ADA: Closed-loop Artificial Pancreas Works

According to Andrew Dauber, MD, of Boston Children's Hospital, and colleagues, overall time in the target range for overnight glucose wasn't significantly different between groups, but significantly fewer children -- all of whom were under age 7 -- spent time above 300 mg/dL during the night with the closed-loop system compared with open-loop delivery. "We think kids under 7 are particularly amenable to closed-loop control, since they have unpredictable eating patterns, erratic activity levels, and higher rates of hypoglycemia and nocturnal hypoglycemia," Dauber said while reporting the findings during a late-breaking session at the American Diabetes Association meeting.

However, he noted that current algorithms are designed for older patients, and that one "designed for a 40-year-old won't necessarily work on a 4-year-old."

So he and colleagues developed a model in which insulin was administered via a pump that was operated manually by a physician who interpreted data from continuous glucose monitors. The researchers used two monitors per patient in this trial in case the child accidentally pulled one off.

All 10 patients were under age 7, had type 1 diabetes for at least 6 months, and had been on pump therapy for at least 6 weeks.

The 2-day, in-hospital, cross-over trial involved children being admitted at 3 p.m. with an afternoon snack, followed by dinner at 5 p.m., and another snack at 8 p.m. Open- or closed-loop delivery began at 10 p.m. and continued overnight. The next night, the children crossed over to the opposite form of therapy. Overnight target blood sugar was 150 mg/dL, and was 120 mg/dL during the day.

The mean age was 5.1 years. Patients had diabetes for about 2 years, a mean HbA1c of 8.1%, and an average daily insulin dose of 0.72 units/kg.

Dauber and colleagues found that the time in the target range for glucose overnight (110 mg/dL to 200 mg/dL) wasn't significantly different between groups, but significantly fewer kids spent time above 300 mg/dL overnight with closed-loop delivery (0.18% versus 1.3%, P=0.035).

"Overnight closed-loop values were closer to target" despite the nonsignificant difference, Dauber noted.

Kids also had better lunchtime glycemic control with closed-loop delivery (189 mg/dL versus 273 mg/dL, P=0.009), and there were no significant differences in terms of peak post-prandial glucose, even though kids didn't receive a pre-meal insulin bolus during closed-loop therapy -- just in case the child didn't eat the full meal. "That's quite impressive considering we gave no pre-meal insulin," he added.

Finally, the number of interventions for hypoglycemia was the same in both groups, leading the researchers to conclude that closed-loop therapy "decreases the degree of nocturnal hyperglycemia in young children without increasing the incidence of hypoglycemia," Dauber said.

"Closed-loop therapy has the potential to improve diabetes care for very young children," he added. "We believe this is an underrepresented group in this area of research."

• In this study, overnight delivery of insulin via a pump in young children did not change the time in the target range for glucose, but significantly fewer kids spent time above 300 mg/dL overnight with closed-loop delivery.
• This study was published as an abstract and presented at the ADA conference. These data and conclusions should be considered to be preliminary until published in a peer-reviewed journal.

Dauber A, et al "Closed-loop insulin therapy improves nocturnal glycemic control in children under 7 years" ADA 2012; Abstract 153LBOR.