Skip navigation

Massage for delayed-onset muscle soreness

Clinical bottom line: There is insufficient evidence to determine whether massage is effective in treating delayed-onset muscle soreness. The studies were of poor quality and validity.

Delayed-onset muscle soreness is a painful condition which often occurs after unaccustomed exercise. It is particularly bothersome in athletes. Massage may help by increasing local blood flow and decreasing swelling.

Systematic review:

Ernst E. Does post-exercise massage treatment reduce delayed onset muscle soreness? British Journal of Sports Medicine 1998;32:212-214

Date review completed: July 1997

Number of trials included: Seven

Number of patients: Total 132 patients

Control groups: no massage treatment or rest; placebo massage (i.e. no pressure).

Main outcomes: pain; soreness

Inclusion criteria were: controlled study, outcome measures to quantify delayed onset muscle soreness (DOMS), treatment group which included massage.

Medline, Embase and the Cochrane Library were searched (all years to July 1997) in addition to the reviewers own database and the experts' files of alternative therapies. No language restrictions were made. Data were extrated in a standardised, predefined manner. A descriptive analysis was conducted.


Seven trials were included. Four were randomised, three were not. The patient population was male athletes in one study (nine patients) and healthy/untrained volunteers in six (123 patients). Types of massage used were: vibrational massage with a mechanical device (one study), retrograde massage (one), unspecified massage (one), unspecified massage by a professional therapist (two), manual massage (two).

Randomised studies

One trial in nine healthy volunteers showed a statistical improvement in pain with daily massage treatment compared with no massage after 48 hours, but no difference at other observations up to 96 hours. The other studies showed massage to be more effective than control, but it was not stated if the difference was statistically significant.

Nonrandomised studies;

Two out of three studies reported massage to be more effective than control but, again, it was not stated whether these differences were statistically significant. The other reported no statistically significant difference between massage and control.

Adverse effects

Adverse effects were not mentioned.


The reviewer correctly stated that these studies were methodologically flawed. They were too small to be capable of detecting a true difference between massage and control treatment, they were all open to observer bias because of the lack of blinding, and the three studies which were not randomised were also open to selection bias. One study used a warm up and stretching period before exercise, followed by massage so it was unclear whether any improvement was due to the pre-exercise intervention or to the massage. This was not an appropriate study design to detect whether massage therapy was effective.

Further reading

Related topics