Abstract
C
The automated feedback control of closed-loop systems allows insulin delivery to be stepped up or modulated in response to sensor glucose level, allowing greater flexibility and compensation of miscalculated meal boluses. Is it, therefore, possible to omit bolusing for meals altogether in a closed-loop system? Several clinical studies have attempted to answer this question. 7,8 Although this approach is appealing from a user perspective, as it alleviates the burden of self-management, results from fully automated closed-loop studies, without meal bolusing or meal announcement, and carbohydrate counting, showed prolonged postmeal hyperglycemia and/or late postprandial hypoglycemia. 7,8 The risk of deskilling the user, in terms of self-management such as carbohydrate counting, may also arise, and although not of concern when closed loop is in use, this may become undesirable during “closed-loop holidays” when the user opts out from using sensor. A concessionary approach includes a simplified bolus strategy with reduced user input, while maintaining satisfactory glucose control. 6 Gingras et al. in this issue of Diabetes, Technology & Therapeutics 9 assessed the application of a simplified meal-bolus strategy compared against conventional counting, using both single- and dual-hormone systems in a randomized controlled outpatient clinical trial. The simplified bolus approach utilized a bolus calculator based on semiquantitative carbohydrate content sizes, for which a fixed amount of carbohydrate was assumed.
The authors reported a similar mean sensor glucose level between carbohydrate-matched and simplified bolus strategy while using single-hormone and dual-hormone closed-loop systems. Exploratory analyses showed that the simplified bolus strategy, despite the addition of glucagon during dual-hormone closed loop use, increased time spent in hypoglycemia, defined as sensor glucose <4.0 mmol/L, compared with sensor-augmented pump therapy. The authors concluded that further adjustments to the control algorithm are needed to support the simplified bolus approach, with either less aggressive insulin delivery or increased glucagon dosing. Larger sufficiently powered studies will be needed to validate these proposals.
The present study highlights that optimal postprandial management during closed loop remains a challenge. This may likely be due to the complexity of meal-induced glucose excursions, which can be affected by several factors such as insulin kinetics, including delayed insulin absorption and action, 10 timing and dosing of meal boluses, as well as amount and type of carbohydrate and other macronutrients such as fat and protein. 11 Alternative meal-bolusing strategies are currently being investigated by other closed-loop researchers. Some have adopted built-in conventional bolus calculators, requiring the user to count carbohydrates. 2,3 Others utilize more simplified approaches involving body weight-adjusted partial boluses, or categorizing meals according to portion size and carbohydrate content, or a combination of both. 6,7,9,12,13
In line with the findings of Gingras et al., one of the remaining challenges to closed-loop systems that utilizes simplified bolus approaches includes the potential of overbolusing and consequently increased risk of hypoglycemia. Episodes of hypoglycemia during closed loop may still occur because of delays in absorption and action of subcutaneous insulin, or user overcorrection resulting in excessive insulin on board. This has led to some closed-loop research groups to evaluate the use of dual-hormone systems as mitigation against further hypoglycemia. 5 In a head-to-head comparison between dual-hormone and single-hormone closed-loop systems, median time spent within target range from 4.0 to 10.0 mmol/L was comparable. Addition of glucagon tended to further reduce percentage of time spent in hypoglycemia below 4.0 mmol/L, although this was not significant at the corrected significance level. 4 This may partly be due to the fact that glucagon's effectiveness in preventing hypoglycemia may be limited in the presence of excessive insulin on board above 40 mU/L. 14 In addition, fluctuation in glycemia may also occur as a result of undesirable competition between insulin and glucagon in the postprandial period, highlighting the delicate balance of glucagon–insulin relationship. 5
The single-hormone approach is currently the furthest along in terms of development and nearing commercialization. 15 Current published evidence from single-hormone studies is more comprehensive, as long-term studies have been performed under free-living unsupervised conditions and showed that extended use of a single-hormone closed-loop system increased time in target and improved HbA1c compared with sensor-augmented pump therapy. 2,3 The 670G system (Medtronic MiniMed), a hybrid single-hormone closed-loop system, in which the user additionally administers prandial insulin, is seeking premarket approval from the U.S. Food and Drug Administration. 16 Dual-hormone closed-loop systems have shown increased time in target and reduction in hypoglycemia compared with insulin pump therapy; 6 however, these studies were of relatively short duration and performed under supervised conditions with restricted geographical movement, limiting applicability and generalizability to daily life. The increased complexity and costs may delay dual-hormone systems coming to the market. Finally, other hurdles to commercialization of dual-hormone systems still remain. Glucagon is an unstable protein in aqueous solution and conventional formulation is unsuitable for use in a portable pump. Multiple efforts are underway to create a stable liquid glucagon formulation and dual-chamber pumps are being developed. 17
Thus the question remains whether the addition of glucagon to closed-loop insulin delivery will be needed for the wider general type 1 diabetes population. Pharmaceutical innovations with ultrarapid insulin 18 that better emulate the physiological timing of prandial insulin action and adjunctive therapies, for example, amylin and GLP-1 analogues, which attenuate postprandial glucose excursions, 19 may potentially reduce the need or incremental benefits of dual-hormone closed-loop systems. However, dual-hormone closed-loop systems may still have a role in a specific situation such as exercise. During exercise, glucose levels can fall rapidly due to administered subcutaneous insulin onboard as well as increased peripheral glucose uptake in response to insulin-independent translocation of GLUT-4. 20 Exercise-induced insulin sensitivity may persist for several hours, leading to delayed onset hypoglycemia. This is compounded by blunted counter-regulatory responses due to antecedent exercise. 21 Glucagon due to its immediate glucose-raising effect 22 may prevent exercise-induced hypoglycemia, especially when exercise was announced to the control algorithm by the user. 23 Glucagon effectiveness, however, may be reduced in patients with dysfunctional glycogen metabolism due to poor glycemic control 24 or diet-related low glycogen stores. 25 Possible side effects of glucagon include nausea and vomiting. A possible appetite-suppressive effect of glucagon may have contributed to significantly lower breakfast intake during dual-hormone closed loop use in an outpatient trial of preadolescents with type 1 diabetes. 26 Glucagon use was generally well tolerated, however, with comparable occurrence of gastrointestinal symptoms. Safety of chronic glucagon delivery needs further assessment in long-term studies, as studies have linked in vivo glucagon exposure with infusion site reaction, electrolyte disturbance, lipolysis, and cardiac toxicity. 27 –30
In conclusion, significant progress has been made, with a hybrid single-hormone system expected to be clinically available soon. Closed-loop systems may allow flexibility and compensation of user errors during prandial bolusing. Improvements in control algorithms (e.g., with enhanced learning adaptation) are still needed along with pharmaceutical innovations to achieve optimal postprandial control with simplified bolus approaches. Although dual-hormone systems have shown favorable trends for hypoglycemia mitigation, the use of glucagon may be limited to specific situations, and requires longer “real-world” studies to confirm its usability and efficacy in clinical practice.
Footnotes
Author Disclosure Statement
R.H. reports having received speaker honoraria from Eli Lilly and Novo Nordisk, serving on advisory panel for Eli Lilly and Novo Nordisk; receiving license fees from BBraun and Medtronic; having served as a consultant to BBraun, and patents and patent applications related to closed-loop. L.B. and H.T. declare no competing financial interests exist.