Wearable Fitness Sensors in Automated Insulin Delivery Systems For Diabetes Management

Automated insulin delivery (AID) systems have quickly become a vital part of clinical care for patients with diabetes. Despite being a solution for many aspects of diabetes management, limitations exist with currently available technology, such as the inability to automatically respond to stimuli like meals or exercise to avoid hypoglycemia.

A recent study examined the potential of wearable fitness sensors in combination with AID systems to minimize hypoglycemic events among patients with type 1 diabetes.¹ The AID systems used real-time fitness data from wearable sensors to adjust insulin dosing and found that these combined systems resulted in improved glucose outcomes.

Lead study author Peter G. Jacobs, PhD, discusses these study results and the potential future implications of this technology for diabetes management. Dr Jacobs is an Associate Professor in the Department of Biomedical Engineering at the Center for Health and Healing at Oregon Health and Science University and the Artificial Intelligence for Medical Systems Lab, in Portland, Oregon.

Consultant360: To begin, could you briefly discuss the current role of automated insulin delivery (AID) systems in preventing hypoglycemia among patients with diabetes?

Peter Jacobs: There are currently multiple AID systems available on the market for people to choose from. In general, the use of these AID systems has been shown to reduce HbA1c by 0.2% to 0.5%, which is an important but also modest amount. There are multiple reasons for this. These AID systems have primarily shown benefit in improving glucose outcomes during the overnight period, when meals and exercise events are not occurring. AID systems do not provide significant benefit during times when a person is consuming meals because there is a significant delay in kinetics of subcutaneous insulin delivery of about 1 hour, thereby leading to large post-meal glucose spikes. Exercise, and especially aerobic exercise, is also a challenge because it can cause steep drops in glucose during and following the physical activity, which can lead to hypoglycemia. Current AID systems cannot automatically detect that exercise is occurring and so they cannot reduce insulin in response to the onset of exercise. Therefore, hypoglycemia can still occur even when people are using an AID.

C360: Can you tell us more about how this study came about? Why now?

PJ: Since current AID systems do not automatically detect and respond to exercise, we wanted to build and test a system that uses wrist-worn fitness sensors to automatically detect the exercise and adjust the AID’s insulin dosing when exercise was happening. Furthermore, we wanted physical activity data to be utilized by the AID all the time, not just when a person was doing typical exercise. When a person is doing physically active housework or yardwork for example, a drop in glucose is also possible due to this physical activity. Since commercial fitness sensors provide heart rate data, we were able to utilize this information to inform an AID control algorithm to automatically respond to exercise and reduce insulin delivery.  

C360: The results of your study showed that both algorithms examined here had comparable outcomes, but that one algorithm required interaction from the participant whereas the other did not. Why is this an important finding?

PJ: Most people have busy lives and they do not always remember to adjust the insulin settings on an AID. People with type 1 diabetes using AIDs oftentimes forget to report meals to the system and they may forget to adjust their insulin dosing before and or exercise. To remove additional burden for people with type 1 diabetes, we designed the system to respond to exercise in a fully automated way.   

C360: Your study focused on the use of wearable sensors in AID systems among people completing aerobic exercise and noted that future studies will need to be done with a variety of exercise types, such as resistance or interval exercise. Do you expect the type, duration, or intensity of exercise to impact the performance of these systems? Why or why not?

PJ: We did a study recently² that showed that aerobic exercise causes the most significant drops in glucose under free-living conditions compared with resistance (eg, weights) or interval (eg, cross-fit or soccer). In fact, resistance exercise resulted in a 50% less drop in glucose compared with aerobic exercise. Therefore, it’s important for an exercise-aware AID to automatically detect the type of exercise so that it can respond appropriately to improve glucose outcomes.

C360: What is the next step for research on the use of wearable sensors and automated AID systems in diabetes technology?

PJ: Glucagon is a hormone that stimulates endogenous glucose production. Delivery of glucagon can be used to help prevent hypoglycemia as it acts very quickly to increase glucose levels, whereas shutting off insulin oftentimes does not result in a rapid increase in glucose. We have done research in the past³ showing that use of the hormone glucagon within a multi-hormone exercise-aware AID can significantly reduce low glucose (<70 mg/dL) and eliminate severe low glucose events (<54 mg/dL) when glucagon was delivered during aerobic exercise compared with an exercise-aware AID that only delivers insulin. There are several multi-hormone pumps coming on the market soon that could make use of glucagon to help prevent hypoglycemia during and after exercise. Glucagon is also available now in liquid stable form, thereby making it appropriate for use in a pump.

References:

1. Jacobs PG, Resalat N, Hilts W, et al. Integrating metabolic expenditure information from wearable fitness sensors into an AI-augmented automated insulin delivery system: a randomized clinical trial. Lancet Digit Health. 2023;5(9):e607-e617. doi:10.1016/S2589-7500(23)00112-7.
2. Riddell MC, Li Z, Gal RL, et al. Examining the acute glycemic effects of different types of structured exercise sessions in type 1 diabetes in a real-world setting: the type 1 diabetes and exercise initiative (T1DEXI). Diabetes Care. 2023;46(4):704-713. doi: 10.2337/dc22-1721
3. Wilson LM, Jacobs PG, Ramsey KL, et al. Dual-hormone closed-loop system using a liquid stable glucagon formulation versus insulin-only closed loop system compared with a predictive low glucose suspend system: an open-label, outpatient, single-center, crossover, randomized controlled trial. Diabetes Care. 2020;43(11):2721-2729. doi: 10.2337/dc19-2267

© 2023 HMP Global. All Rights Reserved.
Any views and opinions expressed are those of the author(s) and/or participants and do not necessarily reflect the views, policy, or position of Consultant360 or HMP Global, their employees, and affiliates.