The best diet for health: What the science really says

On this blog I’ve argued in favor of a whole food, mainly plant-based diet. My argument so far has been based on evolution. In this blogpost, I’m going to review the results of scientific studies on the health effects of such diets.

Randomized controlled studies versus observational studies

To compare the health effects of whole food diets to other diets, we can look at the results of scientific studies. In observational studies one compares the health of groups of people who stick to that diet to the health of another group of people who stick to some other diet. In randomized controlled studies, one recruits people for a diet study and assigns the participants randomly to one or the other diet.

In the scientific literature, randomized controlled studies are considered to be far more reliable than observational studies, as the latter type of studies can be affected by many confounding factors. Unfortunately, it’s almost impossible to do a randomized controlled study comparing a strict whole food diet to a normal Western-type diet where a significant fraction of the calories comes from refined fats and sugars.

The volume of the food for a fixed amount of calories will be much larger in case of a whole food diet compared to a Western-type diet. Just 8 spoons of olive oil contains the same amount of energy as 2 pounds of potatoes. It would take many months for people to get adapted to a strict whole food diet, making a randomized controlled study impractical.

Observational studies are also hard to perform. Here the problem is that very few people eat a strict whole food diet. A large fraction of the people who do eat such diet are on weight loss programs or have suffered cardiovascular problems and use this diet as a complementary treatment. These and other confounding factors make it difficult to get to reliable results.

One can also attempt to extract the health effects of a whole food diet by studying the few populations on Earth that stick to a whole food diet. The problem with this is that these populations live in Nature far from civilization. There is then a lack of reliable medical data. Also the members of such populations have a low life expectancy due to a lack of medical facilities.

Nevertheless, despite these difficulties a few studies have been performed. Let’s start with looking at a few population-based studies.

Health and diet of traditional Ugandans

When Uganda was a British colony, the African population was eating a reasonably strict whole food diet, while the Asian population there was eating a diet containing significant amounts of refined oils and fats. A study published in 1959 in the Lancet (free reprint), showed that there were huge differences in the incidence of heart disease in both populations:

In the African population of Uganda coronary heart disease is almost non-existent. This statement is confirmed by adequate necropsy evidence1. In the Asian community, on the other hand, coronary heart disease is a major problem.


The diet of the Africans was as follows:

The staple foods, green plantain and sweet potatoes, are steamed in banana leaves; cassava, yams, maize, and millet are also staple commodities in particular of the non-Baganda groups, while pumpkins, tomatoes, and green leafy vegetables are taken by all. The adequacy of protein in the diet depends almost entirely on the extent to which pulses, groundnuts, and cereals are used. Most meals are served with a sauce made of groundnuts, beans, and a mixture of vegetables, and occasionally meat or fish, and these are fried in very small amounts of fat. 

The amount of fat varied between 16 to 20 grams for the poorer people to about 40 grams for richer people on a 2000 Kcal diet. So, at most the Africans were getting less than 20% of their calories from fat. The difference in cardiovascular disease incidence between the Asians and Africans was spectacular, and according to the article, this was due to the low fat content of the African diet leading to low cholesterol values that don’t increase with age.

Health and diet of the Tsimané people

The Tsimané people of Bolivia have until recently stuck to a whole food diet. We can read here:

The Tsimane diet was characterized by high energy (2422–2736 kcal/d), carbohydrate (376–423 g/d), and protein (119–139 g/d) intakes; low fat intake (40–46 g/d); and low dietary diversity relative to the average US diet. Most calories (64%) were derived from complex carbohydrates. 

Kraft, T. S., Stieglitz, J., Trumble, B. C., Martin, M., Kaplan, H., & Gurven, M. (2018). Nutrition transition in 2 lowland Bolivian subsistence populations. The American Journal of Clinical Nutrition, 108(6), 1183–1195. 

Their diet is now changing due to the availability of Western-type foods. Like in case of the Ugandan population, the incidence of heart disease is very low. A recent study done using mobile CT scanners revealed that the Tsimané have very low levels of arterial plaque:

Despite a high infectious inflammatory burden, the Tsimane, a forager-horticulturalist population of the Bolivian Amazon with few coronary artery disease risk factors, have the lowest reported levels of coronary artery disease of any population recorded to date. 

Kaplan, H., Thompson, R. C., Trumble, B. C., Wann, L. S., Allam, A. H., Beheim, B., … Thomas, G. S. (2017). Coronary atherosclerosis in indigenous South American Tsimane: a cross-sectional cohort study. The Lancet, 389(10080), 1730–1739.

The diets of the Tsimané and Ugandans are quite similar and the health effects as far as cardiovascular health is concerned is also similar. The incidence of cardiovascular disease in both populations is an order of magnitude less than in the Western population.

Other population-based studies

Similar effects have also been noted in other indigenous populations who survive on a whole food, mainly plant-based diet, but not in populations whose diets include large amounts of animal products. Recently the widely held belief that Eskimo’s rarely get heart disease was debunked.

Studies of populations living in the civilized world, such as the Adventist Health Studies and the Okinawan diet studies are consistent with the general pattern that a whole food diet that’s mainly plant-based is the best diet for health. However, in these studies less spectacular reductions in heart disease are found compared to the findings of studies in indigenous populations. The difference between the whole food diets eaten by populations in the civilized world and indigenous populations is that the civilized world-variants of this diet contains a lot more refined fats.

Controlled studies

Dr. Esselstyn has performed a clinical study of a strictly whole food plant-based diet. Heart patients were put on a very strict whole food plant-based diet. The results were quite spectacular:

Of the 198 patients with CVD, 177 (89%) were adherent. Major cardiac events judged to be recurrent disease totaled one stroke in the adherent cardiovascular participants—a recurrent event rate of .6%, significantly less than reported by other studies of plant-based nutrition therapy. Thirteen of 21 (62%) nonadherent participants experienced adverse events.

Dr. Esselstyn demonstrates the changes in the angiogram of one of his patients in this video:

Reversal of coronary artery disease via plant-based nutrition

Other such studies with smaller groups of patients have yielded similar results when the diet was as strict as in this study, while less spectacular results have been obtained using diets that are less strict. For example, this randomized controlled study by Dean Ornish found significantly less cardiac events in the diet group compared to the control group.

Other studies

The vast majority of the studies looking into the relation between health and diet have focused on variants of the Western diet. Only a small fraction of the World’s population eats the RDA for fruits and vegetables. This RDA is already much lower than the quantities one would need to eat when on a whole food plant-based diet. Many of these studies have been conducted in scientifically very rigorous ways, the results of such studies then end up forming the basis of the official guidelines for the diet.

The problem is then that the information obtained from these studies is then only valid for diets that are close to Western-type diets. It’s not always possible to extrapolate the results to diets such as a whole food diet that are very different from the Western diet. Suppose that the relationship between a disease risk and the diet is as follows:

Example of disease risk vs. diet that can be extrapolated to pwhole-food region from conventional study results
Fig. 1. Example of a disease risk vs. diet in a case where conventional study results will get the global picture qualitatively correct

Then studies performed at diets that are between 15% and 25% similar to an optimal whole food diet would miss the global picture but the conclusions extrapolated toward 100% similarity would still be qualitatively valid. Such results have e.g. been found for the relation between dietary fiber and cardiovascular disease. But it’s entirely possible that far away from the true optimum one or more local optima exist.

Rigorous scientific results can be misleading

The relationship between a disease risk and diet can look like this:

Fig. 2. Example of a disease risk vs. diet where conventional study results will miss the correct global picture.

Studies done on people sticking to Western-style diets between, say, 15% and 25% similarity to optimal whole food diets will then only detect the region near the sharp local minimum. Many of the recent controversies in the debate on diets may be caused by such an effect. Take e.g. the recent debate on fat intake in relation to heart disease or whether a moderate salt intake is better than a low salt intake.

Reaching a local optimum far away from the true global optimum for a very complex system like the human body that have multiple redundant mechanisms, will in general require some fine tuning. The optimum then isn’t very broad, the way we need to eat to get to such a result will then differ from person to person and will also change over time. Exactly such a result has been found in the recent Predict Study. It suggests that we would all need personalized diets for optimal health.

Conclusion on best diet for health

Scientific results when properly interpreted, overwhelmingly support the idea that whole food, mainly plant-based diets are the best for health. The results are fully consistent with what one can guess based on the arguments from evolution that I’ve elaborated in in the previous blogposts here:

Living organisms are extremely robust physical systems, they’ve been optimized under natural conditions where they would eat foods they can find in Nature. This suggests that the best diet for us should have the following properties:

  • High in whole grains and starches
  • High in fruits and vegetables
  • Moderate amounts of nuts and seeds
  • Low in refined fats
  • Low in refined carbs
  • Low in animal sources of foods
  • Low in salt

In Nature, we would be able to find carb-based foods more easily than foods rich in fats such as nuts and seeds. We would have to do without any refined oils and carbs. We wouldn’t get much salt either.

Now,there are many possible variants of this type of diet. One can have preferences for particular types of vegetables, whether or not it’s a strict vegan diet, whether or not refined oils are totally banned or allowed in limited quantities and many other details. Do we expect such details to matter? The Tsimané study suggests not, they have excellent cardiovascular health despite eating a rather one-sided diet.

It’s no surprise that the optimal diet should not be sensitive to the precise details. Animals have evolved to get adapted to an environment that isn’t constant. The food animals can find will fluctuate due to seasonal changes, prolonged droughts or other weather events. So, we should expect that we’re optimally adapted to a wide range of natural whole food diets.

Conversely, a type of diet that requires getting a lot of the details right in order to get enough of the essential vitamins, minerals and amino acids, is unlikely to be a healthy diet. For optimal health we need to get a lot more nutrients than just the essential nutrients, many new discoveries are made on a regular basis.

For example, getting up to 2.5 grams per kg bodyweight of protein is likely better than the RDA of 0.8 grams per kg bodyweight. On a whole food diet you would have gotten this larger amount of protein all along, while on most other diets, you would likely have gotten a lot less. And it’s not easy to get enough protein on these other diets without using protein powder supplements.

We can easily verify that a whole food diet does indeed yield enough of the essential nutrients without requiring it to be fine tuned. We can use online tools such as the meal plan calculator to compose meals based on only whole foods. When the total amount of calories reaches 2500 Kcal, then the amount of protein, vitamins and minerals is typically going to be more than the RDA. An exception here is vitamin A, this will fluctuate significantly. It’s then no surprise that the body stores vitamin A.

In the next blogposting I’ll write about the best way to get started on a whole food diet. If you want personal assistance to start on this type of diet, contact me via this contact form if you also need to lose weight, or via this form if weight loss is not the primary concern.