EPIC-Oxford 2019 + Adventist Health Study-2 + Ornish trials

Vegan vs Vegetarian Heart Disease: EPIC-Oxford and AHS-2

Cardiovascular disease is the single largest cause of death globally. Plant-based eating patterns have some of the strongest published evidence among lifestyle interventions for reducing heart attack risk and (in selected patients) reversing established coronary stenosis. This page works through the cohort numbers, the mechanisms, and the important stroke-risk caveat that does not show up in the headlines.

The short answer. EPIC-Oxford 2019: vegetarians have 22% lower ischaemic heart disease risk than meat-eaters, vegans 24% lower. AHS-2: vegans have approximately 42% lower CVD mortality than non-vegetarians. The stroke story is the caveat: vegetarians and vegans have approximately 20% higher total stroke rate, mostly haemorrhagic. Net cardiovascular benefit remains substantial but smaller than the heart attack effect alone.

The EPIC-Oxford 2019 BMJ paper, in detail

The EPIC-Oxford cohort recruited 65,000 UK adults between 1993 and 1999 with deliberate over-representation of vegetarians, vegans, and fish-eaters. The 2019 paper by Tong, Appleby, Bradbury, Perez-Cornago, Travis, Clarke, and Key in the BMJ analysed cardiovascular outcomes over 18 years of follow-up in 48,188 of these participants who had complete dietary data and were free of CHD or stroke at baseline.

The findings, after multivariable adjustment for age, sex, smoking status, alcohol, physical activity, education, supplement use, BMI, and other potential confounders:

Diet group	IHD risk vs meat-eaters	Stroke risk vs meat-eaters	Haemorrhagic stroke risk
Meat-eaters (reference)	1.00	1.00	1.00
Fish-eaters	0.87 (13% lower)	1.04 (no clear difference)	1.21 (21% higher, NS)
Vegetarians + vegans combined	0.78 (22% lower)	1.20 (20% higher)	1.43 (43% higher)
Vegans alone	0.76 (24% lower)	~1.20 (similar to vegetarians)	~1.40 (similar to vegetarians)

In absolute numbers: over 10 years, the IHD reduction in vegetarians and vegans was about 10 fewer cases per 1,000 people; the stroke excess was about 3 additional cases per 1,000 people. The net cardiovascular benefit per 1,000 people over 10 years was therefore approximately 7 cases prevented, dominated by the IHD reduction. The size and direction of these effects is consistent with other major cohorts.

Adventist Health Study-2 corroboration

The Adventist Health Study-2 is a US cohort of approximately 96,000 Seventh-day Adventists recruited from 2002, with substantial vegetarian and vegan representation. The Orlich 2013 paper (JAMA Intern Med) reported on cardiovascular and all-cause mortality across diet groups. After multivariable adjustment, vegans had approximately 42% lower cardiovascular mortality and 26% lower all-cause mortality than non-vegetarians; lacto-ovo vegetarians had approximately 30% lower cardiovascular mortality. The effect sizes were larger in men than women.

The AHS-2 cohort is interesting because the comparison group (Adventist non-vegetarians) is itself a low-mortality population (the Adventist tradition limits alcohol and tobacco use across all members), which makes the vegan-vs-comparison-group difference particularly meaningful. The findings imply the dietary effect is on top of an already health-conscious baseline.

Mechanisms: what is actually doing the work

LDL cholesterol. Plant-based diets reduce LDL by 13 to 15 mg/dL on average per the Yokoyama 2017 meta-analysis. The mechanism involves replacement of saturated fat with unsaturated fat, increased soluble fibre intake (which binds bile acids), and removal of dietary cholesterol. Lower LDL maps to lower atherosclerotic plaque burden over years.

Blood pressure. Vegans have about 4 to 8 mmHg lower systolic BP than meat-eaters in pooled analyses. Mechanisms: higher potassium intake (from fruits and vegetables), lower sodium (when whole foods dominate), lower BMI, higher fibre, and possible effects of nitric oxide pathways from leafy greens. A 5 mmHg systolic reduction maps to approximately 10% lower cardiovascular event risk.

BMI and metabolic health. Vegans average BMI 22 to 24; meat-eaters 26 to 28 in the same cohorts. Lower BMI correlates with better insulin sensitivity, lower triglycerides, higher HDL, and lower visceral adiposity. The visceral fat reduction may matter most for cardiometabolic risk.

Inflammation markers. Plant-based diets reduce CRP, fibrinogen, and other inflammatory markers, possibly via gut microbiome effects on TMAO production, reduced dietary AGEs, and higher polyphenol intake. The mechanism is less precisely characterised but the direction is consistent.

Stroke caveat mechanism. The haemorrhagic stroke excess in vegetarians and vegans is not fully understood. Possible contributors: lower vitamin B12 elevating homocysteine (an independent stroke risk factor); lower LDL having a U-shaped relationship with haemorrhagic stroke (very low LDL associated with higher haemorrhagic risk in some studies); lower omega-3 EPA and DHA reducing platelet function; lower vitamin K2 affecting vascular calcification. The interventions that may mitigate are good B12 status, adequate omega-3 (algae oil), and avoiding inadvertent extreme LDL reduction.

The healthy-vs-unhealthy plant-based distinction

The Satija 2017 analysis from the Harvard Nurses' Health Study and Health Professionals Follow-up Study (J Am Coll Cardiol) introduced a now widely-used distinction between healthy and unhealthy plant-based diets. Healthy plant-based diet emphasises whole grains, fruit, vegetables, legumes, nuts, and tea or coffee. Unhealthy plant-based diet emphasises refined grains, sugar-sweetened beverages, sweets and desserts, and fried foods (all of which can be vegan).

Findings over 25 years of follow-up: a higher score on the healthy plant-based diet index was associated with 25% lower coronary heart disease risk. A higher score on the unhealthy plant-based index was associated with 32% higher CHD risk. The unhealthy plant-based diet was worse than a moderate omnivore diet for heart health. The implication: vegan is necessary but not sufficient for cardiovascular benefit; the composition matters as much as the category.

Practical translation: a vegan whose diet is built on lentils, beans, vegetables, whole grains, nuts, fruit, and tofu has the cardiovascular profile described above. A vegan whose diet is built on white bread, vegan cheese, vegan ice cream, vegan crisps, and sugary plant milks does not. The category alone does not determine the outcome.

The reversal trials: Ornish and Esselstyn

The Lifestyle Heart Trial (Ornish D et al., Lancet 1990; 336: 129-133; JAMA 1998; 280: 2001-2007) randomised 48 patients with moderate to severe coronary heart disease to a low-fat (under 10% energy from fat) whole-food vegetarian diet plus stress management and moderate exercise versus standard care. Quantitative coronary angiography at 1 year showed regression of stenosis in the intervention group (average lumen diameter increased) and progression in controls. At 5 years, the difference was maintained, with the intervention group having one-third the cardiac event rate of controls. The trial is one of the most cited cardiovascular lifestyle intervention studies.

The Esselstyn case series (J Fam Pract 1995, 2014 follow-up) followed 22 and then 198 patients with established CHD on a near-no-oil WFPB diet. The published outcomes included angiographic regression of stenosis and zero cardiac events in patients who maintained the diet. The case series design carries methodological limitations (no control, motivated participants, possible regression to mean) but the directional finding aligns with the Ornish randomised data.

Discuss before changing. If you have established cardiovascular disease, do not undertake major dietary changes without coordination with your cardiology team. Sudden dietary changes can affect medication efficacy (warfarin, statins, ACE inhibitors) and may need parallel adjustments. The published reversal trials were intensive supervised interventions, not unsupervised self-experiments. See the WFPB page for the dietary framework.

Frequently asked questions about vegan and vegetarian cardiovascular outcomes

How much lower is heart disease risk on a vegan or vegetarian diet?

The EPIC-Oxford 2019 analysis by Tong and colleagues (BMJ 2019; 366: l4897) followed 48,188 participants for 18 years. Compared to meat-eaters, fish-eaters had 13% lower ischaemic heart disease risk, vegetarians had 22% lower, and vegans had 24% lower (all after adjustment for BMI, smoking, alcohol, education, physical activity). The Adventist Health Study-2 (Orlich 2013, JAMA Intern Med) reported similar effect sizes in a US cohort with even stronger results for cardiovascular mortality. The directionality is consistent across major cohorts; the absolute size varies depending on which comparison group and what adjustments are made.

Is there any cardiovascular downside to plant-based diets?

The same EPIC-Oxford 2019 analysis found vegetarians and vegans had a 20% higher rate of total stroke compared to meat-eaters, driven primarily by haemorrhagic (not ischaemic) stroke. The absolute excess was small (about 3 additional strokes per 1,000 people over 10 years), but it is statistically robust. The mechanism is not fully understood; lower B12 status, lower LDL (which has a U-shaped relationship with haemorrhagic stroke risk), and lower omega-3 EPA and DHA have all been proposed. The net cardiovascular benefit (heart attack reduction minus stroke excess) remains positive but smaller than the heart attack effect alone.

What causes the heart disease reduction on plant-based diets?

Multiple mechanisms contribute. Lower LDL cholesterol on plant-based diets reduces atherosclerosis progression; the Yokoyama 2017 meta-analysis showed vegan diets reduce LDL by 13 to 15 mg/dL on average. Lower blood pressure: vegans have about 4 to 8 mmHg lower systolic BP than meat-eaters in pooled analyses, mostly attributable to lower saturated fat, higher potassium, higher fibre, and lower sodium in vegan eating patterns. Lower body weight in vegans (typical BMI 22 to 24 vs 26 to 28 in meat-eaters) reduces cardiometabolic risk. Higher fibre intake improves insulin sensitivity and gut microbiome health. Lower TMAO production from red meat reduces atherogenic signalling. The whole-foods composition of typical plant-based diets matters too; an ultra-processed vegan diet would not deliver the same benefits.

Can a plant-based diet reverse coronary heart disease?

In selected patients with established CHD, yes, on the published trial evidence. The Ornish Lifestyle Heart Trial (Lancet 1990, JAMA 1998) randomised 48 patients with moderate to severe coronary stenosis to a low-fat whole-food vegetarian diet plus stress management and exercise versus standard care. The intervention group showed angiographic regression of stenosis at 1 and 5 years; the control group showed progression. The Esselstyn case series (n=22 in 1995, n=198 in 2014) showed similar angiographic improvement in patients on a near-no-oil WFPB diet. The trials have methodological limitations (small sample size, intervention bundles including more than just diet) but the directional finding is striking enough that the diet is now accepted in cardiac rehabilitation programmes that follow the lifestyle medicine model.

Does the type of plant-based diet matter?

Yes, substantially. The Satija 2017 analysis from Harvard distinguished healthy plant-based diet (whole grains, fruit, vegetables, legumes, nuts) from unhealthy plant-based diet (refined grains, sugary drinks, sweets, fried foods that happen to be vegan). The healthy plant-based pattern was associated with 25% lower CHD risk; the unhealthy pattern was associated with 32% higher CHD risk. The vegan label alone does not guarantee cardiovascular benefit; the composition does. A vegan diet built on vegan junk food may be no better than an omnivore diet for heart health. The Esselstyn-style WFPB diet, Mediterranean diet, and DASH diet all have strong RCT evidence; the unifying feature is whole plant foods dominance.

Should someone with existing heart disease go vegan?

Discuss with your cardiology team. The Ornish and Esselstyn evidence supports WFPB-type dietary intervention as part of comprehensive cardiac rehabilitation, but the dietary change should be coordinated with medication management, particularly for blood thinners (vitamin K intake from leafy greens affects warfarin dosing) and statins (rapid lipid changes may need monitoring). Sudden major dietary changes in elderly patients with multiple comorbidities are not advisable; gradual transition over weeks with dietitian support is the safer route. The American Heart Association does not specifically endorse vegan diets but does endorse plant-forward eating patterns including Mediterranean and DASH.

Sources cited. Tong TYN et al. Risks of ischaemic heart disease and stroke in meat eaters, fish eaters, and vegetarians over 18 years of follow-up: results from the prospective EPIC-Oxford study, BMJ 2019; 366: l4897; Orlich MJ et al. Vegetarian dietary patterns and mortality in Adventist Health Study 2, JAMA Intern Med 2013; 173: 1230-1238; Ornish D et al. Can lifestyle changes reverse coronary heart disease? The Lifestyle Heart Trial, Lancet 1990; 336: 129-133; Ornish D et al. Intensive lifestyle changes for reversal of coronary heart disease, JAMA 1998; 280: 2001-2007; Esselstyn CB Jr et al. A way to reverse CAD?, J Fam Pract 2014; 63: 356-364; Yokoyama Y et al. Association between plant-based diets and plasma lipids: a systematic review and meta-analysis, Nutr Rev 2017; 75: 683-698; Satija A et al. Healthful and unhealthful plant-based diets and the risk of coronary heart disease in U.S. adults, J Am Coll Cardiol 2017; 70: 411-422; American Heart Association dietary guidance. All values as of May 2026.