A body-by-body review of the clinical literature behind cold and heat therapy from post-surgical knees to chemotherapy-induced neuropathy including the exact temperatures and timing protocols the studies used, and where the evidence still hasn't caught up.
PurHeart · Clinical Evidence Reference
Cold & heat therapy, by body region
What the published literature supports for cryotherapy and thermotherapy — including target temperature and device timing — across chronic pain, acute injury, and the peri‑operative window. Organized by the same clinical departments as the hospital capability statement.
Compiled 27 Aug 2026
24 sources reviewed
Research synthesis — not regulatory‑cleared claims language
Cold — general range
0–15°C / 32–59°F
Static packs run 0–5°C; regulated cold‑flow devices run warmer, 10–15°C, for longer safe dwell time.
Heat — general range
40–45°C / 104–113°F
Vasodilation begins ~40°C; therapeutic effect plateaus by 42–45°C. Above this range, burn risk rises sharply.
Universal skin‑safety rule
Never below 2°C / 35°F on bare skin. Always a cloth barrier. Check skin every 15–20 min.
Evidence strength
Strong — Cochrane / multiple RCTs
Moderate — RCT(s), smaller base
Limited — consensus / case reports
None found in this review
| Region / Application | Modality | Temperature | Timing & duration | Evidence | Key sources | Note |
|---|---|---|---|---|---|---|
| Oncology | ||||||
| Hands & feet (peripheral nerves) | Cold | −10°C to −25°C across trials; most cluster ~−20°C to −24°CTrial protocols | On 15 min before infusion → worn through infusion → off 15 min after. Total wear 60–90 min; first 60 min carries most of the cooling effect. | Moderate | Frozen gloves meta‑analysis Cryotherapy & CIPN prevention | Frozen gloves/socks for chemo‑induced peripheral neuropathy, mainly taxane regimens in breast cancer. Temperature varies by device — no single standardized setpoint yet. |
| Scalp | Cold | Coolant ≈ −5°C; target scalp surface 18–22°C (64–72°F)Device spec | Pre‑cool 30 min (up to 45 min for thick/coarse hair) → continuous through infusion → post‑cool 20–90 min, varies by drug (e.g. weekly Taxol: 60 min post; docetaxel: 20 min post). | Strong | Paxman system specs COHAIR trial | Only application in this map with FDA‑cleared devices and a Medicare coverage determination — and the most precisely specified temperature protocol of any row. |
| Oral / Maxillofacial – ENT | ||||||
| Jaw / perioral (third‑molar extraction) | Cold | Standard ice/gel pack, 0–5°C, cloth‑wrappedGeneral post‑op norm | 15–20 min on / 15–20 min off, repeated through waking hours for the first 24–48 h; benefit concentrated in that window per the cryotherapy meta‑analysis. | Strong | RCT, third‑molar surgery Meta‑analysis, inflammatory markers | Trials support the practice reliably; the specific minute‑by‑minute cadence comes from standard post‑op instruction sheets rather than a single dosing RCT. |
| Ophthalmology | ||||||
| Periorbital / eye (cataract surgery) | Cold | Not specified in source | Short, intermittent applications per surgeon direction; source did not report a standardized minute count. | Moderate | Am. J. Ophthalmology | Direct evidence for reduced inflammation/comfort, but temperature and dwell time weren't reported precisely. Warm compress is for a separate indication (meibomian gland). |
| Pediatrics | ||||||
| General | Cold / heat | No dedicated data found | No dedicated data found | None found | — | No site‑specific pediatric trial surfaced. Likely extrapolated from adult protocols at reduced intensity. |
| Aesthetic / Cosmetic Surgery | ||||||
| Face & trunk (facelift, liposuction) | Cold | General cold pack, 0–5°CSurgeon guidance | 15–20 min on/off, first 24–48 h — surgeon patient‑guidance pattern, not RCT‑derived. | Limited | — | Weakest evidence tier in this map. Use “commonly recommended,” not “clinically proven,” here. |
| Reconstructive Surgery | ||||||
| Perineum (vaginal delivery) | Cold | Not specified across trials | 10–20 min per application, within the first 2 days postpartum.Cochrane review | Strong | Cochrane CD006304 | Longest‑running evidence trail in this map (since the late 1990s). Timing is consistent across trials; temperature was never standardized or reported. |
| Abdomen (cesarean incision) | Cold | Not detailed in source | Not detailed in source | Moderate | Suwannalert 2021 RCT | Only one dedicated RCT identified; full protocol would need pulling from the paper directly if this becomes claims‑relevant. |
| Breast (mastectomy / BCS) | Cold | General cold pack, 0–5°CPost‑op guidance | ~15–20 min on/off, several times daily for the first 48–72 h. Avoid direct skin contact — post‑surgical numbness masks cold‑injury warning signs. | Moderate | Frozen gel packs, axillary surgery Cold therapy, BCS pain | Trials confirm reduced NSAID/narcotic use; the numbness caveat is a genuine safety note worth keeping in patient‑facing copy. |
| Orthopedic | ||||||
| Knee (total knee arthroplasty) | Cold | Device/tissue target 10–15°C (50–59°F); never below 2°C (35°F) on skinDevice protocol | Days 1–3: 20–30 min every 2 h. Days 4–14: 3–5×/day, 15–20 min, especially around PT sessions. Weeks 3–6: as‑needed. | Strong | Cochrane CD007911 (2025) Temp‑controlled cold‑flow RCT | Best‑studied application in this entire map — four rounds of Cochrane review, plus dedicated temperature‑controlled device trials. |
| Shoulder (arthroscopy / rotator cuff) | Cold | Device‑regulated, ~10–15°C rangeConsistent with ortho norm | Continuous or cycled per device instructions through the acute post‑op days; exact minute‑by‑minute cadence isn't uniformly reported across trials. | Strong | RCT, cold vs. cold compression Cold unit & narcotic use | Continuous cold units reduce narcotic use vs. static ice packs; the “continuous” element seems to matter more than any single exact temperature. |
| Low back & neck (spine) | Heat favored / cold | Cold: 0–5°C. Heat: ~40°C (104°F) sustained low‑level heatNadler RCTs | Cold: 15–20 min, up to 2–3×/day, as needed. Heat: continuous low‑level heatwrap for up to 8 h — shown superior to ibuprofen/acetaminophen for acute low back pain. | Strong | Cochrane CD004750 Nadler heatwrap RCT | The one region where heat and cold aren't interchangeable, and where heat's protocol (long, low, continuous) is the opposite shape of a typical cold protocol (short, cold, cyclic). |
| Neurosurgery | ||||||
| General | Cold / heat | No dedicated data found | No dedicated data found | None found | — | Same flag as pediatrics — no site‑specific trial surfaced in this review. |
Pattern worth noting: evidence strength tracks inversely with distance from bone/joint surgery. Orthopedic and oral surgery carry the most rigorous RCT/Cochrane support and the most precisely specified temperatures; oncology's two applications (CIPN, scalp) have exact device specs because they're built into cleared hardware; cosmetic surgery has neither a trial base nor a standardized temperature — it runs on surgeon convention.
Safety note baked into nearly every protocol above: direct skin contact below ~2°C (35°F) risks frostbite/nerve injury, especially where post‑surgical numbness masks the warning signs (breast, jaw, perineum). Every credible protocol pairs cold with a cloth or gel‑pack barrier and a scheduled off‑cycle.
PurHeart · Clinical Evidence Reference
Cold & heat therapy, by body region
What the published literature supports for cryotherapy and thermotherapy — including target temperature and device timing — across chronic pain, acute injury, and the peri‑operative window. Organized by the same clinical departments as the hospital capability statement.
Compiled 27 Aug 2026
24 sources reviewed
Research synthesis — not regulatory‑cleared claims language
Cold — general range
0–15°C / 32–59°F
Static packs run 0–5°C; regulated cold‑flow devices run warmer, 10–15°C, for longer safe dwell time.
Heat — general range
40–45°C / 104–113°F
Vasodilation begins ~40°C; therapeutic effect plateaus by 42–45°C. Above this range, burn risk rises sharply.
Universal skin‑safety rule
Never below 2°C / 35°F on bare skin. Always a cloth barrier. Check skin every 15–20 min.
Evidence strength
Strong — Cochrane / multiple RCTs
Moderate — RCT(s), smaller base
Limited — consensus / case reports
None found in this review
| Region / Application | Modality | Temperature | Timing & duration | Evidence | Key sources | Note |
|---|---|---|---|---|---|---|
| Oncology | ||||||
| Hands & feet (peripheral nerves) | Cold | −10°C to −25°C across trials; most cluster ~−20°C to −24°CTrial protocols | On 15 min before infusion → worn through infusion → off 15 min after. Total wear 60–90 min; first 60 min carries most of the cooling effect. | Moderate | Frozen gloves meta‑analysis Cryotherapy & CIPN prevention | Frozen gloves/socks for chemo‑induced peripheral neuropathy, mainly taxane regimens in breast cancer. Temperature varies by device — no single standardized setpoint yet. |
| Scalp | Cold | Coolant ≈ −5°C; target scalp surface 18–22°C (64–72°F)Device spec | Pre‑cool 30 min (up to 45 min for thick/coarse hair) → continuous through infusion → post‑cool 20–90 min, varies by drug (e.g. weekly Taxol: 60 min post; docetaxel: 20 min post). | Strong | Paxman system specs COHAIR trial | Only application in this map with FDA‑cleared devices and a Medicare coverage determination — and the most precisely specified temperature protocol of any row. |
| Oral / Maxillofacial – ENT | ||||||
| Jaw / perioral (third‑molar extraction) | Cold | Standard ice/gel pack, 0–5°C, cloth‑wrappedGeneral post‑op norm | 15–20 min on / 15–20 min off, repeated through waking hours for the first 24–48 h; benefit concentrated in that window per the cryotherapy meta‑analysis. | Strong | RCT, third‑molar surgery Meta‑analysis, inflammatory markers | Trials support the practice reliably; the specific minute‑by‑minute cadence comes from standard post‑op instruction sheets rather than a single dosing RCT. |
| Ophthalmology | ||||||
| Periorbital / eye (cataract surgery) | Cold | Not specified in source | Short, intermittent applications per surgeon direction; source did not report a standardized minute count. | Moderate | Am. J. Ophthalmology | Direct evidence for reduced inflammation/comfort, but temperature and dwell time weren't reported precisely. Warm compress is for a separate indication (meibomian gland). |
| Pediatrics | ||||||
| General | Cold / heat | No dedicated data found | No dedicated data found | None found | — | No site‑specific pediatric trial surfaced. Likely extrapolated from adult protocols at reduced intensity. |
| Aesthetic / Cosmetic Surgery | ||||||
| Face & trunk (facelift, liposuction) | Cold | General cold pack, 0–5°CSurgeon guidance | 15–20 min on/off, first 24–48 h — surgeon patient‑guidance pattern, not RCT‑derived. | Limited | — | Weakest evidence tier in this map. Use “commonly recommended,” not “clinically proven,” here. |
| Reconstructive Surgery | ||||||
| Perineum (vaginal delivery) | Cold | Not specified across trials | 10–20 min per application, within the first 2 days postpartum.Cochrane review | Strong | Cochrane CD006304 | Longest‑running evidence trail in this map (since the late 1990s). Timing is consistent across trials; temperature was never standardized or reported. |
| Abdomen (cesarean incision) | Cold | Not detailed in source | Not detailed in source | Moderate | Suwannalert 2021 RCT | Only one dedicated RCT identified; full protocol would need pulling from the paper directly if this becomes claims‑relevant. |
| Breast (mastectomy / BCS) | Cold | General cold pack, 0–5°CPost‑op guidance | ~15–20 min on/off, several times daily for the first 48–72 h. Avoid direct skin contact — post‑surgical numbness masks cold‑injury warning signs. | Moderate | Frozen gel packs, axillary surgery Cold therapy, BCS pain | Trials confirm reduced NSAID/narcotic use; the numbness caveat is a genuine safety note worth keeping in patient‑facing copy. |
| Orthopedic | ||||||
| Knee (total knee arthroplasty) | Cold | Device/tissue target 10–15°C (50–59°F); never below 2°C (35°F) on skinDevice protocol | Days 1–3: 20–30 min every 2 h. Days 4–14: 3–5×/day, 15–20 min, especially around PT sessions. Weeks 3–6: as‑needed. | Strong | Cochrane CD007911 (2025) Temp‑controlled cold‑flow RCT | Best‑studied application in this entire map — four rounds of Cochrane review, plus dedicated temperature‑controlled device trials. |
| Shoulder (arthroscopy / rotator cuff) | Cold | Device‑regulated, ~10–15°C rangeConsistent with ortho norm | Continuous or cycled per device instructions through the acute post‑op days; exact minute‑by‑minute cadence isn't uniformly reported across trials. | Strong | RCT, cold vs. cold compression Cold unit & narcotic use | Continuous cold units reduce narcotic use vs. static ice packs; the “continuous” element seems to matter more than any single exact temperature. |
| Low back & neck (spine) | Heat favored / cold | Cold: 0–5°C. Heat: ~40°C (104°F) sustained low‑level heatNadler RCTs | Cold: 15–20 min, up to 2–3×/day, as needed. Heat: continuous low‑level heatwrap for up to 8 h — shown superior to ibuprofen/acetaminophen for acute low back pain. | Strong | Cochrane CD004750 Nadler heatwrap RCT | The one region where heat and cold aren't interchangeable, and where heat's protocol (long, low, continuous) is the opposite shape of a typical cold protocol (short, cold, cyclic). |
| Neurosurgery | ||||||
| General | Cold / heat | No dedicated data found | No dedicated data found | None found | — | Same flag as pediatrics — no site‑specific trial surfaced in this review. |
Pattern worth noting: evidence strength tracks inversely with distance from bone/joint surgery. Orthopedic and oral surgery carry the most rigorous RCT/Cochrane support and the most precisely specified temperatures; oncology's two applications (CIPN, scalp) have exact device specs because they're built into cleared hardware; cosmetic surgery has neither a trial base nor a standardized temperature — it runs on surgeon convention.
Safety note baked into nearly every protocol above: direct skin contact below ~2°C (35°F) risks frostbite/nerve injury, especially where post‑surgical numbness masks the warning signs (breast, jaw, perineum). Every credible protocol pairs cold with a cloth or gel‑pack barrier and a scheduled off‑cycle.