The health impact of carbohydrate-rich foods has been the subject of much discussion and even controversy, particularly the use of the Glycemic Index, described forty years ago. The publication, a few days ago, of an article in the New England Journal of Medicine sheds new light on this topic, which we will examine in detail.
The use of carbohydrates as an energy source has likely accompanied the human species since its appearance. Different populations, at different times, have used, and still use, quite diverse proportions of this macronutrient: the dietary model practiced in southern Japan and some areas of Korea consists of about 70-75% of energy derived from this source, while the diet practiced by hunter-gatherers in certain periods of prehistory would have a much lower proportion, which could be around 35% of the total energy value.
This apparent ability of the human organism to adapt to such different proportions of its energy supply must certainly have been an asset in its capacity for adaptation and, consequently, survival as a species.
Moving forward to the present day, we find ourselves in an environment that bears little or no resemblance to the one that led to the metabolic adaptation of humankind, which naturally raises some questions and problems. It is certainly no surprise that a large part of the inhabitants of this Blue Planet suffer from lifestyle-related diseases, namely obesity, diabetes, and cardiovascular diseases.
It is precisely by studying the (highly variable) response of diabetic individuals to different carbohydrate-containing foods that Professor David Jenkins' group at the University of Toronto, Canada, proposed the creation of the Glycemic Index (GI)as a way to better predict blood glucose behavior after consuming these foods. Thus, the GI was defined as a ratio between the glycemic impact of equal-carbohydrate amounts of a given food and the glycemic impact of a standard food, which could be anhydrous glucose or, for greater convenience, white bread. The glycemic impact was measured by measuring the incremental area under the blood glucose curve up to 2 hours after food consumption.
These initial trials revealed a reality that, in 1981, was still little known: the way blood glucose levels rose varied substantially from one food to another. Even foods containing the same type of carbohydrate can have very different glycemic indices (GI). Examples include legumes, whose glycemic impact is much lower than that of bread, for example, with starch being the predominant carbohydrate in both foods. Some mono- and disaccharides also show quite different GIs; for example, lactose has much lower values than glucose. Even sucrose, common sugar, has a lower GI than white bread, an apparently paradoxical effect, but one that finds an explanation in its chemical composition, as it contains fructose and glucose in equal proportions.
Equally interesting are the conclusions that reveal that the glycemic impact of starchy foods (those that contribute most to carbohydrate intake in our species, such as cereals, legumes, or tubers) depends significantly on their degree of processing. Heating, hydration, or grinding, for example, decisively affect the glycemic index (GI), and this is probably one of the factors that may contribute to the different health impact of processed foods compared to their more "natural" counterparts.
We have thus reached the point of trying to understand the true value of the GI for our health, and the path taken since that date (forty years ago) has not been particularly linear. This is yet another example demonstrating the immense difficulty in translating simple laboratory data into long-term health effects, especially in the area of nutrition and diet. This is because some of these effects take years or even decades to manifest. Furthermore, the magnitude of the effects is often small (but extremely important because it affects a very large number of individuals), which requires studies involving samples of tens or hundreds of thousands of people. These studies are rare, due to the tremendous logistical and economic involvement they represent, but fortunately the last decade has begun to bring us very consistent data at this level. In addition, studying the diet of individuals for a long time presents another problem, which is the constant variation in the eating habits of populations, even more so in a constantly changing world like the one we live in.
If we look at the guidelines for the treatment of diabetic patients from the American Diabetes Association (ADA) in conjunction with the European Association for the Study of Diabetes (EASD), published in December 2018, we can conclude that the term "Glycemic Index" appears only once, mixed with other characteristics of the diet or eating patterns that may be potentially beneficial, without referring to any numerical value or way to implement its use in practice. The more detailed version on dietary and nutritional intervention, from 2019, even mentions the existence of ambiguous data in the scientific literature, and concludes that the GI does not affect long-term glycemic control (more than 12 weeks). This still recent consensus from two of the most important scientific societies in the field of diabetes reflects well the lack of solid knowledge in this area.
But, as we mentioned, science doesn't stop, and work with large cohorts continues to offer invaluable data and conclusions, with a regularity that is sometimes difficult to absorb! This is the case with the study that justifies this text, published on February 24th of this year in the prestigious New England Journal of Medicine.
Headed by Professor David Jenkins, this work results from the analysis of a large international cohort, designated PURE (Prospective Urban Rural Epidemiology), involving more than 137,000 participants from twenty countries across five continents, and studied the impact of GI on cardiovascular disease.
Dividing the sample by ginning quintiles of the participants' meals, the association between higher ginning glycemic index (GI) levels and the occurrence of cardiovascular events, such as myocardial infarctions, strokes, or heart failure, was notable (and statistically significant). Furthermore, the increased risk was greater in individuals with a BMI > 25 kg/m2, representing a 38% increase in risk for the last quintile compared to the first quintile, or in individuals with pre-existing cardiovascular disease at the start of the study, representing a 51% increase in risk for the last quintile compared to the first quintile.
The authors conclude that those who consume diets with a lower GI have a lower risk of developing cardiovascular disease than those who consume a diet with a higher GI. While seemingly simple, this is a conclusion that was clearly missing and had already been previously suggested in the 2020 meta-analysis. However, we cannot forget that this type of cohort study does not constitute the strongest methodology, and the scientific community continues to await clinical trials in this area that will unequivocally establish the GI as an important methodology in meal planning and dietary guidelines.
We can then question whether something in our practice should change from now on based on this new data. We believe that, notwithstanding the aforementioned methodological limitations, these new studies end up reinforcing some notions that have been around for a while and that we should have already implemented: a smaller quantity of processed cereal products, less added sugars, and greater consumption of vegetables, legumes, and fruit. However, precise knowledge of the magnitude of this effect allows for a new perspective and renewed attention to the GI and certainly stimulates new research on the subject. How to plan meals and dietary days according to this index? How do the other components of the meal affect the GI, whether macronutrients (whose effect is already known) or other non-nutritional components? In addition to cardiovascular disease, what will be the impact of the GI on other pathologies, such as obesity, diabetes mellitus, or dyslipidemias?
These are questions that large epidemiological studies will certainly help to answer, thus allowing nutritionists to adopt a practice even more grounded in scientific evidence. Until then, let us always apply the principle of moderation, not rushing to alter our thinking and practice based on the latest scientific news. For now, the glycemic index (GI) seems to be resurfacing from some neglect, giving, in our view, renewed importance to the choice of carbohydrate source in the diet. Knowing the importance of the processing of cereal products in determining their GI, these results further reinforce the importance of choosing products with the least possible processing.
