Finite Element Modeling of Dermally-implanted Enzymatic Microparticle Glucose Sensors

With the rising prevalence of diabetes, effective means of successful management of blood glucose levels are increasingly important. To improve on the ease of measurements, new technology is being developed to enable less invasive measurements. Some recent efforts have focused on the development of optical microscale glucose sensing systems based on the encapsulation of glucose oxidase within microspheres coated with polyelectrolyte multilayer nanofilms. In such sensors, a phosphorescent oxygen indicator is also co-encapsulated with the enzyme inside so that when glucose is present, glucose oxidase within the sensor reduces the local oxygen levels, causing a corresponding change in the luminescence intensity of the sensors. To test the aforementioned factors, a two-substrate, 2D FEM model of microscale optical glucose sensors in the dermis was developed. The model was used to predict the response time and sensitivity of glucose sensors with varying number and spacing of particles distributed in the dermis and varying physiological characteristics of the surrounding tissue; specifically, capillary density, blood vessel location relative to sensor, and glucose and oxygen consumption in tissue. Simulations were conducted to determine the magnitude of the change in the response time of sensors. Because the steady-state oxygen concentration within the sensors for a given blood glucose level determines the signal output, steady-state concentration of oxygen within sensors and the surrounding tissue for the entire physiological glucose range was evaluated. The utility of the model to predict the performance and efficacy of the sensors in the event of a host response to the foreign body implant was also evaluated. Simulations were performed to evaluate changes in sensor response and sensitivity in the occurrence of inflammation and progression of fibrous encapsulation of various thickness and density. The results from these simulations have provided knowledge on the impact of physiological factors that can potentially degrade sensor function in vivo. Our results indicate that upon the occurrence of a host response, sensitivity is reduced while range is extended. Furthermore, using the model we have been able to determine which conditions in vivo improve response time, sensitivity, and the linear response range for these sensors.