What Is an Artificial Pancreas, Really?

Bionic vs Artificial Pancreas

Anne L. Peters, MD: Hi. I'm Dr. Anne Peters at the American Diabetes Association (ADA) meeting in San Francisco, California. I am speaking with Dr. Howard Zisser, Medical Director of the Insulet Corporation in Bedford, Massachusetts.

Everyone is very excited about the "bionic pancreas."^[1] Can you explain what that is and how it differs from the artificial pancreas?

Howard C. Zisser, MD: Technically we are talking about nomenclature. Something that is "bionic" replaces something that already exists in the body but is malfunctioning or not working properly. If you want a bionic pancreas, you need alpha cells, beta cells, and the entire exocrine pancreas as well. An "artificial pancreas" is not really an artificial pancreas; it is probably not even a bionic pancreas. However, if people talk about an artificial beta cell or another system that has glucagon and insulin (an artificial beta cell and alpha cell), it gets confusing. An artificial pancreas is not any one thing.

Some people have the idea that they are getting an artificial pancreas, like a new kidney -- that they will just put it on and walk away into the sunset and everything will be fine. However, this is more of a staged approach. Some of the things that we will get out of this research will have nothing to do with fully automated control of blood glucose. But along that path, we are going to make some unique discoveries and help people control their blood sugar.

Don't Picture a Pancreas

Dr. Peters: I want to visualize what we are talking about. There is a New England Journal of Medicine article^[2] about the bionic pancreas. What does it look like?

Dr. Zisser: I can tell you what it doesn't look like. It doesn't look like your pancreas.

When making a system to control blood glucose, you need to be able to sense glucose. This is currently accomplished with a continuous glucose monitor, which is a small device that goes slightly under the skin. We take those data and run them on a "brain," using an algorithm. The "brain" can be a laptop, a cell phone, or an insulin pump. We decide whether more insulin, less insulin, or no insulin is needed every 5 minutes, and we do this every 5 minutes around the clock.

Currently, it would look like a sensor and an insulin pump, both of which are attached to the body, and then a "brain" (cell phone, laptop, tablet, etc.). It doesn't really matter. Eventually it should be back on the pump, so it is a system. It controls the output of your insulin pump.

Dr. Peter: You are using an existing pump and an existing sensor, or even 2 pumps in the case of using glucagon? The "brain" is presumably better than the human brain at adjusting doses, because it can see the trends and the rises and falls in blood sugar that a human being with diabetes might not perceive.

Overnight Success

Dr. Zisser: The advantage of this system, like any automated monitoring system, is that it is looking at the data all the time. Even the most compliant patients, the patients who control their diabetes well and make all the right decisions, when they are sleeping or exercising they can't possibly look at all of these data and make decisions in real time. Some aspects are really exciting, such as overnight control, which these systems can do pretty easily. There is no "disturbance" to the system. You are not waking up in the middle of the night to eat, although in using these systems I diagnosed a patient with a sleep eating disorder.

Dr. Peters: Do you mean they didn't know they were eating at night? They didn't remember?

Dr. Zisser: The patient was on Ambien (zolpidem), was eating in the middle of the night and not remembering, and the blood sugars were going up to 350 mg/dL and coming back down. We changed medications and this patient stopped eating in the middle of the night.

Usually in the middle of the night, nothing is happening. Patients are not watching their blood sugars. If they have hypoglycemia unawareness, they don't respond to hypoglycemia. In most of the controllers, the "brain" portions are different; they were designed by different control engineers. They don't work perfectly, but they work pretty well overnight if you can trust the sensor.

This has changed over the past year or two. The sensors are becoming more accurate and more precise. You can believe the sensor, and there is less variability in the manufacturing of sensors.

Dr. Peters: During your research, what did patients say when they began to use the device and it took away some of the need to fuss with their blood sugars?

Dr. Zisser: That is as variable as the number of patients there are. Most patients are more than happy to hand over their blood sugar control to a system, if they can trust it. I spoke about this at TEDx: "Should We Trust Our Lives to Technology?" We can trust technology, but early adopters of those technologies need to realize that their risks will be higher because there are situations that haven't happened yet. Physicians and engineers try to design devices to be as safe as possible. However, when you go from 10 patients to 100 patients, or to 1000 patients, you start to see new situations.

Taking a Break From Diabetes

Dr. Zisser: Usually patients are very happy. My patients call it their "diabetes spa day." They come in and work on scrapbooks, sew, quilt, or read a book. They get a break from having to control their diabetes all the time. Some of the patients who are very meticulous about their diabetes management have a hard time initially handing over control. For our inpatient trials, we have a big screen so they can see what is happening with their insulin, glucose, and meals. We sit there and watch it and they feel very comfortable.

Dr. Peters: In the study that was published recently^[2] about the individuals -- the kids in camp and the adults who had 5 days of "free living" -- they were still monitored very closely, weren't they?

Do you perceive this as a step closer to making these devices real for a broader population? How do you see this developing?

Dr. Zisser: The process, as I know from previous work, is that the first few years are spent taking a widget that has already been approved and improving on it -- making it smaller, trying to make it redundant, or making it a little safer. We are trying to figure out whether it can be done.

The next step is to determine whether we can do it all the time, for everyone. That is a big hurdle. We have to figure out all the ways the system can fail. We are getting there. The steps of remote monitoring for patients came out of this research. For example, when someone with diabetes says, "I will be traveling and will send you my glucose data; if I don't respond, call the hotel" -- all of this came out of taking information that was in the patient's pocket and using it in real time. The data can go anywhere. All kinds of companies are sprouting up to shoot these data to the cloud. We have to figure out the best way to deal with it.

Still Recruiting

Dr. Peters: In terms of development, where are we now? What is the next step? Can my patients in Los Angeles join the study?

Dr. Zisser: Yes. There are multiple research sites in the United States, mainly on the East and West Coasts -- not a lot in the middle, although the Mayo Clinic has recently joined some of the research. There are multiple sites in Europe and a couple of sites in Australia as well. Patients can search at clinicaltrials.gov to find out where those trials are taking place. We even have patients coming from out of town to be in these studies. Every time we recruit a new patient and do another trial, we learn something new. That is what it is all about.

The push in the past year and a half to 2 years has been to take these very controlled inpatient trials (where we use intravenous fluids, glucagon, 50% dextrose, etc) and move out of that very safe environment to the real world.

Dr. Peters: My patients certainly want this now. They want to put it on and wear it and not have to worry about all the details of managing type 1 diabetes.

Dr. Zisser: It is going to take some time, because every patient is different. There is a common denominator of not making enough insulin, or no insulin, but responses to meals, exercise, stress, and illness are all unique to the individual and can change over time. We need to take the current state of the art -- fingerstick glucoses and subcutaneous insulin -- and compress the signal and make it a little bit tighter. Eventually we would like it to be even tighter, but for that we will probably need faster-acting insulins.

Part of what comes out of this research is taking it as far as we can and finding out what the rate-limiting step is. It might be insulin, so how can we make insulin act faster?

Dr. Peters: It is wonderful and overwhelming all at the same time. I love the steps, and I think we are moving closer. It clearly is going to take us a while to get to that holy grail of being able to make a truly bionic pancreas.

Dr. Zisser: I am probably on the conservative end when it comes to giving an opinion. We have insulin right now, which is a fantastic drug for people with type 1 diabetes. They didn't have that in 1921, but they had it in 1922. The insulins have gotten a lot better, but that was a long time ago. Let's see whether we can compress that signal and make lives better for patients with type 1 diabetes.

Dr. Peters: Thank you very much. This has been Dr. Anne Peters for Medscape.

Comments

Commenting is limited to medical professionals. To comment please Log-in.

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.