In March 2008, a chiropractor in Florida noticed that several of his patients had the same complaints: stomach trouble, and hair coming out. They had all bought the same liquid supplement from his office. Some of them had doubled the dose in response, and then got worse.
The bottle said it contained 200 micrograms of selenium per fluid ounce. When the FDA tested it, the product held 40,800 micrograms per ounce, about 200 times the label.
Selenium is a mineral your body genuinely needs. The federal recommended intake for an adult is 55 micrograms a day and the federal ceiling is 400, a gap that sounds roomy until two things complicate it. A single ounce of Brazil nuts averages 544 micrograms. And two respected agencies cannot agree on where the ceiling belongs.
The margin, in two numbers
The NIH Office of Dietary Supplements publishes the federal nutrient fact sheets. Its selenium page, updated 4 September 2025, gives two figures that answer different questions.
One is the Recommended Dietary Allowance, or RDA, the target intake; article 322 covers where that value comes from and what it is for. For men and women it is 55 micrograms a day at 19 to 50 and 55 again at 51 and older. Pregnancy raises it to 60 micrograms and lactation to 70. The Food and Nutrition Board at the National Academies set them, and one oddity is worth flagging: no year is attached to any figure in the published table.
Its companion ceiling, the Tolerable Upper Intake Level or UL, stands at 400 micrograms a day for adults 19 and older, pregnancy and lactation included. The Office of Dietary Supplements states these limits cover selenium from food and supplements together, and that they were set from the amounts associated with hair and nail brittleness and loss. One exception is spelled out: the limits do not apply to people taking selenium under the care of a physician.
Divide the ceiling by the recommended intake and you get a ratio of roughly 7.3. That is worth stating carefully, because selenium is often described as the mineral with the tightest safety margin and the ratio does not show that. Zinc’s adult male ratio, 40 mg over 11 mg, is about 3.6, which is tighter. What makes selenium unusual is not the ratio. It is that an ordinary food clears the ceiling in one serving, that the ceiling itself is disputed, and that the United States has a documented mass poisoning to show what the top of the range does. Article 323 covers what upper limits are and how they get built.
The Daily Value behind the percentages on a Supplement Facts panel, set by FDA in its food labeling regulation at 21 CFR 101.9, is 55 micrograms for adults and children 4 and older.
Who is short of selenium
Before the harms, the base rate. The Office of Dietary Supplements states that selenium deficiency is very rare in the United States and Canada.
Selenium’s job explains why it matters. It is built into 25 selenoproteins, including the glutathione peroxidases and thioredoxin reductases, and it is heavily involved in thyroid hormone metabolism, including conversion of the storage hormone T4 into the active form T3.
Soil selenium varies enormously by region, and plant foods reflect the soil they grew in. The reason that does not produce American deficiency is logistics: the fact sheet notes that extensive transport of food across the country lets people in low-selenium areas get enough anyway. Serum selenium still runs higher in Midwestern and Western residents than in the South and Northeast.
The groups the fact sheet names as at risk are specific: people living in selenium-deficient regions, people on kidney dialysis, and people living with HIV. The clearest deficiency disease, Keshan disease, is an endemic cardiomyopathy identified in 1935 in parts of China with low-selenium soil. Article 321 covers how deficiency gets measured.
Two upper limits, both official, 145 micrograms apart
In January 2023, the European Food Safety Authority’s panel on nutrition published its own review of selenium and arrived at a different ceiling.
The panel picked alopecia, meaning hair loss, as the critical endpoint, calling it an early observable feature and a well-established effect of selenium excess. It identified a lowest-observed-adverse-effect level of 330 micrograms a day from a large randomized human trial, applied an uncertainty factor of 1.3, and concluded: “A UL of 255 μg/day is established for adult men and women (including pregnant and lactating women).”
So there are two upper limits in circulation for the same mineral and the same adults. The U.S. Food and Nutrition Board says 400 micrograms a day. The European panel says 255. The difference is 145 micrograms, which is more than two and a half times the entire recommended daily intake.
The strange part is that the Office of Dietary Supplements prints the European figure on its own page, alongside its own, without reconciling the two or explaining which a reader should use.
Neither number is wrong. They came from different endpoints and different reasoning, which is what happens when two expert committees read the same literature. But the margin changes. Against 400 micrograms, the recommended intake fits about 7.3 times. Against 255, about 4.6. The European panel also flagged who was most likely to pass its limit: regular users of high-dose selenium supplements, and regular eaters of Brazil nuts.
One trial doing double duty
The large randomized trial the European panel used was SELECT, the Selenium and Vitamin E Cancer Prevention Trial.
Article 326 sets out how large SELECT was, where it ran and over what window it enrolled, and uses it as a general lesson about evidence. What this article needs is the selenium arm. To enroll, a man had to be 50 or older if African American or 55 or older otherwise, with a prostate-specific antigen level of 4 ng/mL or less. The selenium dose was 200 micrograms a day as L-selenomethionine.
The 2008 primary report, by Scott Lippman and colleagues in JAMA, found no cancer prevention benefit. At a median follow-up of 5.46 years as of 23 October 2008, the hazard ratio for prostate cancer in the selenium group was 1.04, with a 99 percent confidence interval from 0.87 to 1.24. Longer follow-up through July 2011, reported by Eric Klein and colleagues, put selenium’s prostate cancer hazard ratio at 1.09, interval 0.93 to 1.27, an absolute increase of 0.8 cases per 1,000 person-years against placebo.
Buried in the 2008 report is something the federal fact sheet does not carry at all.
The paper reports a statistically nonsignificant increased risk of type 2 diabetes in the selenium group: a relative risk of 1.07, with a 99 percent confidence interval from 0.94 to 1.22, and a P value of .16. That P value matters. By the usual conventions.16 is well inside what chance could produce. It is a signal, not a finding, and reporting it as proof that selenium causes diabetes is overreading it.
But the Office of Dietary Supplements selenium page does not mention diabetes anywhere. Not as a finding, not as a nonfinding, not as a signal worth watching. A reader relying only on the federal fact sheet would never know the question had been asked.
Now the part worth naming out loud. The European panel derived its 255-microgram upper limit from the hair-loss data in SELECT. The same trial that produced the unresolved diabetes signal is the evidentiary base for one of the two competing safety ceilings. One trial, two jobs. That is not misconduct; large well-run trials get reused for exactly this reason. It is simply a narrower foundation than two independent numbers would suggest.
A manufacturing error, and what it did to hundreds of people
The 2008 outbreak was documented by Jennifer MacFarquhar and colleagues from the CDC and state health departments, in the Archives of Internal Medicine in 2010, and it is the most detailed account of selenium poisoning in the American record.
The mechanics are mundane and that is the point. The product was a liquid supplement sold for the whole family, blending 16 vitamins, 12 elements, 58 trace elements, 18 amino acids and other ingredients. The label listed 200 micrograms of selenium per fluid ounce as sodium selenite. FDA’s testing found 40,800 micrograms per ounce. Chromium was also elevated, at 17 times the labeled amount, though investigators judged that below the level known to cause harm.
The cause was employee error at one of the ingredient suppliers. The distributing company had recently changed manufacturers, and the misformulated lots were the first produced after the change. The product had been on the market 12 years with no reported problems. The lots came to 1,200 bottles, made in September 2007, with distribution starting that January and a recall in late March 2008. On 14 April 2008, the CDC issued a Health Alert Network notice naming the two products involved and confirming FDA’s findings on selenium and chromium content.
The investigators’ closing judgment, quoted exactly: “Had the manufacturers been held to standards used in the pharmaceutical industry, this outbreak may have been prevented.”
The paper contradicts itself
Read the abstract and the body of the same paper and you get different numbers. Almost every secondary account picks one set without saying so.
The abstract reports 201 cases identified in 10 states, a median estimated dose of 41,749 micrograms a day, and a mean initial serum selenium concentration of 751 micrograms per liter in eight patients. Its symptom percentages are fatigue 75 percent, hair loss 72 percent, joint pain 70 percent and nausea 58 percent.
The body reports a questionnaire given to 227 affected people identified in 9 states, which it names: Florida, Georgia, Kentucky, Michigan, North Carolina, Pennsylvania, Tennessee, Texas and Virginia. Of those 227, 201, or 89 percent, met the case definition, which resolves the 201-versus-227 difference. The state count does not resolve. Neither does the dose: the body gives a median of 41,585 micrograms a day among the 156 patients with usable data, range 3,400 to 244,800. Its mean serum figure is 761 micrograms per liter, median 664, range 321 to 1,500, against a reference mean of 125. Its symptom percentages run slightly lower: fatigue 72 percent, hair loss 70 percent, joint pain 67 percent, nausea 57 percent, with diarrhea highest at 78 percent.
Both sets are published. The body’s account is the more granular, and it is the one that names the states. Report either, but know which you are using.
What it did to people
The affected group had a median age of 54, with a range from 4 to 92, and 60 percent were women. They had been taking the product for a median of 29 days, with a range from 1 to 109. One person was hospitalized. Nobody died.
Scalp hair loss ranged from 10 percent to 100 percent of the head, with a median of 50 percent. Of 78 patients reporting hair loss, 14 lost all the hair on their scalp and one lost all body hair. Of 74 reporting fingernail discoloration or brittleness, 28 lost nails.
Ninety days or more after symptom onset, more than half still had fingernail discoloration and loss. Fatigue persisted in 35 percent, hair loss in 29 percent, joint pain in 26 percent, memory loss and muscle aches in 22 percent each. Investigators also recorded fever in about 23 percent, a symptom they noted had not previously been linked to selenium toxicity.
Two details deserve to travel with the numbers. When investigators reached them, 41 percent of affected people had not brought their symptoms to any medical attention. And patients often never told their doctors about the supplement, because it did not occur to them that a health product was the cause. Some raised the dose, trying to feel better.
What selenosis looks like
Selenosis is the clinical name for the effects of chronically high selenium intake, and the Office of Dietary Supplements describes it the same way the outbreak played out.
It is most commonly characterized by hair loss and nail brittleness or loss. The early indicators are a garlic odor in the breath and a metallic taste in the mouth. Other signs include skin rash, nausea, diarrhea, fatigue, irritability and nervous system abnormalities. Acute selenium toxicity can cause severe gastrointestinal and neurological symptoms and, in rare cases, death.
The persistence has a mechanism. Selenium is stored in different tissues and leaves slowly, which is why symptoms outlast the exposure by months rather than days. Investigators tracking the 2008 patients watched urine levels return to normal within a week or two while serum levels only began falling at week six.
Brazil nuts, the food that can pass the ceiling
Most federal upper limits are hard to reach from food alone. Selenium’s is not, and the reason is a snack.
The Office of Dietary Supplements puts Brazil nuts at 68 to 91 micrograms of selenium per nut, and lists a one-ounce serving of six to eight nuts at 544 micrograms, which the page marks as 989 percent of the Daily Value. That single serving is above the 400-microgram upper limit on its own. Working from the per-nut range, five nuts at the top end reach 455 micrograms, past the U.S. ceiling. Three reach 273, past the European one. The fact sheet says plainly that Brazil nuts could cause selenium toxicity if eaten regularly.
It also warns that values from other analyses vary widely, and that caveat has a number behind it. A 2017 study in Chemosphere by E.C. Silva Junior and colleagues sampled Brazil nuts and the soils beneath the trees across the Amazon basin and found median selenium concentrations ranging from 2.07 mg per kilogram in nuts from Mato Grosso state to 68.15 mg per kilogram in nuts from Amazonas state, roughly a 33-fold difference driven by origin. In the authors’ own framing, a single nut could supply anywhere from 11 percent to 288 percent of an adult man’s daily selenium requirement depending on which state it grew in. They used a 70-microgram requirement, higher than the U.S. RDA of 55, so against the American figure the spread runs wider still. Soil acidity, not just total soil selenium, drove the difference.
Nothing on a bag of Brazil nuts tells you which it is.
What a selenium label cannot tell you
Built from the sources above rather than copied from any of them. These are the gaps between a Supplement Facts panel and the actual question.
- Which ceiling the percentage is measured against. The percent Daily Value uses 55 micrograms. It says nothing about 400, nothing about 255, and nothing about which of those two an expert committee would apply.
- What form the selenium is. The 2008 product used sodium selenite, an inorganic form. SELECT used L-selenomethionine, an organic one. The outbreak investigators noted that organic selenomethionine produces much higher serum selenium concentrations than inorganic forms do.
- What the rest of the day contributes. The upper limit counts food and supplements together, and a handful of Brazil nuts is not a rounding error.
- Whether the contents match the panel. The 2008 outbreak is the proof: no amount of careful label reading would have caught a product holding 200 times what it claimed.
- Whether the limit applies to you at all. The Office of Dietary Supplements excludes people taking selenium under a physician’s care from the upper limit entirely.
One more gap, and it is the fact sheet’s own admission. Its interactions section names exactly one medication, cisplatin, and says outright that one example is being provided. It is not a complete list and should not be read as one. Article 677 covers supplement and medication interactions.
Where a selenium question stops being a label question
Three situations, and one thing to bring to the appointment.
Hair coming out, or nails that have changed color, gone brittle or lifted, in someone taking any selenium-containing product. In the 2008 outbreak these were the signs that eventually sent people looking for answers, and they were also the slowest to resolve.
A garlic smell on the breath or a persistent metallic taste. The Office of Dietary Supplements describes both as early indicators of excess intake, which makes them worth mentioning out loud rather than ignoring.
Anyone taking selenium for a diagnosed condition or under a physician’s direction. The published limit does not apply there, and the figure that does is a clinical judgment.
And the lesson the outbreak paid for in hair and fingernails: tell the clinician what is in the cabinet. Forty-one percent of the affected people had not sought care at all, and those who did often did not mention the supplement, because they did not connect a health product with being ill. The most useful thing a patient brought to those appointments was the bottle.
What selenium demonstrates
Selenium is the clearest case in the supplement aisle of a nutrient where the useful dose and the harmful dose live close together, and where ordinary food can cover the distance by itself.
Two agencies publish ceilings 145 micrograms apart. One ounce of a common snack food passes both. And the worst American episode on record did not happen because anyone took too many pills. It happened because a supplier made an error, nobody tested the result, and the number on the label was off by a factor of 200. That is a quality-control problem, not a consumer-vigilance problem, and no amount of careful reading solves it.
Sources
- NIH Office of Dietary Supplements. “Selenium: Fact Sheet for Health Professionals.” ods.od.nih.gov
- NIH Office of Dietary Supplements. “Zinc: Fact Sheet for Health Professionals (used for the zinc RDA and UL comparison).” ods.od.nih.gov
- Centers for Disease Control and Prevention. “CDC Alert on Adverse Effects Associated with Consuming "Total Body Formula" and "Total Body Mega Formula" (Health Alert Network, CDCHAN-00274, April 14, 2008).” stacks.cdc.gov
- U.S. Food and Drug Administration (eCFR). “21 CFR 101.9, Nutrition labeling of food (Reference Daily Intakes).” ecfr.gov
- PubMed. “MacFarquhar JK, Broussard DL, Melstrom P, et al. "Acute selenium toxicity associated with a dietary supplement." Archives of Internal Medicine, 2010. DOI: 10.1001/archinternmed.2009.495.” doi.org
- PubMed. “EFSA Panel on Nutrition, Novel Foods and Food Allergens (Turck D, et al.). "Scientific opinion on the tolerable upper intake level for selenium." EFSA Journal, 2023. DOI: 10.2903/j.efsa.2023.7704.” doi.org
- PubMed. “Lippman SM, Klein EA, Goodman PJ, et al. "Effect of selenium and vitamin E on risk of prostate cancer and other cancers: the Selenium and Vitamin E Cancer Prevention Trial (SELECT)." JAMA, 2009. DOI: 10.1001/jama.2008.864.” doi.org
- PubMed. “Klein EA, Thompson IM, Tangen CM, et al. "Vitamin E and the risk of prostate cancer: the Selenium and Vitamin E Cancer Prevention Trial (SELECT)." JAMA, 2011. DOI: 10.1001/jama.2011.1437.” doi.org
- PubMed. “Silva Junior EC, Wadt LHO, Silva KE, et al. "Natural variation of selenium in Brazil nuts and soils from the Amazon region." Chemosphere, 2017. DOI: 10.1016/j.chemosphere.2017.08.158.” doi.org