Pumping Insulin's IOB recommendations

Type 1 and LADA
1

I am reading through the latest (6th) edition of Pumping Insulin and am struggling a bit with his method of counting IOB. He gives an example of someone who takes a 10U bolus for dinner, and then at 2 hours, with 5U still on board, looks at his blood sugar and decides how much to bolus for dessert. According to his i:c or carbF, the dessert should be covered with 5U. However depending on the IOB and 2hour PP, Walsh says that dessert can be in some cases completely covered by the IOB. Basically, he’s only adding insulin to correct the high BG, and doesn’t add anything new for the dessert.

I am really not sure if I’m understanding him right. :blush:

  • Why does IOB that is being used for carbs that are still being digested somehow also able to do double duty and cover the carbs in dessert?
  • Is the real thing going on here less about IOB and more about most likely the carbs in the meal were overestimated on nights 1 and 2?
  • I notice he’s correcting high BG at 2 hours PP. I usually wait longer, until at least 3 hours since my insulin is barely working at 2 hours, and then I see a drop later. I’m unsure if he’s saying “correct at 2 hours” or if I should be looking somewhere else in the book for the rules on when to correct a high PP?

Here’s the page in question. I hope it’s okay to post it!

IMG_2654.JPG2016Ă—1512 1 MB

2

This is, I think, a really interesting question. We have been at this for less than a year, so better experts may give a better answer.

We have the same book and have worked hard at applying the IOB concept over the past few months (my son does not have a pump but we calculate the IOB by hand when we need to make corrections). This has been my read:

  • The assumption the book makes is that, after 2 hours, the carb peak has done its work (I find such IOB calculations don’t work as well when there is high fat in the meal).
  • so, the leftover insulin after 2 hors is “extra” and can be applied to the upcoming dessert - essentially, on Night 1 and 2, the patient is going to go low and overestimated his insulin need.
  • The patient did overestimate his needs on nights 1 and 2.

Using the IOB concept has been very useful to us when making correction calculations, even 1 hour after injections. We just spent a week in Louisiana, and we had on average one meal per day where we had to make multiple corrections due to bad guesses. I calculated the IOB every time and used it to make a better guess on the outcome. The fact that we had multiple corrections shows that my guesses were not as good as they could have been all the time:-)

1 Like
3

Let me take a stab at answering these questions.

This is a good question. It seems to me also, that the original 10 units was just for dinner. I think possibly that is a mistake by the author.

To test this, I put his EXACT numbers into my pump (correction factor, target BG, carb:insulin ratio, the BG he shows, the carbs in the meal, the desert) Everything was the same has he describes, and I didn’t get the same recommendations from my pump that is listed in his table! My pump showed what you expected, that the dinner bolus was for dinner only.

I think the author is just trying to say that sometimes the bolus recommendations miss factors. He specifically mentions that the user eats the same thing, so I don’t think it is about mistakes. He also mentions that the user exercises for example, and the bolus calculator on the pump doesn’t take that into account. The point of the section is just to not always blindly follow your pump recommendations.

No, the author is just using 2 hours to make the math easier to see. The user in his example probably has a duration of 4 hours, so he uses 2 hours to make the math easy (4 divided by 2 is one half, so it is easy to represent that half of the 10 unit bolus is 5 units left on board).

I don’t think there is any rule that is always right on “when” to correct. It just depends on so many things. I definitely do not think the author is saying to always correct at 2 hours!

I would be hesitant to apply an “always” rule to that. So many factors, like what you had to eat (high fat for example), how quickly you are spiking, whether you are going to bed soon, or if it is in the middle of the day, or if you have easy access to sugar to fix a low… All kinds of factors there!

1 Like
4

Hi Michel,
Since you are doing this manually, can I ask you what formula you use to calculate IOB? I have seen some really good ones, and I have seen some that are not so good. Many pumps use a linear formula, which is NOT a very accurate representation. Insulin does not get absorbed at the same rate from beginning to end. The ones that reflect the different rates from onset to peak, and from peak to end, are better. You are a math type, right? So I am betting you use a good one…

Maybe someone has some good IOB formulas they can post. I don’t go by formula, I use “the force”. :wink:

2 Likes
5

haha I love that. :slight_smile: I am doing IOB by hand since I haven’t started on my pump yet and also use the force.

6

Thanks, I think you are right. So much of the book encourages following rules and formulas that I was missing the point of the example, that the formulas don’t always get it right.

1 Like
7

As far as following a set of written rules, I am quite possibly the worst example that a CDE or Endo would ever see. I don’t use the calculator or follow the IOB. I think there are so many things that just aren’t in there!

On my pump (and probably a lot of others):
The calculator factors current BG, but it doesn’t look at which way it is moving and how quickly. That makes a big difference.

It has a food library with carbs and fiber and fat, but it doesn’t show sugar grams. The glycemic index of a food makes such a difference in how quickly you spike. So I will hit the bolus much sooner for a high GI food. And I will anticipate more of a spike with a high GI food.

The calculator doesn’t look at activity, what you’ve been doing, what you will be doing.

The calculator doesn’t know when your next meal will be, but you may know. And that can affect how aggressively you bolus.

They don’t consider how well the insulin has been working for you that day. That has so much to do with various factors, like previous day’s foods, exercise, stress, hormones, etc. The Tandem pump lets you easily have different profiles setup for that sort of thing, but I think on some of the pumps it is not as easy to change those values constantly.

There are always so many adjustments. As Captain Barbossa famously said, "The code is more what you’d call guidelines than actual rules."

2 Likes
8

You made my day with your use of the Force, Eric!! And I am guessing that with it you are probably more accurate than me with the math:-)

I do use a formula. The first step to determine, same as with a pump, is to find out the Duration of Insulin Actio (DIA) for yourself, i.e. exactly how long insulin acts on you. It is quite tricky to measure because, after a bit of lag time, the insulin acts quickly, then its action slows down more and more but keeps on working for quite a while, almost imperceptibly at the end. To calculate the DIA you need to check how long it acts, all the way through to its end. For instance, with my son, right now his “fast” insulin action ends up typically after about 95 minutes, but goes less fast for another 45 minutes, then it creeps down for another 160-175 minutes roughly, so a total that I round up to 5 hours.

Once you have your DIA, there are formulas in Pumping Insulin (the book quoted by the OP) for each DIA duration. For 5 hours, this is the formula I use, slightly adapted from the book for simplicity of calculation:

  • During hour 1: 10% insulin used
  • During hour 2: 27.5% insulin used (I typically round this down or up to 25 or 30% depending upon the time of observation for calculations)
  • During hour 3: 27.5% insulin used (see comment above)
  • During hour 4: 20% insulin used
  • During hour 5: 15% insulin used

So, here are examples of how I would calculate that:

A. Say my son injected 4 units, and, 3 hours later, he needs to do a correction. In hours 4 and 5 there is 20% + 15% insulin to be used, for a total of about 1/3 of the total, so that means that at the end of hour 3 there are 4/3 units of IOB (insulin on board) = 1.33 units IOB left.

B. Say my son injected 5 units, and, 1 hour and 50 minutes later, he needs to make a correction. In hour 1 he used 10%, in hour 2 he would use 27.5%, but, after 1 hour 50 minutes he is a bit short of 2 hours, so I use 25% instead, for a total over 2 hours of 10+25= 35% of insulin used, about 1/3 is already used. So, after 1 hour and 50 minutes, he has about 2/3 insulin left, i.e. 5*2/3 IOB = 3.33 units IOB.

I will look up the formulas for other DIAs (3, 4, and 6 hours) and put them up on line as well.

As a note, I also have a look at the actual curve I plotted for my son. This is his actual data, based on his CGM curve:

  • insulin action lag: 25 minutes
  • fast action period: 70 minutes, 60% of insulin used in this period.
  • medium action period: 45 minutes 20% of insulin used.
  • slow action period: 160 minutes, 20% insulin used.
    You will note that the numbers don’t quite match the formula I use.

You may ask why I don’t use my son’s actual formula. It is because, to determine that formula in the right manner, you would have to take a meal, wait about 5 hours so that there is no actual active digestion (and glucose generation), then apply a large dose of insulin, and keep fasting for the DIA, so another 5 hours. I did not submit my 12-year-old to this regimen, so my curve is approximate,and I don’t quite trust my own numbers as much as the “standard” formula. But I may be wrong.

Sorry for the very long post. I tried to be as clear as I could:-)

2 Likes
9

Based on Pumping Insulin table 8.5 on page 98 (the numbers are rounded for simpler calculations):

  • For a DIA of 3 hours:
    Hour 1 = 30% insulin used in that hour
    Hour 2 = 45% insulin used in that hour
    Hour 3 = 25% insulin used in that hour

  • For a DIA of 4 hours:
    Hour 1 = 15% insulin used in that hour
    Hour 2 = 35% insulin used in that hour
    Hour 3 = 30% insulin used in that hour
    Hour 4 = 20% insulin used in that hour

  • For a DIA of 5 hours:
    Hour 1 = 10% insulin used in that hour
    Hour 2 = 27.5% insulin used in that hour
    Hour 3 = 27.5% insulin used in that hour
    Hour 4 = 20% insulin used in that hour
    Hour 5 = 15% insulin used in that hour

  • For a DIA of 5.5 hours
    Hour 1 = 10% insulin used in that hour
    Hour 2 = 25% insulin used in that hour
    Hour 3 = 25% insulin used in that hour
    Hour 4 = 20% insulin used in that hour
    Hour 5 = 10% insulin used in that hour
    Hour 6 = 10% insulin used in that hour

Please note that (a) there is no calculation for a DIA of 6 hours, but (b) the numbers for the 5.5 hour DIA don’t seem to make sense since the 6th hour should really be 1/2 hour (there is too much insulin for 1/2 hour compared to the 5th hour), so I figure that the numbers for a DIA of 5.5 hours may be used for a DIA of 6 hours, really.

In another note - the book mentions that many pumps show programmed DIAs of 3 hours, but that almost everyone is between 4.5 and 6.5 hours, and that studies comparing children and adults have found little difference in DIA between the two. It also mentions that, for large boluses, longer DIAs are better approximating actual use (I think, from memory, the book mentions for that prupose DIAs of 5.5 to 6.5 hours, but this is from memory). FYI, from experience, my son, 12 years old, has a DIA of approx 5 hours (although it could be 5.5 hours because it is really difficult to know when it ends).

Hope this makes sense. Of course, this is NOT medical advice:-)

1 Like
10

If I were to translate my son’s curve to the same notation, using the exact timing I have for him (I rounded the numbers for easier presentation) it would be:
Hour 1: 30% insulin used in this hour
Hour 2: 45% insulin used in this hour
Hour 3: 10% insulin used in this hour
Hour 4: 7.5% insulin used in this hour
Hour 5: 7.5% insulin used in this hour

But, as I mentioned before, since the experimental method I used to get to these numbers is not perfect (glucose production was not exactly zero during the experiments), I doubt that these numbers are quite accurate.

1 Like
11

These numbers seem very reasonable. It sounds like you are doing it better than many of the pumps, because you are applying a non-linear formula, and using the different periods of fast, medium, and slow. That makes the most sense to me.

I think that with enough time, coming up with your own numbers will always be better than going strictly out of the book. Because, while those books are written for millions, you are figuring it out specifically for only one!

I think finding the DIA is probably always the toughest thing to do. As you stated, nobody wants to go through those periods figuring it out.

1 Like