Testing the glycaemic index of foods: in vivo, not in vitro

The Glycemic Index and Human Health with an Emphasis on Potatoes

Diabetes and obesity are associated with the excessive intake of high-glycemic index (GI) carbohydrates, increased glycemic load (GL) foods, and inactive lifestyles. Carbohydrate-rich diets affect blood glucose levels. GI is an indicator of the impact of a specific food on blood glucose, while GL represents the quantity and quality of carbohydrates in the overall diet and their interactions. There are in vitro and in vivo methods for estimating GI and GL. These values are useful human health markers for conditions such as diabetes, obesity, and pregnancy. Potato is a major starchy vegetable, which is consumed widely and is the fourth most important crop globally. However, the GI of diets rich in starchy vegetables such as potatoes has not been studied in detail. The GI values in potatoes are affected by external and internal factors, such as methods of cooking, methods of processing, resistant starches, cultivation methods, mixed meals and food additions, and hormone levels. This review summarizes how these factors affect the GI and GL associated with diets containing potatoes. Understanding the impacts of these factors will contribute to the development of new and improved potato varieties with low GI values. The consumption of low-GI foods will help to combat obesity. The development of low-GI potatoes may contribute to the development of meal plans for individuals living with diabetes and obesity.

A Review of In Vitro Methods for Measuring the Glycemic Index of Single Foods: Understanding the Interaction of Mass Transfer and Reaction Engineering by Dimensional Analysis

The Glycemic Index (GI) has been described by an official method ISO (International Organization for Standardization) 26642:2010 for labeling purposes. The development of in vitro methods for GI measurement has faced significant challenges. Mass transfer and reaction engineering theory may assist in providing a quantitative understanding of in vitro starch digestion and glycemic response from an engineering point of view. We suggest that in vitro GI measurements should consider the mouth and the stomach in terms of fluid mechanics, mass transfer, length scale changes, and food-solvent reactions, and might consider a significant role for the intestine as an absorption system for the glucose that is generated before the intestine. Applying mass transfer and reaction engineering theory may be useful to understand quantitative studies of in vitro GI measurements. The relative importance of reactions and mass-transfer has been estimated from literature measurements through estimating the Damköhler numbers (Da), and the values estimated of this dimensionless group (0.04–2.9) suggest that both mass transfer and chemical reaction are important aspects to consider.

Structural breakdown of starch-based foods during gastric digestion and its link to glycemic response: In vivo and in vitro considerations

The digestion of starch-based foods in the small intestine as well as factors affecting their digestibility have been previously investigated and reviewed in detail. Starch digestibility has been studied both in vivo and in vitro, with increasing interest in the use of in vitro models. Although previous in vivo studies have indicated the effect of mastication and gastric digestion on the digestibility of solid starch-based foods, the physical breakdown of starch-based foods prior to small intestinal digestion is often less considered. Moreover, gastric digestion has received little attention in the attempt to understand the digestion of solid starch-based foods in the digestive tract. In this review, the physical breakdown of starch-based foods in the mouth and stomach, the quantification of these breakdown processes, and their links to physiological outcomes, such as gastric emptying and glycemic response, are discussed. In addition, the physical breakdown aspects related to gastric digestion that need to be considered when developing in vitro-in vivo correlation in starch digestion studies are discussed. The discussion demonstrates that physical breakdown prior to small intestinal digestion, especially during gastric digestion, should not be neglected in understanding the digestion of solid starch-based foods.

Bread making with sourdough and intact cereal and legume grains - effect on glycaemic index and glycaemic load

The concept of glycaemic index (GI) has led to efforts to develop low-GI foods. Bread contributes around one-quarter of carbohydrate intake in the Swedish diet. In this study, we sought to develop low-GI bread prototypes and examined the effects of bread making on content of total dietary fibre (TDF) and resistant starch (RS). Five bread prototypes were made in a commercial bakery, using sourdough fermentation and intact cereal and legume kernels. Predicted (p-GI) and in vivo GI values were determined, and TDF and RS were quantified. The p-GI value of the five prototypes was between 56 and 68. The confirmed in vivo GI value was 65 and 67 for two of the breads. The TDF content (≥17%) was not affected by bread making, but RS content was increased by three-fold. All breads were categorised as medium-GI, but with low glycaemic load (GL).

Glycemic index of pulses and pulse-based products: a review

Pulses are a major source for plant-based proteins, with over 173 countries producing and exporting over 50 million tons annually. Pulses provide many of the essential nutrients and vitamins for a balanced and healthy diet, hence are health beneficial. Pulses have been known to lower glycemic index (GI), as they elicit lower post prandial glycemic responses, and can prevent insulin resistance, Type 2 diabetes and associated complications. This study reviews the GI values (determined by in vivo methodology) reported in 48 articles during the year 1992-2018 for various pulse type preparations consumed by humans. The GI ranges (glucose and bread as a reference respectively) for each pulse type were: broad bean (40 ± 5 to 94 ± 4, 75 to 93), chickpea (5 ± 1 to 45 ± 1, 14 ± 3 to 96 ± 21), common bean (9 ± 1 to 75 ± 8, 18 ± 2 to 99 ± 11), cowpea (6 ± 1 to 56 ± 0.2, 38 ± 19 to 66 ± 7), lentil (10 ± 3 to 66 ± 6, 37 to 87 ± 6), mung bean (11 ± 2 to 90 ± 9, 28 ± 1 to 44 ± 6), peas (9 ± 2 to 57 ± 2, 45 ± 8 to 93 ± 9), pigeon peas (7 ± 1 to 54 ± 1, 31 ± 4), and mixed pulses (35 ± 5 to 66 ± 23, 69 ± 42 to 98 ± 29). It was found that the method of preparation, processing and heat applications tended to affect the GI of pulses. In addition, removal of the hull, blending, grinding, milling and pureeing, reduced particle size, contributed to an increased surface area and exposure of starch granules to the amylolytic enzymes. This was subsequently associated with rapid digestion and absorption of pulse carbohydrates, resulting in a higher GI. High or increased heat applications to pulses were associated with extensive starch gelatinization, also leading to a higher GI. The type of reference food used (glucose or white bread) and the other nutrients present in the meal also affected the GI.

Impact of maternal obesity on offspring adipose tissue: lessons for the clinic

Maternal obesity is a major risk factor for the subsequent development of obesity and Type 2 diabetes in the child. This relationship appears to be driven largely by the exposure of the fetus to an increased nutrient supply during critical periods of development, which results in persistent changes in the structure and function of key systems involved in the regulation of energy balance, appetite and fat deposition. One of the key targets is the fat cell, or adipocyte, in which prenatal overnutrition programs a heightened capacity for fat storage. The increasing prevalence of maternal obesity has led to an urgent need for strategies to break the resulting intergenerational cycle of obesity and metabolic disease. This review will discuss the relationship between maternal obesity and poor metabolic health of the offspring, with a particular focus on the involvement of adipose tissue, recent clinical studies examining potential strategies for intervention and priority areas for further research.

Nutritional approaches to breaking the intergenerational cycle of obesity

The link between poor maternal nutrition and an increased burden of disease in subsequent generations has been widely demonstrated in both human and animal studies. Historically, the nutritional challenges experienced by pregnant and lactating women were largely those of insufficient calories and severe micronutrient deficiencies. More recently, however, Western societies have been confronted with a new nutritional challenge; that of maternal obesity and excessive maternal intake of calories, fat, and sugar. Exposure of the developing fetus and infant to this obesogenic environment results in an increased risk of obesity and metabolic disease later in life. Furthermore, increased caloric, fat, and sugar intake can occur in conjunction with micronutrient deficiency, which may further exacerbate these programming effects. In light of the current epidemic of obesity and metabolic disease, attention has now turned to identifying nutritional interventions for breaking this intergenerational obesity cycle. In this review, we discuss the approaches that have been explored to date and highlight the need for further research.

Barley cultivar, kernel composition, and processing affect the glycemic index

Barley has a low glycemic index (GI), but it is unknown whether its GI is affected by variation in carbohydrate composition in different cultivars and by food processing and food form. To examine the effect of these factors on GI, 9 barley cultivars varying in amylose and β-glucan content were studied in 3 experiments in separate groups of 10 healthy participants. In Expt. 1, 3 barley cultivars underwent 2 levels of processing: hull removal [whole-grain (WG)] and bran, germ, and crease removal [white pearled (WP)]. GI varied by cultivar (CDC Fibar vs. AC Parkhill, [mean ± SEM]: 26 ± 3 vs. 53 ± 4, respectively; P < 0.05) and pearling (WG vs. WP: 26 ± 4 vs. 35 ± 3, respectively; P < 0.05) with no cultivar × pearling interaction. In Expt. 2, the GI of 7 WG cultivars ranged from 21 ± 4 to 36 ± 8 (P = 0.09). In Expt. 3, WG and WP AC Parkhill and Celebrity cultivars were ground and made into wet pasta. The GI of AC Parkhill pasta (69 ± 3) was similar to that of Celebrity pasta (64 ± 4) but, unlike in Expt. 1, the GI of WP pasta (61 ± 3) was less than that of WG pasta (72 ± 4) (P < 0.05). Pooled data from Expts. 1 and 2 showed that GI was correlated with total fiber (r = -0.75, P = 0.002) but not with measures of starch characteristics. We conclude that the GI of barley is influenced by cultivar, processing, and food form but is not predicted by its content of amylose or other starch characteristics.

In vitro and in vivo assessment of the glycemic index of bakery products: influence of the reformulation of ingredients

PURPOSE

To evaluate whether the modification of ingredients of two bakery products, muffins and bread, reduces their glycemic index, by means of in vitro and in vivo procedures.

METHODS

In vitro and in vivo glycemic index were evaluated for two types of bread and two types of muffins including one standard product for each category. For the in vitro determination, kinetics of starch digestion method was used. For the in vivo procedure, postprandial glucose measured as IAUC was obtained in a group of eighteen healthy volunteers (ten did the test with muffins and eight with breads).

RESULTS

In in vitro, a reduction in the expected glycemic index regarding the control muffin was achieved with the partial substitution of wheat flour by a mixture of resistant starch, dextrin and lentil flour. In breads, with the partial substitution of wheat flour by a mixture of resistant starch and dextrins, a decrease in the expected glycemic index was also observed. In in vivo, a reduction in GI was also achieved both in muffin and in bread. All the obtained GI was higher in in vitro method.

CONCLUSIONS

Despite the fact that in vitro overestimate in vivo method, the trend in the reduction in GI seems to be similar in both methods. With the substitution assayed, a reduction in the expected glycemic index and the glycemic index were obtained both in muffins and in breads.

Glycemic index and glycemic load used in combination to characterize metabolic responses of mixed meals in healthy lean young adults

BACKGROUND

Glycemic index (GI) and glycemic load (GL) have been independently assessed with regard to their metabolic effects and their associations with disease risk. Because a low-GI high-carbohydrate and a high-GI low-carbohydrate food/meal can produce the same GL, we propose that these 2 concepts should be used in conjunction to characterize the carbohydrate quality of mixed meals.

OBJECTIVE

The aim of this study was to measure postprandial glucose, insulin, and cortisol responses of meals differing in both GI and GL over a period of 3 hours (every 15 minutes for the first hour, and every 30 minutes subsequently), and to investigate the validity of methods of calculating GI and GL by measuring the cumulative incremental area under the curve (iAUC) for glucose and insulin (0-2 hours and 2-3 hours postprandially).

METHODS

A total of 10 healthy lean young adults (5 males, 5 females) participated. Breakfast meals were designed to differ in terms of GI and GL based on a 2 × 2 grid with a single elaboration: low-GI high-GL (M1); high-GI high-GL (M2a) of similar GL to M1; high-GI high-GL (M2b) of similar macronutrient composition to M1 (the elaboration); low-GI low-GL (M3); and high-GI low-GL (M4). The 5 breakfast meals were administered in a double-blind randomized crossover design. Repeated measures analysis of variance was performed to investigate differences in metabolic response between low- versus high-GI and between low- versus high-GL breakfast meals (and GI × GL interactions), with GI and GL used as within-subject factors.

RESULTS

High-GL meals increased glucose iAUC and insulin iAUC in both immediate (0-2 hours) (p < 0.01, p < 0.001, respectively) and middle postabsorptive periods (2-3 hours) (p = 0.04, p = 0.02, respectively) compared with low-GL meals. GI meals were not associated with glucose iAUC 0 to 2 hours (p = 0.37) and 2 to 3 hours (p = 0.81) postprandially. GI meals were not associated with insulin iAUC 0 to 2 hours postprandially (p = 0.81); in contrast, high-GI meals increased insulin iAUC 2 to 3 hours postprandially (p = 0.03) compared with low-GI meals. No significant differences were noted in cortisol responses, GI × GL interactions, or effect modification by gender.

CONCLUSIONS

These findings highlight the need for further investigation of meals/diets differing in both GI and GL to characterize metabolic responses and potential health effects.

Glycemic index, glycemic load and insulinemic index of Chinese starchy foods

AIM: To determine the glycemic index (GI), glycemic load (GL) and insulinemic index (II) of five starchy foods that are commonly used in Chinese diets.

METHODS: Ten healthy subjects aged between 20-30 years were recruited. Each subject was asked to consume 50 g of available carbohydrate portions of test foods and reference food. Finger capillary blood samples were collected at the start of eating and 15, 30, 45, 60, 90 and 120 min after consumption. The GI and II of foods were calculated from the ratio of incremental area under the glucose/insulin response curves of test and reference foods. The GL for each test food was determined from its GI value and carbohydrate content.

RESULTS: The results showed that brown rice elicited the highest postprandial glucose and insulin responses, followed by taro, adlay, yam and mung bean noodles, which produced the lowest. Among the five starchy foods, brown rice evoked the highest GI and GL at 82 ± 0.2 and 18 ± 0.2, followed by taro (69 ± 0.4, 12 ± 0.2), adlay (55 ± 0.4, 10 ± 0.2), yam (52 ± 0.3, 9 ± 0.0) and mung bean noodles (28 ± 0.5, 7 ± 0.2), respectively. The II values of the test foods corresponded with GI values. Similarly, brown rice gave the highest II at 81 ± 0.1, followed by taro (73 ± 0.3), adlay (67 ± 0.3), yam (64 ± 0.5) and mung bean noodles (38 ± 0.3). All five starchy foods had lower GI, GL and II than reference bread (P < 0.05).

CONCLUSION: The GI, GL and II values of starchy foods provide important information for the public to manage their diet and could be useful for the prevention of lifestyle-related diseases such as diabetes mellitus.

Glycemic impact as a property of foods is accurately measured by an available carbohydrate method that mimics the glycemic response

The relative glycemic impact (RGI), the weight of glucose that would induce a glycemic response equivalent to that induced by a given amount of food, is preferably expressed for reference amounts of foods customarily consumed per eating occasion. But because customarily consumed portions of different foods deliver different glycemic carbohydrate doses, methods for determining their RGI need to allow for homeostatic responses to different glycemic carbohydrate loadings. We tested the accuracy of an in vitro method for measuring the RGI of customarily consumed portions that allows for homeostasis, using 24 foods. Glucose equivalents released during simulated gastrointestinal digestion were adjusted by the glycemic potency of contributing sugars to obtain cumulative glycemic glucose equivalents (GGE) and multiplied by food portion weight. Corresponding dose-dependent blood glucose clearance was calculated and subtracted from GGE, giving net GGE compared with time curves reminiscent of blood glucose response curves. RGI values (GGE content) for the food portions were obtained by comparing incremental areas under the curves for foods with that for a white bread reference of known GGE content. The correlation between in vivo values calculated from glycemic index values for the same foods and in vitro values was: in vivo GGE = 1.0 in vitro GGE - 0.5; R2 = 0.90. Bland-Altman methods comparison analysis showed close agreement: in vivo GGE = -0.055 in vitro GGE + 1.16; R2 = 0.027. The results suggest that a modified available carbohydrate determination can economically provide valid RGI values for consumer and industry use.

Variable classifications of glycemic index determined by glucose meters

The study evaluated and compared the differences of glucose responses, incremental area under curve (IAUC), glycemic index (GI) and the classification of GI values between measured by biochemical analyzer (Fuji automatic biochemistry analyzer (FAA)) and three glucose meters: Accue Chek Advantage (AGM), BREEZE 2 (BGM), and Optimum Xceed (OGM). Ten healthy subjects were recruited for the study. The results showed OGM yield highest postprandial glucose responses of 119.6 ± 1.5, followed by FAA, 118.4 ± 1.2, BGM, 117.4 ± 1.4 and AGM, 112.6 ± 1.3 mg/dl respectively. FAA reached highest mean IAUC of 4156 ± 208 mg × min/dl, followed by OGM (3835 ± 270 mg × min/dl), BGM (3730 ± 241 mg × min/dl) and AGM (3394 ± 253 mg × min/dl). Among four methods, OGM produced highest mean GI value than FAA (87 ± 5) than FAA, followed by BGM and AGM (77 ± 1, 68 ± 4 and 63 ± 5, p<0.05). The results suggested that the AGM, BGM and OGM are more variable methods to determine IAUC, GI and rank GI value of food than FAA. The present result does not necessarily apply to other glucose meters. The performance of glucose meter to determine GI value of food should be evaluated and calibrated before use.

Relative glycaemic impact of customarily consumed portions of eighty-three foods measured by digesting in vitro and adjusting for food mass and apparent glucose disposal

Practical values to guide food choices for control of postprandial glycaemia need to refer to entire foods in amounts customarily consumed. We tested an in vitro method for determining the relative glycaemic impact (RGI) of customarily consumed portions of foods. Sugars released during in vitro pancreatic digestion of eighty-three foods were measured as glucose equivalents (GE) per gram of food, adjusted by the glycaemic indexes of the sugars to obtain glycaemic GE (GGE) per gram and multiplied by food portion weight to obtain the GGE contribution of the food portion, its RGI. The results were compared with clinical GGE values from subjects who consumed the same food amounts. In vitro and in vivo GGE values were significantly correlated, but the slope of the regression equation was significantly less than one, meaning in vitro GGE values overestimated in vivo GGE values. Bland-Altman method comparison showed the in vitro-in vivo disparity to increase as mean GGE increased, suggesting the need to allow for different rates of homeostatic blood glucose disposal (GD) due to different GGE doses in the customarily consumed food portions. After GD correction, Bland-Altman method comparison showed that the bias in predicting in vivo GGE values from in vitro GGE values was almost completely removed (y = 0.071x - 0.89; R2 0.01). We conclude that in vitro food values for use in managing the glycaemic impact of customarily consumed food quantities require correction for blood GD that is dependent on the GGE content of the food portions involved.