Objective: We previously reported the mean average relative difference (MARD) of the sensor glucose (SG) of the first-generation FreeStyle Libre with the original algorithm, an intermittent scanning continuous glucose monitoring (isCGM) device, was 15.6% in the Effect of Intermittent-Scanning Continuous Glucose Monitoring to Glycemic Control Including Hypoglycemia and Quality of Life of Patients with Type 1 Diabetes Mellitus Study (ISCHIA Study). In the present study, we aimed to further analyze its accuracy in detail by conducting a post-hoc analysis of the study.
Methods: The ISCHIA Study was a multicenter, randomized, cross-over trial to assess the efficacy of isCGM. The SG levels of isCGM and the measured capillary blood glucose (BG) levels of 91 participants were used for the analysis.
Results: Bland-Altman analysis showed bias of -13.0 mg/dl when the SG levels were compared to the BG levels, however no proportional bias was observed (r = 0.085). MARD of the participants without and with contact dermatitis were 15.0 ± 6.0% and 27.4 ± 21.4% (P = 0.001), respectively.
Conclusion: There was negative bias in the SG levels of isCGM compared to the BG levels. There is a possibility that the complication of the contact dermatitis during isCGM use may be related with deteriorated accuracy of the SG levels.