Blood Glucose Response Curve Research Articles

Influence of test interval length on the variability of glycemic response tests

The aim of this study was to investigate whether the washout length between glycemic response tests influences their reliability. A total of 3 men and 12 women performed eight identical blood glucose tolerance tests: four tests on consecutive days (short interval) and four tests spread over 20-30 days, with 5-10 days between the tests (long interval). No difference was observed in the coefficient of variation (P=0.32) of the incremental area under the blood glucose response curve between the short and long interval, and there was no drift within the short (P=089) and long interval (P=0.20). The first test did not differ from any of the subsequent tests (P>0.99). In conclusion, glycemic response testing on consecutive days does not seem to influence the variability of glycemic response tests compared with longer intervals and it does not cause any data drift under conditions of earlier diet and habitual exercise control. In addition, familiarization trials do not seem to be necessary for glycemic response tests.

Glycemic response of mashed potato containing high-viscocity hydroxypropylmethylcellulose

Potatoes generally have one of the highest glycemic index values of any food. Relatively small differences in the glycemic response (GR) of regularly consumed starch foods have shown beneficial effects on health. Lowering the GR of a potato-based meal has potentially wide-reaching health benefits. High-viscosity hydroxypropylmethylcellulose (HV-HPMC) is a modified cellulose dietary fiber extensively used in the food industry. We hypothesized that the GR of a high-glycemic index product such as mashed potato would be lower with the addition of HV-HPMC. In a nonblind, randomized, repeat-measure, crossover controlled trial, 15 healthy adults consumed portions of mashed potato with different doses (0%, 1%, 2%, and 4%) of a specially selected and optimized HV-HPMC and a reference food (glucose) on separate occasions. Five subjects were excluded from the final analysis due to noncompliance with study procedures. Capillary blood glucose was measured in fasted subjects and at 15, 30, 45, 60, 90, and 120 minutes after starting to eat. For each sample, the incremental area under the blood glucose response curve was calculated and the GR determined. There was a significant lowering effect of HV-HPMC on GR (P < .001) of mashed potato. Glycemic responses for all mashed potato samples with the HV-HPMC were significantly lower than the standard mashed potato: 1% level (P < .05), 2% level (P < .05), and 4% level (P < .05). However, there was no significant effect of the HV-HPMC dose on GR. We conclude that addition of select HV-HPMC to mashed potato blunts GR.

Determination Of The Glycemic Index Of Selected Chinese Traditional Foods Using Different Glucose Analyzers

PURPOSE: To determine the glycemic index (GI) of 4 Chinese traditional foods and compare the agreement between portable glucose monitoring system (PoG) and the YSI glucose analyzer in determining the GI of the test foods. METHODS: Thirteen healthy adults (8 males and 5 females, age: 26.0 ± 1.3 y, BMI: 21.5 ± 0.6 kg·m-2) volunteered to participate the experiment. All subjects were required to refrain from alcohol consumption and vigorous physical activities 24h before test. After 10-14h overnight fast, the subjects consumed either reference (50g anhydrous glucose) or test foods containing 50g of available carbohydrate. The food included baked barbecued pork puff (BBPP), fried rice in Yangzhou-style (FRYS), fried fritter (FF), and "Mai-lai" cake (MLC). The finger-prick blood samples were collected immediately before, 15, 30, 45, 60, 90, and 120min after food consumption, and were analyzed with portable glucometer (Glucometer Elite, Bayer Diagnostics, Japan) and YSI glucose analyzer (YSI 1500, USA). The GI was calculated by expressing the incremental area under the blood glucose response curve (IAUC) for each test food as a percentage of each subject's average IAUC for the glucose. RESULTS: The mean glucose concentrations measured by PoG were higher than that measured by YSI (5.58 ± 0.05 vs. 5.39 ± 0.05 mmol·L-1, P<0.01). However, there were no differences in average IAUC (136.3 ± 8.3 vs. 135.8 ± 7.7 mmol·min·L-1) and GI (62.7 ± 4.6 vs. 68.8 ± 4.6) between PoG and YSI. There were no differences in the mean within-subject CV for the 13 subjects between IAUCYSI and IAUCPoG (47.3 ± 6.5 vs. 47.9 ± 3.2 %), GIYSI and GIPoG (47.9 ± 7.4 vs. 52.8 ± 3.1 %). Bland-Altman plots showed that the 95% limits of agreement for IAUC was -74.6 to 73.6 mmol·min·L-1, and for GI values was -45 to 57. No differences were observed in the mean GI values of the 4 food items determined by YSI and PoG (BBPP: 57 ± 7 vs. 54 ± 8; FRYS: 79 ± 7 vs. 73 ± 9; FF: 74 ± 9 vs. 69 ± 11; MLC: 66 ± 12 vs. 54 ± 9). CONCLUSIONS: The PoG and YSI glucose analyzer produced similar results in determining the GI of 4 Chinese traditional foods.

Variability in measurements of blood glucose response to foods in human subjects is not reduced after a standard breakfast

Measurements of blood glucose response to food are highly variable. We determined whether within-individual variability in data for blood glucose responses were reduced if individuals consumed a standard meal 2 hours before testing and investigated the effect of serving size. Blood glucose responses to muesli and macaroni cheese were determined in 13 individuals by taking 2 fasting capillary blood samples. Food was then consumed, and capillary blood samples were taken every 15 minutes for the first hour and every 30 minutes for the second hour. The incremental area under the blood glucose response curve was determined, and glycemic glucose equivalents (GGEs) were calculated. The GGE values were not significantly different whether the muesli and macaroni cheese were fed fasting or after a standard breakfast (29.2 vs 34.5 g for muesli and 11.0 vs 14.6 g for macaroni cheese). Within-individual coefficients of variation were not significantly different whether the food was consumed fasting or after a standard breakfast (24.9% and 32.5% for muesli and 38.1% and 59.4% for macaroni cheese). Differences in GGE between measured and estimated half serving size for macaroni cheese were 0.8 g (P = .6) and for muesli, 3 g (P = .2); for the double serving size for macaroni cheese, 1.7 g (P = .7); and for muesli, 6.7 g (P = .06). The GGE values for foods and variability in blood glucose response within individuals were not significantly different whether individuals fasted or consumed a standard breakfast before testing. However, blood glucose levels tended to differ significantly after consumption of the double serving size of muesli compared with other serving sizes.

Dietary glycaemic index: A review of the physiological mechanisms and observed health impacts

Carbohydrates (CHOs) are the most important energy source in human diets and are often classified by their molecular size as sugar, oligosaccharides, polysaccharides, and polyols (hydrogenated CHOs). However, the relevance of this structural classification has been questioned and interest in an alternative property of CHOs has grown. The glycaemic index (GI) is a contribution of Jenkins and co-workers in 1981 to classify CHO containing foods according to their impacts on the body?s postprandial glycaemic response. GI is defined as ?The incremental area under the 2-hour blood glucose response curve of a test food containing 50 g of glycaemic (available) CHOs expressed as the percentage of the response to the same amount of glycaemic CHOs from a standard food (either white bread or glucose) taken by the same subject?. Although white bread and glucose both give valid values, glucose may be the preferred control because of its stable composition. When, for any reason, white bread is used as reference, the obtained GI value needs to be divided by 1.4 to get the GI value contrast to glucose.

Role of Low-Glycemic-Index Foods in Improving Overall Glycemic Control in Type 1 and Type 2 Diabetic Patients and Correcting Excessive Postprandial Hyperglycemia

Increasing postprandial plasma glucose and insulin excursions are assumed to increase severity of diabetes and to be independent predictors of risk for atherosclerotic diseases and adiposity. Many possible connections have been found between postprandial events and the development of diabetes complication [1]. Lowering postprandial plasma glucose and insulin responses is relevant in preventing and managing diabetes mellitus [2,3]. Therefore, intervention towards reducing postprandial plasma glucose and insulin spikes is an essential topic in therapeutic strategy for diabetic patients, and may reduce the risk of developing cardiovascular complications. There is growing recognition that the postprandial glycemic [4,5] and insulinemic [6] responses to different foods may vary despite equal amounts of total absorbable carbohydrates, thus favoring low-glycemic-index (GI) carbohydrates. The notion of glycemic index was proposed by David Jenkins and et al. [5] more than twenty years ago as a practical way of classifying carbohydrate-containing foods according to their effect on postprandial blood glucose rather than according to carbohydrate content alone. The Gl is defined as the incremental area under the blood-glucose response curve of a 50 carbohydrate test food portion expressed as the percentage of the response to the same amount of carbohydrate from standard food (glucose, or white bread) taken by the same subject [7]. In practice, the actual carbohydrate load from a normal portion size varies considerably. Both type and amount of carbohydrate have now widely accepted influences on glycemic response [8,9]. In order to address this problem, the concept of glycemic load (GL) was introduced. GL, calculated as the amount of carbohydrate in one serving multiplied by the GI of the food, allows comparisons between likely glycemic effects of different foods in realistic portions [10]. However, GL values should be considered with caution, as both a low-GI/high-CHO food and a high GI/low-CHO food can have the same GL with different metabolic effects on β-cell function and plasma lipids [8]. Separate use of GL may lead to habitual consumption of low-carbohydrate diets. While short-term beneficial effects could be induced by low-carbohydrate diet, the long-term effects might be harmful. All the centers over the world that have tested experimental and clinical use of low-GI foods are enthusiastic about diet counseling to diabetic or hyperlipidemic patients, and even normal subjects at risk of cardiovascular diseases. A large number of studies have indeed demonstrated the efficiency of diet counseling and the use of low glycemic index (low-GI) foods in these patients. More striking effects were noted in the improvement of postprandial blood glucose excursions and, consequently, in glycated hemoglobin, fasting plasma lipids - particularly triglyceride levels - and, marginally, total and LDL cholesterol. We will give some insights on a few of our clinical and experimental results on this topic. At this stage, we would like to point out that the use of low-GI foods is only part of a more general strategy to improve postprandial hyperglycemia, and thus overall blood glucose control.

Determining the glycemic glucose equivalent value of foods in humans

Glycemic glucose equivalent is calculated from measurements of the incremental area under the blood glucose response curve (AUC) for a food of a specified weight divided by the AUC in response to a specified amount of glucose. In this study, we tested the assumption that the blood glucose response for ingested glucose is a linear function of glucose intake over a range of normal intakes and, therefore, that a reference drink of one concentration can be used and a correction factor applied to account for the fact that the reference drink and food are of different weight. Ten individuals, 9 female and 1 male with a mean age of 29 years, consumed 5 different glucose drinks—25, 50, 60, 80, and 100 g on 5 separate occasions. Two fasted blood samples were collected by capillary blood sampling, then blood samples were collected at 15, 30, 45, 60, 90, and 120 minutes after the drink was consumed, then every half hour until individuals' glucose levels returned to baseline, for a maximum of 4 hours. The AUC for each of the glucose drinks was calculated. Up to a glucose load of 65 g, the AUC increased linearly with the amount of glucose, though above this level, there was a plateau effect. Therefore, glycemic glucose equivalent can be determined using a single reference concentration and a correction factor for foods containing less than 65 g of carbohydrate.

Glycaemic index and glycaemic load values of commercially available products in the UK

The objective of this paper is to provide glycaemic index (GI) and glycaemic load (GL) values for a variety of foods that are commercially available in the UK and to compare these with previously published values. Fasted subjects were given isoglucidic (50 or 25 g carbohydrate) servings of a glucose reference at least two to three times, and test foods once, on separate occasions. For each test food, tests were repeated in at least eight subjects. Capillary blood glucose was measured via finger-prick samples in fasting subjects (0 min) and at 15, 30, 45, 60, 90 and 120 min after the consumption of each test food. The GI of each test food was calculated geometrically by expressing the incremental area under the blood glucose response curve (IAUC) of each test food as a percentage of each subject's average IAUC for the reference food. GL was calculated as the product of the test food's GI and the amount of available carbohydrate in a reference serving size. The majority of GI values of foods tested in the current study compare well with previously published values. More importantly, our data set provides GI values of several foods previously untested and presents values for foods produced commercially in the UK.

The effect of walking before and after breakfast on blood glucose levels in patients with type 1 diabetes treated with intensive insulin therapy

Objective: We examined the effect of walking at different timing on carbohydrate metabolism in patients with type 1 diabetes. Subjects and Methods: Subjects included six non-obese patients treated with intensive insulin therapy. The blood glucose profile was determined with and without walking for 30 min before (ExBM) or after (ExAM) breakfast. Results: Mean blood glucose values at 07:00 h in the control, ExBM, and ExAM were 9.0, 8.0, 8.8 mM, respectively. Glucose levels gradually increased after meals up to 13.6, 15.0, and 15.3 mM, respectively, at 09:00 h (0.5 h after meals). At 09:30 h, glucose levels significantly fell to 11.0 mM during walking in the ExAM (P=0.039 vs 09:00 h values). The area under blood glucose response curve was significantly lower only in the ExAM when compared with that in the control (P=0.043) (11.8, 17.8. and 3.8 h mM in the control, ExBM and ExAM, respectively). Conclusion: These results might suggest that walking after meals improves glycemic control in patients with type 1 diabetes being treated with intensive insulin therapy consisting of the basal-bolus (NPH-human regular) insulin regimen.

INFLUENCE OF HIGH AND LOW GLYCEMIC INDEX CARBOHYDRATE MEALS ON ENDURANCE CAPACITY

1404 The purpose of this study was to examine the effect of high and low glycemic index (GI) carbohydrate (CHO) meals on endurance running capacity. Eight active subjects (5 males, 3 females) ran on a treadmill at 68.0 ± 1.0% VO2max to exhaustion on two occasions separated by 7 days. Three hours before the run, after an overnight fast, each subject was given in a single-blind, random order isoenergetic meal (850 ± 21 kcal; 67% carbohydrate, 30% protein and 3% fat) containing either high (HGI) or low(LGI) GI foods providing 2.0g CHO/kg body weight. Our preliminary findings indicated that ingestion of the HGI meal resulted in a 290% and 243% greater incremental area under the 1-hour blood glucose response curve and 3-hour serum insulin response curve respectively. Performance times were not different between the HGI and LGI trials (111±6 min and 113±5 min respectively). During the first 80 minutes of exercise in the LGI trial, CHO oxidation was 12% lower and fat oxidation 118% higher than HGI trial. Although serum insulin concentrations did not differ at the onset of and during exercise, blood glucose at 20 min into exercise in the HGI trial was 3.9 ± 0.3 mmol.l-1 compared to 4.4 ± 0.3 mmol.l-1 during LGI (p<0.05). There was also a trend towards decreased FFA concentrations during exercise in the HGI trial. Blood lactate and serum cortisol concentrations were not different between the two trials. It is likely that the enhanced CHO oxidation during the high GI trial is a consequence of increased muscle and liver glycogen stored before exercise. However, the low GI meal may have provided less CHO to the muscle for subsequent utilisation because of delayed digestion and absorption. Thus, despite the CHO sparing effect of the low GI meal, endurance running capacity did not differ.

Controlled Portions of Presweetened Cereals Present no Glycemic Penalty in Persons with Insulin-Dependent Diabetes Mellitus

Objective to determine metabolic responses to commercially sweetened flaked corn cereal, unsweetened flaked corn cereal, glucose, and sucrose in teenagers and young adults with insulin-dependent diabetes mellitus (IDDM). Design A crossover design in which each subject consumed test meals in random order on 4 separate days at least 72 hours apart. Setting The inpatient setting of the General Clinical Research Center of the Indiana University Medical Center Hospital. Subjects Sixteen males and eight females, aged 14 to 25 years, with IDDM. Interventions After fasting overnight, each subject underwent challenge tests with 50 g carbohydrate per 1.73 m 2 of body surface area from sweetened flaked corn cereal, unsweetened flaked corn cereal, sucrose, and glucose. All subjects were maintained on continuous intravenous infusion of insulin overnight (euglycemic goal=3.9 to 6.7 mmol/L), with a constant basal insulin dose infused before and throughout a 3-hour postprandial period. Main outcome measures Plasma glucose, free insulin, triglycerides, and free fatty acid levels measured at baseline and 15, 30, 45, 60, 90, 120, 150, and 180 minutes after meals. Statistical analyses performed Comparisons among the four meals were made using two-way repeated measures analyses of variance followed by the Newman-Keuls multiple comparison procedure to identify specific differences among meals. The areas under the response curves were compared using one-way repeated measures analysis of covariance, adjusted for baseline values. Results The response to glucose for the area under the 3-hour blood glucose response curve was significantly greater than the response to sucrose ( P=.006 by repeated measures analysis of variance); the areas for the two cereals (not significantly different from one another) were between the glucose and sucrose areas. At 3 hours, glycemia differed significantly among three of the meals: unsweetened flaked corn cereal>sweetened flaked corn cereal>sucrose ( P<.001). Glucose at 3 hours was greater than sucrose ( P<.001). There were no significant differences for free insulin, triglycerides, or free fatty acids. Applications Equivalent gram amounts of carbohydrate as presweetened breakfast cereals are not detrimental to persons with IDDM compared with unsweetened cereals. Therefore, presweetened cereals can be used in the correct portion sizes and based on the number of carbohydrate or starch servings in a person's diabetic meal plan. J Am Diet Assoc. 1996; 96:458-463.

A New Bioassay For Insulin in Conscious Rabbits by Continuous Measurement of Glycemic Responses

Because of the substantial variability often observed in the determination of insulin potency by the United States Pharmacopeia (USP) bioassay method, a large number of rabbits are required in order to attain an accuracy of ±6% and each bioassay needs 2–3 weeks to be completed. In this report, an improved bioassay method has been developed. This bioassay was conducted in conscious healthy rabbits. The decline in blood glucose levels, following intravenous injection of an insulin preparation, was monitored (in uninterrupted manner) by the continuous glucose monitoring system developed. A glucose response curve was generated, and from this response curve, various pharamcodynamic parameters were easily determined. The whole procedure could be completed in 1 day and also achieved an accuracy where upon only one rabbit is needed to determine accurately the insulin potency. To validate the method, a total of nine healthy rabbits were studied to determine the inter‐ and intra‐animal reproducibility. The values of insulin potency determined from four pharmacodynamic parameters were compared, and the potency calculated from the ABGC (area of the blood glucose response curve under baseline) was found to be the most accurate: a mean (±SEM) value of 102.3(±1.2)% was determined by inter‐animal study (n= 9) and a value of 100.6(±1.3)% by intra‐animal study (n= 4). The inter‐bioassay variability among the random‐dose bioassays was 1.2% for inter‐animal and 1.3% for intra‐animal studies. The insulin potency in an insulin sample determined by this bioassay had attained an overall mean (±SEM) value of 101.6(±0.8)% (n= 17). Apparently, the bioassay method developed is suitable for the determination of insulin potency in the insulin preparations and a high degree of accuracy can be attained even with the use of fewer animals, also within a shorter period of assay time, than the conventional USP method.

Day-to-day variation of blood glucose and insulin responses in NIDDM subjects after starch-rich meal.

To study day-to-day variation of postprandial blood glucose and insulin increments in non-insulin-dependent diabetes mellitus (NIDDM) subjects and to analyze intra- and interperson variance of response. Ten NIDDM subjects attending the outpatient clinic at Aarhus Kommunehospital were studied. The subjects ate three meals of 90 g of white bread, with 7 days between tests. Mean +/- SD areas under the blood glucose response curve (above basal) over a 3-h period were 557 +/- 60, 569 +/- 74, and 565 +/- 67 mM x 180 min (NS), and areas under the insulin-response curve were 3350 +/- 448, 2815 +/- 359 and 3551 +/- 679 mU/L x 180 min (NS) on each of the three occasions. The 95% confidence intervals of blood glucose and insulin areas for the test meal repeated three times were 564 +/- 120 mM x 180 min and 3240 +/- 1645 mU/L x 180 min, respectively. Intra- and interperson components of variance were 25 vs. 75% (glucose) and 78 vs. 22% (insulin) of the total variance. The intraperson components of variance included all sources of variation other than between-person variation. There was no significant correlation between blood glucose and insulin response areas. A valid estimate of the glycemic response in a single patient is obtained after a single meal. Because of the large between-person variation, paired data should preferably be used when comparing glycemic responses to different foods.

Low glycemic response to traditionally processed wheat and rye products: bulgur and pumpernickel bread

To look at the effect of processing wheat and rye on blood glucose responses with special reference to bulgur and pumpernickel bread, groups of 9–12 Noninsulin-dependent (NIDDM) and 5–6 Insulin-dependent diabetic volunteers (IDDM) were fed test meals containing 50 g carbohydrate portions of four wheat and three rye products. Glycemic indices for IDDM and NIDDM combined, calculated as the incremental area under the blood glucose response curve, where white bread = 100, demonstrated values of 96 ± 5 for wholemeal wheat bread, 89 ± 6 for wholemeal rye bread, 78 ± 3 for pumpernickel bread, 65 ± 4 for bulgur, 63 ± 6 for whole wheat kernels and 48 ± 5 for whole rye kernels. Results for IDDM and NIDDM were similar (r = 0.96, p < 0.01). It is concluded that traditional processing of cereals, such as parboiling (bulgur) or the use of wholegrains in bread (pumpernickel) may result in the low GI value associated with the unmilled cereal. Cereal foods processed in these ways may form a useful part of the diet where a reduction in postprandial glycemia is required.

The dynamic property of glucagon secretion in response to blood glucose--analysis with the aid of the control theory using an artificial endocrine pancreas

The precise dynamic property of glucagon secretion in response to glucose concentration has not yet been elucidated, since in in vitro studies using pancreatic islets, the co-existence of pancreatic B cells modifies the mode of glucagon secretion, and in vivo studies, the exogenous insulin administration greatly affects pancreatic A cell function. In this study, to clarify the dynamic property of glucagon secretion in response to glucose concentration and the ability of glucagon to raise hepatic glucose production, an artificial endocrine pancreas was employed as a research tool. Our originally developed artificial endocrine pancreas prepared the glucagon and/or glucose infusion algorithm as the counterregulatory system. The principle of glucagon and glucose infusion algorithm is set as the proportional plus derivative modes of action to blood glucose concentration with time delay constant, as follows; Gn.I.R.(t) = Gp[BGp - BG(t - tau)] + Gd[-delta BG(t - tau)] + Gc G.I.R.(t) = Cp[BGp - BG(t - tau)] + Cd[-delta BG(t - tau)] where Gn.I.R.(t) and G.I.R.(t) are glucagon infusion rate and glucose infusion rate, respectively. BGp is the projected value of blood glucose concentration, and BG(t) and delta BG are blood glucose concentrations and the rate of change in blood glucose concentration at time t respectively. Gp and Gd are coefficients for glucagon infusion, and Cp and Cd are those for glucose infusion. Gc is the constant for basal glucagon supplementation. tau(min) is the time delay constant for glucagon and glucose infusion. In a depancreatized dog, the blood glucose concentration was maintained at the normoglycemic level with intraportal insulin infusion using this artificial endocrine pancreas for at least one hour, then hypoglycemia was induced by iv bolus insulin injection (0.1 U/kg). Glucagon was infused intraportally or glucose was infused into the peripheral vein when the counterregulatory system was operated according to each of these algorithm by variously changing the parameters. In the intraportal glucagon infusion algorithm, with the optimal parameters based on proportional plus derivative modes of action with a 10-min time delay (Gp/Gd/Gc/tau = 0.2/0.4/0.6/10), both the blood glucose response curves and plasma glucagon profiles simulated those seen in normal dogs. On the other hand, glucagon infusion based only on the proportional action failed to simulate the blood glucose response and plasma glucagon profile of normal dogs. When glucose was infused on the basis of the proportional action with a 20-min time delay (Cp/Cd/tau - 0.2/0/20), the insulin-induced hypoglycemia in depancreatized dogs could be restored to normoglycemia in the sa

Relationship between rate of digestion of foods and post-prandial glycaemia.

The amount of carbohydrate released at 1 and 5 h by digestion in vitro of 2 g carbohydrate portions of 14 foods by human digestive juices was compared with the area under the 2-h blood glucose response curve when 50 g carbohydrate portions were fed to groups of five to ten healthy volunteers. A significant relationship was found between the amounts of sugars and oligosaccharides liberated at 1 and 5 h and the food blood glucose area expressed as a percentage of the blood glucose area for 50 g glucose (r = 0.8627 and 0.8618, p less than 0.001). A significant relationship was also found between the glycaemic index and the food fibre content (p less than 0.02) and between the glycaemic index and the glucose trapping capacity of the foods (p less than 0.05). Legumes as a group liberated 56% less sugars and oligosaccharides (p less than 0.01) than the eight cereal foods over 5 h. It is suggested that such studies in vitro may help to identify food of use for diabetic patients, and at the same time throw further light on factors which affect post-prandial glycaemia.

Scope and specificity of acarbose in slowing carbohydrate absorption in man.

Fifty-gram carbohydrate tolerance tests were performed on healthy volunteers to test the activity and specificity of an alpha-glucoside hydrolase inhibitor, acarbose (BAY g 5421). Two hundred milligrams acarbose reduced the area under the blood glucose response curve by 89% (P less than 0.001) after sucrose by 80% (P less than 0.002) after starch, by 19% (N.S.) after maltose, with no effect on glucose. Breath hydrogen measurements indicated an almost complete malabsorption of the sucrose. At 50 mg acarbose, some reduction in blood glucose and insulin response to sucrose was still seen, but no significant hydrogen production. It is suggested that at lower doses, acarbose may prolong the time course over which carbohydrate is absorbed as does dietary fiber; as with fiber, it may be a useful adjunct to diabetic therapy.