American Journal of Medicine, internal medicine, medicine, health, healthy lifestyles, cancer, heart disease, drugs

How Pure is PURE? Dietary Lessons Learned and Not Learned From the PURE Trials

vegetable juices and veggies on a table

The Prospective Urban Rural Epidemiology (PURE) cohort studies add a new level of understanding of some key environmental components of health.1 The population is diverse, and represents individuals from various socioeconomic levels with a long period of follow-up. The dietary macronutrient analysis of this cohort conducted by the study group2 has yielded interesting findings, undoubtedly warranting additional analysis.

The study concludes that high carbohydrate intake is associated with increased mortality. By contrast, a higher intake of both saturated and unsaturated fats were reported to be associated with lower total mortality. The authors note that a higher fat and lower carbohydrate diet is more optimal and that saturated fats do not necessarily have to be restricted.

We commend the authors of the PURE trial for putting together a comprehensive assessment of international dietary habits. However, there are some important points to consider. First, this study’s conclusions were based on food frequency questionnaires, which are historically notable for underreporting of food intake and, notably, of fat intake, thus obscuring dietary data. Making broad statements about dietary guidelines based on food recall is difficult and always a limitation of studies based on surveys.3 Further, the fact that the questionnaires were done only at baseline leaves a large assumption that subjects’ dietary habits did not change over time.

With these caveats in mind, the authors noted a significant amount of carbohydrate intake around the globe. In this trial, 52% of the population consumed an average of at least 60% of energy from carbohydrates, especially in the low- and middle-income countries. However, the macronutrient analysis did not qualify the type of dietary carbohydrate as either simple (with low glycemic index) or complex (high in fiber and low glycemic index). In fact, many of the low-income countries that had high consumption of carbohydrates obtain them from carbohydrates from refined sources, such as white rice and white bread, which appear to correlate with the obesity epidemic4and the rapid increase of diabetes in low-income populations.5 Consumption of foods with a high glycemic index is also associated with a linear increase in coronary heart disease risk and an increased risk of total mortality and cardiovascular events.6 This trial similarly noted carbohydrates to be associated with worse overall mortality (hazard ratio [HR] 1.28) and non-cardiovascular disease (CVD) mortality (HR 1.36) overall, consistent with prior data.

Two studies have recently been published with important dietary lessons. One study showed that plant-based diets composed of healthful complex carbohydrates were associated with decreased CVD, whereas plant-based diets with more refined foods were associated with increased CVD.7 Another recent analysis that looked at differences in diets and mortality noted that changes in diet toward whole grains, vegetables, fruit, and fish or other omega-3 fatty acids was associated with decreased mortality.8 Both of these important studies emphasize the value of complex carbohydrates, fruits, and vegetables as the fundamental components of an optimal diet.

Also notable in this study was the role of fats in the diet in relation to CVD. Lower total mortality (HR 0.86), stroke (HR 0.79), and non-CVD death (HR 0.86) were noted in groups with the highest intake of saturated fats, but interestingly, stroke was not significantly lowered by higher intake of monounsaturated fats or polyunsaturated fats. Diets higher in fat in indigent populations, however, may reflect a need for any type of fat in the diet to treat nutritional deficiencies. Diets with the highest protein content showed the lowest rates of overall death and non-CVD death, but no difference in incidence of CVD or CVD death, again hinting that adequate nourishment in the diet was likely the reason for less death and non-CVD death. If based on the findings of the study, we are to conclude that saturated fats do not have a negative health impact on population health, we would also have to ignore much of the earlier data showing a correlation between saturated fats and increased cardiovascular disease,9 including the powerful data on animal protein that comes coupled with saturated fat linking the downstream products (trimethylamine N-oxide) to the development of atherosclerosis.10

It is also important to delve into the potential effects of socioeconomics on the results reported in the trial. The authors extensively adjusted for education, household income and wealth, income level of the country, and rural and urban location, yet this did not alter results, making the association of refined carbohydrates with poor outcomes more likely to be valid. However, the majority of the subjects eating refined carbohydrates were poor, and therefore, confounders associated with being poor cannot fully be ruled out. In addition, access to medical care was not specifically examined, although it would be expected that it would be closely tied to income and urban vs rural location, which were controlled for.


The findings of the PURE trial are eye-opening with respect to the nutritional concerns that are likely different in areas of poverty. This is of even greater concern at a time when low and middle income countries such as Brazil and India are encountering the incorporation of Westernized, highly processed, inexpensive food to their diet, leading to significant increases in obesity, diabetes, and heart disease on top of the health concerns of poverty and famine. This type of poor-quality carbohydrate as the bulk of the diet is certainly an unfair comparator to saturated fat. In fact, the question of whether processed carbohydrates or saturated fats are worse for cardiovascular health does not warrant scientific evaluation because we have adequate evidence to indicate that both are harmful (and likely so) and should be limited, if not eliminated from the diet.

In addition, oversimplification of macronutrients (fat, carbohydrates, protein) as good or bad sets the field back through misuse by the lay public and potentially, clinicians who are guiding the public. This same concept of oversimplification may be what triggered an epidemic of obesity and diabetes in the past several decades, driven largely by conclusions from observational data. Moving forward to advance the field of nutritional science, we absolutely agree with the suggestion by Dr. Marion Nestle that “What we really need are well designed studies of dietary patterns—the ones done to date suggest that largely plant-based diets are associated with excellent health and longevity”.11

In sum, these trials have largely compared many differences and can therefore only guide us on relative, not absolute, dangers or benefits of nutritional components. Consistent with several recent publications indicating higher mortality with refined grains and sugar, this study demonstrates that consuming highly refined carbohydrate-laden diets exceeded the mortality rate associated with eating fat. This is not a call to increase fat consumption, but to decrease consumption of low-fiber, low-quality, high-sugar items, to which so many Americans subscribe daily, and which people in countries of lower economic status use for sustenance.

To read this article in its entirety please visit our website.

-Eugenia Gianos, MD, Kim A. Williams, MD, Andrew M. Freeman, MD, Penny Kris-Etherton, PhD, RD, Monica Aggarwal, MD

This article originally appeared in the May issue  of The American Journal of Medicine.

Comments are closed.