Safety of New Algorithms for Premeal Insulin Boluses in High Glycaemic Index Meals in Persons with Type 1 Diabetes Mellitus Using Insulin Pumps

Aims: Consumption of glucose or foodstuff s with high glycaemic index (GI) in persons with type 1 diabetes mel-litus (PWD1) is a hot topic in present diabetology. The aim of our pilot prospective study was to assess the effi ciency of empirically suggested simple algorithms for premeal boluses in PWD1 using insulin pumps and continuous glucose monitoring (CGM). 2 , mean±SD) on insulin pumps Paradigm 522/722 with RT-CGMS sensors (Medtronic MiniMed, Northridge, CA) underwent a 12-week CGM. In one week, subjects consumed 50g of carbohydrates in eleven alternative meals (rice squares, dark chocolate, white bread, honey, glucose, ravioli with meat and Eidam cheese, mashed potatoes with fi sh fi ngers, apricot dumplings with butter, spa waffl es, spalta squares, and tomato soup with pasta) eaten for breakfasts, lunches, snacks and dinners in order to calculate their GI. The insulin boluses were adjusted according to empirically defi ned algorithms. Average glucose levels and daily insulin doses over three one-week periods (before testing, testing and after testing) were compared. IU/d, respectively) were similar. One-week consumption of high GI foodstuff s had only a negligeable eff ect on average glucose levels. Conclusion: The suggested algorithms for premeal insulin boluses appear to limit the risk of potential hyperglycae-mia resulting from intake of high GI foodstuff s.


INTRODUCTION
Glycemic index (GI) is a measure of the ability of a food to raise glucose levels after the food consumption.GI is defi ned by the incremental area under the blood glucose curve (AUC or IAUC) after the ingestion of a test food containing 50g of carbohydrates, expressed as the percentage of the AUC of a reference food containing equal amount of carbohydrates (generally glucose or white bread) 1 .
The foodstuff s with low GI are preferred in diabetic diets in order to avoid postprandial and postabsorption hyperglycemia and to achieve improvements in overall plasma glucose control [2][3][4][5] .On the other hand the advances in intensive insulin therapy with fast-acting insulin analogs 6,7 and integrated real-time continuous glucose monitoring system and insulin pump (Medtronic Mini-Med Paradigm REAL-Time System; Medtronic MiniMed, Inc., Northridge, CA, USA) 8 , allow more fl exible meal plans with the potential to revolutionise diabetic care.
However, to date we still lack data on the consequences of high GI foodstuff consumption on glucose control in PWD1 on insulin pump therapy.Even though the glycemic indexes in persons with diabetes type 1 or type 2 and in healthy persons appear to be identical 9 , it is useful to perform independent GI determination for both groups.There is the question as pure glucose taken in order to calculate the GI of foods is acceptable in PWD1.Therefore, the aim of our pilot study was to evaluate changes in average glucose levels and assess the eff ect of a one-week consumption of meals with higher GI on glucose control in PWD1 treated by an insulin pump and real-time glucose sensor.

MATERIALS AND METHODS
The study was conducted between January 2007 and June 2007 at the Faculty of Medicine and Dentistry, Palacký University, Olomouc, Czech Republic (as a part of project Degif -IGA NR 7825-3).

Subjects
Six PWD1 on insulin pump therapy (3 males, 3 females, aged 46.2±15.09years, diabetes duration 14.5±9.65 years, HbA1c 6.3±1.59%, BMI 23.6±1.67 kg/m 2 ) provided written informed consent with participation in this study, which was approved by the Local Ethics Committee and performed in accordance with the guidelines of the Helsinki Declaration on human experimentation.Subjects were involved in a 12-week continuous glucose monitoring using insulin pump Paradigm RT 522/722 and RT-CGMS sensors with transmitters, Medtronic MiniMed, Northridge, CA, USA.This insulin pump enables to read the actual glucose concentration on its display any time.(Figs. 1, 2) foodstuff s for snacks at 4 p.m.Each foodstuff was consumed in three replicates, except tomato soup with pasta, which was consumed twice only.
The defi ned meal plan including daily intake of nutrients and energy was the same as in a similar study with healthy persons 10,11 .The foodstuff s were selected by the following criteria: (1) carbohydrate content in each portion should be 50g ± 5 %; (2) daily energy intake ranging from 6000kJ to 9000 kJ; (3) carbohydrates, lipids and proteins should be in the ratio 5:1, 5:1,5; and (4) simple preparation of servings and reasonable costs making the meal plan fl exible.
In every subject, average glucose levels were evaluated and compared over the period of three weeks, comprising the week before, during, and after the GIs testing.
One hour before and immediately before each meal test, subjects were encouraged to read the plasma glucose concentration on the display of insulin pump and to adjust the insulin boluses according to the empirically defi ned algorithms (Table 1, Table 2).The subjects were also instructed to enter the boluses in a log-book and keep recordings about any hypoglycaemia, food intake, meal times, exercise, and level of stress.Glucometer Advance (Hypoguard, Woodbridge, UK) was applied twice a day to calibrate the sensor.10.0 -11.9 1 12.0 -13.9 2 ≥ 14.0 3   3 gives the results of CGM over the study period.Average glucose levels during the week before, during, and after GIs testing were similar.Average total daily

Study design
During the assigned week, subjects consumed eleven alternative meals, each serving containing 50g of carbohydrates (rice squares, 70% dark chocolate, rolls, honey, glucose, ravioli with meat and Eidam cheese, mashed potatoes with fi sh fi ngers, apricot dumplings with butter, spa waffl es, wheat squares, and tomato soup with pasta) eaten for breakfasts, lunches, snacks and dinners.Foodstuff s for breakfasts and dinners were consumed at 7 a.m. and 8 p.m., respectively, foodstuff s for lunches at 12 a.m. and Safety of new algorithms for premeal insulin boluses in high glycaemic index meals in persons with type 1 diabetes mellitus using insulin pumps     dose of insulin during the week before, during, and after GIs testing were similar as well (Table 3, Figs.3, 4).There were no disturbances in general condition and meal plan; no serious hypoglycaemia was registered.

DISCUSSION
This pilot study deals with six case reports focused on the potential hazards of consumption of pure glucose in order to investigate the glycaemic indexes of foods in PWD1.
Awareness of GIs of various foodstuff s is an important part of diabetes management, particularly in persons with type 1 diabetes mellitus.It has been suggested that mean GI values for foods are very similar in persons with type 1 and type 2 of diabetes mellitus [9][10][11][12][13][14][15][16][17][18] .
Accurate measurement of GI requires a standardized protocol; glucose concentrations are measured using whole blood or serum/plasma samples obtained from a vein, artery, or capillary taken at baseline and at 15-and 30-minute intervals for two hours after the food consumption.Most recent studies have utilized capillary blood to determine glucose values and calculate the AUC and, subsequently, the GI.Recently, a study conducted by Chlup 19,20 found that there was no signifi cant diff erence between GI determined using the continuous glucose monitoring (CGMS; Medtronic MiniMed, Northridge, CA) as compared to conventional methods.The long-lasting use of transcutaneous sensors had no serious adverse events [21][22][23][24] .
In our six case reports, the CGMS results show similar average glucose levels during the period of GIs testing (even after glucose and honey consumption) compared to week before and after GI testing.However, the intensive instruction of PWD1 carried out by professional persons should be emphasized 25 .
Based on these preliminary fi ndings we might assume that consumption of pure glucose by PWD1 involves only little risk for diabetes control.On the other hand, not only the average glycemia, but also glucose variability should be taken into consideration.Several studies have shown glucose variability is a risk factor for complications independent of HbA1c in PWD1 26,27 as well as in type 2 diabetes 28,29 .Study by Monnier et al. demonstrated that glucose fl uctuations during postprandial periods and, more generally, during glucose swings exhibited a more specifi c triggering eff ect on oxidative stress than chronic sustained hyperglycemia 30 .
Our results suggest that insulin boluses adjustment according to actual blood glucose values is necessary in order to maintain normal or nearly normal glucose control.The fi ndings also show that real-time continuous glucose monitoring helps PWD1 on insulin pump therapy to make more informed treatment decisions and supports their confi dence in diabetes management.
Hence, one-week consumption of foodstuff s with high GI does not seem to increase the risk of reduced diabetes control in informed PWD 1 using insulin pumps and sensors.Further studies are in progress.

Fig. 3 .
Fig. 3. Average total daily dose of insulin during the week before, during, and after GIs testing.

Fig. 4 .
Fig. 4. Average glucose levels during the week before, during, and after GIs testing.

Table 1 .
Algorithms for adjustment of insulin boluses 1 hr before meal test.Plasma glucose levels 1hr before a meal-test [mmol/l] Correction bolus [IU]

Table 2 .
Algorithms for adjustment of insulin boluses immediately before meal test.

Table 3 .
Characteristics of 6 PWD 1 and infl uence of GI testing on insulin doses and average glucose levels.