Accuracy of a Fourth-Generation Subcutaneous Continuous Glucose Sensor

Background: This study evaluated the accuracy and performance of a fourth-generation subcutaneous glucose sensor (Guardian(™) Sensor 3) in the abdomen and arm. Methods: Eighty-eight subjects (14–75 years of age, mean ± standard deviation [SD] of 42.0 ± 19.1 years) with type 1 or type 2 diabetes participated in the study. Subjects wore two sensors in the abdomen that were paired with either a MiniMed(™) 640G insulin pump, or an iPhone(®) or iPod(®) touch(®) running a glucose monitoring mobile application (Guardian Connect system) and a third sensor in the arm, which was connected to a glucose sensor recorder (GSR). Subjects were also asked to undergo in-clinic visits of 12–14 h on study days 1, 3, and 7 for frequent blood glucose sample testing using a Yellow Springs Instrument (YSI) reference. Results: The overall mean absolute relative difference (MARD ± SD) between abdomen sensor glucose (SG) and YSI reference values was 9.6% ± 9.0% and 9.4% ± 9.8% for the MiniMed 640G insulin pump and Guardian Connect system, respectively; and 8.7% ± 8.0% between arm SG and YSI reference values. The percentage of SG values within 20% agreement of the YSI reference value (for YSI >80 mg/dL) was 90.7% with the MiniMed 640G insulin pump, 91.8% with the Guardian Connect system, and 93.1% for GSR-connected arm sensors. Mean functional sensor life, when calibrating 3–4 times/day, was 145.9 ± 39.3 h for sensors paired with the MiniMed 640G insulin pump, 146.1 ± 41.6 h for sensors paired with the Guardian Connect system, and 147.6 ± 40.4 h for sensors connected to the GSR. Responses to survey questions regarding sensor comfort and ease of use were favorable. Conclusions: The Guardian Sensor 3 glucose sensor, whether located in abdomen or the arm, provided accurate glucose readings when compared with the YSI reference and demonstrated functional life commensurate with the intended 7-day use. ClinicalTrials.gov: NCT02246582