The InsulPatch is a novel integrated patch-pump device used to deliver drugs, especially macromolecular drugs that are difficult to deliver through an oral pathway and that require transdermal delivery. The patch-pump is a promising replacement for conventional syringes and battery-powered pumps because it is slim, powerless, painless, and relatively inexpensive. The majority of this thesis focuses on the fabrication and testing of microfluidic devices for the delivery of insulin, which is a model drug that is widely used and needs to be delivered transdermally.

In this thesis, we demonstrate the fabrication of the patch-pump, which includes an insect-mimetic microfluidic pump fabricated using photolithography and replica molding, and a microneedle array fabricated using 3D printing. The microfluidic pump is used to drive the fluid flow powered by pressurized air or the user’s pulse, and the microneedle array is used to inject the fluid through the skin painlessly. Using pressurized air-driven flow testing, we have tested the flow rate across microfluidic pumps of various flow channel widths over a range of physiologically relevant actuation frequencies and pressures. We have found that for the specific channel design we have been using, the flow rate generally positively correlates with the actuation pressure. For devices with wider flow channels, the flow rate generally negatively correlates with the actuation frequency, whereas the flow rate increases and then decreases with increasing actuation frequency for devices with narrower flow channels. This property of these devices is beneficial in insulin delivery because the demand for insulin is generally reduced in vigorous exercise (with elevated heart rate/actuation frequency) and increased in hypertension patients (with elevated blood/actuation pressure).

A major future direction of the study is to test a wide range of device designs in a sample of human subjects by attaching the device onto the wrist and measuring the pulse-driven flow across the device. We can further change the channel design parameters of the device so that it will be ideal for insulin delivery. Using the ex vivo flow testing and human subject data, we can further tailor the device design to specific patients using a genetic algorithm-guided optimization based on the heart rate and blood pressure of the patient and the desired flow rate. We will also perform computational modeling using COMSOL Multiphysics to predict the flow across devices of different designs as well as to understand the physics behind the pulse-driven flow. Finally, a 3D-printed insulin reservoir will be incorporated into our patch-pump system for the storage of U-500 insulin.