90th Anniversary Commentary: Consumption of Sweetened Beverages Predicts the Occurrence of Type 2 Diabetes

Carbohydrate and sugar are recurring themes in the history of diet and disease. After the agricultural revolution 10,000 y ago, nutrition changed from a hunter-gatherer way of life to one in which food could be grown in one locale, avoiding the challenges of continuing to move to find food. The domestication of animals and the development of many grains and other crops allowed civilizations to develop and reduced most effects of seasonality on food supplies for many countries (1, 2). The agricultural revolution also led to the growth of cities and, with it, a group of infectious diseases that dominated the health of people for thousands of years and led to the continued development of many new caloric foods and beverages over the last 10,000–12,000 y (3).

With the advent of the industrial revolution several hundred years ago, the techniques for growing and processing foods changed again, with a more varied food supply being one consequence of industrial techniques. After World War 2, as infectious diseases were claiming fewer lives, the focus on how the foods we ate could influence the progress of noncommunicable diseases such as diabetes, heart disease, cancer, and obesity came to be the focus of nutrition and public health.

The concept that sugar consumption might be related to the onset of diabetes was suggested more than a century ago by Sir Richard Havelock Charles (4), who noted its relation to the increasing incidence of diabetes among natives of Calcutta, India. This idea and the concept that diabetes was the result of “overnutrition” or obesity, proposed by Elliott Joslin the eminent diabetologist in the early 20th century (5), became competing concepts. The idea that sugar intake is related to the high incidence of diabetes and obesity has been a recurrent theme of Indian medicine as Gulati and Misra noted (6). Similarly, John Yudkin (7, 8) pushed this idea in the 1950s and 1960s. But it is only in the past several decades that many scientists have studied sugar, it components, its role in our metabolism, and its impact on human health in great detail.

Because obesity has increased rapidly in the past 40 y, diabetes has followed in its footsteps. Like other noncommunicable diseases, nutritional factors, in addition to obesity, play an important role. One of these nutrients is sugar. The diet contains many different sugars, including the disaccharides, such as sucrose (usual form of table sugar), lactose, and maltose, as well as the monosaccharides glucose and fructose. The article by Montonen et al. (9) in this Journal examined a cohort of Finnish adults for the relation of the intake of this broad group of sugars to the development of diabetes.

As background, we must remember major breakthroughs in science that led us to understand that the beverages we drink may not affect the quantity of food we eat (i.e., the energy in beverages may be less visible than in other foods). Led by Mattes and Rolls (10–12), a literature grew to show that caloric beverages did not adequately reduce the energy of the food we consumed so that energy balance was maintained (13). This was followed by a large literature that explored the health effects of consuming sugar-sweetened beverages. Subsequently, the role of fructose, particularly the major source of fructose in beverages (as high-fructose corn syrup or sucrose), was hypothesized to be a major cause of disease (14). Additional studies showed that fructose in any form has detrimental effects on satiety and cardiometabolic function (15–18).

It is in this context of exploring the varying roles of different sugars that the study by Montonen et al. (9) raised key issues. As Montonen et al. noted, fructose has a lower glycemic index than glucose—that is, it does not increase plasma glucose very much nor does it stimulate insulin release. The reason that fructose has so little effect on the glycemic index is that as fructose enters the hepatic circulation from the intestines it is almost completely cleared by the liver. Moreover, its metabolic route in the liver enhances lipid formation. It is for this reason that fructose leads to abnormal concentrations of lipids. Glucose, on the other hand, is only partly cleared by the liver, and in the liver the metabolic pathway is regulated by the need for energy or ATP. In clinical studies of beverages comparing fructose with other sugars, fructose produces changes that are similar to the metabolic syndrome, whereas beverages without fructose have little or no effect (19, 20). A very large literature has emerged on fructose (21–27).

In contrast to the literature on sugar-sweetened beverages, studies focusing on sugar consumption per se are fewer in number and sometimes contradictory. For example, 2 large recent prospective cohort studies came to different conclusions concerning the association between added-sugar consumption and cardiovascular disease mortality (28, 29). One major review on sugar did find adverse effects of sugar in food (30, 31).

Thus, the article by Montonen et al. (9) is of particular importance. These authors reported that 2 sugars, glucose and fructose, which are extremely highly correlated (r = 0.99), were associated with an increased risk of type 2 diabetes in a cohort of Finnish adults who were aged 40–60 y when examined between 1967 and 1972. This article reported on those individuals who were free of diabetes. The sample included 4284 individuals for whom all measurements were available. There were 177 incident cases of diabetes that developed during the 12 y of follow-up. In contrast to the highly significant relation of the monosaccharides, glucose and fructose, none of the disaccharides, including sucrose, lactose, or maltose, showed any significant relation to the development of diabetes.

The study by Montonen et al. (9) provides clear support for the relation of fructose to the incidence of diabetes. However, it is contrary to the proposition put forward by Khan and Sievenpiper (32) that “an association between fructose-containing sugars and cardiometabolic risk including weight gain, cardiovascular disease outcomes and diabetes only [occurs] when restricted to sugar-sweetened beverages and for sugars from other sources.” It is similarly at odds with the statement by Rippe and Angelopoulos (33), who stated “singling out added sugars as unique culprits for metabolically based diseases such as obesity, diabetes, and cardiovascular disease appears inconsistent with modern high-quality evidence and is very unlikely to yield health benefits.” If Montonen et al. are correct, targeting fructose intake might indeed be productive.

Glucose is a key metabolic product in the body, and it is tightly regulated by varying the secretion of insulin. When this regulation breaks down, diabetes develops. No studies, to our knowledge, have implicated this molecule in the onset of cardiometabolic diseases such as diabetes or obesity. We wonder whether it is the high association (r = 0.99) with fructose that is being detected.

Given the strong relation between sugar-sweetened beverages using either sucrose or high-fructose-corn syrup and the development of obesity and diabetes in other studies (19), it is unclear why sucrose did not have a relation to the incidence of diabetes in the study by Montonen et al. (9). Could it be the smaller number of cases and thus of incident diabetes? This is certainly plausible. Of the 11 studies included in the meta-analysis by Malik et al. (34), that by Montonen et al. (9) was the smallest. In contrast to their 2360 individuals, 6 of the other studies had >10,000 individuals. Could it be the adjustment for energy intake? Among the 11 studies reviewed by Malik et al. (34) only 3 adjusted for energy intake (including Montonen et al.), a concern that may represent another potential explanation for their different results. Because sucrose is a 50–50 dimer of glucose and fructose that can be split into the monosaccharides before ingestion or by the enzyme sucrose in the intestine after ingestion, the study by Montonen et al. probably represents the exception to the rule that increased sugar intake increases the risk of diabetes.

Overall, this study fits into the large recent history of added sugars and our diet. We have found increasing amounts of sugar added not only to beverages but also to foods (35). This has led many nations and the WHO to lead a push to reduce sugar in our diets, particularly from beverages (36–38), and to tax sugary beverages in an increasing number of countries and cities.

Notes

Supported by NIH grants R01DK108148 and R01DK098072. BMP reports no food or beverage industry funding.

Author disclosures: GAB and BMP, no conflicts of interest.