The Problem With Diet Studies: We Don't Eat Macronutrients

Jocelyne R. Benatar, MBChB, MD


September 05, 2018

Low-carbohydrate diets have become increasingly popular as a health and weight-loss strategy, especially in the West. Although short-term randomized studies support low-carbohydrate diets for weight loss,[1] effects on mortality are not known.

Two observational studies suggest that both high and low carbohydrate intake are associated with increased mortality. In the large, multinational PURE study,[2] high-carbohydrate diets were associated with increased mortality, whereas a meta-analysis of large cohort studies in North America and Europe found that low carbohydrate intake was associated with increased mortality.[3]

The ARIC Study

The Lancet Public Health recently published an article reporting results from the observational Atherosclerosis Risk in Communities (ARIC) study, which attempted to evaluate both matters in more detail.[4] In this observational study, more than 15,400 people in the United States were followed up for a median 25 years. Dietary information was gleaned from interviews that included a 61-item food frequency questionnaire at baseline and again at 6 years, with a focus on assessing the source or quality of proteins and fats consumed in low-carbohydrate diets. Outcomes data were from linked national health data, suggesting good follow-up.

A U-shaped association was seen between mortality and carbohydrate consumption, similar to the U-shape seen in many observational studies assessing mortality and risk factors such as alcohol intake,[5] body mass index,[6] fertility,[7] and sleep.[8] Diets both low (< 40% of energy) and high (> 70% of energy) in carbohydrates were associated with increased mortality, whereas moderate consumers of carbohydrates (50%-55% of energy) had the lowest risk for mortality. However, the results for the high-carbohydrate diet were not statistically significant, with wide confidence intervals. This finding was confirmed in a subsequent meta-analysis undertaken by the authors.

Unlike those of previous studies, the investigators of ARIC evaluated food sources in more detail, specifically for participants on low-carbohydrate diets. Replacing carbohydrates with animal-based proteins and fats was associated with a greater risk for mortality, whereas the inverse was true for plant-based proteins and fats, though this cohort was small. This is consistent with other large observational and randomized studies that suggest that a diet high in plant matter is healthier.[9,10,11,12]

For participants in the high-carbohydrate group, the source of carbohydrates is not known. This cohort may have consumed a large amount of processed foods and free sugars to achieve such a high intake of carbohydrates, and mortality may have been affected through other mechanisms—for example, a high salt intake.

This exposes the limitations of focusing studies on a macronutrient. People do not eat a macronutrient in isolation; they consume food as part of an overall dietary pattern. The carbohydrate content of a large bowl of fresh fruit and vegetables may be similar to that of a bowl of cornflakes, but the effects on health may be different on the basis of nutrients consumed.

The ARIC study had a number of other limitations. As an observational study, causality cannot be inferred. Food patterns are part of overall health behaviors that cannot be fully accounted for. Cohorts within the study were different; for example, those who consumed fewer carbohydrates were more likely to be male, smoke, have diabetes, and eat more animal-based foods and less plant-based foods, but they were more educated and from higher socioeconomic groups. The difference in mortality between groups was modest. And finally, intake of food was self-reported, which is prone to inaccuracy.[13]

This study does fit in with the narrative that "moderation is key." It may be that carbohydrates need to be consumed above a minimal level to provide energy for exercise and prevent excessive consumption of fat and protein, and too much carbohydrate increases the risk for diseases like obesity and diabetes. This study supports this narrative, but does not prove it.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: