Diabetes Research and Clinical Practice, 15 (1992) 135-142 CD 1992 Elsevier Science Publishers B.V. All rights reserved 0168-8227/92/$05.00
135
DIABET 00596
The effect of muesli or cornflakes at breakfast on carbohydrate metabolism in type 2 diabetic patients Alain
Golay
*, Brigitte
Koellreutter
Bloise ‘, Jean-Philippe 2, Donatella Pierre Wiirsch 2
Assal ’ and
’ Unite d%nseignement pour Diabktiques. Hbpital Cantonal Universitaire. Geneva, Switzerland and ‘NestlP Research Centre, Nestec Ltd., Vers-chew-les-Blanc. Lausanne, Switzerland (Received 17 April 1991) (Revision accepted 23 August 1991)
Summary Fourteen overweight insulin-treated type 2 diabetic patients ate a breakfast, consisting of either muesli (slow release starch: SRS) or cornflakes (fast release starch: FRS), in either case with milk (46 g carbohydrate), during two consecutive randomized crossover periods of two weeks. The rest of the diet remained unchanged. At the end of each period the patients underwent a glucose tolerance test after an overnight fast without their usual evening insulin injection. Both mean plasma glucose response curves were identical after the two dietary periods, but plasma insulin was significantly lower at zero ( - 17”/,, P < 0.05) and 2 h ( - 21x, P < 0.05) at the end of the muesli (SRS) period as compared to the cornflakes (FRS) period. The mean day-long plasma glucose level (four measurements) at the end of the muesli period was 21”/, (P= 0.023) lower than after the cornflakes period. These results show that switching, at breakfast only, from standard cereals to slow release starch cereals improves the carbohydrate metabolism of diabetic patients. In addition, the fact that diabetic patients could reduce their insulin requirement (P< 0.05) with concomitant reduction of their daily blood glucose level implies that sensitivity to insulin was improved by slow release starch foods consumed at breakfast.
Key words: NIDDM;
Breakfast;
Cereal; Diet; Insulin;
Current recommendations from the ADA for treatment of patients with diabetes includes the Correspondence to: P. Wiirsch, Nestle Research Centre, Nestec Ltd., Vers-chez-les-Blanc, CH-1000 Lausanne 26, Switzerland. Abstract presented at the 13 ICC Congress, Wien, Austria, May 28-31, 1990.
Sensitivity
use of high-carbohydrate, low fat diets in order to reduce the risk of atherosclerotic heart disease [ 11. However, it has been pointed out that an increase in the carbohydrate intake may lead to postprandial hyperglycemia, increased glycosuria, hyperinsulinemia, elevated fasting plasma triglyceride levels and reduced HDL cholesterol in NIDDs treated with diet alone or diet plus sulphonylureas [2]. A rich carbohydrate diet which
136
nonetheless produces blunted glycemic and insulinemic responses has therefore been recommended. Incorporation of low rather than high glycemic index foods into the diet improved the carbohydrate metabolism of diabetic patients and often decreased serum lipid concentrations [ 3-81. Diabetic patients require more insulin for a mixed meal at breakfast, than for an isocaloric mid-day or evening meal [ 9, IO]. However, it has been suggested that insulin sensitivity during the second meal may be related to the effect from the preceding glucose challenge. Furthermore, a high fat breakfast results in impairment of the glucose tolerance to a standard lunch [ 111. Thus, better insulin sensitivity following a slowly absorbable carbohydrate eaten at breakfast might promote enhanced tolerance for the carbohydrate intake at lunch, without producing an excessive blood glucose excursion after the breakfast. This type of reduction of blood glucose excursion was achieved after the consumption of a muesli containing raw rolled wheat and instant bean flakes compared with cornflakes [ 121. The aim of the present study was to determine whether two weeks of a diet including slow release starch at breakfast compared with a diet including conventional breakfast cereal, could improve glucose tolerance or insulin sensitivity in type 2 diabetic patients, with insulin requirement.
Subjects Fourteen adults with type 2 diabetes participated in the study (Table 1). All the patients were treated with insulin on average for 2.7 + 1.2 years and diet only. All the patients studied were NIDDM requiring insulin for optimal blood glucose control. They were on a regimen of two injections of intermediate insulin without short acting insulin. Doses ranged from 8 to 24 U in the morning and 4 to 22 U in the evening. None of them had long term complications of diabetes. The dose of insulin injected was adapted by the patients themselves when they felt the need, but they were advised to do it only once a week in each period.
I-ABLE
1
Physical characteristics
(n = 14, mean f SD)
62 fr Age (years) 83 t Weight (kg) 168 + Height (cm) 29 * BMI (kg/m’) Energy intake (kJ/day) 6926 k Carbohydrate (Y’,,daily energy intake) 45 f .~______~.._
11 146 7 5 2110 8.6
The subjects received dietary instructions during the two weeks of training in our diabetes teaching unit before starting the one week stabilisation period at home. The blood glucose monitoring was carried out at home, using a glucometer (Reflolux II M, Boehringer Mannheim, F.R.G.).
Study design The study was conducted on an outpatient basis. The experimental protocol was approved by the Ethics Committee of the Department of Medicine of Geneva University Hospital and the subjects gave their written informed consent to their participation in the study. All the patients selected were on insulin. The study consisted of one week of metabolic stabilisation (fasting plasma glucose below 10 mmol/l followed by two weeks of dietary trials separated by a one week washout interval. The order of the two periods was randomized. The patients were asked to maintain their usual personal diet throughout the study apart from the breakfasts and morning snacks, which were provided in individual portions. They were strictly followed by a dietitian during the whole study. Their diets were recorded and found to be very constant during their two two-week periods. They were asked to monitor their blood glucose fasting at 7 am and at 11 am, 2 pm and at 5 pm, on the last two days of each period. All patients underwent an oral glucose tolerance test (OGTT) after each dietary period. The patients fasted overnight and omitted their usual insulin injection the evening before the test and the morning of the test. During the 75-g OGTT,
137
blood samples were drawn at time 0, 30, 60, 120, 180 min and plasma glucose and insulin were measured. Plasma cholesterol, HDL-cholesterol, triglycerides and glycosylated hemoglobin were also measured at time 0. Plasma was separated immediately, frozen and subsequently analyzed for plasma glucose (Beckman Glucose Analyzer, Palo Alto, CA), and cholesterol [ 131, triglycerides [ 141, insulin (RIA Pharmacia Upsala, Sweden). The patients did not have significant antibodies to disturb the measurements.
Diets The breakfasts provided 46 g carbohydrate or roughly 27% of the daily carbohydrate intake (Table 2). The size of the meal was fixed prior to the study by testing it with diabetic patients. The carbohydrate distribution was comparable in both meals e.g. 26 g starch, 14.4 g sucrose, honey, dextrin and 5.6 g lactose. The energy content of the cornflakes meal was 10% lower due to lower protein and lipid content. The SRS breakfast was composed of 65 g muesli in 120 ml whole milk. The starchy ingredients in the muesli were selected from among those having a very low digestion index of the starch as described by Schweizer et al. [ 151 i.e., raw rolled wheat and white bean flakes [ 16-181.
TABLE
2
Composition
of the breakfasts SRS muesli + milk
Carbohydrate (g) Starch (g) Sugars (g) Protein (g) Lipid (g) Dietary fibre (g) Energy (kJ)
46 26 20 11 6.2 6.7 1200
FRS cornflakes + milk 46 26 20 I 4.4 1.2 1080
The other breakfast (FRS) was composed of 35 g cornflakes, 12g sugar and 120 ml whole milk. The sugar level was the same in both meals. The snack at 10 am during both test periods was composed of 20 g wholemeal bread and 20 g cottage cheese (protein: 4.8 g, carbohydrate: 11 g, lipid: 2.2 g, total energy: 350 kJ). The patients were free to double this amount but they had to keep it unchanged during the whole study. Starch and sucrose in the meals were analysed as described by Holm et al. [ 193, fibre by using the AOAC method [20] and the other components calculated from food tables [ 211. Caloric intakes were based on the patients’ dietary record made during the three days of the stabilisation period. Total energy intake was 6929 + 2110 kJ/day and carbohydrate accounting for 45 * 8.6%.
Statistical analysis Data are presented as means with standard error of the means. Area under the glucose and insulin was determined according to the method described by Wolever and Jenkins [ 221. Means were compared by the paired t-test. For the daily blood glucose levels and for the areas under the glucose and insulin curves, two way of variance (ANOVA) was used for repeated measurements. For this analysis, the two independent variables were time and treatment. If there is significant interaction between different treatment and time, the statistical significance among groups at any point was analyzed.
Results The patients’ weight remained stable during the study (beginning 82.5 f. 3.7 kg; end muesli period: 82.4 + 3.8 kg; end cornflakes period: 82.5 k 3.8 kg). The patients reduced their daily insulin dose during the muesli period from a mean of 28 + 4 U to 24 k 4 (P < 0.05) at the end of the period. The results of the glucose tolerance tests
138
0
60 TIME
120 (
180
minutes I
I _” 400
-
I ?
-
300
k__
J
I
0
60 TIME
1
120 ( minutes
180
I
Fig. I. Mean ( + SEM) plasma glucose and insulin reaponscs for I80 min after ingestion of 75 g glucose at the end of two weeks of consumption of an SRS muesli (-) or FRS cornflakes (---) in 14 diabetic patients. Statistical significances are between both diets. *: P < 0.05.
are shown in Fig. 1 and Table 3. Fasting plasma glucose concentrations as well as the glucose response profiles were identical at the end of each diet. However, the fasting insulin level was 17% lower after the muesli than after the cornflakes period (P < 0.05). The insulin response curve also tended to be lower, and significance was reached only at 120 min (P < 0.05). Mean blood glucose levels measured at four time points (7, 11 am and 2, 5 pm) at the end of each period are presented in Fig. 2. The mean blood glucose levels during the muesli period were always lower by 18, 14,29 and 20%, respectively than during the cornflakes period (mean of all values, - 2 1%, P = 0.023). During the control period the mean values were in between, and were significantly higher than those associated with the muesli diet at 5 pm (P < 0.05). The effect of treatment -type of breakfast - was neither dependent on the order of the breakfast, nor on the period. Mean glycosylated hemoglobin levels were similar after the two study periods and during the stabilization period., Thus, the reduction of the daily
TABLE 3 Metabolic parameters (mean 2 SEM)
after SRS-muesli and FRS-cornflakes
and the oral glucose tolerance test (OGTT) in 14 diabetic patients
P
FRS
SRS
_
Mean daily blood glucose level,’ (mM) Glycosylated hemoglobin” (‘I,, ) Cholesterol (mM ) HDO-cholesterol (mM 1 l‘riglycerldes (mhl 1
7.1 7.5 5.7 I.0 1.7
+ i f f *
0.7 0.5 0.4 0.06 0.3
9.7 7.1 5.4 1.0 1.8
* f * + +
0.9 0.4 0.3 0.06 0.4
< 0.02 NS NS NS NS
9.1 f 19.3 + 22.8 + 150 f 324 k 768 +
0.8 1.0 1.4 54 78 180
NS NS NS < 0.05 NS NS
OGI-I
fasting blood glucose (mM) max. glucose (mM ) Incremental area under glucose curve (mM. 3 h) fasting insulin (PM) max. insulin (PM) arca under insulin curve (PM. 3 h) ,’ Before the study: X.8 k 0.60, P < 0.05 compared
8.9 k 0.6 19.4 * I.0 21.7 k 2.0 126 2 48 268 & 53 660 k 132
with SRS; h before the study: 7.9 k 0.4.
139
2
10
E
-3 ,......... I . . . . . . . . . . . . .
I . .
*;
r /.....
. . . . . . . . . .“>‘....._**_
I
_......a
* T**
z **
8 3 -I0
i
5
t
_ 07.00
.
_
,
_
.
_
14.00
11.00 TIME
.
( hours
_
.
_ 17.00
)
Fig. 2. Day-long mean ( i SEM) plasma glucose level at the end ofthe 2 weeks muesli ( + ) and cornflakes (0) breakfasts. *: P < 0.05; **: P < 0.02.
mean blood glucose level during the muesli period was not reflected by a decrease in glycosylated hemoglobin, probably because of the brevity of the study period. Total cholesterol, HDL-cholesterol and total triglycerides remained unchanged throughout the study. The patients filled in a questionnaire at the end of each period. Ten out of the fourteen subjects found that the muesli meal appeased their hunger for longer than the cornflakes. Nobody reported this advantage with the cornflakes.
Discussion This study shows that breakfast containing slow release starch cereals appears to improve glycemic control and to reduce insulin requirement in type 2 diabetic patients. Modem dietary recommendations for increased carbohydrate, reduced fat intake for diabetic patients specify that the carbohydrate-rich food should be of the ‘unrefined’ type. However, several recent comparative dietary studies using similar fibre contents suggest that fibre per se is not the major determinant of improved metabolic control [23]. When the foods were selected according to the criteria of the low amplitude of glycemic response, significant changes in carbohydrate metabolism were observed. Extensive substitution of the diet by low glycemic index foods such as whole grain
bread, leguminous seeds, parboiled rice and pasta was very beneficial to the glucose control of diabetic patients [ 3-7,241. Even modest dietary changes from bread and potato to rice, biscuits and pasta for three weeks in IDDM subjects resulted in reduced insulin requirement, better glucose control and lower concentrations of triglycerides [ 81. Brand et al. [3] obtained an improvement of the glycemic control in NIDDM after 12 weeks on a reduced glycemic index diet. The practical nature of our study design allowed the subjects to take major responsibility for their own diabetic control. Their degree of glycemic control may therefore be assumed to be representative of insulin treated type 2 diabetic out-patients. Although the selection of the low or high glycemic food was limited to breakfast, a significant difference in the glucose metabolism was achieved. During the low glycemic index breakfast period (raw cereal flakes and bean flakes), both insulin requirement and daily blood glucose were lower than during the high glycemic index breakfast period (cornflakes). No difference in basal glucose level was observed on the day of the glucose tolerance test, but the basal insulin level was 17% lower. This may be because the patients did not inject their intermediate insulin on the evening before the test. After two weeks of treatment, the glucose responses to the OGTT were similar following both types of diet. However, the insulin response to the OGTT tended to be lower after the SRS and the difference was significant at baseline and at 2 h. The association of reduced insulin response with a similar glucose response implies improved sensitivity to insulin during the glucose challenge performed after the SRS period. The difference between the diets was insufficient to modify the blood lipid profile. Roth et al. [25], who tested the effect of high fibre cereals against cornflakes at breakfast on 74 healthy adults for 12 weeks observed that the high fibre cereals did not significantly improve the blood cholesterol and triglycerides levels, but increased the HDL-cholesterol.
140
High amylose meals are associated with lower glucose and insulin responses in healthy volunteers than high amylopectin meals [26]. Comparison over a 5-week period of diet containing two thirds of the carbohydrate either as high amylose starch or high amylopectin starch, showed that amylose was associated with a significant reduction in triglyceride and cholesterol levels, and a reduction of the insulin response 1 h after a glucose load [ 271. Similar lipid differences were also observed in NIDDM subjects, between high rice and high potato diets [4] or high pasta and high pasta-guar diets [28]. These observations suggest that blood lipids may decrease if enough of the dietary carbohydrate in the diet is of the slow release type. The breakfasts tested in the study, contained little fat (less than 20% of total energy). High carbohydrate meals produce higher postprandial glucose levels than meals containing more fat [30], because fat delays gastric emptying and increases insulin response [29]. However, Collier et al. [ 111, have shown that the simultaneous ingestion of fat and carbohydrate in a meal impairs glucose tolerance to subsequent meals. As our study design precluded modification to the remainder of the diet, we tried to avoid an unfavourable carry-over effect. It has also been shown, that breakfast including low glycemicindex food reduces the glycemic excursion of the next carbohydrate meal [31,32]. All these elements may have contributed to the metabolic differences observed. The muesli breakfast also appeased hunger for longer than cornflakes. This has already been reported for high-fibre cereal breakfasts of reduced glycemic index [33] and beans [34]. In conclusion, breakfast including slow release starch cereals appears to improve glycemic control and to reduce insulin requirement in NTDDM patients. The concomitant decrease of blood glucose and insulin requirement implies that sensitivity to insulin has been improved.
References 1 American Diabetes Association Nutritional recommendations and principles for individuals with diabetes mellitus: 1986, 1987. Diabetes Care 10, 126-132. 2 Coulston, A.M., Hollenbeck, C.B., Swislock, A.L.M. and Reaven, G.M. (1989) Persistence of hypertriglyceridemic effect of low-fat high carbohydrate diets in NIDDM patients. Diabetes Care 12, 94-101. 3 Brand, J.C., Colagiuri, S., Crossman, S., Allen, A., Roberts, D.C.K. and Stewart Truswell, A. (1991) Lowglycemic index foods improve long-term glycemic control in NIDDM patients. Diabetes Care 14, 95-101. 4 Andersen, E., Hellstrom, P., Karlander, S.G. and Hellstrom, K. (1984) Effects of rice-rich versus a potatorich diet on glucose lipoprotein and cholesterol metabolism in non-insulin-dependent diabetics. Am. J. Clin. Nutr. 39, 598-606. 5 Jenkins, D.J.A., Wolever, T.M.S., Buckley, G. et al. (1988) Low-glycemic-index starchy foods in the diabetic diet. Am. J. Clin. Nutr. 48, 248-254. 6 Collier, G.R., Guidici, S., Kalmusky, J. et al. (1988) Low glycaemic index starchy foods improve glucose control and lower serum cholesterol in diabetic children. Diabetes Nutr. Metab. 1, 1l-19. 7 O’Dea, K., Traianedes, K., Ireland, P. et al. (1989) The effects of diets differing in fat, carbohydrate and fibre on carbohydrate and lipid metabolism in Type II diabetes. J. Am. Diet. Assoc. 89, 1076-1086. 8 Fontvieille, A.M., Acosta, M., Rizkalla, S.W. et al. (1988) A moderate switch from high to low glycaemic-index foods for 3 weeks improves the metabolic control of Type I (IDDM) diabetic subjects. Diabetes Nutr. Metab. 1, 139-143. 9 Nestler, J.E.. Gebhart, S.S.P. and Blackard, W.G. (1984) Failure of a midnocturnal insulin infusion to suppress the increased insulin need for breakfast in insulin-dependent diabetic patients. Diabetes 33, 266-270. 10 Vlachokosta, F.V., Piper, C.M., Cleason, R., Kinzel, L. and Kahn, CR. (1988) Dietary carbohydrate, a Big Mac, and insulin requirements in type I diabetes. Diabetes Care 11, 330-336. 11 Collier, G.R., Wolever. T.M.S., Jenkins, D.J.A. et al. ( 1987) Concurrent ingestion offat and reduction in starch content impairs carbohydrate tolerance to subsequent meals. Am. J. Clin. Nutr. 45, 963-969. 12 Wtirsch, P., Koellreutter, B., Haesler, E., Felber, P. and Golay, A. (1991) Metabolic effects of slow release starch in non-insulin dependent diabetic patients. Diabetes Nutr. Metab. 4, 195-199. 13 Wahlfeld, A.W. (1974) Triglyceride determination after enzymatic hydrolysis. In: H.V. Bergmeyer (Ed.), Methods of Enzymatic Analysis, Academic Press, New-York, pp. 1831-1835.
141
14 Allain, C.A. (1974) Enzymatic determination of total cholesterol. Clin. Chem. 20, 470-475. 15 Schweizer, T.F., Reimann, S. and Wtirsch, P. (1989) Definition and measurement of a starch digestion index and a study of factors determining starch digestion rates in foods. Lebensm.-Wiss Technol. 22, l-6. 16 Wtirsch, P.. Del Vedovo, S. and Koellreutter, B. (1986) Cell structure and starch nature as key determinants of the digestion rate of starch in legume. Am. J. Clin. Nutr. 43, 25-29. 17 Golay, A., Coulston, A.M., Hollenbeck. C.B., Kaiser, L..L.. Wtirsch, P. and Reaven, G.M. (1986) Comparison of metabolic effects of white beans processed into two different physical forms. Diabetes Care 9. 260-266. 18 Bornet. F.R.J.. Fontvieille, A.-M., Rizkalla, S. et al. (1989) Insulin and glycemic responses in healthy humans to native starches processed in different ways: correlation with in vitro a-amylase hydrolysis. Am. J. Clin. Nutr. 50. 315-323. 19 Helm, J., Bjdrk. I.. Drews, A. and Asp, N.G. (1986) A rapid method for the analysis of starch. Starch/Starke 38, 224-226. 20 Prosky, L., Asp, N.-G., Furda, I., De Vries, J.W.. Schweizer, T.F. and Harland, B. (1985) Determination of total dietary fiber in foods and food products: collaborative study. J. Assoc. Off. Anal. Chem. 68, 677-679. 21 Paul, A.A. and Southgate, D.A.T. (1978) McCance and Widdowson’s The Composition of Foods, 4th Edit., Elsevier/North Holland Biomedical Press, Amsterdam. 22 Wolever. T.M.S. and Jenkins, D.J.A. (1986) The use of the glycemic index in predicting the blood glucose response to mixed meals. Am. J. Clin. Nutr. 43, 167-172. 23 Hollenbeck, C.B., Coulston, A.M. and Reaven, G.M. (1986) To what extent does increased dietary fiber improve glucose and lipid metabolism in patients with non-insulin-dependent diabetes mellitus (NIDDM)? Am. J. Clin. Nutr. 43. 16-24. 24 Kinmonth,A.L., Angus,R.M., Jenkins, P.A., Smith, M.A. and Baum, J.D. ( 1982) Whole foods and increased dietary
25
26
27
28
29
30
31
32
33
34
fibre improve blood glucose control in diabetic children. Arch. Dis. Child. 57, 187-194. Roth, G. and Leitzmann, C. (1985) Long-term effect on human blood lipids of breakfast cereals rich in dietary fibre. Aktuelle Ernahr. 10, 106-109. Behall, K.M., Scholfield, D.J. and Canary, J.J. (1988) Effect oftype of starch on metabolic parameters in human subjects. Am. J. Clin. Nutr. 47, 428-432. Behall, A.M., Scholfield, D.J., Yuhaniak. I. and Canary, J. (1989) Diets containing high amylose vs amylopectin starch: effects on metabolic variables in human subjects. Am. J. Clin. Nutr. 49, 337-344. Gatti, E., Gatenazzo, G.. Camisaca. E.. Torri, A., Denegri, E. and Sirtori, CR. (1984) Effects of guarenriched pasta in the treatment of diabetes and hyperlipidemia. Ann. Nutr. Metab. 28, l-10. Cunningham, K.M. and Read, N.W. (1989) The effect of incorporating fat into different components of a meal on gastric emptying and postprandial blood glucose and insulin responses. Br. J. Nutr. 61. 285-290. Collier. G., McLean, A. and O’Dea, K. (1984) Effect of co-ingestion of fat on the metabolic responses to slowly and rapidly absorbed carbohydrates. Diabetologia 26, 50-54. Nestler, J.E., Barlascini, C.O., Clore. J.N. and Blackward. W.G. (1988) Absorption characteristics of breakfast determines insulin sensitivity and carbohydrate tolerance for lunch. Diabetes Care 11, 755-760. Jenkins, D.J.A., Wolever, T.M.S., Taylor, R.H. et al. (1982) Slow release carbohydrate improves second meal tolerance. Am. J. Clin. Nutr. 25, 1339-1346. Levine, A.S., Tallman, J.R., Grace, M.K., Parker, S.A., Billington, C.J. and Levitt, M.D. (1989) Effect of breakfast cereals on short-term food intake. Am. J. Clin. Nutr. 50, 1303-1307. Leathwood, P. and Pollet, P. (1988) Effects of slow release carbohydrates in the form of bean flakes on evolution of hunger and satiety in man. Appetite 10, l-l 1.
135
DIABET 00596
The effect of muesli or cornflakes at breakfast on carbohydrate metabolism in type 2 diabetic patients Alain
Golay
*, Brigitte
Koellreutter
Bloise ‘, Jean-Philippe 2, Donatella Pierre Wiirsch 2
Assal ’ and
’ Unite d%nseignement pour Diabktiques. Hbpital Cantonal Universitaire. Geneva, Switzerland and ‘NestlP Research Centre, Nestec Ltd., Vers-chew-les-Blanc. Lausanne, Switzerland (Received 17 April 1991) (Revision accepted 23 August 1991)
Summary Fourteen overweight insulin-treated type 2 diabetic patients ate a breakfast, consisting of either muesli (slow release starch: SRS) or cornflakes (fast release starch: FRS), in either case with milk (46 g carbohydrate), during two consecutive randomized crossover periods of two weeks. The rest of the diet remained unchanged. At the end of each period the patients underwent a glucose tolerance test after an overnight fast without their usual evening insulin injection. Both mean plasma glucose response curves were identical after the two dietary periods, but plasma insulin was significantly lower at zero ( - 17”/,, P < 0.05) and 2 h ( - 21x, P < 0.05) at the end of the muesli (SRS) period as compared to the cornflakes (FRS) period. The mean day-long plasma glucose level (four measurements) at the end of the muesli period was 21”/, (P= 0.023) lower than after the cornflakes period. These results show that switching, at breakfast only, from standard cereals to slow release starch cereals improves the carbohydrate metabolism of diabetic patients. In addition, the fact that diabetic patients could reduce their insulin requirement (P< 0.05) with concomitant reduction of their daily blood glucose level implies that sensitivity to insulin was improved by slow release starch foods consumed at breakfast.
Key words: NIDDM;
Breakfast;
Cereal; Diet; Insulin;
Current recommendations from the ADA for treatment of patients with diabetes includes the Correspondence to: P. Wiirsch, Nestle Research Centre, Nestec Ltd., Vers-chez-les-Blanc, CH-1000 Lausanne 26, Switzerland. Abstract presented at the 13 ICC Congress, Wien, Austria, May 28-31, 1990.
Sensitivity
use of high-carbohydrate, low fat diets in order to reduce the risk of atherosclerotic heart disease [ 11. However, it has been pointed out that an increase in the carbohydrate intake may lead to postprandial hyperglycemia, increased glycosuria, hyperinsulinemia, elevated fasting plasma triglyceride levels and reduced HDL cholesterol in NIDDs treated with diet alone or diet plus sulphonylureas [2]. A rich carbohydrate diet which
136
nonetheless produces blunted glycemic and insulinemic responses has therefore been recommended. Incorporation of low rather than high glycemic index foods into the diet improved the carbohydrate metabolism of diabetic patients and often decreased serum lipid concentrations [ 3-81. Diabetic patients require more insulin for a mixed meal at breakfast, than for an isocaloric mid-day or evening meal [ 9, IO]. However, it has been suggested that insulin sensitivity during the second meal may be related to the effect from the preceding glucose challenge. Furthermore, a high fat breakfast results in impairment of the glucose tolerance to a standard lunch [ 111. Thus, better insulin sensitivity following a slowly absorbable carbohydrate eaten at breakfast might promote enhanced tolerance for the carbohydrate intake at lunch, without producing an excessive blood glucose excursion after the breakfast. This type of reduction of blood glucose excursion was achieved after the consumption of a muesli containing raw rolled wheat and instant bean flakes compared with cornflakes [ 121. The aim of the present study was to determine whether two weeks of a diet including slow release starch at breakfast compared with a diet including conventional breakfast cereal, could improve glucose tolerance or insulin sensitivity in type 2 diabetic patients, with insulin requirement.
Subjects Fourteen adults with type 2 diabetes participated in the study (Table 1). All the patients were treated with insulin on average for 2.7 + 1.2 years and diet only. All the patients studied were NIDDM requiring insulin for optimal blood glucose control. They were on a regimen of two injections of intermediate insulin without short acting insulin. Doses ranged from 8 to 24 U in the morning and 4 to 22 U in the evening. None of them had long term complications of diabetes. The dose of insulin injected was adapted by the patients themselves when they felt the need, but they were advised to do it only once a week in each period.
I-ABLE
1
Physical characteristics
(n = 14, mean f SD)
62 fr Age (years) 83 t Weight (kg) 168 + Height (cm) 29 * BMI (kg/m’) Energy intake (kJ/day) 6926 k Carbohydrate (Y’,,daily energy intake) 45 f .~______~.._
11 146 7 5 2110 8.6
The subjects received dietary instructions during the two weeks of training in our diabetes teaching unit before starting the one week stabilisation period at home. The blood glucose monitoring was carried out at home, using a glucometer (Reflolux II M, Boehringer Mannheim, F.R.G.).
Study design The study was conducted on an outpatient basis. The experimental protocol was approved by the Ethics Committee of the Department of Medicine of Geneva University Hospital and the subjects gave their written informed consent to their participation in the study. All the patients selected were on insulin. The study consisted of one week of metabolic stabilisation (fasting plasma glucose below 10 mmol/l followed by two weeks of dietary trials separated by a one week washout interval. The order of the two periods was randomized. The patients were asked to maintain their usual personal diet throughout the study apart from the breakfasts and morning snacks, which were provided in individual portions. They were strictly followed by a dietitian during the whole study. Their diets were recorded and found to be very constant during their two two-week periods. They were asked to monitor their blood glucose fasting at 7 am and at 11 am, 2 pm and at 5 pm, on the last two days of each period. All patients underwent an oral glucose tolerance test (OGTT) after each dietary period. The patients fasted overnight and omitted their usual insulin injection the evening before the test and the morning of the test. During the 75-g OGTT,
137
blood samples were drawn at time 0, 30, 60, 120, 180 min and plasma glucose and insulin were measured. Plasma cholesterol, HDL-cholesterol, triglycerides and glycosylated hemoglobin were also measured at time 0. Plasma was separated immediately, frozen and subsequently analyzed for plasma glucose (Beckman Glucose Analyzer, Palo Alto, CA), and cholesterol [ 131, triglycerides [ 141, insulin (RIA Pharmacia Upsala, Sweden). The patients did not have significant antibodies to disturb the measurements.
Diets The breakfasts provided 46 g carbohydrate or roughly 27% of the daily carbohydrate intake (Table 2). The size of the meal was fixed prior to the study by testing it with diabetic patients. The carbohydrate distribution was comparable in both meals e.g. 26 g starch, 14.4 g sucrose, honey, dextrin and 5.6 g lactose. The energy content of the cornflakes meal was 10% lower due to lower protein and lipid content. The SRS breakfast was composed of 65 g muesli in 120 ml whole milk. The starchy ingredients in the muesli were selected from among those having a very low digestion index of the starch as described by Schweizer et al. [ 151 i.e., raw rolled wheat and white bean flakes [ 16-181.
TABLE
2
Composition
of the breakfasts SRS muesli + milk
Carbohydrate (g) Starch (g) Sugars (g) Protein (g) Lipid (g) Dietary fibre (g) Energy (kJ)
46 26 20 11 6.2 6.7 1200
FRS cornflakes + milk 46 26 20 I 4.4 1.2 1080
The other breakfast (FRS) was composed of 35 g cornflakes, 12g sugar and 120 ml whole milk. The sugar level was the same in both meals. The snack at 10 am during both test periods was composed of 20 g wholemeal bread and 20 g cottage cheese (protein: 4.8 g, carbohydrate: 11 g, lipid: 2.2 g, total energy: 350 kJ). The patients were free to double this amount but they had to keep it unchanged during the whole study. Starch and sucrose in the meals were analysed as described by Holm et al. [ 193, fibre by using the AOAC method [20] and the other components calculated from food tables [ 211. Caloric intakes were based on the patients’ dietary record made during the three days of the stabilisation period. Total energy intake was 6929 + 2110 kJ/day and carbohydrate accounting for 45 * 8.6%.
Statistical analysis Data are presented as means with standard error of the means. Area under the glucose and insulin was determined according to the method described by Wolever and Jenkins [ 221. Means were compared by the paired t-test. For the daily blood glucose levels and for the areas under the glucose and insulin curves, two way of variance (ANOVA) was used for repeated measurements. For this analysis, the two independent variables were time and treatment. If there is significant interaction between different treatment and time, the statistical significance among groups at any point was analyzed.
Results The patients’ weight remained stable during the study (beginning 82.5 f. 3.7 kg; end muesli period: 82.4 + 3.8 kg; end cornflakes period: 82.5 k 3.8 kg). The patients reduced their daily insulin dose during the muesli period from a mean of 28 + 4 U to 24 k 4 (P < 0.05) at the end of the period. The results of the glucose tolerance tests
138
0
60 TIME
120 (
180
minutes I
I _” 400
-
I ?
-
300
k__
J
I
0
60 TIME
1
120 ( minutes
180
I
Fig. I. Mean ( + SEM) plasma glucose and insulin reaponscs for I80 min after ingestion of 75 g glucose at the end of two weeks of consumption of an SRS muesli (-) or FRS cornflakes (---) in 14 diabetic patients. Statistical significances are between both diets. *: P < 0.05.
are shown in Fig. 1 and Table 3. Fasting plasma glucose concentrations as well as the glucose response profiles were identical at the end of each diet. However, the fasting insulin level was 17% lower after the muesli than after the cornflakes period (P < 0.05). The insulin response curve also tended to be lower, and significance was reached only at 120 min (P < 0.05). Mean blood glucose levels measured at four time points (7, 11 am and 2, 5 pm) at the end of each period are presented in Fig. 2. The mean blood glucose levels during the muesli period were always lower by 18, 14,29 and 20%, respectively than during the cornflakes period (mean of all values, - 2 1%, P = 0.023). During the control period the mean values were in between, and were significantly higher than those associated with the muesli diet at 5 pm (P < 0.05). The effect of treatment -type of breakfast - was neither dependent on the order of the breakfast, nor on the period. Mean glycosylated hemoglobin levels were similar after the two study periods and during the stabilization period., Thus, the reduction of the daily
TABLE 3 Metabolic parameters (mean 2 SEM)
after SRS-muesli and FRS-cornflakes
and the oral glucose tolerance test (OGTT) in 14 diabetic patients
P
FRS
SRS
_
Mean daily blood glucose level,’ (mM) Glycosylated hemoglobin” (‘I,, ) Cholesterol (mM ) HDO-cholesterol (mM 1 l‘riglycerldes (mhl 1
7.1 7.5 5.7 I.0 1.7
+ i f f *
0.7 0.5 0.4 0.06 0.3
9.7 7.1 5.4 1.0 1.8
* f * + +
0.9 0.4 0.3 0.06 0.4
< 0.02 NS NS NS NS
9.1 f 19.3 + 22.8 + 150 f 324 k 768 +
0.8 1.0 1.4 54 78 180
NS NS NS < 0.05 NS NS
OGI-I
fasting blood glucose (mM) max. glucose (mM ) Incremental area under glucose curve (mM. 3 h) fasting insulin (PM) max. insulin (PM) arca under insulin curve (PM. 3 h) ,’ Before the study: X.8 k 0.60, P < 0.05 compared
8.9 k 0.6 19.4 * I.0 21.7 k 2.0 126 2 48 268 & 53 660 k 132
with SRS; h before the study: 7.9 k 0.4.
139
2
10
E
-3 ,......... I . . . . . . . . . . . . .
I . .
*;
r /.....
. . . . . . . . . .“>‘....._**_
I
_......a
* T**
z **
8 3 -I0
i
5
t
_ 07.00
.
_
,
_
.
_
14.00
11.00 TIME
.
( hours
_
.
_ 17.00
)
Fig. 2. Day-long mean ( i SEM) plasma glucose level at the end ofthe 2 weeks muesli ( + ) and cornflakes (0) breakfasts. *: P < 0.05; **: P < 0.02.
mean blood glucose level during the muesli period was not reflected by a decrease in glycosylated hemoglobin, probably because of the brevity of the study period. Total cholesterol, HDL-cholesterol and total triglycerides remained unchanged throughout the study. The patients filled in a questionnaire at the end of each period. Ten out of the fourteen subjects found that the muesli meal appeased their hunger for longer than the cornflakes. Nobody reported this advantage with the cornflakes.
Discussion This study shows that breakfast containing slow release starch cereals appears to improve glycemic control and to reduce insulin requirement in type 2 diabetic patients. Modem dietary recommendations for increased carbohydrate, reduced fat intake for diabetic patients specify that the carbohydrate-rich food should be of the ‘unrefined’ type. However, several recent comparative dietary studies using similar fibre contents suggest that fibre per se is not the major determinant of improved metabolic control [23]. When the foods were selected according to the criteria of the low amplitude of glycemic response, significant changes in carbohydrate metabolism were observed. Extensive substitution of the diet by low glycemic index foods such as whole grain
bread, leguminous seeds, parboiled rice and pasta was very beneficial to the glucose control of diabetic patients [ 3-7,241. Even modest dietary changes from bread and potato to rice, biscuits and pasta for three weeks in IDDM subjects resulted in reduced insulin requirement, better glucose control and lower concentrations of triglycerides [ 81. Brand et al. [3] obtained an improvement of the glycemic control in NIDDM after 12 weeks on a reduced glycemic index diet. The practical nature of our study design allowed the subjects to take major responsibility for their own diabetic control. Their degree of glycemic control may therefore be assumed to be representative of insulin treated type 2 diabetic out-patients. Although the selection of the low or high glycemic food was limited to breakfast, a significant difference in the glucose metabolism was achieved. During the low glycemic index breakfast period (raw cereal flakes and bean flakes), both insulin requirement and daily blood glucose were lower than during the high glycemic index breakfast period (cornflakes). No difference in basal glucose level was observed on the day of the glucose tolerance test, but the basal insulin level was 17% lower. This may be because the patients did not inject their intermediate insulin on the evening before the test. After two weeks of treatment, the glucose responses to the OGTT were similar following both types of diet. However, the insulin response to the OGTT tended to be lower after the SRS and the difference was significant at baseline and at 2 h. The association of reduced insulin response with a similar glucose response implies improved sensitivity to insulin during the glucose challenge performed after the SRS period. The difference between the diets was insufficient to modify the blood lipid profile. Roth et al. [25], who tested the effect of high fibre cereals against cornflakes at breakfast on 74 healthy adults for 12 weeks observed that the high fibre cereals did not significantly improve the blood cholesterol and triglycerides levels, but increased the HDL-cholesterol.
140
High amylose meals are associated with lower glucose and insulin responses in healthy volunteers than high amylopectin meals [26]. Comparison over a 5-week period of diet containing two thirds of the carbohydrate either as high amylose starch or high amylopectin starch, showed that amylose was associated with a significant reduction in triglyceride and cholesterol levels, and a reduction of the insulin response 1 h after a glucose load [ 271. Similar lipid differences were also observed in NIDDM subjects, between high rice and high potato diets [4] or high pasta and high pasta-guar diets [28]. These observations suggest that blood lipids may decrease if enough of the dietary carbohydrate in the diet is of the slow release type. The breakfasts tested in the study, contained little fat (less than 20% of total energy). High carbohydrate meals produce higher postprandial glucose levels than meals containing more fat [30], because fat delays gastric emptying and increases insulin response [29]. However, Collier et al. [ 111, have shown that the simultaneous ingestion of fat and carbohydrate in a meal impairs glucose tolerance to subsequent meals. As our study design precluded modification to the remainder of the diet, we tried to avoid an unfavourable carry-over effect. It has also been shown, that breakfast including low glycemicindex food reduces the glycemic excursion of the next carbohydrate meal [31,32]. All these elements may have contributed to the metabolic differences observed. The muesli breakfast also appeased hunger for longer than cornflakes. This has already been reported for high-fibre cereal breakfasts of reduced glycemic index [33] and beans [34]. In conclusion, breakfast including slow release starch cereals appears to improve glycemic control and to reduce insulin requirement in NTDDM patients. The concomitant decrease of blood glucose and insulin requirement implies that sensitivity to insulin has been improved.
References 1 American Diabetes Association Nutritional recommendations and principles for individuals with diabetes mellitus: 1986, 1987. Diabetes Care 10, 126-132. 2 Coulston, A.M., Hollenbeck, C.B., Swislock, A.L.M. and Reaven, G.M. (1989) Persistence of hypertriglyceridemic effect of low-fat high carbohydrate diets in NIDDM patients. Diabetes Care 12, 94-101. 3 Brand, J.C., Colagiuri, S., Crossman, S., Allen, A., Roberts, D.C.K. and Stewart Truswell, A. (1991) Lowglycemic index foods improve long-term glycemic control in NIDDM patients. Diabetes Care 14, 95-101. 4 Andersen, E., Hellstrom, P., Karlander, S.G. and Hellstrom, K. (1984) Effects of rice-rich versus a potatorich diet on glucose lipoprotein and cholesterol metabolism in non-insulin-dependent diabetics. Am. J. Clin. Nutr. 39, 598-606. 5 Jenkins, D.J.A., Wolever, T.M.S., Buckley, G. et al. (1988) Low-glycemic-index starchy foods in the diabetic diet. Am. J. Clin. Nutr. 48, 248-254. 6 Collier, G.R., Guidici, S., Kalmusky, J. et al. (1988) Low glycaemic index starchy foods improve glucose control and lower serum cholesterol in diabetic children. Diabetes Nutr. Metab. 1, 1l-19. 7 O’Dea, K., Traianedes, K., Ireland, P. et al. (1989) The effects of diets differing in fat, carbohydrate and fibre on carbohydrate and lipid metabolism in Type II diabetes. J. Am. Diet. Assoc. 89, 1076-1086. 8 Fontvieille, A.M., Acosta, M., Rizkalla, S.W. et al. (1988) A moderate switch from high to low glycaemic-index foods for 3 weeks improves the metabolic control of Type I (IDDM) diabetic subjects. Diabetes Nutr. Metab. 1, 139-143. 9 Nestler, J.E.. Gebhart, S.S.P. and Blackard, W.G. (1984) Failure of a midnocturnal insulin infusion to suppress the increased insulin need for breakfast in insulin-dependent diabetic patients. Diabetes 33, 266-270. 10 Vlachokosta, F.V., Piper, C.M., Cleason, R., Kinzel, L. and Kahn, CR. (1988) Dietary carbohydrate, a Big Mac, and insulin requirements in type I diabetes. Diabetes Care 11, 330-336. 11 Collier, G.R., Wolever. T.M.S., Jenkins, D.J.A. et al. ( 1987) Concurrent ingestion offat and reduction in starch content impairs carbohydrate tolerance to subsequent meals. Am. J. Clin. Nutr. 45, 963-969. 12 Wtirsch, P., Koellreutter, B., Haesler, E., Felber, P. and Golay, A. (1991) Metabolic effects of slow release starch in non-insulin dependent diabetic patients. Diabetes Nutr. Metab. 4, 195-199. 13 Wahlfeld, A.W. (1974) Triglyceride determination after enzymatic hydrolysis. In: H.V. Bergmeyer (Ed.), Methods of Enzymatic Analysis, Academic Press, New-York, pp. 1831-1835.
141
14 Allain, C.A. (1974) Enzymatic determination of total cholesterol. Clin. Chem. 20, 470-475. 15 Schweizer, T.F., Reimann, S. and Wtirsch, P. (1989) Definition and measurement of a starch digestion index and a study of factors determining starch digestion rates in foods. Lebensm.-Wiss Technol. 22, l-6. 16 Wtirsch, P.. Del Vedovo, S. and Koellreutter, B. (1986) Cell structure and starch nature as key determinants of the digestion rate of starch in legume. Am. J. Clin. Nutr. 43, 25-29. 17 Golay, A., Coulston, A.M., Hollenbeck. C.B., Kaiser, L..L.. Wtirsch, P. and Reaven, G.M. (1986) Comparison of metabolic effects of white beans processed into two different physical forms. Diabetes Care 9. 260-266. 18 Bornet. F.R.J.. Fontvieille, A.-M., Rizkalla, S. et al. (1989) Insulin and glycemic responses in healthy humans to native starches processed in different ways: correlation with in vitro a-amylase hydrolysis. Am. J. Clin. Nutr. 50. 315-323. 19 Helm, J., Bjdrk. I.. Drews, A. and Asp, N.G. (1986) A rapid method for the analysis of starch. Starch/Starke 38, 224-226. 20 Prosky, L., Asp, N.-G., Furda, I., De Vries, J.W.. Schweizer, T.F. and Harland, B. (1985) Determination of total dietary fiber in foods and food products: collaborative study. J. Assoc. Off. Anal. Chem. 68, 677-679. 21 Paul, A.A. and Southgate, D.A.T. (1978) McCance and Widdowson’s The Composition of Foods, 4th Edit., Elsevier/North Holland Biomedical Press, Amsterdam. 22 Wolever. T.M.S. and Jenkins, D.J.A. (1986) The use of the glycemic index in predicting the blood glucose response to mixed meals. Am. J. Clin. Nutr. 43, 167-172. 23 Hollenbeck, C.B., Coulston, A.M. and Reaven, G.M. (1986) To what extent does increased dietary fiber improve glucose and lipid metabolism in patients with non-insulin-dependent diabetes mellitus (NIDDM)? Am. J. Clin. Nutr. 43. 16-24. 24 Kinmonth,A.L., Angus,R.M., Jenkins, P.A., Smith, M.A. and Baum, J.D. ( 1982) Whole foods and increased dietary
25
26
27
28
29
30
31
32
33
34
fibre improve blood glucose control in diabetic children. Arch. Dis. Child. 57, 187-194. Roth, G. and Leitzmann, C. (1985) Long-term effect on human blood lipids of breakfast cereals rich in dietary fibre. Aktuelle Ernahr. 10, 106-109. Behall, K.M., Scholfield, D.J. and Canary, J.J. (1988) Effect oftype of starch on metabolic parameters in human subjects. Am. J. Clin. Nutr. 47, 428-432. Behall, A.M., Scholfield, D.J., Yuhaniak. I. and Canary, J. (1989) Diets containing high amylose vs amylopectin starch: effects on metabolic variables in human subjects. Am. J. Clin. Nutr. 49, 337-344. Gatti, E., Gatenazzo, G.. Camisaca. E.. Torri, A., Denegri, E. and Sirtori, CR. (1984) Effects of guarenriched pasta in the treatment of diabetes and hyperlipidemia. Ann. Nutr. Metab. 28, l-10. Cunningham, K.M. and Read, N.W. (1989) The effect of incorporating fat into different components of a meal on gastric emptying and postprandial blood glucose and insulin responses. Br. J. Nutr. 61. 285-290. Collier. G., McLean, A. and O’Dea, K. (1984) Effect of co-ingestion of fat on the metabolic responses to slowly and rapidly absorbed carbohydrates. Diabetologia 26, 50-54. Nestler, J.E., Barlascini, C.O., Clore. J.N. and Blackward. W.G. (1988) Absorption characteristics of breakfast determines insulin sensitivity and carbohydrate tolerance for lunch. Diabetes Care 11, 755-760. Jenkins, D.J.A., Wolever, T.M.S., Taylor, R.H. et al. (1982) Slow release carbohydrate improves second meal tolerance. Am. J. Clin. Nutr. 25, 1339-1346. Levine, A.S., Tallman, J.R., Grace, M.K., Parker, S.A., Billington, C.J. and Levitt, M.D. (1989) Effect of breakfast cereals on short-term food intake. Am. J. Clin. Nutr. 50, 1303-1307. Leathwood, P. and Pollet, P. (1988) Effects of slow release carbohydrates in the form of bean flakes on evolution of hunger and satiety in man. Appetite 10, l-l 1.
