A trace element found in fruit, vegetables and nuts, not established as essential for humans. It is sold for bone health, joint comfort and — more prominently in fitness marketing — testosterone.
Boron sits in an unusual place: it has no established essential function in humans, no RDA, and yet a Tolerable Upper Intake Level of 20 mg/day, because there is enough toxicity data to set a ceiling but not enough functional data to set a requirement. The human research consists mostly of small metabolic studies from a single research programme in the 1980s and 1990s, reporting effects on calcium and magnesium excretion and on steroid hormone levels, plus a handful of small hormone studies at 10 mg/day lasting a week. None of it has been followed into a clinical outcome, and the testosterone claims that dominate its marketing rest on short-term hormone measurements in a few dozen men.
Sold as
Capsule, Tablet, Component of bone and joint formulas, Component of testosterone-support blends
Also known as
Boron citrate, Sodium borate, Boron glycinate
Regulation: No RDA or Adequate Intake has been established because boron is not established as essential for humans. The Tolerable Upper Intake Level is 20 mg/day for adults. Sold as a dietary supplement with no pre-market efficacy review.
What the research used
The amounts, forms and durations published trials of Boron actually administered.
These are the amounts published trials administered — not a recommendation, and not a dose to copy. Trial participants were screened, supervised and given a characterised preparation. What is in a retail bottle is frequently not what was studied, and the amount that suits one person is not general advice.
Studied for Typical retail supplement
3 mg
Not enough
Capsule, daily
Roughly two to three times typical dietary intake and well below the 20 mg/day upper limit. There is no established requirement for comparison.
Studied for Hormone and mineral metabolism studies
3–10 mg
Limited
Sodium borate or boron citrate, daily · 1 week–7 weeks
Small metabolic studies, typically with fewer than 30 participants and often in controlled feeding settings with deliberately low-boron baseline diets, which makes the findings hard to apply to ordinary eating.
Studied for Tolerable Upper Intake Level, adults
20 mg
Strong
Set on reproductive and developmental toxicity in animal studies. Some products supply 6–10 mg, which leaves less headroom than the numbers suggest once diet is added.
This is general information about food, not medical advice. It cannot diagnose, treat or replace guidance from a clinician who knows your history.
When it was taken
Each window says whether a trial compared it against an alternative, or whether it is convention and pharmacology. Most timing advice is the second kind.
None of the timings below were directly compared in a trial. They reflect how the research happened to dose, or what the pharmacology implies — which is not the same as a timing being better.
🕐Any time
Not enough
No timing has been studied. Boron is well absorbed and rapidly excreted in urine.
Timing was not directly compared. This reflects convention or how the compound behaves in the body, not a trial showing this window works better.
The evidence in depth
Best supported: Not enough
What each association actually rests on, and the papers behind it. A rating describes the state of the research, not how promising the supplement sounds. How we rate evidence.
Small metabolic studies reported reduced urinary calcium and magnesium loss with boron supplementation in people on low-boron diets. No trial has measured bone density or fractures, so the bone claim rests entirely on short-term mineral balance.
Not enough evidence. Not enough human research to say anything useful yet.
No source is attached to this association yet, so it is rated not enough evidenceby default. It is listed because people commonly ask about it — not because there is evidence for it.
The osteoarthritis claim rests on ecological observations correlating regional soil boron with arthritis rates, plus one very small pilot trial. That is not a basis for any conclusion.
Not enough evidence. Not enough human research to say anything useful yet.
Boron is sold heavily on hormonal claims: one very small study reported higher free testosterone after a week at 10 mg/day in a handful of men, and a few case-control analyses have linked dietary boron with prostate outcomes. None of it has been replicated, and no trial has examined a clinical outcome.
Not enough evidence. Not enough human research to say anything useful yet.
Deep research mode adds each paper’s own conclusion and stated limitations.
Things to know
Interactions, contraindications and dose limits, most serious first. Concentrated extracts interact with medication in ways the whole plant does not, so this section carries more weight here than it does on a food page.
Too much
Reproductive and developmental toxicity at high doses
The 20 mg/day upper limit is based on testicular and developmental toxicity in animal studies. High-dose boron is not advisable at any stage of pregnancy or when trying to conceive, and the hormone-modifying marketing makes this an easy caution to overlook.
Too much
Acute toxicity from borate compounds
Borax and boric acid ingestion causes nausea, vomiting, diarrhoea, headache and, at high doses, seizures and kidney injury. Boric acid pest products and household borax are not supplements and have caused poisonings, particularly in children.
Worth knowing
May raise oestrogen as well as testosterone
The same small studies used to sell boron for testosterone also reported changes in oestradiol. Anyone taking it for hormonal reasons should know the reported effects are not one-directional and are based on very small samples.
Worth knowing
Kidney impairment
Boron is cleared by the kidneys, so reduced kidney function allows it to accumulate. Supplementation is not advisable in that setting.
In the food catalogue
The whole-food form, where one exists. The evidence for a concentrated extract rarely transfers to the food, or the other way round — these are cross-links, not equivalents.