Expert Blog

Can We 'Close the Loop' With an Artificial Pancreas?

Aaron Kowalski, PhD
Diabetes Researcher
For many years now, the concept of an artificial pancreas – a machine that could replace the functions of the pancreas lost due to diabetes – has been appealing and sought after. Conceptually, it is so simple, and I often get asked: “If we could land a man on the moon 40 years ago or if we land jumbo jets with autopilots, why can’t we simply have a computer help dose insulin?” The answer is both simple and complex. Here, I’ll describe the challenges, why I think we can “close the loop” in the near future, and the flurry of clinical research activity that is moving this field forward.

The concept of an artificial pancreas is quite simple: We measure glucose levels and dose the appropriate amount of insulin at the right time based upon those levels. Today, we envision four key components to an artificial pancreas: an insulin pump, a continuous glucose monitor, a control algorithm, and insulin. So why the hold up? It really boils down to one overarching and critical concern – safety. We need to make sure that any such artificial pancreas system is safe and that we don’t cause hypoglycemia through a malfunctioning system. That said, we do know today that diabetes comes with significant risks, and the lifetime risk of death due to hypoglycemia is still very high, with recent estimates of 6-10% lifetime risk (Cryer et al., 2011).[1] Furthermore, we can and must do better in reducing hyperglycemia exposure. In the JDRF Continuous Glucose Monitor (CGM) trial (JDRF Study Group, 2008) in which I was fortunate to participate on the Operations and Steering Committee, the data was quite startling.[2] Even CGM users with good A1c levels (in the low 7’s) still spend eight to 10 hours a day with hyperglycemia (glucose levels above 180 mg/dL). Real-time CGM is a powerful tool to help improve glucose control, but data supports that adding computer-controlled automation may revolutionize diabetes management.

JDRF launched the Artificial Pancreas Project in 2005, and we have since seen tremendous progress. Many studies in the United States and abroad have shown that artificial pancreas systems can automatically reduce both hyperglycemia and hypoglycemia exposure (please visit for an extensive list of publications from the JDRF Artificial Pancreas Consortium). For example, Buckingham et al. (2009) showed that using predictive algorithms to suspend insulin delivery could prevent up to 84% of nocturnal hypoglycemia events and Hovorka et al. (2010) demonstrated that an artificial pancreas system could significantly increase the time children spend in the normal blood sugar range, reduce hypoglycemia exposure and reduce risk for severe hypoglycemia.[3][4]

The next big step that will happen this year in artificial pancreas research is the move to outpatient trials of these systems. JDRF has been working closely with the clinical community and the Food and Drug Administration (FDA) to determine a clear and reasonable pathway – with a particular focus on the safety of these trials – so that we can test these artificial pancreas systems in real-world settings. We are pleased that the FDA has committed to issuing draft guidance to industry and academia on federal requirements for such studies, with an eye on ultimately making safe and effective products available to patients with diabetes.
The potential impact of the artificial pancreas is huge. Recent studies have demonstrated that artificial pancreas technologies could both lower the risk of dysglycemia and ease the burden of diabetes for patients. Many studies are justifying the optimism that the long-sought realization of a “closed-loop” system is near. I look forward to the FDA producing draft guidance this fall, and to the first outpatient studies this year. The future for diabetes technologies is bright.