Research

EPIC-Oxford cohort: what 17 years of plant-based observational data shows

Lower CVD risk, lower diabetes risk, modestly higher fracture risk concentrated in the lower-calcium tail.

By Ifeoma Obi, MS Nutrition · Published January 15, 2026 · Last updated April 11, 2026

The EPIC-Oxford cohort is the largest and longest-running prospective study of vegetarians and vegans in the published literature. Approximately 55,000 men and women in the United Kingdom were enrolled in the early-to-mid 1990s with the explicit recruitment goal of including a high proportion of vegetarians and vegans, and the cohort has been followed for an average of 17.6 years with multiple published analyses on chronic disease outcomes.

The findings have shaped current plant-based clinical guidance materially, particularly the calcium-and-fracture-risk relationship that informs the calcium-intake recommendations in the AND position paper and in our calcium piece. This summary covers the major outcome categories, the strengths and limitations of the cohort, and the practical clinical implications.

Cohort design

EPIC-Oxford recruited approximately 55,000 participants with an oversampled fraction of vegetarians and vegans. The diet pattern categories tracked are: meat-eaters (consuming meat and fish), fish-eaters (consuming fish but not meat), vegetarians (no meat or fish; dairy and eggs allowed), and vegans (no animal products). Diet pattern was assessed at baseline and at follow-up; participants who changed pattern during follow-up were handled with sensitivity analyses.

The cohort’s strengths are size, duration, and the high proportion of vegetarians and vegans (which gives statistical power that smaller studies lack). The cohort’s limitations are observational design (cannot establish causation), self-reported diet, and a UK-specific food environment.

Cardiovascular disease findings

The headline cardiovascular finding is that vegetarians and vegans had lower ischemic heart disease risk than meat-eaters. The pooled risk reduction was approximately 22 percent lower for ischemic heart disease in vegetarians and vegans combined, after adjustment for body mass index, smoking, alcohol, and other lifestyle factors. The finding has been replicated in subsequent analyses and is consistent with the cardiovascular literature on plant-based dietary patterns more broadly.

Stroke findings have been more nuanced. Some analyses found higher stroke risk in vegetarians and vegans (specifically hemorrhagic stroke), with the effect attributed to lower vitamin B12 status and lower total cholesterol. The clinical interpretation is debated: lower total cholesterol is generally a CVD-protective factor, but extremely low cholesterol may increase hemorrhagic stroke risk. The effect, where present, has been modest in absolute terms.

Type 2 diabetes findings

Vegetarians and vegans had lower type 2 diabetes incidence than meat-eaters in EPIC-Oxford. The effect persisted after adjustment for BMI, suggesting that the dietary pattern itself contributes beyond its weight effects. The finding is consistent with the broader literature on plant-based dietary patterns and diabetes risk.

Cancer findings

EPIC-Oxford has reported on multiple cancer outcomes. The headline finding is that vegetarians and vegans had lower overall cancer risk than meat-eaters, with the strongest effects on cancers historically associated with red and processed meat (colorectal, certain GI cancers). Effects on hormone-sensitive cancers (breast, prostate) have been more modest and less consistent.

Bone fracture findings

The fracture finding is the most-cited from EPIC-Oxford for plant-based clinical work. Vegans had a 43 percent higher overall fracture risk than meat-eaters in the pooled analysis, with the effect concentrated in hip fractures. Stratification by calcium intake showed the effect was driven by the lower-calcium-intake tail: vegans consuming below approximately 525 mg/day calcium had substantially elevated fracture risk; vegans consuming above that threshold did not show meaningful elevation.

The clinical interpretation has been to recommend that vegans hit at least 1000 mg/day calcium reliably and ideally land in the upper quintile of plant-based calcium intake. This recommendation is woven through the AND position paper and through clinical practice in plant-based RD work.

Mortality findings

Long-term mortality findings from EPIC-Oxford have shown overall similar all-cause mortality across diet groups after adjustment for lifestyle factors, with some specific cause-of-death differences (lower CVD mortality, possibly slightly higher fracture-related mortality in the lower-calcium-vegan tail). The differences are modest in absolute terms.

Methodological notes

EPIC-Oxford is observational and cannot establish causation. The cohort’s strengths (size, duration, UK plant-based-specific recruiting) are tempered by limitations:

Self-reported diet is subject to recall bias and misclassification.
The UK food environment in the 1990s and 2000s differs from current US plant-based food availability (fewer fortified plant milks, fewer mock meats, less attention to vegan supplementation in early-cohort years).
The plant-based community has become more attentive to supplementation and to nutrient-target hitting since the cohort enrolled. Findings on outcomes that depend on B12 or D status may not generalize fully to current plant-based eaters who supplement.

The mitigation has been replication in additional cohorts (Adventist Health Studies, NHANES analyses, smaller European studies) which generally agree with the EPIC-Oxford direction on cardiovascular disease, diabetes, and cancer outcomes, with fracture findings less consistently replicated.

Practical clinical implications

For plant-based clinical work the EPIC-Oxford findings translate to:

Cardiovascular benefit is real and supported. Plant-based diets are CVD-protective on average. The clinical case for plant-based dietary patterns in cardiovascular prevention is one of the stronger evidence bases in nutrition science.
Diabetes risk reduction is supported. Plant-based diets are associated with lower type 2 diabetes incidence beyond weight effects.
Calcium intake matters specifically for vegans. The fracture finding’s stratification is the practical guidance: hit at least 1000 mg/day calcium reliably, with attention to food sources that contribute meaningfully (calcium-set tofu, fortified plant milks, low-oxalate greens).
B12 status matters specifically for vegans. The hemorrhagic stroke signal, where present, has been linked to B12 status. The clinical case for B12 supplementation, already strong, is reinforced.
Whole-food plant-based emphasis is supported. The cohort’s findings are most favorable for vegetarians and vegans whose diets emphasize whole grains, legumes, fruits, vegetables, and nuts/seeds. Diets relying heavily on refined grains and ultra-processed plant-based products do not show the same benefits.

Citation

EPIC-Oxford cohort, multiple publications in American Journal of Clinical Nutrition, European Journal of Clinical Nutrition, BMJ, Public Health Nutrition, and other journals. The Travis et al. and Tong et al. fracture analyses are the most-cited for the calcium-and-fracture findings.

For related summaries see the AND position paper, plant-protein quality literature, and recent B12 literature.

Topics: EPIC-Oxford vegan · EPIC-Oxford fracture · vegan cardiovascular disease · plant-based cohort study