Sweat, Death, and the 95% Your Blood Test Cannot See
Men who sweated four to seven times a week died at a rate 37 percent lower than men who sweated once a week. Lead exposure is linked to 412,000 American deaths annually — ten times the official estimate. And the blood test your doctor uses to rule out heavy metal toxicity is measuring about five percent of what is actually in you
Three bodies of research. Almost never cited together. Here is what happens when you put them in the same room.
The Finnish Anomaly
Start with the death data, because that is where the argument has to begin.
In 1984, researchers at the University of Eastern Finland enrolled 2,315 middle-aged men in what became the Kuopio Ischaemic Heart Disease Risk Factor Study.
Among the many variables recorded at baseline was one that no American cohort study would have thought to ask: how often do you use the sauna, and for how long?
They followed these men for a median of 20.7 years. By the end, 929 of them were dead.
When Laukkanen and colleagues published the analysis in JAMA Internal Medicine in 2015, the numbers were not subtle. Compared with men who used the sauna once a week, men who used it four to seven times a week had a 63 percent lower rate of sudden cardiac death (HR 0.37, 95 percent CI 0.18–0.75) and a 50 percent lower rate of fatal cardiovascular disease (HR 0.50, 95 percent CI 0.33–0.77).
Look at the absolute figures, which are more honest than hazard ratios. Over the follow-up period, all-cause mortality was 49.1 percent in the once-weekly group, 37.8 percent in the two-to-three-times group, and 30.8 percent in the four-to-seven-times group.
Roughly half the infrequent sweaters were dead. Roughly a third of the frequent sweaters were dead.
And there was a second dose-response gradient, independent of the first. Compared with sessions under 11 minutes, sessions longer than 19 minutes carried a hazard ratio of 0.48 for sudden cardiac death. Not just how often. How long.
The pattern did not stop at the heart. In 2017, the same cohort was analyzed for neurocognitive outcomes in Age and Ageing: four-to-seven weekly sessions were associated with a hazard ratio of 0.34 for dementia and 0.35 for Alzheimer’s disease, and the associations held after adjustment for physical activity and socioeconomic status.
A 2018 extension in BMC Medicine followed 1,688 participants — this time including women — and found cardiovascular mortality risk declining linearly with sauna frequency, with no threshold effect.
Now hold that thought, because the next study is the one that breaks the easy explanation.
What the Easy Explanation Misses
The standard reading of the Finnish data is straightforward and mostly correct: sauna is cardiovascular exercise you can do sitting down. Heart rate climbs to 100–150 bpm. Blood pressure falls.
Arterial compliance improves. Heat shock proteins are induced. Vagal tone rises. C-reactive protein drops. These are real mechanisms and they are well documented.
But notice what the Laukkanen model already controlled for: systolic blood pressure, LDL cholesterol, smoking, alcohol, prior myocardial infarction, type 2 diabetes, resting heart rate, socioeconomic status — and cardiorespiratory fitness, and physical activity.
The effect survived all of it.
Then in 2018, Kunutsor and colleagues published the study that makes the point unavoidable. Following 2,277 men for a median of 26.1 years across 1,124 deaths, they crossed objectively measured cardiorespiratory fitness with sauna frequency:
Read the bottom two rows carefully. Being fit was powerful. Being fit and sweating frequently was more powerful still. Sauna added protection on top of measured aerobic fitness — which means whatever sauna is doing, aerobic conditioning does not fully account for it.
So here is the question that the cardiology literature does not ask, and that I want to put on the table: what if part of what a lifetime of sustained sweating does is not add something — but subtract something?
The Thing That Kills You That Nobody Is Measuring
In March 2018, Bruce Lanphear and colleagues published an analysis in The Lancet Public Health that should have reorganized preventive medicine and did not.
They took 14,289 American adults from NHANES-III, measured blood lead at baseline, and followed them for a median of 19.3 years. Comparing the 90th percentile of blood lead (6.7 µg/dL) with the 10th (1.0 µg/dL), they found:
- All-cause mortality: HR 1.37 (95 percent CI 1.17–1.60)
- Cardiovascular mortality: HR 1.70 (1.30–2.22)
- Ischaemic heart disease mortality: HR 2.08 (1.52–2.85)
The population attributable fraction of blood lead for all-cause mortality was 18.0 percent — equivalent to 412,000 deaths per year in the United States, including an estimated 256,000 premature cardiovascular deaths.
That figure is ten times the estimate then carried by the Institute for Health Metrics and Evaluation. The reason for the discrepancy is the entire point: previous estimates assumed a threshold below which lead does nothing.
Lanphear found no threshold. The risk curve rose most steeply at the lowest levels of exposure. And the association held when the analysis was restricted to people below 5 µg/dL — the levels a physician would glance at and call normal.
In his own words: there is no safe level.
Sit with the magnitude. If this analysis is even approximately right, low-level lead is a mortality risk factor on the order of tobacco, operating invisibly, in a population that has been told the problem was solved when they took it out of gasoline.
And now the part that almost nobody has followed through on.
The 95 Percent Problem
Lanphear measured blood lead.
Blood lead has a half-life in adults of roughly 28 to 36 days. It is a snapshot of the last month. It is also, structurally, a snapshot of a tiny minority compartment.
Per the ATSDR’s own toxicological profile and the foundational work of Barry and Mossman, more than 95 percent of the total lead burden in an adult human is stored in the skeleton, where its half-life is measured not in days but in years to decades — up to 25 to 30 years in cortical bone.
Restate that plainly: the standard clinical test for lead exposure measures a compartment holding roughly one-twentieth of your lead, with a residence time roughly one three-hundredth as long as the compartment holding the rest.
Bone is not an inert vault, either. Skeletal lead is remobilized into circulation during pregnancy, lactation, menopause, calcium deficiency, and the accelerating bone turnover of aging. You are, in late middle age, being re-dosed from the inside by exposures you sustained decades earlier.
So what happens if you measure the real burden instead of the snapshot?
In 2024, researchers publishing in Science of the Total Environment did exactly that. Using a validated algorithm to estimate bone lead from blood lead and covariates, they re-analyzed 11,628 NHANES-III adults over a median 26.8 years of follow-up.
Comparing 90th versus 10th percentiles for cardiovascular mortality:
- Blood lead: HR 1.63 (1.25–2.14)
- Patella (trabecular) bone lead: HR 2.42 (1.56–3.76)
- Tibia (cortical) bone lead: HR 3.32 (1.93–5.73)
The paper’s title asks the question directly: Do we underestimate risk of cardiovascular mortality due to lead exposure?
When you measure what is actually stored in the body rather than what is briefly circulating, the mortality signal roughly doubles.
This is not a fringe claim from an integrative medicine conference. This is mainstream environmental epidemiology, published in a Q1 journal, using the same national dataset that generated the 412,000 figure.
And its implication is devastating for standard practice: a “normal” heavy metal panel is not evidence of a low body burden. It is evidence that your last month was unremarkable.
The instrument has been hiding the effect. And the effect is death.
Sweat as a Second Window
Which brings us, finally, to perspiration — and I want to be precise about what the sweat literature does and does not establish, because the precision is what makes it powerful.
The foundational work is the Blood, Urine, and Sweat (BUS) series by Stephen Genuis and colleagues. In the 2011 paper in Archives of Environmental Contamination and Toxicology, 20 participants — 10 healthy, 10 with chronic illness — had all three fluids analyzed for toxic elements.
The finding:
“Many toxic elements appeared to be preferentially excreted through sweat… Presumably stored in tissues, some toxic elements readily identified in the perspiration of some participants were not found in their serum.”
And the conclusion that matters most:
“Biomonitoring for toxic elements through blood and/or urine testing may underestimate the total body burden of such toxicants.”
This was not an isolated result. The pattern repeats across the compound classes:
- Bisphenol A (2012): BPA appeared in sweat in some individuals who had no detectable BPA in either serum or urine.
- Phthalates (2012): MEHP was found at roughly twice the concentration in sweat as in urine, and in several participants appeared in sweat while absent from serum — consistent with tissue retention.
- Organochlorine pesticides (2016): DDT/DDE were detected in nearly all participants and endosulfan in over half. Sweat showed more frequent detection and higher levels than blood or urine. Critically, there was little correlation between the three compartments in individuals.
- PBDE flame retardants (2017): the congeners were not found in urine at all. 100 percent of participants tested positive for BDE-47 in blood or sweat. The authors concluded that blood testing alone provides only a partial picture.
Independently, the 2012 systematic review by Sears, Kerr and Bray in the Journal of Environmental and Public Health pooled 24 studies and reported that in individuals with higher exposure or body burden, sweat concentrations generally exceeded plasma or urine, and dermal excretion could match or surpass 24-hour urinary excretion.
Arsenic dermal excretion ran severalfold higher in exposed individuals than controls. Cadmium was more concentrated in sweat than in blood plasma.
And there is a beautiful old experiment that anchors the mechanism. In 1988, Lilley and colleagues applied radioisotope-labeled lead to workers’ skin. Sweat lead rose. Blood and urine did not move.
Their conclusion: strong evidence that the skin is itself capable of excreting certain heavy metals.
Note what this establishes and what it does not. The biomonitoring claim is now well supported: sweat sees things blood and urine miss. That claim alone indicts standard practice, because it means the reassurance patients receive from a clean serum panel is, at minimum, incomplete — and at worst, an artifact of looking in the wrong compartment.
The 10x Finding: Why How You Sweat Matters
If sweat is a route of elimination, the obvious next question is whether all sweat is created equal. In 2022, Kuan, Chen and Liu answered it in the International Journal of Environmental Research and Public Health, and the result deserves far more attention than it received.
Twelve healthy young adults, dietary intake controlled for a week beforehand, each tested under two conditions in counterbalanced order: 20 minutes on a treadmill at 25°C, versus 20 minutes seated in a 45°C sauna cabinet.
Humidity matched. Sweat rates matched at approximately 1.8 mg/cm²/min, so this is not a dilution artifact. Rigorous skin pre-cleaning. Paired t-tests.
Sweat concentrations, in µg/L:
Exercise-induced sweat carried roughly eleven times the lead and eleven times the nickel of passively induced sweat at the same sweat rate.
The proposed mechanism is elegant. Dynamic exercise raises core temperature and drives sympathetic cholinergic activation of the secretory coil while dramatically increasing perfusion — sweat contents are expelled from blood directly through active secretion.
Passive heat above skin temperature produces mostly local cutaneous vasodilation and a dilute, hypotonic sweat optimized for evaporative cooling and very little else.
Now look at the mercury row, because I think it is the most important line in the table.
Mercury did not move under either condition. And there is a specific physiological reason: roughly 90% of body mercury is sequestered inside red blood cells as methylmercury; only about 10% is inorganic, and only half of that resides in plasma.
There is almost nothing available at the secretory interface for the gland to move.
Contamination does not discriminate by protein binding. If these findings were simply skin surface residue, sebum, and epidermal debris — the standard critique — mercury would be no more exempt than lead.
The selectivity is the signature of a real transport phenomenon operating on the accessible fraction. Which is also why the next section is necessary.
What Sweating Does Not Do
Any framework that explains everything explains nothing. So here is where the evidence says sweat fails.
PFAS. The “forever chemicals” are the defining toxicological problem of this decade, and induced perspiration does not touch them. Genuis and colleagues tested this directly in 2013: none of the perfluorinated compounds found in serum appeared to be efficiently excreted into perspiration.
Not PFOS. Not PFOA. Not PFHxS. The ATSDR toxicological profile states it flatly: perfluoroalkyls do not appear to be eliminated in sweat, and induction of perspiration by exercise or sauna does not alter their clearance.
The same study found only some PCB congeners released into sweat, not all.
For PFAS the evidence points somewhere else entirely — toward bile acid sequestrants such as cholestyramine and colesevelam, which have now been replicated across multiple trials including the Ronneby cohort in Sweden, and toward therapeutic phlebotomy, which Genuis’s group documented lowering serum PFAA concentrations over four to five years in a highly exposed family.
Different toxicant, different exit. Anyone selling you a sauna as a universal solvent is not reading the papers.
The Study Nobody Will Fund
Here is the trial that would settle it, and it is not complicated.
Take a cohort with measurable burden. Establish baseline with K-shell X-ray fluorescence — a non-invasive 30-minute scan of the shin or kneecap that reads the lead deposited in bone itself, integrating decades of exposure rather than the past month.
It is the gold standard for cumulative dose, it exists in a handful of research labs, and it has never been commercialized as a clinical test.
The instrumentation exists. The endpoints are validated. The population is everyone.
It has not been done, and the reason is structural rather than scientific. There is no molecule to patent at the end of it. The intervention costs nothing and can be performed in a gym, a bathtub, or a Finnish shed.
No sponsor recovers a return, so no sponsor writes the check, so the evidence base stays permanently “insufficient” — and the insufficiency is then cited as though it were a finding.
This is the same architecture that lets a system generate 412,000 attributable deaths a year while confidently reassuring patients on the basis of a test that examines five percent of the relevant compartment.
The blind spot is not an accident of science. It is a consequence of who pays for science. The obvious alternative — identify the poisons, then remove them — is not on the table, because nobody owns it.
See more here substack.com