To evaluate glucose metrics in a large group of children with T1D using glucose sensors in the real life.
During a planned visit, all children aged <18y with T1D for more than 1y, wearing CGM, were recruited during January-May 2019 by 11 Italian paediatric centres. According to their diabetes management, patients were divided in four groups: MDI&isCGM, MDI&rtCGM, IP&isCGM, IP&rtCGM. Predictive Low Glucose Management and Hybrid Closed-loop systems were excluded. We analysed: Time in Range (TIR), Time Below the Range (TBR), Time Above the Range (TAR), Coefficient of Variation (CV), HbA1c. Kruskal-Wallis test was used to perform comparisons and variables were summarized by median and IQR.
Seventy percent of children were using CGM. 665 cases (51% males), median age 12y, median diabetes duration 5y, were analysed. Significant differences in glucose metrics were found between groups (Figure). The lowest median CV values were found in MDI&rtCGM, 36.2% (32.8-40.8) and in IP&rtCGM, 36.8% (34.0-39.9) and significantly lower than MDI&isCGM, 39.4% (37.1-43.4) and in IP&isCGM, 40.5% (37.4-45.1), p<0.001. Regardless from CGM, groups using IP had significantly lower HbA1c values (MDI: 60 mmol/mol, 51-66; IP: 56 mmol/mol, 50-62; p<0.001).
Among a large group of children and adolescents with an average good metabolic control, the best glucose metrics are achieved with the use of rtCGM and IP.