The artificial pancreas—a promising approach for the treatment of type 1 diabetes mellitus (T1DM)—has demonstrated superiority over conventional insulin pump therapy in clinical trials; however, until now, the benefits, if any, of the dual-hormone system (which delivers glucagon and insulin) over the single-hormone system (which delivers insulin only) were unknown. A new study reports the first direct comparison of the two systems and shows that the dual-hormone artificial pancreas does not significantly increase the proportion of time spent in the target range of glucose levels over that already achieved by the single-hormone system. The findings suggest that the single-hormone artificial pancreas is probably sufficient for hypoglycaemia-free overnight glycaemic control.

The researchers, led by Ahmad Haidar, conducted an open-label randomized three-way controlled crossover trial, which compared the dual-hormone artificial pancreas, the single-hormone artificial pancreas and conventional insulin pump therapy in patients with T1DM who were aged 12 years or older. Participants were assigned to one of the three interventions in random order and attended an inpatient facility for three 24 h periods. The primary end point of the study was the percentage of time during which the patients' plasma glucose levels were in the target range (4.0–10.0 mmol/l for 2 h postprandially and 4.0–8.0 mmol/l otherwise).

Credit: Modified from johavel/iStock/Thinkstock

Among the 40 patients enrolled initially, 30 patients (adults and children) completed the study. The mean proportion of time the patients' plasma glucose levels were in the target range over 24 h was 63%, 62% and 51% for patients using the dual-hormone artificial pancreas, the single-hormone artificial pancreas and conventional insulin pump therapy, respectively. Although both artificial pancreas systems provided better glycaemic control than conventional insulin pump therapy, no statistically significant difference was observed in the time spent in the glycaemic target range between patients who used the dual-hormone system and those who used the single-hormone system.

Compared with conventional insulin pump therapy, the single-hormone artificial pancreas reduced the proportion of time spent in hypoglycaemia (plasma glucose levels <4.0 mmol/l) from 13.3% to 3.1%. Interestingly, the dual-hormone artificial pancreas reduced this time even more, to 1.5%, but there was little evidence of glucagon benefits in reducing the number of hypoglycaemic events requiring treatment. Boris Kovatchev of the University of Virginia Center for Diabetes Technology, Charlottesville, VA, USA, who was not involved in the study, comments: “This direct comparison study confirms what has been long suspected and anecdotally reported in the field: namely, that adding glucagon to an insulin control system does not significantly improve average glycaemia or control metrics such as time within the target range.” He continues: “At best, closed-loop control involving glucagon has the potential to marginally reduce mild hypoglycaemia, but there is no evidence that it can reduce significant hypoglycaemic episodes.” In response, Haidar explains: “Our trial was not powered to assess reductions in significant hypoglycaemic events. Moreover, the dual-hormone artificial pancreas was designed to deliver the same amount of insulin as the single-hormone artificial pancreas and thus we would not have expected reductions in average glycaemia with the dual-hormone system.” He continues: “Our findings should not be interpreted as an absence of benefits of glucagon.”

“The next step is to conduct studies over multiple days in an outpatient setting; we have actually just finalized two outpatient studies focusing on night control of glycaemia (over multiple nights), and we have a day and night (60 h) study currently ongoing,” explains Haidar. Kovatchev agrees that further studies are warranted, but in addition to evaluating glucose control he recommends that future studies collect metrics about system usability, the logistics of glucagon delivery, the complexity of system operation, and the patients' perception of the relative efficacy of the two artificial pancreas systems.

Overall, the findings confirm the superiority of the artificial pancreas over conventional insulin pump therapy. Whilst Haidar believes that the results will accelerate the development of the dual-hormone artificial pancreas, he concedes that this course of action will depend on the development of a stable glucagon formulation and a dual-chamber pump.