Comparison of serum glucose, insulin, and glucagon responses to different types of complex carbohydrate in noninsulin-dependent diabetic patients13

Transcription

1 Comparison of serum glucose, insulin, and glucagon responses to different types of complex carbohydrate in noninsulin-dependent diabetic patients13 Phyllis A. Crapo, R.D., Jonathan Insel, M.D., Mark Sperling,4 M.D., and Orville G. Kolterman, M.D. ABSTRACT We have studied the acute effects of oral ingestion of dextrose, rice, potato, corn, and bread on postprandial serum glucose, insulin, and glucagon responses in 2 diabetic subjects with nonketotic, noninsulin requiring fasting hyperglycemia. The carbohydrate loads were all calculated to contain 5 g of glucose. The data demonstrate that 1) dextrose and potato elicited similar postprandial serum glucose responses whereas rice and corn elicited lower responses, with bread intermediate; 2) postprandial insulin responses were relatively flat but rice ingestion led to significantly lower insulin responses than did potato; 3) urinary glucose excretion during the 3 h after carbohydrate ingestion was greatest following dextrose and least after rice and corn. In conclusion, there is a range in the magnitude of postprandial hyperglycemia after ingestion of different complex carbohydrates in diabetic patients with fasting hyperglycemia and emphasis on the use of the less hyperglycemic starches could be of therapeutic value in controlling hyperglycemia. Am. J. Clin. Nutr. 34: , KY WORDS Complex carbohydrate, diabetes, serum glucose, insulin, and glucagon responses Introduction Traditional concepts of dietary therapy of diabetes (1) and other diseases (2) have included the idea that consumption of glucose and other simple sugars produces a more rapid rise in blood glucose levels than does ingestion of longer chain or more complex carbohydrates. However, our recent studies (3, 4) and those of others (5-9) have indicated that this premise requires modification since variable responses to different complex carbohydrates (3, 4, 8, 9) as well as variable responses to different simple carbohydrates (3, 5-7) have been seen. For example, in normal individuals ingestion of potato yields postprandial glucose and insulin responses indistinguishable from those after the ingestion of a comparable amount of dextrose, while rice yields relatively flat plasma glucose and insulin response curves (3, 4). Likewise, ingestion of sucrose leads to lower blood glucose concentrations than does dextrose (3, 5, 6), and fructose ingestion has resulted in relatively flat blood glucose response curves as compared to either sucrose or dextrose (7). It seemed possible that similar variations in postprandial responses to different carbohydrates would exist in diabetic individuals. In our studies of different complex carbohydrates (4) we found that, within the range of From the Department of Medicine and General Clinical Research Center, University of Colorado Health Sciences Center, and Denver Veterans Administration Hospital, Denver, Colorado; and Department of Pediatrics, University of Cincinnati College of Medicine, Cincinnati, Ohio. 2 Supported in part by the National Institute of Arthritis, Metabolism and Digestive Diseases Grant 5 ROl 2524, National Institute of Child Health and Human Development Grant 2 ROI HD12613, and General Clinical Research Centers Grant RR-5 1 from the Division of Research Resources, National Institutes of Health, Bethesda, Maryland. Address reprint requests to: Phyllis Crapo, RD., University of Colorado Health Sciences Center, Department of Medicine; Bl51, 42 ast Ninth Avenue, Denver, Colorado Recipient of a Research Career Development Award (5K4HD278) from the National Institute of Child Health and Human Development, of the National Institutes of Health. 184 The American Journal of Clinical Nutrition 34: FBRUARY 1981, pp Printed in U.S.A American Society for Clinical Nutrition

2 RSPONS TO CARBOHYDRAT IN DIABTICS 185 normal, those subjects with the worst glucose tolerance had greater differences in their responses to the different starch loads and those with the best glucose tolerance exhibited the smallest differences. This suggested that even greater differences between postprandial glucose responses to various complex carbohydrate loads would be seen in individuals with overt diabetes, and if so, these differences could have clinical significance. In order to evaluate the postprandial responses to different starches in diabetic individuals and to test the hypothesis that greater differences between the least and most glycemic starches would be seen in diabetic patients, we have measured serum glucose, insulin, and glucagon levels as well as urinary glucose excretion in 2 subjects with nonketotic, noninsulin requiring diabetes mellitus after the ingestion of four different kinds of dietary starch (potato, rice, corn, and wheat). Materials and methods Twenty volunteers, 5 males and 15 females, were studied, and all procedures followed were in accord with the ethical standards of the Committee on Human xperimentation of the University of Colorado Health Sciences Center. Mean age of the 2 subjects was 55 (range 33 to 69) yr. Mean weight was 182 lb. and mean relative weight according to the Metropolitan Life Tables was 1.18, with a range of.76 to All of the volunteers had significant fasting hyperglycemia (fasting serum glucose levels> 14 mg/ 1 ml on three separate occasions) and none were ingesting any drugs known to affect glucose or insulin metabolism. During the course of the study, each person consumed a weight-maintenance solid-food diet that included a minimum of 2 g of carbohydrate each day. Five different carbohydrates were studied. The exact composition of the five test loads has been previously published (4). Briefly, they consisted of: 1) 5 g of dextrose, 2) 61 g of Uncle Ben s Converted Brand long grain white rice (dry weight before boiling for 2 to 3 mm), 3) 317 g of Russet Burbank potato (raw weight before baking for 45 to 6 mm in foil), 4) 279 g of heated diet-pack (Featherweight) whole kernel canned corn (drained weight), and 5) 12 g of white bread (Butternut). The content of the test substances was independently determined by chemical analysis, with carbohydrate content being calculated by difference. The test doses were administered so that the glucose load in all of them was 5.-g equivalent. The protein content of the rice, potato, corn, and bread loads was 4.9, 6.3, 7.9, and 8.7 g, and the fat content was.8, 1.7, 1.8, and 2.9 g, respectively. All studies were begun at 8 AM after an overnight fast, and their order was randomized. Because the dextrose was given in a 5 ml volume, the subjects consumed 5 ml of water along with each starch meal in order to minimize meal volume as a variable. The entire test load was ingested in 15 min, and blood samples were drawn for measurement of serum glucose, insulin, and glucagon at time zero, and at 15, 3, 45, 6, 12, and 18 min after the beginning of the period of consumption. Three-hour urine samples for the determination of urinary glucose excretion were obtained during each of the studies in six subjects. Analytic methods Serum and urinary glucose determinations were performed with a Beckman Glucose Analyzer (Beckman Instruments, Fullerton, CA) by the glucose oxidase method. Serum immunoreactive insulin was measured by the method of Desbuquois and Aurbach (1). Plasma glucagon was determined in duplicate by radioimmunoassay using antibody Gl-5, which is highly specific for pancreatic glucagon (1 1). Statistical analysis was carried out by the use of the paired t test for dependent means. Results The mean (±SM) serum glucose and insulin results after each of the oral carbohydrate loads at all time points are given in Table 1. In Figure 1 the i serum glucose responses to the different carbohydrates can be seen. Dextrose and potato give similar responses as compared to the other carbohydrates. Bread results in an intermediate response at 6 mm and a higher response than the other carbohydrates at 18 min. The postprandial glucose responses to rice and corn are similar and markedly lower when compared to the other carbohydrates. The magnitude of the differences between the most glycemic and least glycemic starches was substantial, ranging from 53 to 36 mg/loo ml from 6 to 12 min. The peak glucose responses to rice and corn occurred 15 mm earlier than the peak responses to dextrose, potato and bread. Relative weights ranged widely in the twenty patients with eight individuals exceeding a relative weight of 1.2, however, there was no correlation between the degree of obesity and the degree of difference in postprandial responses between the various carbohydrates. There were also no observable differences in the responses of the females as compared to the males. Figure 2 displays the serum insulin responses to the various carbohydrates. Since subjects with fasting hyperglycemia are usually characterized by decreased postprandial insulin responses, the curves are all relatively flat. Nevertheless, small differences between the various carbohydrates can be seen with

3 186 CRAPO T AL. TABL 1 Mean (± SM) glucose and insulin responses during the oral carbohydrate tolerance tcstsc Serum glucose (mg/loo ml) F Dextrose 237±16 Potato 224 ± 14 Rice 23 ± 16 Corn 23 ± 15 Bread 231 ± ± ± 14t 244 ± l6t 245 ± l5 244 ± ± ± 13t 276 ± 16t 282 ± 14t 274 ± 18t 346± ± 14t 299 ± l7t 34 ± 1St 32 ± 19t 373±18 35 ± 1St 32 ± l7t 34 ± 17t 32 ± 21t 333±21 31 ± l6t 284 ± l8t 274 ± l6t 32 ± ±2 263 ± ± ± l7t 283 ± 2t Serum insulin (jiu/ml) Dextrose 27±5 Potato 29 ± 5 Rice 34±8 Corn 34±8 Bread 3±7 38±8 41 ± 7 36±7 44±9 38±9 51±13 59 ± 13 44±8 49±11 46±7 6±17 67 ± 13 48±1 57±15 56±1 63±16 73 ± 2 42±8t Sl±l3 55±1 55±16 61 ± 17t 39±7 51±13 56±13 43±12 46 ± 14 38±6 43±11 52±l6t C Statistical comparisons indicate the significant differences between the dextrose values and the values of the various starches. All analyses were carried out by the use of the paired t test for dependent means. t Represents p <.1. Represents p <.5. Represents p <.1. ) 8) C/) 8) (I) Time FIG. 1. Mean z serum glucose responses to dextrose (#{149}), potato (Ls), rice (), corn (A), and bread (W. (mm) potato causing the greatest insulin response and rice the least. Post-potato and -rice values are significantly different at 3 (p <.5), 45 (p <.1), 6 (p <.1), and 12 (p <.5) minutes. Although a few post-starch insulin values were significantly different from dextrose (Table 1), no obvious trend existed. It is quite apparent that because of the marked decrease in insulin secretory capacity of these patients, that serum glucose better distin-

4 RSPONS TO CARBOHYDRAT IN DIABTICS 187 guishes the differences in postprandial responses elicited by the different carbohydrates. Since the basal glucose level accounts for a sizable proportion of the total area under the response curve, the magnitude of the differences between the carbohydrates can be better appreciated by evaluating the total incremental areas under the response curves. The incremental areas under the postprandial glucose (A) and insulin (B) response curves are summarized in Figure 3. Rice, corn, and bread yield serum glucose responses which are 11 to 41% lower and serum insulin responses which are 26 to 66% lower than those elicited by potato. It should be noted that differences between the area glucose response curves would be much less striking if total rather than incremental area were calculated, while comparable results would be seen if total insulin area were calculated. However, since only the response above basal can be altered by the acute ingestion of different foods, we believe that the incremental area is the critical value. The mean (±SM) urinary glucose excretion in six of the subjects during the oral carbohydrate tests can be seen in Figure 4. Ingestion of dextrose led to more glucose excretion than the other carbohydrates; rice and corn resulted in the least glycosuria and potato and bread gave intermediate levels. (Rice resulted in 41% less glycosuria than did potato.) The rather large standard errors for these measurements are due to the fact that fasting serum glucose levels ranged from 18 to 293 mg/l ml between the different patients, and this obviously affects the absolute amount of glycosuria between subjects. The amount of glycosuria is directly related to the serum glucose values seen with the different carbohydrates and, thus, urinary glucose loss does not account for the differences in the postprandial glucose responses. On the other hand, it is apparent that the differences between the different starches tend to be dampened by the greater degree of glycosuria after the most glycemic starch (i.e., potato), and that postprandial differences are even greater when this factor is considered. Fasting and postprandial glucagon values are presented in Table 2. As can be seen, glucagon levels are relatively nonsuppressible after any of the test loads. Although the 12- mm response to dextrose is significantly lower than the fasting value (p <.25), no consistent differences exist in the postprandial responses to any of the carbohydrates. Therefore, differences in glucagon levels cannot explain the data presented in Figures 1 to 3. Discussion We have studied the effect of orally administered loads of rice, potato, bread, and corn (5-g dextrose equivalent) on postprandial #{149} Dextrose fr- Potato o.-o Rice a-. Corn #{149}-. Bread 6 C/) 5 C o Time FIG. 2. Mean serum insulin responses to dextrose (#{149}), potato (Ls), rice (), corn (A), and bread (). (mm) 18

5 188 CRAPO T AL , A NS obi.o oi PS NS.1 1.f.1 f -F 14.C 12 1 B NS.5.5 NS.25.5 N S NS.#{212} , ) 8) C V 8).) C FIG. 3. Incremental (±SM) areas under the serum glucose (A) and serum insulin (B) response curves for dextrose, potato, rice, corn, and bread. serum glucose, insulin, and glucagon responses, and urinary glucose excretion in subjects with significant fasting hyperglycemia. The results demonstrate that the serum glucose responses after potato are similar to the responses after dextrose. Rice and corn elicited similar and much lower glucose responses than dextrose and potato. These results are comparable to previous reports in normals (3, 4, 8), and in insulin-dependent diabetic children (9). If these data are compared to our previously published results in normal subjects (4), the difference in peak glucose responses between potato and rice are 3 times greater in the individuals with fasting hyperglycemia (35 to 32 = 48 mg/ 1 ml) as compared to normal subjects (138 to 122 = 16 mg/l ml (4)). If urinary glucose loss is taken into consideration, the difference in glycemic effect between the different starches is even greater than reflected in serum glucose values alone. It thus appears that the differences in postprandial responses elicited by the different starches are magnified in subjects with fasting hyperglycemia. The explanation for the different responses to various types of complex carbohydrates is not yet known, although we have previously hypothesized that the rate of digestion and absorption may not be the same for all orally ingested starchy foods. Wahlquist et al. (12) have reported that chain length alone does not affect the rate of glucose absorption significantly. The polysaccharides evaluated in their work were processed and refmed, and were only 4 to 15 glucose molecules long (with a mean of only 5), but nevertheless, their data suggest that the differences we observed with natural starchy foods are not related to size of the molecule. If the size of the starch molecules or the granule size were critical to our results, one might expect that large starch molecules would lead to lower postprandial glucose responses. However, the starch molecules in roots and tubers (i.e., potato) tend to have larger molecular weights than the starch molecules in cereals, (rice, wheat, and corn); additionally, of the starches evaluated, potato has the largest reported granule size (13, 14). The differences seen are most likely a function of some other physical or chemical characteristic of the natural foodstuffs. For example, the gelatinization (or swelling) of the starch granules during cooking may be important since the gelatinized

6 form of a starch molecule could be more accessible to enzymatic digestion. Although gelatinization is affected by several factors, including the presence of protein, fat and other minor constituents which are closely associated with the starch in the plant cell, it is interesting to note that, of the starches evaluated, potato gelatinizes at lower temperatures (55 to 66#{176}C)than rice (68 to 78#{176}C). This would indicate that in the cooking process the potato starch granule would begin to swell at a lower temperature than would the rice starch granule. During baking the potato is subjected to higher and longer heating than C,, ).) U) Co C x 8) FIG. 4. Mean (±SM) 3-h urinary glucose excretion after ingestion of dextrose, potato, rice, corn, or bread. RSPONS TO CARBOHYDRAT IN DIABTICS 189 is the rice during boiling. A greater degree of swelling could make the potato starch more accessible to enzyme attack and the action more rapid. It should be emphasized that amylases are present in plant systems and increase in activity with heating, until they are destroyed by heat denaturation. Thus, endogenous amylase action during preparation must be considered; thermal degradation of the starch molecule may also occur. However, we have measured the free glucose levels in our potato and rice samples and have found them to be negligible. The starches may have been broken down into dextrins or smaller molecules (but not into significant amounts of free glucose) in varying degrees during the cooking process. It has been suggested that plant fiber alters intestinal transit time, and fiber has been shown to suppress postprandial glucose responses when added to meals (15). Therefore, one might suggest that differences in fiber content lead to the divergent metabolic responses. However, based on published tables (16), plant fiber can be estimated to be 11 g for the potato load, 1.5 g for the rice load, 13 g for the corn load, and 3 g for the bread load. Therefore, one would expect potato and corn, with their higher plant fiber content, to generate lower plasma glucose responses than rice and bread; in fact, the opposite was the case. Thus, the plant fiber content of the starches has no correspondence to the ordering of plasma glucose responses. Other factors affecting gastric emptying time, physical availability of the starch to hydrolytic enzymes, or the stimulation of gastrointestinal insulinogenic hormones may be important or contributing determinants in our results. In conclusion, our data suggest that if control of blood sugar is important, diets en- TABL 2 Mean (± SM) glucagon responses during the oral carbohydrate tolerance tests Se rum glucagon (pg/mi) F Dextrose 184 ± ± ± ± ± ± ± 29 Potato 157 ± 21 2 ± ± ± ± ± ± 18 Rice 28 ± ± ± ± ± ± ± 22 Corn 157 ± ± ± ± ± ± ± 2 Bread 194 ± ± 31C 234 ± 39C 229 ± ± 3S 183 ± ± 21 C These values are significantly different from dextrose with p <.5.

7 19 CRAPO T AL. riched in the less glycemic starches, at the expense of the more glycemic starches, may have a therapeutic advantage for diabetic individuals. Since these are ordinary foods, diets enriched in them are feasible and this approach could represent an adjunct to dietary therapy of diabetes. However, all of our studies have involved acute tests of single foods. Before extrapolating from these results to the free living state, chronic feeding studies in which different amounts and mixtures of the least glycemic starches are incorporated into otherwise normal diets are necessary. If the differences seen in the acute studies are present when the foods are fed for a longer period of time and in different mixtures, then emphasis on the less glycemic starches could provide an adjunct to dietary measures aimed at controlling hyperglycemia. l References I. Fajans SS. Current unsolved problems in diabetes management. Diabetes l972;21(suppl 2): Fredrickson DS, Levy RI, Bonnell M, rnst N. Dietary management of hyperhipoproteinemia, A Handbook for Physicians and Dietitians. Washington, D.C.: United States Department of Health, ducation and Welfare, Public Health Service, United States Government Printing Office, Crapo PA, Reaven GM, Olefsky JM. Plasma glucose and insulin responses to orally administered simple and complex carbohydrates Diabetes 1976;25: Crapo PA, Reaven G, Olefsky JM. Postprandial glucose and insulin responses to different complex carbohydrates Diabetes 1977;26: S. Swann DC, Davidson P, Albrink Mi. ffect of simple and complex carbohydrates on plasma nonesterified fatty acids, plasma-sugar, and plasma-insulin during oral carbohydrate tolerance tests Lancet 1966; 1: Arvidsson-Lenner R. Studies of glycemia and glycosuria in diabetics after breakfast meals of different compositions Am J Chin Nutr l976;29:7l Lamar CP. Comparative oral glucose and fructose tolerance tests in normal subjects and in diabetic patients. J Fla Med Assoc 19596: Vailer 5, Hanssen KF, Aagenaes. Plasma glucose and insulin responses to orally administered carbohydrate-rich foodstuffs Nutr Metab 198;24: Akerblom HK, Koivukangas T. Comparison of postprandialglycemia in diabetic children after breakfast meals containing as the starch component of carbohydrates either bread or potato Diabetologia 1977;13: Desbuquois B, Aurbach DF. Use of polyethylene glycol to separate free and antibody bound peptide hormones in radioimmunoassays J Chin ndocrinol Metab 197 l;33: Sperling MA, Delamater PV, Kazenelson M, Fisher RH, Fisher DA. Development and application of a radioimmunoassay for plasma glucagon. Clin Chem 1974;2: Wahlqvist ML, Wilmshust G, Murton CR, Richardson N. The effect of chain length on glucose absorption and the related metabolic response. Am J Chin Nutr l978;31: Whistler RL, Smart CL. Polysaccharide chemistry. New York: Academic Press, 19S Kerr RW. Chemistry and industry of starch. New York: Academic Press, 19S. 15. Anderson JW, Chen WL. Plant fiber carbohydrate and lipid metabolism. Am J Chin Nutr l979;32: Anderson JW, Lin W-J, Ward K. Composition of foods commonly used in diets for persons with diabetes. Diabetes Care 1978; 1: