Part one: what sweat is, and what it does

The glands

You have roughly two to four million eccrine sweat glands, the ones that do thermoregulation. Density is highest on palms and soles, around 250 to 550 per square centimetre, and two to five times lower on the face, trunk and limbs, which nonetheless do most of the work because they cover so much more area.

The total number is fixed by about age two or three. It does not increase. Most of the variation in how much people sweat comes from output per gland, not from having more glands.

There is a second type, apocrine, limited to the armpits, breasts, face, scalp and groin. These open into hair follicles rather than onto the skin surface, start working at puberty, and produce a thicker, lipid-rich secretion containing proteins, sugars and ammonia.

What is in it

Sweat is mostly water and salt. Sodium and chloride dominate by a wide margin. Below those, at lower millimolar concentrations: lactate, ethanol, urea, ammonia, bicarbonate and potassium. At micromolar levels: calcium, magnesium, iron, zinc, copper, vitamin C, glucose and individual amino acids. Cortisol and thiamine are nanomolar. Cytokines are picomolar.

So sweat is not "just salt water." That framing is inaccurate, and we are not going to use it. Things do leave the body in sweat.

Two details are worth having, because they explain a lot.

Sweat lactate is largely made by the gland itself. A tracer study found that lactate excreted in sweat derived from blood glucose, not from blood lactate. Sweat lactate typically exceeds blood lactate and falls as sweating rate rises.

Sweat glands cannot ramp up excretion the way kidneys can. Vasopressin, the hormone that lets kidneys concentrate urine, does not regulate sweat water loss. Plasma vasopressin does not correlate with sweat rate or sweat sodium, and drugs acting on that system have no effect on it. The duct does not meaningfully concentrate sweat by reclaiming water.

That last point matters more than it sounds. The kidney's superpower is that it can concentrate. A sweat gland cannot.

How much

Whole-body sweating runs from about 0.2 to 3 litres an hour, depending on person, activity and conditions. Light activity like walking produces roughly 0.3 to 0.6 L/h. Rates above 3 L/h happen in under 1% of athletes; the highest well-known figure is 3.7 L/h, recorded for a marathon runner in competition.

In a measured study of a commercial far-infrared sauna, whole-body sweat rate over a 45-minute session was 0.46 L/h, with participants losing about half a percent of body weight. That is a real number from a real device, and it is more modest than the experience suggests.

Where the skin ranks

The central comparison comes from the major 2019 review of sweat physiology by Lindsay Baker, who puts it directly: sweating's role in eliminating waste products and toxicants "seems to be minor compared with other avenues of excretion via the kidneys and gastrointestinal tract." Her reason is mechanical: sweat glands cannot adapt to excrete more, either by concentrating sweat or by raising the overall rate.

On mechanism she is blunter still. No known process allows sweat glands to preferentially concentrate BPA, persistent organic pollutants or trace metals in order to move them out of the body, which is why, as she puts it, direct evidence for sweating as an effective detoxification method is lacking.

For heavy metals specifically, that review states the primary route, based on tracer studies, is faecal. For BPA, a chemical central to the popular version of this story, 84 to 97% is eliminated in urine within 5 to 7 hours of exposure, and 100% within 24 hours.

One disclosure, because it is relevant: Baker is employed by the Gatorade Sports Science Institute, part of PepsiCo's R&D organisation. We note it because affiliations should be named. The direction of that interest does not obviously favour minimising sweat's importance: a sweat-science group has an interest in electrolyte replacement, not in arguing that sweat matters less.

Part two: the sweat-and-toxins research, examined

There is a body of published work reporting that metals and industrial chemicals appear in sweat, sometimes at higher concentrations than in urine. It is the evidentiary backbone of the entire claim, and it is worth understanding exactly what it is.

It is one study, not a series

The papers are usually cited as though they were multiple independent investigations. They are not. They draw on a single dataset (the "Blood, Urine and Sweat" study) of 20 participants, each of whom gave one blood sample, one urine sample and one sweat sample. The several published papers are re-analyses of those same specimens for different chemicals.

Half the participants were recruited from the first author's own clinical practice.

How the sweat was collected

From the methods, in the researchers' own description: participants collected perspiration from any site on their body, directly into a glass jar, either by holding the jar against pre-washed skin or by scraping it off with a stainless steel spatula. No specification was given as to how long sweating had been going on before collection. Sweat induction was self-selected: ten used an infrared sauna, seven a steam sauna, three exercise.

This is unsupervised, unstandardised, self-collected sampling, and it includes scraping, which matters most.

Why scraping matters

Four independent findings converge on the same problem.

Scraped sweat contains four to ten times more lipid than cleanly collected sweat. Lipophilic chemicals concentrate in lipid. If you scrape the skin's surface, you collect skin oils along with sweat, and skin oils are exactly where fat-soluble compounds would sit.

Skin cell contamination can inflate mineral readings two- to four-fold. The outer layer of skin holds substantial calcium; incorporated into a sample it can swamp the actual sweat concentration.

Some chemicals were reported in sweat while being absent from blood. Sweat derives from interstitial fluid, which derives from blood. A chemical that is not in the blood has no obvious route into sweat, which points toward the skin's surface as the source rather than the body's interior.

Two laboratories measuring the same thing differ by orders of magnitude. One group reported copper in sweat at 2.59 µg/L during strenuous exercise. The other reported 652.49 µg/L under sauna conditions. For lead, 0.63 versus 25.67 µg/L. That is roughly 250-fold for copper and 40-fold for lead. Method, not biology, is the likelier explanation.

The direct test, and it came back negative

A separate group had 50 people wear validated sweat patches for seven days alongside daily urine samples. The result: the person with the highest urine BPA in the study, 195 ng/mL after deliberate exposure, had no BPA detectable above background in their sweat patches.

What even the enthusiastic research concedes

The most important quotation in this entire topic comes from the researchers who produced the pro-sweat data themselves:

We cannot draw conclusions with respect to the rate at which induced perspiration will diminish total body PBDE burden as there is currently no means to accurately assess PBDE load in the spectrum of human tissues.

That sentence is about one class of flame retardants, but the limitation it describes applies across the whole series: there is no validated way to measure how much of any of these substances is stored in a living person. And from the same paper's limitations: the study did not assess health outcomes, and it was not possible to determine whether the measured perspiration was tainted by material originating from sebum or from skin tissue.

That is the crux of the whole subject: every study detects substances in sweat, and not one measures whether sweating reduces how much is in the body. The authors say so themselves.

The same group also published against their own hypothesis, which is to their credit and should be reported: for the PFAS family of chemicals, they found that sweating does not appear to facilitate clearance of the most common ones found in the human body.

The systematic review, and a problem with it

There is one systematic review on this question, from 2012, and it is routinely cited as independent validation. Its conclusion is favourable: that sweating "offers potential and deserves consideration."

Two things should be known about it.

Its own stated limitations are substantial. Most of the research it identified was over twenty years old; collection methods varied widely; sweat concentrations measured in research settings are "not well validated" and varied by body site, collection method and day-to-day factors. It specifically notes that lead in sweat may partly originate from material absorbed within the skin that pre-test cleaning did not remove. And it calls for "robust trials, appropriately sized to assess clinical outcomes," which is a way of saying they do not exist.

It is not independent of the work it reviews. The review and the primary sweat studies were published in the same journal, the same year, in the same special issue. We retrieved the guest editorial for that issue: the lead author of the primary studies identifies himself as lead editor of the issue, and the author of the systematic review is one of the guest editors. The editorial announces both papers. The two have also co-authored together elsewhere.

None of that makes the findings wrong. It does mean the review should not be described as independent corroboration, and we are not going to describe it that way.

So: does sweating remove anything meaningful?

On the available evidence, nobody knows, because the experiment has not been done.

The honest statement in the sceptical direction:

Small amounts of many substances (including some metals and industrial chemicals) can be measured in sweat. But no study has yet shown that sweating meaningfully lowers the amount of any of them stored in the body. The kidneys, liver and gut do that work, and sweat glands have no known machinery for selectively pulling contaminants out of tissue. Several of the most-cited studies used methods that could not separate sweat from skin oils and skin cells, and a careful study using validated patches found no BPA in sweat even in the person with the most in their urine.

And the honest statement in the other direction, which readers are equally owed:

This is not a settled question that has been put to rest. The relevant experiment (measure body burden, sweat people properly under controlled conditions, measure body burden again) has not been done. Researchers on both sides say it should be. Until then, "sweating detoxifies you" is an untested idea rather than a disproven one, and anyone who tells you science has definitively settled it either way is ahead of the evidence.

We are not going to tell you sweating eliminates nothing, because that is not true: substances measurably leave in sweat, and in one randomised crossover study pesticide metabolites were detected in 85% of pooled sweat samples. The honest argument is about magnitude and meaningfulness, not about presence.

Part three: saunas and health, a separate subject

This section is about a different question, and we want to be explicit about why it is separated.

There is a substantial body of research associating regular sauna use with better cardiovascular outcomes. None of it concerns detoxification.

That is not our interpretation. We checked. The most comprehensive review by the researchers who built that evidence base, a 25-page paper on the health effects of passive heat therapies, contains zero occurrences of "detox," "toxin," "heavy metal," "eliminate," "purify" or "cleanse." The only instance of "toxic" is the word cytotoxic, in an unrelated passage about oncology.

The mechanisms those researchers propose are cardiovascular and thermoregulatory: heat shock proteins, increased autophagy, reduced systemic inflammation, effects on mitochondrial capacity in muscle, and lowered blood pressure.

The single most likely way an article like this goes wrong is by letting the credibility of the sauna mortality data quietly transfer to a detox claim those studies never made. So: it does not transfer. These are two separate bodies of work with no bridge between them.

What the sauna research found

The core data come from a Finnish cohort, with baseline examinations between 1984 and 1989.

In 2,315 middle-aged Finnish men followed a median of 20.7 years, those using a sauna 4 to 7 times a week had an adjusted hazard ratio for sudden cardiac death of 0.37 (95% CI 0.18 to 0.75) compared with once a week, with a significant trend. By session length, over 19 minutes gave a hazard ratio of 0.48 (0.31 to 0.75) against under 11 minutes.

A second cohort of 1,688 men and women followed 15 years found a fully adjusted cardiovascular mortality hazard ratio of 0.23 (0.08 to 0.65) for 4 to 7 sessions weekly. The 2 to 3 sessions estimate, 0.75 (0.52 to 1.08), lost statistical significance on full adjustment.

Across the same research programme, frequent versus infrequent use gave all-cause mortality 0.84 (0.72 to 0.97) and cardiovascular mortality 0.74 (0.59 to 0.94).

How much weight that carries

These are observational cohort studies, and there are real reasons for caution.

They are Finnish, almost entirely men, recruited in the 1980s, in a culture where sauna use is near-universal. Sauna frequency plausibly tracks with income, housing, leisure time and social connectedness, and with being well enough to use a sauna several times a week. Reverse causation is a live concern: people in early undiagnosed illness use the sauna less.

The investigators adjusted for physical activity, socioeconomic status and existing heart disease, which helps. Adjustment is not randomisation.

In the same cohort, cardiorespiratory fitness was a stronger factor than sauna frequency, with a hazard ratio of 0.48 for fitness against 0.67 for sauna use.

Randomised evidence exists and is modest. A meta-analysis in heart failure found significant improvements in a cardiac biomarker, cardiothoracic ratio and ejection fraction, but no significant effect on heart dimensions or blood pressure, and the authors graded the strength of evidence from "moderate to insufficient." A review of regular dry sauna found that of 40 clinical studies covering 3,855 people, only 13 were randomised and most had fewer than 40 participants. A review of hot spring and sauna effects on blood lipids found reductions in adults under 60 (graded a weak recommendation) and no significant change in adults 60 and over.

And one useful corrective to "sauna is like exercise": a randomised crossover in 10 healthy women comparing rest, infrared sauna and cycling found no significant differences in blood pressure, arterial stiffness or heart rate variability between any of them, concluding the effects are thermoregulatory rather than exercise-mimetic.

Infrared specifically

You will often read that infrared penetrates 3 to 4 centimetres into tissue, reaching deeper than a traditional sauna.

That claim has now been tested directly, with a multisensor probe placed inside the quadriceps muscle of 10 adults during a 45-minute session in a commercially available far-infrared sauna. Muscle temperature rose 3.0 °C at 1.4 cm, 1.9 °C at 2.4 cm and 1.1 °C at 3.4 cm. Core temperature did not change at all (0.0 °C, p = 0.94).

The authors' conclusion: commercially available infrared saunas provide only superficial heating of peripheral tissues, with the thermal effect having lessened by 63% at 2.4 cm.

There is also no head-to-head outcome evidence that infrared is better than traditional sauna for anything. The long-term Finnish cohorts all used traditional Finnish saunas.

Sauna safety

Proportionate framing first: in Finland (roughly two million saunas for five million people), deaths occurring in a sauna ran at fewer than 2 per 100,000 people per year. This is not a dangerous activity.

That said:

Alcohol is the dominant risk. Half of Finnish sauna deaths involved alcohol, which magnifies the risk of low blood pressure, cardiac complications and accidents, and was involved in 40% of sauna-related burns. Do not use a sauna after drinking, and do not leave an intoxicated person alone in one. Sauna during a hangover is specifically flagged as carrying significant health concerns.

Burns are the more common real harm. One seven-year series, as reported in a review, attributed 26% of burn cases to saunas, usually from contact with the stove or steam.

Ask a clinician first if you have unstable coronary artery disease, unstable angina, recent heart attack, severe aortic stenosis, uncontrolled high blood pressure or severe valve disease. Pregnancy, older age, fever and skin infections also warrant a conversation. Rapid hot-to-cold plunging carries specific risk for people with existing cardiovascular conditions.

On salt, not "minerals." The real electrolyte consideration is sodium chloride loss, which matters for people whose sweat is unusually salty during prolonged heavy sweating. The same review that sets out sweat composition concludes that sweat-induced deficiencies appear to be of minimal risk for trace minerals and vitamins, so we are not going to suggest you need to replace trace minerals after a sauna, because the evidence says the opposite.

One more thing worth knowing

The regulators have been here, and the precedent is specific.

In 2009 the Federal Trade Commission charged the marketers of detox foot pads with deceptive advertising, alleging they claimed the pads would remove heavy metals, metabolic wastes, toxins, parasites, chemicals and cellulite from the body, and specifically that they had falsely claimed to have scientific proof that the pads removed toxic materials. A ban followed in 2010.

In a separate 2009 matter, a company settled for $3 million over health claims spanning numerous serious diseases; among the products, an infrared sauna was marketed to treat cancer. More than 5,700 refund cheques were later issued to consumers who had bought infrared saunas and supplements from that company.

And the National Center for Complementary and Integrative Health, part of the NIH, states plainly that there is no compelling research to support the use of "detox" diets for weight management or eliminating toxins from the body.

We mention this because it is the clearest available illustration that citing real papers does not convert an unsupported claim into a substantiated one, and not as a warning to anyone.

The short version

Sweat is mostly water and salt, plus small amounts of a great many things. You cannot concentrate it the way a kidney concentrates urine, and there is no known mechanism for selectively pulling contaminants into it. Substances do appear in sweat, but no study has measured whether sweating reduces the amount in the body, and the researchers most enthusiastic about the idea say so themselves. Much of the supporting data was collected by scraping skin, which picks up skin oils and cells along with sweat.

Saunas are a separate topic with real research behind them, none of which is about detoxification. Infrared heating is superficial, not deep. Alcohol is the main hazard.

Sitting in a hot room is pleasant and, in Finland at least, associated with living longer. That is a good enough reason on its own.

This article is for general education. It is not medical advice, and nothing here is intended to diagnose, treat, cure or prevent any disease. If you have a heart condition, high blood pressure, are pregnant, or take medication, speak with your doctor before regular sauna use.