Amygdalin is a natural compound in the kernels (seeds) of bitter apricots and other stone fruits. “Laetrile” is a related, semi-synthetic version of it, and “vitamin B17” is a marketing name for the same chemistry. They are used interchangeably in alternative-cancer circles, though they are not exactly the same substance. All of them are promoted on one core claim: that they release a cancer-killing agent selectively inside tumours. That claim has never been shown to be true in people. What the chemistry actually does, reliably, is release hydrogen cyanide, which is why apricot kernels and amygdalin supplements keep sending people to hospital, and have killed. This is not an unproven promise. It is a claim that has been examined and a hazard that is documented.
Why the claim sounds convincing
The idea arrives wrapped in a persuasive story: a cheap, natural molecule from fruit kernels, a “vitamin” the body supposedly lacks, and a cure the establishment is said to have buried. That story has kept the name alive for decades, helped along by cultural memory such as the actor Steve McQueen’s widely reported laetrile treatment in Mexico in 1980.
It does not survive contact with the evidence. The US National Cancer Institute states plainly that laetrile has shown little anticancer activity in animal studies and no anticancer activity in human clinical trials, is not approved by the Food and Drug Administration, and carries a real risk of cyanide poisoning.[4] The “vitamin B17” label is not science: the term was coined to market it, and there is no evidence amygdalin is required for normal metabolism or behaves as a vitamin.[4]
The real history behind the apricot-kernel claim
The kernels do have a genuine medicinal history, which is part of why the modern claim feels credible. But traditional practice turned on a distinction the modern story ignores: sweet and bitter kernels were used for entirely different things. Sweet kernels were eaten as food and pressed for an edible oil. The bitter kernels, the ones loaded with the cyanide-releasing compound, were the medicinal ones, and largely for external use. In Chinese medicine, bitter apricot seed (ku xing ren) has been used for roughly two thousand years, but for coughs, asthma and constipation, and as a carefully processed preparation rather than raw kernels eaten by the handful.[7] In Himalayan communities such as Ladakh, the edible oil came from sweet kernels, while bitter-kernel oil was reserved for the skin, for massage, aches, and cosmetic and household use.[8] In none of these traditions is cancer a recorded use, and none involved swallowing large numbers of raw bitter kernels for their amygdalin.
The most persistent version of the story attaches to the Hunza of northern Pakistan, said to eat apricot kernels, live past 120 and never get cancer. Apricots genuinely are central to that region’s food culture, but the rest does not hold up. The extreme-age claims were never supported by reliable birth records and are now treated as a classic geographic longevity myth,[9] and the “cancer-free” claim has circulated in questionable-cancer-treatment lore for decades without any population study behind it.[10] The habit of eating raw bitter kernels for their amygdalin is a twentieth-century Western idea, not an ancient tradition. We tell that fuller story — the real Chinese, Himalayan and Hunza traditions, and where the cancer claim actually came from — in The Apricot Kernel: A History of Food, Medicine and Mountain Life.
The clinical record: no controlled trial, and no benefit in the largest study
Most alternative cancer remedies are argued about because they were never properly studied. Laetrile is unusual: it was studied, and the record is unusually clear about what is and is not there.
A systematic review that set out to collect every kind of clinical report found 36 in total — case reports, best-case series and small case series — and among them, no controlled clinical trials at all. None of the 36 established that laetrile works.[11] Earlier, the National Cancer Institute had asked practitioners and pro-laetrile groups to submit their best cases; from an estimated 70,000 or more users, only 93 records came in, and a blinded panel of specialists judged just six to be responses — “no definite conclusions”.[12]
That weak signal led to the largest clinical study ever done, published in 1982. Researchers treated 178 cancer patients with amygdalin plus the full “metabolic therapy” package of diet, enzymes and vitamins, at doses representative of real laetrile practice. It is important to be precise about what this study was: an uncontrolled evaluation, not a randomised or controlled trial. Its finding was nonetheless decisive.
Nothing since has changed the picture. A 2015 Cochrane review found no study meeting basic quality criteria for benefit and called the risk–benefit balance “unambiguously negative”.[2] A 2025 independent review, published after four more decades of laboratory interest, again concluded there is no clear evidence of cancer-treatment efficacy, particularly at the clinical level.[13]
Half a century, the same conclusion
The idea is packaged. Ernst Krebs Jr promotes a preparation he calls “laetrile” and later rebrands amygdalin as “vitamin B17,” despite it having no vitamin properties.[4]
The best cases are collected, and fall short. The National Cancer Institute’s review of solicited “success” cases finds nothing supporting anticancer activity.[12]
The largest clinical evaluation. 178 patients; no substantive benefit; cyanide toxicity observed. “A toxic drug that is not effective as a cancer treatment.”[1]
Formal review. Cochrane finds no reliable evidence of benefit and calls the risk–benefit balance unambiguously negative.[2]
The harm is still current. The FDA warns that commercial bitter-apricot-seed products contain amygdalin capable of fatal cyanide toxicity,[14] and a 2025 report describes 14 children treated for kernel/seed cyanide poisoning.[6]
The mechanism sold as a cure
Amygdalin is a cyanogenic glycoside: when an enzyme called β-glucosidase breaks it apart, one of the products is hydrogen cyanide. Cyanide shuts down the part of the cell that uses oxygen to make energy, which is why serious poisoning causes headache, confusion, seizures, coma, a dangerous build-up of acid in the blood, and, in the worst cases, death. Treatment relies on antidotes such as hydroxocobalamin and sodium thiosulfate.[15]
The body does have a defence against small amounts. An enzyme called rhodanese, working mainly in the liver and kidneys, attaches cyanide to a sulfur atom to make thiocyanate, a far less harmful substance the kidneys excrete. But the pathway has a limit, and the limit is not the enzyme. It is the body’s supply of sulfur, so rhodanese can only clear so much cyanide at a time. A large dose can overwhelm it, and repeated dosing can draw the sulfur reserves down, which is part of why ongoing use can cause slow, cumulative harm rather than a single dramatic crisis. It is also why one of the antidotes is sodium thiosulfate: it supplies the sulfur the enzyme needs.[15]
Advocates answer the obvious objection with a theory of selectivity: cancer cells, they say, are rich in the enzyme that unlocks the cyanide and poor in the one that detoxifies it, so the poison is released precisely where it is wanted. The trouble is that this has never been shown in patients, and there was never just one version of the theory. When researchers tested human tissues directly, they found the amygdalin-splitting enzyme active in the lining of the small intestine but could not detect it releasing cyanide in any of the human tumour samples examined — the opposite of what selective tumour-killing would require.[3] An earlier review reached the same conclusion about the enzymes involved.[16]
Why the oral route is the dangerous one
Most of the body’s own cells carry little of the enzyme that frees cyanide from amygdalin. The gut is the exception: the lining of the small intestine can do it, and gut bacteria do a great deal of it. So the dangerous way to take amygdalin is orally — swallowed, it reaches the large intestine, where bacteria unlock the cyanide.[17,18,3] In the early human study, the same compound given intravenously was largely excreted unchanged and caused no toxicity, while the oral route produced significant blood cyanide.[19]
Intravenous dosing is not simply harmless, though. In a later study of 55 patients who were already taking amygdalin orally, an intravenous dose still raised average blood cyanide measurably.[20] The takeaway is not that injection is safe, but that the familiar, gentle-seeming pill is the more hazardous form.
“A kernel” is not “a dose”
The words “apricot kernel,” “amygdalin” and “B17” hide very different exposures. When the FDA measured amygdalin in marketed apricot kernels and almonds, the content varied roughly ten-thousand-fold across the samples.[21] The route, the product and what is taken alongside all change how much cyanide is released.
| Form | What we actually know | What not to assume |
|---|---|---|
| Raw bitter apricot kernels | European regulators estimate that, in one serving, even a few small kernels can push a child past the safe cyanide limit, and less than half of one large kernel can do so for an adult.[22] As few as three kernels put a three-year-old into a coma.[23] In 2024 the FDA warned that specific commercial seed products could cause fatal cyanide toxicity.[14] | “It’s just a seed” is not reassurance. Bitter kernels are a genuine acute-cyanide source, especially for children. |
| Amygdalin / Laetrile (“Vitamin B17”) tablets | The highest-risk consumer form. This covers purified “laetrile”/amygdalin tablets and bitter-apricot-kernel or kernel-extract capsules alike: they differ in what else is in them, but all release cyanide once gut bacteria act on them. Tablet and supplement ingestions have caused life-threatening poisoning.[24,25] | A supplement label does not tell you the cyanide a given batch will release in your gut. |
| Intravenous amygdalin | Much lower acute toxicity, but not inert — it still raised blood cyanide in patients already taking it orally.[19,20] | Lower toxicity by this route is not benefit; the same research is part of the record showing no efficacy. |
| Taken with high-dose vitamin C | Frequently used with amygdalin/laetrile, vitamin C can increase the conversion of amygdalin to cyanide and deplete a substance the body uses to detoxify it, and may raise the danger.[26] | Two “natural” products together can be more dangerous than either alone. |
The harm is documented
Because the mechanism is cyanide, the harm is not theoretical. It appears across the decades, across the forms people use, and in more than one shape.
- Deaths. A fatal cyanide poisoning was reported in an infant who swallowed amygdalin kept at home for a relative’s cancer,[5] and a rapid death after oral laetrile in a young patient.[27] In another case, a patient who already had liver disease went into a coma after a three-gram dose and died of massive liver failure that the authors judged possibly related to laetrile.[28]
- Slow and progressive nerve damage. One patient developed a progressive nerve-and-muscle disorder with raised cyanide and thiocyanate levels after laetrile, which improved once the treatment stopped.[29]
- Modern cases, adult and child. Severe cyanide toxicity from internet-bought “vitamin B17” in a woman with breast cancer;[30] a four-year-old with cancer given intravenous and oral “B17” plus apricot kernel who arrived unresponsive;[31] tablet and supplement poisonings in 2017 and 2022;[24,25] and a 2025 report of 14 children treated for apricot-seed cyanide poisoning.[6]
- Chronic, hidden exposure. A man who took apricot-kernel extract daily for five years was found to have chronic cyanide toxicity, discovered by chance during surgery.[32]
Where does ordinary food end and danger begin? European regulators set an acute safe limit for cyanide of about 20 micrograms per kilogram of body weight, whatever the source.[33] Everyday foods rarely reach it, but deliberately eating bitter kernels “for health” is a different act. EU law now caps the cyanide in apricot kernels sold to consumers at 20 milligrams per kilogram of kernel, and requires sellers to prove they comply.[34] As the paediatric cases show, it can still take very little.
What the poison centres record
The individual cases noted above are just the ‘tip of the iceberg’. National poison-centre records demonstrate the exposure burden is far larger. French poison centres logged 154 apricot-kernel cases in under six years;[35] Canada’s logged 195 enquiries in five;[36] and the New South Wales centre in Australia took 120 amygdalin-product calls in eight years, a third of them serious enough to send the person to hospital.[37] While no standard method exists between nations to track these cases, they all point the same way: far more people are exposed, and poisoned, than the few published case reports suggest. And it’s a pattern documented worldwide for decades.[38]
| Where and when | What was recorded | Outcome severity |
|---|---|---|
| France, 2012–2017 | 154 apricot-kernel cases in the national poison-centre database; 12% taken for “anti-cancer” purposes.[35] | No deaths; two more serious cases (hypotension after 50 kernels; a cardiac event after 40). |
| Canada, 2013–2017 | 195 poison-centre enquiries about apricot-kernel consumption.[36] | No deaths in this set; Health Canada later capped the cyanide allowed in kernels sold as food. |
| United Kingdom, 2008–2019 | 437 plant-material cyanide ingestions among 1,252 suspected cyanide cases.[39] | 40 deaths across the whole cyanide dataset, but mostly from smoke inhalation, not kernels. |
| Australia (NSW), 2015–2022 | 120 amygdalin / “B17” product exposure calls.[37] | A third had symptoms or a history serious enough for hospital referral. |
| Turkey, 2005–2009 | 13 children admitted to intensive care with apricot-seed cyanide poisoning.[40] | Four needed a ventilator and two were comatose; all survived with treatment. |
Most of those exposures were mild, and that matters: eating a kernel or two may not be a medical emergency, and it would be dishonest to pretend otherwise. But the French data show why “nothing happened to me” is not reassurance. The chance of symptoms climbed steadily with the number of kernels eaten, from about 5% below five kernels to 80% at fifty or more.[35] And people really do take them in quantity: among the French cases, fifteen involved regular, repeated use, most of it for cancer, at a median of 10 kernels a day.[35] Set that against the dose promoted online, where ANSES and EFSA both found sites recommending 10 kernels a day for “prevention” and up to 60 a day for people with cancer.[35,22] The advertised cancer dose sits squarely in the range where four out of five real-world exposures caused symptoms.
What about the laboratory studies?
Supporters point to cell studies in which amygdalin damages cancer cells, for example in lung, bladder and breast cancer cells.[41,42,43] That literature is real, but several things keep it from meaning what advocates claim.
What the lab studies are
- Cells in a dish, or tumours in mice, often exposed to high concentrations of amygdalin.
- Reasons to ask a question, which is exactly what the 1982 study then did.
- Even the most-cited of them says plainly that testing in a living body “must follow.”[42]
What they are not
- Not consistent: some kernel-extract studies actually stimulated cancer-cell growth at certain doses.[44]
- Not all reliable: at least five amygdalin cancer papers have been formally retracted, including a review.[45]
- Not human evidence: a 2025 review still finds no clear clinical efficacy.[13]
None of this means scientists have given up on the molecule. Amygdalin is still studied, and a 2025 review found the laboratory interest continuing.[13] But it is worth being clear about what that research is trying to do. It is not testing whether apricot kernels or “B17” capsules treat cancer. It is trying to turn amygdalin into something it currently is not: a stable, well-absorbed, tumour-targeted agent that does not simply expose the patient to cyanide. A 2026 study, for instance, sealed amygdalin inside an engineered nanoparticle to test in mice, precisely because ordinary amygdalin’s “cyanide toxicity, poor stability, and minimal bioavailability” make it unusable as it is.[46]
That effort is itself a kind of verdict on the product people actually buy. The frontier is not confirming that raw kernels or capsules work; it is trying to escape the very problems they carry. And even these engineered results are early animal experiments that have not produced a proven treatment in a single human being.[13]
The reassuring counter-argument, and why it fails
An honest article should show the other side. A 2016 review from a group that has studied amygdalin argued that, while there is no convincing evidence it shrinks tumours in patients, there is also no evidence that purified amygdalin at a “therapeutic” dose causes toxicity.[47] A later review even proposed a supposedly avoidable oral dose range.[48]
The problem is the phrase “therapeutic dose.” It assumes a dose has been shown to help at acceptable risk — but no controlled study has ever established such a dose. And the point is not hypothetical: the National Cancer Institute’s own study used pharmaceutical-grade amygdalin whose purity it checked, and still recorded cyanide toxicity, with some patients’ blood cyanide approaching dangerous levels.[1,4] A “safe therapeutic dose” you cannot define, and cannot deliver reliably from products that vary ten-thousand-fold, is not a safety argument.
The real decision is an evidence asymmetry
You do not need to settle every laboratory debate. You only need to weigh what is known about benefit against what is known about harm.
Potential cancer benefit
- Cell and animal studies, inconsistent and partly retracted
- A selectivity theory never shown in patients
- A “vitamin” that is not a vitamin
- One weak collection of hand-picked “best” cases
- No controlled trial, ever
- The largest study found no substantive benefit
Known harm
- The reliable effect of oral use is cyanide exposure
- Worst by the oral route, the one people use
- May be made more dangerous by high-dose vitamin C
- Documented deaths, including a child
- Poisonings from the 1970s to 2025
The evidence for cancer benefit is weak and uncontrolled. The cyanide is real.
What this means if you or someone you love has cancer
- Do not use apricot kernels, amygdalin or laetrile as a cancer treatment. No controlled trial supports it, the largest study found no benefit, and it can cause cyanide poisoning.[1,4]
- Treat the oral form as the dangerous one. Kernels, tablets and capsules release cyanide in the gut in a way intravenous dosing does not.[17,19]
- Never combine it with high-dose vitamin C, which may turn a dose into a medical emergency.[26]
- Keep kernels and “B17” products away from children. A very small amount can be dangerous.[22,23]
- Tell your oncology team about anything you are taking. They can check for interactions and give you accurate guidance.
The conclusion is firmer than for most compounds this site examines, and it does not depend on proving a negative. There is no controlled evidence that apricot kernels, amygdalin or laetrile treat cancer, the largest study found nothing, and the selectivity the whole claim rests on has never been shown — while the cyanide, and the poisonings, are documented. For anyone facing cancer, that asymmetry is not close.
Last reviewed: August 2026.