Stepwise progress toward an artificial pancreas.

KEYSTONE, COLO. -- Regulatory approval of a closed-loop medical equipment-based artificial pancreas will probably come in stages--and the first piece to get the nod will likely be automatic shutoff of an insulin pump in response to a continuous glucose monitor's nighttime warning of a pending low blood sugar level, Dr. H. Peter Chase predicted.

In early July, the Food and Drug Administration approved a 450-page Investigational Device Exemption cowritten by Dr. Chase for study of five potential mathematical control algorithms designed to do exactly that.

[ILLUSTRATION OMITTED]

The project is supported by the Juvenile Diabetes Research Foundation, which has declared the development of an artificial pancreas to be a top priority.

"I don't think initially the FDA is going to allow insulin to be given on the basis of continuous glucose monitor values. So that part of the closed loop is probably 5-10 years away. But for safety, given that there are deaths every year from hypoglycemia, I think the FDA will allow us to turn off a pump if the sensor is giving an alarm of a pending low, just to prevent a severe hypoglycemic episode or death," said Dr. Chase, professor of pediatrics at the University of Colorado, Denver.

He and his colleagues have shown that diabetic teens using an external continuous glucose monitor (CGM) measuring interstitial glucose often fail to respond to the sensor's nighttime alarm of a pending low blood sugar level. They doze fight through the warning, or partially awaken and then quickly slip back to sleep before they can take corrective action.

In preliminary studies conducted in 42 patients in Denver and at Stanford (Calif.) University, the investigators have shown that programming a pump to turn off for 90 minutes in response to a CGM alarm of a pending low enables 80% of hypoglycemic episodes to be avoided. Reassuringly, patients don't become ketotic when the pump is off for that length of time, Dr. Chase continued at a conference on the management of diabetes in youth.

Audience members argued that it would be more physiologic to program the pump to give a fraction of the normal insulin dose in response to a pending low, rather than shutting the pump down altogether. Dr. Chase agreed, but noted that FDA approval of the Investigational Device Exemption for temporary pump shutdown required exhaustive documentation running to 450 pages.

"If you want to talk about also having the pump deliver 10% or 50% of the normal insulin, you're up to around 1,300 pages. We're going to have to look at one thing at a time, but eventually we'll get there," he said.

Diabetes experts are eagerly anticipating the development of a safe and reliable artificial pancreas featuring auto control of the external insulin pump in response to real-time CGM data. They expect that closing the loop and removing the human factor will result in greatly improved blood glucose control and a resultant marked reduction in the major long-term complications of diabetes, along with a new sense of freedom for patients because of their lessened disease-management burden.

Although the artificial pancreas is more realistic than islet cell transplantation--with its attendant immunosuppressive therapy--as a solution to the traditional problems of diabetes management, improvements in CGM technology will be a prerequisite, Dr. Chase stressed. Sensor accuracy needs to be boosted. And the CGM calibration process must be simplified; patients now have to perform a finger-stick blood glucose test several times per day to reset the sensor or else risk getting misleading results.

"CGM is a work in progress," he asserted. "It can add a lot, and yet today we're not anywhere near where we are in terms of blood glucose monitoring and other aspects of diabetes care."

[ILLUSTRATION OMITTED]

Another speaker, Dr. Marian Rewers, forecast that creation of a true closed-loop artificial pancreas is going to take longer than many observers have hoped. That's because it will likely require development of a glucagon-releasing system functioning akin to the pancreatic alpha cell, in addition to insulin release in response to CGM data.

"What we are now trying to accomplish is an artificial beta cell, not an artificial pancreas," he said. "People without diabetes can control blood glucose levels so tightly because we have both beta and alpha cells working on it. I think we'll see more development in terms of adding glucagon, perhaps in a dual-channel pump approach.

"This will be needed for two reasons: to improve the first phase of insulin secretion and to protect patients from postprandial hypoglycemia," continued Dr. Rewers, who is a professor of pediatrics and preventive medicine at the University of Colorado, Denver, and clinical director of the university's Barbara Davis Center for Childhood Diabetes, which cosponsored the conference along with the Children's Diabetes Foundation at Denver and the university.

Considerable excitement at the meeting surrounded a recently published pilot study comparing a fully automated closed-loop insulin delivery system to a semiautomated one. The Yale University study involved 17 adolescent pump and CGM users with well-controlled type 1 diabetes who were assigned to 24 hours of either fully closed-loop blood glucose control or semiautomated hybrid control, which involved auto control supplemented by small manual priming insulin boluses given 15 minutes before meals in order to improve postprandial glycemic excursions.

The fully closed-loop artificial pancreas proved "feasible and remarkably effective," and the hybrid approach was even more so, according to the investigators (Diabetes Care 2008;31:934-9).

Of all CGM glucose measurements, 85% were in the target 70-180 mg/dL range during fully closed-loop therapy, compared with just 58% obtained during standard open-loop pump therapy.

But peak postprandial glucose levels averaged 194 mg/dL in the group receiving small manual priming insulin boluses, significantly better than the 226 mg/dL with the fully closed loop.

BY BRUCE JANCIN

Denver Bureau