797-P: Impact of Carbohydrate Counting on Glycaemic Control in People with Type 1 and Type 2 Diabetes on Intensified Insulin Therapy

This study involved 154 subjects with type 1 and type 2 diabetes (84.6% with type 1 diabetes, age 46.6 ±13.6 years; 52% female, A1c 8.3 ± 0.9 %) who were undergoing intensified insulin therapy. Participants completed a questionnaire (SMART) that measures the accuracy of estimating carbohydrates in 27 foods or in meals, using the multiple-choice method. At the same time, glucose data collected by a flash glucose monitoring system were analysed in the two weeks prior to completion of the SMART questionnaire. The level of estimation error regarding carbohydrates correlated significantly with glucose variability (CV; r=-0.29; p<0.001), range of glucose values (r=0.33, p<.001) and maximum glucose values (r=0.32, p<.001). These significant effects remained even after adjustment for HbA1c. Interestingly, there was no significant correlation between estimation error and mean glucose values (r=0.012, p=0.882), time in range (r=-0.018; p=0.815) or percentage of hyperglycemic (>180 mg/dl; r=0.015, p=0.843) or hypoglycemic phases (<70 mg/dl; r=0.005; p=0.953). The negative correlation coefficient indicates that underestimation of the amount of carbohydrates in a meal is more significant in terms of glucose variability than overestimation of the amount of carbohydrates. The results suggest that a good estimation ability regarding carbohydrate content is associated with lower glucose variability but does not necessarily affect glucose levels.