Wang · Journal of rehabilitation medicine 2022 · network meta-analysis of randomized controlled trials · n=59 studies (1,367 participants)

Effect of cold and heat therapies on pain relief in patients with delayed onset muscle soreness: A network meta-analysis.

Cited 38 times in the scientific literature.

Level 1 - systematic review of randomized trials

Systematic review and network meta-analysis of randomized controlled trials

PubMed 34636405 · doi:10.2340/jrm.v53.331 · record verified 2026-08-30

What was done

Eight Chinese and English databases were searched up to May 31, 2021, to identify randomized controlled trials evaluating cold and heat therapies for delayed onset muscle soreness (DOMS). Risk of bias was evaluated using the Cochrane risk-of-bias tool. A network meta-analysis was conducted across 59 studies (1,367 participants) comparing 10 interventions: hot pack, cold pack, contrast water therapy, phase change material, novel modality cryotherapy, cold-water immersion, hot/warm-water immersion, ice massage, ultrasound, and passive recovery.

What was found

The abstract reports comparative rankings without effect estimates or confidence intervals. For pain relief within 24 hours post-exercise, hot pack ranked most effective, followed by contrast water therapy. Within 48 hours, hot pack remained the highest-ranked intervention, followed by novel cryotherapy modalities. Beyond 48 hours post-exercise, novel cryotherapy modalities ranked first for pain relief.

Why it matters

This review provides a comparative ranking of common thermal modalities across different time windows, suggesting that heat application may be preferable in early post-exercise phases (0–48 hours) while advanced cryotherapy modalities may be more effective for delayed recovery (>48 hours).

Limits

The abstract notes that the underlying evidence base had limited methodological quality. Numerical effect sizes, credible intervals, and safety/adverse event outcomes are not reported in the abstract. Blinding of participants is inherently unfeasible for physical thermal modalities, presenting a high risk of performance bias.

Cited by