pressure. Some studies showed effects at 4 weeks, but the meta-analysis suggests 8+ weeks is more reliable.

For metabolic markers (glucose, cholesterol), **12–16 weeks** may be needed.

### What to measure (specific metrics)

**Primary metric:** Resting heart rate — measure it first thing in the morning before getting out of bed, using a heart rate monitor or smartphone app. Take 5 morning readings over 1 week at baseline, then weekly during the intervention.

**Secondary metric:** Evening cortisol — this is harder to measure at home, but you can approximate it with a subjective stress rating (1–10 scale) taken at 8 PM each evening. Alternatively, some companies sell at-home salivary cortisol test kits.

**Tertiary metric:** Blood pressure — use an automated home blood pressure monitor. Measure at the same time each day (morning, after sitting quietly for 5 minutes). Take 3 readings and average them.

**Optional:** Fasting blood glucose and lipid panel — these require blood draws, but you can get them through your doctor or a direct-to-consumer lab service. Measure at baseline and after 12 weeks.

**Subjective measure:** Perceived Stress Scale (PSS-10) — a validated 10-item questionnaire available free online. Complete it at baseline and every 2 weeks.

### Key confounds to control for

**Time of day:** Always measure at the same time. Cortisol follows a diurnal rhythm (peaks ~30 min after waking, drops through the day). Heart rate and blood pressure also vary by time of day.

**Sleep:** Poor sleep elevates cortisol and heart rate. Track your sleep quality (duration and subjective quality) daily. If sleep changes during your experiment, it could confound results.

**Diet:** Caffeine, alcohol, and high-sugar meals can spike cortisol and heart rate. Try to keep your diet consistent during the experiment. Avoid caffeine for 2 hours before measurements.

**Exercise:** If you're already doing other exercise, keep it constant. Don't start a new running program while testing yoga — you won't know which intervention caused the changes.

**Menstrual cycle (if applicable):** Cortisol and heart rate vary across the menstrual cycle. If possible, start and end your experiment at the same phase of your cycle (e.g., both measurements in the follicular phase).

**Stressful life events:** Major stressors (job loss, relationship problems, illness) will swamp any yoga effect. If something major happens during your experiment, note it and consider extending the experiment.

**Expectation effects:** You know you're doing yoga, which can create a placebo effect. To partially control for this, include a "sham" control period (e.g., 4 weeks of gentle stretching without the breathing