We studied how to achieve both the target time in range (70-180 mg/dL) [TIR] and target coefficient of variation (CV). In a cross-sectional study, we analyzed 24-h glucose levels measured using CGM (iPro2) for 150 patients with type 2 diabetes. We calculated the “mean of all absolute differences between glucose levels and the ‘reference glucose levels (R)’” (M1, in mg/dL), corresponding to 21 R (110, 111⋯130 mg/dL), as a metric to achieve the target TIR. We also calculated the “percentage of the mean absolute deviation to mean glucose levels” (M2, in %) as a metric to achieve the target CV. The R values for which M1 most correlated to TIR were 117 and 118 mg/dL (r = -0.95, p < 0.001). M2 correlated to CV (r = 0.99, p < 0.001; M2 = 0.85 × CV - 0.43). The optimal cutoff value (OC) for TIR > 70% in “M1 for R of 117.5 mg/dL” (M1, 117.5) was 44.6 mg/dL (sensitivity, 96%; specificity, 96%; AUC, 0.99; p < 0.001). The OC for CV < 36% in M2 was 31.5% (sensitivity, 100%; specificity, 100%; AUC, 1). The OC for “both TIR > 70% and CV < 36%” in “M1, 117.5 × M2,” was 842.0 (sensitivity, 90%; specificity, 89%; AUC, 0.96; p < 0.001). Putting all glucose levels close to 117.5 mg/dL may be the most efficient method to increase the TIR as much as possible. Based on our results, the simple target values are as follows: “M1, 117.5 of 42.5 mg/dL” (75-160) [if all glucose levels are within 75-160 mg/dL, M1, 117.5 is sure to be < 42.5 mg/dL] × M2 of 20% = 850 (≈ 842)” (T1); “M1, 117.5 of 27.5 mg/dL (90-145) × M2 of 30% = 825 (≈ 842)” (T2). For T1, if the range that satisfies both 75-160 mg/dL and “‘24 h mean glucose levels on the day before’ ± ‘20 % of the mean glucose levels’” (mean ± 20%) is used as an alarm threshold in a personal CGM, keeping all glucose levels within the range as much as possible may achieve both TIR > 70% and CV < 36%. For T2, if the range that satisfies both 90-145 mg/dL and mean ± 30% is used as an alarm threshold, keeping all glucose levels within the range as much as possible may achieve both TIR > 70% and CV < 36%.

Disclosure

S. Takeishi: None. T. Inoue: None.

Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at http://www.diabetesjournals.org/content/license.