Safflower oil: two strong studies pointing opposite ways

Across 54 randomised trials that swapped one fat for another at equal calories in adults, safflower oil ranked first of thirteen fats for lowering LDL cholesterol, with a ranking probability of 82 percent. Replacing butter with any of those oils lowered LDL by 0.23 to 0.42 mmol/l per 10 percent of calories exchanged. A ranking is not an effect size. That 82 percent means safflower oil most often came out on top in trials that were small and mostly short, and blood lipids are a marker rather than an outcome anyone lives or dies of. The one randomised trial that fed safflower oil to men after a heart attack points the other way. It followed 458 men aged 30 to 59 in Sydney between 1966 and 1973, for a median of 39 months. Deaths from all causes were 17.6 percent in the safflower group against 11.8 percent in the controls, hazard ratio 1.62 with a 95 percent confidence interval of 1.00 to 2.64. When those recovered data were added to a meta-analysis of linoleic acid trials, the pooled hazard ratio for coronary death came out at 1.33, 95 percent CI 0.99 to 1.79, which does not reach significance. That trial was single-blind and its data were pulled off magnetic tape decades later. Its safflower arm also received polyunsaturated margarine of a kind that carried trans fat in those years. This page cannot resolve the two findings, and nobody has run the trial that would. Two different oils share this name, and the FDA claim on a label covers only the high oleic kind. The ordinary linoleic oil most bottles hold is outside it. The one severe harm on this page is not from cooking. Three adults who took safflower oil as a weight loss supplement developed acute liver failure, all three needed a liver transplant and one died 12 days after the operation.

The numbers, per 100 g

How much is actually in it

• Safflower oil carries 45.95 mg of vitamin E per 100 g, which is 306 percent of the Daily Value, but across eight USDA Foundation samples the vitamin E ran from 32.67 to 58.25 mg and the vitamin K from 0.8 to 19.05 mcg. — USDA Foundation Foods record for Oil, safflower (FDC 1750350), 8 analysed samples, against the CoFID 2021 record for the same oil (n = 8 samples) Evidence E source
  Vitamin E 45.95 mg per 100 g with samples 32.67 to 58.25, saturated fat 7.651 g with samples 6.94 to 8.16, vitamin K 5.344 mcg with samples 0.8 to 19.05

What a real portion looks like

• A tablespoon of safflower oil weighs 14 g and a cup weighs 224 g, so every per 100 g figure on a label describes about seven tablespoons of oil. — household measures recorded for safflower oil in the food code used for US dietary recalls (n = 3 measures) Evidence E source
  1 tablespoon 14 g, 1 cup 224 g

What your body absorbs

• The apolipoprotein B-48 response after a meal differed between oils eaten by the same people, and it was higher after olive oil than after safflower oil, P = 0.003, so these oils are handled differently at the absorption step. — adults fed palm oil, safflower oil, a fish and safflower oil mixture, or olive oil in a sequential two-meal protocol, each person receiving every oil Evidence B source
  Apolipoprotein B-48 response differed between oils with olive oil highest after both meals, P = 0.003, and the triacylglycerol to apo B-48 ratio was lowest after olive oil, P = 0.02

What it does to your health

• Across 54 randomised trials comparing thirteen fats head to head, safflower oil ranked first for lowering LDL cholesterol with a ranking probability of 82 percent, and replacing butter with any of these oils lowered LDL by 0.23 to 0.42 mmol/l. — adults aged 18 and over in 54 randomised trials of at least three weeks, each swapping one fat for another at equal calories (n = 54 trials) Evidence A source
  SUCRA 82% for LDL cholesterol and 90% for total cholesterol, and LDL cholesterol 0.42 to 0.23 mmol/l lower than butter per 10 percent isocaloric exchange

Safety, limits and interactions

• Seven weeks of 15 to 20 g of linoleic acid a day as safflower oil raised systolic blood pressure by 2.61 percent while fish oil lowered it by 1.81 percent, and neither oil beat the other on any circulating inflammatory marker. — 39 adults aged 30 to 70 with abdominal obesity, 16 women and 23 men, crossover with a 9-week washout (n = 39) Evidence B source
  Systolic blood pressure relative change fish oil -1.81% against safflower oil +2.61%, P = 0.003, and no between-treatment difference in any circulating cytokine or chemokine
• Three adults who took safflower oil as a weight loss supplement developed acute liver failure, all three needed a liver transplant and one died 12 days after the operation, and four such cases were already in the literature. — three adults treated for acute liver failure at a Brazilian university hospital, with viral hepatitis, autoimmune disease and other drug causes excluded (n = 3) Evidence C source
  Acute liver failure in all three, liver transplantation in all three, one death 12 days after transplantation

What people believe that the data does not support

• In the one trial that replaced saturated fat with safflower oil in men who had already had a heart attack, deaths from all causes were higher in the safflower group, 17.6 percent against 11.8 percent, and adding those recovered data to a meta-analysis of linoleic acid trials left no cardiovascular benefit. — 458 men aged 30 to 59 with a recent coronary event, median follow-up 39 months, Sydney 1966 to 1973, plus an updated meta-analysis of linoleic acid trials (n = 458) Evidence A source
  All cause death 17.6% against 11.8%, hazard ratio 1.62, 95% CI 1.00 to 2.64; updated meta-analysis hazard ratio 1.33, 95% CI 0.99 to 1.79, for coronary death
• FDA lets labels say that about 1.5 tablespoons, 20 g, of a high oleic oil such as high oleic safflower oil may reduce heart disease risk when it replaces fats higher in saturated fat, and the wording FDA requires calls the evidence supportive but not conclusive. — US food labels for oils containing at least 70 percent oleic acid per serving, qualified health claim of 19 November 2018 Evidence E source
  20 g a day of a qualifying oil, which supplies about 15 g of oleic acid, and the oil must carry at least 70 percent oleic acid per serving

Who this works differently for

• 30 ml of high oleic safflower oil a day for 28 days did not significantly change blood lipids in 12 postmenopausal women, while coconut oil in the same women raised total cholesterol by 18.2 mg/dL. — 12 postmenopausal women aged about 59, each taking 30 ml of one oil for 28 days with a 28-day washout, keeping their usual diet (n = 12) Evidence B source
  Safflower oil no significant change in total cholesterol, LDL or HDL, coconut oil total cholesterol +18.2 mg/dL, LDL +13.5 mg/dL, HDL +6.6 mg/dL
• In 11 randomised trials of 2,200 infants where safflower oil was one of the comparison oils, enteral lipids may have little to no effect on mortality, risk ratio 1.19 with a 95% CI of 0.90 to 1.57, and the review rates that evidence very low certainty. — 2,200 mostly preterm infants on parenteral nutrition, at risk of parenteral nutrition associated liver disease, in the 11 randomised trials pooled for mortality (n = 2200) Evidence A source
  Mortality risk ratio 1.19, 95% CI 0.90 to 1.57, I2 0%, 11 studies; cholestasis risk ratio 0.65, 95% CI 0.13 to 3.14, 3 studies, all very low certainty

The bottom line

Two different products share the name. The FDA qualified health claim covers only oil with at least 70 percent oleic acid per serving. It lets a label say that about 1.5 tablespoons, 20 g, of such an oil may reduce the risk of coronary heart disease when it replaces fats higher in saturated fat. The wording the agency requires calls that evidence supportive but not conclusive. A qualified claim is the weakest thing FDA grants, and it exists because the evidence fell short of an authorised claim. Ordinary linoleic safflower oil, the kind most bottles hold, sits outside that claim entirely. In 12 postmenopausal women of about 59, taking 30 ml a day of high oleic safflower oil for 28 days did not significantly change total cholesterol, LDL or HDL. Coconut oil in the same women raised total cholesterol by 18.2 mg/dL. Twelve women is small enough that a modest real effect could hide in the noise, so that flat result is weak evidence of no effect rather than proof of none. The doses in these trials are supplement doses. Thirty-nine adults aged 30 to 70 with abdominal obesity took 15 to 20 g of linoleic acid a day as safflower oil for seven weeks, which is 24.5 to 32.7 ml of oil and far more than cooking puts on a plate. Their systolic blood pressure rose by 2.61 percent while fish oil in the same people lowered it by 1.81 percent, P = 0.003. Neither oil beat the other on any circulating inflammatory marker. A tablespoon of the oil weighs 14 g, so every per 100 g figure on a label is describing about seven tablespoons. That column gives 45.95 mg of vitamin E, 306 percent of the Daily Value. Across the eight USDA samples behind it the vitamin E ran from 32.67 to 58.25 mg and the vitamin K from 0.8 to 19.05 mcg. British CoFID data for the same oil put vitamin K at 3.4 mcg against the USDA 5.344 mcg, a 36 percent gap that nothing in either dataset resolves. The caution here is about the oil sold as a slimming supplement. Three adults treated for acute liver failure at a Brazilian university hospital had all been taking safflower oil for weight loss, with viral hepatitis, autoimmune disease and other drugs excluded. All three needed a liver transplant and one died 12 days after the operation, and four similar cases were already in the literature. A series of three rests on excluding other causes rather than on a comparison group, and the dose and the brand of supplement are not reported. This is concentrated oil swallowed as a supplement rather than oil in a pan. For infants the answer is that nobody knows. In 11 randomised trials of 2,200 mostly preterm infants on parenteral nutrition, where safflower oil was one of the comparison oils, the review could not tell whether enteral lipid supplements make any difference. The mortality risk ratio was 1.19 with a 95 percent CI of 0.90 to 1.57, and the Cochrane authors rate that estimate and every other one in the review very low certainty. One more difference between oils shows up before any of this. Adults in a crossover trial ate palm oil, safflower oil, a fish and safflower mixture and olive oil in turn. The apolipoprotein B-48 response differed between the oils, with olive oil highest after both meals, P = 0.003. That is a step in how fat is absorbed and packaged rather than a health outcome.

Sources

1. saff-r1-09 · food composition dataset
   Oil, safflower (FDC 1750350, Foundation): vitamin E 45.95 mg, saturated fat 7.651 g, vitamin K 5.344 mcg per 100 g; Foundation samples n=8 vitamin E 32.67-58.25, saturated fat 6.94-8.16, vitamin K 0.8-19.05
   USDA Foundation Foods, FoodData Central, record 1750350 · checked 2026-09-24
2. saff-r1-10 · portion dataset
   Safflower oil (2710189): 1 tablespoon 14 g, 1 cup 224 g
   USDA FNDDS, Safflower oil, record 2710189 · checked 2026-09-24
3. saff-r1-04 · randomised crossover
   The pattern of the apo B-48 response differed significantly among the dietary oils, with olive oil resulting in higher concentrations after both meals (P = 0.003).
   Am J Clin Nutr, 2002 · checked 2026-09-24
4. saff-r1-01 · network meta-analysis
   Safflower oil had the highest SUCRA value for LDL-C (82%) and TC (90%), followed by rapeseed oil (76% for LDL-C, 85% for TC); whereas, palm oil (74%) had the highest SUCRA value for TG, and coconut oil (88%) for HDL-C.
   J Lipid Res, 2018 · checked 2026-09-24
5. saff-r1-03 · randomised crossover
   While significant differences between treatments in relative change scores were found for systolic blood pressure (n-3 vs. n-6: -1.81% vs. 2.61%, P = 0.003), no differences between n-3 and n-6 were found for any circulatory inflammatory markers.
   J Lipid Res, 2025 · checked 2026-09-24
6. saff-r1-06 · case series
   All had a history of ingestion of this oil for weight loss...All 3 patients underwent liver transplantation: 2 had good postoperative evolution, and 1 died 12 days after the procedure.
   Transplant Proc, 2018 · checked 2026-09-24
7. saff-r1-02 · trial plus updated meta-analysis
   The intervention group (n=221) had higher rates of death than controls (n=237) (all cause 17.6% v 11.8%, hazard ratio 1.62 (95% confidence interval 1.00 to 2.64), P=0.05; cardiovascular disease 17.2% v 11.0%, 1.70 (1.03 to 2.80), P=0.04; coronary heart disease 16.3% v 10.1%, 1.74 (1.04 to 2.92), P=0.04).
   BMJ, 2013 · checked 2026-09-24
8. saff-r1-08 · regulatory decision
   Supportive but not conclusive scientific evidence suggests that daily consumption of about 1½ tablespoons (20 grams) of oils containing high levels of oleic acid, when replaced for fats and oils higher in saturated fat, may reduce the risk of coronary heart disease.
   FDA, Qualified Health Claim for Oleic Acid and Coronary Heart Disease, 19 November 2018 · checked 2026-09-24
9. saff-r1-05 · randomised crossover
   VCO significantly raised total cholesterol, TC (+18.2 ± 22.8 mg/dL), low-density lipoprotein (+13.5 ± 16.0 mg/dL), and high-density lipoprotein, HDL (+6.6 ± 7.5 mg/dL). SO did not significantly change lipid values.
   J Med Food, 2017 · checked 2026-09-24
10. saff-r1-07 · Cochrane review
    Enteral lipids may have little to no effect on mortality up to discharge and the first year of life, but the evidence is very uncertain (RR 1.19, 95% CI 0.90 to 1.57; I² = 0%; 11 studies, 2200 participants; very low-certainty evidence).
    Cochrane Database Syst Rev, 2026 · checked 2026-09-24