Endocrine Abstracts

There is little information regarding inter-individual variability of accuracy for CGM and flash glucose monitor (FGM) sensors. We studied the possibility of inter-individual variability of accuracy for CGM and FGM sensors. This is a prospective study. During hospitalization, 10 patients with type 2 diabetes wore a CGM [iPro2] and FGM [FreeStyle Libre Pro] in parallel over 24 h while every pre- and post-prandial “venous blood glucose levels” (VBG) were measured and used to calibrate CGM (n of VBG measurement = 60). We evaluated mean absolute difference (MAD) and mean absolute relative difference (MARD), derived from sensor glucose levels (SG) for CGM, measured every 5 min (MAD “5 min” and MARD “5 min”) and extracted every 15 min (MAD “15 min” and MARD “15 min”), and derived from SG for FGM, measured every 15 min (MAD “FGM” and MARD “FGM”). The study participants correlated to the distribution of VBG (correlation ratio: η² = 0.38 P = 0.003). The study participants correlated to the distribution of MAD “5 min” and that of MAD “15 min” (η^{2 =} 0.27 P = 0.04998, η² = 0.27 P = 0.047, respectively); however, the study participants did not correlate to the distribution of MARD “5 min” and that of MARD “15 min”. However, the study participants correlated to the distribution of MAD “FGM” and that of MARD “FGM” (η² = 0.57 P < 0.001, η² = 0.52 P < 0.001, respectively). The result of VBG may have been caused by patient-to-patient differences in glycemic control. The results of MAD for CGM may have been caused by the fact that higher VBG may lead to higher MAD. The results of MARD for CGM may have been caused by the fact that VBG are not associated with MARD because MARD is percentage of MAD to VBG. Compared of the result of MARD for FGM to the result of MARD for CGM, the correlation between the study participants and distribution of MARD “FGM” may suggest the inter-individual variability of accuracy for FGM sensors.