jeramie irwinJeramie has just completed a JDRF-funded artificial pancreas (AP) trial over 12 weeks at the University of Virginia, one of the very first to be completed in an out-of-hospital setting. He’s also a JDRF staff member from the USA’s Kentucky & Southern Indiana Chapter. Jeramie blogged on his experience with the AP he nicknamed ‘Artie’ during the trial, and the result, ‘Me & My Broken Pancreas‘, has been read by more than 11,000 people globally. Jeramie has kindly allowed us to republish his work for our JDRF Australia readers.

Here is the final chapter in the trial series, where Jeramie provides a ‘day in the life of’ and summary of the AP trial.

So today I am saying goodbye to Artie, my artificial pancreas. I will have to say I am a little sad. I am going to miss the nights of restful sleep, waking up feeling rested, and having more energy. Over the past 12 weeks, Artie and I have been through a lot. Some days Artie was my best friend, and other days I wanted to run him over with my car. Overall though, it was a very positive experience. I am not looking forward to going back to my traditional insulin pump and CGM. Though I would love to be able to keep the system, I understand it is more important for them to use it with someone else to continue to further research. This trial has given me much hope toward a bright future for all of us with type 1 diabetes (T1D).

As promised, I will be sharing my experiences on the artificial pancreas during a typical day and while exercising. I will also give an overview of the trial and share my thoughts on the DiA’s In-Control artificial pancreas system both good and bad.

Artificial Pancreas trial

Daytime and mealtime with Artie

The Diabetes Assistant (DiAs) In-Control artificial pancreas system has to face many challenges to account for all the variables of daily life like diet and exercise. With these challenges, the DiAs acts less aggressive during daytime than overnight. With a target of 8.9mmol/L, it seeks to avoid hypoglycaemia (below 4mmol/L). The DiAs does not treat hyperglycaemia until I rise above 10mmol/L. If DiAs predicts I am heading out of range (4-10mmol/L) it will give a correction or reduce/suspend insulin. Since it is less aggressive, my numbers did run higher than I would normally prefer. With Artie the artificial pancreas I tended to stay in the 8-10mmol/L range for much of the day.

This does lead to more stable numbers and less of a roller coaster throughout the day which is the ultimate goal.

Having the large swings and variability in blood sugar can become exhausting or can eventually do a lot of harm to your body. So the DiAs artificial pancreas is designed to give you much better control and ultimately reduce long term complications.

During meal time, it is still required to enter carbs and blood glucose into a bolus calculator just like a regular insulin pump. When starting on the DiAs artificial pancreas I was surprised to learn I still had to manually count carbohydrates. I had always envisioned that the artificial pancreas would do all this work for me. The first artificial pancreas system when it comes to market will be “hybrid” closed-loop instead of fully-automated closed-loop. It will still require some input from the user. I was told the reason for this is because none of the current fast-acting insulins on the market work fast enough. On traditional pump therapy, I normally have to bolus up to twenty minutes before I eat to keep my numbers below 10mmol/L. If I had not used a bolus calculator to tell Artie that we are eating and how much we ate than we would get a tremendous spike before the artificial pancreas could predict a high, make a treatment, and then the insulin could start working properly. So it’s not necessarily the fault of the system itself but a limitation of our current rapid acting insulins.

Once we have a faster acting insulin it will be game changer with the artificial pancreas system.

I have found that just like life with no Artie, Artie also has trouble controlling blood sugars when eating diabetes-unfriendly food. The more carbs and fatty food I eat, the worse my blood sugars are going to be. What I have found though, is that my highs are less high and tend to also come down much faster. I also have a much softer landing. My blood glucose under Artie rarely ever went above 14mmol/L and not below 3mmol/L. The only time it did go above 10 was because of a site issue. I have greatly reduced my amount of hypoglycaemia during the day. I have reduced my lows from greater than 6% of my Dexcom readings, to about 2%.

Exercising with Artie
Artie has an exercise button for physical activity that raises your target blood glucose. The artificial pancreas become less aggressive as I exercise. It reduces insulin and only gives a correction at over 10mmol/L.

While using this feature I never had a hypoglycaemic episode.

The problem is because the artificial pancreas does not deliver glucagon, I would still have to supplement with some kind of glucose during exercise. While doing intense cardio like cycling and running I found Gatorade worked very well. During exercise, I monitor my blood glucose with my CGM during my workout and drink Gatorade for glucose as needed to keep me my blood sugar stable. Many people feel to have a true artificial pancreas system it must deliver both insulin and glucagon. Many others disagree and feel insulin only is the best route. I can see the advantages and drawbacks to both systems but I cannot really accurately comment on an artificial pancreas system with glucagon since I have never tested one. Using the exercise button while supplementing glucose during my workout worked very well. Over my 12 week trial, I had at least 40 workouts, and during exercise never once experienced a low blood glucose. The exercise option works very well on the DiAs system, though not perfect.

Biggest takeaways:

night time on artificial pancreas

  1. Night time control is amazing. I wish the artificial pancreas could work as well the 16 hours I am awake as during the 8 hours I sleep at night. I would choose Artie over my traditional insulin pump and CGM if nothing else but for the improved night time control. I almost never went low while asleep and almost always woke up between 5-6.7mmol/L. I definitely felt much more rested and felt like I had more energy while on the artificial pancreas. See my previous blog post No Rest for Artie, for more on how the AP keeps us safe while sleeping.
  2. The AP reduces lows and decreases glucose variability. The first generation DiAs artificial pancreas won’t eliminate your low blood glucose but definitely will reduce lows by predicting lows and reducing or turning off basal insulin. Since DiAs doesn’t use glucagon, it is almost impossible for it to completely eliminate lows but the system does a good job of keeping your blood glucose in a safe range. For example, if you are driving your car and come upon stopped traffic your instinct is to hit your brakes and stop your car. If you are going fast you might not be able to completely stop your car and still might have an accident. Without the brakes though, your accident would have more than likely been a lot worse. The DiAs system controls the brakes of your artificial pancreas.

    It may not be able to stop you from going low if your glucose is dropping too rapidly, but it will definitely give you a much softer and safer landing.

  3. The DiAs system seems to prefer keeping my blood glucose between 8.3-10mmol/L during the day time. I am much more aggressive with my diabetes care than DiAs. If my blood glucose is above 8mmol/L, I am giving myself a correction. DiAs will not provide a correction until I am predicted to go above 10. For this reason alone, my average blood glucose and A1C actually increased during the study. My A1c increased from 6.8 to 7.2 and my average blood glucose went from 7.3 to 7.9 mmol/L. This was my biggest complaint about the system. I would like to see it a little more aggressive in correcting my blood glucose.
  4. Bluetooth connectivity can be a challenge. The DiAs system relies on Bluetooth connectivity for the communication from the Dexcom CGM to the Smartphone and from the Smartphone to the insulin pump. The Bluetooth was unreliable and required a lot of trouble shooting, repairing, and rebooting. Many times these devices would come unpaired and I would have to stop what I was doing and reconnect the devices. If they became disconnected and I did not immediately tend to the devices I would receive a very loud and disturbing alarm, much like a fire alarm, which would keep alerting until I tended to the issue. Also, these alarms could not be deactivated or turned to vibrate or silent. This was sometimes a challenge if I was at work, in a meeting, or busy somewhere where I couldn’t respond to the alarm right away. It became very obnoxious. I especially experienced a lot of communication issues at places with large groups of people like at airports and at stadiums. Sometimes I would have to change back to “open-loop” mode until I get to a place without the communication interference.artificial pancreas trial
  5. Further integration is a must. Having T1D is not convenient in any way, shape, or form. We have lots of stuff we have to carry around and keep up with to manage our diabetes. The last thing we need is a system that is not user-friendly or integrated and requires us to carry around several devices. In this study, I had to carry the Dexcom receiver, an Android phone (not my smartphone), a blood glucose meter, and my insulin pump. There were too many moving parts to keep up with. It became a hassle trying to keep up with all the parts to the DiA’s system and always having to have them on me. I could not leave the smartphone or the CGM EVER. If I did walk away from the system more than a few feet I would lose my Bluetooth connection. I had to keep the CGM receiver, pump, and smartphone on me 24/7. Definitely not convenient, or user friendly.

Final thoughts

As much as I found to critique about the DiAs AP system, I would trade my Omnipod and Dexcom G5 back in to wear it again today. As much as the system is imperfect, it is still better than anything on the market today. The first artificial pancreas systems will be very good, but I guarantee they will just keep getting better and better. Look at today’s insulin pumps compared to what we had 10-15 years ago. The first generation insulin pumps that came to market were massive and had little functionality. Also, we have to realise that the DiAs system I used was a trial version and not the final commercialised product. I am confident the commercialised product will be much more user-friendly and much of the connectivity issues will be corrected. The purpose of the study is to test the safety of the system, not to test the user ability and convenience.

Overall it was an amazing experience and I feel very blessed to have been part of this trial. I am currently waiting on word to hopefully start a new artificial pancreas trial with the University of Virginia in the coming months. I very much look forward to reconnecting with Artie, and testing the newest version of the DiAs system.
You can see more about Artie on Jeramie’s blog,  Me & My Broken Pancreas





Leave a Reply