Evaluation of the New Software Program Degifxl4 in the Determination of the Glycaemic Indices of Foodstuffs

Aims: There is no standardized protocol for measuring glycemic index (GI) that takes time-of-day eff ects into account. The software DegifXL2 and Medtronic-Minimed's CGMS TM and Solutions TM , makes the GI calculation at breakfast and dinner time possible. The aim of this study was to assess the enhanced data processing software (DegifXL4) enabling the GI calculation at breakfast, lunch, afternoon snack and dinner times. Methods: The glucose levels of 20 healthy volunteers were monitored after they consumed either 50 g of glucose or one of ten alternative foodstuff s either for breakfast and dinner or for lunch or snack. Within the 9-day test period, 10 such meals were monitored in 3 replicates for each volunteer. Specifi cally, CGMS TM was used to monitor plasma glucose levels at 5-minute intervals for a period of 120 min following the foodstuff ingestion. Results: Using the enhanced spreadsheed DegifXL 4, a total of 640 profi les were obtained and 491 (77 %) accomplished the criteria for further processing. The percentage of successful tests in each foodstuff varied from 57 to 87 %. Conclusions: The use of the new software DegifXL4 off ers accurate GI estimates for foodstuff s eaten for breakfast, lunch, snacks and dinners in three replicates. In combination with the CGMS Solutions Software TM is DegifXL4 an enhanced effi cient and comfortable way to routinely measure GI values.


INTRODUCTION
The glycaemic index of food (GI) is conventionally defi ned as the ratio of the area under the fi ngerprick 7-point glycaemic curve (AUC) of the tested food to the AUC of the standard (glucose); both areas are calculated for a 120 min-interval after the food ingestion.Within this period, the commercially available continuous glucose monitoring system (CGMS TM ) gives a total of 25 values of an individual's interstitial fl uid glucose concentration (ISFG) 1 .There is no signifi cant diff erence between the GI values obtained by the CGMS TM and values obtained by more time-consuming conventional methods, and both methods have high variability 1 .Using the software DegifXL2 and Medtronic-Minimed's CGMS TM and Solutions TM , the GI calculation at breakfast and dinner times is possible 2 .To enhance the capacity of the GIs investigating centre, DegifXL was adapted to enable the detemination of the GI also at lunch-and snack-times.The aim of this study was to assess the new data processing software (DegifXL4) enabling the GI calculation at breakfast, lunch, afternoon snacks and dinners.

MATERIAL AND METHODS
The procedures used in this study were in accordance with the Helsinki Declaration of 1975 as revised in 2000, and were approved by the local Ethics Committee.
Twenty-fi ve healthy volunteers entered the study and 20 aged 21.94 ± 1.39 y (mean ± SE), 7 men, 13 women, completed it.Informed written consent was obtained from all volunteers.
The tested software program DegifXL4 is based on the principles of the program DegifXL2.Both programs were developed at the Faculty of Medicine, Palacký University, Olomouc, Czech Republic 2 , using Microsoft Excel 2000 (Microsoft Corporation, Redmond, WA, USA).To use this program, the following is required: 1) A team of trained investigators (comprising physicians, educators, dietitians, lab workers, IT support workers); 2) Trained testers who are familiar with the use of CGMS TM and the terms of the study protocol (at least one tester is required); 3) Exact portions of foods to be tested, each containing 50 g of absorbable carbohydrates; 4) The necessary hardware: a PC, CGMS TM , glucometers; 5) The necessary software (MS Excel 2000, CGMS Solutions Software TM and DegifXL4).

Test period design
In this study, each volunteer consumed 50 g of glucose or one of ten alternative foodstuff s for breakfast and dinner, for lunch and for afternoon snacks.Exceptionally, juice or chips were consumed, if other foods were not available.Thirty-two tests were performed on each of the 20 volunteers; the test period began in the evening of the fi rst day and continued for the following 8 days, one test in the morning (breakfast) , at noon (lunch), one test in the afternoon (snack) and one test in the evening (dinner).
The subjects were encouraged to consume one portion of the test meal within 30 minutes.During the day they were not allowed to eat any other meal.They were asked to fast for 4 hours before breakfast, lunch, snack and dinner tests, respectively.All meals were designed for scheduled tests only.

CGMS TM operation
The operations are similar to those in DegifXL2 as was described elsewhere 3 .The CGMS sensor was inserted before the fi rst test meal and remained inserted throughout the whole test period.At the beginning of each test meal, the tester entered the event 'FOOD' into the monitor.If this step was omitted, the meal start time had to be entered into 'DATA' at a later time, since tests lacking the meal start time can not be processed.

Transfer of data to the DegifXL spreadsheet
Subjects' ISFG data was exported from the CGMS Solutions Software TM 7310 v. 3.0C (Medtronic-Minimed, Northridge, CA, USA) at the end of the test period, which in this study was day 9.Each subject was identifi ed by a number or name.

Mathematical principles and formulas
The IAUC was determined using calculus (integration, as the sum of the trapezoid areas) according to the formula (Fig. 1): IAUC = ∑S i i=1,…,24, where t 2 G i : glucose concentration at a particular time; G 0 : starting glucose concentration; and Δt = 5 min.For G i < G 0 , S i = 0.
The GI calculation for each test food was carried out in every test subject separately according to the formula: The group-averaged GI for the test food can be calculated from all the individuals' calculated GIs.

Data processing in DegifXL
The current version of DegifXL consists of four spreadsheets, named SETUP, DATA, LIST OF PROBANDS, AND STATISTICS.Having opened the SETUP sheet of DegifXL (Fig. 2), the user has to fi ll in or revise the following data: -The total number of tests planned for one individual -The exact path to the directory ".fst" -The lower limit for ISIG (Input Signal of Glucose) in the CGMSTM.-The limit for missing ISFG values in one test -The list of tested foods (  , b).
The DATA spreadsheet (Fig. 3) imports data from a .fstfi le when the user clicks on the button [Import CGMS].The missing start times must be fi lled in manually, according to tester's protocol.Any events introduced into the CGMS monitor by mistake of the tester (e.g.event FOOD at lunch time when no test meal was served) have to be deleted.By clicking on the button [Resulting IAUC], the values of IAUC, which are calculated from the ISFG values of the individual tests, appear on the screen.The data of the next tester can then be added as described above.The table in the upper part of the DATA spreadsheet shows the number of successful tests with each type of food, the number of persons successfully completing all planned tests (in this study, 3 tests with each food), and the number of people completing one or two tests.
The spreadsheet PROBAND LIST (Fig. 4) shows an overview of the number of successful tests carried out with individual foods selected in this sheet and in the SETUP spreadsheet.The selection of probands (e.g., according to sex and/or diabetes type and/or other user-defi ned parameters) is performed by entering the fi lter parameter (to be designed by the user) into the row [Filter].

Assessment of DegifXL 4
The newly developed DegifXL4 software was asessed by processing the CGMS TM data obtained from 20 volunteers.The effi ciency of using the CGMS TM /DegifXL software was considered to be the percentage of successfully processed tests from the total number of tests.

RESULTS
To evaluate the software and protocols used in this study, we determined the number of tests performed from which we obtained usable data (i.e. the effi ciency of the test).Using the CGMS TM /DegifXL4 System, there were 491 successful tests (77 %) from the 640 tests performed in the group of 20 healthy volunteers (Table 1).Each of the 11 foods (except soup) was tested in 3 replicates in each of the 20 volunteers, i.e. 60 tests were performed with each food.However, only complete tests performed under standardized conditions were processed.The number of successfully processed tests for the particular foods ranged from 34 (56.7 %) to 52 (86.7 %).The test failures were tester-related (n = 47; i.e. 7 %) or CGMS TMrelated (n = 102; i.e. 16 %).
Conventional methods for calculating GI have been described and widely used, studying 10 healthy subjects on multiple occasions in the morning after a 10-14h overnight fast.Subjects are asked to do no unusually vigorous activities on the day before the test, to drink no alcohol and not to smoke for 24h before the test.After a fasting blood sample, subjects eat a test meal and have further blood samples at 15, 30, 45, 60, 90 and 120 minutes after starting to eat.Capillary blood is obtained by fi nger-prick and whole blood glucose determined with an automatic analyzer using the glucose oxidase method.Each test meal contains 50g available carbohydrate (total carbohydrate minus dietary fi ber).The GI is valid as a method of classifying the plasma glucose responses of high carbohydrate foods.To test low carbohydrate foods, meals containing smaller amounts of carbohydrate may be used, but the amount of reference food should be adjusted so that it contains the same amount of available carbohydrate as the reference food.A drink of the subject's choice is served with each test meal.The subject can chose to have 1-2 cups of water, coff ee or tea, with 30 ml 2 % milk per cup if desired.However, the drink chosen by the subject is the same for every test performed.Test meals are consumed within 10 minutes.Each subject conducts one trial of each test food and 3 trials of the reference food.The reference food can be anhydrous glucose or white brad.Typically the reference food trials are done at the beginning middle and end of the series of tests, with the order of the test foods randomized between the reference foods.If large numbers of foods are being tested, a reference food trial should be done for every 5-6 test foods (to ensure no changes in subject's glucose responses with time).There are many ways to calculate the AUC, and the method used aff ects the GI value obtained.Assuming that at times t 0 , t 1 , ... t n (here equalling 0, 15 ... 120 min, respectively) the blood glucose concentrations are G 0 , G 1 , ... G n , respectively.

x=1 AUC= A x n
Where A x = the AUC for the xth time interval and the xth time interval is the interval between times t (x-1) and t x .
At the Faculty of Medicine, Palacký University, Olomouc, new software DegifXL was developed and succesfully used to determinate GI at breakfast and dinner times 3 .Published GI tables 11 show group-averaged GIs only, and the large variability in the GIs is ascribed to various factors [12][13][14] .The terms "group-averaged GI" and "subject-related GI" were used in order to distinguish GIs obtained from a group versus a GI obtained from an individual.Determination of individual subject-related GI has improved our understanding of metabolic processes 9,[15][16] .It is also a useful concept in relation to physical exercise 17 and to occupational activities 18 .
In the present study, in order to increase the capacity of the investigational process, the enhanced software DegifXL4 involves not only breakfasts and dinners but also lunches and snacks [19][20] .Nevertheless, the accuracy of the GIs determined at lunch-snack-and dinner times remains to be established.As it was not easy to consume the demanded amount of dark chocolate within 10 minutes, the period of food intake was prolonged to 30 minutes.The Glucometer Advance System 21 was used to calibrate the CGMS TM .The CGMS TM sensor remained inserted for up to 9 days [22][23][24][25][26] without serious adverse reactions.
Clinically, it is important to evaluate the glycaemic profi les of individuals with diabetes in response to a given food/mixed meal in combination with a particular treatment [27][28][29][30][31][32][33] -i.e. the glycaemic response to a complex food such as pizza.
The strength of this new protocol is that it automates GI calculation for four meals a day taken in a group of about 20 or more volunteers in a test period of several days.The data can be analyzed in greater detail using a statistical program such as SPSS v. 14 (SPSS Inc., Chicago, USA).Use of DegifXL4 is quite simple.However, routine determination of GI should be carried out in specialized centers staff ed by qualifi ed professionals who have been properly trained 34 .
To summarize, the use of the new software DegifXL4 off ers accurate GI estimates for foodstuff s taken in three replicates for breakfasts, lunches, snacks and dinners.DegifXL4 in combination with the CGMS Solutions Software TM is an enhanced effi cient and comfortable way to routinely measure GI values.Further practice-oriented studies are necessary to fi nd the relation between the GI at diff erent meal times.

Fig. 1 .Fig. 2 .
Fig. 1.Calculating the incremental area under the curve (IAUC).G 0 : glucose concentration before meal intake; G i : glucose concentration at a particular time; Δt: time interval between estimations (in CGMS TM , always 5 min); S i : surface of the respective trapezoid.

Fig. 3 .
Fig. 3. Example of the DATA spreadsheet, which is generated by clicking on [Import CGMS].

Fig. 4 .Fig. 5 .
Fig. 4. Example of the PROBAND LIST spreadsheet.Probands are selected in the table on the right.

Table 1 )
and their abbreviations (Marks).-Test meal times for morning, noon, afternoon and evening tests into a separate table [Test meal time].-The number of the test meal in the test schedule must be input into [Order of meal in test].This function allows the averaged GI values to be calculated for specifi c foods from a given test.-Optional marks into [Other rank] when needed to select meals according to specifi c criteria (e.g., GI for fresh meals); the GI value is calculated if the mark is entered into the raw [Filter values] and may be found in the STATISTICS spreadsheet (Figs.5a

Table 1 .
Number of tests with individual foodstuff s which have reached the evaluation criteria and number of probands which have successfully accomplished 3, 2 or one tests.All tests that have not fullfi lled the evaluation criteria are automatically excluded from further processing.