Heat or Cold for DOMS? What the Best Evidence Actually Says

For soreness after a hard workout, cold-water immersion is the safer evidence-based default. Heat packs produced the largest pain-relief estimate in a 32-trial meta-analysis, but that estimate was so imprecise that it could reflect anything from a modest benefit to an implausibly huge one. A larger network analysis ranked heat first through 48 hours and extreme cryotherapy first after 48 hours, yet those rankings rely partly on indirect comparisons.

The useful answer is therefore conditional: use cold water when you want a protocol backed by a more precise estimate; consider heat for pain relief in the first two days if it feels good and you accept shakier evidence; and do not mistake less soreness for restored athletic performance.

The ranking changes with the clock

A 2022 network meta-analysis by Wang et al. compared 10 temperature-based approaches across 59 studies: cold-water immersion (CWI), phase-change-material garments, contrast water therapy, hot or warm-water immersion, novel cryotherapy, cold packs, ice massage, hot packs, ultrasound, and passive recovery. Network meta-analysis can compare treatments that were not all tested against one another in the same trial, which expands the picture but adds assumptions.

Hot packs ranked first for pain relief at up to 24 hours and up to 48 hours. After 48 hours, "novel cryotherapy" ranked first. That category meant whole-body, partial-body, or air-pulsed cryotherapy using dry air from −30°C to below −110°C. Contrast water therapy, defined here as cold water at 10°C or less alternating with water at 35–40°C, ranked second through the early windows.

Those rankings do not establish that a hot pack will beat an ice bath for every person after every workout. The published material does not provide the full effect-size and uncertainty table for every modality and time window, so adjacent ranks cannot be independently judged for size or precision. Ranking first can mean "best estimate in this network," not "settled winner."

The striking heat result has a precision problem

A 2021 systematic review and meta-analysis by Yutan Wang et al. pooled 32 randomized controlled trials with 1,098 participants. It found that hot packs had the largest numerical effect on DOMS pain at or before 24 hours: standardized mean difference (SMD) −2.31 versus control, with a 95% confidence interval of −4.33 to −0.29 (P = 0.03). Beyond 24 hours, hot packs remained statistically significant at SMD −1.78, 95% CI −2.97 to −0.59 (P = 0.003).

Those are eye-catching numbers. They are also uncertain. The early hot-pack confidence interval spans 4.04 SMD units. Its upper end still favors heat, but only modestly; its lower end implies an effect far larger than most recovery interventions produce. A wide interval usually points to few studies, small samples, inconsistent results, or all three. It does not make heat useless. It means the headline estimate is a poor basis for pretending we know its exact size.

CWI had a smaller but tighter early estimate: SMD −0.48, 95% CI −0.84 to −0.13 (P = 0.008). Across all cold approaches, the early pooled estimate was SMD −0.57, 95% CI −0.89 to −0.25; across heat approaches, it was SMD −1.17, 95% CI −2.62 to −0.09. Yet the direct comparison between the broad heat and cold categories was not statistically significant (P = 0.16). The data support pain relief from both categories; they do not prove that heat is categorically superior.

A large point estimate is not automatically a reliable estimate. For choosing a routine, CWI's smaller, more precise effect is easier to trust than the hot-pack result.

What a practical CWI protocol looks like

A 2025 dose-response network meta-analysis identified 5–10°C as the most effective CWI temperature range for reducing DOMS pain and creatine kinase after acute exercise-induced muscle damage. Creatine kinase is a muscle-damage biomarker, not a direct measure of how well you can train tomorrow, but the pain finding gives the temperature range practical value.

The 2022 network analysis had defined CWI as water at 15°C or below within one hour of exercise. The later dose-response analysis refines that ceiling: "below 15°C" is not a target in itself. Based on the available evidence, 5–10°C is the more defensible range for soreness relief.

Cold also has a more recently tested evidence base. A 2024 network meta-analysis of hydrotherapy and cryotherapy, a 2023 review comparing CWI with other recovery methods, and a 2026 review of different immersion regions all evaluated cold-based recovery. The last analysis included 30 RCTs and found lower DOMS with CWI than passive control (g = −0.40, 95% CI −0.64 to −0.16). These reviews do not verify hot-pack superiority because they did not include hot-pack arms.

Soreness and performance are different endpoints

Less pain can make recovery feel better without restoring strength, jump performance, or sprint capacity. In a 2022 randomized trial by Angelopoulos et al., 60 male amateur athletes were assigned after plyometric training to CWI, sports massage, both treatments, or control. CWI improved pain compared with the combined-treatment and control groups on days two and three, but the study found no significant group differences in most tested functional measures.

That distinction matters most when an athlete plans to compete or train hard again. Use pain relief as pain relief. Do not use it as proof that muscle function has recovered.

The placebo issue limits every thermal claim

Heat and cold are hard to blind: participants can plainly feel which one they received. That matters because DOMS studies commonly use subjective visual-analog or Likert pain ratings. In a 2023 randomized, counterbalanced crossover trial, 12 semi-professional soccer players completed a shuttle-test protocol and then received CWI for 15 minutes at 11°C, a placebo recovery drink, or passive rest on separate weeks. The placebo condition produced partial recovery, showing that expectation can influence recovery outcomes.

This does not mean CWI is "all placebo," and it does not erase the meta-analytic findings. It does mean that subjective pain estimates probably combine physiological effects with expectation. That concern applies at least as strongly to comforting heat packs, whose largest estimate came from unblinded trials.

A decision tree that matches the evidence

• Within 24 hours: Choose CWI at 5–10°C if you want the most reproducible option. Its early pain estimate was statistically significant and much more precise than the hot-pack estimate.
• Between 24 and 48 hours: Heat is reasonable when pain relief is the goal. Hot packs retained the top network rank, and their pooled estimate beyond 24 hours was statistically significant. Treat that as a plausible option, not a proven champion.
• After 48 hours: Extreme dry-air cryotherapy ranked first in the 2022 network analysis. For most people, chamber access and cost make this a less practical recommendation than CWI.
• Do not rely on stretching for DOMS: A 2021 meta-analysis of randomized trials found that post-exercise stretching did not significantly reduce DOMS in short-term or delayed recovery windows.

The evidence does not justify crowning either a heat pack or an ice bath as the universal recovery tool. CWI has the more dependable early estimate and a specific temperature range. Heat may reduce pain more, especially in the first 48 hours, but its apparent advantage rests on sparse, unblinded, imprecise evidence. If a hot pack helps you move comfortably, use it. If you want the least speculative default after hard training, use cold water at 5–10°C and keep your expectations proportional to an SMD of about half a standard deviation.