Time lag characterization of two continuous glucose monitoring systems

Abstract Aim To evaluate the system time lag associated with subcutaneous interstitial glucose (IG) sensing and venous blood glucose (BG) of two continuous glucose monitoring (CGM) systems, the SEVEN ® (DexCom™, San Diego, CA) and the Navigator ® (Abbott Diabetes Care, Alameda, CA), in adults with type 1 diabetes. Methods Fourteen subjects wore both CGM systems concurrently during the 15-day study. Reference YSI (Yellow Springs, OH) and CGM data from the in-clinic sessions, conducted on day 5, 10 and 15 of the study were evaluated. The system time lag of CGM system was estimated using various regression method related statistical estimators. Results The estimated time lags based on different statistical measures are similar within each CGM system. The time lag based on correlation coefficient criteria is estimated as 4.5 ± 5 min (median ± IQR) for the SEVEN ® , and 15 ± 7 min for the Navigator ® . The ranges of these estimators of two CGM systems were different (2–6 min for SEVEN ® and 14–15 min for Navigator ® ). Conclusions The study findings suggest that commonly accessible statistics, such as correlation statistics, offer estimates that are comparable to complicated approaches. Different time lags were observed with two CGM systems.