Supplementary Material for: High glycemic variability is associated with worse Continuous Glucose Monitoring metrics in children and adolescents with type 1 diabetes

Objective: The primary aim of this study was to quantify the prevalence of children and adolescents with T1D who achieve the recommended target for CV identifying the determining factors to reach this target. The secondary aim was to examine the relationship between CV, the other metrics derived from CGM data and clinical parameters. Method: CGM data were collected from 805 children/adolescents with T1D. Several CGM metrics and patients' characteristics were evaluated. Participants were stratified by CV≤36% and CV>36%. Binary logistic regression analysis was run to identify the determining factors of high CV. Results CV was positively correlated with %TBR<70mg/dL, %TBR<54 mg/dL, %TAR>250 mg/dL, LBGI, HBGI and negatively with %TIR. CV≤36% was found in 31.4% of the subjects. The CV>36% group spent less time in %TIR, more time in hypoglycemia and hyperglycemia with lower proportion of subjects using rtCGM and CSII. Percentage of TBR<70mg/dL and TAR>250mg/dL were significant predictors of CV>36%, whereas age, gender, BMI, duration of diabetes, type of CGM device, type of insulin therapy administration and %TIR were not significant predictors (p<0.001,R2Nagelkerke=0.48). Conclusions: CV identifies children and adolescents with worse glycemic control at higher risk of both hypoglycemia and hyperglycemia.研究目标：本研究首要目标为量化1型糖尿病（T1D）儿童及青少年群体中，达到血糖变异性（CV）推荐控制目标的人群比例，并明确实现该目标的影响因素。次要目标为探究血糖变异性（CV）、持续葡萄糖监测（CGM）衍生的其他指标与临床参数之间的关联。 研究方法：本研究共纳入805名1型糖尿病（T1D）儿童及青少年，收集其持续葡萄糖监测（CGM）数据，对多项CGM指标及患者临床特征进行评估。以CV≤36%与CV>36%为界对受试者进行分组，采用二元logistic回归分析识别高CV的影响因素。 研究结果：血糖变异性（CV）与<70mg/dL低血糖时间占比（%TBR<70mg/dL）、<54mg/dL低血糖时间占比（%TBR<54mg/dL）、>250mg/dL高血糖时间占比（%TAR>250mg/dL）、低血糖指数（LBGI）及高血糖指数（HBGI）呈正相关，与目标血糖范围内时间占比（%TIR）呈负相关。共有31.4%的受试者CV≤36%。CV>36%组的目标血糖范围内时间占比更低，低血糖与高血糖的持续时间更长，使用实时持续葡萄糖监测（rtCGM）与皮下持续胰岛素输注（CSII）治疗的受试者比例更低。<70mg/dL低血糖时间占比与>250mg/dL高血糖时间占比是CV>36%的显著预测因素，而年龄、性别、体重指数（BMI）、糖尿病病程、CGM设备类型、胰岛素治疗方案类型及目标血糖范围内时间占比（%TIR）均无显著预测价值（p<0.001，Nagelkerke R²=0.48）。 研究结论：血糖变异性（CV）可有效识别出血糖控制较差、低血糖与高血糖风险均更高的1型糖尿病（T1D）儿童及青少年群体。