Miriam E. Tucker
June 27, 2018
Closed-loop automated insulin delivery offers the potential for improving glycemic control among hospitalized noncritically ill patients with type 2 diabetes, new research suggests.
Findings from a randomized open-label trial of a prototype fully closed-loop system in 136 adults with type 2 diabetes were presented June 25 here at the American Diabetes Association (ADA) 2018 Scientific Sessions by Lia Bally, PhD, of Inselspital, Bern University Hospital, Switzerland. The results were also simultaneously published in the New England Journal of Medicine.
"Use of a fully automated subcutaneous closed-loop insulin delivery system in patients receiving noncritical care resulted in significantly better glycemic control compared to conventional treatment. This was notably achieved without increasing the risk of hypoglycemia. Closed-loop might therefore represent a novel approach to manage inpatient hyperglycemia in a safe and effective way," Bally said during her presentation.
The system used in the study, FlorenceD2W-T2, comprises a model predictive control algorithm residing on a Dell tablet computer linked via USB cable to a continuous glucose monitoring (CGM) receiver (FreeStyle Navigator II, Abbott). The tablet communicates with the study pump (Dana R Diabecare, Sooil) via Bluetooth, automatically adjusting the insulin dose based on the CGM readings. Such systems are often referred to as a closed-loop system or artificial pancreas.
Unlike currently marketed hybrid closed-loop systems, this one is fully automated so it doesn't require meal announcements. It is initialized using only the patient's weight and total daily insulin dose.
More Work Needed to Assess Practicality, but Easy to Use
"Compared to conventional treatment approaches, the advantage of the fully automated closed-loop insulin delivery system is the finely tuned, instantaneous and glucose-responsive automation of insulin which addresses the rapidly changing needs of a diverse and complex hospital population, including patients receiving hemodialysis, glucocorticoid treatment, and nutrition support," Bally explained.
In an interview with Medscape Medical News, Bally also noted, "I think there is still more work to do to assess the practicality, but in terms of time and effort necessary, the system can be installed within 15 minutes, and I think it's really easy to use. Nurses or healthcare assistants could be trained."
Asked to comment, session moderator Elena Toschi, MD, of the Joslin Diabetes Center, Boston, Massachusetts, said that Bally "showed great data" adding that the system "really improved all the measures we want to improve: time in range with no hypoglycemia."
However, if such systems are to be introduced in hospitals, there would definitely be a need to develop protocols, including for staff training, and very clear guidelines, said Toschi. "But the appealing part is that it might be cost-effective, so the hospital might like that. Having less hypoglycemia and more time in range has been shown in the [intensive care unit] especially to improve outcomes and reduce hospitalizations," she added.
Toschi also noted that such a system would need to be a fully closed loop, as used in the study, rather than the currently marketed hybrid closed-loops, which "might involve too much input from patients or personnel. I think it's less realistic."
Increased Time in Range, No Rise in Hypoglycemia
The study was conducted at one center each in the UK and Switzerland, and 136 adults with type 2 diabetes who required subcutaneous insulin therapy were randomized to the closed-loop system (n = 70) or conventional insulin injections (n = 66) without other glucose-lowering medications.
Infection/sepsis was the most common reason for hospital admission, in just over 40% of both groups, followed by renal, cardiac, malignancy, and gastrointestinal issues.
Ten participants in the closed-loop system group and nine participants in control group were on hemodialysis; 12 and 18, respectively, received glucocorticoid medication; and five and four, respectively, received parenteral or enteral nutrition.
The closed-loop system was stopped in four patients because of transfer to intensive care and in two because of participant discomfort while wearing the device.
The primary endpoint, percentage of time in hospital-adapted target glucose range (100–180 mg/dL) for up to 15 days or until hospital discharge, was 65.8% with the closed-loop system versus 41.5% with conventional treatment, a significant difference (P < .001).
Mean glucose values were also significantly different at 154 mg/dL versus 188 mg/dL (P < 0.001).
Mean time spent in hyperglycemia (glucose ≥ 180 mg/dL) was also significantly lower in the closed-loop system vs control group (23.6% vs 49.5%; P < .001), and median time spent with glucose levels below 54 mg/dL was zero throughout the study in both groups.
The median amount of insulin delivered also didn't differ between groups, at 44.4 units/day with the closed-loop system versus 40.2 units/day for controls (P = .50).
When asked for feedback, 98% of participants said they were happy to have their glucose controlled automatically by the closed-loop system and 100% would recommend it to others.
Bally told Medscape Medical News that the system "has the potential to reduce the workload, but this is not a conclusion we can actually take from this study. This needs further research."
When asked whether she thought the majority of American hospitals would embrace such a novel approach to managing glycemia, Toschi replied, "It might take several years but I think it's something we'll see more and more. There's already a shift away from sliding scale to use of meal bolus with corrections and for allowing patients with type 1 diabetes to wear pumps ... So I think we can progress."
Bally has reported receiving grants from the Swiss National Science Foundation, University Hospital Bern, University of Bern, and Swiss Diabetes Foundation. Toschi has reported no relevant financial relationships.