Apricot kernel oil: an ordinary fat with a cyanide question attached

A Cochrane review screened over 200 references and found no randomized trial at all of laetrile or amygdalin in people with cancer, and amygdalin is the apricot kernel compound sold, misleadingly, as vitamin B17. The authors judged its balance of risk and benefit unambiguously negative because of cyanide poisoning. That compound can reach the oil. In a lab paper that tested a single cold-pressed oil and no refined one, oil from a low-amygdalin Italian variety held 10 mg of amygdalin per kg, and the same pressing of other varieties has reached 400 mg/kg. The paper calls 5 mg/kg safe, a figure it borrows from another paper with no agency behind it. EFSA sets the acute limit for cyanide at 20 micrograms per kg of body weight, whatever the food. As food, USDA lists the oil as 100 g fat per 3.5 oz (100 g), with 6.3 g saturated fat and 4 mg vitamin E, from one old record with no spread between samples. We found no trial of people eating the oil.

The numbers, per 3.5 oz (100 g)

How much is actually in it

• USDA lists apricot kernel oil as 100 g fat per 3.5 oz (100 g), with 6.3 g saturated fat and 4 mg vitamin E. — USDA SR Legacy food record Oil, apricot kernel, FDC 171032 (n = 1 record) Evidence E source
  Fat 100 g, saturated fat 6.3 g, vitamin E 4 mg per 3.5 oz (100 g); no sample spread
• Laboratory analysis found 10 mg of amygdalin per kg in a cold-pressed apricot kernel oil from a low-amygdalin Italian cultivar, above the 5 mg/kg the authors treat as safe, and other cultivars pressed the same way can reach 400 mg/kg. — Cold-pressed oil from kernels of the Farclo apricot variety, 2018 harvest, Italy (n = 1 oil) Evidence D source
  10 mg/kg amygdalin in this oil; up to 400 mg/kg reported for other cultivars; safe level cited as below 5 mg/kg
• Sweet apricot kernels from three Central Asian cultivars held 40 to 50% fat, and sweet kernels carry far less amygdalin than bitter ones. — Sweet apricot kernels of the Suhany, Kandek and Kurmayi cultivars (n = 3 cultivars) Evidence D source
  Fat 40.0-49.9%; amygdalin in sweet kernels about 0.06-0.88 g/100 g

What your body absorbs

• According to EFSA, intact amygdalin is barely absorbed from the gut, and its cyanide is released mainly by gut bacteria, so 1.5 g a day for 21 days gave blood cyanide near 80 nmol/mL. — Humans taking pure amygdalin, summarized by EFSA Evidence E source
  Plasma amygdalin peak 1.1 nmol/mL versus whole-blood cyanide about 80 nmol/mL at 1.5 g/day for 21 days
• According to EFSA, eating plant enzymes with amygdalin raised peak blood cyanide from 3.4 to 10.0 micromol/L in a volunteer given 120 mg amygdalin. — One human volunteer, 120 mg isolated amygdalin with and without sweet almonds (n = 1) Evidence E source
  Peak blood cyanide 3.4 uM after 60 min alone versus 10.0 uM after 30 min with sweet almonds

What cooking and storage change

• Laboratory work reports that grinding, soaking, fermentation, enzymes, ultrasound and microwave treatment can lower amygdalin in apricot kernel oil, with microwaves degrading it. — Apricot kernel oil and kernels, laboratory processing Evidence D source
  Direction only: amygdalin reduced; size not given in these sentences

What it does to your health

• In an uncontrolled study, 18 Slovak women who ate 60 mg of bitter apricot seeds per kg body weight daily for 42 days had lower LDL cholesterol at the end. — 18 Slovak women of reproductive age, mean 41.6 years, 42 days (n = 18) Evidence C source
  LDL-C decreased (P < 0.05) after 42 days; size not given in the abstract; total cholesterol and triglycerides rose at day 21
• In a meta-analysis of 9 RCTs, replacing dietary palmitic acid with oleic acid lowered LDL cholesterol by 6.2 mg/dL (0.16 mmol/L). — adults in RCTs of isoenergetic fatty acid replacement lasting at least 14 days, mean age 39.9 (n = 9 RCTs) Evidence A source
  LDL-C -6.2 mg/dL (95% CI -11 to -1.2; metric: -0.16 mmol/L, -0.28 to -0.03), I2 89.6%; no effect on HDL-C, triacylglycerol, glucose, CRP

Safety, limits and interactions

• According to EFSA, the acute reference dose for cyanide is 20 micrograms per kg body weight, set in 2016 for raw apricot kernels and confirmed in 2019 for every dietary source. — General population, all age groups (EU) Evidence E source
  ARfD 20 ug cyanide per kg body weight per eating occasion
• A 41-year-old woman became weak and short of breath within 20 minutes of eating apricot kernels from a health food store and was comatose on arrival with cyanide poisoning. — One 41-year-old woman, USA (n = 1) Evidence D source
  Onset within 20 minutes; coma, hypothermia, persistent metabolic acidosis; recovered with antidotes
• An Australian poisons center logged 120 calls about amygdalin products from 2015 to 2022, and 32% involved patients referred to or already in hospital. — Calls to the New South Wales Poisons Information Centre, 2015 to 2022 (n = 120) Evidence C source
  38 of 120 calls (32%) needed hospital referral or came from a treating hospital

What people believe that the data does not support

• Amygdalin is sold misleadingly as vitamin B17, and a 33-year-old woman who swallowed 20 g of an apricot amygdalin supplement developed cyanide toxicity that returned 12 hours after antidotes. — One 33-year-old woman, intentional overdose of an apricot amygdalin supplement (n = 1) Evidence D source
  Lactate 14.1 mmol/L, QTc 538 ms, intubation; rebound acidosis 12 h later

Who this works differently for

• According to EFSA, one small raw apricot kernel exceeds the cyanide limit for a toddler, and more than three small kernels exceed it for an adult. — Toddlers and adults (EU risk assessment of raw apricot kernels) Evidence E source
  0.06 g kernel in toddlers and 0.37 g in adults reach the ARfD
• Of 13 children treated in one Turkish intensive care unit for cyanide poisoning from apricot seeds, four needed mechanical ventilation and all recovered. — 13 children admitted to a pediatric ICU in Turkey, 2005 to 2009 (n = 13) Evidence C source
  Symptom onset mean 60 min; 4 ventilated, 2 comatose, 9 with lactic acidosis; all discharged within mean 3.1 days

What is still unknown

• A Cochrane review screened over 200 references and found no randomized or quasi-randomized trial of laetrile or amygdalin in cancer, so the anti-cancer claim for apricot kernel amygdalin has no trial support. — People with cancer; trials of laetrile or amygdalin searched to 2014-2015 (n = 0 included trials) Evidence A source
  No eligible trials; risk-benefit judged unambiguously negative because of cyanide poisoning

The bottom line

The case for the fat is borrowed. A meta-analysis of 9 randomized trials found that swapping palmitic acid for oleic acid lowered LDL cholesterol by 6.2 mg/dL (95% CI -11 to -1.2; metric: 0.16 mmol/L, -0.28 to -0.03), with high inconsistency between trials and no change in HDL or triglycerides. That is evidence about oleic acid, and the USDA row we hold does not give this oil's oleic share. The one feeding study we have gave 18 Slovak women bitter apricot seeds, 60 mg per kg of body weight a day for 42 days, with no control group. Their LDL fell by the end, by an amount the abstract does not give. That daily total, split into 8 to 12 doses, is several times the 0.37 g of kernel that EFSA calculates takes an adult to the cyanide limit at one sitting. Who should be careful: the danger sits in the kernels, and we have no measurement of how much of it a spoonful of oil carries. EFSA calculated that one small raw kernel takes a toddler past the cyanide limit, and more than three small ones take an adult past it. Kernel enzymes speed the release. In one volunteer given 120 mg of pure amygdalin, peak blood cyanide was 3.4 micromol/L alone and 10.0 when sweet almonds were eaten with it. A 41-year-old woman became weak and short of breath within 20 minutes of eating kernels from a health food store and was comatose on arrival at hospital. In one Turkish intensive care unit, 4 of 13 children poisoned by apricot seeds needed a ventilator, and all went home. An Australian poisons center took 120 calls about amygdalin products from 2015 to 2022, and 32 percent involved hospital care, counting calls rather than confirmed poisonings. Lab work lists grinding, soaking, fermentation, enzymes, ultrasound and microwaves as ways to lower amygdalin in the oil, with no size given. Keep apricot kernels and anything sold as vitamin B17 away from children. For the oil itself, the open question is how much amygdalin is in the bottle in front of you.

Sources

1. apko-r7-01 · USDA FoodData Central SR Legacy, dataset
   Oil, apricot kernel (SR Legacy, FDC 171032): Total lipid (fat) 100 G; Fatty acids, total saturated 6.3 G; Vitamin E (alpha-tocopherol) 4 MG per 100 g
   USDA FoodData Central, SR Legacy, FDC 171032 · checked 2026-09-30
2. apko-r7-02 · laboratory analysis
   Results highlighted that the concentration of the considered metabolite was 10 mg/kg of oil, which is a value below the one obtained with other cultivars, but still too high to be considered safe. Indeed, using the same extraction procedure, amygdalin content can reach 400 mg/kg of oil [19], while the concentration considered as safe is below 5 mg/kg of oil [20].
   Int J Mol Sci, 2025 · checked 2026-09-30
3. apko-r7-14 · laboratory analysis
   The kernels contained 3.3%-4.7% moisture, 40.0%-49.9% fat, 17.9%-18.9% protein, 8.1%-14.7% crude fibre and 2.2%-2.7% ash. ... The amygdalin content in sweet apricot kernels is generally low (~0.06-0.88 g/100 g) compared to bitter kernels and is safe for consumption.
   Int J Food Sci, 2026 · checked 2026-09-30
4. apko-r7-06 · EFSA CONTAM Panel opinion
   For example, Ames et al. (1981) reported that the ingestion of 1.5 g (3.28 mmol) of pure amygdalin per day for 21 days gave rise to only marginal levels of unchanged amygdalin (peak at 1.1 nmol/mL) in blood plasma but much higher levels of cyanide (ca. 80 nmol/mL) in whole‐blood. This finding is in agreement with the animal studies discussed above, indicating that intact amygdalin is virtually not absorbed from the gastrointestinal tract but partially degraded to cyanide, probably by the gut microflora.
   EFSA CONTAM Panel, EFSA Journal, 2019 · checked 2026-09-30
5. apko-r7-07 · EFSA CONTAM Panel opinion
   After ingestion of 120 mg isolated amygdalin containing 6.8 mg cyanide, a peak cyanide level of 3.4 μM was reached after 60 min, indicating some minor degradation of amygdalin (by the intestinal flora) occurring in the human body even in the absence of the plant enzymes. A distinct higher level of 10.0 μM was reached after 30 min when sweet almonds (containing the degrading plant enzymes but no amygdalin) were ingested together with the same dose of isolated amygdalin.
   EFSA CONTAM Panel, EFSA Journal, 2019 · checked 2026-09-30
6. apko-r7-03 · laboratory analysis
   To mitigate this risk, various processing methods can be employed to reduce amygdalin content in kernel oil, including grinding and soaking, as well as more advanced approaches such as fermentation, enzymatic treatment, ultrasound, or microwave processing [5,8].
   Int J Mol Sci, 2025 · checked 2026-09-30
7. apko-r7-13 · uncontrolled clinical study
   Volunteers consumed 60 mg.kg-1 of body weight of bitter apricot seeds divided into 8-12 doses daily for 42 days. ... The blood levels of T-C, HDL-C and T-C did not change significantly (P > 0.05), however, the level of LDL-C decreased significantly (P < 0.05) after 42 days.
   J Environ Sci Health B, 2021 · checked 2026-09-30
8. apko-r7-15 · meta-analysis of RCTs
   We observed reductions in LDL-cholesterol concentrations after the replacement of palmitic acid (16:0) with UFAs (-0.36 mmol/L; 95% CI: -0.50, -0.21 mmol/L; I2 = 96.0%, n = 18 RCTs) or oleic acid (18:1n-9) (-0.16 mmol/L; 95% CI: -0.28, -0.03 mmol/L; I2 = 89.6%, n = 9 RCTs), with a similar impact on total cholesterol and apoB concentrations.
   Adv Nutr, 2022 · checked 2026-10-01
9. apko-r7-04 · EFSA CONTAM Panel opinion
   The CONTAM Panel concluded that there are no data indicating that the acute reference dose (ARfD) for cyanide of 20 μg/kg bw, established in 2016, should be revised and that it is applicable for acute effects of cyanide regardless of the dietary source.
   EFSA CONTAM Panel, EFSA Journal, 2019 · checked 2026-09-30
10. apko-r7-10 · case report
    A 41-year-old woman ingested apricot kernels purchased at a health food store and became weak and dyspneic within 20 minutes.
    Ann Emerg Med, 1998 · checked 2026-09-30
11. apko-r7-12 · retrospective review
    There were 120 unique exposure calls. Eighty-two (68%) were regarding minor exposures, with the remaining 38 (32%) of calls involving patients who had either a signifcant history or symptoms to prompt referral to hospital or were already seeking advice from a treating hospital clinican.
    Emerg Med Australas, 2024 · checked 2026-09-30
12. apko-r7-09 · case report
    Amygdalin, marketed misleadingly as supplement "Vitamin B17," is a cyanogenic glycoside.
    Clin Toxicol (Phila), 2020 · checked 2026-09-30
13. apko-r7-05 · EFSA CONTAM Panel opinion
    The quantity of apricot kernels that can be consumed without exceeding the ARfD was estimated to be 0.06 and 0.37 g in toddlers and adults, respectively. The Panel concluded that the ARfD would be exceeded by consumption of one small kernel in toddlers and by more than three small kernels in adults or less than half of a large kernel.
    EFSA CONTAM Panel, EFSA Journal, 2019 · checked 2026-09-30
14. apko-r7-11 · case series
    Thirteen patients admitted to the Pediatric Intensive Care Unit (PICU) of Erciyes University between 2005 and 2009 with cyanide intoxication associated with ingestion of apricot seeds were reviewed retrospectively. ... In addition to signs of mild poisoning related to cyanide intoxication, there was severe intoxication requiring mechanical ventilation (in four cases), hypotension (in two), coma (in two) and convulsions (in one).
    Ann Trop Paediatr, 2010 · checked 2026-09-30
15. apko-r7-08 · Cochrane systematic review
    We located over 200 references, 63 were evaluated in the original review, 6 in the 2011 and none in this update. However, we did not identify any studies that met our inclusion criteria. ... The claims that laetrile or amygdalin have beneficial effects for cancer patients are not currently supported by sound clinical data.
    Cochrane Database Syst Rev, 2015 · checked 2026-09-30