Fiasp Day 17 update

Our Fiasp experience started October 11th.  Here’s what I was expecting:

• increased insulin resistance
• fast insulin

Optimistically, we gave our first Fiasp bolus.  Full of excitement.

After 24 hours, here’s how I summarized our experiences:

1. If Anna started to rise with pretty significant IOB, I adjusted basals up. The last adjustment was leveling out beautiful, then started to climb for three consecutive readings about +3 or +4 each time, but the important part is that it was happening with 1.8 u IOB. That means she was really short on basals since it was nearly 3 hours after she had last eaten.

2. Also, she climbed more than 150 mg/dl from eating one single uncooked spaghetti noodle. Do I really think that was all noodle? NOPE. Definitely another sign that basals were too low. Shouldn’t be spiking that bad from small food.

3. Only one adjustment was to lower basals…when she had dropped significantly and had negative IOB. That was the adjustment right before the steady line at night. I had apparently gone too far with the earlier basal increase.

4. We prebolused the first meal by about 8 minutes I’d say. A peanut butter sandwich that she overcounted carbs on. That’s the only low we corrected on this graph. (Fiasp rebounds quickly from inadequate basals). Tonight’s sandwich is the same meal, but bolused with a stronger carb ratio, and no prebolus. Looking better.

In short, an awful lot more insulin needed almost right away. Like A LOT more. BUT, the no prebolus thing is real. And Loop’s fiasp curve is working better and better as I get my settings better fleshed out.

We had a couple good days, but then things started to get a little worse for wear around day 4.  We were having an awful time of:

• volatile BGs, more lows to treat/more highs to stare at and wonder
• fast insulin
• insulin sensitivity
• frustration with looping
• distrust of looping (both Loop and OpenAPS, I tried both)

Our days just got progressively worse.

Then on the evening of October 18th, we just had a lot of stubborn lows and I was sick of spending so much time on diabetes again.  We were treating a lot of lows in advance on those graphs.  I had a choice between throwing the Fiasp into the trash (considered mailing it to a friend but wondered if that would even be considered “nice”) or *gasp* suck it up and go back-to-basics.  I opted for back-to-basics…aka open loop test all our settings.

It didn’t take long to find my first problem on open loop.  Notice on the screenshot above, the low near midnight after a manual correction?  WOW, that correction brought her down over twice as much as I’d expected and I needed to treat a low.  Basals were too strong and I suspected my ISF was way off, too.  I lowered basals as I treated the low but didn’t adjust ISF at the time (after all, I was still open looping so ISF wasn’t actively being used other than by my brain if I wanted to do a correction).

For two days, I opened my loop so that I was in control.  I tested basals, tested ISF, tested carb ratios, watched my IOB, and watched Loop predictions during all of this.  Things got better pretty fast.  After a day and a half, things had smoothed out quite a bit with a lot of adjustments.

The screenshot above shows a few things I learned.  That dip at middle near noon…Anna’s PE class.  She is very sensitive to her lunch bolus since PE comes right after.  We are still working on that.  It’s not a fiasp issue, per se.

But, the more subtle observation?  See the Loop prediction?  Loop was predicting that she would start coming back up around 12:30pm.  But instead, she was still heading down.  This started me wondering if either my carb ratio was still too strong, and/or if I needed to maybe shorten our default carb absorption times.

It became pretty obvious that my carb ratio was still too strong in the next few meals.

Basically after two days of open loop use, I learned that my basals had been too high, my ISF had been too weak, my default carb absorption times needed to be lowered, and my carb ratio was too strong.

The odd part to me was that you’d think that with all those indicating that I would need less insulin…we would’ve only been battling lows while closed looping the week before.  But, we weren’t.  We were fighting lows and highs.  I think that all the suspends to keep us from going low were leading to some strange rebounds with fiasp.  It was really hard to see through all the looping noise to figure it out though.

In the end our average numbers by comparison have ended up as:

• Novolog: ISF = 40, Basal = 1.0, Carb ratio = 7.5
• Fiasp: ISF = 58, Basal = 0.85, Carb ratio = 10.5

Our settings now are working much better.  We closed loop again and are happily looping on Loop.

Other than the standard things we learned on open loop regarding basals, ISF, and carb ratio…we also decided to shorten the 1.5x carb absorption multiplier that is default in dynamic carb absorption.  Basically, we shortened our default carb absorption buttons to 1, 2, and 3 hours for lollipop, taco, and pizza.  Other than that we are using the standard Loop settings.

Why did I adjust the default carb times?  Because of the way Loop calculates bolus recommendations, the quicker peak time of Fiasp will predict an early low after eating meals compared to a similar meal with Novolog.  So, even if everything else is equal, a meal bolused with Fiasp will tend to get less of a bolus typically for an average meal than a bolus with novolog in Loop, and this effect gets more pronounced the longer the carb absorption is entered.  Therefore, that 1.5 multiplier will tend to lower the recommended bolus even more.  What we found happening was a smaller upfront bolus would be followed by a high temp/high BG, and then we would crash later as the later carbs (from the rest of the 1.5 multiplier area) wouldn’t be there to support the earlier high temps.  So the upfront and later parts of the meal were being affected by the multiplier. (*screenshot from a pizza meal below and part of this screenshot was also affected by our settings still not being tuned, so take it with a grain of salt.)

On novolog, we never had this issue I suspect because the peak time of novolog was so slow that dynamic carbs had enough time to adjust the predictions before the high temps came on and couldn’t be recovered from.  And we prebolused a lot, so the insulin was pretty active by the time she ate to help prevent spikes.

So, in summary, I’m stoked on Fiasp.  It’s been great now.  BUT, it didn’t behave like others had experienced…so keep your mind open.  Perhaps we will find resistance later, after longer use.  If we do, I’ll be sure to document.  And, if you find yourself slumping into confusion initially with the change from novolog/humalog…don’t be afraid to open loop to get your feeting solidly beneath you again.  A day or two of open looping can save you from wanting to poke eyeballs out.  And, sadly, some meals can still really benefit from a 5-7 min prebolus even on Fiasp.  The really fast carbs are still faster than Fiasp.  Simply announcing those carbs won’t be enough for my teen’s fast digestion.  We are learning which foods need prebolus and for how long…but the list is A LOT SHORTER than with novolog (novolog list included just about everything but water and ice).

---

Side note (because I also love the not-perfect-examples, too): Fiasp still doesn’t save you completely from a really poor carb count and a busy teen.  Example, she ate 20g uncovered just before I picked her up from karate.  I didn’t know that and she was just distracted.  We almost immediately went to In-n-Out where she scarfed a double-double with a HUGE french fries basket.  We were way off on carb counts (originally this graph only had 80g on it because I didn’t know about the previous 20g and I didn’t know she was getting fries, too).  We adjusted carbs a couple times as I found out about things…and gave the suggested corrections as we adjusted.

But, the recovery from such a bad carb count, no prebolus, and eating 20g without any bolus for that portion…really quite fantastic and quick.  3 hours after that meal was eaten we are recovered…and I don’t think we would look like this with novolog.

(As with all things, don’t take my word as gospel on Fiasp.  I’m still experimenting. YDMV.)