This study investigated the feasibility and accuracy of a new automated insulin delivery system combining an insulin patch pump with a continuous glucose monitor.

This study was the first time testing this combined device that integrates a patch insulin pump with a glucose-sensing insulin delivery cannula in humans. We hoped to determine whether combining these two functions could reduce the burden of managing type 1 diabetes.

This study was completed in two phases. Phase one involved 3 participants who completed one mixed meal test (MMT) for 12 hours while wearing the device. Phase two involved 15 participants wearing the device over at least 3 days, completing meal tests on day 1 and day 3, with one supervised free-living day at a hotel. Our aims for this study were to test was accuracy of the CGM device, investigate the device survival duration, and analyze time in range while wearing the device.

We had 18 people complete this study, all of whom had experience with insulin pumps and CGMs. Throughout the study, the glucose readings of the device were compared against a YSI glucose analyzer (a laboratory reference standard), blood glucose meter readings, and the participants’ own commercial CGM systems, to ensure participant safety and device reliability.

Overall, we used 26 devices over the duration of the study, with some devices being excluded or replaced due to manufacturing issues or immediate device failure. Of these 26 devices, 17 produced usable data. Devices that were successfully inserted lasted nearly 3 days on average, and 12 of 14 devices in the main phase completed the full 3-day study without premature removal. No serious adverse events occurred in this study, and the device was tolerated well by the vast majority of participants.

The glucose sensing cannula of this device performed similarly to commercially available CGM systems, meaning that the devices were reliable and could be safe to use for insulin dosing decisions. No insulin delivery failures were recorded throughout this study, and the average bolus time was approximately 8 minutes.

We concluded that this study demonstrates that integrating a patch pump with a glucose-sensing insulin delivery cannula in a single device is feasible and safe. However, as some insertions worked much better than others, improvements in manufacturing and insertion processes could make the device more accurate and consistent.

The full publication can be accessed here