The purpose of this analysis with the results from all performed clinical studies was to evaluate the CoG NI performance compared to the YSI 2300STATPlus reference method.
Data collected from 224 participants from four clinical studies were included into this analysis (52% male / 48% female, 26.3% type 1, 56.3% type 2, and 17.4% healthy subjects). The results were collected during in-clinic visits (standardized meal tests) and during 12 weeks of home use with comparisons to YSI during rapid glucose changes and stable glucose levels.
Meta-analysis provided 3,798 paired NI/YSI data points from the in-clinic meal tests and 18,654 non-invasive measurements from the 12 weeks of in-home use. More than 99% of the noninvasive readings were in the A and B zones of the Consensus Error Grid. The total MARD (vs. YSI) of the entire meta-analysis data set of NI measurements was 15.2%. Forty-three possible interfering factors were tested, and no interference was detected. Fifty-four percent fewer hypoglycemic episodes were detected (p<0.01) with stable glycemic control (no change in HbA1c) during the three-month home use period. Patients were testing more frequently with the non-invasive module at home and reported a high level of treatment satisfaction.
The CoG device showed the acceptable performance and may help to improve routine care of patients with type 1 and type 2 diabetes.