Glycemic Index and Glycemic Load
Summary
- The glycemic index (GI) is a measure of the blood glucose-raising potential of the carbohydrate content of a food compared to a reference food (generally pure glucose). Carbohydrate-containing foods can be classified as high- (≥70), moderate- (56-69), or low-GI (≤55) relative to pure glucose (GI=100). (More information)
- Consumption of high-GI foods causes a sharp increase in postprandial blood glucose concentration that declines rapidly, whereas consumption of low-GI foods results in a lower blood glucose concentration that declines gradually. (More information)
- The glycemic load (GL) is obtained by multiplying the quality of carbohydrate in a given food (GI) by the amount of carbohydrate in a serving of that food. (More information)
- Prospective cohort studies found high-GI or -GL diets to be associated with a higher risk of adverse health outcomes, including type 2 diabetes mellitus and cardiovascular disease. (More information)
- Meta-analyses of observational studies have found little-to-no evidence of an association between high dietary GI and GL and risk of cancer. (More information)
- Lowering the GL of the diet may be an effective method to improve glycemic control in individuals with type 2 diabetes mellitus. This approach is not currently included in the overall strategy of diabetes management in the US. (More information)
- Several dietary intervention studies found that low-GI/GL diets were as effective as conventional, low-fat diets in reducing body weight. Both types of diets resulted in beneficial effects on metabolic markers associated with the risk of type 2 diabetes mellitus and cardiovascular disease. (More information)
- Lowering dietary GL can be achieved by increasing the consumption of whole grains, nuts, legumes, fruit, and non-starchy vegetables, and decreasing intakes of moderate- and high-GI foods like potatoes, white rice, white bread, and sugary foods. (More information)
Glycemic Index
Glycemic index of individual foods
In the past, carbohydrates were classified as simple or complex based on the number of simple sugars in the molecule. Carbohydrates composed of one or two simple sugars like fructose or sucrose (table sugar; a disaccharide composed of one molecule of glucose and one molecule of fructose) were labeled simple, while starchy foods were labeled complex because starch is composed of long chains of the simple sugar, glucose. Advice to eat less simple and more complex carbohydrates (i.e., polysaccharides) was based on the assumption that consuming starchy foods would lead to smaller increases in blood glucose than sugary foods (1). This assumption turned out to be too simplistic since the blood glucose (glycemic) response to complex carbohydrates has been found to vary considerably. The concept of glycemic index (GI) has thus been developed in order to rank dietary carbohydrates based on their overall effect on postprandial blood glucose concentration relative to a referent carbohydrate, generally pure glucose (2). The GI is meant to represent the relative quality of a carbohydrate-containing food. Foods containing carbohydrates that are easily digested, absorbed, and metabolized have a high GI (GI≥70 on the glucose scale), while low-GI foods (GI≤55 on the glucose scale) have slowly digestible carbohydrates that elicit a reduced postprandial glucose response. Intermediate-GI foods have a GI between 56 and 69 (3). The GI of selected carbohydrate-containing foods can be found in Table 1.
Measuring the glycemic index of foods
To determine the glycemic index (GI) of a food, healthy volunteers are typically given a test food that provides 50 grams (g) of carbohydrate and a control food (white, wheat bread or pure glucose) that provides the same amount of carbohydrate, on different days (4). Blood samples for the determination of glucose concentrations are taken prior to eating, and at regular intervals for a few hours after eating. The changes in blood glucose concentration over time are plotted as a curve. The GI is calculated as the incremental area under the glucose curve (iAUC) after the test food is eaten, divided by the corresponding iAUC after the control food (pure glucose) is eaten. The value is multiplied by 100 to represent a percentage of the control food (5):
GI = (iAUCtest food/iAUCglucose) x 100
For example, a boiled white potato has an average GI of 82 relative to glucose and 116 relative to white bread, which means that the blood glucose response to the carbohydrate in a baked potato is 82% of the blood glucose response to the same amount of carbohydrate in pure glucose and 116% of the blood glucose response to the same amount of carbohydrate in white bread. In contrast, cooked brown rice has an average GI of 50 relative to glucose and 69 relative to white bread. In the traditional system of classifying carbohydrates, both brown rice and potato would be classified as complex carbohydrates despite the difference in their effects on blood glucose concentrations.
While the GI should preferably be expressed relative to glucose, other reference foods (e.g., white bread) can be used for practical reasons as long as their preparation has been standardized and they have been calibrated against glucose (2). Additional recommendations have been suggested to improve the reliability of GI values for research, public health, and commercial application purposes (2, 6).
Physiological responses to high- versus low-glycemic index foods
By definition, the consumption of high-GI foods results in higher and more rapid increases in blood glucose concentrations than the consumption of low-GI foods. Rapid increases in blood glucose (resulting in hyperglycemia) are potent signals to the β-cells of the pancreas to increase insulin secretion (7). Over the next few hours, the increase in blood insulin concentration (hyperinsulinemia) induced by the consumption of high-GI foods may cause a sharp decrease in the concentration of glucose in blood (resulting in hypoglycemia). In contrast, the consumption of low-GI foods results in lower but more sustained increases in blood glucose and lower insulin demands on pancreatic β-cells (8).
Glycemic index of a mixed meal or diet
Many observational studies have examined the association between GI and risk of chronic disease, relying on published GI values of individual foods and using the following formula to calculate meal (or diet) GI (9):
Meal GI = [(GI x amount of available carbohydrate)Food A + (GI x amount of available carbohydrate)Food B +…]/ total amount of available carbohydrate
Yet, the use of published GI values of individual foods to estimate the average GI value of a meal or diet may be inappropriate because factors such as food variety, ripeness, processing, and cooking are known to modify GI values. In a study by Dodd et al., the estimation of meal GIs using published GI values of individual foods was overestimated by 22 to 50% compared to direct measures of meal GIs (9).
Besides the GI of individual foods, various food factors are known to influence the postprandial glucose and insulin responses to a carbohydrate-containing mixed diet. A recent cross-over, randomized trial in 14 subjects with type 2 diabetes mellitus examined the acute effects of four types of breakfasts with high- or low-GI and high- or low-fiber content on postprandial glucose concentrations. Plasma glucose was found to be significantly higher following consumption of a high-GI and low-fiber breakfast than following a low-GI and high-fiber breakfast. However, there was no significant difference in postprandial glycemic responses between high-GI and low-GI breakfasts of similar fiber content (10). In this study, meal GI values (derived from published data) failed to correctly predict postprandial glucose response, which appeared to be essentially influenced by the fiber content of meals. Since the amounts and types of carbohydrate, fat, protein, and other dietary factors in a mixed meal modify the glycemic impact of carbohydrate GI values, the GI of a mixed meal calculated using the above-mentioned formula is unlikely to accurately predict the postprandial glucose response to this meal (3). Moreover, the GI is a property of a given food carbohydrate such that it does not take into account individuals' characteristics like ethnicity, metabolic status, or eating habits (e.g., the degree to which we masticate) which might, to a limited extent, also influence the glycemic response to a given carbohydrate-containing meal (11-14).
Using direct measures of meal GIs in future trials — rather than estimates derived from GI tables — would increase the accuracy and predictive value of the GI method (2, 6). In addition, in a recent meta-analysis of 28 studies examining the effect of low- versus high-GI diets on serum lipids, Goff et al. indicated that the mean GI of low-GI diets varied from 21 to 57 across studies, while the mean GI of high-GI diets ranged from 51 to 75 (15). Therefore, a stricter use of GI cutoff values may also be warranted to provide more reliable information about carbohydrate-containing foods.
Glycemic Load
The glycemic index (GI) compares the potential of foods containing the same amount of carbohydrate to raise blood glucose. However, the amount of carbohydrate contained in a food serving also affects blood glucose concentrations and insulin responses. For example, the mean GI of watermelon is 76, which is as high as the GI of a doughnut (see Table 1). Yet, one serving of watermelon provides 11 g of available carbohydrate, while a medium doughnut provides 23 g of available carbohydrate.
The concept of glycemic load (GL) was developed by scientists to simultaneously describe the quality (GI) and quantity of carbohydrate in a food serving, meal, or diet. The GL of a single food is calculated by multiplying the GI by the amount of carbohydrate in grams (g) provided by a food serving and then dividing the total by 100 (4):
GLFood = (GIFood x amount (g) of available carbohydrateFood per serving)/100
For a typical serving of a food, GL would be considered high with GL≥20, intermediate with GL of 11-19, and low with GL≤10. Using the above-mentioned example, despite similar GIs, one serving of watermelon has a GL of 8, while a medium-sized doughnut has a GL of 17. Dietary GL is the sum of the GLs for all foods consumed in the diet.
It should be noted that while healthy food choices generally include low-GI foods, this is not always the case. For example, intermediate-to-high-GI foods like parsnip, watermelon, banana, and pineapple, have low-to-intermediate GLs (see Table 1).
Disease Prevention
Type 2 diabetes mellitus
The consumption of high-GI and -GL diets for several years might result in higher postprandial blood glucose concentration and excessive insulin secretion. This might contribute to the loss of the insulin-secreting function of pancreatic β-cells and lead to irreversible type 2 diabetes mellitus (16).
A US ecologic study of national data from 1909 to 1997 found that the increased consumption of refined carbohydrates in the form of corn syrup, coupled with the declining intake of dietary fiber, has paralleled the increased prevalence of type 2 diabetes (17). In addition, high-GI and -GL diets have been associated with an increased risk of type 2 diabetes in several large prospective cohort studies. A recent updated analysis of three large US cohorts indicated consumption of foods with the highest versus lowest GI was associated with a risk of developing type 2 diabetes that was increased by 44% in the Nurses' Health Study (NHS) I, 20% in the NHS II, and 30% in the Health Professionals Follow-up Study (HPFS). High-GL diets were associated with an increased risk of type 2 diabetes (+18%) only in the NHS I and in the pooled analysis of the three studies (+10%) (18). Additionally, the consumption of high-GI foods that are low in cereal fiber was associated with a 59% increase in diabetes risk compared to low-GI and high-cereal-fiber foods. High-GL and low-cereal-fiber diets were associated with a 47% increase in risk compared to low-GL and high-cereal-fiber diets. Moreover, obese participants who consumed foods with high-GI or -GL values had a risk of developing type 2 diabetes that was more than 10-fold greater than lean subjects consuming low-GI or -GL diets (18).
However, a number of prospective cohort studies have reported a lack of association between GI or GL and type 2 diabetes (19-24). The use of GI food classification tables based predominantly on Australian and American food products might be a source of GI value misassignment and partly explain null associations reported in many prospective studies of European and Asian cohorts.
Nevertheless, conclusions from several recent meta-analyses of prospective studies (including the above-mentioned studies) suggest that low-GI and -GL diets might have a modest but significant effect in the prevention of type 2 diabetes (18, 25, 26). Organizations like Diabetes UK (27) and the European Association for the Study of Diabetes (28) have included the use of diets of low GI/GL and high in dietary fiber and whole grains in their recommendations for diabetes prevention in high-risk individuals. The use of GI and GL is currently not implemented in US dietary guidelines (29).
Cardiovascular disease
Observational studies
Numerous observational studies have examined the relationship between dietary GI/GL and the incidence of cardiovascular events, especially coronary heart disease (CHD) and stroke. A meta-analysis of 14 prospective cohort studies (229,213 participants; mean follow-up of 11.5 years) found a 13% and 23% increased risk of cardiovascular disease (CVD) with high versus low dietary GI and GL, respectively (30). Three independent meta-analyses of prospective studies also reported that higher GI or GL was associated with increased risk of CHD in women but not in men (31-33). A recent analysis of the European Prospective Investigation into Cancer and Nutrition (EPIC) study in 20,275 Greek participants, followed for a median of 10.4 years, showed a significant increase in CHD incidence and mortality with high dietary GL specifically in those with high BMI (≥28 kg/m2) (34). This is in line with earlier findings in the Nurses' Health Study (NHS) showing that a high dietary GL was associated with a doubling of the risk of CHD over 10 years in women with higher (≥23 kg/m2) vs. lower BMI (35). A similar finding was reported in a cohort of middle-aged Dutch women followed for nine years (36).
Additionally, high dietary GL (but not GI) was associated with a 19% increased risk of stroke in pooled analyses of prospective cohort studies (32, 37). A meta-analysis of seven prospective studies (242,132 participants; 3,255 stroke cases) found that dietary GL was associated with an overall 23% increase in risk of stroke and a specific 35% increase in risk of ischemic stroke; GL was not found to be related to hemorrhagic stroke (38).
Overall, observational studies have found that higher glycemic load diets are associated with increased risk of cardiovascular disease, especially in women and in those with higher BMIs.
GI/GL and cardiometabolic markers
The GI/GL of carbohydrate foods may modify cardiometabolic markers associated with CVD risk. A meta-analysis of 27 randomized controlled trials (published between 1991 and 2008) examining the effect of low-GI diets on serum lipid profile reported a significant reduction in total and LDL-cholesterol independent of weight loss (15). Yet, further analysis suggested significant reductions in serum lipids only with the consumption of low-GI diets with high fiber content. In a three-month, randomized controlled study, an increase in the values of flow-mediated dilation (FMD) of the brachial artery, a surrogate marker of vascular health, was observed following the consumption of a low- versus high-GI hypocaloric diet in obese subjects (39).
High dietary GLs have been associated with increased concentrations of markers of systemic inflammation, such as C-reactive protein (CRP), interleukin-6, and tumor necrosis factor-α (TNF-α) (40, 41). In a small 12-week dietary intervention study, the consumption of a Mediterranean-style, low-GL diet (without caloric restriction) significantly reduced waist circumference, insulin resistance, systolic blood pressure, as well as plasma fasting insulin, triglycerides, LDL-cholesterol, and TNF-α in women with metabolic syndrome. A reduction in the expression of the gene coding for 3-hydroxy-3-methylglutaryl (HMG)-CoA reductase, the rate-limiting enzyme in cholesterol synthesis, in blood cells further confirmed an effect for the low-GI diet on cholesterol homeostasis (42). Well-controlled, long-term intervention studies are needed to confirm the potential cardiometabolic benefits of low GI/GL diets in people at risk for CVD.
Cancer
Evidence that high-GI or -GL diets are related to cancer is inconsistent. A recent meta-analysis of 32 case-control studies and 20 prospective cohort studies found modest and nonsignificant increased risks of hormone-related cancers (breast, prostate, ovarian, and