New metrics derived from continuous glucose monitoring systems for the prediction of cardiovascular risk factors: cross-sectional analysis in a large cohort of children and youths with type 1 diabetes.

Background: Cardiovascular diseases (CVD) are the leading cause of death in subjects with Type 1 Diabetes (T1D). Reducing cardiovascular risk factors (CVRFs) exposure in children and youths with T1D is critical for CVD prevention. Poor long-term glycaemic control, measured by HbA1c, had been recognised as the main factor affecting CVRFs profile. To date, the possible association between short-term glycaemic control and variability measured by continuous glucose monitoring (CGM) metrics and CVRFs has been explored by a limited number of studies with inconsistent results. Aim: To test the hypothesis that short-term glycaemic control and variability independently contribute to CVRFs exposure in children and adolescents with T1D. To verify this hypothesis, the relationship between CGM metrics and three main CVRFs (overweight, LDL cholesterol, blood pressure [BP]) was analysed in a large cohort of paediatric subjects with T1D. Subjects and Methods: Eight hundred and ninety-five children and youths (age range 9-18 years) with T1D were enrolled. Anthropometric, blood pressure and biochemical (HbA1c, lipid profile) parameters were collected. Metrics of glycaemic control and variability were calculated from four and two weeks of CGM data. The association between CGM metrics and CVRFs was explored with bivariate correlation analysis. Binary multivariable logistic regression analyses were performed to test independent associations between CVRFs (cut-off defined according to the current International Society for Paediatric and Adolescents Diabetes [ISPAD] guidelines: BMI>85th percentile, LDL-c>100 mg/dL, BP>90thpercentile) and CGM metrics according to gender and adjusting for confounding factors. Results: In both genders, metrics of hypoglycaemia and glycaemic variability (coefficient of variation [%CV]) positively correlated with BMI percentile. LDL-c positively correlated with mean glucose and metrics of hyperglycaemia. A negative correlation was found between LDL-c and time in range (TIR). No significant correlations were found between CGM metrics and BP percentiles. In both genders, TIR<70% was significantly associated with LDL-c>100 mg/dL (OR 3.2 in males, 2.1 in females). In females, CV>36% was significantly associated with overweight (OR 2.1). Conclusions: CGM metrics of glycaemic control and variability were significantly associated with the risk of overweight in females and high LDL-c in both genders.