Objective: To assess whether the addition of viscous fiber at an amount recommended by the US FDA to allow a ‘low saturated fat, cholesterol, soluble fiber and coronary heart disease’, health claim label on a food package (1.7 g psyllium) and/or fat (30 g sunflower oil and 3 g sodium propionate) to a pasta meal would affect gastric emptying, postprandial glucose, insulin and GLP-1 concentrations.
Design: Ten subjects participated in a two-by-two single blind randomized crossover study. Four meals containing 50 g of available carbohydrate were consumed: pasta with or without psyllium enrichment served with a tomato sauce with (520 kcal per meal) and without (240 kcal per meal) fat. Blood samples were taken for 240 min following the meal and all subjects consumed a buffet meal at the end of the study. Gastric empting was measured using the paracetamol absorption test. Blood was analysed for glucose, insulin, GLP-1. Visual analog scales were used to record feelings of hunger, pleasantness and nausea.
Results: The psyllium-enriched pasta had no significant effect on gastric emptying or the incremental area under the curve (IAUC) for GLP-1, insulin or glucose compared with the control pasta. The addition of polyunsaturated fat and sodium propionate significantly increased the IAUC for GLP-1 (P<0.001), delaying gastric emptying (P<0.002), and decreasing glucose (P<0.002).
Conclusions: A dose of 1.7 g psyllium did not evoke measurable effects on gastric emptying, postprandial GLP-1, insulin or glucose metabolism. However the addition of 30 g of oil and 3 g of sodium propionate to the pasta did reduce gastric emptying, increase GLP-1 and reduce glucose and insulin concentrations. While this short-term study may have implications in terms of reducing the risk of diabetes and improving coronary risk factor profiles the long term effects of these nutrients need to be studied.
Sponsorship: This study was supported by Kellogg Company.
The prevalence of overweight and obesity are increasing worldwide (WHO, 2001) and the incidence of diabetes increases concomitantly (Seidell, 2000). Satiating foods may help people to reduce energy intake by earlier termination of meals. For people with diabetes, the consumption of smaller meals may be beneficial since the consumption of smaller, more frequent meals has been associated with improvements in blood glucose management (Jenkins et al, 1989).
Although the mechanism is not fully understood, slowing carbohydrate absorption appears to influence the insulin sensitivity of adipose tissue in vitro and reduce free fatty acid output, factors which may influence in vivo insulin sensitivity (Frost et al, 1996, 1998) It is possible that the lowered insulin demand accompanying slowly absorbed carbohydrates may be significant since even short durations of hyperinsulinemia induce insulin resistance in healthy subjects (DeFronzo & Ferrannini, 1991).
The composition and physicochemical properties of foods affect satiation and postprandial metabolic responses. Studies have demonstrated that the addition of viscous, soluble fibers, like psyllium, to food slows carbohydrate absorption and improves glucose regulation (Jenkins & Jenkins 1985a; Wolever et al, 1991). The inclusion of short chain fatty acids, eg lactic, acetic acid and sodium propionate, to carbohydrate-based products reduces glycemia and insulin demand (Liljeberg et al, 1999a). The addition of butter to potatoes decreases postprandial glycemia (Gannon et al, 1993), as does the addition of fat to pasta (Normand et al, 2001).
The addition of fat to meals is known to slow gastric emptying rates (Liljeberg & Bjorck 1998; Normand et al, 2001). These effects of fat on postprandial glycemia appear to reflect changes in gastric emptying rate. Part of the improvement in postprandial responses observed with the addition of soluble fibers to foods is thought to reflect increased meal viscosity and slower gastric emptying (Jenkins & Jenkins, 1985b).
The addition of fat to a meal may similarly alleviate pancreatic insulin requirement by decreasing gastric emptying rate and slowing carbohydrate absorption.
The incretin hormones GIP (glucose dependednt insulinotropic factor) and GLP-1 (glucagon-like peptide 1) are thought to account for 100% of incretin activity. Both peptides have strong effects on gastric emptying and are thought to be insulinotropic. Infusion of exogenous GIP fails to stimulate insulin (Elahi, McAloon Dyke et al, 1994) thus in this study we concentrated on the effects of GLP-1.
Glucagon-like peptide 1 (GLP-1) is a hormone secreted from L cells of the intestine in response to carbohydrates, lipids and mixed meals. It has been shown to affect the regulation of appetite and food intake and to reduce the rate of entry of glucose into the circulation by a reduction in gastric emptying rate (Flint et al, 2001; Verdich et al, 2001). GLP-1 also stimulates insulin secretion, decreases glucagon secretion and may partially reverse the defect in non-insulin-mediated glucose uptake in older individuals with diabetes mellitus (Meneilly et al, 2001).
The first aim of this study was to test the hypothesis in healthy volunteers that the addition of viscous fiber to pasta improves postprandial glycemic and insulin responses primarily by decreasing carbohydrate absorption within the small intestine, while assessing the effect on gastric emptying rate or postprandial GLP-1 concentrations. Secondly, it was hypothesized that fat enrichment of a meal would improve postprandial glycemic and insulin responses by increasing GLP-1 secretion and slowing gastric emptying.
Ten subjects (six women and four men; age 33.8±3.5 y; body weight 74.2±20.3 kg: body mass index (BMI) 25.1±4.4 kg/m2) participated in the study. All subjects were healthy with no known medical condition as assessed by a medical screen prior to the study, which included a medical history, physical examination, standard blood tests, and ECG. The Hammersmith Hospitals research ethics committee approved the study and all volunteers gave informed written consent.
The study was conduced using a two by two single-blind randomized cross-over factorial design. All subjects were studied on four occasions, separated by at least a week. The meals consisted of freshly cooked noodles as previously described (Hoebler et al, 1998). Subjects consumed pasta meals with or without psyllium enrichment served with a warmed tomato sauce with or without fat (30 g sunflower oil and 3 g sodium propionate (Table 1). A subjective dose ranging study was undertaken in 4 of the volunteers prior to the start of the study in order to check that the amount of propionate, oil and tomato sauce that was being added to the pasta was palatable and realistic.
The day before the test days, each subject consumed his or her own standard diet. Participants were asked to standardize their exercise and to refrain from consuming alcohol the day before each test. The evening meal prior to the overnight fast was standardized throughout the study by supplying participants with a frozen meal. Volunteers were asked to fast for 12 h prior to the test meal.
A single meal was given following an overnight fast (Table 1). A cannula was placed in the antecubital vein and blood samples were taken at the following times: −20, −10, 0, 10, 20, 30, 40, 50, 60, 90, 120, 180 and 240 min. All meals were consumed within 5 min of the baseline 0 min sample. At 240 min subjects were offered a buffet meal where all items had been previously weighed. Prior to the study subjects were asked for likes and dislikes, and the buffet meal did not contain any item that the subject disliked. After the subjects had finished eating, the remaining food was weighed and the nutrient intake calculated.
Visual analog scales
Visual analog scales (VAS) were completed at times −20, 10, 60, 120, 180 and 240 min. These required the subjects to rate their feelings of fullness, hunger, pleasantness and nausea on separate scales. Each was rated 0–10 and ranged, for example, for fullness from ‘as hungry as it is possible to be’ to ‘as full as it is possible to be’.
Gastric emptying studies
Paracetamol was used to measure gastric emptying rates as described by Heading et al (1973)). The paracetamol (1.5 g) was dissolved in 100 ml of water and consumed halfway through the meal to ensure mixing with the meal. Plasma paracetamol levels were measured using an enzymatic colorimetric assay using the Olympus AU600 analyzer.
Plasma glucose was measured by using a glucose oxidase-based autoanalyzer (Technicon; Axon Bayer Diagnostic, Newbury, UK). Insulin and GLP-1 concentrations were measured using in-house radioimmunoassays as previously described (Kreymann et al, 1987). These assays were capable of detecting 2 pmol/l with a 95% confidence. All samples where included in one assay and analysed in duplicate after the first freeze–thaw.
Sample size was estimated from a previous acute study in healthy volunteers using a power of 80% and an alpha of 5% with an estimated difference in gastric emptying of 30%. This suggested a sample size of 10 paired results. All results are presented as mean±s.e.m. unless otherwise stated. The incremental area under the curve (IAUC) was calculated using the trapezoidal rule. Comparison of IAUC, peak (Cmax) postprandial plasma glucose, Cmax and time of maximum concentration (Tmax) of plasma paracetamol levels, and absolute differences between the food consumption were by ANOVA with post-hoc analysis only when significant. Visual analog scores were compared by the Wilcoxon signed ranks test.
Fasting plasma concentrations of glucose, insulin and GLP-1 are shown in Table 2. There was no significant difference in fasting values between any of the four test days.
Gastric emptying rate was significantly delayed with the addition of fat and sodium propionate to the tomato sauce but unaffected by the presence of psyllium in the pasta (Table 2, Figure 2). One subject vomited 10 min after consuming the meal containing fat and sodium propionate-enriched tomato sauce. On that occasion the test was stopped and the data set was excluded from the analysis.
The postprandial response of glucose, insulin and GLP-1, are described in Figure 1. The incremental areas for each meal (IAUC), Cmax (maximum concentration) and Tmax (time of maximum concentration) are given in Table 2. IAUC GLP-1 was not significantly different after the psyllium-enriched meal and Cmax was significantly lower (psyllium 35±6 vs control 43±5 pmol/l; P<0.02), suggesting a longer slower release of GLP-1 in the plasma. The addition of fat and sodium propionate to the test meals significantly increased IAUC GLP-1 (Table 2, Figure 1). GLP-1 Tmax was 35 min later and maximum concentration was also significantly higher (Table 2) after the addition of fat.
Although the addition of 1.7 g psyllium to the pasta had no apparent effect on gastric emptying rate or postprandial IAUC glucose or insulin responses, the combined addition of psyllium to the pasta and fat and sodium propionate to the tomato sauce significantly reduced postprandial IAUC glucose compared to the control meal (Table 2). This combined addition (psyllium, fat and sodium propionate) delayed the time to maximal GLP-1, glucose and insulin concentrations but not gastric emptying rate (Table 2).
The results of the visual analog scales for change in appetite, pleasantness and nausea are shown in Figure 3. A significant increase in nausea was reported after the two fat-containing meals (P<0.05).
There was no difference in the amount of energy (kcal) consumed at the buffet meal after any of the fixed pasta meals (control 1149±106 vs psyllium 1235±106 kcal vs control plus fat 1156±123 kcal vs psyllium plus fat 1121±151 kcals P=NS) despite the two-fold greater energy intake with the high-fat meals.
The study was conducted so that the primary source of glucose in the meals was derived from the pasta. The addition of sodium propionate and polyunsaturated oil had dramatic effects on postprandial events. There was a significant increase in GLP-1 concentration with a significant fall in gastric emptying rate and a decrease in postprandial glucose. These effects agree with the reports of decrease glucose and insulin and decrease in gastric emptying from the sodium propionate enrichment of bread (Liljeberg et al, 1999b). GLP-1 is known to have effects on gastric empting and may explain some of the effects on insulin and glucose. A confounding factor is the reported increase in feeling sick after the fat supplemented meal. It is possible that the increase in GLP-1 secretion and the resulting decrease in gastric emptying result in the increase in nausea. This fat-induced feeling of nausea seemed to be transient. Subjects consumed the same amount of energy at the buffet meal despite eating twice as many calories with the high-fat pasta meal they also reported no differences in the amount they thought they could eat, hunger or how pleasant it would be to eat at 240 min before the start of the buffet meal.
Although propionate is recognised to have an independent effect on glucose metabolism, acting as a gluconeogenic substrate it is unlikely that there is any background effect from propionate produced from fermentation of carbohydrate by intestinal flora (Wolever, Brighenti et al, 1989).
Soluble fibers, like psyllium, that hydrate rapidly and develop significant viscosity in vitro are proposed to reduce the rate of carbohydrate absorption (Jenkins & Jenkins 1985b) when these viscous fibers are intimately mixed with foods (Wolever et al, 1991). In our study we were unable to show any significant effect of psyllium enrichment. However the amount of psyllium contained in the meal was 1.7 g, the amount required in a food for the US FDA to allow a health claim label on a food package stating that the consumption of this food may reduce the risk of heart disease. This is far below the 5–10 g a meal that has been used in most other studies in people with diabetes where effects have been seen on blood glucose and insulin concentrations. Also there are few studies that report effects of psyllium on postprandial glucose concentrations in normal volunteers. Jarjis et al (1984) failed to report any effect of psyllium on the blood glucose or insulin concentrations in normal volunteers. There was no effect on gastric emptying or postprandial GLP-1 concentrations but the greater delay until maximum concentration (Tmax) and lowered maximal concentrations (Cmax) for glucose and insulin on the combined high-fat meal with psyllium-enriched pasta may suggest that psyllium does affect the intestinal milieu to influence carbohydrate digestion and glucose uptake from the small intestine with slower rates of gastric emptying, other factors may also be involved that were not measured in this study such as postprandial lipids. This is an interesting observation that deserves further attention to develop foods for individuals with diabetes mellitus.
The glycemic index (GI) was introduced to classify carbohydrate foods based on postprandial glycemic responses to a constant carbohydrate load. As a research tool, the glycemic index clearly identified the importance of botanical integrity (including particle size), amylose/amylopectin ratio of the starch, and the fructose/glucose content of the available carbohydrate. Recently, the Food and Agricultural Organization and the World Health Organization advocated the increased consumption of low-glycemic index foods. This study demonstrates a possible shortcoming of using the glycemic index in isolation as a public health tool since all four tests meals contained the same amount of available carbohydrate but it was the high-fat meal, with the highest energy content, that had the lowest glycemic response. These results indicate that the glycaemic index should not be used in isolation but in the context of the desired macronutrient profile of the diets. Consumer surveys consistently show that taste is more important than nutrition. The fact that the glycemic index and glycemic load of foods can be reduced by adding fat or substituting starch with fructose or other non-glucose-yielding carbohydrates that increase triglyceride concentrations reduces the public health value of the glycemic index. With increasing incidence of obesity and diabetes, a simpler and more effective public health goal may be to advocate energy balance and increased physical activity.
In conclusion it appears that the addition of 1.7 g psyllium to a normal portion of pasta did not affect gastric emptying rate or postprandial GLP-1, insulin, and glucose metabolism. However the addition of 30 g of oil and 3 g of sodium propionate to the pasta stimulated GLP-1 secretion, reduced gastric emptying rate and attenuated postprandial glucose changes, independently of changes in carbohydrate content of the meal.