Format

Send to

Choose Destination
J Diabetes Complications. 2015 Jan-Feb;29(1):93-8. doi: 10.1016/j.jdiacomp.2014.09.001. Epub 2014 Sep 16.

Factors affecting the success of glucagon delivered during an automated closed-loop system in type 1 diabetes.

Author information

1
Harold Schnitzer Diabetes Center, Oregon Health and Science University, 3181 SW Sam Jackson Park Road, Portland, OR 97239.
2
Harold Schnitzer Diabetes Center, Oregon Health and Science University, 3181 SW Sam Jackson Park Road, Portland, OR 97239. Electronic address: elyoussj@ohsu.edu.
3
Department of Biomedical Engineering, Oregon Health and Science University, 33030 SW Bond Ave., Portland, OR 97239.
4
Oregon Institute of Occupational Health Sciences, Oregon Health and Science University, 3181 SW Sam Jackson Park Road, Portland, OR 97239.

Abstract

BACKGROUND:

In bi-hormonal closed-loop systems for treatment of diabetes, glucagon sometimes fails to prevent hypoglycemia. We evaluated glucagon responses during several closed-loop studies to determine factors, such as gain factors, responsible for glucagon success and failure.

METHODS:

We extracted data from four closed-loop studies, examining blood glucose excursions over the 50min after each glucagon dose and defining hypoglycemic failure as glucose values<60 mg/dl. Secondly, we evaluated hyperglycemic excursions within the same period, where glucose was>180 mg/dl. We evaluated several factors for association with rates of hypoglycemic failure or hyperglycemic excursion. These factors included age, weight, HbA1c, duration of diabetes, gender, automation of glucagon delivery, glucagon dose, proportional and derivative errors (PE and DE), insulin on board (IOB), night vs. day delivery, and point sensor accuracy.

RESULTS:

We analyzed a total of 251 glucagon deliveries during 59 closed-loop experiments performed on 48 subjects. Glucagon successfully maintained glucose within target (60-180 mg/dl) in 195 (78%) of instances with 40 (16%) hypoglycemic failures and 16 (6%) hyperglycemic excursions. A multivariate logistic regression model identified PE (p<0.001), DE (p<0.001), and IOB (p<0.001) as significant determinants of success in terms of avoiding hypoglycemia. Using a model of glucagon absorption and action, simulations suggested that the success rate for glucagon would be improved by giving an additional 0.8μg/kg.

CONCLUSION:

We conclude that glucagon fails to prevent hypoglycemia when it is given at a low glucose threshold and when glucose is falling steeply. We also confirm that high IOB significantly increases the risk for glucagon failures. Tuning of glucagon subsystem parameters may help reduce this risk.

KEYWORDS:

Artificial pancreas; Closed-Loop; Glucagon; Hypoglycemia; Type 1 diabetes

PMID:
25264232
PMCID:
PMC4281277
DOI:
10.1016/j.jdiacomp.2014.09.001
[Indexed for MEDLINE]
Free PMC Article

Supplemental Content

Full text links

Icon for Elsevier Science Icon for PubMed Central
Loading ...
Support Center