Factory-Calibrated Continuous Glucose Sensors: The Science Behind the Technology

Abstract

The use of commercially available continuous glucose monitors for diabetes management requires sensor calibrations, which until recently are exclusively performed by the patient. A new development is the implementation of factory calibration for subcutaneous glucose sensors, which eliminates the need for user calibrations and the associated blood glucose tests. Factory calibration means that the calibration process is part of the sensor manufacturing process and performed under controlled laboratory conditions. The ability to move from a user calibration to factory calibration is based on several technical requirements related to sensor stability and the robustness of the sensor manufacturing process. The main advantages of factory calibration over the conventional user calibration are: (a) more convenience for the user, since no more fingersticks are required for calibration and (b) elimination of use errors related to the execution of the calibration process, which can lead to sensor inaccuracies. The FreeStyle Libre^™ and FreeStyle Libre Pro^™ flash continuous glucose monitoring systems are the first commercially available sensor systems using factory-calibrated sensors. For these sensor systems, no user calibrations are required throughout the sensor wear duration.

Keywords: Calibration; Continuous Glucose Monitoring; Factory calibration; Glucose sensor; Subcutaneous.

FIG. 1.

Correlation between mean in vitro sensitivity and mean in vivo sensitivity of sensors from sensor lots used in Study 1 and 2. Mean in vitro sensitivity (horizontal axis) is the lot average of the individual in vitro sensor responses (sensor signal in nA divided by glucose concentration in mM). The in vitro sensor response was determined by testing sensors in glucose solution (20 mM phosphate buffered saline) with glucose concentrations ranging from 1 to 30 mM. The corresponding mean in vivo sensitivity (vertical axis) was obtained from clinical data. Individual in vivo sensor responses were calculated using capillary BG values and time paired sensor values. Correlation between in vitro and in vivo sensor sensitivity makes it possible to predict the in vivo sensor response from in vitro sensor testing (factory calibration). BG, blood glucose.

FIG. 2.

Per-sensor percentile (5th, 25th, 50th, 75th, and 95th) distribution of normalized sensitivity by day. Data from 72 subjects wearing 2 sensors simultaneously were collected together with capillary BG values over a 14-day sensor wear period to calculate the in vivo sensitivities.

FIG. 3.

Per-sensor in vivo sensitivities from three sensor lots are presented as separate distributions. Each dot represents one sensor. The in vivo sensitivity values (horizontal axis, sensor signal in nA divided by glucose concentration in mM) for the sensors are sorted from the lowest to the highest in a cumulative distribution function (cdf).The midpoint of the sorted values on the vertical axis (50th percentile) is the median value. For each of the three lots, all of the sensor sensitivities are within 20% of their corresponding median value.