Automated Insulin Delivery System ‘Getting Better and Better’


Medtronic’s next-generation automated insulin delivery system offers significant improvements over the currently available 670G hybrid closed-loop, particularly in young people with type 1 diabetes, new data suggest.

Automated insulin delivery systems are comprised of an insulin pump, continuous glucose monitor (CGM), and an automated insulin dosing algorithm.

Data from three trials of such systems using Medtronic’s advanced hybrid closed-loop (AHCL) algorithm (tradename SmartGuard) were presented June 12 during the virtual American Diabetes Association (ADA) 80th Scientific Sessions. The AHCL is the algorithm used in Medtronic’s new MiniMed 780G system, which received a CE Mark on June 11 for the treatment of type 1 diabetes in people aged 7 to 80 years.

One trial, presented by Bruce W. Bode, MD, of Atlanta Diabetes Associates, Georgia, was the US pivotal safety study that will be submitted to the US Food and Drug Administration for approval of the Medtronic 780G.

Another trial, presented by Richard M. Bergenstal, MD, executive director of the International Diabetes Center at Park Nicollet, Minneapolis, Minnesota, was a separate comparison of the AHCL with the 670G. (The AHCL-based system used in the three trials was identical to the 780G except it didn’t include Bluetooth, which will be a feature of the final product.) 

A third trial, presented by Martin de Bock, PhD, of the University of Otago, New Zealand, included the CE Mark dataset for the 780G.

In contrast to the 670G, the 780G adds automated correction boluses for high blood glucose levels (rather than simply adjusting the basal infusion) and allows for adjustment of target glucose levels down to 100 mg/dL rather than a minimum of 120 mg/dL.

Taken together, the data from the three trials showed that the AHCL-based system improved glycemic time-in-range with no increased risk for hypoglycemia, including in children and teenagers, with high patient-reported satisfaction. And specifically compared to the 670G, the AHCL-based system reverts to open-loop far less often because it only exits closed-loop mode when the sensor stops working or during sensor changes, but not during hyperglycemia even above 300 mg/dL.

Asked to comment, session moderator Timothy S. Bailey, MD, president and CEO of the AMCR Institute, Escondido, California, told Medscape Medical News: “Automated insulin delivery systems are getting better and better.”

“None of these devices is perfect, but they are a substantial improvement over what we’ve had…They all take people from where they are now to better time-in-range, less time with hypoglycemia, and most important, they might make the quality of their lives better. That’s really underappreciated.”

One factor that has allowed for the improvements, Bailey said, is the recognition by regulatory bodies that the hybrid closed-loop devices are generally safer than current open-loop type 1 diabetes management so that fewer “safety” device features that interfere with tight glycemic control are necessary. 

With first-generation closed-loop systems, “If a wide variety of conditions occur, users get kicked off [hybrid closed-loop mode]. Originally it was perceived by the regulatory agencies as a safety feature because they perceived the standard of care as safe. The new system was allowed to have fewer rules.”

Pivotal Trial: Time-in-Range Improved, 96% Say System Easy to Use

The goal of the AHCL system is to maximize the time-in-range of blood glucose between 70-180 mg/dL. Automated basal delivery of insulin is programmed to a set-point of 100 or 120 mg/dL, with dosing every 5 minutes.

The US pivotal trial was a single-arm, 16-center, in-home trial of 157 people with type 1 diabetes, including 39 adolescents aged 14-21 years and 118 adults aged 22-75 years. All had type 1 diabetes for at least 2 years, A1c levels < 10%, and had been using insulin pumps for at least 6 months, with or without CGMs. 

After a 14-day run-in, they wore the systems with a 100 or 120 mg/dL set-point for 45 days, then switched to the other setpoint for another 45 days. Average A1c dropped from 7.5% to 7.0%, with the proportions having an A1c ≤ 7.0% increasing from 34% to 61%.

Overall time-in-range was 75% compared to 69% at baseline, with time below range (< 70 mg/dL) of 1.8%. Overnight time-in-range was 82%, with 1.5% below range. Time-in-range increased from 62% to 73% in the adolescents and from 71% to 75% in the adults. 

There were no incidences of severe hypoglycemia or diabetic ketoacidosis, and no device-related serious adverse events. 

Participants reported being in hybrid closed-loop, or auto-mode, 95% of the time, compared with 33% for those who had been previously using the 670G.

The number of AHCL exits was 1.3 per week, significantly less than with the 670G. Of those, 29% were user-initiated while the rest were implemented by the device, most often when the sensor wasn’t working.

In a study questionnaire, 96% reported that the system was easy to use.

AHCL vs 670G: Major Improvements Seen

Bergenstal presented data from the Fuzzy Logic Automated Insulin Regulation (FLAIR) study, funded by the National Institute of Diabetes and Digestive and Kidney Disease, comparing Medtronic’s AHCL-based system with the currently marketed 670G hybrid closed-loop, in 113 individuals with type 1 diabetes aged 14-29 years.

“This age group has traditionally been the most difficult group in which to optimize glucose management,” Bergenstal said.

FLAIR is believed to be the first-ever study comparing an investigational automated insulin delivery system with a commercially approved system, he noted. All participants used each automated insulin delivery system for 3 months in the randomized crossover trial.

The primary outcome, time spent above 180 mg/dL during the day combined with time below 54 mg/dL over 24 hours at baseline with the 670G and AHCL went from 42% to 37% to 34%, respectively, for the former and from 0.46% to 0.50% to 0.45%, respectively, for the latter. 

The percentage time-in-range over 24 hours went from 57% at baseline to 67% with the AHCL versus 63% with the 670G. A1c levels dropped from 7.9% at baseline to 7.6% with the 670G and 7.4% with AHCL.

“Remember, these are the adolescents who are the toughest of the tough, yet there was a 10% increase in time-in-range…this is very clinically significant,” Bergenstal said.

Even among 14 patients who had been using multiple daily injections without CGM prior to the study, a group often excluded from closed-loop studies, time-in-range improved from 45% at baseline to 63% with the 670G to 65% with AHCL.

“I’m making a plea not to exclude people just because they haven’t previously used technology,” Bergenstal said.

One patient who had dosed with extra insulin manually had a severe hypoglycemia event with AHCL. No patient had diabetic ketoacidosis.

The proportion of insulin given as auto-correction boluses was 36%, which is important as it means that the system was compensating for missed meal doses, a common phenomenon among teenagers, Bergenstal noted.

“There is still room for further improvement in glycemic control in this population of patients with type 1 diabetes, but AHCL represents a significant step forward,” he concluded.

New Zealand Study: More Data in Youth Show AHCL Benefits

Unlike the US study populations of just teens aged 14 and older, and adults, the study data used for approval in the EU — from New Zealand — included a total of 60 patients with 20 children aged 7-15 years. It, too, was a 10-week randomized crossover clinical trial comparing the AHCL to a sensor-augmented pump system with an algorithm only for predictive low-glucose management (PLGM) and no adjustments for high blood glucose.

Time-in-range was 59% at baseline and 58% with PLGM, compared to 70.4% with AHCL, and most of the time-in-range improvement occurred at night. Time below 70 mg/dL dropped from 3.1% to 2.5% to 2.1%, respectively.

Similar to the US studies, participants spent 96% of the time in closed-loop mode with only 1.2 exits per week. On a questionnaire, 95% of patients agreed that the system was easy to use and 85% that the system improved their quality of life.

De Bock showed a slide with some quotes, including one from a parent saying, “We didn’t have to be fearful at night or have that thought when we opened her bedroom door in the morning that she might not be conscious,” and from a patient, “I forgot I had diabetes today.”

Bailey commented: “Of course these devices are not free. So, the challenge is how do we make them available, less expensive, and easy to use? We have our work cut out for us, but this is heartening data. Everything has gotten better but we’re not out of a job yet.”

Bailey has reported receiving research support from Abbott, Capillary Biomedical, Dexcom, Diasome, Eli Lilly, Kowa, Lexicon, Medtronic, Medtrum, Novo Nordisk, REMD, Sanofi, Senseonics, ViaCyte, vTv Therapeutics, Zealand Pharma, and consulting or speaking honoraria from Abbott, LifeScan, Novo Nordisk, Sanofi, and Medtronic. Bode has reported receiving consulting and speaker fees from Medtronic. Bergenstal has reported participating in clinical research, being an advisory board member, and/or serving as a consultant for Abbott Diabetes Care, Ascensia, CeQure, Dexcom, Eli Lilly, Hygieia, Senseonics, and United Healthcare. De Bock has reported receiving honoraria or expenses from Novo Nordisk, Sanofi, Pfizer, Medtronic, and Lilly, and research funds from Novo Nordisk and Medtronic.

For more diabetes and endocrinology news, follow us on Twitter and Facebook.

Leave a Reply

Your email address will not be published. Required fields are marked *