Best post-prandial testing time

I’m a little confused.

According to the Joslin Diabetes Deskbook, and many other sources, one should check one’s blood sugar 2 hours after meals. But if insulin (such as Humalog and Novolog, and Apidra in some cases) lasts for 3-4 hours, what information will this give?

When I was at a program at Joslin Diabetes Center a month ago, they emphasized checking 3 hours after a meal. They said one could check 2 hours after, but one could not act on that information. 3 or 4 hours after the meal, and one could correct if necessary. (Correcting after only 2 hours would result in insulin stacking.)

So, what ought one to do?

Well, it depends on the information you want to get. When I first started using insulin, I wanted to know when my peak was, and how high it was, so I tested 1/2 hr, 1 hr, 1 1/2 hrs. and 2 hours to find out when MY peak was. It turned out to be 1 1/2 hrs, so the 2 hr. advice would not have served me well. Knowing how high I peak is important to me, because it teaches me what foods to avoid, and how much to limit portions and carbs in order to keep the peaks as low as possible.

But when you want to know when to correct, it depends on how long the insulin lasts for you. It gets out of the system faster for some people than others. The way I figured it out was with my CGM – after 1 1/2 hrs. the peak starts coming down, but at 3 hours, it levels off, even if it hasn’t gotten down as low as I want it to. So I figure insulin lasts 3 hours for me. So between 3 and 4 hours, I can test and correct if need be. But you could figure that out with your meter and strips, too, if you have enough. Keep testing every half hour, and see what your BG does. But do remember that high fat+carb (like pizza) meals keep on raising your BG for a lot longer than a low-carb, high fat and protein meals does. So you have to have a couple of different action plans depending on the composition of the meal. I haven’t got the pizza part down yet, because I square bolused for it tonight, and ended up going low. Gonna try plan Q (when plans A, B, C, etc. don’t work!) next time! :slight_smile:

Like Natalie said, it depends on what you want to know. If you don’t test until 2 or 3 hours after you eat, then you don’t know if a food is spiking you or not. There are a lot of foods that will spike you when you first eat but you are back in range a couple hours later. Those spikes are just as bad as staying high all the time. I test an hour after I eat. If I eat something I have never eaten before, then I test ½ hour after I eat. I also split my shots up so I do take insulin at 1 and 2 hours after I eat. I use Apidra and it is out of my system in 2 ½ hours.

As Natalie and Kelly said, it depends on the information you want. I never knew I peaked a bit more after an hour after eating almost anything because I tested at 2 hours and was fine if I’d calculated correctly. When I got my CGM I was truly surprised. So, I learned that I should pre-bolus by maybe 1/2 hour. I don’t always do so 'cause I’m staaaarving and can’t wait, in which case I try to eat more slowly. I also changed my insulin action from 4 to 3 hours.

I like to test every couple of hours, partially because my insulin pump suggests it, but it helps prevent suprises from lurking in your undergrowth and can provide a lot of data. I don’t mind stacking insulin very much as I only usually do it in the evening and the cure is at hand:

Why not “spread the joy” if the purpose is to size post-prandial peaks without necessarily deciding on a correction bolus?

The basic premise to my argument, is that there is no “magic number” which is how many hours afterwards to check. We want to be able to keep the bg in check no matter how many hours after a meal.

So if you abandon the “magic number” premise then you can test at a variety of different times after the meal and it doesn’t have to be the same magic number every time.

You ask "if insulin lasts for 3-4 hours what information will a [2 hour check] give? The answer, just as it is for all the different possible timings, is that it’ll give a range of numbers. Personally I have to worry about not just “peaks” but also “valleys” (going hypo) after a meal (when for example the insulin is kicking in too fast and the food might be absorbed on the slow side.) If I can keep the range limited so that I’m not going hypo after a meal, that is a hugely valuable tool for me. And some random samplings after a variety of different meals helps me get grok the effects that different foods have in terms of rate of carb absorption.

Having an idea of the rate of change is also useful. W/ a CGM, I will see 110–111—112 as a lot different than 110—119—130 and perceive it differently, dependiing on what else is going on. Doing a test like 15-20 minutes after test #1 can help with this. I still do this all the time in the morning to see what sort of DP/ hormone/ whatever stuff is floating around as it doesn’t seem to be particularly consistent and its useful to know the rate, if any, at which my BG is changing.

Advocating waiting a long time between tests helps health insurance companies save lots of money but I think there’s useful data in more frequent testing? If there were a methodological approach to how to use this data, I think that it would be great but the medical industry isn’t there yet, perhaps distracted by the constant quest for a cure?

I agree with Alan that a 2 hour correction need not lead to stacking. I was taught to test at 3 hours when I went to the Joslin DOIT program and that worked pretty well but when I started pumping I switched to testing at 2 and 4 hours which is what Walsh seemed to suggest in “Pumping Insulin”. I’ve discovered that I often have a rise in readings between the 2nd (or 3rd) and 4th hour and I usually wait until hour 4 to correct unless I’m clearly off course at hour 2.

If you don’t low carb, the correction wizard on many pumps will give suggest silly corrections at hour 2. When I first began pumping, I totally miscalculated the carbs at a dinner and tested 250 at two hours. My pump suggested that I test for ketones and suggested a correction of about .05 units. Sure. I corrected to 150 which worked pretty well although I needed a small snack 4 hours later. It was bedtime and I had fallen to the 60s.


Thank you, everyone, for your comments! They really helped to elucidate the issue.

Like everyone says, it depends on what you want to know. I find that I peak significantly at the 1 hour mark, and some foods definitely cause this to be worse. In addition, the action time for those analogue insulins really vary depending on the person. The 3-4 hour range is just an average; some folks actually get longer action time and some folks (like me) get less. I am closer to the 2-3 hour mark, especially if I’ve been really active on a given day.

So, I have some follow-up questions:

In Smart Pumping, Wolpert says:

Two hours after eating, you should expect your blood glucose to be higher than your pre-meal glucose since your food bolus is still at work. It's perfectly normal to see an increase of aboout 40-80 mg/dl. After 4 hours, your blood sugar should be at, or very close to, your target range.
(Wolpert, Howard. Smart Pumping. Alexandria, Virginia: American Diabetes Association, 2002, p. 53.)

When my ratios were working, my blood sugar 3 or 4 hours would be close to normal. But there would be an immense spike in the middle. Evidently, this is not good. How would I adjust my I:C ratio to avoid the spike so that I wouldn’t go low after 3-4 hours after the meal?

Also, how to people deal with the dawn phenomenon? It’s not always at the same time, not always the same rise in blood sugar, and I know action needs be taken before it happens to avoid it (as insulin takes some time to kick in)? Either I developed dawn phenomenon recently or never noticed it until I got a CGMS.


The reason for the 2-hour check is that is when many people spike, so that is the time when you can best catch the spike if it is happening (With testing you can see if this is, in fact, your own personal spike time, some people spike sooner or later). I think it is over-conservative to say you should never correct at two hours and leads to maintaining blood sugar at too high a level unnecessarily. If you are on the pump, your pump keeps track of the IOB to prevent stacking. When I was on shots I had knew my own duration of action and just divided to see how much IOB existed at the 2 hour mark so I could subtract it from my correction. That isn’t an exact science as it is not used evenly, but it works ok.

After testing at your “spike point” and correcting, if you start to see a pattern you might want to adjust your I:C ratio. But I would keep records for at least a couple weeks before doing this. If it is not a pattern, you don’t want to go and alter your I:C ratio if, for example, you only spiked occasionally or when you ate a certain food. Then, if you decide to alter the I:C ratio (by just one point at a time is my suggestion unless your results are wildly off the mark), you should alter it to the point where you are in target range at the two hour (or whenever) spike point more often than not. There is no reason you should drop low 3-4 hours later unless your starting point was low. In other words if you’re starting at say 65, you may rise to say 95 and then drop to 59. If I was 65 before a meal I would do one of several things: take glucose tablets and wait, bolus right before or even after starting eating, reduce the bolus a bit. If you find you are regularly dropping low several hours after a meal (and haven’t over-corrected a high from your meal) you might want to look to tweaking your basal rates for these times.

Hi - You manage dawn pheomenon by adjusting your basal a few hours before you wake up. Have you done basal testing yet? If you haven’t, you really need to do that first. Otherwise, all your pump settings will be hit and miss.

If you are spiking early you might try any of the following methods to improve matters:

  1. Go low carb. I don’t but if you have an interest Alan and others will be more than happy to give pointers.
  2. Add more fat and fiber to your diet to slow the rise in blood sugar.
  3. Prebolus and/or eat the high carb/high glycemic foods late in your meal so that your insulin has a chance to catch up.
  4. Schedule your exercise soon after meals. I actually wait 3 hours to exercise because if I don’t I tend to go low but if you’re running high the first couple hours exercise may help quite a bit.


You guys are simply amazing!

Because I have consistently had great trouble bringing my blood sugar into a normal range, I made a change to my sensitivity factor (which seems to work well) and my I:C ratio. It’s almost hour 1 after lunch and my levels have been steady. Which is good (as it tells me I needed to make some changes) but not so good because I was around 150 and have been hovering around there, meaning more changes may be needed.

What really frustrated me was that I worked out after breakfast, but the effect on my blood sugar was minimal. (Although I figure if I didn’t work out, my levels may have risen even more.) I made changes to my sensitivity factor, did a correctional bolus, and it brought me down although not as far as I would have liked. But I’ll work slowly to make the necessary changes. I’m still monitoring my blood sugar to see what will happen as time goes by - will my blood sugar rise? go lower? plummet? Then I’ll figure out on how to tweak it so that it will deal with the food and bring it down.

That said, I don’t think if I could do this stuff without a CGMS. Or, I could, but it would be very painful.

You don’t say, but if you started out before lunch at 150 you should have included your correction along with your carb bolus. Otherwise there is nowhere to go but up! (Sorry if you already knew this).

LOL! Problem is that there was a correctional along with the bolus for carbs. Which is why my current trend isn’t all good news. But it means my settings are closer to what I should be at than before. However, that said, part of the correctional was deducted because of IOB (I use the OmniPod, if that makes a difference) and I’m thinking that keeping it at 4 hours duration for insulin may not be accurate.
By the way, I can’t begin to express my appreciation for y’all. You’ve made an extremely frustrating situation so much more bearable and tolerable. As long as I know there are things I can do (and what those things are), my mind will be at ease.
I love TuDiabetes!

If you are frequently high before meals, I agree with Maurie, you should see if your basals are set right before you work too hard on the I:C ratios. We all start our meals high sometimes and need to correct, but if you are regularly high between meals the basals are too low at least for some of your “time zones”.

Yeah, I’d give up my doctor before I’d give up TuD! (If it weren’t for those pesky prescriptions - when I lived in Guatemala I didn’t even need doctors for that!)

I agree with the others – take a look at your basal rates. If they are not correct, nothing will work. Do you know how to do basal testing? Your duration might be wrong also. They have a tendency to start everyone off at 4 hours but 4 hours doesn’t work for everyone.

Even if you don’t want to go low or lower carb, take a look at the foods you are eating. Maybe at least change some to lower glycemic foods.

When I was on Humalog, I always ended up low a couple hours after my meal. I know it wasn’t basals because I did basal testing and if I changed my I:C ratio and took less for my meal, I ended up high. I know a few other people here have said the same thing and just planned on having a small snack at the point they would go low. I know a lot of people don’t approve of that, but it was the only way to keep my BS in range so it worked for me.

I am putting links for 2 groups here. The first is the Flatliners so you can see there are people that do manage to keep their BS in range. It is not a perfect flatline, but just have to stay within the lines of your CGMS. The other is for TAGgers. TAG is total available glucose. Some of us bolus for protein & fat and that helps keep us in range also.

Like Natalie, I too peak at 1 1/2 hours (with Humalog), and it levels off at around 2-2 1/2 hours. If I am trending much higher than my target and I can see I won’t have enough insulin, I correct at the 1 1/2 hour point. If I am only one-three points higher, I leave it alone and check again at the two hour mark.