Fully Closed AID systems Effective Without Meal Announcements

Fully closed loop automated insulin delivery (AID) systems could be a game-changer for people with diabetes in the near future. The systems require no meal announcements, yet advanced computing allows them to provide excellent blood sugar control. At the ATTD 2025 conference on diabetes technology, presenters revealed some of the first study results on how these systems perform.

In one study, an open-source AID system that runs on an Android smartphone was shown to achieve time in range about 66 percent of the time despite users making no meal announcements. 

The AndroidAPS system is capable of making small insulin dosing adjustments as blood sugar rises and falls, based on readings from a continuous glucose monitor (CGM).

Neale Cohen, a professor at Baker Heart and Diabetes Institute researchers in New Zealand presented the CLOSE IT study, which involved 75 participants with type 1 over three months.

"A reliable, fully closed loop system would change the paradigm for people living with type 1 diabetes," Cohen said. "I think we're close."

In the study, half of the participants announced meals and manually bolused insulin, while the other half used the system in fully closed loop mode and all dosing was automatic. At the end of the trial, the hybrid closed loop group achieved time in range 69% of the time. And the fully closed loop participants fared nearly as well, at 66% percent time in range. A1C did not significantly differ between the two groups, averaging 6.8-6.9%.

In addition, there was no significant difference in time below range, Cohen said. 

"We did a lot of work adjusting the algorithm, and I know that was something that took a lot of time, quite frankly," Cohen said. "But I think that was the reason for the success of what we saw."

For example, initially users saw higher blood sugars after meals at dinner, which the researchers addressed by tuning the insulin sensitivity factor on the system.

"In the evening, where the meals were larger, we had to adjust sensitivity factors really tight for a few hours," he said. "And that was one of the things we found very useful, was just in that time between, say, 6 p.m. to 10 p.m." 

Cohen reported that user satisfaction with the system was still being analyzed but was favorable.

"There's a pretty positive reaction generally, and some of the metrics were strongly positive," he said.

A different approach

In another study, research showed an AID system that uses a form of artificial intelligence and no user input was as effective at controlling blood sugar as a hybrid-closed loop system where users made meal announcements and manually bolused.

Dr. Laya Ekhlaspour, a pediatric endocrinologist at the University of California, San Francisco, presented results from the FCL@Home study.

The system uses neural networking technology, a type of advanced computing modeled on the human brain.

In the study, users of the fully closed, neural-network system had an average time in range of 62%, compared to 49% for the hybrid closed loop group (one participant open-looped, meaning they used no automated dosing). Time above range was lower on average for those using the fully closed loop – 37% to 50%. Average blood sugar level in adults, young adults, and adolescents were all better for those who used the advanced fully closed loop system.

"We want a fully closed loop algorithm because we want to decrease the user input," Ekhlaspour said. "We want to improve glycemic outcomes, and we want to expand the use. As a pediatrician, I'm very excited to be able to introduce the system to our patients who can benefit."

The bottom line

Studies show that fully closed loop systems can achieve time in range comparable to hybrid closed-loop systems, without requiring user input for meals. While further refinements are needed, especially for reducing mealtime blood sugar spikes in the evening when meals tend to be larger, the results indicate progress.

Advancements in artificial intelligence, such as neural network-based insulin delivery, show promise in managing sugar levels. Researchers said their ultimate goal is to create systems that work for all users, regardless of their varying needs.

Researchers and people with diabetes who have tested the systems in studies have responded positively to fully closed-loop AID systems, which could soon help people with diabetes achieve better blood sugar control with less effort. These systems could potentially expand access to people with diabetes who providers may otherwise not consider a good fit for automated insulin delivery.

 Learn more about diabetes technology:

• Fully Closed Loop: Automated Insulin Delivery Takes the Next Step
• Tech Watch: The Latest in Diabetes Tech News
• Guide to AID Systems