Your donations help T1Determined #keepgoing.
Exercise and the Medtronic 670g (with Auto Mode): What I’ve Learned So Far
UPDATE: In November 2018, I switched from my almost-new Medtronic 670g to the Tandem t:slim X2 and Dexcom package, at my own expense.
* * * * * * * * *
My original Medtronic 670G review:
Back in March 2018, after the announcement that Animas would be shutting its doors, I had to make a decision about when to migrate from using my trusty Animas Vibe, which when I got it in April 2016 was the only IPx8 waterproof insulin pump on the market, to the Medtronic 670g, which is now the one of the only two REMAINING Ipx8 waterproof insulin pumps on the market.
I decided to make the cutover near the beginning of March, giving myself 3 months to get used to the new system before doing a solo 339-mile run across Iowa, rather than placing another order for supplies for the Animas, which would have forced me to cut over right before the Big Run.
In the end, it’s a mixed bag.Here, I’ll try to give an honest review of the features and capabilities I value as an athlete, along with the foibles and downright letdowns that annoy me to no end.
- Auto Mode: There is absolutely nothing to compare to the pump’s noticing highs and lows and shaving the sharp edges off them. This may sound like a minor thing, until it’s the day after a 50-mile training run, and you’re glycogen depleted and super insulin sensitive. Just watching the pump / sensor combo catch the low and back off the basal insulin rate is a wonder to behold. I’ve had virtually NO dangerous lows since getting the new pump.
- Less variation in BG over the course of a week, month, etc. I’m not sure yet what my A1C will be for the first month after using the new pump, but I can tell you that the standard deviation AND number of BG variances has dropped strikingly. In that sense, and with the addition of auto mode, type 1 starts to feel more like type 2, in the sense that your “pancreas” sort of halfway functions, but is slower than normal to adjust to carb intake.
- The sensors, while they appear fairly inaccurate during the “learning week” at the start, appear to be fairly accurate and last about 6-8 days. That’s not the 10 I was promised, but it’s good enough.
- The dressings. I don’t know what they’re made out of, but they stick really well, especially with Skin-Tac, and mostly don’t come off. This is important if you’re swimming or sweating for 17-20 hours at a time.
- The transmitter for the CGM is rechargeable using a special charging station and a AAA battery. I appear to get a longer, better charge out of lithium AAAs than the recommended / included Energizer. That said, the first transmitter I was shipped was faulty and wouldn’t hold a charge for longer than 3 days. Which raises the question: how is it that the Dexcom G4’s battery was good for up to 9 months?
- I love the way the system gracefully slips from auto mode to “safe basal” mode (sort of “what it remembers about recent auto mode settings”), which is keeps up for 90 minutes until you put it back into auto-mode OR ignore things and let it slip into manual mode. Even manual mode has high and low alerts and basal auto-suspend before / on low. Pretty darn automatic for manual mode!
- It’s incredibly fussy about calibrations. Medtronic tells you that it takes 3 sensor readings and looks at the two that agree the most closely, but I can tell you that unlike the Dexcom CGM (continuous glucose monitor), you can’t just throw a number from your glucose meter at the sensor calibration screen and have it correct its guess. Instead, you get what’s called the “cycle of death” or something like that, in which the pump repeatedly tells you that the BG (blood glucose) reading from your meter is “unacceptable” (as opposed to what? LYING to the CGM just to get it to shut up?). This actually UNDERMINES the confidence of a typical type 1 diabetic, since a slightly-inaccurate-but-in-the-ballpark number is better than “—” or “???”. It forces you to test very frequently, using up glucose test strips, until you’re sure your glucose is level enough to allow a calibration. If you’re out on a 40 mile run, the last thing you want is to be running low on strips AND not have any idea of what your CGM thinks your sugar is. The best strategy I’ve found so far is to ignore most of the calibration warnings until it’s just the perfect time to punch in a new number. It’s a little more forgiving with “Enter BG” prompts, which, strangely, AREN’T the same as calibration requests.
- The difference between “Calibrate Now” and “Enter BG” is unintelligible. I’ve gone through training AND read all the documentation, and I STILL don’t know the difference, EXCEPT that you don’t EVER want to calibrate until you have to, and you can “Enter BG” fairly frequently without paying a penalty.
- It takes FOREVER for micro-boluses to correct a high. Hint: if you’ve badly underestimated carbs, go ahead and enter them as soon as you can and let the pump dose for you. It will usually be much better than trying to chase a high with a correction bolus, and it feels better.
- In auto mode, the system “learns” your basal rate — which is fine if you eat and do the same thing every day. When your training schedule is 40 miles today, sauna tomorrow, 30 miles the next day, the system learns to set the basal high, then when you’re out on your run, it has to back off that rate when you’re exercising; OR, it learns to set it low, and on rest days or non-cardio days, you run around 200 mg/dl most of the day. That’s not great.
- The form factor for the Guardian CGM transmitter. It’s STILL not very body-conformant and it’s much too easy to accidentally rip off a sensor site and potentially (for instance, during a swim) lose $1200 worth of transmitter.
- The sensors are a pain to put in. They really only work well with this bulky, orthopedic sen-serter (sensor inserter), and the automatic retraction mechanism for the introducer needle RELIABLY fails if you’ve used Skin-Tac on the site before injecting the sensor. So you’re standing there with a needle that won’t come out of your belly, jimmying with the release, pulling on it, trying to get the needle out without it going back in to the same site at an angle and destroying the sensor cannula. Annoying, and has a high ick factor that’s on par with cleaning up vomit after tending to a sick pet or family member.
- It’s not really possible to get a clear answer from Medtronic about just how waterproof the sensor / transmitter combo are. They say it’s “waterproof”, then follow that up with “you can swim for up to 30 minutes at X depth.” OK, I understand it’s not dive proof, but in 30 minutes, I can swim maybe… 1000 meters, 3/4 of a mile? What am I supposed to do on 3 mile swims? Pray it holds up? I literally can’t get a straight answer FROM THE COMPANY. From type 1 diabetic athlete friends, I’m told it lasts pretty much the whole day in the water for a lake swim. Let’s hope so. That’s my sense as well, but then… why the cautions about only 30 minutes? If the pump is IPx8 waterproof but you can only have the transmitter in the water for 30 minutes, then that’s not exactly useful.
I hope this is useful to those of you who are #diabadasses and looking at this pump. For what it’s worth, I looked at the Tandem T-slim and basically, it’s “dunk proof” but not waterproof. I also looked at the Omnipod and saw too many complaints about delivery failures, occlusions, and other issues; plus my one-on-one experience with reps at trade shows felt a little… well, uncomfortable. I know there are a lot of podders out there, and I respect that it’s a great and widely used pump, but even with my complaints about Medtronic, they’ve always been a reputable company with a solid support infrastructure, and while their technology sometimes feels like an awkward teenager that’s late to the prom, it is generally VERY reliable and the benefits of working with the people, specifically the local sales support and training reams, are often understated and under-reported.