Sauna: 201 Finnish Men and a Very Long Sweat

Four to seven times a week, or you are doing it wrong

If you have spent any time in the longevity corner of the internet this year, you have been handed a number. Sauna, four to seven times a week, roughly twenty minutes at around 80°C. The number is attached to a claim: Finnish men who did this had about 40% lower all-cause mortality, and about two-thirds lower dementia risk. It is repeated on podcasts, in newsletters and in the marketing copy for home cabins that cost as much as a used car.^1,2

It is a good story, and some of it is true. Heat exposure does real things to the body. The cohort data are consistent. The mortality curve is real. What I want to test is whether the story survives contact with the details, because the details are where the 40% lives. The group behind the headline number is small, it comes from one region of one country, and the randomized trials that should confirm it are lukewarm at best.

What heat actually does

A sauna heats the skin quickly and pushes core temperature up. Heart rate rises substantially. Blood vessels in the skin dilate, blood pressure dips by the end of the session, and plasma volume falls as you sweat. In a recent physiology study of 12 adults in a far-infrared cabin at 65°C, rectal temperature rose 1.4°C, heart rate went from 74 to 153, mean arterial pressure dropped by 9 mmHg and plasma volume fell 11.6% over 45 minutes. The sauna was supplied by a manufacturer.²⁴

The hypothesis follows naturally. Repeated heat stress looks a bit like mild exercise to the cardiovascular system, so perhaps it improves endothelial function, lowers blood pressure and reduces arterial stiffness. That is plausible biology. It is also a hypothesis, and the clean way to test it is to randomize people to heat or no heat and measure what happens. We will get to that, because the answer is less flattering than the cohort data.

The Kuopio Ischaemic Heart Disease study

Almost every hard-endpoint claim about sauna traces back to one place: the Kuopio Ischaemic Heart Disease Risk Factor Study (KIHD) in eastern Finland. It enrolled 2,315 men aged 42 to 60 between 1984 and 1989, asked them how often they used a sauna, and followed them for a median of 20.7 years.¹

The same questionnaire has since been mined for a long list of outcomes. In the same men, 4 to 7 sessions per week was linked to a hazard ratio of 0.34 for dementia (0.16 to 0.71) and 0.35 for Alzheimer's disease.² Among 1,621 of the men who did not have high blood pressure at the start, it was linked to a hazard ratio of about 0.54 for new hypertension.⁴ In a separate KIHD sample of 1,628 men and women aged 53 to 74, it was linked to 0.39 for stroke,⁵ and to 0.23 for fatal cardiovascular disease in a largely overlapping group of 1,688 men and women from the same examination.⁶

I would be more impressed if I did not notice that the same single question protects against nearly everything. When one lifestyle exposure is linked to a dozen unrelated diseases, the more economical explanation is that it marks a type of person rather than a type of therapy.

Who has a sauna in 1985?

Observational data cannot tell you whether sauna produced the benefit or whether the kind of person who can sauna four to seven times a week was already healthier. Several problems stack up here.

First, reverse causation. People who are becoming ill stop going to the sauna. A second Finnish study followed 13,994 men and women for 39 years. Those using the sauna 9 to 12 times a month had a dementia hazard ratio of 0.47 in the first 20 years, but only 0.81 across the whole follow-up.³ That fits what you would expect if early decline drives people out of the sauna, though the authors read their results as consistent with protection, and the study did not test reverse causation.

Second, confounding. Home saunas in Finland track with rural living, housing and income, and frequent bathers differ in ways a questionnaire misses. Two letters published after the 2015 paper argued the association may be noncausal for these reasons.⁸ The authors adjusted for fitness, physical activity and socioeconomic status, which helps, but adjustment on self-reported covariates is a blunt instrument.

Third, the exposure itself. These are Finnish men in 80 to 100°C dry rooms, usually followed by a cool-down. Whatever the data say about that, they say nothing directly about a 55°C infrared cabin, a steam room or a heated blanket.

What happens when you randomize heat

Here the story cools. The most useful recent synthesis pooled 20 randomized trials of passive heating lasting 2 to 15 weeks. It found no significant effect on flow-mediated dilation, pulse wave velocity, heart rate variability, glucose, HbA1c, lipids or C-reactive protein. Systolic blood pressure fell by 2.46 mmHg, but the confidence interval ran from -5.02 to +0.10, and heterogeneity was high.¹¹ One author discloses grants and honoraria from supplement and food-industry companies, which is not relevant to the finding but is relevant to a newsletter that counts conflicts.

In heart failure there is a more favorable signal. A Japanese multicenter trial of far-infrared sauna at 60°C for 15 minutes daily over 10 days in 149 patients missed its primary endpoint (BNP) but improved NYHA class and 6-minute walk distance.⁹ A meta-analysis found short-term improvement in BNP and ejection fraction, with GRADE certainty from moderate to insufficient.¹⁰ That is a supervised clinical therapy in sick patients, not a lifestyle habit in healthy people. A broader systematic review found 40 studies and 3,855 participants, of which only 13 were randomized and most had fewer than 40 people.¹⁴ I found no Cochrane review. Nothing randomized touches mortality or dementia.

What this means for the stratum corneum

Since this is a dermatology newsletter, I looked for the skin data. There is almost none. The one human study that keeps getting cited compared 41 people, regular sauna users versus non-users. Users had a more stable barrier, better stratum corneum hydration, faster recovery of transepidermal water loss and pH, and less forehead sebum.²⁰ It was not randomized, and regular sauna users differ from non-users in many ways. I could not find any trial on wrinkles, photoaging or skin firmness, and no controlled psoriasis trial.

On the other side of the ledger, heat is a recognized trigger in atopic dermatitis. In a survey of 587 adults with the condition, sweat was the second most common itch trigger at 30.5%, and heat was reported by 24.0%.²¹ If you have eczema, rosacea or heat-sensitive melasma, I would treat the sauna as a potential flare, not a spa treatment. That last point is clinical judgment rather than trial evidence, and I will say so plainly.

Mostly safe, with real exceptions

For healthy adults who stay hydrated, skip alcohol and do not bathe alone, regular sauna use appears low risk. That is a real finding and worth saying clearly. The harm profile is not the problem. The sales pitch is.

The $8,000 cabin is not the same room

Home sauna marketing leans on the Finnish mortality data and then sells something else. Infrared cabins from the premium makers run from roughly $7,500 to more than $9,000 on retailer listings, and infrared blankets sell for around $700. They operate at 45 to 65°C rather than 80 to 100°C, so the cohort evidence does not transfer cleanly.

The FDA regulates infrared lamps as Class II devices, cleared for temporary relief of minor muscle and joint pain and local circulation. In 2022 it sent a warning letter to a far-infrared sauna seller for cancer, kidney detox and blood-pressure claims that the clearance did not cover.²² The word "detox" is doing a lot of work in this market, and sweat is not a validated route for clearing meaningful toxin loads.

The influencer dose gets the same treatment. The widely quoted "57 minutes a week" figure comes from a 2021 Cell Reports Medicine study, where it described the self-reported sauna habits of 8 winter-swimming men (7 analysed) compared with 8 controls in a study about brown fat. It was not a sauna dosing trial.²³ It cannot support a protocol.

Real, modest, and not yet proven