Eat more plants, reduce your risk of death and disease? Original paper

In this meta-analysis of prospective cohort studies, higher adherence to a plant-based diet was associated with a reduced risk of death and chronic diseases.

The association between adherence to a plant-based diet and the risk of (i) all-cause mortality, (ii) cardiovascular disease, (iii) type 2 diabetes, and (iv) cancer.

A total of 2,230,443 participants (average ages of 25–87).

A meta-analysis of 76 prospective cohort studies was performed. The studies were conducted in North America, Europe, Asia, and Australia, and the follow-up duration ranged from 2 to 36 years. The vast majority of studies assessed dietary intake using food frequency questionnaires, while the rest used 24-hour dietary recalls.

Associations were analyzed according to (i) adherence to an overall plant-based diet, irrespective of the healthfulness of the foods consumed (i.e., the more plant-based foods consumed and the less animal-based foods consumed, the higher the overall plant-based diet adherence score) and (ii) the healthfulness of the plant-based diet. Higher adherence to an unhealthy plant-based diet indicated a higher intake of refined grains, starchy vegetables (e.g., white potatoes), and sugar (e.g., sweets/desserts, sugar-sweetened beverages), whereas higher adherence to a healthy plant-based diet indicated a higher intake of vegetables, fruits, whole grains, and legumes.

Compared to lower adherence to a plant-based diet, higher adherence to a plant-based diet was associated with a reduced risk of type 2 diabetes (−18%), all-cause mortality (−16%), cardiovascular disease (−10%), and cancer (−9%).

The magnitude of risk reduction slightly increased, except for all-cause mortality, when analyses were performed according to higher adherence to a healthy plant-based diet. In contrast, higher adherence to an unhealthy plant-based diet was associated with an increased risk of the examined outcomes.

Because all of the included studies were observational in nature, the results may be influenced by residual or unmeasured confounders, particularly because each of the included studies adjusted for a different set of confounders.