Soybean oil: the best evidence is about a hospital drip, not your frying pan

The strongest studies that name soybean oil are about a drip in a hospital. In a network meta-analysis of 47 randomized trials in hospitalized adults fed intravenously, fish oil emulsions cut the odds of infection against pure soybean oil emulsions (odds ratio 0.43, 90% credibility interval 0.29 to 0.63). In a meta-analysis of 26 trials in 1733 critically ill adults, less soybean oil in the drip did not change mortality (risk ratio 0.91, 95% CI 0.76 to 1.10) and shortened hospital stay by about 7 days. None of this tells you anything about salad dressing. For the oil you cook with, the evidence is two small feeding trials. In 90 Chinese women at high cardiovascular risk, three months of cooking with soybean oil left body weight no different from olive or camellia oil, and olive oil raised HDL cholesterol slightly more. The only US label claim is a qualified one. FDA says about 1.5 tablespoons (20.5 g) a day may lower heart disease risk if it replaces saturated fat without adding calories, and calls its own evidence supportive but not conclusive.

The numbers, per 3.5 oz (100 g)

As people eat it, per 3.5 oz (100 g)

How much is actually in it

• In 5 USDA samples, soybean oil held 12.2 mg of alpha-tocopherol (vitamin E) per 3.5 oz (100 g), ranging from 10.3 to 14.2 mg. — 5 samples, USDA Foundation Foods, Oil, soybean (FDC 748366) (n = 5) Evidence E source
  Vitamin E 12.2 mg per 3.5 oz (100 g) (range 10.3 to 14.2, median 12.6)
• USDA samples of soybean oil carried 153 mg of beta-sitosterol and 58 mg of campesterol per 3.5 oz (100 g), with up to 38% spread between samples. — 5 samples, USDA Foundation Foods, Oil, soybean (FDC 748366) (n = 5) Evidence E source
  Beta-sitosterol 153 mg (135 to 168), campesterol 58.0 mg (46.2 to 68.2) per 3.5 oz (100 g)

What a real portion looks like

• US dietary surveys count a tablespoon of soybean oil as 14 g and a cup as 224 g. — FNDDS record Soybean oil (food code 82108000, FDC 2710191), WWEIA category Salad dressings and vegetable oils (n = 3 portion measures) Evidence E source
  1 tablespoon = 14 g; 1 cup = 224 g; quantity not specified = 14 g

What your body absorbs

• In a randomized trial in 12 women, adding more soybean oil to a salad dressing (0 to 32 g) raised absorption of carotenoids and fat-soluble vitamins from the vegetables, highest at 32 g. — 12 young women, 5 single test salads with 0, 2, 4, 8 or 32 g soybean oil (n = 12) Evidence B source
  Absorption of all carotenoids and fat-soluble vitamins highest with 32 g oil (P < 0.002); linear dose response for alpha-carotene, lycopene, phylloquinone; wide individual variation

What it does to your health

• In a 3-month randomized feeding trial in 90 Chinese women at high cardiovascular risk, cooking with soybean oil did not change body weight differently from olive or camellia oil, while olive oil raised HDL slightly more. — 90 middle-aged and elderly Chinese women at high cardiovascular risk, 3 months, lunch and dinner provided (n = 90) Evidence B source
  Weight change: soybean 0.68 lb (0.31 kg), olive -0.29 lb (95% CI -1.4 to 0.79; metric: -0.13 kg, -0.62 to 0.36), camellia -1.6 lb (95% CI -3 to -0.15; metric: -0.72 kg, -1.38 to -0.07), no significant group difference; HDL higher with olive vs soybean (P = 0.03)
• In a 4-week randomized crossover pilot in 10 adults, 30 g a day of soybean oil in foods tended to raise blood linoleic acid without changing body weight or energy intake. — 4 men and 6 women aged 25 to 76, BMI 26 to 41, 30 g oil a day in 3 foods, 4 weeks per arm (n = 10) Evidence B source
  Plasma linoleic acid +2.35 (P = 0.07), erythrocytes +1.09 (P = 0.05) on soybean oil; no change in body weight (P > 0.40)

Safety, limits and interactions

• The FDA-approved Intralipid label warns that its soybean oil contains vitamin K1, which may counteract warfarin. — Patients on intravenous soybean oil emulsion taking vitamin K antagonists Evidence E source
  Anticoagulant activity may be counteracted
• A case report described warfarin failing to work in a woman receiving high-dose propofol, which contains 10% soybean oil, until the infusion stopped. — One 39-year-old woman with severe Crohn's disease (n = 1) Evidence D source
  Warfarin up to 30 mg daily did not anticoagulate until propofol stopped

What people believe that the data does not support

• FDA allows a qualified US label claim that about 1.5 tablespoons (20.5 g) of soybean oil a day may reduce coronary heart disease risk only if it replaces saturated fat without adding calories. — US food labeling Evidence E source
  Qualified health claim, supportive but not conclusive evidence

Who this works differently for

• In a network meta-analysis of 47 RCTs in hospitalized adults on intravenous nutrition, fish oil emulsions cut infection odds versus pure soybean oil emulsions (OR 0.43). — Hospitalized adults receiving parenteral nutrition, 47 RCTs (n = 47 RCTs) Evidence A source
  Infection OR 0.43 (90% CrI 0.29 to 0.63), sepsis OR 0.22 (0.08 to 0.59), hospital stay -2.31 days (-3.14 to -1.59) for fish oil vs soybean oil emulsions
• In a meta-analysis of 26 RCTs in 1733 critically ill adults, lowering omega-6 (soybean oil) in intravenous lipids did not reduce mortality but shortened hospital stay. — 1733 critically ill adults in 26 RCTs (n = 1733) Evidence A source
  Mortality RR 0.91 (95% CI 0.76 to 1.10); hospital stay -6.88 days (95% CI -11.27 to -2.49)
• A Cochrane review of 29 studies in 2037 preterm infants found no clear difference in liver cholestasis between fish oil and other lipid emulsions, including soybean oil ones. — Parenterally fed preterm infants under 37 weeks, 29 studies (n = 2037) Evidence A source
  Cholestasis RR 0.61 (95% CI 0.24 to 1.56), 4 studies, n = 328, low-quality evidence
• In a randomized pilot of 20 followed-up infants with intestinal failure, soybean oil versus mixed-oil intravenous lipids showed no difference in development at 3 years. — Neonates with intestinal failure on parenteral nutrition, soybean oil 1 g/kg/day vs mixed oil 3 g/kg/day, 3-year follow-up (n = 20) Evidence B source
  No difference in mean scores or developmental delay in all 5 ASQ-3 domains

What is still unknown

• FDA based the soybean oil heart claim on only four small randomized trials of 15 to 36 people each. — FDA review of evidence for the soybean oil qualified health claim (n = 4 trials) Evidence E source
  4 intervention studies, 15 to 36 subjects per study or group

The bottom line

Soybean oil helps you absorb what is in your vegetables. In a crossover trial in 12 young women, each ate five single salads dressed with 0, 2, 4, 8 or 1.1 oz (32 g) of soybean oil. Absorption of every carotenoid and fat-soluble vitamin was highest at 32 g, and some of the women barely responded at all. A tablespoon, which US surveys count as 14 g, carries about 1.7 mg of vitamin E. That figure comes from 5 USDA samples averaging 12.2 mg per 3.5 oz (100 g), ranging from 10.3 to 14.2 mg, and no second food table gave us a number to check it against. In a 4-week pilot in 10 adults, 30 g a day in foods left body weight unchanged. The rise in blood linoleic acid did not reach significance in plasma (P = 0.07). The heart claim stands on thin ground. FDA based it on four small randomized trials of 15 to 36 people each. We found no trial of what frying or reheating does to soybean oil itself, so this page says nothing about cooking losses. The warfarin warning that exists is about the intravenous form. The FDA-approved Intralipid label says the vitamin K1 in its soybean oil may counteract warfarin. A single case report describes warfarin failing in a woman on high-dose propofol, which is 10% soybean oil, until the infusion stopped. Neither involved eating the oil, and a clinical effect of kitchen amounts has not been shown. This page has no evidence on soy allergy and the oil, so it cannot say whether the oil is safe for people allergic to soy. For preterm babies fed intravenously, a Cochrane review of 29 studies found no clear difference in liver cholestasis between fish oil and other emulsions (risk ratio 0.61, 95% CI 0.24 to 1.56, low-quality evidence).

Compared with other foods

Sources

1. sbol-daily-r3-01 · dataset
   Oil, soybean (Foundation Foods 2026-04-30): Vitamin E (alpha-tocopherol) 12.2 MG, range 10.3-14.2, median 12.6, samples 5
   USDA FoodData Central, Foundation Foods, release 2026-04-30, FDC 748366 (Oil, soybean) · checked 2026-09-29
2. sbol-daily-r3-02 · dataset
   Oil, soybean (Foundation Foods 2026-04-30): Beta-sitosterol 153.0 MG, range 135.0-168.0, median 152.0, samples 5; Campesterol 58.0 MG, range 46.2-68.2, median 58.3, samples 5
   USDA FoodData Central, Foundation Foods, release 2026-04-30, FDC 748366 (Oil, soybean) · checked 2026-09-29
3. sbol-daily-r3-03 · dataset
   Soybean oil (FNDDS 82108000): 1 cup 224 g; 1 tablespoon 14 g; Quantity not specified 14 g
   USDA FNDDS 2021-2023, food code 82108000 (Soybean oil, FDC 2710191) · checked 2026-09-29
4. sbol-daily-r3-04 · RCT
   Women (n = 12) each consumed 5 vegetable salads with salad dressings containing 0, 2, 4, 8, or 32 g soybean oil. ... Absorption of all carotenoids and fat-soluble vitamins was highest with 32 g oil (P < 0.002).
   Am J Clin Nutr, 2017 · checked 2026-09-29
5. sbol-daily-r3-05 · RCT
   We found no significant difference in the changes of body weight among the SO group (mean change, 0.31 kg; 95% CI, -0.88 to 0.27), the OO group (mean change, -0.13 kg; 95% CI, -0.62 to 0.36), and the CSO group (mean change, -0.72 kg; 95% CI, -1.38 to -0.07). For secondary outcomes, the OO group showed a slight increase in HDL cholesterol (P = 0.03), while the CSO group showed greater reduction in aspartate aminotransferase (AST) (P = 0.02) when compared with the SO group.
   Food Funct, 2022 · checked 2026-09-29
6. sbol-daily-r3-06 · RCT
   After 4 wk of consuming 3 foods/d, plasma linoleic acid decreased during the palm oil intervention (-1.60, P = 0.04), whereas it tended to increase in plasma (2.35, P = 0.07) and erythrocytes (1.09, P = 0.05) during the soybean oil intervention. ... There was no difference in energy intake between the 2 interventions (P = 0.65) and no change in bodyweight during either intervention (P > 0.40).
   J Nutr, 2026 · checked 2026-09-29
7. sbol-daily-r3-13 · drug label
   Soybean oil in Intralipid contains vitamin K 1 which may counteract the anticoagulant activity of vitamin K antagonists such as warfarin.
   Intralipid (i.v. fat emulsion), FDA-approved label (DailyMed), effective 2025-11-05 · checked 2026-09-29
8. sbol-daily-r3-14 · case report
   High-dose propofol infusion was associated with reversal of enteral and intravenous warfarin anticoagulation in a 39-year-old woman with severe Crohn's disease. Despite increasing the daily dose of warfarin to 30 mg, anticoagulation was not achieved until propofol was discontinued.
   Pharmacotherapy, 1997 · checked 2026-09-29
9. sbol-daily-r3-07 · regulation
   Supportive but not conclusive scientific evidence suggests that eating about 1 ½ tablespoons (20.5 grams) daily of soybean oil, which contains unsaturated fat, may reduce the risk of coronary heart disease. To achieve this possible benefit, soybean oil is to replace saturated fat and not increase the total number of calories you eat in a day.
   FDA, Qualified Health Claim: Soybean Oil and Reduced Risk of Coronary Heart Disease, letter of enforcement discretion, Docket FDA-2016-Q-0995, 2017 · checked 2026-09-29
10. sbol-daily-r3-09 · network meta-analysis
    1651 publications were retrieved in the original search, 47 RCTs were included in the NMA. For FO-ILEs, very highly credible reductions in infection risk versus SO-ILEs [odds ratio (OR) = 0.43 90% credibility interval (CrI) (0.29-0.63)], MCT/soybean oil-ILEs [0.59 (0.43-0.82)], and OO-ILEs [0.56 (0.33-0.91)], and in sepsis risk versus SO-ILEs [0.22 (0.08-0.59)], as well as substantial reductions in hospital length of stay versus SO-ILEs [mean difference (MD) = -2.31 (-3.14 to -1.59) days] and MCT/SO-ILEs (-2.01 (-2.82 to -1.22 days) were shown.
    Clin Nutr, 2023 · checked 2026-09-29
11. sbol-daily-r3-10 · meta-analysis
    Omega-6 FA sparing PN in comparison with traditional lipid emulsions did not decrease overall mortality (20 studies; risk ratio [RR] 0.91; 95% CI 0.76, 1.10; p = 0.34) but hospital length of stay was substantially reduced (6 studies; weighted mean difference [WMD] - 6.88; 95% CI - 11.27, - 2.49; p = 0.002).
    Crit Care, 2022 · checked 2026-09-29
12. sbol-daily-r3-11 · meta-analysis
    We included 29 studies (n = 2037) in this review. ... In preterm infants, meta-analysis found no evidence of a difference in the incidence of PNALD/cholestasis (Cbil cut-off: 2 mg/dl) between fish oil-LEs and all non-fish oil LEs (typical risk ratio (RR) 0.61, 95% confidence interval (CI) 0.24 to 1.56; typical risk difference (RD) -0.03, 95% CI -0.08 to 0.02; 4 studies; n = 328; low-quality evidence).
    Cochrane Database Syst Rev, 2019 · checked 2026-09-29
13. sbol-daily-r3-12 · RCT
    For this follow-up analysis, 20 patients were included, 10 in each group. ... There was also no difference when comparing the mean scores or rate of developmental delay for all five Ages & Stages Questionnaire-3 domains between treatment groups.
    JPEN J Parenter Enteral Nutr, 2026 · checked 2026-09-29
14. sbol-daily-r3-08 · regulation
    All four intervention studies were small (15-36 subjects per study or treatment group) randomized controlled trials of high or moderate methodological quality.
    FDA, Qualified Health Claim: Soybean Oil and Reduced Risk of Coronary Heart Disease, letter of enforcement discretion, Docket FDA-2016-Q-0995, 2017 · checked 2026-09-29