• Diabetes Mellitus
• Diabetes & Endocrinology Clinical Trials
• Hyperglycemia/Hypoglycemia

Closed-loop systems consist of a continuous glucose sensor that is connected to a computer that contains a glucose control algorithm and this algorithm advises the rate of the insulin pump. Arianne van Bon, MD, Academic Medical Center, Amsterdam, The Netherlands, presented data comparing a bihormonal closed-loop system to control postprandial and post exercise glucose with an open-loop system (usual care, insulin dosing performed by the patient). Overall, the bihormonal closed-loop system (automated glucagon plus insulin delivery) successfully controlled the glucose values in type 1 diabetic patients.

The first prototype, (APPEL 1) was a pilot study that evaluated the feasibility of the bihormonal closed-loop system in 6 subjects with type 1 diabetes postprandially. The closed-loop consisted of subcutaneous continuous glucose monitor (CGM) based on microdialysis, self-learning proportional derivative algorithm built in a personal computer and two D-Tron+ pumps for subcutaneous insulin and glucagon administration [van Bon A et al. J Diabetes Sci Techno 2010].

The algorithm had three operating ranges: 1) administer insulin if the glucose level was >7 mmol/L; 2) add glucagon if the glucose level was <3.2 mmol/L; 3) issue an eating alert for glucose levels <5 mmol/L).

One subject was excluded due to technical failure of the CGM. Overall mean venous glucose values were similar between the two systems (open-loop 11.4 mmol/L [5.2 to 14.7]; closed-loop in 8.7 mmol/L [7.1 to 8.8]). There was an initial postprandial rise in glucose with the closed-loop system followed by a drop in glucose values to <5mmol/L. There were four hypoglycemic events (glucose <3.9 mmol/L) in the closed-loop group compared with one event in the open-loop group. These results showed that the technique was feasible, but adjustments were needed.

In the follow-up study (APPEL 2) adjustments were made: a needle type CGM was used instead of microdialysis CGM; two CGMs (one primary and one back up sensor) were used and there was a change in the algorithm. Insulin was administered if glucose levels >6.5 mmol/L, glucagon was given at glucose levels <6.5 mmol/L, and eating alerts were issued when glucose levels were <3.5mmol/L. Rescue glucagon bolus was given when glucose was <4.5 mmol/L. Also, exercise (30 minutes on a home trainer) was introduced to increase stress on the system. Included were 10 patients (8 men and 2 women, mean age 55.4 years) with type 1 diabetes treated with an insulin pump. Subjects had a mean HbA1C of 8.0%, mean duration of diabetes of 34.6 years, and mean pump use of 11.2 years. The glucose was controlled 2 hours after breakfast, during and one and a half hour post exercise, and four hours after lunch.

There were no overall differences in venous and sensor glucose concentrations between the closed-loop system and usual care (open-loop; Table 1). Significantly higher venous glucose levels (p=0.001) were noted in the closed-loop system post exercise. Significantly higher postprandial breakfast (p=0.001) and significantly lower post exercise (p=0.01) glucose concentrations (AUC) were noted in the closed-loop system. The postprandial lunch glucose control was not different. There were no incidents of severe hypoglycemia. Glucose levels <3.5 mmol/L were observed in two patients in the open-loop system and four in the closed-loop system. All patients were given glucagon. In conclusion, the closed-loop system was efficient particularly after lunch when all glucose levels were <6.5 mmol/L suggesting that the glucose level before the meal influences the performance of the algorithm.

• © 2011 MD Conference Express®