Monday, July 24, 2017

When to Measure Postmeal Glucose Levels

What's the best time to test your blood glucose (BG) after meals?

Of course that depends on what you want to learn from testing after meals. If you want to know the after-meal peak, then you should test about an hour after you eat. The peak will differ a bit with different foods; fats slow down gastric emptying, and liquids pass through faster than solids. It also can differ with different people.

So if you want to know where your own peaks are, you should test every 15 minutes or so starting about 45 minutes after eating and continuing until the numbers start coming down.

If you want to know if you're able to return to normal, or close to normal, BG levels within a few hours, then you should test at 2 or 3 hours.

"After eating" is also ambiguous. Should you test X minutes after you start to eat or after you stop eating? Again, it depends on your habits. If you wolf your meal down, it doesn't much matter. If you eat leisurely and take 30 minuts to finish a meal, then it does. What you really want to know is differences between different meals, so the important thing is to test about the same way every time. Don't compare one meal you ate in 4 minutes with another one that you took an hour to finish. Most people measure the time after starting to eat.

If you ask CDEs or your doctor when to test, they'll usually tell you to test at 2 hours. This is because most research papers about postprandial (after meal) BG numbers use the 2-hour reading. But this may not be the best.

One researcher, Antonio Ceriello, recently published a paper proposing that it's time to switch to a one-hour postprandial measurement. He kindly sent me a copy of the full text of the paper.

Ceriello says that there's evidence that the one-hour measurement has even stronger power than the two-hour measurement for identifying impaired glucose tolerance. He said this number is also related to the risk for cardiovascular complications. In vitro experiments have shown that just one hour of high BG levels is enough to cause endothelial dysfunction that can then lead to coronary vascular disease, as well as reactive oxygen species (strong oxidants), he said.

Endothelial function is worse at one hour than at two hours both during oral glucose tolerance tests and after meals, he said.

If all this isn't enough, Ceriello said short-term high BG levels can impair beta-cell function.

When I've measured BG levels in nondiabetic friends and relatives, I've sometimes found one-hour readings of 160 or so, but the numbers come down to baseline by two hours. Testing only at two hours wouldn't identify these people, who might be at risk of developing diabetes in the future.

Ceriello's recommendations apply to clinical studies, but there's no reason you couldn't measure at one hour if you so chose. The best, of course, would be to measure at both one hour and two hours if you could afford enough strips. Then you'd know which time made most sense for you.

Even if most clinicians accept Ceriello's recommendations, it will take time before they become standard. So for now most studies will continue to use the two-hour numbers. But when  you see a study mentioning postprandial (or postmeal) numbers in the future, you should check to see what they mean, if you can (abstracts may not specify).

If anyone wants references to the studies Ceriello was citing, let me know and I'll send them to you. If I get tons of requests, I'll just edit this post to put the links in.


  1. Oh this is GOOD!

    In my experience most but not all forms of Type 2 start with a 1 hour spike - reduced Phase 1 insulin - but some show up first at 2 hours - probably when the pancreas is going down in a different way and Phase 2 is lost first.

    For example my late mother spiked nearer to 2 hours - but then her diabetes came from taking steroids while being well over 90.

    chris c

  2. I have a friend who was in some study and asked the researcher how many people with type 2 have reduced Phase 1 insulin response, and he said, "100%."

    Time of spike depends on what you eat as well as individual physiology and presence or absence of gastroparesis.

    1. Yes and I'll bet most "prediabetics" will also show a reduced Phase 1 insulin. Probably a lot of "nondiabetics" who are actually just "not diabetic yet" also. IMO the only thing that would be more predictive of health consequences would be insulin itself, or c-peptide.

      Now how can we get this information out to actual doctors???

      chris c

    2. Chris, I agree. It's possible that those who end up with type 2 have had a reduced phase I most of their lives and these spikes are what wear out the beta cell.

      I think measuring insulin is expensive, and it has a very short half-life, which is why they developed the c-peptide test. But unless someone is at very high risk (type 2 runs in the family and the patient has a weight problem), most doctors probably wouldn't prescribe.

      I asked my endo if she'd ever measured her own A1c, and she said no because it was expensive, which I found odd.

  3. Oh they DO come up with some bizarre stuff! Mostly here in the UK c-peptide is only used on Type 1s after they have been diagnosed, not on Type 2s at all although it would be useful in unpicking the increasing numbers of adult-onset Type 1/LADA who are routinely misdiagnosed as Type 2.

    I was told by the nurse that I could NOT have an A1c "because I was not diagnosed diabetic (yet)" and also not to test. A few years later and it has become the ONLY diagnostic test used. Maybe someone designed a cheaper lab analyser?

    Likewise the nurse told me "We don't measure your "cholesterol" again once you are on your statin" and in many places trigs are no longer given. Cheap lab equipment again? I had to see the doctor to get mine done and her lab actually used to use *measured* LDL, the current lab uses calculated - but to calculate it you need to measure HDL and trigs . . .

    chris c