Frequency of flash glucose monitoring and glucose metrics: real-world observational data from Saudi Arabia

Abstract

Background: This real-world data study analyzed glucose metrics from FreeStyle Libre^® flash glucose monitoring in relation to scanning frequency, time in range (TIR) and estimated A1c (eA1c) in Saudi Arabia.

Methods: Anonymized reader data were analyzed according to scanning frequency quartiles, eA1c categories (<7%,≥7%‒≤9% or>9%) and TIR categories (<50%,≥50%‒≤70% or>70%). Sensors, grouped by reader, were required to have≥120 h of operation. Differences in scanning frequency, eA1c, TIR, time in hypoglycemia and hyperglycemia, and glucose variability (standard deviation [SD] and coefficient of variation [CV]) were analyzed between groups.

Results: 6097 readers, 35,747 sensors, and 40 million automatic glucose measurements were analyzed. Patients in the highest scanning frequency quartile (Q4, mean 32.0 scans/day) had lower eA1c (8.47%), greater TIR (46.4%) and lower glucose variation (SD 75.0 mg/dL, CV 38.2%) compared to the lowest quartile (Q1, mean 5.2 scans/day; eA1c 9.77%, TIR 32.8%, SD 94.9 mg/dL, CV 41.3%). Lower eA1c and higher TIR were associated with greater scanning frequency, lower glucose variability and less time in hyperglycemia.

Conclusions: Higher scanning frequency in flash glucose users from Saudi Arabia is associated with lower eA1c, higher TIR, lower glucose variability and less time in hypoglycemia or hyperglycemia.

Keywords: Flash glucose monitoring; FreeStyle Libre®; Real-world data.

Fig. 1

Relationship between scanning frequency and glucose metrics Mean values are connected by solid lines, 75th and 25th percentile values are connected by dashed lines. For percent time below 54 and 70 mg/dL, median values are connected by solid lines.