Massage Evidence by Goal: Relaxation, Musculoskeletal Pain, Recovery and Mobility

How to Read This Page

Massage is hard to test fairly. It is difficult to keep participants and therapists unaware of who is receiving massage, which opens results to expectation effects, and a convincing sham massage is difficult to design. The Cochrane review of massage for low-back pain names this as the most common source of bias. [1] Most benefits below are short term and rated low in certainty. Evidence labels: promising but limited, mixed, insufficient evidence, no good evidence of benefit. See also understanding evidence in complementary health.

Relaxation

A 2004 meta-analysis of 37 randomized studies found that one massage reduced state anxiety, blood pressure and heart rate but not negative mood, immediate pain or cortisol. Multiple sessions reduced trait anxiety and depression, with benefits similar in size to psychotherapy in that analysis. [2] A review in pregnancy found moderate reductions in anxiety and depression compared with usual care, but no trial had a low risk of bias. [8] After surgery, a meta-analysis of 10 trials (1,157 participants) found less post-operative pain after a single session, with low methodological quality. [9] For a mix of other ways to unwind, see relaxation techniques and how to relieve stress.

Musculoskeletal Pain

Low-back pain. A Cochrane review of 25 trials (3,096 people) found massage better than inactive controls for pain and function in subacute and chronic low-back pain in the short term, but not long term; the authors had "very little confidence" that it is effective, and no serious adverse events were reported. [1] NICE in England says to consider manual therapy, including soft tissue techniques such as massage, only as part of a package that includes exercise. [10] The American College of Physicians 2017 guideline included massage for acute or subacute low-back pain but not for chronic pain. [11] A systematic review for that guideline, though, rated evidence for massage in chronic low-back pain as low to moderate strength, with small to moderate, usually short-term effects. [12]

Neck pain. The 2024 Cochrane review (33 studies) found low-certainty evidence that massage probably makes little to no difference to pain, function or quality of life compared with placebo for subacute to chronic neck pain at up to 12 weeks. A subgroup analysis suggested a clinically important difference with a higher dose: at least 8 sessions over four weeks of at least 30 minutes. [4] Most placebo-comparison trials used single sessions. [4]

Shoulder and knee. For neck and shoulder pain, a 2013 review found massage helped more than inactive therapies but not more than other active therapies, with short-term shoulder benefits only. [13] For knee osteoarthritis, 5 of 6 small studies reported short-term pain relief. [13]

Headache. Few studies, with inconsistent results; in a trial of trigger-point massage for tension headache, an inactive treatment worked as well. [13]

Exercise Recovery

A meta-analysis of 29 studies (1,012 participants) found no improvement in strength, jump, sprint, endurance or fatigue, but small improvements in flexibility and delayed onset muscle soreness. [6] Another of 11 studies found lower soreness ratings at 24, 48 and 72 hours and lower creatine kinase, a muscle-damage blood marker, versus no intervention. [5] In short, massage may ease soreness; it has not been shown to make you perform better.

Mobility

A meta-analysis of 7 studies (237 participants) found massage improved shoulder flexion and abduction (raising the arm forward and sideways). [7] We did not find a comparable review for other joints that we could read; proper evidence not found there.

Condition-Specific Research

• Fibromyalgia. Massage lasting at least 5 weeks improved pain, anxiety and depression, but not sleep disturbance (9 trials, 404 people). [14] NCCIH adds that Swedish massage may be an exception (2 studies, no benefit). [13]
• Cancer. Evidence for pain and anxiety relief is not strong; a 2016 evaluation of 19 studies found very low quality evidence. [15] Guidelines for breast cancer care include massage as one approach for stress, anxiety, depression, fatigue and quality of life. [13]
• Labor. A Cochrane review found low-quality evidence that massage reduced pain in the first stage of labor compared with usual care, and very low-quality evidence of no clear benefit on length of labor. [16]
• Premature infants. Most of 20 studies (1,250 infants) showed improved weight gain; no clear benefit was shown for healthy full-term infants. [13]
• HIV. Some evidence for anxiety, depression and quality of life, from small studies. [13]

What This Means for You

Massage is a reasonable choice for comfort and short-term relief if you enjoy it and it is safe for you. It is not proven to treat disease. For persistent back pain, NICE pairs manual therapy with exercise (above), and the NHS lists manual therapy, in which a trained therapist massages and moves muscles, bones and joints, among treatments if pain does not improve after a few weeks. [17] See massage safety before booking.