Changing DIA?

So, my son has been having some pretty bad lows the past day – 3 below 55 in the last 24 hours alone!! What’s weird is that his lows seem to come towards the end of what should be the DIA, and then they seem to really accelerate in a way we’re not used to.

If it happens 2 or 3 hours after a meal I usually assume the carbF is too strong. If it seems to happen within an hour or two of a basal change I tend to assume it’s an overly high basal rate. I also tend to assume it’s an overly high basal rate if he’s going low when he technically has negative IOB or no IOB…to me that’s an indication that the baseline (what’s assumed to be 0 basal) is too low.

But as a sanity check i usually look at the TDD – if his overall need for insulin went down then I assume it stands to reason that his basal insulin needs have decreased.

But this time his TDD is actually on the higher end – 8.5 units to 10 units, as opposed to his typical 6.5 to 8. Obviously some of that TDD is to deal with some of the lows, but not that much of it.

To me the logical conclusion is that maybe the insulin is just acting in his body longer than we typically assume. We have his DIA set to 4 hours and am tempted to try switching it to 5. But is that weird? Do people’s DIAs typically change a lot over time, or is that a fairly stable variable.

Also, if I change his DIA, I assume I’d also need to change his ISF – are there formulas that can relate DIA and ISF so that he is net getting almost the same amount of insulin, but with a stronger ISF and a longer DIA?


Tia, I am sure there will be some others with more experience by in a bit that will have a more technical solution, what I will say is that lately my son has been running almost autonomously, his diabetes care, and this has resulted in not much changing. i.e. he doesn’t stress over his reports and tracings and look for changes unless I initiate, and we have had a few days out of the last two weeks where we have had a burst of strange lows that are unexplained and not due to changes in what he eats or how much insulin he is delivering. So I might fall into the camp of lots of strange things happen, and they don’t always have an explanation that is understandable with our current level of data inputs. If they are like my sons they will straighten out in a bit.

A number of things can affect the DIA. The most common are:

  1. location of the infusion site
  2. number of days you’ve been connected to the infusion site
  3. temperature
  4. activity level
  5. use of muscles located near the infusion site

It can vary with the factors I mentioned, and also just seasonally based on different things happening with your body. The best thing to do is to make note of different factors so you can try to see if there is a pattern with any of the things I listed. If the change appears only on occasion, then see if you can figure out a pattern. And if the change seems to be a new normal, by all means, adjust the setting.

Start with infusion site and length of time. For example, look for patterns like - it does not always happen when it is on his arm, but it never happens on his leg. Or - it never happens on day 1, but sometimes on day 2 and frequently on day 3. Just things like that. Because it will probably not be consistent, so you have to look for frequency.

1 Like

I’m not educated enough to answer the heart of your question unfortunately, but as you’ve helped Erin and I understand…watch for the patterns. If there isn’t a pattern, it really could be anything at our kids ages.

1 Like

infusion site is definitely a possibility. I’ll take a look at the day-since-insertion pattern.

1 Like

I also think some of this may be due to what I call the “Because It’s Tuesday Phenomena”, i.e. there are so many other minute metabolic processes and kinetic uncertainties playing off one another that we are not able to quantify and/or control.

It’s so maddening to be constantly shooting at a moving target…

Parents of kids with D (and everyone trying to manage this beast to the best of their ability) should not forget to give themselves the credit that is rightfully due for all the time we do manage to get it “right”.


I should have asked this initially, what kind of insulin do you use?

We use Novolog. We were using Humalog but it had a really long tail in our
son and didn’t onset quickly enough, for whatever reason.

Have you actually tested for DIA, or are you going by what the endo is saying?

My experience is that the best way to figure it out is at night, after dinner. What I do is that I look for changing behavior after a certain number of hours of the insulin injection, typically a change in trend or in “look”. During DIA, the curve is flatter, and has very few upwards bumps.

For my son, DIA seems to vary between 4 hours 30 and 5 hours 45 depending upon ??. I use 5 hours as a reference for calculations (but we do not pump, it is for our own stacking calculations).

We have tested DIA extensively in the past, as it’s a major factor in the
success of openAPS. The issue is that we got pretty comfortable wit a
4-hour DIA after finding that a 5-hour DIA was too long several months ago.
Medtronic only allows DIA in 1-hour increments. I’m just wondering if the
true DIA varies all that much. It’s just one more variable that must be
factored in with openAPS and it’s already hard enough to debug these issues
when you’re factoring in the calculations of the algorithm on top of the
settings you’re tweaking.

This is probably well known by everyone, but just want to emphasize - duration is never linear.

It’s much more like the blue line than the green line (apologies for my primitive drawing).

A lot of pumps and calculators do not consider this, but I do not know why…

UPDATE: So we’re going to try lengthening the DIA and then strengthening the ISF by 20% – basically the idea is our overall insulin dosing seems mostly right, but the timing is wrong, so I don’t want to cut back on how much insulin he’s getting total. I’ll keep people posted; it could be a disaster.


What have you seen so far?

It’s actually been good! We haven’t had any of those lows that start accelerating at negative IOB…he has still had lows but not nearly as bad. And it also seems like decreasing the ISF has made the openAPS algorithm work a little bit better. But it’s still early days; it’s hard to know whether he’s just had a good few days’ run or what.


One explanation for this is “anti-insulin antibodies”…

Antibodies are molecules manufactured by the immune system to latch onto foreign bodies. Antibodies that latch onto insulin molecules are called anti-insulin antibodies.

There are various types of these things, but certain types of them simply render the insulin molecule inactive for several hours until the antibody “breaks off” again. This means that the action of an insulin bolus is stretched out over a longer period. In practice this means an unexpected high, which, if treated, is followed by stubborn lows.

Unfortunately there is very little research about this problem. My endocrinologists says that there’s not much to be done except (risky) immunosuppression. However, I find that reducing my total daily insulin and using short extended (square wave) boluses seems to make it less likely for the immune system to “decide” to produce lots of these antibodies.

After I learned about these things, diabetes made a lot more sense and was easier to manage. When I get an unexpected high I know that one of the possible explanations is antibodies and that I may have a lot of insulin in my system that will hit me hours late.

1 Like

There has actually been a fair amount of research concerning insulin antibodies. Here is a good explanation of what they are and of the significance in insulin therapy:

This article describes a case in which severe insulin resistance was due to insulin antibodies:

One more describing testing for insulin antibodies, written for the patient:


Hmm - I like how reading posts helps me rethink trends I’ve seen. Case in point, I’ve notice when I take even mild antihistamines it “seems” like I need less insulin for a 8 - 10 hours (my I/C ratio goes up). I don’t take antihistamines often but enough that I’ve noticed this.

@docslotnick Wondering if this article you posted has any correlation?

1 Like

@Millz I don’t think that anti insulin antibodies would really cause mild, transient insulin resistance. From my reading this type of insulin resistance in T1’s is pretty rare and dramatic.

I do believe that most transient “insulin resistance” in T1’s has more to do with variability of insulin action and other external factors. Of course other drugs, like antihistamines could be one of those external factors.


Thanks for the feedback.