INTRODUCTION — The type and amount of carbohydrates consumed may have important implications for health. With widespread promotion of low-fat diets in the United States in the 1980s and 1990s, there was a reduction in the percentage of dietary calories from fat and an increase in carbohydrate intake. Despite these dietary changes, however, the prevalence of obesity and type 2 diabetes mellitus continued to rise [1]. Diets that emphasize low or no carbohydrates are becoming increasingly popular for weight loss.

In the scientific community, however, serious concern has been raised over dietary recommendations for high-carbohydrate, low-fat diets. High-carbohydrate diets reduce high-density lipoprotein (HDL) cholesterol and raise blood concentrations of glucose, insulin, triglycerides, and blood pressure, contributing to a metabolic profile called insulin resistance syndrome that is conducive to the development of coronary heart disease and type 2 diabetes mellitus [2]. (See "Metabolic syndrome (insulin resistance syndrome or syndrome X)".)

In addition, due to differences in chemical structure, fiber content, and degree of processing (refining and gelatinization), different forms of carbohydrates have varied biological functions and effects on health; these effects are discussed here. Overall recommendations for a healthy diet and dietary therapy for obesity are reviewed elsewhere. (See "Healthy diet in adults" and "Obesity in adults: Dietary therapy", section on 'Choosing a diet'.)

CLASSIFICATION OF CARBOHYDRATES — Carbohydrates are traditionally classified as simple starches (sugars, ie, mono- and disaccharides) or complex starches, (ie, polysaccharides) on the basis of the number of sugar molecules in their chemical structures. It had been assumed that complex carbohydrates cause smaller rises in blood glucose than simple carbohydrates. Dietary guidelines for individuals with diabetes mellitus that recommend the use of complex rather than simple carbohydrates to control blood glucose levels reflect this assumption [3,4].

A large body of data, however, contradicts this notion. In the mid-1970s, studies showed that starchy foods differed in their ability to induce plasma glucose and insulin responses [5-8]. These findings initiated a new era of research on the physiologic effects of carbohydrates. Collectively, these data demonstrate that digestion of carbohydrates, particularly starch, is not a rate-limiting event and is often quite rapid because of the presence of sufficient amylolytic capacity in the intestine [9]. As an example, postprandial glycemic responses to potato and white bread are similar to the response to pure glucose [10,11], indicating that complex carbohydrates may not differ from simple sugars in their effect on plasma glucose level. From a clinical and public health perspective, carbohydrates may therefore be classified by their physiologic functions (formally called "nutritional classification"), as proposed by an expert panel convened by the World Health Organization (WHO) [12].

Historically, the quality of dietary carbohydrates has been characterized by its fiber content, which refers to the components of the plant-cell wall that are not digestible [12]. As originally defined, dietary fiber is also highly correlated with numerous minerals, vitamins, and other biologically active constituents within a structurally intact cell wall that is not readily accessible to enzymatic reaction. Further, dissolving a carbohydrate in water has important effects on satiety. Thus, whether carbohydrates are consumed in solid or liquid form needs to be considered [13].

GLYCEMIC INDEX

Definition — The glycemic index (GI) is a measure of the relative impact of carbohydrate-containing foods on serum glucose. A particular food's GI is determined by evaluating the incremental rises of blood glucose after ingestion of a food that contains 50 g of available carbohydrate compared with the same amount of carbohydrate from a reference food, usually white bread or glucose [14].

The following values are generally applied for defining the GI of a particular food (using glucose as a reference); the GI ratio is unitless, and as such, must be considered in conjunction with the quantity of carbohydrate content (see 'Glycemic load' below):

●Low GI – 55 or less

●Medium GI – 56 to 69

●High GI – 70 or more

Determination — The GI of a food depends upon the rapidity of digestion and absorption of its dietary carbohydrates, which is largely determined by both its physical and chemical properties [15-18]. Typically, foods with a low degree of starch gelatinization (eg, pasta), and those containing a high level of viscous soluble fiber (eg, whole-grain barley, oats, and rye), have slower rates of digestion and lower GI values. In addition, cooling starch after cooking (such as refrigerating boiled potatoes or parboiling rice) can also lower the GI.

Another important influence on GI is the ratio of amylose (a linear-chain molecule of 300 to 600 glucose residues linked by alpha-1,4-glucosidic bonds) to amylopectin (a larger, highly branched polymer of 6000 glucose residues linked by both alpha-1,4- and alpha-1,6-glucosidic bonds). Foods with a higher amylose/amylopectin ratio (such as legumes and some varieties of rice) tend to have lower GI values, likely due to the compact structure of amylose, which physically slows enzymatic reactions. By contrast, the branched structure of amylopectin is more open to enzymatic attack and is thus easily digested.

Utility — Use of the GI has shown that many complex carbohydrates induce glycemic and insulinemic responses nearly as high as those induced by pure glucose, thereby further casting doubt on the usefulness of the "simple versus complex carbohydrate" classification system. (See 'Classification of carbohydrates' above.)

The utility of the GI concept has been debated since its introduction in the 1980s [19,20]. The debate originated from evaluating what kind of dietary carbohydrate is best for controlling postprandial glycemia in the management of individuals with diabetes [21]. Other concerns regarding the clinical usefulness of the GI have revolved around [3,19,22]:

●The usefulness of the GI when consuming mixed meals

●The lack of long-term studies

Mixed meals — In the early 1980s, three studies showed that when individual carbohydrate foods were taken as part of a mixed meal, differences in glycemic responses between foods no longer existed [22-24]. The authors postulated that these findings were due to the effects of fat and protein on glycemic responses. Although the methodology of these studies has been criticized [19], the results were used by the 1986 National Institutes of Health (NIH) consensus conference on diet and exercise in type 2 diabetes mellitus to reject the use of the GI [3].

Since then, many studies aimed at addressing these concerns have been conducted [2,25,26], and there are now abundant data to support the GI concept [27,28]. In particular, several studies have shown that although fat and protein affect the absolute glycemic response, they do not affect the relative differences between foods [15]. In other words, given the same amount of protein and fat in a meal, high glycemic foods still lead to a higher glycemic response than low glycemic foods. Studies using standardized methodology have indicated that the correlation between the GI of mixed meals and the average GI values of individual component foods ranges from 0.84 to 0.99 [15,28].

Long-term data — Controversy regarding the clinical usefulness of the GI continues, mainly due to the lack of long-term high quality randomized trials [21,29]. Most metabolic studies of GI are of very short duration (several meals or several days). Long-term studies are needed to assess the clinical usefulness of the GI. (See 'Glycemic index/glycemic load' below.)

GLYCEMIC LOAD — Another argument against the glycemic index (GI) concept is that it does not capture the entire glucose-raising potential of dietary carbohydrates, because the blood glucose response is influenced not only by the type but by the quantity of carbohydrate consumed. To address this concern, the concept of glycemic load (GL) was introduced [30]. Defined as the product of the GI value of a food and its carbohydrate content, GL incorporates both the quality and quantity of carbohydrate consumed and the interaction between them [12,31,32].

With the use of a reference standard carbohydrate source (ie, white bread or glucose), the GL of any carbohydrate-containing food quantifies the serum glucose-raising potential of that specific food; each unit of dietary GL represents the equivalent glycemic effect of 1 g of carbohydrate standardized to either white bread or glucose. Thus, the GL allows for a direct comparison of all carbohydrate-containing foods with the same common physiological response. In general, carbohydrate-rich foods with low fiber content have high GL values; examples include potatoes, refined cereal products, and many sugar-sweetened beverages. By contrast, intact whole grains, legumes, fruits, and vegetables with high fiber content provide low to very low GLs per serving.

Many low GI foods are not necessarily high in fiber (eg, white pasta, basmati rice, and dairy products). Some high-fiber, whole-meal bread and breakfast cereals (eg, raisin bran and whole-wheat flakes) are high in GI. This highlights the importance of quantifying actual GI values of food products through direct testing. Such testing has distinguished two types of available glucose from foods (most foods contain a mixture of starch and sugars), termed rapidly available glucose (RAG) and slowly available glucose (SAG) [33]. In studies that examined over 60 foods, RAG values correlated with both their direct glycemic response and published GI values, supporting the physiologic relevance of this categorization [33,34].

The following values are generally applied for defining the GL of a particular food (per serving, using glucose as a reference):

●Low GL – 10 or less

●Medium GL – 11 to 19

●High GL – 20 or more

The following values are generally applied for defining the GL per day:

●Low GL – Less than 80

●High GL – More than 120

WHOLE VERSUS REFINED GRAINS — Carbohydrates from grains may also be classified by whether the grain is whole or has been refined by removing all or part of the bran and germ. The majority of the fiber, vitamins, minerals, and phytochemicals reside in the bran and germ fractions, and most of these are removed with refining. The evidence for health benefits of dietary fiber is based on consumption of fiber in food and not the purified fiber that lacks all of these potentially beneficial components that accompany the fiber.

The US Food and Drug Administration (FDA) allows almost all fiber (purified and synthetic) to be included on "Nutrition Facts" food labels. However, this information can be misleading as many manufacturers are adding these synthetic fibers to foods that are otherwise unhealthy [35], marketing them as healthy foods. In addition, products are allowed to be marketed as whole grain if they are at least 51 percent whole grain (and thus contain as much as 49 percent refined grain).

Another important distinction is whether the whole grain is intact or has been milled into flour. The intact whole grain tends to act like a slow-release capsule, reducing the surface area for enzymatic digestion of the starch, while the milling process greatly increases the surface area, thus increasing the GI of the grain.

DIETARY CARBOHYDRATES AND WEIGHT

Dietary fiber and whole grains — Due to their high fiber and water content, whole-grain foods contain fewer calories per gram than the isovolumic amount of corresponding refined grain foods. Whole grains may protect against weight gain through several mechanisms involving effects on satiety, post-load glycemia and insulin responses, antioxidants, and minerals [36].

Several large, prospective studies have found that intake of dietary fiber and whole grains, compared with refined grains, is inversely associated with body weight and fat distribution [37-39]. As examples:

●In the Coronary Artery Risk Development in Young Adults Study (CARDIA), intake of whole grains was inversely associated with body mass index (BMI) at all levels of fat intake and predicted insulin levels, 10-year weight gain, and other cardiovascular disease risk factors including blood pressure [38]. Participants in the highest category of whole-grain intake (>21 g/2000 kcal) gained approximately 3.6 kg (8 lb) less in weight than those in the lowest category of intake (<12 g/2000 kcal).

●In the Nurses' Health Study (NHS) of middle-aged females, participants with the greatest increase in intake of whole grain gained an average of 1.52 kg less than those with the smallest increase in intake, independent of body weight at baseline, age, and changes in covariate status [37]. Those in the highest quintile of whole-grain intake had a 49 percent lower risk of major weight gain than those in the highest quintile of refined-grain intake (odds ratio [OR] 0.51, 95% CI 0.39-0.67).

●A prospective investigation of dietary patterns among 120,877 people apparently free of chronic diseases reported that increasing whole-grain intake by one serving/day was associated with a modest 0.37 pound lower weight gain over a four-year period [40].

Furthermore, several systematic reviews and meta-analysis of prospective cohort studies and trials confirm the beneficial effects of greater dietary fiber and whole-grain intake on weight control in adults [41,42].

Glycemic index/glycemic load — The manipulation of dietary glycemic index (GI) or glycemic load (GL) has emerged as an important approach to weight control. Since the duration of post-meal satiety is generally related to postprandial glycemia, low GI foods have been hypothesized to reduce hunger signals and delay the onset of the next meal. (See "Obesity in adults: Dietary therapy", section on 'Low-carbohydrate diets'.)

This concept was examined in a randomized trial that compared the hormonal and metabolic changes associated with isocaloric meals differing only in GI and GL [43]. Blood glucose and fatty acid concentrations were significantly lower, while counterregulatory hormones were elevated in high GI diets three to five hours after ingestion.

Similarly, a randomized trial compared the effects of a low GL diet (total daily GL 82) and a low fat diet (total daily GL 205) on resting energy expenditure (REE) and coronary heart disease (CHD) risk factors in 39 young adults during a diet designed to provide 60 percent of energy requirements [44]. After subjects had lost 10 percent of their baseline body weight, REE had decreased less in those on a low GL diet (-96 versus -176 kcal/day), and hunger was also less on the low GL diet. The diets had similar effects on high-density lipoprotein (HDL) and low-density lipoprotein (LDL) cholesterol levels, but those on the low GL diet had greater decreases in triglycerides (-3.5 versus +16.2 percent), in C-reactive protein (-47.7 versus -5.1 percent), and on a measure of insulin resistance (-33.9 versus -15.8 percent). There was also a trend toward a larger decrease in systolic and diastolic blood pressure on the low GL diet. These results suggest that a low GL diet can reduce some of the physiologic adaptations to a low-calorie diet (decreased REE, hunger) that interfere with sustained weight loss, while achieving an improved CHD risk profile.

Other trials also suggest weight loss and CHD risk benefits with low GI diets [45-47]. In a meta-analysis of six trials (202 participants) comparing a low GI diet with a higher GI diet or a conventional, calorie-restricted, reduced-fat diet, short-term weight loss (weighted mean difference 1.1 kg) and reduction in LDL cholesterol (-0.24 mmol/L) were significantly greater in the low GI diet group [48]. There were no trials lasting more than six months, and therefore the long-term metabolic effects of a low GI diet are unknown.

Baseline insulin secretion may predict weight loss with low GL diets. This was illustrated in a randomized trial of obese young adults assigned to a low GL or low-fat diet after baseline assessment of insulin secretion (serum insulin concentration measured 30 minutes after a 75 g oral glucose load) [49]. At 18 months of follow-up, patients with higher insulin concentrations lost more weight and body fat with the low GL diet than with the low fat diet. In contrast, in patients with lower insulin secretion, both diets produced similar results.

Very low carbohydrate diets — The definition of a "very low carbohydrate diet" differs between practitioners but in general refers to any diet that contains from 0 to <60 grams of carbohydrates, or approximately 0 to 10 percent of total dietary energy derived from carbohydrates.

Trials of very low carbohydrate diets demonstrate that they are more effective for short-term weight loss than low-fat diets. A more detailed review of the impact of very low carbohydrate diets on weight loss are reviewed in detail elsewhere. (See "Obesity in adults: Dietary therapy", section on 'Low-carbohydrate diets' and "Obesity in adults: Dietary therapy", section on 'Choosing a diet'.)

DIETARY CARBOHYDRATES AND DISEASE — Several large prospective cohort studies demonstrate strong inverse associations between dietary fiber and whole-grain intake and the risk of cardiovascular disease, with an estimated daily intake of approximately 30 grams associated with 20 to 40 percent reduction in risk [41,50]. Although studies demonstrate that an increased intake of dietary fiber reduces plasma cholesterol and risk of metabolic syndrome, high-quality data demonstrating that manipulation of dietary fiber, GI, or GL will prevent or delay the development of cardiovascular events, diabetes incidence, cancer, or other major disease outcomes are lacking.

Diabetes and metabolic syndrome — In prospective studies that have controlled for other risk factors, higher intakes of total or cereal fiber and whole grains have been associated with lower risks of type 2 diabetes [36,37,42].

In multiple meta-analyses of prospective cohort studies, a high GI or high GL diet was associated with an increased risk of type 2 diabetes [51-53]. In one meta-analysis, there was a dose-response relationship between GL and the risk of type 2 diabetes, in which dietary GL was associated with greater risk of type 2 diabetes when intake levels were higher than 95 g/2000 kcal [51]. The dose-dependent relationship was stronger in females than males. In another meta-analysis, however, the associations between GI and GL to the risk of type 2 diabetes were less impressive [42]. The relationship between dietary patterns and risk of type 2 diabetes is reviewed in more detail separately. (See "Type 2 diabetes mellitus: Prevalence and risk factors", section on 'Dietary patterns'.)

Resistance to insulin-mediated glucose transport is a fundamental early defect in the pathogenesis of type 2 diabetes [54] (see "Pathogenesis of type 2 diabetes mellitus"). Among individuals with impaired insulin action, glucose homeostasis can be maintained through increased insulin production. In some individuals, hyperglycemia and/or compensatory hyperinsulinemia (increased insulin production) make the pancreatic beta cells eventually lose their responsiveness to glucose, ultimately leading to the development of clinical diabetes.

While mechanisms for the so-called insulin resistance-pancreatic exhaustion phenomenon remain incompletely understood, it has been demonstrated in animal models that postprandial hyperglycemia and glucose toxicity directly contribute to insulin resistance and defects in insulin secretion [55,56]. Among individuals with diabetes, an elevated plasma glucose concentration causes defects in insulin secretion and may lead to insulin resistance [57]. Thus, dietary factors that increase plasma glucose and insulin responses may aggravate the effects of insulin resistance that, in turn, may increase the risk of type 2 diabetes. Conversely, dietary factors that minimize both glycemic and insulinemic responses may protect against the development of type 2 diabetes [26].

Animal studies have demonstrated adverse effects of diets very high in fructose or sucrose on insulin sensitivity and hypertriglyceridemia [58]. Similarly, a study in which partially pancreatectomized rats were given diets with identical nutrients except for the type of starch (either high or low GI) found that rats given the high GI food had almost twice the body fat of the rats given low GI food after nine weeks [59]. In addition, the high GI group had greater increases over time in blood glucose and plasma insulin after oral glucose and higher plasma triglyceride concentrations.

Because of limited data in humans, however, controversies remain regarding the role of sugars in the development of insulin resistance and type 2 diabetes [1,60]. Similarly, few studies have examined the relation between starch types and insulin resistance in humans. Partly because of the difficulties involved in assessing whole-body insulin sensitivity, studies of dietary GI and GL in humans have focused primarily on indicators of insulin resistance, including levels of postprandial glucose, insulin, C-peptide, and fasting triglycerides. Most studies show reduced glycemic and insulin responses to a low GI meal as compared with a high GI and GL meal. (See "Type 2 diabetes mellitus: Prevalence and risk factors", section on 'Dietary patterns'.)

In addition, reductions in plasma levels of glucose, insulin, or C-peptide, and improvement in insulin sensitivity have been shown among patients with diabetes, coronary heart disease (CHD), or postmenopausal females who switched from a high GI to a low GI diet [25,26,61-63]. In meta-analyses of trials comparing low and high GI diets in individuals with either type 1 or type 2 diabetes, low GI diets reduced glycated hemoglobin (A1C) by 0.4 to 0.5 percentage points [64,65]. This compares favorably to the carbohydrate exchange concept in terms of glycemic control among patients with diabetes. (See "Nutritional considerations in type 2 diabetes mellitus", section on 'Nutritional content'.)

Although the reduction in A1C is modest, it is comparable with the changes in A1C typically achieved with the drug acarbose [66,67], which delays carbohydrate digestion through inhibiting glucosidase in the small intestine. Thus, dietary approaches to the prevention and control of diabetes should thus be given at least as much consideration as drug therapies since a low GI diet may improve glycemic control to the same extent as some pharmacologic agents. (See "Alpha-glucosidase inhibitors for treatment of diabetes mellitus".)

The results of metabolic as well as prospective cohort studies suggest that long-term consumption of low GI carbohydrates should reduce the risk of type 2 diabetes. Indirect evidence using glucosidase inhibitors (which reduce postprandial glycemia by blocking the digestion of starch, thus converting a high GI food to a low GI food without changes in fiber or micronutrient intake) to prevent type 2 diabetes mellitus and CHD support this postulate [68]. (See "Prevention of type 2 diabetes mellitus".)

In the National Women's Health Initiative (WHI) including over 11,000 postmenopausal females, low circulating levels of sex hormone-binding globulin (associated with type 2 diabetes [69,70], cardiovascular disease, and hormone-dependent cancers) were found to be associated with high dietary GL and GI intake but not total carbohydrates [71]. In a subsequent review of a meta-analysis, consideration of mechanistic, clinical, and epidemiologic evidence supports a causal relation between high dietary GI and GL and risk of type 2 diabetes independent from of the effects of dietary fiber [72].

Cardiovascular disease — In large prospective cohort studies, higher intake of dietary fiber, cereal fiber, and whole grains have been associated with lower risk of CHD [42], although the specific contributions of fiber and the micronutrients and phytochemicals associated with the fiber in these foods are difficult to disentangle. In addition, most, but not all cohort studies show that dietary GI and GL are associated with an increased risk of CHD, major cardiovascular events, and cardiovascular mortality [32,73-78]. As examples:

●In a multinational prospective study including approximately 137,000 participants followed for 9.5 years, comparing the highest and lowest GI quintiles, a high GI diet was associated with increased risk of major cardiovascular events and cardiovascular mortality; the increased risk was identified among subjects with and without cardiovascular disease at baseline (hazard ratio [HR] 1.51, 95% CI 1.25-1.82; and HR 1.21, 95% CI 1.11-1.34, respectively) [78].

●In a meta-analysis including 10 prospective studies and 240,000 participants, females in the highest compared with lowest GI quantile had an increase in CHD risk (relative risk [RR] 1.26, 95% CI 1.12-1.41); no increased risk was noted for males [79]. Similar increased risk was noted for females in the highest compared with lowest GL quantile (RR 1.55, 95% CI 1.18 - 2.03), but not for males.

●In the Nurses' Health Study (NHS) including 82,800 healthy females followed for 20 years, high GL was associated with an increased CHD risk (RR 1.90, 95% CI 1.15-3.15) [75]. A significant interaction with body mass index (BMI) was observed; there was little association between dietary GL and CHD risk among the leanest females (BMI less than 23 kg/m²), but a strong positive association was observed among heavier individuals [32]. A separate analysis including almost 79,000 subjects found a positive association between total carbohydrate intake with the risk of hemorrhagic but not ischemic stroke (RR 3.84, 95% CI 1.27-12.66), and only among those with a BMI of ≥25 kg/m² [80].

●In another large meta-analysis including observational studies and randomized trials, GI and GL were not clearly associated with risk of CHD [42]. However, methodologic issues, including the heterogeneity of populations studied and dietary assessment methods, may limit the study’s conclusions. [41,52,73].

In some [81,82], but not all [83], short-term trials, low GI/GL diets appear to have beneficial effects on CHD risk factors.

●In a meta-analysis of randomized trials, low compared with high GI diets reduced total and low-density lipoprotein (LDL) cholesterol independently of weight loss, while no effects on high-density lipoprotein (HDL) cholesterol or triglycerides were observed [84].

●In a subsequent randomized, crossover-controlled trial designed to evaluate the effects of different GI diets on cardiovascular disease risk factors, 163 adults who were overweight were randomly assigned to one of four healthy diets with high or low GI and high or low carbohydrate content [83]. Diets with low compared with high GI did not improve biomarkers of cardiovascular disease risk (insulin sensitivity, lipid levels, systolic blood pressure). The higher GI diet did increase post prandial glucose levels [85].

●Beneficial effects on CHD risk factors were observed in several randomized trials examining effects of Mediterranean and low GL diets on lipids [86], as well as the components of the metabolic syndrome, including waist circumference, systolic and diastolic blood pressure, and triglycerides [87].

Although the observational data support an association between GI and GL and CHD, there are no direct clinical trial data to show whether manipulation of dietary GI or GL will prevent or delay the development of CHD in individuals with or without diabetes. However, in a randomized trial including adults with impaired glucose tolerance, treatment with acarbose (which delays carbohydrate digestion) reduced the risk for cardiovascular events and hypertension (HR 0.47, 95% CI 0.24-0.90; and HR 0.62, 95% CI 0.45-0.86, respectively), suggesting adverse effects with a high GI and GL diet [88].

Cancer — Hyperinsulinemia and/or insulin resistance may influence carcinogenesis. In the mid-1990s, it was noted that dietary and lifestyle risk factors for developing insulin resistance, such as physical inactivity, obesity, and positive energy balance, were also risk factors for developing colorectal cancer [89,90]. A high GI and low-fiber diet may lead to (or interact with) the state of insulin resistance such as hyperinsulinemia, hyperglycemia, and hypertriglyceridemia in influencing the carcinogenesis of colorectal cancer [91,92].

There are several case-control and prospective cohort studies that have examined the direct association between dietary GI and GL with colorectal cancer risk, and findings are inconsistent [93-100]. As examples:

●The prospective Women's Health Study found positive associations between dietary GL and overall GI and colorectal cancer incidence in this cohort of 39,876 females ≥45 years of age at baseline and followed for nearly eight years [96].

●A large, prospective study including 73,061 Chinese females in the Shanghai Women's Health Study reported no significant association of overall GI and GL to colorectal cancer risk (multivariable hazard ratio for the highest versus lowest quintile of dietary GL 0.94, 95% CI 0.71-1.24) [98].

●Similarly, little relationship was seen between GI or GL and risk of colorectal cancer in other United States cohorts [99] or among Swedish females [100].

The relations of GI and GL with the risk of other cancers have not been extensively examined in prospective studies. One prospective study linked dietary GL to increased risk of pancreatic cancer [101], while two others failed to show any overall association with breast cancer risk [102,103]. There are few data examining the possible link to prostate cancer, but in a study of 52,000 males, there was little relation between intakes of GI, GL, fiber, and whole grain and risk of this cancer [104].

DIETARY SOURCES OF CARBOHYDRATES — Prevalence estimates from a 2015 to 2016 National Health and Nutrition Examination Survey (NHANES) indicate that consumption of vegetables is below the recommended levels in 90 percent of the United States population, and the intake of whole grains was below the recommended amount in 98 percent [105-108]. To some extent, the public health campaign to lower fat consumption in the United States has been translated by food manufacturers and consumers into a potentially harmful set of food choices. Instead of replacing high-fat foods with fruits, vegetables, legumes, and whole-grain foods, consumers have often increased their consumption of low-fat or "fat-free" varieties of naturally high-fat foods, such as fat-free snack foods. The result is often an increase in diets with high glycemic load (GL), refined carbohydrates, and sugar.

The prevailing dietary pattern of the United States in the late 20^th and early 21^st century mainly consists of white bread, white rice, potatoes, sugar-sweetened beverages, and sugar-laden snacks. The United States food supply has become reliant on highly refined carbohydrates as significant sources of energy since these carbohydrate-dense foods are economical and easily consumed in fast-food restaurants or otherwise highly processed and prepackaged. The quality of these carbohydrates is considerably different from those consumed before the beginning of the 20^th century since the refining process has changed their compositions. As an example, processing whole grains into white flour actually increases the caloric density by 10 percent, reduces the amount of dietary fiber by 80 percent, and reduces the amount of dietary protein by almost 30 percent, leaving a dietary substance that is nearly pure starchy carbohydrate with far fewer nutrients [1].

The table lists the glycemic indices (GIs) for some major carbohydrate-contributing foods consumed by participants in the Nurses' Health Study (NHS), the GL values per serving, and their percent contribution to overall dietary GL (table 1). There are now a variety of low GI foods available to implement a low GI diet [27]. Using a standardized in vivo testing methodology endorsed by the World Health Organization (WHO) [28], researchers around the world continue to identify low GI foods. A useful resource is the University of Sydney’s glycemic index website. This site allows the user to enter many different foods and determine both the GI and GL.

Fructose — In the United States, high-fructose corn syrup sweeteners were not commercially available until 1960s but now comprise more than 20 percent of total daily carbohydrate intake and 10 percent of daily total energy intake, representing an increase of more than 2100 percent. These sweeteners have surpassed sucrose as the leading sweetener in the United States food industry, especially for soft drinks, accounting for much of the rebound increase in carbohydrate consumption after the mid-1960s, largely replacing the losses due to whole grains [1]. An ecological assessment of national data between 1909 and 1997 found that intake of corn syrup increased by 2100 percent and fiber intake decreased by 40 percent, paralleling the upward trend of obesity and diabetes prevalence in the United States [1]. (See "Type 2 diabetes mellitus: Prevalence and risk factors", section on 'Sugar-sweetened beverages' and "Obesity in adults: Etiologies and risk factors", section on 'Dietary factors'.)

The dominant type of commercially available high-fructose corn syrup contains 55 percent fructose and 45 percent glucose, a chemical composition similar to sucrose (50/50 fructose and glucose). In theory, enzyme-catalyzed isomerization of dextrose to fructose (in making high-fructose corn syrup) may be physiologically different from that of sucrose (which is not broken down until it reaches the gut). Available evidence indicates that high intake of sugar-sweetened foods in general contribute to weight gain due to their relatively high-energy density [109], GL [110], and palatability [111] in comparison to high-fiber and low GI foods. Moreover, these sugar-sweetened foods are frequently consumed in liquid rather than solid forms [112]. Although the specific metabolic effects of fructose on weight gain and insulin resistance have been reported, human data are mixed [113-116]. Further, there is no good evidence that replacing high-fructose corn sweetener with sucrose will have any health benefits.

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Basics topics (see "Patient education: Counting carbs if you do not use insulin (The Basics)")

SUMMARY AND RECOMMENDATIONS

●Carbohydrate types – Due to differences in chemical structure, fiber content, and degree of processing, different forms of carbohydrates have varied biological functions and effects on health. There is no clear evidence that the percent of dietary energy from total carbohydrates has an important effect on health outcomes, although the nature of the carbohydrate source may differentially affect postprandial glycemia, metabolic response, bodyweight, and major health outcomes. (See 'Introduction' above and 'Classification of carbohydrates' above.)

●Glycemic index – Glycemic index (GI) is an in vivo measure of the relative impact of carbohydrate-containing foods on blood glucose. The following values are generally applied for defining the GI of a particular food (table 1) (see 'Glycemic index' above):

•Low GI – 55 or less

•Medium GI – 56 to 69

•High GI – 70 or more

●Glycemic load – Glycemic load (GL) is the product of the GI value of a food and its carbohydrate content. The concept of the GL was developed in response to concerns that the GI cannot capture the entire glucose-raising potential of dietary carbohydrates, because the blood glucose response is influenced by not only the nature but also the quantity of the carbohydrate consumed. The following values are generally applied for defining the GL of a particular food (table 1) (see 'Glycemic load' above):

•Low GL – 10 or less

•Medium GL – 11 to 19

•High GL – 20 or more

The following values are generally applied for defining the GL per day:

•Low GL – Less than 80

•High GL – More than 120

●Weight benefits of high-fiber, whole grain foods – Due to their high fiber and water content, whole-grain foods contain fewer calories per gram than the isovolumic amount of corresponding refined grain foods. Whole grains may protect against weight gain through several mechanisms involving effects on satiety, post-load glycemia, and insulin responses. (See 'Dietary carbohydrates and weight' above.)

●Dietary fiber and disease – Observational data demonstrate strong inverse associations between dietary fiber and whole-grain intake with the risk of cardiovascular disease. Although studies demonstrate that an increased intake of dietary fiber reduces plasma cholesterol and risk of metabolic syndrome, high-quality data demonstrating that manipulation of dietary fiber, GI, or GL will prevent or delay the development of cardiovascular events, diabetes incidence, cancer, or other major disease outcomes are lacking. (See 'Dietary carbohydrates and disease' above.)