Objectives: To study the effect of baseline glucose variability (GV) on the time to stability and interweek variability (IWV) of continuous glucose monitoring (CGM)-derived glycemic indices. Materials and Methods: Anonymized CGM data (median duration 32 weeks, ≥70% data coverage) of 85 adults with type 1 diabetes; age (41 ± 12 years), 66.3% women, HbA1c (7.5% ± 1.2%, 58 mmol/mol) were analyzed. We evaluated the time to stability, that is, the minimum duration of data that provided a close (r2 ≥ 0.9) correlation with data taken across the whole sampling period and IWV. We also evaluated the impact of baseline variability on the time to stability. Results: For the whole data set, all indices achieved stability (r2 ≥ 0.9) by 9 weeks (range 5-9). Time to stability progressively increased from the lowest quartile to the highest quartile of baseline coefficient of variation (CV). Time above range (TAR) and time below range (TBR) had higher IWV than time in range (TIR) (%CVIWV: TIR-16%, TAR-31%, TBR-62%). Conclusion: Baseline GV and IWV of indices affect the time to stability of glycemic indices. We recommend a minimum of 9 weeks of data to represent long-term CGM data.
Keywords: Coefficient of variation; Continuous glucose monitoring; Glucose variability; Time in range; Type 1 diabetes.