Starting Eversense 365 with Loop DIY

[Terry]

When Dexcom announced earlier this year that it would discontinue the G6 CGM, it prompted me to examine my tech situation. I had been using Dexcom CGM products daily since 2009 beginning with the Seven Plus CGM. I then progressed to the G4, G5, and G6 products. I had hung back using the G4 as I didn’t think the G5 was performing as well as the G4. When Dexcom stopped making the G4, I was forced onto the G5 briefly and then moved on to the G6.

While I was overall pleased with the Dexcom CGM products, I began to find fault with the accuracy and consistency. I integrated the CGM into Loop DIY in 2016 and that focussed my attention on just how important CGM accuracy and consistency was not only to metabolic measures but also to quality of life.

Just what do I mean by “quality of life”? My classic pet peeve and the single circumstance that totally alienated me with the Dexcom products was this scenario. My Dexcom low alarm would wake me up, often around 3 am. Many times I just treated with a glucose tab and try to quickly slip back into sleep. I soon began to suspect that these low alarms were not accurate as I forced myself to finger-stick and be certain the event justified a glucose tab.

What I soon began to conclude is that many sensors, not all, would report CGM BGs in the low 60s while the actual BG was in the mid 70s. This irritated me to no end. A mid-70s BG during the 3-7 am was not dangerous, it was ideal for that time period. This false low also had the effect of turning off my basal insulin to counteract the low. This in turn exacerbated my subsequent BG rise due to the circadian release of cortisol at that time of day.

This didn’t just happen once in a while but with as many as half the sensors I used, especially on day 1 and days 8-10. I put up with it because when the Dex behaved, my BG experience was excellent. With Dexcom’s end of G6 announcement, I decided to consider other options. I heard that Eversense was compatible with Loop DIY, although it required loading a development branch, something that many users had already done and found workable.

I decided in late June that I was going to try the ES 365 and scheduled an implant of July 15. It’s only a small cut, about 1/4 inch, done in doctor’s office and no stitches were necessary. My wound healed well and when the steri-strips came off 9 days post-procedure, I could hardly see where the cut was. The sensor was implanted in my right arm about halfway between my shoulder and elbow and rotated about halfway between my side and front. I chose this spot to accommodate side sleeping and to give me good visibility and access of my left hand. Due to a stroke, my right hand has impaired dexterity.

What I didn’t know and soon learned is that the app that Eversense publishes and the Loop app that lives on my iPhone cannot both communicate with the ES transmitter. Loop DIY has incorporated all the necessary functions that the ES app does. I’m aware that the Twiist loop system is set up to allow commination with both the pump and the ES app.

ES produces vibratory alerts for highs, lows, and lost connections. I usually take off the ES transmitter and charge it when I shower. The Loop app provides a “placement guide” to reposition the transmitter after the recharge session. The placement guide provides a bar graph visual depicting a 0-100, poor to excellent, feedback.

One day I accepted a less than excellent placement guide feedback figuring the positioning did not have to be perfect. The next day, before my usual wake-up time, I was awakened by a vibratory alert signaling a lost connection. The alert instructed me to call customer service.

Customer service was helpful and professional but my call was longer than need due to my DIY Loop circumstance. What I learned is that I should only accept an excellent 100 from the placement guide and when I do, no misconnections happen.

It’s been about 3 weeks since I’ve put ES 365 in the “driver’s seat” to steer Loop DIY. I’ve discovered that ES lags another 15 minutes or so when compared to when compared to the G6 when coming from a low BG. The ES has been very good to excellent in accuracy overall, especially at the alert-consequential low 70s to upper 60s range.

Here’s a BG trace for just after my last meal yesterday till this morning through my typical dawn-phenomena/feet-on-floor cortisol spike. It shows an average BG of 87 mg/dl (4.8) and 10% CV, an exceptional range. I still woke up at 3 am for an hour, however, but that is still a work in process.

My current 14-day AGP report shows 86% TIR (65-120) with an average BG of 97 and time below 54 at 0%. CV = 20.5%.

I’m happy with these results, especially the disappearance of the false low alarms I experienced in far too many Dexcom sensor sessions

[Eric]

Hi @Terry
Thanks for the Eversense report. It’s interesting.

How does it compare with your meter? How quickly does it detect a rise or fall?

The Dexcom is fine for me when my BG is flat, but it takes a while for it to catch up to changing BG.

[Terry]

Thanks for the question, @Eric. I’ve only had about 3 1/2 weeks of ES data and I think this trend should hold but things may change over a longer period. Plus I’m certain my perception of the data may become clearer as time goes by.

The biggest thing, as my post states, is that precision, accuracy, and consistency at the low end (54-70) is much better than the Dexcom G6 was. In one particular episode that the ES was reporting a hypo in the sub-54 range, my frequent finger-sticks agreed. I perceived that the ES tends to time-lag the G6 by about 15 minutes or so. This was true in at least the one hypo case that I mentioned cause I was treating with glucose tabs and used multiple finger-sticks to prevent over or under-treating.

That incident was unusual in that it occurred just after a full meal but before my Loop-advised delayed meal dose. I ended up delaying that dose for about an hour while I was also taking several single glucose tabs about 10-15 minutes apart. I’ve often delayed my meal dose by 15 minutes or so after finishing eating but an hour delay was highly unusual.

The biggest change for me is that the false lows stopped. I have little patience for these situations as it often meant needlessly losing sleep. Most of the time, the G6 was dependable but accuracy, precision, and consistency degraded on days 8-10. Not every sensor but often enough to undermine confidence.

It’s only where I’ve treated a hypo or hyper and use the finger pokes to slow down my tendency to over treat. In short, I think the ES sensor, for me, lags the G6 by 10-15 minutes both going up or going down. Since CV% is usually held to 20 or under I don’t know the condition of a sustained rise or fall of significant duration.

[Eric]

@Terry

In short, I think the ES sensor, for me, lags the G6 by 10-15 minutes both going up or going down.

Thanks for the reply @Terry.

Just to make sure I understand this, when you say the Eversense “lags” the G6, are you are saying that the Eversense is 10-15 minutes worse than the Dexcom?

Did you ever have them both on at the same time? To me, the true test would be to do a meter check and compare it with both the Dexcom and Eversense readings all at the same time.

[jim26]

Thanks for sharing your experience!! I’m excited about the technology and what is could be. What if it advanced to the point of only needing to be replaced every 5 years or 10 years?

Curious if there was a hassle with getting insurance approval? Dexcom at this point is a standard of care, so no hassles in getting prescriptions.

[Terry]

@Eric

To me, the true test would be to do a meter check and compare it with both the Dexcom and Eversense readings all at the same time.

You’re right, Eric. I didn’t run the G6 in parallel with the Eversense. In the process of getting the ES running, I deleted the G6 transmitter. I did, however, do many finger-stick readings with the ES running but I did not gather the data in any organized way.

It was my intention to run the G6 at the same time. I had mistakenly thought I could run the ES app and the Loop app at the same time. Turns out only one of these two apps can communicate with the ES transmitter. While deleting the ES app, I unfortunately deleted the G6 CGM. The G6 CGM deletion is one-way event that is not reversible.

[Terry]

@jim26

Curious if there was a hassle with getting insurance approval?

It took a little longer than necessary. Eversense needed my endocrinologist to complete a medical document but they were not happy with the first answer and had to ask my endo for some clarifications. The PA who did the insertion needed a Rx from my endo.

I started an email conversation in the middle of April with a local area rep as well as another Senseonics official. I estimate that it took about a month to get the application completed and the medical info from my doctors. I chose to wait till July 15 as that would be 90 days after Medicare paid for my last G6 shipment. That added another two weeks.

I expected that getting this process moving would take some persistence from me & I had to follow up with phone calls and emails. It’s like any bureaucracy; the more people involved in the process, the longer it takes.

The Medicare billing is not finished yet but I expect for my traditional Medicare (not Advantage!) to pay for 80% of the hardware/software/service as well as the insertion service. I expect my supplemental insurance to pay for the remaining 20%. I just checked at Medicare.gov and the claim has not been submitted yet.

I trust it will go through OK as Senseonics was careful to ensure Medicare will cover me.

[Terry]

It’s now been one month since an Eversense sensor was implanted in my right arm. I looked back, comparing 14-day AGP reports, one over the last two weeks, with a G6 report from July 1 to the 14th. The results for both periods were each excellent and doesn’t show the ES 365 any more accurate than the Dexcom G6 it replaced.

Yet I see the ES 365 as clearly superior in many regards. First of all is the quality of life (QOL) it gives me. My practice with the G6 was to insert it 12-24 hours before the expiration of the current sensor. Not a large responsibility yet does require some added attention that day and energy given to my electronic calendar and the prompts I built into being reminded so that I didn’t forget.

My usual practice of placing the G6 sensor was due to reduce the first day inaccuracy of a sensor placed immediately before turning it on. This usually worked as intended but not always.

As I said in the original post, I found increasing inaccuracy frustrating, including the maddening false low alerts, on days 8-10. This required attention and careful recalibration. Many times, careful recalibration, was met with “stubbornness” on the part of the algorithm to accept my thoughtful and well considered calibrations. I tried not to over calibrate, never made large calibrations (> 30 mg/dL), and did not calibrate at times when my actual BG was changing quickly.

The actual time and attention I allocated to these routines now seems more significant when contrasted with no comparable ES routines. For someone like me, who used tech (pump, CGM, AID system) to both exceed the usual standards of care and provide good QOL, these efforts for many years seemed a reasonable trade-off. Now that I’ve been relieved of the required sensor attention on days 1 and 8-10, I am grateful for the reduction of responsibility. When diabetes chores go down without sacrificing performance, I count it as a win!

[RachelMaraii]

@Terry

Now that I’ve been relieved of the required sensor attention on days 1 and 8-10, I am grateful for the reduction of responsibility.

That does sound really nice… I’m on day 1 of a new G7 sensor and just had to calibrate from 47 to the real meter reading of 130, and always knowing that the first day frustration like this will happen does stink.

[Terry]

@RachelMaraii

I’m on day 1 of a new G7 sensor and just had to calibrate from 47 to the real meter reading of 130,

Happy to read that you’re keeping the technology honest as you negotiate pregnancy with T1D!

[TomH]

That’s a great report! Thanks for posting it! I have looked at the ES and put it on my short list of CGMs to consider. With the comments from a CEO interview, I’m hoping the “no transmitter” version gets the ok soon; supposed to be about twice as long, go straight to app (also to Loop/Trio AIDs), but the G8 is supposed to release next year also and supposedly good features as well. We’ll see!