SAT098 Method For Predicting Time-below-range From Ambulatory Glucose Profile

Abstract

Abstract Disclosure: S. Takeishi: None. T. Inoue: None. Time-below-range (TBR) for target duration is described in ambulatory glucose profile (AGP) reports of continuous glucose monitoring (CGM). However, for a method to reduce hypoglycemia, it is necessary to determine the frequency of hypoglycemia and the time zone when hypoglycemia occurs. Thus, in this study, we evaluated a method for predicting TBR from AGP. This is a cross-sectional study. We analyzed glucose levels (GL) measured using CGM (FreeStyle Libre Pro) over 24 hours for 13 days (from 00:00 on Day 2 to 00:00 on Day 15 [CGM attachment: Day 1]) for 100 outpatients with type 2 diabetes, whose diabetic treatment did not change through the CGM-wearing duration. We proposed the metric, “eTBR<70AGP” as follows: ① when 95 percentile values were <70 mg/dL, GL<70 mg/dL were defined to exist at 100% (percentages were defined as “hypoglycemia existing rate [HER]”). When 95 percentile values were not <70 mg/dL, and 75 percentile values were <70 mg/dL, GL<70 mg/dL were defined to exist at 95%. When 75 percentile values were not <70 mg/dL and 50 percentile values were <70 mg/dL, GL<70 mg/dL were defined to exist at 75%. When 50 percentile values were not <70 mg/dL and 25 percentile values were <70 mg/dL, GL<70 mg/dL were defined to exist at 50%. When 25 percentile values were not <70 mg/dL and 5 percentile values were <70 mg/dL, GL<70 mg/dL were defined to exist at 25%. However, when 5 percentile values were not <70 mg/dL, GL<70 mg/dL were defined as nonexistent for convenience. ② We applied the definition mentioned in ① to the percentile values in the 13 days for each 96 timepoint of measurement every 15 minutes over 24-hours and determined the HER. ③ We calculated the eTBR<70AGP as the average of HER of the 96 timepoint. We also calculated “TBR (<70 mg/dL)” [TBR<70] for the 13 days. In patients overall (n=100), eTBR<70AGP correlated to TBR<70 (r=0.997, p<0.001). The optimal cutoff value of TBR<70 for “eTBR<70AGP - TBR<70 > 0” was 1.0% with sensitivity of 97% and specificity of 91% in receiver operating characteristic analysis [Area under the curve (AUC): 0.98, p<0.001]. In patients with TBR<70>1% (n=34), eTBR<70AGP was clinically equivalent to TBR<70 as values (11.3% vs 9.0%). However, eTBR<70AGP was statistically significantly higher than TBR<70 (p<0.001). In patients with TBR<70<1% (n=66), eTBR<70AGP was clinically equivalent to TBR<70 as values (0.04% vs 0.08%). However, eTBR<70AGP was statistically significantly lower than TBR<70 (p=0.02). eTBR<70AGP may be substituted for TBR<70 without underestimating the risk of hypoglycemia. Thinking in the calculation process of eTBR<70AGP may be utilized for assessment of the risk of hypoglycemia at each time zone. Presentation: Saturday, June 17, 2023