ABSTRACT

Introduction: There is no current data about the effects of non-nutritive sweeteners (NNS) about important factors, such as the energy intake, appetite and its relationship in people with diabetes when tasting sweet. It is highly relevant to compare the effects of NNS intake, such as, stevia (steviol glycosides) and sucralose, previous to a mixed food on glycemic response, insulin and plasmatic concentrations of Glucagon-like peptide type 1 (GLP-1) and ghrelin in subjects with type 2 diabetes mellitus (T2DM).

Objective: To compare the effects of non-nutritive sweeteners intake: stevia (steviol glyco-sides) and sucralose previous to mixed food on appetite, glycemia, insulin, ghrelin,incretin plasmatic concentrations GLP-1 in people with T2DM.

Methods: Seventeen subjects with T2DM were studied in 3 different moments and they received 3 treatments: pre-load of water or sucralose or stevia and then offered to consume mixed food as a test, which provided 332 Kcal and 75 grams of available carbohydrates. Blood samples were obtained to measure the dependent variables, glycemic and insulin at times -10, 0, 30, 60, 90, 120, 150 and 180 minutes and GLP-1 with ghrelin, at times -10, 0, 30, 90, and 180 minutes. The analogue visual scale questionnaires (VAS) was conducted every 30 minutes in order to obtain the results of the depend variables: appetite and wish of specific type of food in a subjective way; appetite, satiety, relax, wish to eat any food, craving for something sweet, craving for something salty, something tasty, something fatty. Through food provided ad libi-tum (objective appetite), were obtained the results of: energy, carbohydrates, proteins and lipid intakes. The statistical analysis applied included the Shapiro-Wilk's Normality test, repeated measures ANOVA to assess differences among treatments, Friedman's test followed by Wilcoxon test corrected by Bonferroni as needed. The degree of association between variables was conducted using the Pearson's or Spearman's correlation coefficient tests, as requested. A probability value p <0.05 was considered significant.

Protocol The experimental design was developed 3 times. Each intervention was performed separately for a minimum period of 7 days out and a maximum of 14 days, depending on the availability of the experimental subjects. Cross treatment design (crossover) each subject serves as its own control The first intervention corresponded to the control group, in order to maintain the ignorance of the type of sweetener they were receiving, between day 7 or maximum 14 days post-intervention the order of the test of the NNS was reversed. Therefore, during the second or third intervention they were given to drink water with sucralose or stevia respectively. For each intervention a check list was followed.

Subjects were asked to maintain their normal diet between study days and abstain from strenuous exercise and alcohol intake for 24 hours before each evaluation. The subjects attended with an 8-hour fast, without having ingested metformin or other medication, to the sampling room of the Department of Nutrition (Faculty of Medicine, University of Chile). The study was only begun if they had a capillary glycemia <140%, measured through a glucometer.

Then, they were asked to answer the analogue visual scale (VAS) questionnaire and an intravenous catheter was installed in the antecubital vein to take the blood samples at "time -10". Immediately, they were given to drink 60 ml of water or 60 ml of water with 48 mg of sucralose or an equivalent volume with 96 mg of stevia (steviol glycosides), which they had to consume in less than 2 minutes, and then answer to a question about the "sensation" that came with drinking this preparation, to make an estimate of your predisposition to the taste of the sweetener: sweet, bitter, metallic or other flavor.

After 10 minutes, at "time 0" immediately after the second blood sample, the individuals consumed a mixed test meal consisting of a cup of tea with three teaspoons of sugar, plus 126 grams of a homemade bread. , without fat, using the bread maker, brand; program 3, traditional bread.

The following blood samples were taken at times 30, 60, 90, 120, 150 and 180 minutes after "time 0". In total, 42 ml of blood was extracted, in each time the glycemia and insulin were determined (4 ml was taken at a time). Furthermore, only at time -10, 0, 30, 90 and 180 minutes, the value of GLP-1 and ghrelin was determined (at this time, 2 ml extra was extracted at a time).

In addition, VAS questionnaires were applied every 30 minutes, from "time -10" to 180 minutes for appetite determinations.

After completing the extraction of blood samples at 180 minutes and removing the intravenous catheter, the subjects were offered a meal for 30 minutes, which consisted of a variety of foods of high acceptability, previously weighed and fractionated.

Statistic analysis The normal distribution of the variables was determined by the Shapiro-Wilk test. The natural logarithm was applied to the parameters without normal distribution. The variables were expressed as mean ± standard deviation or median plus interquartile interval (Q1-Q3). The statistical significance between the differences in plasma glucose concentrations, area under the curve (AUC) of glycaemia, energy intake and macronutrients by treatment (preload water, sucralose and stevia), was evaluated through Anova of repeated samples . To analyze the differences between plasma concentrations and AUC of insulin and GLP-1 per treatment (preload water, sucralose and stevia), Friedman's test was used, followed by Wilcoxon and adjusted by Bonferroni.

To identify differences between the perception of sweet, bitter, and metallic taste by treatment (preload water, sucralose and stevia) Friedman test was used, followed by Wilcoxon and adjusted by Bonferroni.

To identify the relationships between glycemic responses and intake of food at will (energy intake and macronutrients by treatment, preload water, sucralose and stevia), the Pearson correlation factor was used, in contrast to insulinemic response, incretin GLP-1 and the visual analogue scale (VAS) and food intake at will per treatment Spearman's correlation factor was used.

The statistical analysis was performed with the SPSS 20.0 computer program (SPSS Inc., Chicago Illinois). A p <0.05 was considered statistically significant for all the analyzes.

Inclusion Criteria:

• Subjects with T2DM
• Treatment with metformin and / or diet
• Diabetes diagnosed more than 1 year and less than 10 years
• Glycosylated hemoglobin (HbA1c) less than 9%
• Body mass index (BMI) ≥25 kg / m2 and <39.9 kg / m2

Exclusion Criteria:

• Subjects who consume any other drug or dietary supplement that may interfere with appetite or satiety or their post-prandial glycemic and hormonal responses
• Subjects with illness; acute, cardio-vascular significant, psychological, neurological, renal, alcohol or drug abuse
• Subjects that have aversion or allergy to foods / sweeteners used in test meals.
• Subjects with eating disorders and who have had gastrointestinal surgery of type: bariatric surgery, gastrectomy, Whip-ple or intestinal resections

the participant has no knowledge of the intervention assignment