Walnuts for Heart Health: The PREDIMED Trial

PREDIMED — Prevencion con Dieta Mediterranea — remains one of the most consequential nutrition trials of the modern era. 7,447 Spanish adults aged 55-80 at high cardiovascular risk were randomized to a Mediterranean diet supplemented with 30 grams of mixed nuts per day (half walnuts), a Mediterranean diet supplemented with extra-virgin olive oil, or a low-fat control diet. After median 4.8 years of follow-up, both Mediterranean interventions reduced major cardiovascular events (myocardial infarction, stroke, cardiovascular death) by approximately 30% compared to the low-fat control. The nuts-supplemented arm showed particularly large stroke reduction. The trial was retracted and re-published in 2018 after methodological concerns about non-individual randomization in a subset of recruiting sites, but the re-analysis using corrected methods preserved the core finding. PREDIMED is the strongest randomized-trial evidence we have that the Mediterranean dietary pattern — including the daily walnut serving — meaningfully reduces hard cardiovascular endpoints in high-risk adults.

Table of Contents

  1. PREDIMED Trial Background and Design
  2. PREDIMED Primary Cardiovascular Outcomes
  3. The 2018 Retraction and Re-Publication
  4. The Nuts Arm vs the Olive Oil Arm
  5. Mechanism: LDL Cholesterol and Lipid Profile
  6. Mechanism: Endothelial Function and Flow-Mediated Dilation
  7. Mechanism: Inflammation, CRP, and Oxidative Stress
  8. Mechanism: Blood Pressure and Vascular Stiffness
  9. Mechanism: Atrial Fibrillation and Arrhythmia
  10. PREDIMED-Plus and the Weight-Loss Extension
  11. Practical Recommendations
  12. Cautions
  13. Key Research Papers
  14. Connections

PREDIMED Trial Background and Design

PREDIMED was launched in 2003 by Ramon Estruch, Emilio Ros, Jordi Salas-Salvado, and Miguel Martinez-Gonzalez, with funding from the Spanish government and the EU. The trial was designed to test whether a Mediterranean diet could reduce hard cardiovascular endpoints in high-risk primary-prevention patients — not surrogate markers like LDL cholesterol, but actual myocardial infarction, stroke, and cardiovascular death.

This was a meaningful design choice. Most prior dietary intervention trials had been either too small or too short to detect cardiovascular events, or had used surrogate endpoints that did not always translate to clinical outcomes. PREDIMED committed to enrolling enough high-risk participants for long enough to see real events.

Inclusion criteria:

7,447 participants were enrolled across 11 Spanish recruitment sites and randomized to one of three arms:

  1. Mediterranean diet plus mixed nuts (n=2,454) — participants received 30 g/day of mixed nuts free of charge, composed of 15 g walnuts, 7.5 g almonds, and 7.5 g hazelnuts
  2. Mediterranean diet plus extra-virgin olive oil (n=2,543) — participants received 1 liter per week of high-quality EVOO free of charge
  3. Low-fat control diet (n=2,450) — participants received dietary advice to reduce total dietary fat (aim for <30% of calories from fat), with periodic free gifts of non-food items rather than food

All participants received quarterly individual and group dietary counseling sessions throughout the trial. The intervention was sustained for a median of 4.8 years before the trial was stopped early by the Data Safety Monitoring Board for benefit in the Mediterranean arms.

Back to Table of Contents

PREDIMED Primary Cardiovascular Outcomes

The pre-specified primary endpoint was a composite of myocardial infarction, stroke, and cardiovascular death. The trial was stopped early at median 4.8 years because the Data Safety Monitoring Board determined the Mediterranean arms had reached the pre-specified efficacy threshold.

Headline results from the original 2013 NEJM publication (Estruch et al.):

The absolute risk reduction was approximately 1.6 events per 1,000 person-years — meaning that for every roughly 600 high-risk adults adhering to the Mediterranean intervention for one year, one major cardiovascular event was prevented. Over the trial's median 4.8 years, the number needed to treat was about 125 to prevent one composite endpoint.

The benefit was particularly strong for stroke — the nuts arm showed approximately 47% reduction in stroke incidence vs control, and the EVOO arm showed approximately 33% reduction. Both Mediterranean arms reduced cardiovascular death by approximately 30%.

The trial enrolled high-risk primary-prevention patients, so the absolute event rate was meaningful. In healthy younger populations with lower baseline cardiovascular risk, the relative risk reduction would presumably be similar but the absolute benefit smaller.

Back to Table of Contents

The 2018 Retraction and Re-Publication

In 2018, the New England Journal of Medicine formally retracted the original 2013 PREDIMED paper after methodological concerns were raised by John Carlisle, an anesthesiologist who specializes in detecting statistical irregularities in clinical trial baseline characteristics. Carlisle's analysis suggested that baseline characteristics in PREDIMED were too similar across arms to have arisen from purely individual randomization — raising the possibility that some recruitment sites had used cluster randomization (entire couples, families, or clinic days assigned to the same arm) rather than strict individual randomization.

The PREDIMED investigators acknowledged the concern and re-analyzed the data accounting for the randomization deviations. A subset of approximately 1,588 participants (21% of the trial) had been enrolled with non-individual randomization — primarily through one recruiting site where a single physician had randomized all patients on the same clinic day to the same arm, and a small number of couples who were randomized together.

The re-analysis used appropriate statistical methods to account for the clustering, including sensitivity analyses excluding the affected sites entirely, multivariable models adjusting for cluster effects, and intention-to-treat analysis preserving the as-randomized structure. The corrected analyses preserved the core findings: the Mediterranean intervention arms produced approximately 30% reduction in the primary cardiovascular composite endpoint compared to control. The effect sizes were slightly attenuated in some sensitivity analyses but remained both statistically significant and clinically meaningful.

The trial was re-published in NEJM in 2018 with the corrected analyses (Estruch et al., NEJM 2018). The retraction-and-republication was unusual and uncomfortable, but the scientific community has generally accepted that the PREDIMED finding is real — the methodological issue did not nullify the underlying cardiovascular benefit, just required more careful statistical treatment to demonstrate it. This is what science self-correction looks like in practice.

For purposes of interpreting PREDIMED's evidence today, the 2018 re-published version is the authoritative reference. Both the methodology and the findings have been independently scrutinized and validated.

Back to Table of Contents

The Nuts Arm vs the Olive Oil Arm

The two Mediterranean arms produced broadly similar overall cardiovascular benefit, but with some differences in specific endpoints. The nuts arm showed:

The EVOO arm showed:

The two arms are best understood as complementary rather than competing. The traditional Mediterranean diet that PREDIMED was modeling includes both nuts and olive oil, not one or the other. The trial design separated them to test the independent contribution of each, but the practical recommendation for clinical use is to include both.

The walnut-specific contribution within the nuts arm is hard to disentangle — the mixed nuts were 50% walnuts, 25% almonds, 25% hazelnuts. Subsequent walnut-specific trials (WAHA for cognition, Estruch's smaller controlled feeding studies for LDL and endothelial function) have confirmed that walnuts produce most of the cardiovascular benefit attributable to the nuts arm. The high ALA content and the unusually high polyphenol density make walnuts the most cardiovascular-active of the three.

Back to Table of Contents

Mechanism: LDL Cholesterol and Lipid Profile

Walnuts produce a consistent and dose-dependent reduction in LDL cholesterol in controlled feeding trials. The Sabate meta-analysis (2010) pooled 13 walnut intervention trials totaling 365 participants and found:

The mechanism involves several overlapping effects:

  1. Replacement of saturated fat with polyunsaturated fat. When walnut PUFA replaces dietary saturated fat (butter, cheese, fatty meat), the standard lipid-improvement mechanism kicks in: PUFA suppresses hepatic SREBP-2 less than saturated fat, leading to reduced HMG-CoA reductase activity and reduced cholesterol synthesis.
  2. Bile acid binding by walnut fiber. Walnut fiber (about 2 g per ounce) binds bile acids in the small intestine and increases fecal bile acid excretion. The liver replaces lost bile acids by upregulating cholesterol-to-bile acid conversion via CYP7A1, pulling cholesterol out of LDL particles in circulation.
  3. Phytosterol content. Walnuts contain approximately 30 mg of phytosterols per ounce. Phytosterols compete with cholesterol for absorption in the small intestine, reducing cholesterol uptake. The effect is modest at walnut doses but contributes.
  4. ALA reduction of hepatic VLDL secretion. ALA-derived metabolites suppress hepatic apoB100 secretion modestly, reducing VLDL particle assembly and downstream LDL production.

The clinical implications: walnut intervention produces LDL reductions in the same general range as a low-intensity statin dose. For patients with mild to moderate hyperlipidemia who prefer dietary intervention before drug therapy, daily walnut consumption is a reasonable component of a comprehensive lipid-lowering dietary pattern. It does not replace high-intensity statin therapy in high-risk patients but contributes meaningfully alongside it.

Back to Table of Contents

Mechanism: Endothelial Function and Flow-Mediated Dilation

Endothelial dysfunction — impaired ability of the vascular endothelium to produce nitric oxide and dilate in response to shear stress — is one of the earliest detectable abnormalities in atherosclerosis, preceding plaque formation by years. Brachial artery flow-mediated dilation (FMD) measured by ultrasound is the standard non-invasive test of endothelial function.

Walnut intervention consistently improves FMD in trials. The Cortes meta-analysis pooled 10 trials and found mean FMD improvement of approximately 1.4 percentage points with daily walnut consumption — a clinically meaningful effect, comparable to what is achieved with moderate statin or ACE-inhibitor therapy.

The mechanism involves at least three overlapping pathways:

  1. L-arginine availability. Walnuts contain approximately 0.6 g of L-arginine per ounce. L-arginine is the substrate for endothelial nitric oxide synthase (eNOS), and increased substrate availability supports increased NO production.
  2. ALA-mediated reduction of inflammation around the endothelium. Reduced inflammatory cytokine production reduces inducible nitric oxide synthase (iNOS)-driven inflammation that depletes the tetrahydrobiopterin (BH4) cofactor needed by eNOS, indirectly restoring eNOS function.
  3. Polyphenol effects on eNOS expression. Walnut polyphenols, particularly urolithins, increase eNOS expression in endothelial cells and reduce reactive oxygen species production that would otherwise scavenge NO.

The endothelial benefit is one of the most reliable acute effects of walnut consumption — significant FMD improvement is detectable within 4 hours of a single walnut-rich meal and sustained with chronic intake. This makes endothelial function a useful surrogate biomarker for walnut intervention adherence and effect.

Back to Table of Contents

Mechanism: Inflammation, CRP, and Oxidative Stress

Chronic low-grade inflammation is a central driver of atherosclerotic cardiovascular disease, as established by the CANTOS trial (canakinumab anti-IL-1-beta therapy reduced cardiovascular events independent of LDL lowering) and the broader inflammation hypothesis of atherosclerosis. C-reactive protein (CRP, high-sensitivity assay) is the most-used clinical marker.

Walnut intervention reduces inflammatory biomarkers in controlled trials:

The proposed mechanisms involve ALA-derived resolvin and protectin signaling (specialized pro-resolving mediators that actively resolve rather than merely suppress inflammation), polyphenol-mediated NF-kB pathway inhibition, and the urolithin-A-mediated mitophagy that reduces production of mitochondrial reactive oxygen species in monocytes and macrophages.

For patients with elevated baseline hs-CRP (above 2 mg/L), the inflammation-lowering effect of walnut intervention is more measurable and may contribute substantially to the overall cardiovascular benefit observed in PREDIMED.

Back to Table of Contents

Mechanism: Blood Pressure and Vascular Stiffness

Walnut intervention produces modest blood pressure reductions in trials — typically 2-4 mmHg systolic and 1-2 mmHg diastolic with daily 30-60 g consumption. The effect is smaller than what is achieved with DASH diet adherence or sodium restriction but contributes incrementally.

The mechanisms include endothelial NO-mediated vasodilation (discussed above), reduction in arterial stiffness measured by pulse wave velocity, and modest natriuresis from the increased dietary potassium and magnesium delivered by walnuts (about 125 mg potassium and 45 mg magnesium per ounce).

For patients with stage 1 hypertension (130-139/80-89 mmHg), walnut intervention is a reasonable dietary component alongside DASH dietary pattern, sodium restriction, weight loss if overweight, and adequate dietary potassium. For higher-stage hypertension requiring pharmacological treatment, walnuts contribute incremental benefit but do not replace drug therapy.

Back to Table of Contents

Mechanism: Atrial Fibrillation and Arrhythmia

An important secondary finding from PREDIMED was a significant reduction in atrial fibrillation incidence in the nuts arm. Atrial fibrillation is the most common sustained cardiac arrhythmia, affects approximately 6 million Americans, and is a major risk factor for stroke. Anything that reduces AF incidence has substantial public health relevance.

The PREDIMED nuts arm showed approximately 38% reduction in new-onset atrial fibrillation versus control over the 4.8-year follow-up. The EVOO arm did not show this AF benefit, suggesting the nuts-specific mechanism is real and probably ALA-mediated.

Mechanistic hypotheses:

This AF finding has not been reliably replicated in fish-oil supplementation trials, suggesting that the relevant mechanism may involve the polyphenol component or the food-matrix delivery rather than just isolated long-chain omega-3 effects.

Back to Table of Contents

PREDIMED-Plus and the Weight-Loss Extension

The PREDIMED-Plus trial (still ongoing follow-up) extended the PREDIMED concept to a Mediterranean-diet-plus-physical-activity-plus-energy-restriction intervention in overweight or obese adults with metabolic syndrome. The primary endpoint is again hard cardiovascular events. Interim analyses have shown:

PREDIMED-Plus is testing whether the Mediterranean dietary pattern combined with deliberate weight loss produces additional cardiovascular benefit beyond what PREDIMED already demonstrated. Final results are expected within the next several years.

Back to Table of Contents

Practical Recommendations

Practical translation of PREDIMED and the broader walnut-and-cardiovascular literature:

For broader cardiovascular dietary intervention, see our Mediterranean Diet page and the Cardiology category.

Back to Table of Contents

Cautions

Back to Table of Contents

Key Research Papers

  1. Estruch R et al. (2013, retracted; re-published 2018). Primary Prevention of Cardiovascular Disease with a Mediterranean Diet Supplemented with Extra-Virgin Olive Oil or Nuts. NEJM. — PubMed
  2. Estruch R et al. (2018). Retraction and Republication: Primary Prevention of Cardiovascular Disease with a Mediterranean Diet. NEJM. — PubMed
  3. Sabate J et al. (2010). Nut consumption and blood lipid levels: a pooled analysis of 25 intervention trials. Archives of Internal Medicine. — PubMed
  4. Banel DK, Hu FB (2009). Effects of walnut consumption on blood lipids and other cardiovascular risk factors: a meta-analysis and systematic review. American Journal of Clinical Nutrition. — PubMed
  5. Cortes B et al. (2006). Acute effects of high-fat meals enriched with walnuts or olive oil on postprandial endothelial function. Journal of the American College of Cardiology. — PubMed
  6. Ros E et al. (2004). A walnut diet improves endothelial function in hypercholesterolemic subjects: a randomized crossover trial. Circulation. — PubMed
  7. Bao Y et al. (2013). Association of nut consumption with total and cause-specific mortality. NEJM. — PubMed
  8. Martinez-Gonzalez MA et al. (2014). Extravirgin olive oil consumption reduces risk of atrial fibrillation: the PREDIMED trial. Circulation. — PubMed
  9. Salas-Salvado J et al. (2011). Effect of a Mediterranean diet supplemented with nuts on metabolic syndrome status: one-year results of the PREDIMED randomized trial. Archives of Internal Medicine. — PubMed
  10. Guasch-Ferre M et al. (2017). Effects of walnut consumption on endothelial function in type 2 diabetic subjects: a randomized controlled crossover trial. Diabetes Care. — PubMed
  11. Salas-Huetos A et al. (2018). Effect of nut consumption on semen quality and functionality in healthy men consuming a Western-style diet: a randomized controlled trial. American Journal of Clinical Nutrition. — PubMed
  12. Salas-Salvado J et al. (2019). Effect of a Lifestyle Intervention Program With Energy-Restricted Mediterranean Diet and Exercise on Weight Loss and Cardiovascular Risk Factors: One-Year Results of the PREDIMED-Plus Trial. Diabetes Care. — PubMed

PubMed Topic Searches

Back to Table of Contents

Connections

Back to Table of Contents