When to treat a low?

80? 70? Something else?

We treat out son when he crosses below 80 (so 79 or lower).



I am not stalking your posts, I promise!

My feeling on the numbers are:
No two 80’s are the same… No two 70’s are the same…

How fast is it dropping? How much insulin is behind it pushing it down. If you leave it alone, will it stop at 70? When is the next meal? How recently did he eat? Is the food that he ate about to start coming in?

So many factors to consider!

I guess the easy way to say it is - the numbers are only part of the story. Sometimes I treat a 90 because I know that if I wait it will turn into an 80, then 70, and keep going. It just depends. Sometimes an 80 will stay there. Sometimes after dinner, I could be at 60 but just ride it out because I know it will turn up in a while.

Sorry, maybe that wasn’t an actual answer!


Yes, I strongly recommend treating a low based on trending arrows or numerical rates of change on Dexcom, plus knowledge of IOB, basal, etc. not just the absolute number. @Eric is absolutely right.

But I do think getting some number guidelines is super helpful.

Here is one way to think about it. Dexcom measures interstitial fluid, not blood. It has an algorithm to try to correct this lag, but in theory, it could be up to 15 minutes lagged behind real blood sugar. It takes 10-15 minutes for blood sugar to rise after a low treatment, right?

So a really conservative treatment threshold would be something like this:

Look at the number right now, how much it’s dropping, and calculate: If he were to drop at this same rate or greater over the next six Dexcom readings, would he be low or not? If yes, then treat.

Let’s say he’s 125 dropping by -10. That means, assuming Dex is really lagged far behind his true drop, he might actually be 95. Then if you were to treat him, you know it takes 10-15 minutes to rise. Over that time he could conceivably drop from 10 to 30 points, meaning he could possibly go low before he starts rising…but just barely, to 65. In this situation though, that’s actually a pretty conservative treatment threshold because with such a slow drop the odds are Dex is actually really close to his true blood sugar value, in which case if you treat him at 125 the lowest he’ll go is 95. So I would wait another reading. Does that make sense?

Take another example: He’s 155 and dropping by 30. Wow, that’s a huge rate of change! Who knows what his true BG is because it’s highly likely the Dexcom just can’t keep up with his fast rate of change. So I would in this case suspect he might actually be 120 or 100. I would treat right then because if he starts at 100 and he’s dropping by 30, then even if his blood sugar starts to rise in 10 minutes, he could plunge to 40 before it starts rising.

(It’s not totally this simple, because as you know a -10 that’s decelerating quickly at 100 is very different from one that’s been steadily ticking for the last 30 minutes at the same rate. I do use the image and the slope of the curve to gauge whether a treatment is necessary at home. AT school we don’t bother with that; it’s too sophisticated for the preschool teachers who are also wrangling 12 other toddlers per teacher.)

I, like Eric, also keep track of IOB for low treatments at home (again, teachers don’t know or care about this number.)

I think you should have fixed in your mind how much insulin a simple low treatment “soaks up.” So let’s say you treat with 4 grams – and your carb ratio is 1:24. That means one 4g snack requires 0.166 units of insulin to process. If he has, say, 0.5 units on board, I’d expect it requires at least two, possibly three, treatments eventually. But that is not enough IOB that I need to give all of that upfront or he’ll start to spike, become insulin resistant and then never come down! So instead I give one 4g snack, wait till it plateaus and starts to drop again, then another, then repeat until all the IOB is soaked up and he’s straight arrow across.

Is that confusing? Please let me know if I’m inundating you. I’ve just found that these types of calculations are incredibly useful for me managing my son’s blood sugar, and that if you’re doing everything you can to avoid lows it will, counterintuitively, help you avoid highs too, because a) not on the roller coaster and b) you have confidence in avoiding lows and so can dose for highs more appropriately.

I’m a very mathematical person so for me i think of things very much in numbers … but let me know if this all sounds like gobbledygook.


… And as to your first question: We treat during the day absolutely any time he drifts below 80, but at night if he has no insulin on board and is dropping reaaaally slowly we may let him ride it out till 70. People are supposed to have slightly lower BGs at night and 70-80 is not truly low, so this allows him to get more sleep (and us too!). Also often avoids a future low in the night because if we treat the low at night he may come crashing down again after the openAPS adds a bunch more insulin to compensate for the quick spike.

1 Like

Nope I totally get it. I’ll dissect later and analyze with my wife to see how we can use all this info for our son.

Thanks so much for the plethora of great information.

There are currently a lot of apps that allow you to see trend predictions, xDrip+, Predict bG (?), and others. I have personally been using xDrip+ as the predictions are usually pretty accurate. I’ve been treating based on the prediction, sometimes at 90 or 100 and have only been burned once.

1 Like

I agree with Eric. The context and history preceding any number is just as important as the number itself…

1 Like

We treat at 70 when no IOB.

When IOB, we treat earlier, for same reason as @TiaG said. When coming down from a hormonal peak, we treat very early, above 100 to CGM, as high as 120 depending upon slope sometimes. More typically, we treat somewhere btw 80 and 100 by CGM depending upon slope.

To figure when, IF WITHIN 2 HOURS OF INSULIN INJECTION, take the slope rate, say 10 per 5 mns, multiply by 2.5, in this case 25, and add to 75. In this case it would mean treat at 100.

[EDIT] You multiply by 2.5 because this is roughly by how much your Dexcom is trailing, assuming that it is about 12.5 mns behind real time, and a steady rate of descent.

If later than 2 hrs after insulin, we treat lower, or do a manual test first, because it is possible that it will inflect or even out.

1 Like

This seems like a good tool for folks who aren’t number obsessed to do a good job responding to Dexcom trends proactively.

LOVE this - great start point for us to begin doing our own analysis of this type. Thx! <3

1 Like

I also look at delta values when determining treatment. Depending on the starting BG, I will correct with fruit snacks or Mt. Dew. There was one occasion where he had a stubborn high and multiple corrections. When the flood gate broke, there were were delta values of -30’s. Even though Dex said he was still in the 200’s, we knew to give him Mt. Dew to slow things down to a nice soft landing.