Night Hypoglycemia Cut as Artificial Pancreas Inches Closer

Miriam E. Tucker

June 22, 2013

CHICAGO — A mechanism that suspends insulin-pump delivery when glucose levels fall too low reduced nighttime hypoglycemic episodes without increasing glycated hemoglobin (HbA1c) levels in patients with type 1 diabetes, a new large open-label randomized controlled trial has found.

Data from the 19-site Medtronic-funded Automation to Simulate Pancreatic Insulin Response (ASPIRE) in-home study were presented by Richard M. Bergenstal, MD, in a poster and at a press briefing here at the annual American Diabetes Association (ADA) 2013 Scientific Sessions and were simultaneously published online June 22 in the New England Journal of Medicine.

The device tested in ASPIRE involves investigational "threshold-suspend" software added to a combined insulin-pump and glucose-sensor system.

With this feature, the pump suspends insulin delivery for 2 hours when the sensor signals a preset low glucose value.

"This is the first step of putting intelligence in the pump to get rid of those severe, dangerous, and devastating low blood sugars," Dr. Bergenstal commented. "I've been saying for 30 years now that we'll have an artificial pancreas in the next 5 to 10 years, and I'm finally confident that it's really going to happen because we have the first phase of it now."

The ASPIRE study represents a major advance, Dr. Bergenstal said. "I think this is really important because it's the first step that shows it's really doable in an outpatient setting in a large number of patients in a well-designed study."

Press briefing moderator Anne L. Peters, MD, CDE, professor of clinical medicine and director, clinical diabetes programs, Keck School of Medicine, University of Southern California, Los Angeles, told Medscape Medical News, "To me this is a very important first step. I'm always a skeptic with new technology… But now that I know it's coming and I've been looking at my patients' [glucose-sensor] tracings, I know that many, many, many of my patients would be able to qualify and I think would benefit."

She cautioned, however, that it will likely still take some time before a fully automated "artificial pancreas" becomes available.

Hypoglycemia Suspend Feature First of a Number of Steps

Dr. Richard M. Bergenstal

The hypoglycemia-suspension feature is viewed as the first of many incremental technological steps toward the ultimate development of a fully automated artificial pancreas, in which direct communication between sensor and pump normalizes glucose levels without the need for user input, a process dubbed "closing the loop."

The product used in the current study, called the MiniMed 430G system, is now in the final phase of review by the US Food and Drug Administration (FDA), Dr. Bergenstal, executive director of the International Diabetes Center at Park Nicollet and clinical professor at the University of Minnesota Medical School, Minneapolis, told Medscape Medical News.

A similar feature known as "low glucose suspend" is already included in Medtronic's Veo sensor-augmented pump system, which is available in more than 50 countries worldwide but not yet in the United States.

Dr. Ann L. Peters

However, speaking to the concept of the fully automated artificial pancreas, Dr. Peters said that "From a research perspective, what a variety of the different groups working on the artificial pancreas are doing is phenomenal."

"But, people aren't going out and free-living with these devices… Until we really start doing those kinds of studies, we're not going to be able to say it's anywhere near getting [a fully automated device] on the market," she noted.

Key in getting to this final stage will be many other studies demonstrating the feasibility of further steps along that pathway. A number of such studies, albeit small ones, were also presented at the ADA meeting and included devices with features that predict low glucose and mitigate hyperglycemia and bihormonal models that incorporate glucagon in addition to insulin.

But Dr. Peters warned that all such systems will "require that people wear [a] sensor all the time… If you don't wear the sensor, it's not going to function." Studies have suggested that a majority of patients who have continuous glucose sensors don't wear them all the time, she noted.

ASPIRE: Reducing Lows Without Raising HbA1c

In ASPIRE, 247 patients 16 to 70 years of age with type 1 diabetes were randomized to wear the Medtronic system with or without the threshold-suspend software feature for 3 months. All of the patients had experienced 2 or more episodes of nocturnal hypoglycemia during a run-in phase, defined as sensor glucose values of 65 mg/dL or below lasting more than 20 minutes from 10:00 PM to 8:00 AM, with no evidence that the patient had interacted with the pump.

The sensor glucose threshold for suspension was initially set at 70 mg/dL but could be adjusted thereafter from 70 to 90 mg/dL.

The mean area under the curve for nocturnal hypoglycemia events — the primary efficacy end point — was 38% lower among the 121 patients who had the threshold-suspend feature compared with the 126 controls, a significant difference (P < .001). The number of hypoglycemia events per week was also significantly reduced, by 32% during the night (P < .001) and 30% for day and night combined (P < .001).

Reductions were seen at all levels of hypoglycemia with the threshold suspend, with episodes of less than 50 mg/dL dropping by 57% at night and by 53% for day and night combined.

And addressing a concern that had been expressed by the FDA, the 2-hour pump-suspension period did not result in excessively high glucose values. At the time of insulin-delivery resumption, the mean glucose value was 92.6 mg/dL. Two hours later, it had risen to a mean of 168.8 mg/dL, Dr. Bergenstal reported.

Indeed, there was no between-group difference in HbA1c, the primary safety outcome. The HbA1c barely changed in the threshold-suspend group (7.26% to 7.24%), while dropping slightly among the controls (7.21% to 7.14%). Those differences did not exceed the noninferiority threshold of 0.4%, he noted.

There were no significant differences between the groups in total insulin dosage, basal/bolus ratio, number of boluses, mean bolus size, use of the bolus calculator, sensor wear, sensor life, sensor calibrations, or change in body weight.

Nor did ketoacidosis occur in either group. Severe hypoglycemia — defined as coma or seizures or requiring medical assistance — occurred in 4 control patients, but in none of the threshold-suspend group.

Small Studies Addressing Further Steps

Speaking about the field in its entirety, Dr. Bergenstal said the Juvenile Diabetes Research Foundation's (JDRF's) Artificial Pancreas Project envisions the technology emerging in a series of steps.

First, this will involve the threshold/low-glucose–suspend feature. Second will be hypoglycemia minimization, in which the system predicts hypoglycemia and suspends insulin-delivery proactively. Next will be hypoglycemia/hyperglycemia minimization, which adds insulin delivery at glucose levels above a high threshold.

Fourth will be automated basal/hybrid closed loop, which requires only meal-time manual bolusing. Then will come the fully automated insulin closed loop, in which the manual meal-time bolus is eliminated, and finally, the fully-automated multihormone closed loop, including both insulin and glucagon.

Small feasibility studies investigating several of those steps were presented separately at the ADA meeting. Among those targeting the final step in JDRF's pathway was a trial demonstrating the effectiveness for up to 51 hours of an autonomously adaptive bihormonal (insulin and glucagon) "bionic" pancreas in 12 children and 12 adult volunteers with type 1 diabetes. That study tested the system's response to 6 carbohydrate-rich meals and a period of structured exercise.

Another study, from France, showed the feasibility of an intraperitoneally implanted artificial pancreas system in 8 patients, aimed at bypassing the problem of slow absorption of subcutaneously infused insulin.

Research looking at earlier steps in the process included a study of 10 patients in which a hyperglycemia-mitigation algorithm was added to the pump/sensor system to create a "control-to-range" process for the entire day. Another trial, of 20 adult patients, demonstrated the feasibility of a "hypoglycemia-hyperglycemia minimizer" for overnight glucose control.

"It's going to be an evolutionary process. There are really good people working on all phases of this," Dr. Bergenstal told Medscape Medical News.

And he advised physicians and patients to "take advantage of the advances as they come. You don't have to wait for the final product that does everything. If you can prevent serious blood sugars at night, do it."

Dr. Peters told Medscape Medical News that she hopes the ASPIRE data will at least help convince Medicare to cover continuous-glucose monitors, given that elderly patients with type 1 diabetes are at particular risk for hypoglycemia. "I really hope that this can help us lobby to get Medicare to cover sensors in type 1s. I can't imagine them denying it if we have proven that we can reduce rates of severe lows."

Dr. Bergenstal disclosed that he is on the scientific advisory board for Medtronic and for several companies that make glucose monitors;  the money goes to his institution. Dr. Peters serves on an ad hoc advisory board for Medtronic and has given talks for Dexcom in the past.

N Engl J Med . Published online June 22, 2013. Article

American Diabetes Association 2013 Scientific Sessions. Abstract 48-LB, Abstract 15-OR , Abstract 16-OR , Abstract 9-OR, Abstract 10-OR, presented June 22, 2103.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.