Heat eases sore muscles as well as ice does
Across 32 randomised trials in 1098 people, both cold and heat applied within an hour of exercise reduced the pain of delayed onset muscle soreness; cold helped mainly within 24 hours, heat helped both within and beyond 24 hours, hot packs showed the largest effects, and there was no significant difference between cold and heat overall.
Compiled by FitTools from the study cited below
The citation, figures and study details on this page are taken mechanically from the source record. No human editor has reviewed it.
Added to Pulse 15 August 2026
- Study design
- Meta-analysis
- Evidence
- well supported
- Published
- 15 August 2026
Key takeaway
What it shows: Trustworthy on direction, fuzzy on size: results pooled from 32 trials with 1098 people, but the trials varied widely in how heat and cold were applied and the authors themselves call for better studies. Head to head, neither therapy clearly beat the other.
Study details
- Design
- Meta-analysis
- Authors
- Wang Y, Li S, Zhang Y, Chen Y, Yan F, Han L, et al.
- Journal
- Phys Ther Sport
- Published
- 2021
- Added to Pulse
- 15 August 2026
Why it matters
Delayed onset muscle soreness, the ache that peaks a day or two after hard or unfamiliar exercise, is one of the most common reasons people skip the next session. Ice baths have become the fashionable answer, while heat is often dismissed as the old-fashioned option. Trials on both are scattered and small, and few people know whether either genuinely reduces soreness or when to apply them. This review pooled the randomised evidence on heat and cold to find out.
What they did
The authors followed a preregistered protocol and searched nine databases up to December 2020 for randomised controlled trials of heat or cold therapy for delayed onset muscle soreness. Thirty-two trials involving 1098 people were included and combined in a pooled analysis. Pain was assessed within 24 hours of exercise and beyond 24 hours. The analyses separated cold water immersion from other cold therapies, and hot packs from other forms of heat.
What they found
Cold therapy applied within an hour of exercise reduced pain within the first 24 hours, with cold water immersion showing a significant effect, but had no clear effect beyond 24 hours. Heat reduced pain both within and beyond 24 hours, and hot packs showed the largest effects of any approach in both time windows. Other heat therapies were not statistically significant. When cold and heat were compared directly, there was no significant difference between them.
Where it fits
This pooled analysis pushes back on the idea that cold is the only serious recovery tool: heat performed at least as well, and hot packs posted the biggest effects. It also sharpens the timing question, since benefits were tied to application within an hour of finishing exercise. The trials varied in methods and quality, and the authors call for higher-quality studies to settle whether heat or cold truly works better. The comparison covers soreness only, not muscle adaptation or performance.
What it means for you
If post-exercise soreness is what you want to blunt, this is evidence that both a hot pack and a cold soak can help, and that the humble hot pack is not the inferior choice. Timing appears to matter, with the benefit tied to applying either therapy within an hour of finishing. Cold's advantage faded after the first 24 hours, while heat kept helping beyond it. There is no basis here for treating ice as obligatory.
The source
Heat and cold therapy reduce pain in patients with delayed onset muscle soreness: A systematic review and meta-analysis of 32 randomized controlled trials. Phys Ther Sport 2021
DOI: 10.1016/j.ptsp.2021.01.004
Read the study →Related tools & guides
Related studies
- Cryotherapy chambers: the evidence still isn't therenot supported · human data
- Ice packs failed to speed recovery, whenever appliedpreliminary · human data
- Beetroot juice blunts muscle damage after hard exercisepreliminary · human data
- Cold plunges eased soreness, not strength recoverywell supported · human data
More research, with the reality check: every study we hold → · this week’s → · or open the full Pulse feed →
