What "70% Lower Risk" Really Means in the Finnish Sauna Study
The sauna headline is a hazard ratio from an observational study with wide error bars. How to read the number, what it can't prove, and a cautious way to start.
Some numbers sound too good to leave alone. "Seventy percent lower risk of dying from heart disease" is one of them, and it has followed the Finnish sauna studies around for years. The reflex I picked up from reading benchmarks is to ask what the baseline was and what exactly got measured. Applied here, those two questions change what the headline means, though not in a way that makes the finding useless.
This is not a retelling of the Finnish sauna story, which is already well covered elsewhere on this site. It is a closer look at how to read the number, so that you can decide for yourself how much weight it deserves.
What the study actually measured
The figures come from the Kuopio Ischaemic Heart Disease cohort, written up by Laukkanen and colleagues in BMC Medicine in 2018. It followed 1,688 people with a mean age of 63, for a median of 15 years. Over that time there were 181 deaths from cardiovascular disease. Participants reported how often they used a sauna, and the researchers compared three groups: once a week, two to three times a week, and four to seven times a week. Once a week was the reference.
The headline figures were hazard ratios. Adjusted only for age and sex, they were 0.71 for two to three sessions and 0.30 for four to seven. Adjusted for established risk factors, physical activity, socioeconomic status and later coronary heart disease, they were 0.75 and 0.23. The authors described the pattern as linear, with no threshold: more sessions, lower risk, all the way up.
What a hazard ratio is, and isn't
A hazard ratio compares rates, not totals. If you think of a long-running service, it is closer to "how often does this fail per hour of uptime" than "how many outages did we count." A hazard ratio of 0.30 says that at any point during follow-up, the four-to-seven group was dying of cardiovascular disease at about 30 percent of the rate of the once-a-week group. That is a 70 percent lower rate.
It does not say that 70 percent of cardiovascular deaths were prevented, and it does not say a person has a 70 percent chance of avoiding a heart attack. The ratio describes a gap between groups under the study's conditions. It tells you nothing, by itself, about how big the risk was to begin with.
Why 0.30 is not 0.23, and why neither is a point
The two numbers for the top group, 0.30 and 0.23, come from different levels of adjustment. The popular version usually quotes the first one. Neither is a precise value. Each comes with a confidence interval, and the intervals are wide: 0.14 to 0.64 for the age and sex model, 0.08 to 0.65 for the fully adjusted one. Read plainly, the second interval says the true effect could be anywhere from about a 92 percent lower rate to about a 35 percent lower rate.
The interval is wide because the top group had few events to count. Cardiovascular deaths are rare in any single group over fifteen years, and when a handful of events drives the estimate, a few more or fewer shift it a lot. It also matters that for the two-to-three group, the fully adjusted interval (0.52 to 1.08) crosses 1. After full adjustment, that middle group is not clearly different from the once-a-week reference.
Relative risk is not absolute risk
Across the whole cohort, 181 of 1,688 people died of cardiovascular causes over the follow-up, roughly 11 in 100. That is an overall figure, not the rate in any one group. To show how the arithmetic works, imagine the reference group had a risk of 10 in 100. A hazard ratio of 0.30 would then correspond to something near 3 in 100. This is illustrative only, since a hazard ratio is not exactly a ratio of risks, but it shows that a large relative change applied to a small base is a modest absolute change.
For an individual, the base is what matters. A 63-year-old with several risk factors sits on a very different base from a healthy 45-year-old, and the same ratio lands in different places.
What an observational design can't tell you
This was an observational study. Nobody was assigned to sit in a sauna. People chose their own habits, and the people who sauna four to seven times a week may differ in ways the questionnaires did not capture. In Finland the sauna is woven into ordinary life, and a habit that frequent can go along with living near a sauna, having the time and a settled routine, feeling well enough to go, and often going with others. Statistical adjustment helps, but it cannot remove what was never measured.
Reverse causation is the other worry. Someone who feels unwell may stop going, so a low sauna frequency can be partly a sign of poorer health rather than a cause of it. The authors adjusted for many things and their conclusions are measured, but the study design cannot prove that the sauna itself lowers risk.
Why a Finnish cohort may not transfer to you
The people in this study used traditional Finnish saunas, and mostly as an established habit. I am inferring that last part, since the paper reports frequency and not how long people had been going. Someone who starts in 2026 with no history of regular heat exposure is a different case from a person who has done it since childhood.
Infrared saunas are a separate question. This cohort studied traditional saunas, so these figures say nothing directly about infrared, and I would not assume the result carries over.
Who should be careful
Heat is a stress on the cardiovascular system, which is part of the proposed reason it might help. It is also why some people should talk to a clinician first. That includes anyone taking blood pressure medication, anyone with a heart condition or unstable blood pressure, and anyone who is pregnant. Alcohol and a sauna are a poor combination. Dehydration is the ordinary risk for everyone else.
A cautious way to begin
This is my own conservative approximation, not the protocol used in the study, and it is no substitute for advice from a clinician who knows your health.
- Frequency: one or two sessions a week to begin, and add more only if it feels easy.
- Time: ten minutes or less per session at the start.
- Heat: choose the cooler end of what the sauna offers, and sit on a lower bench.
- Water: drink before and after, and keep a bottle nearby.
- Alcohol: none before or during.
- Stop rules: leave at once if you feel dizzy, nauseated or have chest discomfort, and cool down and sit for a few minutes before you stand.
FAQ
Does a hazard ratio of 0.30 mean sauna cuts my risk by 70 percent?
No. It means that in this cohort the group using the sauna four to seven times a week had a rate of cardiovascular death about 30 percent of the once-a-week group's rate. It does not tell you what your personal risk is, or that the sauna caused the difference.
Which number should I trust, 0.30 or 0.23?
Both come from the same study at different levels of adjustment, and both have wide intervals. The honest reading is that the study found a lower rate in the most frequent users, and that the exact size is uncertain.
Is two or three times a week enough?
The age and sex model showed a lower rate for that group. In the fully adjusted model the interval included no difference. So the evidence for that middle group is weaker than for the top one.
Do I need a Finnish-style sauna?
The study used traditional saunas, so that is what the evidence covers. Whether other types behave the same way is not something this data answers.
Is it safe to start if I take blood pressure medication?
Ask your clinician first. Heat can affect blood pressure, and some medications change how your body handles it. It is a quick conversation and worth having before the first session.