Skip to content
HabitSeek
Back to results

Replace saturated fat with unsaturated fat

Deliberately swapping the saturated-fat foods in your diet — butter, lard, fatty meat, hard cheese, cream — for unsaturated-fat alternatives such as olive, rapeseed or sunflower oil, nuts, seeds and oily fish, rather than just cutting fat out.

At a glance

  • Evidence: Across 15 randomised trials in about 59,000 participants, reducing saturated fat for at least two years cut combined cardiovascular events by 21 percent.Combined cardiovascular events · Moderate certainty · Randomized intervention
  • What’s uncertain: Trial results vary substantially; the size of the benefit depends partly on how much saturated fat was reduced.
  • First step: Replace rather than simply remove fats: use olive or rapeseed oil instead of butter, ghee or lard where practical.
  • Before you try it: Do not read the evidence as saying that saturated fat is harmless or that any amount of it is fine. All precautions
Read the full takeaway

The substitution

This habit is really about the substitution, not the reduction. The Cochrane review of 15 long-term randomised trials in about 59,000 participants found that reducing saturated fat lowered combined cardiovascular events by 21 percent, with a number needed to treat of 56 in primary prevention and 32 in secondary prevention, and no clear effect on all-cause or cardiovascular mortality.

Replacement foods

What the saturated fat was replaced with mattered: the earlier 2015 review found the event reduction confined to trials that substituted polyunsaturated fat and no effect in trials substituting carbohydrate, and the 2020 update could no longer confirm that contrast, concluding instead that replacing the energy with either polyunsaturated fat or carbohydrate appeared useful while the effects of monounsaturated replacement were unclear. Controlled feeding trials show why the distinction is real — the total-to-HDL cholesterol ratio is unchanged when carbohydrate replaces saturated fat but falls when cis unsaturated fat does, although the same paper warns that these risk markers should not in themselves be treated as changes in risk, and cohort substitution models find myocardial infarction risk rises when saturated fat is replaced with high-glycaemic-index carbohydrate.

Practical interpretation

The practical conclusion is that 'eat less saturated fat' is not a usable instruction on its own, and swapping butter for white bread or sugary foods is not the trade the trials supported.

Olive oil, avocado, walnuts and olives grouped in the center of a kitchen table
NutritionLong-term healthUnder 10 min · Easy · Low
Reviewed Sep 23, 2026
Save

Scientific evidence

What the research shows

Outcome assessments based on research, including its uncertainties and limitations.

Combined cardiovascular events

Finding: Benefit supportedCertainty: Moderatestudies found a benefit from doing this

Evidence basis: randomised trials

Across 15 randomised trials in about 59,000 participants, reducing saturated fat for at least two years cut combined cardiovascular events by 21 percent.

Population: Adults with or without cardiovascular disease in 15 randomised controlled trials (16 comparisons, about 59,000 participants) with intervention lasting at least 24 months; trials that were multifactorial, or in people who were acutely ill, pregnant or breastfeeding, were excluded.

Research details

Summary

The Cochrane review of randomised trials of reducing saturated fat intake, updated in 2020, included 15 trials with 16 comparisons and about 59,000 participants, with interventions lasting at least 24 months. Reducing dietary saturated fat reduced the risk of combined cardiovascular events by 21 percent (risk ratio 0.79; 95% CI 0.66 to 0.93; 11 trials; 53,300 participants of whom 8 percent had a cardiovascular event; I²=65 percent; GRADE moderate-quality evidence). Meta-regression suggested that greater reductions in saturated fat, reflected in greater reductions in serum cholesterol, produced greater reductions in cardiovascular event risk and explained most of the between-trial heterogeneity. The number needed to treat for an additional beneficial outcome was 56 in primary prevention trials over about four years and 32 in secondary prevention trials. There was little or no effect on all-cause mortality (RR 0.96; 95% CI 0.90 to 1.03) or cardiovascular mortality (RR 0.95; 95% CI 0.80 to 1.12), and there was no evidence of harmful effects of reducing saturated fat intake.

Why this certainty

Rated Moderate, matching the GRADE assessment reported by the review. Randomised allocation and cardiovascular event endpoints are genuine strengths and this is the strongest trial evidence in the entry, so the rating is not lower. It is not Strong because heterogeneity was high, the effect was on a composite endpoint rather than mortality, the trials were unblinded dietary interventions with differing comparators, and the authors' own confidence in the replacement-nutrient contrast weakened between the 2015 and 2020 versions of the review.

Limitations

Heterogeneity between trials was substantial (I²=65 percent), which means the trials did not agree closely on the size of the effect, and the reduction was driven by the degree of saturated fat reduction achieved rather than by any single trial. The benefit was on combined cardiovascular events, a composite, and neither all-cause nor cardiovascular mortality fell, so the trial evidence does not show that this habit extends life. Several included trials predate modern statin and antihypertensive treatment and used different background diets, and blinding of a dietary intervention is impossible, so performance bias cannot be excluded. The 2020 update also lost the contrast between replacement nutrients that the 2015 version had reported, which means the mechanism is less clearly established in the trial data than it once appeared.

Effect estimate

Combined cardiovascular events risk ratio 0.79 (95% CI 0.66 to 0.93) with reduction of saturated fat for at least two years; number needed to treat 56 in primary prevention and 32 in secondary prevention.

Sources for this outcome:

Type 2 diabetes and myocardial infarction in nutrient-substitution models

Finding: Mixed evidenceCertainty: Limitedstudies disagree, or the effect depends on who is studied

Evidence basis: observational studies (association, not proof of cause)

Replacing saturated fat with polyunsaturated fat was associated with 14 percent lower type 2 diabetes risk, but the same substitution in the largest trial review had uncertain effects on glucose outcomes.

Population: Adults in prospective cohorts — 83,648 women from the Nurses' Health Study, 88,610 women from Nurses' Health Study II and 41,771 men from the Health Professionals Follow-Up Study for the diabetes substitution analysis, and 53,644 Danish adults for the myocardial infarction analysis — plus participants in the long-term randomised trials of saturated fat reduction for the diabetes diagnosis outcome.

Research details

Summary

In three US prospective cohorts following 214,029 men and women for 4.93 million person-years, with 18,442 incident type 2 diabetes cases, diabetes risk in an isocaloric substitution model was 14 percent lower (95% CI 5 to 21 percent) when linoleic acid replaced saturated fats at 5 percent of energy, and 17 percent lower (95% CI 9 to 24 percent) when it replaced trans fats at 2 percent of energy. The same substitution model found that replacing saturated fat with carbohydrate was also associated with lower diabetes risk (9 percent lower, 95% CI 17 percent lower to 0.1 percent lower, P=0.047), which cuts against a simple reading in which carbohydrate is always the wrong replacement. Conversely, in a Danish cohort of 53,644 adults with 1,943 incident myocardial infarctions over a median 12 years, replacing saturated fatty acids with high-glycaemic-index carbohydrate was associated with higher myocardial infarction risk (hazard ratio 1.33; 95% CI 1.08 to 1.64 per 5 percent increment of energy), while replacement with low-glycaemic-index carbohydrate showed a non-significant inverse association (HR 0.88; 95% CI 0.72 to 1.07). The Cochrane review of saturated fat reduction found little or no effect on diabetes diagnosis and, in its 2015 version, judged that effects of replacement with monounsaturated fat were unclear because only one small trial was located, and that no clear benefit of substituting carbohydrate for saturated fat had been shown.

Why this certainty

Rated Limited. The direction is consistent across cohort substitution models, the diabetes estimate is precise and dose-specific, and the Danish cohort provides an independent test in the opposite direction for carbohydrate replacement, which makes the replacement-nutrient story coherent. Certainty is held at Limited because these are modelled substitutions rather than interventions, the randomised trial evidence for glucose outcomes is null or unclear, and the monounsaturated fat question cannot be answered from the retrieved trials at all.

Limitations

Every estimate here is observational and comes from substitution models fit to food-frequency questionnaire data, which estimate what would happen if one nutrient were swapped for another within the same reported diet rather than what happened when people actually changed what they ate. The diabetes finding is specific to linoleic acid and to a modelled 5 percent of energy swap, and the Danish myocardial infarction result is a single cohort with relatively few events in the substitution categories, so it needs replication. The trial evidence points the other way from a clean effect: the Cochrane review found little or no effect on diabetes diagnosis, and the 2015 version could not establish a monounsaturated fat comparison at all. Diet was self-reported in all of these datasets, so measurement error is likely to dilute rather than exaggerate the associations. The outcome also mixes two designs: the substitution estimates are modelled from cohort data while the diabetes-diagnosis result comes from randomised trials, so the two halves are not directly comparable and the evidence basis recorded here is the observational design that produced the estimates.

Effect estimate

Type 2 diabetes hazard ratio 0.86 (14 percent lower; 95% CI 5 to 21 percent lower) when linoleic acid isocalorically replaced saturated fat at 5 percent of energy; myocardial infarction hazard ratio 1.33 (95% CI 1.08 to 1.64) per 5 percent of energy when high-glycaemic-index carbohydrate replaced saturated fatty acids.

Sources for this outcome:

What people say

Who discusses this?

Public statements, with positions specific to each source and topic. These views do not determine the scientific evidence assessment.

What the research shows ↑

No sourced public statements added for this habit yet. This does not mean no one discusses it.

Practical steps and precautions

How to try it

  1. Change the fat, do not just remove it: use olive or rapeseed oil in place of butter, ghee or lard for cooking and spreading, and nuts, seeds or oily fish in place of processed meat or hard cheese where you would otherwise use them.
  2. For soft spreads, choose a spread high in cis-unsaturated fat rather than a butter blend, which is the closest food-level equivalent to what the substitution trials actually tested.
  3. When you cut back on a saturated-fat food, decide in advance what replaces it. Replacing it with high-glycaemic-index carbohydrate has been associated with higher myocardial infarction risk in one cohort (Jakobsen 2010; low-glycaemic-index carbohydrate showed a non-significant inverse association), while the Cochrane review found no significant difference between polyunsaturated fat and carbohydrate replacement and Zong 2019 found carbohydrate substitution was associated with 9% lower diabetes risk - so the evidence supports avoiding refined carbohydrate specifically, not carbohydrate in general.
  4. Judge the change by the pattern over a week rather than by a single meal; the trials that showed an effect ran for at least two years, and the trial with the largest event reduction sustained the dietary change throughout.

Precautions

  • Do not read the evidence as saying that saturated fat is harmless or that any amount of it is fine. What the trials tested was replacing saturated fat with a defined substitute, not removing fat from the diet or adding more of it, and no included trial tested saturated fat from any particular food.
  • Total fat is energy-dense whatever its type. Adding oils and nuts on top of an unchanged diet increases calorie intake, so make the swap rather than the addition.
  • If you take warfarin or another vitamin K antagonist, a large sustained change in vegetable oil intake can alter vitamin K intake and destabilise anticoagulation; discuss the change with the clinician managing it.
  • People with gallstone disease or fat malabsorption may tolerate dietary fat changes poorly, and anyone with a history of pancreatitis should take dietary fat advice from their clinical team rather than from general guidance.

References

Reduction in saturated fat intake for cardiovascular disease.

Hooper L, Martin N, Jimoh OF, Kirk C, Foster E, Abdelhamid AS · The Cochrane database of systematic reviews · 2020

Cochrane systematic review and meta-analysis of randomised controlled trialsVerified from full text
Open original source

Reduction in saturated fat intake for cardiovascular disease (2015 version).

Hooper L, Martin N, Abdelhamid A, Davey Smith G · The Cochrane database of systematic reviews · 2015

Cochrane systematic review and meta-analysis of randomised controlled trialsVerified from abstract
Open original source

Effects of dietary fatty acids and carbohydrates on the ratio of serum total to HDL cholesterol and on serum lipids and apolipoproteins: a meta-analysis of 60 controlled trials.

Mensink RP, Zock PL, Kester AD, Katan MB · The American journal of clinical nutrition · 2003

Meta-analysis of 60 controlled feeding trialsVerified from abstract
Open original source

Associations Between Linoleic Acid Intake and Incident Type 2 Diabetes Among U.S. Men and Women.

Zong G, Liu G, Willett WC, Wanders AJ, Alssema M, Zock PL, Hu FB, Sun Q · Diabetes care · 2019

Prospective cohort study with isocaloric substitution modelling in three cohortsVerified from full text
Open original source

Intake of carbohydrates compared with intake of saturated fatty acids and risk of myocardial infarction: importance of the glycemic index.

Jakobsen MU, Dethlefsen C, Joensen AM, Stegger J, Tjønneland A, Schmidt EB, Overvad K · The American journal of clinical nutrition · 2010

Prospective cohort study using nutrient substitution modelsVerified from abstract
Open original source

Community

Discussion

Have you tried this? What helped or made it difficult?

Reader experiences are separate from the scientific evidence. Explore all discussions

No discussions yet.

Be the first to ask a question or share your experience.

Sign in to vote or join the discussion.

Related habits

Explore other habits in a shared category. Each has its own evidence and precautions.

Editorial history and review method

Revision 3 · Approved for publication · Codex (publication approval explicitly authorized by site owner; category migration)

Prepared and checked with AI. Publication approval does not imply independent expert review. Request a correction.

Applied the audit corrections for unsaturated-fat-substitution. (1) Two substantive claims were uncited and their papers were absent from content/sources.json: the 2015 Cochrane subgroup contrast used in the verdict and the controlled-feeding total:HDL result. Both have been retrieved and stored (hooper-2015-saturated-fat-cochrane, doi 10.1002/14651858.cd011737; mensink-2003-total-hdl, doi 10.1093/ajcn/77.5.1146) and are cited from the cardiovascular-events outcome that uses them. (2) The card summary previously led with the superseded 2015 subgroup - that the benefit appeared only with polyunsaturated replacement - which the cited 2020 review could not confirm and which contradicts that review's own conclusion that polyunsaturated fat or carbohydrate replacement both appear useful. It now leads with the cited review's result and states what the 2020 update found. (3) Implementation step 3 was overstated; it now distinguishes high-glycaemic-index carbohydrate (associated with higher myocardial infarction risk in Jakobsen 2010) from carbohydrate in general, and records that the Cochrane review found no significant difference between polyunsaturated and carbohydrate replacement while Zong 2019 found carbohydrate substitution was associated with 9% lower diabetes risk. (4) Zong's own carbohydrate result has been added to the substitution outcome so the reader sees the full substitution model rather than only the half that supports the entry's framing. (5) The substitution outcome is renamed to the endpoints actually measured, and its limitations now state that it mixes cohort substitution models with a randomised diabetes-diagnosis result while its evidence basis records the observational design that produced the estimates. (6) Precaution 1's cohort/meat/dairy contrast had no source and its two halves described different studies; it has been replaced by a statement about what the trials did and did not test. (7) The verdict now carries Mensink's own caveat that total:HDL changes are risk markers rather than proven risk changes. UNRESOLVED: the changeNotes previously referred to a 'substitution review giving a 10 percent coronary heart disease reduction per 5 percent energy substitution with polyunsaturated fat'. That paper was never stored and could not be identified with confidence, so the claim has been dropped rather than cited; if it is Mozaffarian 2010 (doi 10.1371/journal.pmed.1000252) a human should retrieve and cite it. Second reviewer must re-check: (a) whether the 2020 review's inability to reproduce the 2015 subgroup reflects lost signal or a changed subgroup definition, which needs both full texts; (b) that the outcome id replacement-nutrient-matters was deliberately left unchanged for link stability even though the display name now describes the endpoints; (c) that no effect estimate here duplicates the LDL claims in soluble-fibre-cholesterol. Publication release 2026-10-04: site owner authorized a bounded AI audit and release of this exact revision. See docs/release-audit-2026-10-04.md for source-check findings and limits. No independent health-professional or peer review is claimed. Category migration 2026-10-04: removed the separate goals field; categories: Nutrition, Long-term health → Nutrition, Long-term health. Research assessments, sources, guidance, precautions, and evidence review dates are unchanged. This revision awaits publication approval; no new evidence review is claimed. Publication approval 2026-10-04: site owner explicitly accepted all current category-migration revisions for deployment and requested that remaining limitations be retained in docs/content-review-backlog.md. Approval applies to this exact content revision. No new evidence review, independent expert review, or clinical validation is claimed; existing precautions and disclosures remain in effect.