The Bionic Pancreas
I have received a number of excited notifications from patients this past week in regard to the bionic pancreas. As such, I did my due diligence to research this new device and noted the recent scholarly paper published by Russell, El-Khatib, Sinha, Magyar, MecKeon, Goergen, Balliro, Hilliard, Nathan, and Damiano in The New England Journal of Medicine entitled “Outpatient Glycemic Control with a Bionic Pancreas in Type 1 Diabetes” (DOI: 10.1056/NEJMoa1314474).
My understanding is that the senior author of the paper, Dr. Damiano, has a son with type 1 diabetes and was working to solve the artificial pancreas “problem” before his son matriculated in college. Indeed, Dr. Damiano has produced a fascinating device that attempts to control blood sugars by use of the two hormones insulin and glucagon. In type 1 diabetes, pancreatic islet cells are unable to produce enough insulin to lower blood sugars effectively, and secretion of glucagon, a hormone that increases glucose, is defective. As such, Dr. Damiano et al have devised a system by which an iPhone serves as the interface between two small pumps that run both insulin and glucagon.
As background, this device has been employed in hospital-based settings over a very limited time period and has not been used in typical outpatient conditions over multiple days. This study is the first to apply the technology in a real life day-to-day environment.
What were the methods?
The study was performed in 20 adults and 32 adolescents with type 1 diabetes using the random-order, crossover method. The authors compared blood sugar control with the bionic pancreas (bionic pancreas period) as described above with an insulin pump (control period) for five days. The “automatically adaptive algorithm of the bionic pancreas received data from a continuous glucose monitor to control subcutaneous delivery of insulin and glucagon.”
What were the results?
- The average blood sugar level over the five-day bionic pancreas period was 128mg/dl and average percentage of time with a blood sugar level less than 70mg/dl was four point eight to five
- After one day of “automatic adaption” by the bionic pancreas, the average glucose level on continuous monitoring was lower than the average level during the control period (133 versus 159 mg/dl, P=0.001).
- The percentage of time with a blood sugar less than 70mg/dl was lower (four point one versus seven point three percent, P=0.01).
- The average blood sugar over the five-day bionic pancreas period was also lower than the control period (138 versus 157mg/dl, P=0.004).
- The percentage of time with a blood sugar less than 70mg/dl was similar during the two periods (six point one percent and seven point six percent, P=0.23).
- Lastly, the average frequency of interventions for hypoglycemia among adolescents was lower during the bionic pancreas period than during the control period (one episode every one point six days versus one episode every point eight days, P less than 0.001).
What may we conclude from these results?
The authors concluded that the bionic pancreas compared to an insulin pump improved average blood sugar levels with less frequent episodes of hypoglycemia between both adolescents and adults with type 1 diabetes.
Why are these results so impressive?
- Users of the bionic pancreas were in an outpatient setting and not in a clinical research unit.
- Meals and physical activity were not prescribed or regulated. The patients were supposed to “announce” meals to the bionic pancreas through the iPhone user interface, although those individuals that were over 18 were NOT reminded if they forgot.
- According to the authors, approximated estimates of meal size were provided, thus eliminating carbohydrate counting. It is of note that less than 30 percent of the insulin that was delivered in the bionic pancreas group was in response to meal announcements. This is directly opposite in usual insulin pump therapy by which omission of meal boluses lead to high blood sugars and more difficult regulation.
- Adults were permitted to moderate alcohol intake (although the authors note that higher intake of alcohol may be problematic for glucagon function).
Limitations of the study
- Alcohol intake (as noted above)
- Due to close supervision of study participants, treatment with carbohydrate for hypoglycemia may have been a bit more aggressive than in real life, unsupervised conditions
- Technical issues with wireless connectivity may have caused occasional missed doses of insulin or glucagon
- Risk of hypoglycemia if there was acetaminophen ingestion (may lead to “overestimation” of blood sugar levels)
- Long-term safety of microdose glucagon administration is yet to be established
FDA approval is still pending at the time of this post.