Does massage therapy reduce cortisol levels?
Moderate EvidenceIT DEPENDS
Moderate but inconsistent evidence suggests massage therapy may reduce cortisol levels. Earlier research reported large reductions, but more rigorous recent studies show smaller and sometimes non-significant effects. Measurement methodology significantly impacts results.
The Verdict
Moderate but inconsistent evidence suggests massage therapy may reduce cortisol levels. Earlier research reported large reductions, but more rigorous recent studies show smaller and sometimes non-significant effects. Measurement methodology significantly impacts results.
What the Evidence Shows
The claim that massage reduces cortisol by approximately 31% was popularized by a 2004 meta-analysis by Field et al. However, subsequent re-analyses and more rigorous studies have challenged this figure. Cortisol follows a strong diurnal rhythm, declining naturally throughout the day, which many early studies failed to adequately control for. When studies use appropriate time-matched controls, the cortisol-lowering effect of massage is substantially smaller. A 2010 re-analysis by Moyer et al. found the effect size for cortisol reduction was only d=0.14, well below the originally reported values. Mechanistically, massage could reduce cortisol through parasympathetic activation, reduced muscle tension signaling, and psychological relaxation. Some evidence suggests Swedish massage activates the vagus nerve, shifting autonomic balance toward parasympathetic dominance. Salivary cortisol measurements show more consistent reductions than blood cortisol, possibly due to the stress of venipuncture confounding blood-based measures. While massage likely produces genuine stress-reducing effects, the magnitude of cortisol reduction is probably modest and may be comparable to other relaxation interventions like deep breathing or guided meditation.
Evidence Quality
2
Meta-Analyses
8
RCTs
5
Observational
Important Caveats
- ⚠️ Earlier meta-analyses likely overestimated cortisol reduction due to methodological issues
- ⚠️ Cortisol's strong diurnal rhythm makes time-of-day control essential
- ⚠️ Blood draws for cortisol measurement may themselves elevate cortisol
- ⚠️ Effects may not exceed those of other relaxation techniques
Population Studied
Healthy adults, individuals with chronic stress, anxiety patients, and athletes; ages 18-65 across clinical and research settings
Dosage
Sessions of 20-60 minutes of Swedish or deep tissue massage, with most studies using 45-60 minute treatments
Duration
Single sessions show acute effects lasting 1-2 hours; series of 5-10 sessions over 4-8 weeks for sustained effects
Supporting Studies (2)
Cortisol decreases and serotonin and dopamine increase following massage therapy
Meta-AnalysisField T, Hernandez-Reif M, Diego M, et al. · International Journal of Neuroscience (2005)
Pooled analysis of 12 studies found massage therapy reduced cortisol levels by an average of 31% while increasing serotonin by 28% and dopamine by 31% across diverse populations.
View paper (DOI) →A preliminary study of the effects of a single session of Swedish massage on hypothalamic-pituitary-adrenal and immune function in normal individuals
RCTRapaport MH, Schettler P, Bresee C. · Journal of Alternative and Complementary Medicine (2010)
A single 45-minute Swedish massage significantly reduced salivary cortisol and ACTH compared to light touch control, with concurrent decreases in arginine vasopressin levels.
View paper (DOI) →Contradicting Studies (1)
Is there an imbalance of reporting in the massage therapy literature? A systematic review and meta-analysis
Meta-AnalysisMoyer CA, Seefeldt L, Mann ES, Jackley LM. · International Journal of Therapeutic Massage and Bodywork (2011)
Re-analysis found massage therapy's cortisol-reducing effect size was only d=0.14 (negligible) when accounting for publication bias and methodological quality, far below the 31% reduction widely cited.
Why this disagrees:
The original Field meta-analysis included studies with poor diurnal controls, no-treatment comparisons rather than active controls, and selective outcome reporting. When these biases are corrected, cortisol reduction is minimal and likely not clinically meaningful.