The AP Workshop – a Meeting of the Great Minds Working to Close the Loop

the fda/nih/jdrf artificial pancreas workshop

A meeting organized by the Food and Drug Administration (FDA), the National Institutes of Health (NIH), and the Juvenile Diabetes Research Foundation (JDRF) was held from July 21-22 to discuss progress on closing the loop to an artificial pancreas. It brought together those in academia, industry, and government working to make the elusive “Holy Grail” of diabetes treatment into a reality.

closing the loop – what is the artificial pancreas?

An artificial pancreas refers to a device designed to respond appropriately to changes in glucose levels in the body, mimicking one of the roles of the normal pancreas. Maintaining glucose levels similar to those in people without diabetes has been shown to reduce major long-term complications of diabetes such as heart attacks, blindness, kidney failure and amputations.

The idea of closing the loop refers to the use of a continuous glucose monitor and an insulin pump that “talk” to each other and allow glucose levels to be appropriately controlled with little (partial closed-loop) or no (fully closed-loop) human intervention. Below are our takeaways about the progress being made toward an AP system.

do not let perfection get in the way of progress

This was an overwhelming theme, repeatedly emphasized by Dr. Aaron Kowalski, director of research at the JDRF. While we may not yet have the tools to create a fully closed loop, we have tools that work better and better as new generations are developed. One example of a first step is the automatic shutoff of basal insulin to reduce the incidence of severe hypoglycemia (See this issue’s NewNowNext). The next step could be a similar development to avoid hyperglycemia. Future steps could include night-time closed-loop since glucose levels are much steadier when people are not eating or moving too much. Additional stepwise approaches could involve the use of glucagon or Symlin (See Issue #2 and #8 Test Drive, and Issue #3 Learning Curve). It’s not a bad idea to crawl, then walk, and then run.

accuracy of current cgm devices may be an obstacle, but it’s not a roadblock

It is true that the accuracy of current sensors still calls for the use of blood glucose finger sticks as a confirmation of CGM glucose values before making insulin adjustments. However, researchers at the meeting pointed out that CGMs used in closed-loop systems will be used differently from current CGMs; glucose readings from closed-loop CGMs will be used to make frequent and very small insulin adjustments. The results of these adjustments on glucose will direct subsequent adjustments in an effort to keep patients within target zones. Closed-loop CGMs will also have fail-safe mechanisms built in to recognize and report potentially inaccurate glucose readings and allow human overrides of insulin dosing decisions. With these two factors in play, we understand that closed-loop development can continue in parallel with the development of improved sensor accuracy – we don’t have to wait to get perfect sensors to make progress on the closed-loop.

virtual testing of closed-loop algorithms may speed the development process

There was much discussion about in silico modeling of closed-loop algorithms. Algorithms are a fundamental part of the closed-loop. They are the instructions that translate glucose readings from the sensor (and other factors such as previous insulin doses) into insulin dosing instructions for the pump. These instructions (algorithms) must also be rigorously tested before the FDA can approve them. In silico modeling is a computer simulated test of these algorithms, and this test was recently cleared by the FDA as an adequate replacement for animal testing, which many believe will save a lot of time and money in getting to a suitable algorithm for a closed-loop. The simulator was open for use by any developer looking to test an algorithm. Dr. Boris Kovatchev (University of Virginia) and Dr. Frank Doyle (UC Santa Barbara) noted that the simulator is meant to challenge these algorithms by seeing the kind of insulin dosing instructions they provide when faced with varying conditions of glucose levels, exercise, ailments and many more. The most robust algorithms should be able to figure out safe actions to take when faced with challenging situations people with diabetes often encounter.

the roles of the jdrf, fda, and nih

In 2005, the JDRF foundation set up a project to speed up the development and availability of new diabetes technology for people with diabetes. During a press conference following the workshop, a number of researchers described the JDRF as instrumental in pulling together different groups in academia, industry, and government and promoting collaboration among these groups, each of which holds a unique part of the closed-loop puzzle.

In March 2006, the FDA included “New Therapies for Juvenile Diabetes” on its Critical Path Opportunities List. We applaud Arleen Pinkos, Chair of the Interagency Artificial Pancreas Working Group (headed by the FDA), and her team for all the hard work and support in bringing this important meeting together.

The NIH is well-known for its role in promoting research and advocacy about diabetes. The National Institute of Diabetes and Digestive and Kidney Disease (NIDDK) has partnered with the National Institue of Child Health and Human Development (NICDH) to create the DirecNet, with the goal of assessing CGM in children with diabetes and using CGM to study hypoglycemia in children. The NIDDK has worked to develop new low-cost accessible technologies relevant to this field. We are especially grateful to the NIH for hosting this meeting and providing a space for so many diabetes experts to come together.