Cold Plunge and the Immune System: What the Evidence Shows

Three different claims sold as one

"Cold plunging boosts your immune system" collapses three separate claims into one sentence, and they have wildly different amounts of support behind them.

Marketing copy takes the third claim, which is easy to demonstrate, and presents it as evidence for the first, which has never been demonstrated. A shift in circulating cell counts is a change in where immune cells are, not proof that they work better. That substitution is the core error in most cold plunge immunity content.

The 3,018-person cold shower trial

Buijze et al. (2016), published in PLOS ONE, is the only large randomised trial in this area and the source of the "29%" figure that circulates everywhere. Between January and March 2015, 3,018 Dutch adults aged 18 to 65 with no routine cold shower habit were randomised to finish their normal warm shower with 30, 60 or 90 seconds of cold water, or to a control group that showered as usual, for 30 consecutive days. Outcomes were tracked for 90 days.

Two details from that trial rarely make it into summaries. First, there was no dose response. Thirty seconds worked about as well as ninety, which is not what you would expect from a graded physiological adaptation. Second, participants who also did regular physical activity showed a larger reduction in sickness absence, roughly 54% when the two were combined, which raises the obvious question of how much of the benefit is downstream of being an active, motivated person who signed up for a cold shower study.

Secondary outcomes were largely null. Work productivity and anxiety scores did not differ between groups. Quality of life showed a small mental component improvement at 30 days that had vanished by 90 days and was described by the authors as too small to be clinically relevant. The most commonly reported subjective effect was an increase in perceived energy.

Sick days are not illnesses

This is the distinction that matters most and the one most articles get wrong. Sickness absence is a decision. Illness days are a symptom report. The trial found a change in the decision and no change in the symptoms.

There are several plausible explanations, and immune enhancement is only one of them. The trial could not be blinded, since you always know whether you just stood under cold water. Participants who signed up for a cold shower study and completed a month of it have a reason to believe it is doing something, and the outcome they influenced most is the one that moved. A daily voluntary discomfort habit plausibly raises tolerance for mild discomfort in general, so the same head cold that would have kept someone home now gets shrugged off. The reported jump in perceived energy fits that reading better than it fits a change in viral clearance.

None of this makes the finding worthless. Fewer days lost to illness is a genuine outcome that employers and individuals care about, and it was found in a large randomised sample. It just is not evidence that your immune system got stronger, and it should not be reported as "cold showers make you 29% less likely to get sick." If you are weighing whether to plunge during an active illness, that is a different question with a different answer, covered in cold plunging when sick.

What actually happens to immune cells

The acute response is well characterised. Brenner et al. (1999, Journal of Applied Physiology) exposed seven healthy men to 5°C air for two hours and measured a leukocytosis and granulocytosis, an increase in natural killer cell count and activity, and a rise in circulating interleukin-6. Pretreatment with exercise in 18°C water amplified the leukocyte, granulocyte and monocyte response. Later work using whole-body cold water immersion reproduces the same pattern on a compressed timeline.

The mechanism is not mysterious and it is not immune building. Catecholamines released during cold shock mobilise cells that were already sitting in the marginal pool along vessel walls and in lymphoid tissue, pushing them into circulating blood where a blood draw can count them. Natural killer cells are especially responsive to this and can rise substantially within the first half hour. Levels drift back toward baseline within hours, and a single immersion produces redistribution rather than production or destruction of cells.

The one study suggesting durable change is Janský et al. (1996), which used 14°C immersion three times weekly for six weeks and reported small increases in monocyte proportion, in lymphocytes expressing the CD25 receptor, and in plasma TNF-alpha. It had roughly ten participants and has not been replicated at scale in the thirty years since. Set against Versteeg et al. (2023), which found nothing over three weeks, the honest reading is that repeated cold exposure has not been shown to move baseline immune status in either direction.

The Wim Hof endotoxin data, and its confound

Kox et al. (2014, PNAS) is the most striking immune result in this literature. Twelve healthy volunteers completed a 10-day training programme combining third-eye meditation, cyclic hyperventilation with breath retention, and immersion in ice cold water. Then all 24 participants, trained and untrained, received an intravenous dose of 2 ng/kg E. coli lipopolysaccharide, a controlled bacterial endotoxin challenge that reliably produces a flu-like inflammatory response.

The trained group, practising their techniques during the challenge, released markedly more epinephrine, produced more of the anti-inflammatory cytokine IL-10, showed a substantially blunted pro-inflammatory cytokine response, and reported fewer flu-like symptoms. As a demonstration that the innate immune response can be voluntarily modulated at all, it was a landmark finding and it deserves the attention it received.

It cannot be used as evidence for cold plunging, because cold was one of three bundled interventions. Zwaag et al. (2022, PLOS ONE) ran the obvious follow-up, separating the components to see which one carried the effect. Both breathing exercises, with and without breath retention, raised epinephrine. Cold exposure training on its own had no significant effect on epinephrine or cytokine levels. On the current evidence, the anti-inflammatory response in the Wim Hof protocol is driven by the breathing, and the ice water is along for the ride.

That does not mean the endotoxin result says nothing about cold. It means anyone citing Kox to sell a plunge tub is citing a study whose own follow-up work points elsewhere. For the separate question of what cold does to inflammatory markers around training, see cold plunge and inflammation, where the evidence is considerably stronger and much narrower than the immunity story.

Evidence strength, claim by claim

Each row is a claim you will see made about cold exposure and immunity, with the study behind it and the reason it does or does not hold.

Where the hormesis argument gets over-claimed

Hormesis is the idea that a small dose of a stressor triggers an adaptive response that leaves the system more robust than before. It is a legitimate concept with good support in cell and animal models, and it is the mechanistic story behind almost every cold plunge immunity claim. The problem is that it is deployed as if it were a finding rather than a hypothesis.

Three specific over-reaches show up repeatedly. The first is treating hormesis as automatically transferable to humans at any dose: benefit in laboratory models does not establish benefit in a 40-year-old taking a three-minute plunge four mornings a week, and the human data remains thin. The second is inconsistency about direction. Cain et al. (2025) found that cold water immersion sharply increased inflammatory markers immediately after immersion and at one hour, yet the same practice is marketed as anti-inflammatory. Both can be true across different timescales, but a mechanism cannot be invoked in whichever direction happens to suit the sales page.

The third is dose blindness. If the stressor drives the adaptation, then more should be better up to a point and worse beyond it. The Buijze trial found no dose response between 30, 60 and 90 seconds, which does not fit a simple hormetic curve. Meanwhile the same logic that predicts benefit at low doses predicts harm from accumulated stress at high ones, which is exactly what shows up in the side effects of overdoing cold exposure.

The 2025 meta-analysis of 11 randomised trials and 3,177 participants is the cleanest statement of where things stand: no significant effect of cold water immersion on immune function measures immediately after exposure or one hour later, with overall certainty of evidence graded D. That is not a refutation. It is an accurate description of a field with few trials, small samples and limited diversity in who gets studied.

What to actually do with this

If immunity is your only reason for starting, the evidence does not justify buying equipment. If you already plunge for recovery, mood or the habit itself, there is no reason to stop and a modest chance you take fewer days off work as a byproduct. Both positions are defensible. What is not defensible is a plunge protocol built around illness prevention.

These numbers come from the protocols used in the research, not from a demonstrated immune dose. Colder and longer is not better here, and the cold shower trial found no advantage to 90 seconds over 30. Full guidance on the ranges sits in our pages on cold plunge temperature and how often to plunge.

For the wider picture of which cold plunge benefits survive scrutiny and which do not, our evidence review of cold plunge benefits grades every major claim the same way this page grades the immune ones.

Frequently asked questions