If you have type 1 diabetes, you likely have heard about the newest management rage: hybrid closed loop systems, or HCL for short. Developmentally speaking, they’ve come a long way in the past decade or so. Basically, they consist of an insulin pump, an integrated continuous glucose monitor (CGM), and an algorithm on an app or associated device that allows the pump to receive data from the CGM and make (some) insulin adjustments for you. In theory, it would be ideal to have an “artificial pancreas” that keeps your blood glucose in a more normal range than what you can do on your own when you’re making all the insulin dosing and timing decisions. In reality, using an HCL system is still more complicated than we would all like.
For starters, you have to have an insulin pump that is compatible with the CGM you use or want to use. Currently, only the Tandem Control IQ insulin pump gives you two CGM choices to integrate with (the Dexcom G7 and Libre 2 Plus, and possibly soon the LIbre 3). A smaller Tandem pump, the Mobi, works with the Dexcom G6 and, pending FDA approval, the G7 version. Medtronic has an updated system, the 780G, integrated with their proprietary CGM, the Guardian 4 (but in Europe, the Simplera Sync with a lower profile is already available as the integrated CGM). The Omnipod system can only pair with the older Dexcom G6 sensors, although they are in the process of getting FDA approval to integrate with G7 sensors. There is a fourth system, the iLET Bionic Pancreas, that only works with the Dexcom G7. While Eversense makes an implantable CGM that works for 6 months (and soon for a year), it doesn’t yet integrate with any insulin pumps.
Confused yet? As a diabetes professional, I know a lot about these systems, and it’s even hard for me to keep up with the latest pump versions and CGM integrations. Also—and probably as important as which system you may choose to use—none of these HCL systems is infallible for multiple reasons. First, let’s start with how smart you have to be to use the systems. For instance, you still have to announce meals and/or choose mealtime and correction boluses (with the help of auto corrections for insulin remaining in your body for some HCL systems). Your pump can respond with additional insulin, but each system handles these a little differently, so you’ll have to learn how yours works, how to give boluses, and how much to give. Some people eat low carb and just increase basal rates aggressively to cover meals, so that’s an alternate possibility.
Second, the biggest issue people have with exercising with type 1 diabetes is having too much “insulin on board,” which can cause lows with activity, and HCL systems imperfectly handle most physical activities because insulin levels typically remain too high if you don’t ramp down insulin delivery an hour or two in advance. Some systems allow you to temporarily override your basal rates or set a different (higher) blood glucose target to try to compensate, but exercise has been the bane of these systems since their inception. Most of the time, you still have to plan ahead whenever possible and make insulin dosing changes yourself.
Third, the systems depend on how well the technologies work, and both pumps and CGMs have their issues. On the glucose monitoring side, most CGM sensors have to be replaced every 7 to 14 days. However, I often have my Dexcom G7 sensors fail miserably early (such as on day 6 out of 10) and be wildly inaccurate both during the first 24 hours and for a day or two prior to the sensor failing. Point in case, just today I put on a new G7 sensor and it has been telling me I’m critically low (under 45 mg/dL) when I have no symptoms and my blood glucose meter reading is 131 mg/dL. You’re limited to two calibrations per sensor session, and I hate to use one in the first 24 hours, but getting low alerts all day (or all night) long with normal blood glucose levels is just not accurate enough for me! (Not everyone has as much trouble as I seem to have, though, plus I have really high standards when it comes to accuracy.) Insurance companies also pay for a limited number of CGMs per month, and having one fail early sets you up to run out of sensors. I call Dexcom to get free replacements when they fail before Day 9 but, apparently, there’s a limit you reach for getting them and I’m already there since only 10% of my sensors last the full 10 days. Then are the compression lows that can stem from lying on the sensor during sleep, etc., along with potential insertion site issues, transmitter range problems, trouble staying on, or skin reactions to their adhesives. So, they’re imperfect—and need to be as accurate as possible when your HCL system is relying on them to make insulin dosing decisions for you.
Fourth, the insulin pumps themselves can have tubing, insertion site, and delivery issues. I wore pumps for 4 years a while ago and had so many problems with scar tissue and infusion site failures that my A1C started to suffer. In my opinion, the biggest drawback of using insulin pumps is that if they fail to deliver the insulin for any reason, it doesn’t take long for DKA to result from severe insulin deficiency. (At least with basal injected insulins, you always have some background insulin still working for you.) I have known some people to take injections of basal insulin and only use pumps to deliver mealtime insulin for that reason, but that would be hard to manage with HCL systems. In addition, only the Omnipod pumps are both tubeless and waterproof, and I swim regularly so that doesn’t leave me a lot of options if I don’t want to take off a pump and leave it off during activities.
Finally, there are issues with the insulin delivery controllers and algorithms themselves that these HCL systems rely on. Most pumps can be smartphone controlled, but others (like the Omnipod 5) still have a separate device that you have to use to deliver insulin if you have an iPhone and not an android one. (FDA approval is pending for their use with iPhones.) The algorithms are also imperfect (but improving over time). Some use AI and try to learn how much insulin you usually require and when and may get better over time.
All this is not to say that you can’t benefit from using an HCL system to better manage your blood glucose readings. The main issue I have with them is that my blood glucose levels are already in a fairly tight range, and I’m not sure that they will be improved by giving control over to an HCL system run by algorithms, given the limitations of the devices involved.
Here’s another true story. I finally decided to try an HCL system using the newest Omnipod 5 tubeless pump because I could get a free 30-day trial and not have to commit to using it. Before starting it, though, I found out that the Omnipod 5 is not yet integrated with my Dexcom G7 CGM, and Omnipod doesn’t know when they will be getting FDA approval for it. Plus, I have an iPhone so I won’t even be able to give boluses from my pod using my phone. Many people I know started using their own HCL system a while ago using a “looping app,” which is basically a DIY HCL system. So, I thought maybe I could download that on my iPhone and still be able to use my G7 sensors with Omnipod 5, only to find out that “looping” doesn’t work with the newer Omnipod 5. So, now I need to contact my doctor and get a new prescription for the Omnipod Dash if I want to try looping on my own. At this point—especially given my problems with pumps in the past and my current issues with my G7 sensors—I’m not sure it’s worth the time or the money to try to cobble together my own HCL system. Maybe I’ll just stay on multiple daily injection (MDI) therapy a while longer until more of the kinks in these systems and the devices that they rely on get worked out.
What you decide to do for yourself is really an individual choice, though, and that is probably my whole point. HCL systems are an exciting advance in glucose management and lifesavers for many people. What doesn’t work for me may work for you just fine, and vice versa. Don’t be afraid to give HCL a try. I know a lot of people who just love being on these systems.