Plant-based diet statistics depend on definition: a U.S. trend study classified a day as plant-based when at least 50% of protein came from plants, while health research also uses vegetarian, vegan, and healthful or unhealthful plant-based indexes. The evidence below separates survey estimates, observational associations, and short randomized trials.
Key Plant-Based Diet Statistics
The most broadly useful figures are:
- In 1999–March 2020, 15.8% of U.S. adults aged 20 and older consumed a study-defined plant-based diet on a given day.
- In 1999–March 2020, 6.8% of U.S. adults aged 20 and older consumed a study-defined vegetarian diet on a given day.
- From 1999–2000 to 2017–March 2020, the U.S. plant-based-diet share rose from 14.4% to 17.2% among adults aged 20 and older.
- From 1999–2000 to 2017–March 2020, the U.S. vegetarian-diet share rose from 6.4% to 8.2% among adults aged 20 and older.
- Among U.S. plant-based-diet consumers, the share aged 60 or older rose from 28.5% in 1999–2000 to 34.4% in 2017–2020.
- Among U.S. plant-based-diet consumers, mean HEI-2020 score rose from 52.1 to 55.8 between 1999–2000 and 2017–March 2020.
- Among U.S. plant-based-diet consumers, ultraprocessed foods supplied 52.7% of energy in 2017–March 2020.
- Across prospective adult cohorts through June 2023, higher plant-based adherence was associated with a pooled relative risk of 0.82 for incident type 2 diabetes.
- Across prospective adult cohorts through June 2023, a healthful plant-based pattern was associated with a relative risk of 0.79 for type 2 diabetes.
- In an 8-week randomized trial of adult identical twins, the vegan diet reduced body weight by 1.9 kg relative to the omnivorous diet.
- Across prospective adult cohorts through June 2023, higher plant-based adherence was associated with a relative risk of 0.90 for incident cardiovascular disease.
- Across prospective adult cohorts through June 2023, higher plant-based adherence was associated with a relative risk of 0.86 for coronary heart disease.
- In an 8-week randomized trial of adult identical twins, the vegan diet lowered LDL cholesterol by 13.9 mg/dL relative to the omnivorous diet.
- Across prospective adult cohorts through June 2023, higher plant-based adherence was associated with a relative risk of 0.88 for total cancer.
- Across prospective adult cohorts through June 2023, plant-based adherence was associated with a relative risk of 0.68 for pancreatic cancer.
- Across prospective adult cohorts through June 2023, higher plant-based adherence was associated with a relative risk of 0.84 for all-cause mortality.
- The 2023 evidence review covered 76 publications from 55 prospective cohorts published through June 2023.
- The included prospective cohorts represented 2,230,443 participants and 174,435 deaths across studies published through June 2023.
Contents
- U.S. plant-based diet adoption statistics
- Plant-based eating trends, diet quality, and processing
- Plant-based diets and type 2 diabetes
- Plant-based diets and cardiovascular outcomes
- Plant-based dietary patterns and cancer
- Plant-based diets, mortality, and randomized trials
U.S. Plant-Based Diet Adoption Statistics
The NHANES trend study used one 24-hour dietary recall, so its estimates describe what adults reported eating on a particular day. Its plant-based definition required at least 50% of protein from plants; that is not the same as identifying as vegan or vegetarian. Trends in Plant-Based Diets among United States Adults, 1999–March 2020
| Measure | 1999–2000 | 2017–March 2020 |
|---|---|---|
| Adults consuming the study-defined plant-based diet | 14.4% | 17.2% |
| Adults consuming the study-defined vegetarian diet | 6.4% | 8.2% |
| Plant-based-diet consumers aged 60+ | 28.5% | 34.4% |
| Mean age of plant-based-diet consumers | 48.4 years | 50.5 years |
Across the full 1999–March 2020 series, 15.8% of U.S. adults aged 20 and older consumed the study-defined plant-based diet on a given day, compared with 6.8% for the study-defined vegetarian diet. The plant-based share had a statistically significant linear increase, although the adjusted trend was weaker and the final cycle was shortened.
Among plant-based-diet consumers, women represented 57.1% in 1999–2000 and 55.5% in 2017–2020. These are descriptive subgroup shares, not a population-wide estimate of adherence differences by sex.
The share of non-Hispanic Black adults consuming the study-defined plant-based diet rose from 10.8% to 14.7% between 1999–2000 and 2017–March 2020. Comparable trends were not statistically significant for non-Hispanic White or Mexican American adults.
Plant-Based Diet Trends, Diet Quality, and Processing
A plant-based label does not establish diet quality. In the NHANES analysis, the mean plant-based-diet HEI-2020 score rose from 52.1 in 1999–2000 to 55.8 in 2017–March 2020, while the vegetarian-diet score rose from 49.7 to 54.8.
HEI is an alignment score, not a clinical outcome.
| Food-pattern measure | Beginning | End |
|---|---|---|
| Ultraprocessed food energy share | 50.7% in 1999–2000 | 52.7% in 2017–March 2020 |
| Processed culinary ingredient energy share | 4.7% in 1999–2000 | 6.5% in 2017–March 2020 |
| Plant-based-diet HEI-2020 | 52.1 | 55.8 |
| Vegetarian-diet HEI-2020 | 49.7 | 54.8 |
The change in ultraprocessed-food energy share was not statistically significant, with P trend 0.34. The increase in processed culinary ingredients was significant, but it describes energy share rather than grams or a measured health effect.
Bread contributed more than 10% of energy in every NHANES cycle among plant-based-diet consumers. That is a food-subgroup contribution, not a standalone judgment about diet quality, and NOVA classifications may misclassify foods when brand or ingredient information is incomplete. NHANES trend study
Plant-Based Diet Statistics for Type 2 Diabetes
A 2023 systematic review and meta-analysis synthesized prospective adult cohorts published through June 2023. It included 60,718 incident type 2 diabetes cases; these cases describe the evidence base, not population prevalence. Associations between plant-based dietary patterns and risks of type 2 diabetes, cardiovascular disease, cancer, and mortality
| Dietary pattern or analysis | Association with incident type 2 diabetes | Evidence context |
|---|---|---|
| Higher plant-based adherence | RR 0.82 (95% CI 0.77–0.86) | Prospective cohorts |
| Healthful plant-based pattern | RR 0.79 (95% CI 0.72–0.87) | Prospective cohorts |
| Unhealthful plant-based pattern | 8% higher risk (95% CI 2%–13%) | Prospective cohorts |
| Increased adherence over time | RR 0.83 (95% CI 0.71–0.96) | Four change-analysis studies |
The healthful indices emphasized vegetables, fruits, whole grains, and legumes, whereas unhealthful indexes included less-healthful plant foods such as refined grains and sugary foods. The overall pooled analysis had I² heterogeneity of 60.9%, and the change analysis had heterogeneity of 71.5%, indicating important differences between studies.
Randomized evidence is shorter-term. In an 8-week trial of adult identical twin pairs, vegan-diet participants reduced body weight by 1.9 kg relative to omnivorous-diet twins and reduced fasting insulin by 2.9 μIU/mL; neither result establishes long-term weight maintenance or a diabetes diagnosis.
In a 74-week randomized trial among adults with type 2 diabetes, HbA1c fell more on a low-fat vegan diet than on a conventional diabetes diet after controlling for medication changes. The accessible review record reports the direction and significance but not one comparable numeric effect.
Plant-Based Diet Data for Cardiovascular Outcomes
Prospective cohort evidence through June 2023 associated higher plant-based adherence with a relative risk of 0.90 for incident cardiovascular disease and 0.86 for coronary heart disease. A healthful plant-based pattern was associated with 0.85 for cardiovascular disease, while an unhealthful pattern was associated with a 14% higher risk; heterogeneity was 49.8% for overall cardiovascular disease and 71.1% for the unhealthful pattern.
- The pooled stroke association was RR 0.92, with a 95% CI of 0.85–1.00; the interval reached 1.00 and heterogeneity was 51.6%.
- The pooled heart-failure association was RR 0.89, with a 95% CI of 0.75–1.06; the interval included no association.
An earlier meta-analysis searched studies through May 2021 and included 410,085 participants across 13 prospective studies. It reported a pooled hazard ratio of 0.92 for cardiovascular mortality, 0.90 for incident cardiovascular disease, and 0.81 for incident cardiovascular disease among vegetarians versus meat-eaters; heterogeneity was high at I² 88.5% and 87.2% for the first two estimates. The Association of Plant-Based Diet With Cardiovascular Disease and Mortality
Short randomized trials measured biomarkers rather than events. In an 8-week identical-twin trial, LDL cholesterol was 13.9 mg/dL lower with the vegan diet relative to the omnivorous diet.
In a 16-week randomized crossover trial of 62 overweight adults, the low-fat vegan diet reduced LDL cholesterol by 15.3 mg/dL, total cholesterol by 18.7 mg/dL, systolic blood pressure by 3.4 mmHg, and diastolic blood pressure by 4.1 mmHg. The Mediterranean comparator reduced systolic pressure by 9.3 mmHg and diastolic pressure by 7.3 mmHg, so those comparator figures are not vegan-diet estimates. A Mediterranean Diet and Low-Fat Vegan Diet
Plant-Based Dietary Patterns and Cancer Statistics
The same prospective-cohort review found an overall relative risk of 0.88 for total cancer through June 2023. A healthful plant-based pattern was associated with 0.86, while an unhealthful pattern was associated with a 7% higher risk; the total-cancer heterogeneity was 30.2%.
| Outcome | Pooled association | 95% confidence interval |
|---|---|---|
| Breast cancer | RR 0.91 | 0.86–0.95 |
| Digestive-system cancer | RR 0.82 | 0.72–0.94 |
| Pancreatic cancer | RR 0.68 | 0.55–0.84 |
| Prostate cancer | RR 0.87 | 0.77–0.99 |
| Colorectal cancer | RR 0.90 | 0.79–1.02 |
| Lung cancer | RR 0.82 | 0.54–1.26 |
| Stomach cancer | RR 1.73 | 0.90–3.31 |
These cancer-specific estimates are observational associations, not randomized prevention results. The colorectal, lung, and stomach confidence intervals include no association; the stomach estimate’s wide interval should not be interpreted as evidence of harm.
The prostate-cancer analysis had heterogeneity of 53.3%, colorectal cancer 61.8%, and the source notes that cancer-site groupings and dietary definitions varied across studies. Systematic review and meta-analysis
Plant-Based Diets, Mortality, and Randomized Trials
The 2023 review covered 76 publications from 55 individual cohorts. The cohorts represented 2,230,443 participants, 157,335 cardiovascular disease cases, 57,759 cancer cases, and 174,435 deaths, with follow-up ranging from 2 to 36 years; participants and cases overlap across outcome-specific publications.
- Higher plant-based adherence was associated with RR 0.84 for all-cause mortality (95% CI 0.78–0.92), with I² heterogeneity of 95.4%.
- A healthful plant-based pattern was associated with RR 0.86 for all-cause mortality (95% CI 0.80–0.92), with I² heterogeneity of 91.9%.
- An unhealthful plant-based pattern was associated with a 16% higher all-cause mortality risk (95% CI 9%–25%), with I² heterogeneity of 84.8%.
- Plant-based adherence was associated with RR 0.79 for cardiovascular mortality (95% CI 0.72–0.87), with I² heterogeneity of 77.0%.
- Plant-based adherence was associated with RR 0.82 for cancer mortality (95% CI 0.71–0.93), with I² heterogeneity of 88.4%.
- In studies measuring dietary change, increased plant-based adherence was associated with RR 0.95 for mortality (95% CI 0.91–1.00); the result was marginal and based on limited change-analysis evidence.
Among the 76 publications, 67, or 88.2%, had study-quality scores of at least 10 on a 14-point assessment. That score is an appraisal measure rather than an outcome, and observational confounding remains possible.
The randomized trials in this evidence base were short and used selected adult samples. Their weight, insulin, lipid, and blood-pressure findings are useful measured outcomes, but they do not establish long-term disease or mortality benefits, and most disease and mortality estimates above should not be read as causal prevention. Systematic review and meta-analysis Identical-twin randomized trial