Artificial Pancreas 'Impressive' in 4-Week, Nightly Home Use Studies

Marlene Busko

September 19, 2014

VIENNA — Researchers have reported striking, positive results with the artificial pancreas — a closed-loop insulin-delivery system — in 40 patients with type 1 diabetes, comparing the device vs a sensor-augmented insulin pump for unsupervised, at-home, overnight delivery of insulin.

In this, the longest, largest analysis of this new device so far, involving 4 randomized crossover studies, adolescents used it for 3 weeks and adults used it for 4 weeks — with no supervision.

When patients used the closed-loop device overnight, their glucose levels remained in the target range (3.9–8.0 mmol/L) for 18% longer than when they used only sensor-augmented insulin pump. Their mean overnight glucose was reduced by 0.8 mmol/L, and they were hyperglycemic or hypoglycemic for less time with the artificial pancreas.

"We have shown that extended use of unsupervised overnight closed-loop delivery of insulin is feasible in adults and adolescents with type 1 diabetes," Hood Thabit, MD, from the University of Cambridge Metabolic Research Laboratories and Wellcome Trust-MRC Institute of Metabolic Science, Cambridge, United Kingdom, said in an oral presentation here at the European Association for the Study of Diabetes 2014 Meeting.

"We can significantly increase the time spent in target and reduce glucose overnight, [and] we can still reduce hypoglycemia, even with a low rate of hyperglycemia," he continued. "We do acknowledge that currently, closed loop is still limited by connectivity between devices, but the hope is that this will improve and further increase the efficacy and feasibility…in the future."

This study is "quite impressive," session cochair Eric M. Renard, from the department of endocrinology, diabetes, and nutrition, at Montpellier University Hospital, France, commented to Medscape Medical News after the talk.

"I think that for nighttime [use], we have very strong proof now [that the 'artificial pancreas' works], and there are ongoing studies that will show [that it works] for several months," he said.

Comfortable to Use, Connectivity Problems Are Last Obstacle

Dr. Renard said that once the connectivity problem is solved, the artificial pancreas will no longer be a dream, "it will be the reality."

Asked how the artificial pancreas compares with the "bionic pancreas" that created a buzz at the recent American Diabetes Association meeting, he clarified that these are 2 different systems.

In the long term it means HbA1c will be lower, because the positive effect in the night is sustained the next day.

The bionic pancreas delivers glucagon as well as insulin, which increases the device "burden" for the patient. "If you have 2 pumps, 1 with glucagon and 1 with insulin, it can become a nightmare. It's good for studies, but not for everyday [use]," he commented.

Moreover, "what is elegant in this study [by Thabit and colleagues] is that the patients had few devices on [their bodies]," he noted. The artificial pancreas has wireless connections between the insulin pump and both the sensor and the platform that runs the algorithm, which makes it quite comfortable for nighttime use. Most patients were quite happy with the device.

"What is really nice, because hypoglycemia was reduced and glucose was better, if you prolong this, in the long term it means HbA1c will be lower, because the positive effect in the night is sustained the next day," Dr. Renard added.

Unsupervised Overnight Home Use

Although there have been a lot of recent papers published about the artificial pancreas, Dr. Thabit said the novelty of the current analysis is that the patients received no supervision or remote monitoring.

It combined results from 3 studies in 24 adults with a mean age of 43 and 16 adolescents with a mean age of 15.6, all with a mean HbA1c of about 8%.

After receiving training on sensor-augmented insulin-pump therapy and on the artificial pancreas, adults were assigned to receive 4 weeks of sensor-augmented pump therapy with the artificial pancreas and 4 weeks of pump therapy without the artificial pancreas, in a random order. Adolescents received 3 weeks of each type of overnight insulin-delivery therapy.

The closed-loop device in the artificial pancreas uses a model predictive control algorithm to direct insulin delivery.

The primary outcome was time when sensor glucose was in the target range between 3.9 and 8.0 mmol/L. All analyses were by intention to treat.

The closed-loop insulin-delivery device was started by participants on their own on 866 nights (89%). The amount of time that sensor glucose was in the target range (3.9–8.0 mmol/L) between midnight and 8 am was increased from 40.5% to 59.2% of the time (P < .001) when patients were using the artificial pancreas.

The artificial pancreas also significantly reduced the mean overnight glucose by 0.8 mmol/L (P < .001), with no difference in glycemic variability.

And with the artificial pancreas, the amount of time that patients had glucose levels above the upper target limit (hyperglycemia) decreased from 53.8% of the time to 37.9% of the time (P = .001).

Similarly, the amount of time they had glucose levels below 3.9 mmol/L (hypoglycemia) decreased from 2.9% of the time to 1.9% of the time (P = .014) during use of the artificial pancreas.

The lower mean overnight glucose while wearing the artificial pancreas was brought about by increased overnight insulin delivery from 6.0 to 7.0 units (P < .001) without changing the total mean daily delivery of 39.4 to 40.3 units (P = .84).

Looking Ahead

Next, the researchers aim to perfect the connectivity between the device components — the sensor, the control algorithm, and the insulin pump — and investigate even longer use, as well as daytime use.

"The next decade will see closed-loop systems progressively entering clinical practice and helping in the management of type 1 diabetes," the researchers predict. "Our current priority is to demonstrate benefits of closed-loop systems when used over several months and in various subpopulations."

The study was supported by Diabetes UK and JDRF. Dr. Thabit disclosed that he received speaker honoraria from AstraZeneca.

European Association for the Study of Diabetes 2014; September 18, 2014; Vienna, Austria. Abstract 194


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.