Muscle soreness has no well-evidenced fix, and its standard explanation is shaky

There is no treatment for delayed-onset muscle soreness with strong evidence behind it. The largest attempt to settle the question — an umbrella review published in Sports Medicine in 2025 that pooled 29 systematic reviews covering 863 unique randomised controlled trials and 24 distinct treatments — found that 17 of those 29 reviews were of critically low methodological quality, and only two were rated high [s1]. The mechanism is less settled than the popular explanation suggests as well: in animal experiments, soreness developed under conditions in which muscle damage was largely absent [s2].

What the umbrella review found

The authors searched MEDLINE, Embase, the Cochrane Database of Systematic Reviews, CINAHL, PEDro and Epistemonikos from 1998 to February 2024 for systematic reviews of randomised trials of post-exercise treatments used by physiotherapists in healthy adults [s1]. They graded methodological quality with AMSTAR-2 and classified each treatment's evidence from Class I to Class V, based on the number of cases, effect size, p-value, heterogeneity, Egger's test and tests for excess significance [s1].

Nothing reached Class I. The strongest result at any time point was contrast therapy — alternating hot and cold — for pain immediately after exercise, at Class II [s1]. At 24 hours, massage therapy reached Class III, with cooling therapy, contrast therapy, electrical stimulation, cryostimulation, phototherapy and heat therapy at Class IV [s1]. At 48 hours, compression, contrast therapy, kinesiotaping and cryostimulation reached Class III [s1]. At 72 hours, kinesiotaping reached Class III [s1]. At 96 hours, only compression, phototherapy and contrast therapy showed significant effects, all Class IV [s1].

The authors state that evidence for stretching, exercise and electrical stimulation is weak [s1].

A discrepancy in the source. The review's abstract reports effect sizes ranging from a Hedges' g of 0.36 for cooling therapy to 1.82 for heat therapy — but gives both figures the identical 95% confidence interval of 0.46 to 3.18 [s1]. Two treatments with effect sizes five times apart cannot share an interval. One of the two intervals is a transcription error, and the abstract does not indicate which. The direction of the finding is unaffected; the precision attached to either number should not be relied on.

The mechanism is not what most people are told

The standard account of delayed-onset soreness is microscopic tearing of muscle fibres from unaccustomed lengthening contractions. A 2024 review in The Journal of Physiological Sciences, drawing mainly on rat experiments, complicates that [s2]. The muscular mechanical withdrawal threshold — the pressure at which an animal registers muscle pain — fell one to three days after lengthening contractions, matching the human time course [s2]. But when the researchers varied the speed and range of the stretch, muscle injury seldom occurred except under extreme conditions, and soreness appeared under parameters that produced no muscle damage [s2].

What the review identifies instead are two neurochemical signalling routes: a B2 bradykinin receptor to nerve growth factor pathway, and a COX-2 to glial cell line-derived neurotrophic factor pathway, interacting at two levels [s2]. The repeated-bout effect — the familiar observation that the second time you do an unfamiliar exercise it hurts far less — was seen both in the withdrawal threshold and in nerve growth factor upregulation, with adaptation appearing to occur upstream of B2 bradykinin receptor activation [s2].

This is animal work, and the review says so; some human research is included but most of the mechanistic detail comes from rats [s2]. It does not establish that human soreness involves no muscle damage. It does establish that damage and soreness can be dissociated, which is enough to undercut the idea that how sore you are measures how much you damaged the muscle.

Stretching, specifically, does not help

The most common self-treatment has been tested directly. A Cochrane review of 12 randomised or quasi-randomised trials found that pre-exercise stretching reduced soreness a day later by an average of half a point on a 100-point scale, and post-exercise stretching by about one point — both with confidence intervals spanning zero in either direction [s3]. A single large field trial of 2,377 participants found that stretching before and after exercise reduced peak soreness over a week by an average of four points on the same scale, a statistically significant but, in the reviewers' own word, very small effect [s3]. Their conclusion was that stretching does not produce clinically important reductions in delayed-onset muscle soreness in healthy adults [s3].

Where that leaves the question

Two things are reasonably clear. Soreness is transient and self-limiting, and it adapts fast: the repeated-bout effect means the same session produces markedly less soreness the second time [s2]. And the interventions marketed for it — cold, heat, compression, massage, tape, light — sit in a literature whose own umbrella review describes it as predominantly low quality, with effects that appear at some time points and not others and rarely survive strict grading [s1].

That is not the same as saying none of them work. It is saying that the evidence base cannot tell you which ones do, by how much, or for whom. A treatment supported by Class III or Class IV evidence is one where the finding is real enough to be worth further study and too fragile to build a recommendation on.

This article is informational and is not medical advice. Muscle pain that is severe, that comes with dark urine, or that does not resolve within a few days is not ordinary soreness and warrants medical assessment.

Sources

• Physical Therapies for Delayed-Onset Muscle Soreness: An Umbrella and Mapping Systematic Review with Meta-meta-analysis — Sports Medicine, 2025-03-22
• Neurochemical mechanism of muscular pain: Insight from the study on delayed onset muscle soreness — The Journal of Physiological Sciences, 2024-01-24
• Stretching to prevent or reduce muscle soreness after exercise — Cochrane Database of Systematic Reviews, 2011-07-06