One day, something I started looking into out of casual curiosity expanded into questions like "What does scientific proof actually mean?" and "How do I make my own judgments?"

The topic was contrast bathing — alternating between hot and cold water.

Many people have heard of it as a health practice. But I'd never properly understood the mechanisms behind why it's supposedly good for you. When I dug deeper, it took me to places far more profound than I expected.

First, I Wanted to Understand the Mechanism

The entry point was three terms: Heat Shock Proteins (HSP), mitochondria, and Natural Killer cells (NK cells).

I'd heard each one individually, but I'd never organized the relationships among the three.

HSP are molecular chaperones whose expression increases within cells when stimulated by heat, oxidative stress, and similar factors. They protect cells by helping proteins fold correctly and promoting the repair of damaged proteins.

Mitochondria produce reactive oxygen species (ROS) as a byproduct during energy (ATP) production. When ROS becomes excessive, it damages the mitochondria's own proteins, but HSPs (particularly HSP60 and HSP70) perform protein quality control inside mitochondria, protecting their function. HSPs serve as the "guardians" of mitochondria.

The relationship with NK cells involves two pathways. First, when stressed or cancerous cells display HSP70 on their surface, NK cells recognize this as a "danger signal" and attack and eliminate those cells. Second, HSPs released outside cells induce NKG2D ligand expression on dendritic cells (DCs), which indirectly activates NK cells through this pathway.

And since NK cells require enormous amounts of energy when attacking target cells, higher mitochondrial activity means greater NK cell capability. It's also known that mitochondrial dysfunction leads to decreased NK cell activity.

Organizing the relationship among the three looks like this:

Relationship Diagram: HSP, Mitochondria, and NK Cells

graph TD
    A["đŸ”„ Heat / Oxidative Stress</br>(Exercise, Thermal Stimulation)"]
    A --> B["HSP Production Increases</br>(HSP60 / HSP70)"]

    B --> C["[Mitochondrial Protection]</br>HSP60 manages protein</br>quality control in mitochondria"]
    C --> D["Mitochondria produce</br>ATP (energy) normally"]

    B --> E["[Pathway 1] HSP70 displayed</br>on cell surface</br>(Stressed / Cancer cells)"]
    E --> F["NK cells directly recognize</br>as 'danger signal'"]

    B --> G["[Pathway 2] HSP70</br>released extracellularly"]
    G --> H["Dendritic cells (DC) express</br>NKG2D ligand (MICA)"]
    H --> I["NK cells indirectly activated</br>via NKG2D receptor"]

    D --> J["Supports NK cell activity</br>through energy supply"]
    F --> K["đŸ›Ąïž NK cells attack and</br>eliminate target cells"]
    I --> K
    J --> K
    K --> L["Enhanced immune surveillance"]

    M["⚠ Aging / Chronic Stress"]
    M --> N["Decreased HSP production"]
    N --> O["Mitochondrial dysfunction"]
    O --> P["Decreased NK cell activity"]
    P --> Q["Weakened resistance to</br>infections and cancer"]

    style A fill:#ffe0b2,stroke:#e65100
    style L fill:#c8e6c9,stroke:#2e7d32
    style M fill:#ffcdd2,stroke:#c62828
    style Q fill:#ffcdd2,stroke:#c62828
    style K fill:#b3e5fc,stroke:#01579b

Moderate stress (exercise, thermal stimulation, etc.) → HSP production is promoted → HSPs protect mitochondria while also acting as activation signals for NK cells → Immune surveillance is enhanced.

When this cycle is disrupted by aging or excessive stress, mitochondrial dysfunction, decreased HSP production, and reduced NK cell activity cascade together, potentially weakening resistance to infections and cancer.

The logic is remarkably clean. There was a feeling of it "clicking into place."

So What About the Evidence?

"The logic checks out" and "it's been proven" are different things.

As I investigated further, I found that surprisingly robust research exists.

In 2016, a randomized controlled trial (RCT) involving 3,018 people was conducted in the Netherlands. This study, published in PLOS ONE and dubbed the "Cool Challenge," was it.

The design was simple. Participants were divided into four groups, ending their daily warm shower with cold water for "30 seconds," "60 seconds," or "90 seconds," compared against a control group that continued showering as usual, over 30 days. The cold water temperature was "the coldest available tap water," which averaged 10-12°C during the study period in the Netherlands.

The result: the intervention groups saw a 29% reduction in sick leave compared to the control group (P=0.003).

But here's something interesting. There was no significant difference in "the number of days they were actually sick."

They didn't stop catching colds. But when they did catch one, they recovered enough to not need time off.

How to read this? When I first saw it, I thought, "If the number of sick days doesn't change, does it mean anything?" But with a little reflection, the picture shifted. The more natural interpretation is that "they still get infected, but the body's coping ability has improved." Since cold showers can't reduce viral exposure, it means that even with the same infection, symptoms are milder and recovery is faster.

Another striking finding was that there was no significant difference among the 30-second, 60-second, and 90-second groups. The maximum physiological response to cold water occurs in the first 30 seconds, and this speed suggests that neural pathways are the primary driver. In other words, 30 seconds is enough.

Additionally, regular exercise habits also reduced sick leave by 35%, and combining cold showers with exercise produced a combined 54% reduction in sick leave.

Looking Honestly at the Study's Limitations

However, we need to be cautious about declaring this study as outright "proof." The authors themselves acknowledged several limitations.

No blinding possible. Participants knew which group they were in. This means placebo effects can't be completely ruled out. However, the authors make an interesting point — even if it is a placebo effect, it involves physiological responses including norepinephrine release and activation of specific brain regions, which is qualitatively different from believing in a sugar pill. Since participants are actually being exposed to cold water, something is definitely happening in the body.

All self-reported. Sick days and days of illness were all self-reported through online questionnaires, with no objective verification.

Sampling bias. 96% of participants rated their health as "good" or "excellent," and 85% exercised regularly. This is substantially higher than the Dutch general population (53%). The control group's sick leave rate was also 1.55%, roughly one-third of the Netherlands' average (4.4%) during the same period. These were already very healthy, active people.

However, this can be read two ways. One reading is that generalization requires caution. But another is that "there's an additional 29% effect on top of people who are already healthy." The fact that such a significant difference emerged in a group with already low sick leave could indicate the robustness of the effect.

Overlap with a major flu epidemic. The study was conducted during the 2014/2015 Dutch influenza epidemic, which lasted 21 weeks — the longest in over 40 years (according to the Netherlands' National Institute for Public Health and the Environment, RIVM). You could read this as the effect holding up even in a high-infection-risk environment, but you also can't deny it may have influenced the results.

Structurally, the research has both ends in place — "basic mechanism research" and "large-scale epidemiological outcomes" — but the causal process connecting them is missing. The outer pieces of the puzzle are assembled, but the middle pieces aren't filled in yet. That's where I think we currently stand.

The Question About "Catching a Cold Early On"

A question nagged at me here.

I understood the idea of contrast bathing as a daily habit "raising the immune baseline." But what about when you're already at the "I might be coming down with something" stage?

Initially, my answer was "HSP production takes several hours to days, so it can't help when you're already getting sick right now." But I needed to reconsider this.

Even with a time lag, HSP70 production peaks 2-6 hours after thermal stimulation (research reports show Hsp70 peaks 2-6 hours after heat stimulation at 43°C). If you bathe in the evening and go to sleep, HSP70 increases during sleep and can contribute to NK cell activation. During sleep, growth hormone secretion also increases and immune repair activity becomes more active, so HSP effects may synergize with this.

Rather than "it can't help in time," a more accurate reading is "precisely because the immune response kicks in hours later, there's value in doing it early." What the Dutch study showed was exactly this pattern — "recovering faster even when you do catch a cold."

I Ended Up Testing It on Myself

There was a reason I was thinking about this topic — I hadn't been feeling great since that morning.

I had chills and a dull ache at the base of my joints. The classic sensation of "the beginning of a cold."

So I did contrast bathing around 2 PM and planned to go to bed early that night.

Before bed, my temperature was 37.6°C. The next morning it was 36.9°C — my normal temperature. The symptoms had almost completely disappeared.

Honestly, I was a little surprised.

Going from joint pain and chills to normal temperature with symptoms gone overnight. To my own perception, that felt "fast." Usually these things linger for 2-3 days.

The timeline works out like this: contrast bathing at noon → HSP production peaks during the night → immune activity intensifies during sleep → NK cell activation is boosted, suppressing viral replication. The mechanism is consistent.

But I can't declare this was the effect of contrast bathing.

To be precise, that day I also took kampo medicine (kakkonto) and had an energy drink before sleep. I was in a state of "doing everything I could that day." So I can't claim it was solely the effect of contrast bathing. Still, as a result of applying multiple interventions simultaneously, the fever broke in one day and I recovered afterward. I record this as fact.

The Wall of n=1 Proof

This is a fundamentally difficult topic to research.

The reason it can't be proven at the individual level is clear: the same person cannot simultaneously experience "having done it" and "not having done it." The virus type, viral load, and physical condition differ every time. There's no basis for comparison.

Even at the research level, there are many barriers. Blinding is impossible, everything ends up being self-reported, and tracking the HSP → NK cell pathway in real time within the human body is extremely costly and invasive. Completely measuring the entire process — "this person took a bath this morning, which increased HSP70 by this much, activated NK cells by this amount, and eliminated the virus in this timeframe" — in a living human is not realistic with current technology.

Cold severity itself also varies too much between individuals and viruses, making it difficult to define and quantify "it was mild." The Dutch study's use of "sick leave" as an indirect indicator was a creative solution born from the impossibility of direct measurement.

So my experience is not proof. It's simply a record that "subjectively, recovery felt faster."

"Not Proven, Therefore Won't Do It" Isn't My Approach

This led me to reconsider my own decision-making framework.

The position of "I won't do it because there's no evidence" has its own rationality. But taken to its logical conclusion, you realize that a surprisingly large number of everyday health behaviors are "not strictly proven."

For example, things like these:

Choosing not to take fever reducers. Fever is the immune system intentionally raising body temperature to suppress viral replication. Fever up to around 38.5°C also promotes HSP production. "Unless it's so unbearable you can't sleep, it's better not to interfere with a mild fever for faster recovery" makes sense mechanistically, but a study comparing recovery speed between "groups that took antipyretics and those that didn't" is ethically difficult to conduct, and evidence is thin.

Immediate zinc intake. Zinc is essential for NK cell and T cell function, and several studies suggest taking it within 24 hours of cold symptom onset speeds recovery. Zinc ions have also been suggested to directly inhibit viral RNA replication. But dosage, form, and timing vary widely, and even meta-analyses reach split conclusions.

Prioritizing humidification. When airway mucosa dries out, ciliary movement decreases and the ability to clear viruses drops. Logically, "maintaining room humidity at 50-60% matters more than drinking lots of water" could be argued, but there's virtually no rigorous research on whether "humidification speeds cold recovery" — only epidemiological data showing infection rates rise in dry environments.

Sleep timing. Growth hormone and melatonin secretion peak during the first deep non-REM sleep after falling asleep and are strongly linked to T cell and NK cell activation. "Going to bed earlier to secure the first deep sleep is more immunologically beneficial than sleeping late into the morning" makes mechanistic sense, but virtually no trials have rigorously compared "what time to go to sleep when you have a cold for faster recovery."

What these all share is the same structure as contrast bathing. The mechanistic logic is solid, but clinical proof hasn't caught up.

My Decision-Making Criteria

When I organized my decision-making criteria for situations like these, they came down to three things:

Does the mechanism make logical sense?
Is the risk low?
Is the cost low?

If all three are met, I believe "try it and see" is sufficient. Waiting for perfect proof could take years, and doing nothing in the meantime is also a choice. But if the logic checks out and the risk is low, starting with practice and accumulating personal experience suits me better.

A 29% reduction in sick leave from the Dutch study, comparable to the 35% reduction from exercise — that's sufficient "rational basis to act." If you ask whether exercise being good for health has been "completely proven," there are actually similar research limitations. But most people continue exercising. Through logic, experience, and accumulation.

Not Dismissing "How It Feels"

In the Dutch study, 91% of participants wished to continue after the 30-day trial ended (64% actually continued).

Why? Probably because of an accumulation of subjective experiences like these: "I feel like my body got a little lighter," "I feel like my cold was milder," "I feel like my morning wakefulness changed."

Scientific rigor and personal experience operate on different layers. It's not that one is right and the other wrong. Both deserve to be held side by side as important.

"Not proven" and "having rational grounds for belief" are not contradictory.

Not being able to prove something and making an informed choice are different things.

What I found interesting was that I started out researching contrast bathing, but before I knew it, I was reexamining "how I make my own judgments."

Understanding the mechanism seems to increase the resolution of subjective experience. Rather than "it's vaguely good for you," having the understanding that "this mechanism should be at work" changes how you observe changes in your own body.

How do I face things where answers haven't been reached yet, things that haven't been proven?

What I thought was a conversation about health habits left me with that kind of question.

Related Books

For those who want to learn more about NK cells and immune mechanisms, the following book may be helpful.

[📩 敆擁ăƒȘンク: moshimo-card-uzskD]

References

1. Cold Shower RCT (Primary Study)

Buijze GA, Sierevelt IN, van der Heijden BCJM, Dijkgraaf MG, Frings-Dresen MHW.
"The Effect of Cold Showering on Health and Work: A Randomized Controlled Trial."
PLOS ONE 11(9): e0161749 (2016)

2. HSP70 and NK Cell Activation (Indirect Pathway: DC-NK Cooperation)

Qiao Y, Liu B, Li Z, et al.
"Activation of NK cells by extracellular heat shock protein 70 through induction of NKG2D ligands on dendritic cells."
Cancer Immunology Research (2008)

The previous version listed "Li Z et al." but the first author is Qiao Y. This is the foundational paper for the pathway where HSP70 induces NKG2D ligand (MICA) expression on dendritic cells (DCs), activating NK cells through the NKG2D receptor.

3. HSP70 and NK Cell Activation (Direct Pathway, Membrane-Bound)

Multhoff G.
"Activation of natural killer cells by heat shock protein 70."
International Journal of Hyperthermia 18:576–585 (2002); reprinted 2009

The paper previously cited as "Multhoff G et al. Dissecting the role of hyperthermia..." was actually by Dayanc BE, Beachy SH, Ostberg JR, Repasky EA (2008, same journal). For the representative paper on HSP70's direct effect on NK cells (cytotoxic activity via membrane-bound HSP70), please refer to the Multhoff paper above.

4. Thermal Stimulation and Immune Changes (Contrast Bathing Basics)

Brenner IK, Castellani JW, Gabaree C, Young AJ, Zamecnik J, Shephard RJ, Shek PN.
"Immune changes in humans during cold exposure: effects of prior heating and exercise."
Journal of Applied Physiology 87(2):699–710 (1999)

5. HSP70 Production Timing (Basic Cell Biology)

Morimoto RI et al.
"The role of heat shock factors in stress-induced transcription."
PMC (review)

6. Dutch Influenza Epidemic (2014/2015)

RIVM (National Institute for Public Health and the Environment, Netherlands)
"Flu season in the winter of 2014/2015 more severe and longer than preceding years"