Cold Exposure

A reliable acute stressor with mood and habituation benefits — and several popular claims that don't survive scrutiny. There is no longevity outcome data for cold water immersion in humans, and many "metabolic" claims confuse it with mild cold-air acclimation, which is a different intervention.

Cold water immersion has been one of the most enthusiastically marketed interventions of the past decade. The honest evidence summary: reliable acute physiology, modest hormetic adaptations, real mood effects, and no longevity outcome data. Several common claims (chronic anti-inflammatory effects, "ice baths burn hundreds of calories", brief plunges replicate the metabolic benefits of mild cold acclimation) are not supported by careful reading of the literature.

What "cold" means

Tipton's group operationalized "cold" as water below 15°C — the threshold at which cold-shock physiology peaks.[1] Water conducts heat roughly 24× faster than air, which is why immersion at 14°C is a fundamentally different stimulus than standing in 14°C air.

Dose-response is steep. The foundational dataset comes from Šrámek's controlled 1-hour 14°C head-out immersions[2]:

  • 14°C: noradrenaline +530%, dopamine +250%, metabolic rate +350%, diuresis +163% (central blood pooling suppresses renin/aldosterone). Note that adrenaline (epinephrine) was unchanged — only noradrenaline and dopamine rose, which complicates the popular "adrenaline rush" framing.
  • 20°C: substantially smaller catecholamine response
  • 32°C: essentially no acute response

This dose-response anchors essentially every "cold = alertness/mood" claim in the field.

Whole-body cryotherapy (WBC) is a different modality. The cold-air chambers (−110°C to −195°C, 2–4 min) increasingly sold for recovery drop skin temperature fast but, because air conducts heat so much more slowly than water, do not achieve the deep-tissue or core cooling of immersion. Its evidence base is thinner still: the dedicated Cochrane review pooled just four laboratory trials in 64 physically active adults (mean age 23 — all but four of them men) and found insufficient evidence to determine whether WBC reduces muscle soreness or improves subjective recovery compared with rest or no treatment. Only the one-hour timepoint separated from control, a moderate-to-large drop in self-reported soreness (standardised mean difference −0.77) resting on 20 participants in two crossover trials, while at 24, 48 and 72 hours the confidence intervals straddled zero. All four trials carried a high risk of bias and every outcome was graded very-low-certainty; the only active comparator ever tested was far-infrared therapy, in a single crossover trial of nine runners, so no trial has compared WBC with cold water immersion; and none ran active surveillance for adverse events, which the authors flagged as a real gap given the temperatures involved.[3] Evidence rating: Weak / preliminary. Where this page says "cold immersion," it means water unless stated otherwise.

What cold immersion reliably does

Evidence rating: Strong for the acute physiology (autonomic surge, response habituation); Moderate for the post-exit vagal rebound (consistent in direction, but only half the trials reached significance); Moderate for the acute mood lift and for the chronic antioxidant upregulation.

Acute autonomic activation

  • Sympathetic surge: catecholamine release in the first minute
  • Vagal rebound: parasympathetic dominance after exit — heart-rate variability (HRV) shifts toward its vagal-dominant, high-frequency pattern. The direction is consistent — all twelve randomised trials in a 2025 systematic review reported parasympathetic reactivation after post-exercise immersion — but only half reached statistical significance against passive recovery, and no pooled effect estimate was calculated.[4]
  • Subjective alertness, mood elevation
  • Delayed stress-marker reduction at about 12 hours post-exposure, not immediately — on thinner evidence than it first appears. Pooled stress markers show no significant change at 0, 1, 24 or 48 hours and a large reduction at 12 hours (standardised mean difference −1.00), but every one of those timepoints rests on the same four arms from three papers by a single laboratory, all in young healthy men immersed at about 14°C.[5] The reviewers themselves do not read the 12-hour dip as a hypothalamic–pituitary–adrenal (HPA) shutdown; they describe a gradual handover from sympathetic to parasympathetic control mediated by baroreceptor feedback, and call the adrenal contribution more nuanced than previously assumed. Treat it as a single-laboratory finding awaiting replication rather than a general property of cold water.

Habituation (with repeated exposure)

  • Cold-shock response (heart rate, gasping reflex, hyperventilation) reduces measurably within 4 immersions — confirmed in per-variable meta-analyses (heart rate −14 bpm across 17 groups, respiratory frequency −8 breaths/min across 12, minute ventilation −21 L/min across 10)[6]
  • Persists 7–14 months after as few as six 3-minute immersions

Antioxidant defenses

  • Winter swimmers had higher baseline glutathione and red-cell superoxide dismutase and catalase than people who had never winter-swum (cross-sectional)[7]
  • In 28 men over one season of weekly 3-minute lake dips (before/after, no control group), most antioxidant defences rose and most oxidative-damage markers fell, at baseline and after a dip[8]
  • Textbook hormesis — chronic mild stress upregulates protective machinery

Acute mood improvement

  • A single 18-minute head-out immersion at 13.6°C dropped Profile of Mood States Total Mood Disturbance by 15 points.[9]
  • fMRI changes in default-mode and salience brain networks were detectable after a 5-min 20°C immersion.[10]
  • A 3,018-person RCT of cold-finishing showers cut self-reported sickness absence by 29% (but not actual illness days).[11]

Inflammation: no evidence cold lowers it

Evidence rating: Moderate that cold does not lower inflammation; Weak for any acute rise. The most rigorous recent synthesis — a 2025 meta-analysis of 11 randomised cold-water immersion (CWI) trials — reported inflammatory markers rising rather than falling after immersion, by a large standardised margin immediately afterwards and at one hour (standardised mean differences of 1.03 and 1.26, where a value near or above 1.0 counts as a large effect).[12] Read that number before repeating it. Only two of the eleven trials fed the inflammation estimate — 21 young men between them, both measuring cytokines after a bout of resistance exercise — and the effect is a within-group change from pre-immersion baseline, not a comparison against a control condition. One of those two trials, comparing conditions head to head, found the reverse: cold immersion held the exercise-induced interleukin-6 rise to about a tenth of the control response (a 5% rise against 48%) and cut the tumour-necrosis-factor-alpha rise by roughly two thirds.[13] The review's own authors add that an immediate rise may reflect metabolic responses such as glycogen breakdown rather than systemic inflammation. The defensible reading is that nobody has shown cold immersion lowers inflammation in healthy adults — not that it reliably raises it.

Where an acute rise does appear, it is consistent with muscle and adipose interleukin-6 (IL-6) release during shivering acting as a myokine signal.

A meta-analysis of post-exercise cooling found no meaningful reduction in interleukin-6 (IL-6) at 24, 48 or 72 hours of recovery, and none in C-reactive protein (CRP) at 24 hours. CRP was lower at 48 hours, but the authors discount that result themselves, given the heterogeneity and risk of bias behind it.[14]

The mainstream claim that CWI lowers chronic systemic inflammation is not supported by RCT evidence in healthy adults.

What it does appear to do, in small uncontrolled winter-swimmer studies, is upregulate antioxidant defenses over months — that's the actual hormetic mechanism, not "anti-inflammatory."

Acute immune-cell redistribution is real, but it is best documented for cryotherapy rather than for water. Randomising 26 men to three minutes of partial-body cryotherapy at −110 to −140°C, three minutes of 9°C water, or a 24°C control, cytotoxic (CD8+) T-cell counts rose after exposure — most in the cryotherapy arm, more than after either the cold water or the control — and a rise in circulating natural-killer cells, traced specifically to their cytotoxic subset, was seen after cryotherapy only.[15] This is redistribution (cells trafficking out of their resting reservoirs), not evidence of "boosted immunity". The immersion meta-analysis found no pooled effect on immune measures at all — a difference compatible with chance immediately after exposure and at one hour, at the lowest certainty grade.[16] Whether repeated mobilisation translates into the 29% sickness-absence signal seen in the cold-shower RCT remains mechanistically plausible but unproven.

The famous "cold boosts immunity" study is mostly about breathing, not cold. The most-cited result is a small RCT (n=24) in which Wim Hof Method practitioners, challenged with intravenous endotoxin, showed higher epinephrine, more anti-inflammatory interleukin-10, blunted pro-inflammatory cytokines (tumour necrosis factor-α, IL-6, IL-8) and fewer symptoms.[17] But the method bundles cyclic hyperventilation, and a follow-up dissected the components: 48 healthy men were randomised to cold-exposure training, the breathing exercise, both, or nothing, then given intravenous endotoxin. Cold training alone left both the inflammatory response and the epinephrine surge untouched; the breathing exercise produced both. Cold was not simply inert, though — it significantly amplified the breathing exercise's anti-inflammatory effect, and the combination was the most potent arm, which is the authors' own conclusion.[18] Acute endotoxemia is also a transient experimental challenge, not real infection or autoimmune disease, and a systematic review of the Wim Hof literature (8 trials, 15–48 participants each) rated all studies very low quality, 86% male, all at high risk of bias.[19] Evidence rating: Weak / preliminary, and the breathing does most of the work.

Cellular pathways: what the molecular evidence actually shows

Evidence rating: Weak / preliminary for every pathway below. Each has real molecular data, but none is validated against a human aging or healthspan outcome — the case for cold as a longevity stimulus rests on several pathways that are currently mechanistic rather than outcome-tested in people.

mTORC1 inhibition

Cold thermogenesis is intensely ATP-consuming, and the textbook story is that the resulting rise in the AMP:ATP ratio activates AMP-activated protein kinase (AMPK), the cell's low-fuel sensor, which then suppresses mTORC1 — the master switch that drives cell growth and protein synthesis. The best direct test complicates that chain: in mouse muscle cells, acute cold exposure did reduce mTORC1 signalling and protein synthesis, but it did so independently of AMPK.[20] So the mTORC1 suppression is real and cell-autonomous; the AMPK step usually invoked to explain it is not, on current evidence, how cold gets there. mTORC1 suppression is exactly the molecular feature linked to the healthspan benefits of fasting and rapamycin — and it is exactly the feature that blunts post-resistance-training hypertrophy. If the mouse-cell result holds in people, cold's "longevity mechanism" and its "anti-hypertrophy mechanism" would be the same molecular event, and it would be hard to get one without the other; no human study has tested that.

Autophagy upregulation (one small human trial)

A 2025 trial at the University of Ottawa put ten habitually active young men, average age 23, through up to an hour of seated head-out immersion in 13–15°C water on seven consecutive days and measured autophagy and cell-death markers in their circulating immune cells. Autophagy — the cell's bulk recycling of damaged proteins and organelles — stalled after the first immersion: the cargo protein that piles up when recycling is blocked roughly doubled, with no matching rise in the marker of autophagosome formation, while the executioner enzyme of programmed cell death rose about 2.6-fold. By day 7 the pattern had inverted — autophagosome formation up about 70%, the blocked-recycling marker a third below baseline, and the cell-death signal back to baseline rather than pushed below it. Note that the study measured marker levels rather than autophagic flux proper, and that pre-immersion baselines did not differ across the week, so nothing accumulated between sessions.[21]

Caveats are large: ten young healthy male participants, seven days, surrogate markers in circulating cells rather than tissue, no clinical outcomes — and no control or sham group, since every participant did every session, so day-7-versus-day-1 differences are uncontrolled for repeated blood sampling and laboratory habituation. It is a plausible mechanistic signal, not validation of "cold extends lifespan".

Cold-shock protein RBM3

RBM3 (RNA-binding motif protein 3) is upregulated by mild hypothermia and protects synapses in animal models of neurodegeneration. In mouse models of prion disease and Alzheimer's, cooling-induced RBM3 expression prevented synapse loss and behavioural decline; forcing RBM3 expression alone reproduced the effect.[22] RBM3 is genuinely interesting and the mechanism is well-mapped. The relevant unknown: brief whole-body cold immersion in humans does not reliably drop core temperature to the range that induces RBM3 in animals, and direct human RBM3 induction data from typical plunge protocols is essentially absent.

Conserved longevity sensors (animal models)

In short-lived model organisms lower temperature really does extend life, but the size and the mechanism differ by organism, and the mammalian data point somewhere other than cold baths. In C. elegans, rearing at 15–20°C rather than 25°C greatly lengthens lifespan, and it does so through an active genetic program rather than a passive slowing of chemistry: a cold-sensitive ion channel signals through a calcium-dependent kinase to the longevity transcription factor DAF-16/FOXO, and switching that channel on by itself adds about 16% to mean lifespan at 20°C.[23] In mice, a sustained 0.3–0.5°C reduction in core temperature — engineered centrally, with no cold exposure and no change in food intake — raised median lifespan by 12% in males and 20% in females.[24] Actual cold exposure in mammals is a different experiment with a different answer: rats that, after three months of gradual acclimatisation, stood in 23°C water four hours a day, five days a week until 32 months of age ate 44% more food and lived on average 968 days against 923 for controls — no meaningful longevity gain, although far fewer of them had malignancies at necropsy (24% versus 57%).[25] The conserved fuel sensors the longevity case leans on — AMP-activated protein kinase (AMPK) and the sirtuins SIRT1/SIRT3, which drive autophagy and damp inflammatory (NF-κB) signalling[26] — are genuinely engaged by cold, but no animal experiment has shown that repeated cold immersion extends mammalian lifespan. Two more cold-engaged mechanisms get cited in the longevity case: shivering and non-shivering thermogenesis release the myokine irisin, which boosts SIRT1, AMPK, autophagy and telomerase in models[27] (though irisin's measurement history is contested, so treat it as a soft signal); and cold represses myostatin — a brake on muscle growth — lifting inhibition of the heat-producing protein UCP1 and promoting white-fat browning in a single mechanistic study.[28]

The fair summary: the molecular machinery that long-lived organisms exploit (AMPK, sirtuins, autophagy, cold-shock proteins, irisin) is engaged by cold, and in worms lower temperature genuinely extends lifespan — but in mammals the one direct test of cold exposure found no lifespan gain. The leap that remains entirely untested is from those pathways to any human outcome — whether brief plunge protocols engage them durably enough to matter for human aging is exactly what we don't know.

Cardiovascular effects: bidirectional and time-dependent

Evidence rating: Strong for the acute stress response; Weak for any chronic benefit — no trial shows resting blood-pressure or arterial-stiffness improvement from cold immersion alone.

Acutely, whole-body cold immersion is a major cardiovascular stressor:

  • Heart rate rises sharply: by about 40 bpm in 7 °C water in a field study of 20 recreational swimmers[29]
  • Blood pressure rises, though over an hour at 14 °C the average systolic rise was only about 7%[30]
  • After cold-water or cold-air exposure, pooled across 24 studies, heart rate settles slightly below baseline while mean blood pressure stays slightly raised[31]

Shattock and Tipton's "autonomic conflict" hypothesis: simultaneous sympathetic activation (cold shock) and parasympathetic activation (face/diving reflex) creates an arrhythmogenic substrate. Free-breathing head-out CWI produces arrhythmias in ~2% of healthy young immersions; submersion with breath-holding raises this to 62–82%.[32]

The reassuring real-world counterpoint: a self-controlled field study put 20 middle-aged recreational athletes (11 women, mean age 56) through 64 immersions at a mean water temperature of 7°C with extended rhythm monitoring. Heart rate rose sharply (+40 bpm, ~43%), but only one brief non-sustained run of atrial tachycardia (8 beats) was captured during immersion, and arrhythmias over the whole 10 days ran about three times more frequent per hour during immersion than outside it (0.12 versus 0.037 events per hour) — a gap a study this small could not resolve either way.[33] Read that as an absence of dangerous events rather than a demonstration of equivalence: 17 of the 20 had brief, mostly benign arrhythmias at some point over the 10 days, and 7 of the 20 carried a high 10-year cardiovascular risk score, so this was not a hand-picked healthy sample. It complements rather than contradicts the autonomic-conflict danger: for adults without known cardiac disease who keep their airway clear, real-world arrhythmic risk looks low — it's the submerged breath-hold that turns it dangerous.

A distinct, under-discussed hazard is immersion pulmonary edema (IPE): the hydrostatic shift of blood into the chest plus cold vasoconstriction raises pulmonary arterial and wedge pressure and cardiac afterload, and in susceptible individuals fluid leaks into the alveoli, producing acute breathlessness, cough, and sometimes haemoptysis.[34] It is not rare in the population most studied, and the risk is strikingly unevenly distributed: across 47,573 swim distances at open-water swimming events, 0.44% ended in immersion pulmonary edema — about 1 in 230 — but the rate was 0.75% in women against 0.09% in men, and rose from 0.08% among 18–30-year-olds to 1.1% among those 61 and over.[35] Those figures come from strenuous open-water racing rather than static plunging, so read them as a signal about who is susceptible rather than as a plunge-tub rate. It can occur in otherwise healthy people and is a reason to exit and seek care if breathing becomes laboured during or after immersion.

No RCT has demonstrated chronic resting blood pressure reduction or improved arterial stiffness in normotensive adults through CWI alone.

Implication

  • Healthy adults: low absolute risk
  • Anyone with cardiac history: real concern; medical clearance before regular practice
  • Submerged breath-holding is high-risk — never combine with cold immersion
  • Stop for unexpected breathlessness — possible immersion pulmonary edema

Metabolic claims: what's actually true

Evidence rating: Moderate for sustained mild-cold-air insulin-sensitivity gains; the brief ice-bath version is not supported.

The strongest metabolic finding came from a 2015 trial in which 10 days of mild cold-air acclimation (14–15°C; 2 h on day 1, 4 h on day 2, then 6 h/day) increased peripheral insulin sensitivity by ~43% in eight men with type 2 diabetes.[36][37]

This was mild prolonged cold air, not brief ice baths. Mediated predominantly by muscle glucose-transporter (GLUT4) recruitment, not primarily brown adipose tissue.

Critically, a 2021 follow-up failed to replicate this benefit in nine patients, using a single-arm protocol that was both warmer (16–17°C rather than 14–15°C) and engineered to prevent overt shivering: insulin sensitivity, postprandial glucose and lipid handling, and liver fat were all unchanged.[38]

A separate whole-body cooling study muddies the shivering question rather than settling it: cooling healthy men in a water-perfused suit to ~18°C (shivering excluded) improved peripheral glucose uptake and insulin sensitivity by ~20% with first-phase insulin secretion unchanged, alongside BAT-linked shifts in plasma fatty acids.[39] Note what this is and isn't: prolonged, controlled, whole-body cooling — not a 1–3 minute plunge. Whether shivering is required (Remie's null fits "yes" but is confounded by its warmer protocol; this suggests maybe not) is unresolved, but every positive metabolic dataset in this literature uses sustained mild cold, never a brief ice bath.

Brown adipose tissue (BAT) is real and recruitable. But direct oxygen-consumption PET makes it a minor heat source: a few kcal/day measured in people with sizeable deposits, estimated at 15–25 kcal/day under mild cold for large depots,[40] and almost 1% of the cold-induced rise in whole-body energy expenditure, with deep cervico-thoracic muscles contributing far more.[41] The "cold burns hundreds of calories" framing is overstated for ice-bath protocols.

The Søberg cross-sectional winter-swimmer paper, in just 8 swimmers, is frequently cited for plunge-based metabolic benefit — but it's confounded by simultaneous sauna use and showed no brown-adipose-tissue glucose uptake at thermal comfort, alongside generally warmer skin that the authors read as heat acclimation from the sauna rather than a cold-induced gain.

Brief 1–3 minute ice-bath plunges do not replicate Hanssen's metabolic effects. And within sustained cold, dose does not run in the intuitive direction. In a 2026 randomised crossover, the mild exposure — a water-perfused suit at 15°C, set to raise resting metabolic rate about 1.5-fold, with shivering confirmed by indirect calorimetry and muscle electrical activity — lowered mean 24-hour glucose by 0.6 mmol/L and cut time spent above the glucose target by 10.9 percentage points of the day in 12 people with type 2 diabetes, while the far colder 4°C exposure, which pushed resting metabolic rate to about twice baseline, moved no glucose outcome at all. Neither dose helped 12 people with prediabetes.[42] Colder is not better, and the benefit appears to need established type 2 diabetes to act on.

The "11 minutes per week" claim

Evidence rating: Weak — a description of observed habits, not a tested dose threshold.

Susanna Søberg's "11 minutes per week" target is descriptive, not experimental. It reflects observed habits in the Copenhagen winter-swimmer cohort (~1–2 min × 3 dips × 2–3 days/week, typically paired with ~57 min/week of sauna heat) and was popularised in Søberg's 2022 trade book and Huberman Lab podcast appearances.

The Søberg et al. 2021 Cell Reports Medicine paper did not test this as a threshold. Treat it as a reasonable starting target, not a validated minimum effective dose.

The often-repeated "Søberg principle" — always end contrast therapy on cold so the body rewarms endogenously via sustained BAT/shivering thermogenesis — is mechanistically plausible but has not been directly tested as a head-to-head RCT against "end on heat". File it under reasonable heuristic, not validated protocol.

Cold immersion blunts post-resistance-training hypertrophy

Evidence rating: Moderate — a small, probable penalty across meta-analyses, on trials that are small and rated "fair to poor":

  • Cold immersion straight after resistance training reduces hypertrophic adaptations through documented molecular pathways. Post-exercise CWI suppresses the muscle-protein-synthesis signal downstream of mTORC1 and reduces activation of muscle stem cells (satellite cells) during a training block.[43]
  • A 2024 systematic review with meta-analysis ("Throwing cold water on muscle growth") finds the hypertrophy attenuation across trained and untrained populations — a small penalty, and a less certain one than the bare figure suggests (comparative standardised mean difference −0.22; the 95% credible interval runs from −0.47 to +0.04, and the authors put the probability that the true effect is a real penalty at about 96% — likely, not settled; eight trials, which they rate "fair to poor" in quality, and five of the eight immersed only one limb region rather than the whole body).[44]
  • The effect on strength is about the same size and only just clears significance — and it rests on a different design. A separate meta-analysis of 10 studies found CWI attenuated strength gains only modestly — a small effect that barely cleared significance (effect size −0.23, 95% CI −0.45 to −0.01; a 95% confidence interval is the range where the true effect most plausibly lies, and this one only just excludes zero). Crucially, that penalty came entirely from single-limb immersion; with whole-body CWI there was no significant difference from control (−0.08, 95% CI −0.53 to 0.38 — a range straddling zero, i.e. compatible with no effect at all).[45] Tellingly, one RCT cohort showed CWI blunted muscle size but not strength gains over the same training block.[46] So neither penalty is settled and both are small; what holds either way is the practical message — keep cold away from the lifting session.
  • The mechanism is the same mTORC1 suppression that gives cold its longevity mechanistic case — and, as noted above, the best direct test shows that suppression runs independently of AMPK, not through it. Cold is doing one thing molecularly; whether you frame that as "good" or "bad" depends on whether you're optimising for hypertrophy or for hormetic recycling.

Practical implication: if hypertrophy is your goal, keep cold immersion off the back of a lifting session. Note what the evidence can and cannot tell you about timing: every trial in the pooled hypertrophy evidence applied cold within about 15 minutes of the last set, and the meta-analysts state plainly that the effect of leaving more time between training and immersion is unknown.[47] No study has tested a waiting period, so any specific number of hours is a guess. Separating the two stimuli generously — morning cold and evening lifting, or alternate days — is the practical way to keep both without betting on a threshold nobody has measured. For pure recovery (between events) or delayed-onset muscle soreness (DOMS) reduction in non-hypertrophy contexts (endurance training tournaments, etc.), it remains useful.

Mood and depression: real acute effects, no clinical evidence

Evidence rating: Moderate for the acute mood lift; Weak / none for treating clinically diagnosed depression or anxiety.

  • Acute mood effects are consistently reported, on small, mostly uncontrolled studies.
  • The frequently-cited van Tulleken case — a 24-year-old woman whose major depressive disorder had resisted fluoxetine and then citalopram, who came off medication over a programme of weekly open-water swimming and was still medication-free a year later — is a single uncontrolled case, not causal evidence.[48]
  • No adequately powered RCT exists for CWI in clinically diagnosed depression or anxiety disorders.

Practitioners should not represent cold plunging as a substantiated antidepressant. It may help mood acutely, but use evidence-based treatments (CBT, SSRIs, exercise) for clinical depression.

Sleep and cognition

Evidence rating: Weak / preliminary — small trials, and the direction flips with timing and the degree of core-temperature drop.

  • Whether late-evening cold immersion impairs sleep in people who have not just exercised is untested; the catecholamine arousal is a plausible reason for caution, not a finding.
  • Post-evening-exercise cold immersion may help, with a caveat about what was compared. Twelve trained male runners ran in the early evening and then took 10 minutes at 13.3°C either whole-body or only to the hip: slow-wave sleep made up a greater proportion of the first three hours of the night after whole-body immersion than after partial immersion, and both immersion conditions reduced arousals across the whole night against a dry control. The comparison that reached significance for slow-wave sleep was whole-body against partial immersion, not against not immersing at all.[49] A separate trial of evening post-exercise immersion in 11 cyclists found no difference between conditions on any whole-night sleep measure.[50]
  • Cognition deteriorates as core temperature falls, and it starts earlier than most people expect. When volunteers were cooled in 15°C water and then tested while rewarming — skin warm and comfortable, core still low, so the effect is core temperature rather than distraction — the ability to register new information declined progressively from a core temperature of about 36.7°C, and at 34–35°C roughly 70% of normally retainable material was lost. Speed on a two-digit calculation task fell by about half at 34–35°C, though accuracy held when enough time was allowed, and recall of already-learned material was unaffected.[51] A 1–3 minute plunge rarely moves core temperature that far; this matters for long immersions and open-water swimming, where judgement degrades before the swimmer notices.

Balance and motor control

Evidence rating: Weak — few, small, heterogeneous studies, none of which found balance worse after immersion; the warm-up precaution rests on cooling physiology rather than on balance trials.

Don't plunge right before anything that needs steady hands or feet. Immediately after immersion, cold can transiently degrade fine motor control — the drivers are slowed peripheral nerve conduction, cold-induced muscle stiffness, and cutaneous vasoconstriction blunting proprioceptive feedback.

The evidence base is thin and mixed. A 2025 systematic review of eight small studies (225 healthy young adults, immersion 4–12°C) could not be pooled into a meta-analysis because protocols varied too much; three studies found acute balance improvements, five found no effect, and none found balance worse; the authors' overall verdict leans slightly positive, though contradictions persist, and they suggest repeated exposure may improve postural control over time via neural adaptation.[52] Treat the long-term proprioceptive benefit as a low-certainty hypothesis. The precaution costs nothing: warm up and let neuromuscular function recover before precision or balance-critical activity after cold.

Habituation: what is and isn't preserved

Evidence rating: Moderate.

A common practical question: "If I'm used to it, am I still getting the benefits?" The cleanest answer the literature supports:

  • Sympathetic/respiratory responses habituate. Gasp reflex, hyperventilation and the heart-rate spike all blunt substantially within ~4–6 exposures and stay blunted long-term. This is mostly a safety gain, not a lost benefit.
  • Thermogenic and metabolic responses do not habituate the same way. Winter swimmers show greater cold-induced thermogenesis than naïve controls on identical exposures: resting energy expenditure matched the controls' at thermal comfort but rose by roughly 1,000 kcal/24 h during 30 minutes of cooling, against about 550 kcal/24 h in controls, alongside a warmer supraclavicular skin temperature and a lower thermoneutral core temperature.[53] Brown-fat glucose uptake — the study's primary outcome — rose no more in the swimmers than in the controls, and muscle activity did not differ either; the authors state plainly that brown fat cannot account for the extra heat, so its source is unidentified. The body becomes more efficient at executing thermogenesis without requiring an extreme cardiovascular panic to initiate it. The same cohort also showed a pronounced diurnal peak in supraclavicular skin temperature in the small hours (~4:30–5:30 AM) that was flat in controls, hinting that habitual cold exposure entrains the daily rhythm of supraclavicular (possibly brown-fat) heat production.

So if your endpoint is mood/alertness, you may need to ratchet stimulus (colder, longer, or less frequent) to keep getting the subjective hit. If your endpoint is metabolic — the habituated state is the adapted state, not a worn-out one.

Sex differences

Evidence rating: Weak / preliminary — most cold-immersion research is in young men, and the sex-specific data is sparse.

Most CWI research has been done in young healthy men, but the sex-difference data we do have is meaningful for protocol design:

  • Cooling kinetics track body composition, not sex. How fast core temperature falls is predicted by body fatness and the ratio of surface area to body size; sex by itself adds little. In the one matched head-to-head immersion study, women actually cooled at about half the men's rate — 0.47°C per hour in 18°C water — because they carried more subcutaneous fat, and once men and women were subgrouped for similar body fatness the difference disappeared entirely.[54] The practical implication survives with a different target: a lean, small person of either sex takes a larger thermal stimulus from the same protocol and should start shorter.
  • Metabolic response scales with body size, not sex. In that same study metabolic rate rose to roughly three times resting in both sexes, and total heat production normalised for body mass or surface area did not differ, nor did the share of heat coming from fat oxidation (about 60% in both).
  • Immersion pulmonary edema is far more common in women — 0.75% of open-water race swims against 0.09% in men (about eightfold after adjustment), in strenuous racing rather than plunging (see the cardiovascular section above).[55]
  • Hormone trajectories differ: leptin, adiponectin, and glucose responses to acute cold show statistically distinct curves between sexes.
  • Cognitive performance under cold load in some studies degrades more in women (reaction time, spatial tasks) than men under identical exposures.
  • The Søberg "11 minutes" prescription was derived from a male winter-swimmer cohort. There is no equivalent dose-finding data in women, and extrapolating directly is a known weakness in the literature.

None of this argues against CWI in women — it argues against assuming protocols designed for men transfer 1:1.

A reasonable practical protocol

If you want the documented benefits without overhyping:

  1. Brief cold exposure for mood/alertness: cold shower 1–3 min, water as cold as available; or 2–5 min in 10–15°C water. Morning timing aligns with the natural cortisol awakening response.

  2. Recovery cold immersion (between athletic events, after endurance training): the recovery goal sets the dose. A 2025 network meta-analysis of post-exercise protocols ranked them by outcome (as standardised mean differences, where roughly 0.5 is a moderate effect and 0.8 or above is large)[56]:

    GoalBest-ranked protocolEffect
    Muscle soreness (DOMS)11–15°C, 10–15 minSMD −1.45
    Neuromuscular recovery (jump power)5–10°C, 10–15 minSMD +0.48 (contested — see below)
    Creatine-kinase clearance5–10°C, 10–15 minSMD −0.90 (contested — see below)

    Two of those three rows do not survive a more conservative synthesis. Pooling 22 randomised trials of a single cold immersion against passive rest, a 2026 meta-analysis found no overall effect on jump performance, and a transient impairment of explosive power in the first hour after immersion that had resolved by 24–48 hours; no effect on maximal isometric strength; and a creatine-kinase benefit that vanished once publication bias was corrected for. Only the soreness reduction survived, and at very low certainty.[57] A second 2026 pool of 30 trials reproduced the nulls for both jump and strength.[58] Treat soreness as the one durable row — and do not take cold immediately before anything power-dependent.

    Longer immersions (>15 min) add overcooling risk and poor adherence without proportionate benefit. Pairing CWI with an adjunct (active recovery, stretching, massage, compression) is associated with a larger soreness reduction than cold alone in trained athletes — a large pooled effect against a small-to-moderate one (SMD −1.13 vs −0.47) — but the two figures come from separate comparisons against control rather than a head-to-head trial, because the authors found too few direct comparisons to pool and fell back on indirect ones. The intervals overlap; and across the whole review — not just the athlete subgroup — neither cold alone nor the combination beat control on the objective outcomes, creatine kinase and countermovement jump.[59] These recovery uses are for non-hypertrophy contexts — see the next point.

    Read these effect sizes as upper bounds. Evidence rating: Moderate, downgraded by a near-total inability to blind. The Cochrane review of CWI for soreness included 17 small trials in 366 participants and found CWI beats passive rest/no intervention for DOMS — a moderate benefit versus doing nothing, pooled from 10 of those trials (SMD −0.55, 95% CI −0.84 to −0.27 at 24 h, with substantial heterogeneity) — but insufficient evidence that it beats warm-water immersion, contrast therapy, active recovery or compression, the last two resting on a single trial each — and flagged poor trial quality (inadequate randomisation, allocation concealment and blinding). Its literature search closed in November 2011 and it has never been updated.[60] The deeper problem is the placebo: CWI recovery trials are nearly impossible to blind, and when a credible sham is used the cold-specific benefit largely vanishes. In an RCT where thermoneutral immersion was presented as a "recovery oil," the placebo restored strength, perceived recovery, pain and vigour indistinguishably from genuine CWI.[61] Several subsequent sham-controlled RCTs reproduce "no better than placebo," including a 2025 trial in national-level soccer players.[62] The DOMS and perceived-recovery numbers above are partly inflated by expectation.

  3. Keep cold away from resistance training if hypertrophy is your goal — separate them generously or use alternate days; no specific waiting period has been tested (see above).

  4. If you try late-evening cold, watch how you sleep. Its effect outside post-exercise trials is untested, and even after exercise the sleep benefit is small and inconsistent across trials.

  5. Build slowly. Cold shock can be dangerous in unhabituated individuals. Start with brief cold-finish showers, progress over weeks.

  6. Never breath-hold submerged in cold water.

  7. Have a buddy or be in monitored conditions for any cold-water swimming or extended immersion, and get out slowly — sit or lie down for the first few minutes rather than standing and walking off.

Contraindications

Strong:

  • Cardiac arrhythmia (atrial fibrillation, long QT, prior cardiac arrest)
  • Severe Raynaud's phenomenon
  • Cold urticaria (cold contact can trigger systemic histamine release and anaphylaxis)
  • Cryoglobulinaemia
  • Pregnancy (especially deep immersion)
  • Recent myocardial infarction (MI) or unstable cardiovascular disease, coronary artery disease with unstable angina, New York Heart Association (NYHA) class III–IV heart failure

Use with caution:

  • Hypertension (poorly controlled)
  • Peripheral vascular disease
  • Untreated hypothyroidism (impaired thermoregulation)
  • Diabetic autonomic / peripheral neuropathy — blunted thermoception means dangerous tissue cooling can go unnoticed, and impaired microvascular rewarming compounds the risk
  • History of hypothermia susceptibility

Medication interactions worth flagging with your clinician before regular practice:

  • Anything that prolongs the QT interval — the electrical window in which the heart is most vulnerable to a dangerous rhythm. This is the drug interaction the cold-immersion literature actually flags: some antihistamines, macrolide antibiotics, and antiarrhythmics such as quinidine, procainamide, amiodarone and sotalol.[63] They stack on top of an exposure that can itself lengthen repolarisation — in 20 trained swimmers who completed a long open-water event in 15°C water (cold and hours of exercise together; the change did not track core cooling), the corrected QT interval lengthened by about 20 milliseconds on average within an hour of finishing, and 5 of the 20 came out above 500 milliseconds, the level at which the risk of a malignant rhythm is considered elevated.[64] Take this seriously if you have known long QT syndrome, or a family history of unexplained drowning or sudden death.
  • Beta-blockers suppress the compensatory tachycardia needed to maintain cardiac output against the cold-induced rise in systemic vascular resistance — raising syncope and decompensation risk.
  • ACE inhibitors, angiotensin-receptor blockers (ARBs), calcium channel blockers alter baseline vascular tone; the rebound vasodilation on exit can drive significant BP swings and orthostatic syncope.
  • Diuretics compound the cold-induced diuresis and can leave you volume-depleted, again worsening post-exit syncope risk. They also lower potassium and magnesium, and low potassium lengthens the QT interval.

Risks not to discount

  • Drowning — the way most cold-water immersion deaths are certified, though a 2026 review argues the lethal sequence usually starts earlier. Cold shock in the first seconds can cause aspiration, hyperventilation-driven loss of brain blood flow, and cardiac arrhythmia before core temperature has fallen at all; swimming failure follows as cooled arm and leg muscles stop working while core temperature is still near normal; and hypothermia is usually a late finding rather than the primary cause.[65] Strong swimmers are not exempt: two thirds of drownings occur in competent swimmers, and more than half of those within three metres of safety — figures Shattock and Tipton read as pointing to sudden arrhythmic death rather than exhaustion.[66] Always have a buddy and supervised conditions.
  • Circum-rescue collapse — the moment you leave the water the hydrostatic squeeze on your legs disappears and blood pools. The same review names that abrupt loss of support as a distinct cause of death around rescue from open-water immersion; whether it scales down to stepping out of a plunge tub has not been tested, but the precaution costs nothing. Come out slowly, sit or lie down rather than standing and walking off, and have someone stay with you for the first few minutes.
  • Cardiac arrhythmia — especially with breath-holding or in those with predisposing conditions.
  • Hypoxic blackout — hyperventilating before a breath-hold (as in the Wim Hof breathing exercise) extends time under water but can cause loss of consciousness without warning; of 58 such underwater-swimming blackouts, all in good swimmers, 23 were fatal.[67] Do breathing exercises on dry land only.
  • Hypothermia — extended exposure or insufficient warming after.
  • Frostbite — at very low temperatures or extended skin contact with ice.
  • Immersion pulmonary edema — sudden breathlessness or cough from fluid shifting into the lungs; can strike otherwise-healthy people. Exit and seek care.

Bottom line

Cold plunging is a plausible hormetic stressor with several measured short-term benefits and documented risks. Its longevity case is mechanistic, indirect, and currently untested in humans. No prospective cohort has tested whether habitual cold immersion lowers mortality, in stark contrast to the well-developed Finnish sauna mortality cohort.

If you enjoy it, the safety profile in healthy adults is reasonable. If you're hoping for transformative health effects, the evidence doesn't support that level of claim.

Further reading

  • Tipton MJ et al. Cold water immersion: kill or cure? Exp Physiol 2017.[68]
  • Šrámek P et al. Human physiological responses to immersion into water of different temperatures. Eur J Appl Physiol 2000 — foundational dataset for the 14°C catecholamine/metabolic/diuretic profile.[69]
  • Cain T et al. Cold-water immersion: meta-analysis of inflammation/stress effects. PLOS ONE 2025.[70]
  • Hanssen MJW et al. Short-term cold acclimation and insulin sensitivity in T2D. Nat Med 2015.[71]
  • Roberts LA et al. Post-exercise cold water immersion attenuates anabolic signalling and adaptations. J Physiol 2015.[72]
  • Bleakley C et al. Cold-water immersion (cryotherapy) for preventing and treating muscle soreness after exercise. Cochrane Database Syst Rev 2012;CD008262 — foundational synthesis; CWI beats rest but not other modalities, trial quality poor.[73]
  • Broatch JR, Petersen A, Bishop DJ. Postexercise cold water immersion benefits are not greater than the placebo effect. Med Sci Sports Exerc 2014.[74]
  • Kox M et al. Voluntary activation of the sympathetic nervous system and attenuation of the innate immune response in humans. PNAS 2014 — the "Wim Hof" immune study; effect driven by breathing.[75]
  • Costello JT et al. Whole-body cryotherapy (extreme cold air exposure) for preventing and treating muscle soreness after exercise in adults. Cochrane Database Syst Rev 2015;CD010789 — four small laboratory trials, 64 participants, all outcomes very-low-certainty, no adverse-event surveillance.[76]
  • Barwood MJ et al. Habituation of the cold shock response: a systematic review and meta-analysis. J Therm Biol 2024 — cold-shock response habituates after ~4 immersions.[77]
  • Piñero A et al. Cold water immersion attenuates resistance-training hypertrophy — systematic review and meta-analysis. Eur J Sport Sci 2024.[78]
  • Søberg S et al. Brown fat thermoregulation in winter-swimming men. Cell Rep Med 2021.[79]
  • Galvez-Rodriguez C et al. Cold Water Immersion, Heart Rate Variability and Post-Exercise Recovery: A Systematic Review. Physiother Res Int 2025.[80]
  • Buijze GA et al. The Effect of Cold Showering on Health and Work: A Randomized Controlled Trial. PLOS ONE 2016.[81]
  • Shattock MJ, Tipton MJ. Autonomic conflict. J Physiol 2012.[82]
  • Peretti D et al. RBM3 mediates structural plasticity and protective effects of cooling in neurodegeneration. Nature 2015.[83]
  • King KE et al. The Effect of 7-Day Cold Water Acclimation on Autophagic and Apoptotic Responses in Young Males. Adv Biol (Weinh) 2025;9(2):e2400111 — ten young men, up to 60 min in 13–15°C water on seven consecutive days; the only human trial of cold acclimation and autophagy to date.[84]
  • Kunutsor et al. The untapped potential of cold water therapy as part of a lifestyle intervention for promoting healthy aging. GeroScience 2025.[85]
  • Merkt B et al. Cardiac arrhythmia during cold-water immersion: a self-controlled field study with extended rhythm monitoring. J Exerc Sci Fit 2026 — low arrhythmic burden across 64 immersions at 7°C.[86]
  • Iwen KA et al. Cold-induced brown adipose tissue activity alters plasma fatty acids and improves glucose metabolism in men. J Clin Endocrinol Metab 2017 — ~20% insulin-sensitivity gain with prolonged suit cooling, shivering excluded.[87]
  • Wang H et al. Impact of different doses of cold water immersion (duration and temperature) on recovery from exercise-induced muscle damage — network meta-analysis. Front Physiol 2025.[88]
  • Ma J et al. Comparison of the Effects of Cold-Water Immersion Applied Alone and Combined Therapy on the Recovery of Muscle Fatigue After Exercise — systematic review and meta-analysis. Life (Basel) 2025.[89]
  • Shahani et al. Effect of cold water immersion on balance and postural control in young healthy individuals — systematic review. J Orthop Surg Res 2025.[90]
  • Bárcena ML et al. Role of AMPK and Sirtuins in Aging Heart: Basic and Translational Aspects. Aging Dis 2024 — AMPK/sirtuin signalling in aging; contains no cold-exposure or lifespan data.[91]

— § —