Many diets claim to protect your heart, but few were tested on heart attacks

Key takeaways

• In the Women’s Health Initiative, 48,835 women were followed for about eight years, and coaching to cut total fat didn’t clearly lower heart disease, while their saturated fat intake barely changed¹².
• In a four-week Toronto trial, a “Portfolio” diet of nuts, soy, plant sterols and gel-forming fiber cut LDL cholesterol by 28.6%, close to the 30.9% drop from 20 mg a day of the statin lovastatin³.
• The American Heart Association’s 2023 diet scores measured how well each diet matches its advice, not heart attacks; DASH scored 100 and very low-carb or keto scored 31⁶.
• The AHA’s March 2026 guidance trimmed its list to nine features and now tells you to swap saturated fats for unsaturated ones rather than naming specific oils¹⁰¹¹.
• The 2025–2030 Dietary Guidelines keep saturated fat under 10% of calories, a cap heart groups support, but cardiologists object to the new protein and full-fat dairy advice¹²¹³¹⁴.

In 1993, women in their 50s, 60s and 70s began signing up for a long-term diet experiment. By 1998, 48,835 had enrolled at 40 U.S. clinical centers in the Women’s Health Initiative¹.

Four in 10 were coached to cut fat to 20% of their calories and to eat more fruit, vegetables and grains. The rest got printed diet advice¹.

After about eight years, heart disease rates were nearly identical. By our math from the trial’s case counts, about 5.1% of the coached women developed coronary heart disease, compared with 5.3% of the others, a gap small enough to be chance¹.

That result surprises anyone raised on low-fat advice. It also shows why “heart healthy” on a book cover means little until you ask what kind of evidence stands behind it.

This article sorts heart-diet evidence into three tiers, then looks at how the American Heart Association (AHA) scored popular diets and what changed in 2026.

Three kinds of diet evidence, and why they don’t weigh the same

Hard-outcome trials are the top tier. They are randomized controlled trials, where people are assigned to diets by chance, and they count heart attacks, strokes and deaths. They need thousands of people and years of follow-up, so few exist.

Risk-factor trials are the middle tier. They change what people eat for weeks or months and measure LDL cholesterol or blood pressure. They are faster and smaller, and they show how hard a diet pushes on a known heart risk factor. They can’t count the heart attacks prevented.

Cohort studies follow large groups for decades and compare people who eat differently. They capture long stretches of real life. But they can’t prove cause, because people who eat well may differ in other healthy habits.

A diet earns the most trust when all three tiers point the same way. The scorecard later in this article shows where each popular pattern stands on each tier.

A low-fat trial of 48,835 women cut the wrong fat

The Women’s Health Initiative’s dietary arm is a large hard-outcome test of low-fat eating. Its goal was total fat at 20% of calories, at least five servings of vegetables and fruit, and at least six servings of grains a day¹.

By year six, the coached women’s total fat had fallen by 8.2 percentage points of calories compared with the other group. But saturated fat, the type heart guidance most wants replaced, fell by only 2.9 points¹¹⁰. Their LDL cholesterol dropped by just 3.55 mg/dL¹.

The heart results matched those small changes. The hazard ratio for coronary heart disease was 0.97, meaning about 3% lower risk, with a confidence range wide enough to include no effect at all¹. Stroke and total cardiovascular disease didn’t budge either¹.

Two details point to what went wrong. The researchers saw trends toward more benefit in women who ate less saturated and trans fat and more fruit and vegetables¹. And a 2006 commentary by Carlos Tajer on IntraMed, a Spanish-language medical site, noted that the trial cut total fat without shifting the balance of saturated to unsaturated fat. By his own calculation, its statistical power, the chance of detecting a real effect, was about 40% rather than the 90% planned².

The lesson isn’t that fat doesn’t matter. A fair reading is that cutting all fat equally, rather than swapping saturated fat for healthier fats, missed the target. The AHA made a similar point in 2023 when it flagged low-fat diets for treating all fats alike⁶.

Mediterranean-style eating has also been tested against heart events, in France’s Lyon Diet Heart Study and Spain’s CORDIOPREV, both in heart patients. Our Mediterranean article weighs both, and our diet comparison covers the 2018 republished PREDIMED trial in high-risk adults.

Short trials show how fast food can move cholesterol

The Portfolio diet trial is a striking example of this tier. In late 2002, researchers at St. Michael’s Hospital in Toronto randomly assigned adults with high LDL cholesterol to one of three plans for four weeks³.

One group ate a very low saturated fat diet with a placebo pill. A second ate the same diet plus 20 mg a day of lovastatin, a statin. The third ate a “portfolio” of cholesterol-lowering foods. For every 1,000 calories, that meant 1.0 gram of plant sterols, 21.4 grams of soy protein, 9.8 grams of viscous fiber and 14 grams of almonds³.

Forty-six people finished. LDL fell 8.0% on the low saturated fat diet, 30.9% with the statin and 28.6% on the portfolio, and the statin and portfolio results didn’t differ significantly³.

Now the fine print. The researchers supplied almost all the food, and the trial lasted one month. Some participants found the amount of food too much, and the authors called for studies of people who assemble the diet themselves³. Our cholesterol article gives the 2026 cholesterol guideline’s estimate for the Portfolio diet.

Ask who paid, too. We couldn’t find the 2003 paper’s funding statement. The same team’s larger 2011 Portfolio trial, of 351 adults over six months, was funded mainly by the Canadian Institutes of Health Research, with extra money from Loblaw Brands, Solae and Unilever¹⁵.

Its authors reported ties to dozens of companies and food industry groups, from Unilever and Quaker Oats to the Almond Board of California¹⁵. That doesn’t make the results wrong, but it earns the scrutiny you’d give a drug company’s trial.

DASH, short for Dietary Approaches to Stop Hypertension, belongs in this tier, too. Our articles comparing diets and on blood pressure habits give its trial numbers.

Decades-long studies agree, but can’t settle cause

The third tier is the biggest. In a 2023 study in JAMA Internal Medicine, Harvard researchers tracked 119,315 nurses and other health professionals for up to 36 years⁴. All entered the study free of heart disease and cancer, and they reported what they ate on a questionnaire every four years⁴.

The team graded each person’s eating with four different scores, a Mediterranean score and a healthful plant-based score among them. The scores differ in the details, but each one gives credit for produce, whole grains, beans and nuts⁴.

People in the top fifth on each score had a 14% to 20% lower risk of dying during follow-up than people in the bottom fifth, by our math from the reported hazard ratios⁵. Every score also tracked with fewer deaths from cardiovascular disease⁵. The pattern held for White, Black and Hispanic participants⁴.

Federal grants paid for the work, and some authors had support from the AHA⁴. The limits are the usual ones for cohorts: people chose their food, and their diets came from questionnaires.

Our article on eight healthy eating patterns covers a related Harvard analysis of healthy aging.

The AHA’s diet scores measure fit with its advice, not heart attacks

In April 2023, an AHA writing group led by Stanford’s Christopher Gardner rated 10 popular diets on a 1-to-100 scale⁶. They checked each one against nine features of the AHA’s 2021 dietary guidance, such as whole grains, plant oils and little added sugar or salt⁶⁹.

Tier	Diets and scores
1	DASH 100, pescatarian 92, Mediterranean 89, vegetarian 86
2	Vegan 78, low-fat 78
3	Very low-fat 72, low-carb 64
4	Paleo 53, very low-carb or keto 31

Source: AHA News, April 2023⁶.

The reasons for the scores teach more than the numbers. The Mediterranean diet lost points because it allows moderate drinking and doesn’t address added salt⁶. Vegan eating drew concerns that it’s hard to sustain and can leave people low in vitamin B12⁶. Paleo and keto sank because they restrict whole fruit, legumes and whole grains and don’t limit saturated fat⁷.

This is the part most headlines skipped. The scores measure how closely each diet matches the AHA’s advice⁶. None of them counts heart attacks prevented.

DASH’s perfect score means it fits the checklist best. It doesn’t mean DASH has been shown to prevent more heart attacks than other patterns.

The group also left out intermittent fasting and diets followed for less than 12 weeks⁶. Gardner and his vice chair reported no relevant conflicts of interest⁷.

2026 brought new heart guidance and new federal guidelines

On March 31, 2026, the AHA updated its dietary guidance. Alice Lichtenstein of Tufts University chaired the writing group, and the AHA said the advice hasn’t shifted greatly since 2021¹⁰. Its list shrank from 10 features to nine⁹¹⁰.

The changes are worth knowing¹⁰¹¹:

• Fats: choose unsaturated fats in place of saturated ones, instead of naming liquid plant oils specifically.
• Protein: shift from meat toward plant protein such as beans and nuts, eat fish and seafood regularly, and if you want red meat, pick lean cuts, skip processed meat and keep portions small.
• Dairy: low-fat and fat-free dairy are still preferred, though the AHA notes that the full-fat debate continues.
• Alcohol: if you don’t drink, don’t start, and the AHA now notes federal and WHO warnings that no level is safe for certain cancers.

The federal Dietary Guidelines for Americans, 2025–2030, arrived in January 2026 with a different tone. The document runs about 10 pages, opens with “eat real food,” keeps saturated fat under 10% of calories, advises full-fat dairy without added sugars, and says salting food is fine “if preferred”¹². It sets protein at 1.2 to 1.6 grams per kilogram of body weight a day and limits added sugar to no more than 10 grams per meal¹⁴.

Heart groups welcomed much of it. In a statement issued the day the guidelines came out, the AHA said both it and the federal guidance call for more vegetables, fruits and whole grains and less added sugar, saturated fat and highly processed food¹²¹³. But it warned that the salt and red meat advice could push people past sodium and saturated fat limits, and it still prefers low-fat dairy¹³.

Experts in the American College of Cardiology’s prevention section went further. They wrote that the protein target exceeds the 0.8 g/kg recommended daily amount and that butter and beef tallow raise LDL compared with olive oil¹⁴. Lichtenstein warned that more butter, tallow, meat and whole-fat dairy would make the 10% saturated fat cap hard to meet¹².

Both documents land on the same saturated fat ceiling. By the AHA’s reckoning, someone eating along its nine features would rarely go over it¹⁰.

The heart-diet scorecard

This table is our own summary. “None found” means we found no trial of that kind in our research, not that one could never exist. Ratings use our four-level scale.

Pattern	Trials that counted heart events	Risk-factor trials	Long-term studies	Our rating
Mediterranean	Lyon and CORDIOPREV, plus a larger trial in high-risk adults (see our Mediterranean and diet comparison articles)	See those articles	A Mediterranean score tracked with fewer heart deaths5	Strong evidence
DASH	None found	Blood pressure trials (see our diet comparison)	Not covered here; ACC prevention experts cite extensive evidence for DASH-style eating14	Strong evidence for blood pressure
Portfolio	None found	LDL down 28.6% in four weeks, food supplied3; the 2026 cholesterol guideline also credits it (see our cholesterol article)	Not covered here	Strong evidence for lower LDL; not enough evidence on heart events
Low total fat	Women’s Health Initiative: no clear benefit1	LDL fell only 3.55 mg/dL1	Not covered here	Evidence says no for cutting total fat alone
Vegetarian, vegan	None found	See our cholesterol and vegetarian articles	A healthful plant-based score tracked with fewer heart deaths5	Promising but early
Low-carb	None found	See our low-carb article	Depends on what replaces the carbs (see our low-carb article)	Not enough evidence
Keto, paleo	None found	Can raise LDL when saturated fat isn’t limited7	Not covered here	Not enough evidence
2025–2030 federal pattern	Not tested as a whole	Not tested as a whole	Not tested as a whole	Not enough evidence as written

Your family’s cooking already holds a heart pattern

The patterns that score well across these tiers share a plain core: vegetables, fruit, whole grains, legumes, nuts, fish and plant oils, with little processed meat or added sugar⁴¹⁰¹¹. None of them requires a product, a powder or a branded plan.

That fits what many natural-health readers already believe about food. It also means heart-healthy eating isn’t tied to one cuisine. The AHA says its 2026 guidance can be fitted to a family’s culture, faith, tastes and budget, and it counts canned and frozen produce as healthy, low-cost choices¹⁰.

The 2023 statement went further. It called for research on African, Asian and Latin American eating patterns. It also said programs that lean on family support may help revive food traditions harmed by historical injustice and forced acculturation⁸.

So start with your own kitchen. Keep the dishes your family loves, and shift the parts the evidence flags: more beans and whole grains, more vegetables on the plate, plant oils for much of the cooking, and less processed meat.

Questions readers ask

The bottom line

Few diets have been tested on heart attacks and deaths, and a 48,835-woman test of low-fat eating mostly showed that cutting the wrong fat doesn’t help. Short trials show that plant-rich patterns can move LDL within weeks, and decades of cohort data point the same way.

The AHA’s 2026 guidance and the new federal guidelines agree on the saturated fat cap and on eating more plants and less processed food. They split on protein, dairy and butter. Build your plate from the shared core, judge every diet claim by the tier of evidence behind it, and work with your clinician on any medicine.

Related reading

• A classic Mediterranean diet trial used canola oil, so the pattern travels well
• Eight eating patterns compared, and why the diet you stick with beats the one you pick
• Eight healthy eating patterns faced a 30-year test and mostly agreed
• Your cholesterol panel leaves out two numbers that matter: ApoB and Lp(a)
• Fiber nudges cholesterol and blood sugar down, and processing can blunt the effect

Sources

1. Howard BV, Van Horn L, Hsia J, et al. “Low-Fat Dietary Pattern and Risk of Cardiovascular Disease: The Women’s Health Initiative Randomized Controlled Dietary Modification Trial.” JAMA, 2006;295(6):655–666 (University of Arizona research record). experts.arizona.edu
2. Tajer C. Commentary on the Women’s Health Initiative low-fat diet trial (in Spanish). IntraMed, March 28, 2006. intramed.net
3. Jenkins DJA, Kendall CWC, Marchie A, et al. “Effects of a Dietary Portfolio of Cholesterol-Lowering Foods vs Lovastatin on Serum Lipids and C-Reactive Protein.” JAMA, 2003;290(4):502–510. pubmed.ncbi.nlm.nih.gov
4. Harvard T.H. Chan School of Public Health. “Variety of Healthy Eating Patterns Linked With Lower Risk of Premature Death” (on Shan Z, Wang F, Li Y, et al., JAMA Internal Medicine, 2023). News release, January 9, 2023. hsph.harvard.edu
5. Patient Care. “Healthy Eating Patterns Associated with Reduced Risk of Mortality in Large Cohort Study” (on Shan Z, et al., JAMA Internal Medicine, 2023). 2023. patientcareonline.com
6. American Heart Association News. “Here’s How 10 Popular Diets Scored for Heart Health” (on Gardner CD, Vadiveloo MK, Petersen KS, et al., Circulation, 2023). April 27, 2023. heart.org
7. Cox CE. “Not All Diets Measure Up for Protecting the Heart: AHA.” TCTMD, April 28, 2023. tctmd.com
8. Jackson EA. “Aligning Popular Dietary Patterns With AHA 2021 Dietary Guidance: Key Points.” American College of Cardiology, May 1, 2023. acc.org
9. Fox 29 (from Fox News). “10 Dietary Tips to Reduce Heart Disease and Stroke Risk, According to New Research” (on Lichtenstein AH, et al., “2021 Dietary Guidance to Improve Cardiovascular Health,” Circulation, 2021). November 4, 2021. fox29.com
10. American Heart Association. “Following 9 Key Steps for a Lifetime of Eating Well Can Support Heart Health” (on Lichtenstein AH, et al., “2026 Dietary Guidance to Improve Cardiovascular Health,” Circulation, 2026). News release via EurekAlert, March 31, 2026. eurekalert.org
11. Guideline Central. “AHA 2026 Dietary Guidance to Improve Cardiovascular Health Guideline Summary.” 2026. guidelinecentral.com
12. Neale T. “New Dietary Guidelines for Americans Get Mixed Review From Cardiology Community.” TCTMD, January 12, 2026. tctmd.com
13. American Heart Association. “New U.S. Nutrition Guidance Emphasizes Importance of Healthier Eating.” Heart.org, undated statement (issued January 7, 2026, according to TCTMD’s coverage). heart.org
14. Aggarwal M, Aspry K, Kris-Etherton P, Freeman A, Williams KA Sr, Gianos E. “From the Member Sections: Dietary Guidelines.” Cardiology Magazine, American College of Cardiology, March 1, 2026. acc.org
15. Moon MA. “Emphasize Soy, Nut, Viscous Fiber Consumption to Cut LDL Cholesterol” (on Jenkins DJA, et al., JAMA, 2011;306:831–839). Family Practice News (MDedge), August 23, 2011. mdedge.com