Multinight "bedside" closed-loop control for patients with type 1 diabetes

Abstract

Background: Studies of closed-loop control (CLC) systems have improved glucose levels in patients with type 1 diabetes. In this study we test a new CLC concept aiming to "reset" the patient overnight to near-normoglycemia each morning, for several consecutive nights.

Subjects and methods: Ten insulin pump users with type 1 diabetes (mean age, 46.4±8.5 years) were enrolled in a two-center (in the United States and Italy) randomized crossover trial comparing 5 consecutive nights of CLC (23:00-07:00 h) in an outpatient setting versus sensor-augmented insulin pump therapy of the same duration at home. Primary end points included time spent in 80-140 mg/dL as measured by continuous glucose monitoring overnight and fasting blood glucose distribution at 7:00 h.

Results: Compared with sensor-augmented pump therapy, CLC improved significantly time spent between 80 and 140 mg/dL (54.5% vs. 32.2%; P<0.001) and between 70 and 180 mg/dL (85.4% vs. 59.1%; P<0.001); CLC reduced the mean glucose level at 07:00 h (119.3 vs. 152.9 mg/dL; P<0.001) and overnight mean glucose level (139.0 vs. 170.3 mg/dL; P<0.001) using a marginally lower amount of insulin (6.1 vs. 6.8 units; P=0.1). Tighter overnight control led to improved daytime control on the next day: the overnight/next-day control correlation was r=0.52, P<0.01.

Conclusions: Multinight CLC of insulin delivery (artificial pancreas) results in significant improvement in morning and overnight glucose levels and time in target range, with the potential to improve daytime control when glucose levels were "reset" to near-normoglycemia each morning.

FIG. 1.

Glycemic excursions (mean and quartiles) measured with the continuous glucose monitor overnight. CLC, closed-loop control (black region, solid line); SAP, sensor-augmented pump (gray region, dotted line).

FIG. 2.

Distribution of fasting glucose levels measured by the continuous glucose monitor at 07:00 h. CLC, closed-loop control (black columns); SAP, sensor-augmented pump (gray columns).

FIG. 3.

(Lower panel) Average glucose and (upper panel) corresponding total injected insulin for closed-loop control (CLC) (black columns) and sensor-augmented pump (SAP) (gray panels), by 2-h increments at night.