Discrimination and Precision of Continuous Glucose Monitoring in Staging Children With Presymptomatic Type 1 Diabetes

Abstract

Context: Staging and monitoring of presymptomatic type 1 diabetes includes the assessment for dysglycemia.

Objective: To assess the ability of continuous glucose monitoring (CGM) to differentiate between islet autoantibody-negative controls and early-stage type 1 diabetes and explore whether CGM classifiers predict progression to clinical diabetes.

Research design and methods: Children and adolescents participating in public health screening for islet autoantibodies in Bavaria, Germany, were invited to undergo CGM with Dexcom G6. In total, 118 participated and valid data was obtained from 97 [57 female; median age 10 (range 3-17) years], including 46 with stage 1, 18 with stage 2, and 33 with no islet autoantibodies.

Results: Mean glucose during CGM in islet autoantibody-negative controls was high (median, 115.3 mg/dL) and varied substantially (interquartile range, 106.8-124.4). Eleven (33%) of the controls had more than 10% of glucose values above 140 mg/dL (TA140). Using thresholds corresponding to 100% specificity in controls, differences between controls and stage 1 and stage 2 were obtained for glucose SD, TA140, TA160, and TA180. Elevations in any 2 of these parameters identified 12 (67%) with stage 2 and 9 (82%) of 11 participants who developed clinical diabetes within 1 year. However, there was marked variation within groups for all parameters and poor consistency observed in a second CGM performed in 18 participants.

Conclusion: This study demonstrated the potential of integrating CGM into staging and monitoring of early-stage type 1 diabetes. However, substantial improvement in the precision of CGM is required for its application in routine monitoring practices.

Keywords: continuous glucose monitoring; disease progression; early-stage diagnosis; glucose variability; pediatrics; type 1 diabetes.

Figure 1.

CGM metrics in islet autoantibody-negative youth and youth with stage 1 and stage 2 type 1 diabetes; (A) mean glucose (mg/dL), (B) SD (mg/dL), (C) CV (%), (D) TA140 (%), (E) TA160 (%), (F) TA180 (%). P-values indicate the level of significance between groups (Mann–Whitney U test). Abbreviations: CV, coefficient of variation; CGM, continuous glucose monitoring; TA, time above; TB, time below.

Figure 2.

Receiver operator curve analysis between islet autoantibody-negative control youth and youth with stage 2 type 1 diabetes for (A) SD (mg/dL), (B) TA140 (%), (C) TA160 (%), and (D) TA180 (%). The area under the curve and P-values are indicated. Abbreviations: TA, time above; TB, time below.

Figure 3.

Logistic regression derived PPS values for progression to stage 3 type 1 diabetes. (A) PPS values in multiple islet autoantibody-positive youth who did and did not progress to stage 3 type 1 diabetes within 1 year. (B) ROC of PPS for discriminating multiple islet autoantibody-positive youth who did from those who did not progress to stage 3 type 1 diabetes within 1 year. (C) ROC of 120-minute glucose value in the OGTT for discriminating multiple islet autoantibody-positive youth who did from those who did not progress to stage 3 type 1 diabetes within 1 year. (D) Kaplan–Meier analysis of progression to stage 3 type 1 diabetes within 1 year by PPS. P-values for the Kaplan–Meier analysis were calculated using the log-rank test. Abbreviations: OGTT, oral glucose tolerance test; PPS, Progression Prediction Score; ROC, receiver operating characteristic.

Figure 4.

Comparison of CGM metrics in 18 participants with either stage 1 (left) or stage 2 (right) type 1 diabetes who had 2 periods of CGM assessments. (A) mean glucose (mg/dL), (B) SD (mg/dL), (C) TA140 (%), (D) TA160 (%), (E) TA180 (%), and (F) the number of parameters above 100% specificity values of islet autoantibody-negative controls. Individual participants are connected by lines. For participants with stage 2 type 1 diabetes, the stage 2 status is shown as a filled red circle for each of the 2 CGM assessments. Abbreviations: CGM, continuous glucose monitoring; TA, time above; TB, time below.