As many of you know, I’m currently studying to become a Certified Diabetes Educator. So I was delighted to have the opportunity to attend the American Diabetes Association’s annual Advanced Postgraduate Course in New York City on Feb. 23 — a two-and-a-half-day conference that can be viewed as the ‘little sister’ to the ADA’s huge Scientific Sessions held each summer.
This Postgraduate event hosts just a few hundred doctors, pharmacists, diabetes educators, dieticians and other healthcare professionals, compared to the several thousand attendees at the Scientific Sessions. But it was still action-packed with D-knowledge — focusing on broad issues in technology, treatments for type 1 and type 2 diabetes (such as exercise, insulin pumps and type 2 medications), rather than on the heavy science and clinical studies.
I attended sessions on the new national AADE standards announced recently, plus a presentation on insulin pumps and CGMs that I found disappointing since it was actually very basic information, geared more toward health care professionals unfamiliar with the technology. Another session took a look at how to better utilize type 2 oral and injectible meds, and there was a lively conversation there about how to convince an unwilling type 2 patient to embrace insulin if it’s deemed necessary. My takeaway from that session was that health care professionals might need to be creative with combinations of oral medications, but also do a better job of explaining the benefits of insulin to eliminate the stigma associated with it.
Unlike bigger conferences throughout the year, this one does not include any kind of expo, so attendees were not exposed to real-life demonstrations of any of the new technology discussed. This seemed a little strange to me, but perhaps it is good sometimes to have health care professionals focus on patient data and research rather than sales pitches (!).
A highlight for me was getting the chance to interview legendary endocrinologist Dr. William Tamborlane, who is the Chief of Pediatric Endocrinology at Yale University and the Deputy Director for the Yale Center for Clinical Investigation. Since Yale is a top center for leading diabetes research, Dr. Tamborlane has pretty much seen it all! He and his staff are currently heavily involved in the in-patient clinical trials of the Artificial Pancreas, specifically studying using a closed-loop overnight in pediatric patients.
We chatted about his work on the AP, his thoughts on the FDA’s hesitancy (but hopefully not for much longer!) to approve the Medtronic Minimed Veo with the low-glucose suspend feature, and the impressive results of new Insuline patch studies that involve using heated infusion sets to improve insulin absorption. Here’s what he had to say:
DM) At this point in your research on the closed-loop system, what are some of the obstacles? We know you’ve had a lot of success showing efficacy of the two-hour shut-off protocol (in case of a low), but what are some of the things we still have to overcome?
WT) I think the major overriding issue is still safety. Once you have a closed-loop system that automatically turns on and off, then you have important safety issues. The most serious one is giving too much insulin if for some reason the system wasn’t working right. There are a number of reasons that could be; chiefly it would be sensor errors. We do have sensors that over-read (exaggerate results) a lot. We have issues in loss of communication between the sensor and the pump receiver with the algorithms.
For the Medtronic system, there’s a receiver in the pump and the transmitter from the sensor. So the laptop computer that we’re using has the receiver and algorithm in it, then it tells the pump how much insulin to give. So it goes from the sensor to the laptop, laptop to the pump. Ultimately the algorithm would be in the pump. What was happening was the sensor was sending signals to the laptop that the blood sugar was going up, and the laptop would send a message to the pump to bolus. When the pump got this message, it would send a message to the laptop saying “Message received. Insulin given.” When we switched to a newer system, there was a delay in the pump telling the laptop that it had received messages. So if the laptop didn’t hear anything, it would send another message, which could spell trouble.
So there a stacking effect because the laptop was sending out another message to the pump, telling it to bolus again?
Right. The laptop didn’t get the right message at the right time, so it didn’t think the pump bolused. There are any number of issues like that. We’re still feeling our way, because there are real practical issues on how a patient could use this. We’re asking: How complicated would this be? How much would the patient need to know?
Also, we need to find out if there are ways that we could have the same kind of remote telemetry that we use for pacemakers in the Artificial Pancreas. This way we can perform diagnostic tests to make sure the systems work properly. Then, of course, we need better, more accurate sensors. There’s still a lot of work to do.
There’s a lot of talk about using other hormones in a closed-loop system… like Symlin or a stable form of glucagon. What are your thoughts on incorporating these secondary hormones and medications?
We’ve actually done two pramlintide (Symlin) studies. In the first one, patients came in for a half day to get the system going. They came in the afternoon the day before to put in the sensor, calibrate, and put in the infusion site. Then we turn on the closed-loop overnight so the next morning, they can start right away. The next day, they took pramlintide before meals. Because they hadn’t been exposed before, we had to use a low starting dose, because you get some gastrointestinal side effects if you go too high. We didn’t want lower blood sugars because they couldn’t eat their meals at the risk of getting very nauseous. Plus, treating a low on Symlin can take up to an hour… it’s very dangerous! With how we did this, results were better and we had lower peaks even though there was considerably less insulin given.
In our second study, with more therapeutic levels, patients came in for one and a half days before starting pramlintide to do their control experiment. Then they went home and we titrated up with pramlintide as outpatients and then brought them back and did a closed-loop control on pramlintide. The results were a little better, but I still think we need to work on the algorithm. If there are delays in communication, we need to change the parameters so that it gives a better response. What we are going to try next is liragulatide (Victoza) and that’s attractive because it’s only once a day. With liragultide, we will look at reducing glucagon responses at meals, so that combined with the gastric emptying will reduce the post-meal peaks.
Are there any near-future milestones we can look forward to?
Something that will have an immediate effect is getting the Veo pump approved in the U.S., and eventually getting a system that uses an algorithm running in the background and only activates if you’re actually having a projected low. You could get the pump to shut off for a couple hours without the patient even knowing it, and have it only alarm if the patient actually got low.
Say you want the patient alerted at 60 mg/dL, but the shut off is at 90. If you’re at 180 mg/dL and dropping fast, and in X number of minutes you’d get to under 90, it would shut your pump off. You might keep coming down but you might not actually get low and never have an alarm. That would be a big step.
(Editor’s note: Medtronic is already working on this predictive Low Glucose Suspend feature, which may be in the next version that’s currently awaiting FDA approval.)
What are your thoughts on why the FDA has been delaying the Veo?
Here’s where the regulatory authorities are a little misguided. They were worried that if you were shutting off the pump, it would make your A1C go up. Turns out in real-life use studies in Europe and Australia, it’s only during the night or the day when you’re sleeping that you would have a full 2-hour suspend. Secondly, the added mean glucose of making it flatten out at 50 mg/dL instead of going down to the 40s is misguided. That’s not going not raise your A1C.
Plus, it’s not like you’re going to be low at night every day for three months…
That is true. But also from the JDRF CGM trial, which I was co-chair of, A1C lowering was primarily related to how often you use the sensor. If you use it every day, you drop your A1C .5-.8% if you’re a teenager. The drop-off was pretty dramatic. If you only use it for four or five days, you didn’t get much benefit at all.
Say you went to Europe and got your Veo pump. All of a sudden you have a greater motivation to wear the sensor every day, and if you wear the sensor more, you’d have the sensor values during the day. The Veo pump is more likely to lower A1C than raise A1C because it would give people an added incentive to wear the sensor every day.
What have been your findings so far in your research with the Insuline heated infusion set?
We’ve done these glucose clamp studies — a study of the time-action curve for insulin and insulin sensitivity in humans — in insulin pump patients. They come in the night before, put in an Insupatch infusion set, and they are usually studied on two different days, one where we turn on the heating element, and one without it. We check throughout the night so the patient is around 100 the next morning. Without warming, the peak action of insulin is about two hours. We found with the warming you have peak action within the first 90 minutes.
It also turns out that — what it looks like to us — is that you actually increase the amount of insulin that gets out of the subcutaneous site, otherwise known as the bioavailability. We actually absorb more insulin when the heat is present.
Wow! So does that essentially ‘trick’ the body into thinking more insulin is being infused, when it’s just warming up the infusion site?
Well, when insulin is sitting in a little pool under the skin, there are enzymes that break it down. So what if we gave 1.15 instead of 1.2 units? You get the same peak earlier and it dissipates faster. You also find out in practice that you don’t need as much insulin as you thought you did before.
What’s your next big research focus?
We are still facing that challenge of insulin not responding quickly enough to sensor data in the closed-loop system. So, we want to do clamp studies with new “ultra-fast” insulin action, like that being developed by Halozyme, as a possible remedy. That’s what we’re working on next. It’s the same goal — to speed up action — but just a different way to get the insulin absorbed faster.
Thanks so much to Dr. Tamborlane for sharing this insight and study results. Let’s keep our fingers and toes crossed that the FDA helps keep the pace moving on this research, and approves the Veo real soon!