Sub-second response algorithm for wearable glucose sensors: normalized slope-based calibration and microvolumes differential compensation measurements

Abstract

Traditional methods of blood glucose monitoring are invasive and can cause anxiety, pain and infection, resulting in poor patient compliance. Sweat-based glucose sensing has emerged as a promising non-invasive alternative, but the significantly lower glucose concentrations (10–100 times lower than in blood) pose a challenge for sensor sensitivity and operation. Here, we present different measurement protocols for enzymatic electrochemical glucose sensors with enhanced sensitivity and sub-second response calibration algorithm. The resulting amperometric response accurately reflects glucose concentration, demonstrating the sensor's potential for non-invasive monitoring of glucose in sweat. To enhance the reliability of the measurements and mitigate the variability among sensors arising from differences in sweat composition and secretion, a post-measurement protocol was developed. This protocol exploits a Response Correction Factor (RCF) calculated from the specific sensitivity of each sensor. This approach compensates for variability among different sensors reducing the standard deviation, thereby improving calibration accuracy (R² = 0.995 vs. R² = 0.822 without correction) allowing the prevention of phenomena related to enzyme inactivation or allogeneic reactions that may affect individual sensors in Continuous Glucose Monitoring (CGM) systems. An in-depth analysis was also conducted using sample microvolumes (20 μL), the typical amount of sweat available in wearable devices, to study thin-layer chronoamperometry response. To enhance the linearity of the sensor response, a differential compensation algorithm based on the slope of the response curve was adopted, employing a sensor without enzyme as a reference. This measurement method enhanced the dynamic range of slope values from 0.0085 μA/s to 0.0125 μA/s. The experimental results identified in a reliable way three operational regions: physiological (60–110 μM), warning values (110–160 μM) and alert/risk (>160 μM). The proposed strategies increase the robustness and applicability of sweat-based glucose monitoring for real-world applications.