Butter or peanut butter: the swap nobody has tested

Nobody has put butter and peanut butter side by side in a trial or a cohort. Our search found none. What exists is a set of comparisons that each stop one step short of the question. Butter has been measured against oils. Across 54 feeding trials of at least three weeks, moving 10 percent of daily energy from butter to any of nine plant oils, or even to beef fat, lowered LDL cholesterol by 0.23 to 0.42 mmol/L, or about 9 to 16 mg/dL on a US test. Butter finished last for LDL of everything tested. In 221,054 US health professionals followed for up to 33 years, replacing 10 g of butter a day with the same weight of plant oil went with a 17 percent lower death rate. That is an association in people already eating that way. Neither study used peanut butter. On its own, butter is a weaker story: across 636,151 people, each tablespoon a day was tied to a 1 percent higher death rate, no change in cardiovascular disease and 4 percent lower diabetes, with no randomised trial behind any of it. Those cohorts asked only what happens as butter goes up, with no statement about what goes down to make room for it. The number you get for butter depends on what you compare it with. Peanut butter has its own evidence, and it is thinner. Pooled cohorts tie it to 13 percent lower type 2 diabetes incidence, a result the authors themselves grade as very low certainty. In 120,852 Dutch adults, 10 g or more of nuts a day went with 23 percent lower mortality while peanut butter showed no link at all. The EU refused a label claim that two tablespoons a day help with cholesterol and heart health. So the case for the swap rests on oils and whole nuts standing in for peanut butter, and the one large study that looked at peanut butter by itself found nothing. Every number here comes from the food as people eat it, and a tablespoon is 16 g of peanut butter but 14 g of butter.

What a real portion looks like

• A tablespoon is 16 g of peanut butter but 14 g of butter. On one slice of bread the usual spread is 16 g of peanut butter against 7 g of butter, more than twice the weight. — US household measures, FNDDS 2021-2023 (n = FNDDS 5,432 foods and 22,046 household measures) Evidence E source
  Peanut butter (fdc_id 2707537): 1 tablespoon 16 g, guideline amount per slice 16 g, per sandwich 32 g. Butter, stick (fdc_id 2710155): 1 tablespoon 14 g, 1 pat 7 g, guideline amount per slice 7 g, per regular sandwich 14 g

What your body absorbs

• Adding 42.5 g of peanut butter to a 75 g carbohydrate breakfast lowered the glucose response to the next meal and the desire to eat in 15 obese women, while the same weight of whole peanuts did not reach significance. — 15 obese women at high risk of type 2 diabetes, randomised crossover, breakfast plus standard lunch, 490 minutes of sampling (n = 15) Evidence B source
  Glucose IAUC 240 to 490 min and NEFA IAUC 0 to 240 min lower after peanut butter than control; PYY, GLP-1 and CCK higher; desire to eat lower; whole peanuts similar but not significant
• Grinding peanuts into butter releases most of their fat: on a high-fibre diet 17.8 percent of dietary fat left in the stool after whole peanuts against 7.0 percent after peanut butter, in a feeding study summarised by a 2023 systematic review. — adults in a controlled feeding study (Levine and Silvis, 1980) with high- and low-fibre diets, one of 21 studies in the review (n = 21 studies) Evidence C source
  Fecal fat excretion: whole peanuts 17.8% (high fibre) and 16.8% (low fibre); peanut butter 7.0% and 4.2%; peanut oil 4.5% and 1.8%

What cooking and storage change

• Eating 56 g a day of peanuts as raw, roasted, salted, honey roasted or as peanut butter for 4 weeks made no difference to body weight or fasting lipids between the forms in 118 adults. — 118 adults (47 men, 71 women, mean age 29, mean BMI 30) in Brazil, Ghana and the United States, 4 weeks (n = 118) Evidence B source
  No differential effect of form or processing on body weight or fasting lipids; HDL rose and TC, LDL and TAG fell only in those with elevated baseline lipids

What it does to your health

• Swapping 10 percent of daily energy from butter for any of nine plant oils, or even beef fat, lowered LDL cholesterol by 0.23 to 0.42 mmol/L in a network meta-analysis of 54 trials lasting at least 3 weeks. — adults in 54 randomised feeding trials of at least 3 weeks comparing isocaloric exchanges of oils and solid fats (n = 54 trials) Evidence A source
  LDL-C lower by 0.23 to 0.42 mmol/L per 10 percent energy exchange away from butter; butter ranked last
• Replacing 10 g of butter a day with the same amount of plant oil was associated with 17 percent lower total mortality over up to 33 years in 221,054 US health professionals. — 221,054 US women and men in three cohorts free of cancer, CVD, diabetes and neurodegenerative disease at baseline, followed up to 33 years (n = 221054) Evidence C source
  Substitution of 10 g/day butter with plant oils: total mortality HR 0.83 (0.79 to 0.86), cancer mortality HR 0.83 (0.76 to 0.90)
• Pooled cohorts give a 4 percent lower cardiovascular risk for butter swapped to olive oil and a 6 percent lower all-cause mortality for dairy swapped to nuts, both at moderate certainty and both from 3 cohorts. Neither butter nor peanut butter is counted on its own in that nut estimate. — 24 prospective cohorts in 37 publications, adults, substitution models (n = 24 cohorts) Evidence C source
  Butter to olive oil: CVD SHR 0.96 (0.95 to 0.98), 3 cohorts; dairy to nuts: all-cause mortality SHR 0.94 (0.91 to 0.97), 3 cohorts
• Pooled prospective cohorts tie peanut butter to a 13 percent lower incidence of type 2 diabetes, an association the authors themselves grade as very low certainty. — adults in 5 prospective cohorts and 3 cross-sectional studies, searched to August 2020 (n = 8 studies) Evidence C source
  Peanut butter and T2D incidence RR 0.87 (0.77 to 0.98), I2 50.6%; peanuts and tree nuts: no association

Safety, limits and interactions

• In 30 highly peanut-allergic children, a minute of skin contact with peanut butter or 10 minutes of smelling it caused no systemic or breathing reaction in any child, only local redness or itch at the skin site in 10. The authors hold this to casual contact and did not test eating, larger exposures, peanut flour or roasted peanuts. — 30 children with peanut allergy, median age 7.7 years, median peanut IgE above 100 kIU/L, double-blind placebo-controlled exposures (n = 30) Evidence B source
  0 of 30 systemic or respiratory reactions; skin-site erythema in 3, pruritus in 5, wheal-and-flare in 2; at least 90 percent of such children would not react systemically (96 percent confidence)
• A systematic review of 6 surveys found aflatoxin B1 in Iranian peanuts and peanut products at levels giving a margin of exposure below 10,000, which the review authors read as a chance of a cancer risk. — peanut, peanut butter, peanut oil and snack peanut samples from Iranian provinces, 6 studies selected from 43 (n = 6 studies) Evidence C source
  Margin of exposure below 10,000; contamination highest in Mashhad and Tehran samples

What people believe that the data does not support

• Across 9 cohorts and 636,151 people, each tablespoon of butter a day was tied to a 1 percent higher death rate, no change in cardiovascular disease, and 4 percent lower diabetes incidence. These cohorts measured butter on its own, against no named substitute, and no randomised trial tested any of it. — 636,151 adults in 15 country-specific prospective cohorts, median butter intake 4.5 to 46 g/day, 6.5 million person-years (n = 636151) Evidence C source
  Per 14 g/day: all-cause mortality RR 1.01 (1.00 to 1.03), any CVD RR 1.00 (0.98 to 1.02), diabetes RR 0.96 (0.93 to 0.99)
• In 120,852 Dutch adults, 10 g or more of nuts a day went with 23 percent lower total mortality, but peanut butter showed no such link at all. — 120,852 Dutch men and women aged 55 to 69 at baseline (1986), 10-year mortality follow-up, case-cohort analysis on 8,823 deaths (n = 120852) Evidence C source
  Nuts 10+ g/day vs none: total mortality HR 0.77 (0.66 to 0.89); peanut butter: no association
• The claim that two tablespoons of peanut butter a day helps achieve healthy cholesterol and heart health was refused in the EU register. — EU food labelling Evidence E source
  Claim POL-HC-7733 for peanuts, peanut butter and peanut oil carries the status Non-authorised

Who this works differently for

• 43 g of peanut butter a day for 6 months did not change weight, fat mass, grip strength or gait speed in 120 adults over 65 at risk of falls, though sit-to-stand time improved by 1.2 seconds. — 120 community-dwelling Australians aged 65 and over at risk of falls, mean age 76, 70 percent women, 6 months, intention to treat (n = 120) Evidence B source
  Gait speed (primary): no effect; five times sit-to-stand -1.23 s (-2.09 to -0.37); muscle power +22.0 W (7.1 to 36.9); weight, BMI, fat and lean mass: no difference

The bottom line

If you are choosing for your cholesterol, the trials point away from butter and towards plant oil, and peanut butter was not the thing they poured in its place. If you want peanut butter anyway, here is what has been measured on it. Grinding peanuts releases their fat. On a high-fibre diet 17.8 percent of dietary fat left in the stool after whole peanuts against 7.0 percent after peanut butter, in one 1980 feeding study quoted inside a 2023 review, so more of the fat in the jar reaches you than the same peanuts whole would give. That did not show up on the scales in the one trial that tried. In 118 adults, 56 g a day of peanuts as raw, roasted, salted, honey roasted or as peanut butter for four weeks made no difference to weight or fasting lipids between the forms. Lipids improved only in the people who started high. In 120 Australians over 65, 43 g a day for six months changed neither weight nor fat mass, and gait speed, the measure the trial set out to change, did not move either. What did move was getting out of a chair five times, 1.2 seconds faster, and muscle power by 22 W, in a trial run open-label against usual care. The finding that peanut butter at breakfast blunted the glucose rise after lunch and the urge to eat comes from 15 obese women over a single day of sampling. It is a mechanism worth knowing and too small to plan a diet on. Watch the portion. On a slice of bread the usual spread is 16 g of peanut butter against 7 g of butter, more than twice the weight, so a swap on toast is a swap of amount as well as of food. None of this is for anyone with a peanut allergy. The one trial in highly allergic children tested a minute of touching and ten minutes of smelling, and 0 of 30 had a systemic or breathing reaction, with local redness or itch in 10. It says nothing about eating. A review of Iranian peanut products found aflatoxin B1, a mould toxin, at levels its own authors read as a chance of a cancer risk. We have no card on peanuts sold anywhere else, so we cannot tell you how your jar compares. One more gap is ours. No card under this page says who should limit either spread or by how much, nothing on pregnancy, on young children, on salt, on any drug. Read that as unanswered rather than as safe.

Sources

1. bpb-c1-14 · reference dataset
   FNDDS 2707537 Peanut butter — 1 tablespoon 16 g · Guideline amount per slice of bread/roll 16 g · Guideline amount per sandwich 32 g · FNDDS 2710155 Butter, stick — 1 tablespoon 14 g · 1 pat 7 g · Guideline amount per slice of bread/roll 7 g
   USDA FNDDS 2021-2023 · checked 2026-09-23
2. bpb-c1-07 · randomised crossover clinical trial
   A group of fifteen women participated in a randomised cross-over clinical trial in which 42·5 g of whole peanuts without skins (WP), peanut butter (PB) or no peanuts (control) were added to a 75 g available carbohydrate-matched breakfast meal. ... Postprandial NEFA incremental AUC (IAUC) (0-240 min) and glucose IAUC (240-490 min) responses were lower for the PB breakfast compared with the control breakfast. ... WP led to similar but non-significant effects. ... The greater bioaccessibility of the lipid component in PB is probably responsible for the observed incremental post-ingestive responses between the nut forms.
   Br J Nutr, 2013 · checked 2026-09-23
3. bpb-c1-09 · systematic review with narrative synthesis of human feeding studies and in vitro studies
   A high-fiber diet showed more dietary fat in the feces after consumption of whole peanuts (17.8%) and peanut butter (7.0%) than a low-fiber diet (whole peanuts: 16.8%; peanut butter: 4.2%) [37].
   Adv Nutr, 2023 · checked 2026-09-23
4. bpb-c1-08 · multicentre randomised controlled trial
   One hundred and eighteen adults (forty-seven males and seventy-one females; age 29.2 (sd 8.4) years; BMI 30.0 (sd 4.5) kg/m2) from Brazil, Ghana and the United States were randomised to consume 56 g of raw unsalted (n 23), roasted unsalted (n 24), roasted salted (n 23) or honey roasted (n 24) peanuts, or peanut butter (n 24) daily for 4 weeks. Peanut form and processing did not differentially affect body weight or fasting plasma lipid responses in the total sample.
   Br J Nutr, 2010 · checked 2026-09-23
5. bpb-c1-01 · systematic review and network meta-analysis of randomised trials
   Each 10% of dietary energy from butter replaced with an equivalent amount of safflower, sunflower, rapeseed, flaxseed, corn, olive, soybean, palm, and coconut oil, and beef fat was more effective in reducing LDL-C (−0.42 to −0.23 mmol/l).
   J Lipid Res, 2018 · checked 2026-09-23
6. bpb-c1-03 · prospective cohort study, substitution analysis
   Substituting 10-g/d intake of total butter with an equivalent amount of total plant-based oils was associated with an estimated 17% reduction in total mortality (HR, 0.83; 95% CI, 0.79-0.86; P < .001) and a 17% reduction in cancer mortality (HR, 0.83; 95% CI, 0.76-0.90; P < .001).
   JAMA Intern Med, 2025 · checked 2026-09-23
7. bpb-c1-04 · systematic review and meta-analysis of prospective cohort studies with GRADE
   There was moderate CoE for a lower risk of CVD when substituting processed meat with nuts [SHR (95% CI): 0.73 (0.59, 0.91), n = 8 cohorts] ... and butter with olive oil [0.96 (0.95, 0.98), n = 3]. ... dairy with nuts [0.94 (0.91, 0.97), n = 3]
   BMC Med, 2023 · checked 2026-09-23
8. bpb-c1-06 · systematic review and meta-analysis of observational studies
   Meta-analyses of prospective cohort studies showed an inverse association between peanut butter consumption and T2D incidence (RR: 0.87; 95% CI: 0.77, 0.98; I2 = 50.6%; Pheterogeneity = 0.16), whereas no association was observed between peanuts or tree nuts and T2D. ... The certainty of the evidence using NutriGrade was very low for all the exposures.
   Am J Clin Nutr, 2021 · checked 2026-09-23
9. bpb-c1-11 · double-blind, placebo-controlled randomised exposure challenge
   Thirty children underwent the challenges (median age, 7.7 years; median peanut IgE level, >100 kIU/L; 13 with prior history of contact and 11 with inhalation reactions). None experienced a systemic or respiratory reaction. Erythema (3 subjects), pruritus without erythema (5 subjects), and wheal-and-flare reactions (2 subjects) developed only at the site of skin contact with peanut butter.
   J Allergy Clin Immunol, 2003 · checked 2026-09-23
10. bpb-c1-12 · systematic review of food contamination surveys
    The contamination of AFB1 in peanut was high in Mashhad and Tehran compared with the other cities. The value of MOE was calculated less than 10,000. The results of MOE indicate that there are chances of the risk of developing cancer and these products may not be safe.
    Rev Environ Health, 2022 · checked 2026-09-23
11. bpb-c1-02 · systematic review and meta-analysis of prospective cohort studies
    No RCTs were identified. Butter consumption was weakly associated with all-cause mortality (N = 9 country-specific cohorts; per 14g(1 tablespoon)/day: RR = 1.01, 95%CI = 1.00, 1.03, P = 0.045); was not significantly associated with any cardiovascular disease (N = 4; RR = 1.00, 95%CI = 0.98, 1.02; P = 0.704) ... and was inversely associated with incidence of diabetes (N = 11; RR = 0.96, 95%CI = 0.93, 0.99; P = 0.021).
    PLoS One, 2016 · checked 2026-09-23
12. bpb-c1-05 · prospective cohort study with meta-analysis of cohorts
    When comparing those consuming 0.1-<5, 5-<10 and 10+ g nuts/day with non-consumers, multivariable hazard ratios for total mortality were 0.88, 0.74 and 0.77 [95% confidence interval (CI), 0.66-0.89], respectively (Ptrend = 0.003). ... Peanuts and tree nuts were inversely related to mortality, whereas peanut butter was not.
    Int J Epidemiol, 2015 · checked 2026-09-23
13. bpb-c1-13 · regulatory register entry
    POL-HC-7733 Peanuts, Peanut butter, Peanut oil Consuming a small handful, about 30 grams, of peanuts a day as part of a diet low in saturated fat helps to achieve healthy cholesterol levels and promotes heart health. Consuming two tablespoons, about 30g, of peanut butter a day as part of a diet low in saturated fat helps to achieve healthy cholesterol levels and promotes heart health. Consuming two tablespoons, about 30 grams of peanut oil a day as part of a diet low in saturated fat helps to achieve healthy cholesterol levels and promotes heart health. Non-authorised
    EU Register on nutrition and health claims, claim POL-HC-7733, snapshot 2026-09-21 · checked 2026-09-23
14. bpb-c1-10 · randomised controlled trial, usual care control
    One hundred and twenty older adults (aged ≥ 65 years) at risk of falls were randomly assigned to receive peanut butter (43 g/day, n = 60) or maintain usual care (control, n = 60) for 6 months. ... At 6 months, there were no significant treatment effects on the primary outcome of gait speed or other measures of physical function (p > 0.05), with the exception that 5STS time and muscle power improved significantly more in the peanut butter compared to control group (estimated treatment effect: time, -1.23 s [95% CI, -2.09, -0.37], p = 0.006; absolute power, 22.0 W [95% CI: 7.1 to 36.9], p = 0.004; relative power, 0.27 W/kg [95% CI: 0.10 to 0.45], p = 0.002). Changes in HGS, KES, weight, BMI, total fat mass, total lean mass or appendicular lean mass did not differ between groups.
    J Cachexia Sarcopenia Muscle, 2026 · checked 2026-09-23