Study group

Thirty-two patients with type 1 diabetes were recruited from the Department of Internal Medicine and Diabetology, the Poznan University of Medical Science, Poland. Inclusion criteria included: 1/type 1 diabetes, age above 18 years, 2/intensive insulin therapy treatment using a personal insulin pump, 3/body mass index below 30 kg/m2, 4/HbA1c < 9%, 5/a written consent to participate in the study. Exclusion criteria included: 1/pregnancy, 2/other types of diabetes, 3/eating disorders, 4/food allergies or intolerance to the ingredients of a standardized meal, 5/history of celiac disease, 6/autonomic neuropathy, including gastroparesis, 7/treatment using a personal insulin pump for <3 months.

Body composition analysis was performed using bioelectrical impedance camera BODY COMPOSITION ANALYZER BC-418 MA’s TANITA. Specifically, fat content and free-fat mass was examined.

Moreover, every patient had the following laboratory tests performed: 1/glycated hemoglobin (in whole blood, HbA1c), 2/high-density lipoprotein (HDL), 3/low-density lipoprotein (LDL), 4/total cholesterol (TC), 5/triglycerides.

HbA1c was determined by turbidimetric inhibition immunoassay on Cobas 6000 analyzer (Roche Diagnostics, Basel, Switzerland). Other measurements were performed using enzymatic assayson Cobas 6000 analyzer.

All participants were treated with intensive insulin therapy using a personal insulin pump with a bolus calculator function.

Before starting the study, to minimize the influence of the basal rate on blood glucose levels during the tests, each patient tested a basal rate between 1 and 5 pm. In a situation where a nonoptimal basal rate, during this time, caused fluctuations in blood glucose values, the appropriate modification was performed under medical supervision. In the days of the test meal consumption, the basal rate was not changed.

Study protocol

The Bioethics Committee of Poznan University of Medical Science in Poland approved the clinical study (ethical approval no. 198/18 1.02.2018). The study was designed as a randomized, single-blind crossover study.

Each participant of the study consumed two standardized test meals consisting of long-grain white rice. One test meal was freshly prepared and served immediately after preparation. Another test meal after preparation was cooled for 24 h at 4 °C and then reheated before serving to the patients.

A test meal on a test day could be served only if the patient had no episode of hypoglycemia within 24 h before the study. Before administering the test meal, the participant’s glucose target had to be in the range of 3.9–10 mmol/L. Test meals were always served at the same time, at 2 pm. The time at which the meal was consumed was set at 10 min. The interval between a last meal and a standardized test meal was 5 h. During this time, subjects were only allowed to drink water. Participants were also instructed to avoid unusual vigorous physical activity beginning on the day before each test. Study participants were asked not to exercise before the study until the completion of the test meal.

The patients always consumed the study test meals in the same room and similar circumstances. The patient did not know if the meal consisted of freshly cooked rice or previously chilled. To ensure blindness, the previously cooled test meal was heated to the same temperature as the fresh meal.

10 min before the test meal, patients administered the insulin bolus (Lispro/Aspart) according to carbohydrates exchange factor, correction factor and insulin sensitivity factor. The insulin dose was calculated based on the bolus calculator programmed into the pump. There should be no active insulin left from the previous bolus before serving test meals. All subjects presented stable glycemia, confirmed by the FreeStyle Libre flash glucose monitoring system, before insulin injection and meal consumption.

Test meals

Long-grain white rice was used for the study. The test meals’ preparation and service procedures were standardized. The test meals were prepared by boiling in water using a Silver Crest induction cooking plate model SIKP 2000 E2.

70 grams of dry product and 280 ml of water were used to prepare the test meal. The rice was dipped into boiling water and cooked for 18 min.

The test meal consisted of 200 g of rice (containing 46 g of carbohydrates) and 100 g of tomato sauce (4 g of carbohydrates).

The tomato sauce did not contain any seasoning. In total, the entire meal contained 50 grams of carbohydrates.

A pre-weighed portion of rice, after preparation, was cooled at 4 °C for 24 h in a refrigerator. After cooling, a portion of rice (200 grams) was reheated by immersion in 250 ml hot water for 3 min.

The rice sample was analyzed for total energy, carbohydrate, protein, fat, ash, total starch, fiber, and water content. Both freshly cooked rice and cooled rice were examined for resistant starch content. Protein content was analyzed using the Dumal method [11]. Fat content was investigated using nuclear magnetic resonance [12]. Starch content was analyzed using the polarimetric method [13]. Ash content was examined using a direct/dry method [14]. Carbohydrate content was determined by difference based on following formula: % Total carbohydrate = [100 − %(Protein + Fat + Moisture + Ash + Fiber)] [15]. Fiber content was analyzed using an enzymatic method [16]. Water content was determined using the oven method [17]. Resistant starch content was analyzed in fresh and cooled test rice (after reheating) using the AOAC 2002.02 method [18]. All analyses, except carbohydrate content, were performed twofold. Arithmetic means of two valuesthat were obtained from the analysis were used as the results.

Glucose measurements

The postprandial glycemia was assessed over 3 h using the FreeStyle Libre flash glucose monitoring system. Measurements were taken at 5 min intervals. FreeStyle Libre was applied at least 2 days before the first test meal.

Dietary intake at breakfast before test meals

Patients were instructed to eat the same type of breakfast during all test days to minimize the effect of the first meal on the blood glucose value. Subsequently, the breakfast consumption data before the test meals were collected and analyzed for total energy, carbohydrate, protein, fat and dietary fiber. We used “Dietician 2014” software with the added Polish food database.

Hypoglycemic episodes

In case of the episode of hypoglycaemia during the test, the experiment was stopped and the patients were asked to consume 15–20 grams of quickly absorbed sugar. Then, the blood glucose measurement was repeated for the next 15 min according to the recommendations of Diabetes Poland. The blood glucose readings below 3.9 mmol/L registered by the FreeStyle Libre sensor or symptomatic hypoglycemia episodes were verified by the measurement of capillary blood glucose using an OptiumXido Abbott Diabetes Care glucose meter. The time of the event occurrence was recorded.

The organoleptic assessment

During the study, each participant completed an organoleptic evaluation questionnaire that assessed the taste, visual appeal, smell, and consistency of the test meal.

Hunger feeling, satiety, and desire to eat, assessment

A degree of hunger feeling, satiety and desire to eat was assessed using a proprietary questionnaire based on the Visual Analogue Scale. Participants of the study assessed the intensity of perception of a given trait in numerical values on a scale from 0 to 10. The questionnaire was completed before consuming the test meal and 30, 60, 120, and 180 min after the meal.

Statistical analysis

Data were analyzed with Statistica PL version 13 (StatSoft, Inc., Tulsa, OK, USA) and MedCalc Statistical Software version 19.3.1 (MedCalc Software Ltd, Ostend, Belgium; https://www.medcalc.org; 2020) for graphical presentation of results. Results are presented as median values (IQR, interquartile range) for continuous variables or numbers and percentages for nominal variables. The normality of continuous variables was tested using Shapiro–Wilk test. Normally distributed variables were compared using paired t-test and for continuous variables not meeting its assumptions the Wilcoxon signed-rank test was used.