Background
Studies have shown that overnight closed-loop insulin delivery can improve glucose control and reduce the risk of hypoglycemia and hence may improve metabolic outcomes and reduce burden for children with type 1 diabetes and their families. However, research so far has not reported insulin levels while comparing closed-loop to open-loop insulin delivery in children. Therefore, in this study we obtained glucose levels as well as plasma insulin levels in children with type 1 diabetes to evaluate the efficacy of a model—based closed-loop algorithm compared to an open-loop administration.
Methods
Fifteen children with type 1 diabetes, 6–12 years, participated in this open-label single center study. We used a randomized cross over design in which we compared overnight closed-loop insulin delivery with sensor augmented pump therapy for two nights in both the hospital and at home (i.e., 1 night in-patient stay and at home per treatment condition). Only during the in-patient stay, hourly plasma insulin and blood glucose levels were assessed and are reported in this paper.
Results
Results of paired sample
t
-tests revealed that although plasma insulin levels were significantly lower during the closed-loop than in the open-loop (Mean difference 36.51 pmol/l;
t
(13) = 2.13,
p
= .03, effect size
d
= 0.57), blood glucose levels did not vary between conditions (mean difference 0.76 mmol/l;
t
(13) = 1.24,
p
= .12,
d
= 0.37). The administered dose of insulin was significantly lower during the closed-loop compared with the open-loop
(
mean difference 0.10 UI;
t
(12) = 2.45,
p
= .02,
d
= 0.68).
Conclusions
Lower insulin doses were delivered in the closed-loop, resulting in lower plasma insulin levels, whereby glucose levels were not affected negatively. This suggests that the closed-loop administration is better targeted and hence could be more effective.