Are vegan and vegetarian diets healthy, and what do they miss?

Who counts. Studies group people as meat eaters, fish eaters, lacto-ovo vegetarians (dairy and eggs), and vegans (no animal foods). "Plant-based" can also mean mostly plants with a little meat.

Cholesterol and weight. Across 11 trials, vegetarian diets lowered total cholesterol by 0.36 mmol/L (about 14 mg/dL) and LDL cholesterol by 0.34 mmol/L (about 13 mg/dL), without changing triglycerides⁹. In 22 pairs of identical twins, eight weeks of a healthy vegan diet lowered LDL by 13.9 mg/dL more than a healthy omnivorous diet, and also lowered fasting insulin and weight⁸.

Heart disease, cancer and death. Pooled cohorts show 25% lower ischemic heart disease and 8% lower total cancer in vegetarians, but no clear difference in overall death rates⁷. In Adventist Health Study 2, vegetarians had 12% lower all-cause mortality¹⁰. In the UK's EPIC-Oxford cohort, vegetarians had 22% lower ischemic heart disease but 20% higher stroke rates, mostly hemorrhagic¹¹.

Quality matters. "Plant-based" includes sweets, refined grains and fries. In three US cohorts a healthful plant-based diet went with 25% less coronary heart disease, an unhealthful one with 32% more³. See also the Mediterranean diet and the Blue Zones.

What can be missing. A review of 141 studies found plant-based eaters took in less protein (still adequate) but more fiber, folate, vitamin C, vitamin E and magnesium. Intake and status of B12, vitamin D, iron, zinc, iodine and calcium were generally lower, and so was intake of EPA and DHA. Vegans had the lowest B12, calcium and iodine intakes⁴.

• B12: deficiency was common among vegetarians and higher in vegans¹. See vitamin B12.
• Iodine: no group, omnivores included, reached the optimal range, and vegans were lowest⁵.
• Omega-3: non-fish-eaters' blood levels were closer to fish eaters' than their intake suggested, possibly because they convert more plant omega-3¹².
• Bones: vegans had more hip fractures⁶.

Planning works. The Academy of Nutrition and Dietetics holds that appropriately planned vegetarian and vegan diets are healthful and adequate at every stage of life, with B12 from fortified foods or supplements².

Pooled cohorts and trials on heart health

A meta-analysis of 86 cross-sectional studies and 10 prospective cohorts found lower BMI, total and LDL cholesterol and glucose in vegetarians and vegans. In the cohorts, ischemic heart disease incidence or death was 25% lower (RR 0.75, 0.68 to 0.82) and total cancer incidence 8% lower (RR 0.92, 0.87 to 0.98); total cardiovascular disease, all-cause mortality and cancer mortality did not differ significantly. Vegans had 15% lower cancer incidence (RR 0.85, 0.75 to 0.95) from a small number of studies⁷. The 11-trial lipid meta-analysis found falls in total cholesterol (-0.36 mmol/L), LDL (-0.34), HDL (-0.10) and non-HDL cholesterol (-0.30), and no change in triglycerides⁹. The Stanford twin trial (22 pairs, 44 people) gave meals for the first four weeks; the vegan twin's LDL fell 13.9 mg/dL more (95% CI -25.3 to -2.4), fasting insulin 2.9 µIU/mL more and weight 1.9 kg more. It was small, short and open-label, and one author reported funding from Beyond Meat outside the study⁸.

Large cohorts: mortality, stroke and bones

Adventist Health Study 2 followed 73,308 people for a mean 5.8 years, with 2,570 deaths. All-cause mortality was lower in all vegetarians combined (HR 0.88, 0.80 to 0.97), vegans (0.85, 0.73 to 1.01), lacto-ovo vegetarians (0.91, 0.82 to 1.00) and pesco-vegetarians (0.81, 0.69 to 0.94), but not significantly in semi-vegetarians (0.92, 0.75 to 1.13), with larger effects in men¹⁰. Adventists are a health-conscious group, so the results may not transfer. In EPIC-Oxford (48,188 people, 18.1 years), ischemic heart disease was 13% lower in fish eaters (HR 0.87) and 22% lower in vegetarians (HR 0.78) than in meat eaters, which for vegetarians means about 10 fewer cases per 1,000 over 10 years. The vegetarian benefit weakened to HR 0.90 (0.81 to 1.00) after adjusting for cholesterol, blood pressure, diabetes and BMI. Vegetarians had 20% more total stroke (HR 1.20, 1.02 to 1.40), about three more cases per 1,000, mostly hemorrhagic¹¹. In the fracture analysis of EPIC-Oxford (about 54,900 people, 17.6 years), hip-fracture risk versus meat eaters was higher in vegans (HR 2.31, 1.66 to 3.22), vegetarians (1.25) and fish eaters (1.26); vegans also had more total fractures (HR 1.43). That is roughly 15 extra hip fractures per 1,000 vegans over 10 years, and it persisted after adjusting for calcium and protein⁶.

Plant-food quality

In the Nurses' Health Study, Nurses' Health Study II and Health Professionals Follow-up Study (about 209,000 people, 8,631 coronary events), an overall plant-based index was weakly protective (HR 0.92, 0.83 to 1.01, top versus bottom decile). A healthful index (whole grains, fruit, vegetables, nuts, legumes, oils, tea and coffee) gave HR 0.75 (0.68 to 0.83), while an unhealthful one (juices, sweetened drinks, refined grains, potatoes, sweets) gave HR 1.32 (1.20 to 1.46)³.

Nutrient gaps

A systematic review of 141 studies (its authors work for Unilever, which sells fortified foods) found nutrient inadequacies across all diet patterns, including meat-based ones⁴. A review of 18 articles using functional B12 markers reported deficiency in 62% of pregnant women, 25% to 86% of children, 21% to 41% of adolescents and 11% to 90% of older adults among vegetarians, with higher rates in vegans and in people vegetarian from birth¹. A meta-analysis of 11 studies (4,421 adults) found no group within the optimal urinary iodine range of 100 to 200 µg/L, with vegans lowest, particularly where salt is not iodized⁵. In EPIC-Norfolk (14,422 people with diet diaries; 4,902 with plasma fatty acids), non-fish-eaters took in 57% to 80% as much omega-3 as fish eaters, yet blood differences were far smaller¹².

Where the evidence conflicts

All-cause mortality is lower in Adventists but not significantly different in the pooled analysis. Stroke and fracture signals come mainly from one UK cohort. Vegetarians in these cohorts are often more health-conscious, so confounding is likely, and diets were self-reported. We found no long-term trial of hard outcomes; the twin trial lasted eight weeks.

Bottom line

Confidence is 4 of 5. Well-planned vegetarian and vegan diets reliably lower LDL cholesterol and are associated with less heart disease. They carry real nutrient risks, above all B12 and iodine, and possible bone and stroke risks that need more study.