SYNOPSIS: This large-scale, observational, prospective study investigating types of plant-based diets found an increased risk of cardiovascular disease in adherents to plant-based diets containing foods such as fruit juices, refined grains, sweetened beverages, and desserts.

SOURCE: Satijam A, Bhupathiraju S, Spiegelman D, 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.

It is widely accepted that vegetarian diets are associated with lower risk of coronary heart disease (CHD).1 Or are they? Although many studies have demonstrated this association, Satijam et al noted two crucial limitations of these earlier studies. Specifically, they postulated that all plant-based foods are not equally protective of cardiovascular risk and noted that even incremental shifts in composition of diet may affect the degree of cardiovascular protection. Furthermore, they ascertained that previous studies were not designed to discern these subtleties of diet.

To investigate further, the researchers obtained dietary data from the Nurses’ Health Study2 beginning in 1976, Nurses’ Health Study 22 beginning in 1989, and the Health Professionals Follow-up Study3 beginning in 1986 to analyze any association between specific dietary factors and later development of CHD. After excluding participants with baseline CHD and other specified conditions, baseline data from 160,000 women and 40,000 men were available for analysis.

They created three types of indices for plant-based diets — each index allowed for continuous gradation within the general category. The expectation was that this approach allows analysis of gradual changes in diet. The plant-based diet index (PDI) weights plant-based foods over animal foods; a healthful plant-based diet index (hPDI) weights dietary consumption of whole grains, fruits, nuts and vegetables, and vegetable oils; and an unhealthful plant-based diet (uPDI) emphasizes processed foods, refined grains, sweetened beverages, and sweets. As all the diets allowed for gradation, each included some degree of animal-derived foods.

Satijam et al examined dietary data from the participants’ food frequency questionnaire completed every two to four years and created 18 food groups. The food groups were assigned positive or negative scores depending on the index. (See Table 1.) For example, fruits were assigned positive scores for PDI and hPDI and negative scores for uPDI, while sweets were assigned positive scores for PDI and uPDI and negative scores for hPDI. All animal food products were assigned negative scores. The indices were obtained by summing the scores for the 18 food groups; these were adjustable to account for changes over time. Higher indices indicated lower consumption of animal foods; however, it is important to note participants in the highest PDI decile still consumed an average of three servings a day of animal food (the lowest contained five to six servings of animal food daily). Additionally, the authors conducted a separate analysis by assigning positive values to healthy animal foods, such as fish and eggs, and reverse scores to unhealthy animal foods, including red meat and animal and dairy fat. This was a prospective study, first collecting the dietary information, adjusting for variables over time, and then looking at incidence of CHD over a period of 20 to 30 years.

Table 1: Examples of Food Groups and Scores

Foods assigned positive scores for hPDI/negative scores for uPDI

Foods assigned positive scores for uPDI/negative scores for hPDI

Whole grains, cooked oatmeal, brown rice

Refined grains, white bread, white rice, sweets

Fruits and vegetables

Fruit juices

Nuts, legumes


Tea, coffee

Sugar-sweetened beverages

Selected Results

Each index was divided into deciles reflecting increasing compliance with the diet, and cumulatively averaged over time to evaluate the effect of a long-term diet. Each decile was evaluated separately from each data source (Nurses’ Health Study, Nurses’ Health Study 2, and Health Professionals Follow-up Study); data then were pooled for further analysis.

The pooled results adjusted for multiple variables revealed an inverse association between CHD and adherence to either PDI and hPDI. This inverse association was strongest when comparing the first decile (lowest compliance with diet) to the last decile (highest compliance with diet.) The association was more pronounced for hPDI than for PDI (See Table 2.) Interestingly, the inverse association was evident by the second decile for adherents to both diet types, but appeared to accelerate by the sixth decile. On the other hand, the association between uPDI and CHD was positive in all deciles. Table 2 includes results from decile 10 for each index (most compliant with diet), with pooled data from all three sources and applying multivariable adjustment for age, smoking status, and weight, among other factors.

Table 2: Selected Results of the Top Decile, With HR and P Values Reflecting Comparison of Extreme Deciles Within Each Group


Hazard Ratio (95% CI)

P value

PDI decile 10

0.92 (0.83-1.01)

= 0.003

hPDI decile 10

0.75 (0.68-0.83)

< 0.001

uPDI decile 10

1.32 (1.20-1.46)

< 0.001

When hPDI was modified to score fish, egg, and most dairy positively (rather than lumping these with all other animal products), results changed very little. The row below illustrates these results.

hPDI including egg, fish, most sources of dairy

0.78 (0.71-0.86)

< 0.001


“Eat food. Not too much. Mostly plants.” This plain-spoken advice comes from Michael Pollan, well-known writer about food and health topics and author of several books, including The Omnivore’s Dilemma.4 Many in the general public and in medicine alike view his message (along with confirming medical studies regarding the health benefits of plant-based diets) quite seriously, and strive to adjust diets accordingly. However, the application of a plant-based dietary approach varies widely. For example, in a recent U.S. News & World Report article regarding the “best” plant-based diets, the experts whittled the choices to no less than 12, including Mediterranean, vegan, vegetarian, and flexitarian to name just a few.5 A basic internet search reveals a range of even more vegetarian diets, including pesco-vegetarian and lactovegetarian.6 When faced with a seemingly endless array of choices, finding the type of plant-based diet most beneficial to health seems a formidable task.

Satijam et al helped to clarify and put a scientific spin on this endeavor by investigating what types of “mostly plant” foods are beneficial for cardiovascular health and by trying to understand if partial compliance or gradation of dietary adherence confers cardiovascular benefits. Additional information from this study is the analysis of data pertaining to diets containing animal foods, such as fish, eggs, and some dairy, thought to be beneficial to health.

The numbers behind this study are impressive, with more than 200,000 subjects studied for more than 20 years. However, one limitation is the lack of geographical, racial, occupational, and socioeconomic diversity in subjects — by definition all are U.S. healthcare professionals. Extending the pool to encompass a more diverse group would help confirm or further define the health effects of dietary changes. Likewise, looking at the health effects of dietary change at specific ages can assist in giving patients relevant and applicable information. Other limitations might be the lack of inclusion of fat source(s) or dairy consumption, known contributors to health outcomes.

Other recent studies support the benefits of plant protein intake in particular. In 2016, Song et al examined U.S. health professionals’ diet and all-cause mortality and concluded that higher ingestion of animal protein was associated with higher mortality rates (and the converse for higher intake of plant protein).7

The Satijam et al study as it stands has immediate clinical applications. The results clearly support the premise that all plant-based diets are not “created equal;” that is, every plant-based food does not convey equal health benefits. The inverse association of the hPDI with CHD risk implies that a diet rich in whole grains, fruits, vegetables, nuts, and legumes conveys cardiovascular protection and can be recommended “whole-heartedly.” The positive association with uPDI and CHD implies the opposite — that “vegetarian” foods, such as processed grains and sweetened beverages, do not convey health benefits. Furthermore, it is useful to be able to tell patients that strict adherence to a diet with healthy foods is not necessary to achieve a degree of protection and that this study looked at gradation of adherence, that is with an emphasis on these healthy foods in a diet rather than absolutes.

When working with patients preventively, providing information that stepwise dietary changes can help with cardiovascular protection may make the prospect of dietary modification seem attainable. Reminding patients that protein can be obtained from plant sources, that not all non-animal foods are equally beneficial, and that some animal foods, such as fish, eggs, and nonfat dairy, may convey health benefits are important elements to consider when helping shape a heart-healthy diet.

If a patient has clear cardiovascular risk or has suffered a cardiovascular event, this information grows in importance. Providing patients with clear data and information helps move vague dietary recommendations to clear and concrete implementation — a useful and natural step for the integrative provider in clinical practice.


  1. Dinu M, Abbate R, Gensini GF, et al. Vegetarian, vegan diets and multiple health outcomes: A systematic review with meta-analysis of observational studies. Crit Rev Food Sci Nutr 2017;57:3640-3649.
  2. Nurses’ Health Study. Available at: Accessed Dec. 10, 2017.
  3. Harvard T.H. Chan School of Public Health. Health Professionals Follow-up Study. Available at: Accessed Dec. 10, 2017.
  4. Michael Pollan. Available at: Accessed Dec. 19, 2017.
  5. U.S. News & World Report. Best Diets Overall. Available at: Accessed Jan. 5, 2018.
  6. Vegetarian Society. What is a vegetarian? Available at: Accessed Jan. 5, 2018.
  7. Song M, Fung TT, Hu FB, et al. Association of animal and plant protein intake with all-cause and cause-specific mortality. JAMA Intern Med 2016;176:1453-1463.