Published August 16, 2024
Note: This blog post mostly discusses data measured by a continuous glucose monitor (CGM). CGMs measure the glucose concentration in the interstitial fluid, i.e., the fluid between our tissue cells. This metric correlates strongly with blood glucose concentrations, and we are therefore using the terms ‘interstitial glucose’, ‘blood glucose’, ‘glucose levels’, and ‘sugar levels’ as synonimous.
This blog post provides all blood or interstitial glucose concentrations in mg/dL. To convert from mg/dL to mmol/L, divide these by 18.
In this blog post, I will help you interpret your continuous glucose monitor (CGM) data by discussing what the scientific evidence tells us about normal vs. abnormal blood glucose responses after a meal. CGMs can be an amazing tool, particularly if we use them not just to lower our acute blood sugar responses. That’s pretty easy; we’d just have to eat fewer carbs. No, where CGMs really offer an opportunity that I feel most people are not taking advantage of is to really understand our own glucose tolerance, that is, the ability of our body to keep blood sugar in the normal range, no matter what we eat.
So with this blog post, I’ll help you use your CGM data to understand to which degree you may be glucose intolerant. We will also discuss how to tell the difference between a normal blood sugar response to a meal and a concerning spike, and I’ll evaluate the very popular claim that any increase in blood glucose by 30 mg/dL or more is abnormal and dangerous, in other words, that such an increase by 30 mg/dL or more constitutes a blood sugar spike that we should try to avoid.
Diabetes and Prediabetes
If we discuss normal vs. abnormal blood glucose, we must mention diabetes and prediabetes first, of course, because these are the two conditions of glucose intolerance that are clearly defined clinically.
In most countries, diabetes is diagnosed if someone has fasting blood glucose of 126 mg/dL or higher, HbA1c of 6.5% or higher, or blood glucose concentrations exceeding 200 mg/dL exactly two hours after drinking a beverage containing 75g of pure glucose as part of an oral glucose tolerance test, or OGTT. Please see the table below that summarizes the diagnostic criteria by the American Diabetes Association.
Prediabetes is diagnosed if someone doesn’t meet the criteria for diabetes, but their fasting glucose level is between 100 and 125 mg/dL, HbA1c is 5.7 to 6.4%, or the blood glucose concentration at two hours in an OGTT is between 140 and 199 mg/dL.
If you have been diagnosed by a doctor to have either diabetes or prediabetes, by these criteria, this means that you are definitely glucose intolerant to some degree, and have abnormal blood glucose levels.
If you have not been told that you have diabetes or prediabetes, you may still have one or the other, because the number of undiagnosed people with these conditions is thought to be very high in most countries. In that case, you can use your CGM data to get a sense of whether you should see your doctor for some more testing. Note, however, that you cannot use your CGM to diagnose any condition, because the diagnostic criteria are all based on standardized lab testing. But you can get a pretty solid idea of whether your blood sugar levels are normal or not, and whether any follow-up with a professional is warranted.
For one, take a look at your fasting glucose levels in the morning. Are these always higher than 100 mg/dL, or even higher than 126 mg/dL? These are the thresholds for prediabetes and diabetes, respectively. Note that quite often, our glucose levels rise a little bit just as we wake up and start our day. This is a normal phenomenon called the dawn effect that is triggered by the levels of the hormone cortisol going up in the morning. It’s quite mild in most people, but people with diabetes can have a very pronounced dawn effect. Look at your baseline glucose levels before they start to rise; if that number is consistently higher than 100 mg/dL, I suggest you discuss this with your doctor.
Then take a look at your typical glucose levels exactly two hours after you STARTED eating a meal. Are these still elevated above your normal baseline level? As you can see in the table above, elevated 2-hour glucose levels are a diagnostic criterion for prediabetes and diabetes, but that is in the context of a 75-g OGTT. This test involves drinking a beverage containing 75 of pure glucose, and this much glucose by itself, without any other food, is a much more substantial load to handle than an average meal, even a high-carb meal, so – again – we cannot use our normal blood sugar response to our typical meals to diagnose prediabetes or diabetes. However, if you regularly have levels higher than 140 mg/dL around two hours after starting a meal, I would talk to a doctor and discuss whether some more testing may be indicated. If we are fully glucose tolerant, our glucose levels should be back to our baseline level under 100 mg/dL, or close to it, by the two-hour mark, unless this was a particularly large or extended feast with several courses.
So if your glucose levels after your typical meals look like the graph on the figure above, still quite a bit elevated above 100 mg/dL one and a half and even two hours after starting a meal, this may be indicative of glucose intolerance, and in a case like this, the cause is often insulin resistance. Here are some blog posts explaining what insulin resistance is, and how it’s related to glucose intolerance; how you can figure out if you are insulin resistant; and what common causes of insulin resistance are.
Now, it’s very important to be clear that our glucose responses to our meals can only shed light on whether or not we are glucose tolerant or intolerant IF these meals do contain at least some carbohydrates. If you eat a low-carb meal of just meat and non-starchy vegetables, your blood glucose will not rise meaningfully, even if you are glucose intolerant. There is nothing wrong with eating low-carb or even very low-carb meals, but just be sure not to fool yourself into thinking that your lack of a blood sugar response to these meals means that your glucose tolerance is great. So again, if you want to use your CGM to understand what your glucose tolerance is like, you can do that only if your meals contain at least some carbs, for example in the form of grains, starchy vegetables, or fruit.
It is also important to understand that following a low-carb diet in general will make you somewhat glucose intolerant temporarily because your body will release less insulin on an occasional cheat meal rich in carbs. What happens here is that the amount of pre-formed insulin that is stored in the pancreatic beta-cells is gradually reduced if carbs are not regularly consumed. So if you have been following a low-carb diet, and would like to get a sense of your glucose tolerance, introduce small amounts of carbs slowly over several weeks, and then start eating a normal diet with higher carbohydrate content again. And only then use a CGM to measure the glucose responses to these higher-carb meals.
OK, so we have discussed how we can use our fasting glucose levels, and the glucose levels two hours after we started eating a meal, to get a sense of whether we are glucose tolerant or intolerant. What about that time between our fasting state and the time point two hours after a meal, though? That’s actually the most interesting time, because that’s when the blood sugar level rises to some peak, and then falls back, ideally to baseline. Is there anything that could happen during these two hours immediately following a meal that could alert us that maybe our blood sugar regulation is not perfect? And related to this is the question, is there such a thing as a blood sugar spike, meaning, a blood sugar response to a meal that is too high, even if our fasting and two-hour levels are just fine?
Let’s start this discussion by taking a look at a very common claim on social media.
Is an Increase in Blood Glucose after a meal by 30 mg/dL an Abnormal and Dangerous Spike?
There are several popular influencers and at least one company that markets CGMs who claim that any increase in blood glucose levels by 30 mg/dL or more is an abnormal and dangerous blood sugar spike. What this claim suggests is that if someone has a fasting glucose of around 80 mg/dL and their blood glucose rises to, say, 120 mg/dL after a meal (a 40 mg/dL rise), that this is abnormal and also leads to negative health consequences compared to someone in whom blood glucose doesn’t ever go higher than 30 mg/dL above the baseline. In this case, that would be an increase to no more than 110 mg/dL.
So let’s examine this claim. Is an increase in blood sugar by 30 mg/dL indeed abnormal, meaning, it is not part of normal physiology? And is it dangerous, i.e., do we have evidence to suggest that such a blood sugar spike is harmful to our health in some way?
First, …
Is an Increase in Blood Sugar After a Meal by 30 mg/dL Abnormal?
Well, let’s just review the blood sugar data of normal, healthy people as they live their normal lives. I have reviewed a number of studies in which researchers equipped such healthy people with a CGM. And as I was reviewing these papers, I could immediately see where that 30 mg/dL figure probably came from. But: I do think this is a misinterpretation of these data.
Two studies (from the JDRF CGM Study Group and Rodriguez-Segade et al.) have reported that whenever blood glucose levels rise throughout the day, the average rise is about 30 mg/dL. And I guess that has been interpreted as suggesting that blood sugar levels don’t rise by more than 30 mg/dL in healthy people. This is an incorrect interpretation.
For one, if the average rise in blood sugar is 30 mg/dL, that means that a good portion of these healthy people with normal glucose tolerance experience an increase by 35, 40, or 45 mg/dL regularly. Also, and this is an important point, the 30 mg/dL figure relates to any rise in blood sugar; scientists call this metric MAGE, the mean amplitude of glycemic excursions. Now, blood sugar can rise for any number of reasons, such as after we have eaten a little snack, experienced a bit of stress, or exercised. Or it could be in response to a major meal. The 30 mg/dL figure is the average increase in blood sugar whenever there is any increase beyond a certain threshold. In other words, to calculate MAGE, researchers don’t even pay attention to whether a meal was eaten or not, they just look at the CGM data, and any time the glucose concentration rises by more than, say, 10 mg/dL above the baseline, they call this a glycemic excursion. And then they average how much the glucose levels rise in each of these glycemic excursions. That 30 mg/dL figure therefore contains a lot of occasions where the sugar level rose even if no meal or even snack was consumed. It is, therefore, very likely that the average response to a normal mixed meal is substantially higher than that 30 mg/dL figure.
And indeed, one group of researchers looked into exactly that. Freckman and colleagues investigated the blood sugar responses to normal breakfasts, lunches, and dinners. These investigators enrolled 12 women and 12 men who were young and who had a normal body weight and normal glucose tolerance. The investigators ran OGTTs before enrolling participants and excluded anyone with any sign of prediabetes or diabetes. So these were 24 young, lean, and healthy people who are probably among the more glucose-tolerant people on the planet. And they should be able to provide us with some insight into what the normal blood sugar response to full meals looks like. Take a look at the table below. Prior to each of the main meals, their blood sugar levels were around 80 – 82 mg/dL, and the average increase was between 39 and 50 mg/dL. In some people less than that, but in others, more than that. Some people had peak glucose around 160 mg/dL. Maybe they were not starting at 80 mg/dL, but around 95 mg/dL (because, by definition, their fasting glucose must have been under 100 mg/dL), but that still means they had an increase in glucose well in excess of 60 mg/dL.
So, I ask you: if the average blood glucose increase in such a young, normal-weight and healthy population is 40 to 50 mg/dL, with some people showing an increase by 60 or more mg/dL, how can anyone claim that an increase by 30 mg/dL is abnormal and not part of normal physiology? I cannot explain it, and absolutely do not understand where this argument comes from. I would even go further and call this dangerous fearmongering.
To explain this further, if an influencer explained that he/she feels better if they avoid their blood sugar from rising more than 30 mg/dL, or if they explained some other anecdotal benefit, I would not object to this at all. However, it is commonly claimed that this definition of a blood sugar spike is based on scientific evidence, which it clearly is not. In fact, the cumulative scientific evidence directly undermines the claim that an increase by 30 mg/dL or more is an unnatural spike.
That brings us to address the second part of the claim, namely …
Is an Increase in Blood Glucose by 30 mg/dL or More Dangerous?
Right, so maybe blood sugar does normally rise by 40, 50, or even 60 mg/dL even in the most glucose-tolerant people, but maybe that is not desirable and even dangerous for our long-term health?
Well, here’s where I can make it even quicker: I have worked in clinical nutrition and diabetes research for more than twenty years, and I know of no paper that shows or at least suggests that an increase in blood glucose by, say, 40 or 50 mg/dL is associated with short-term or long-term health risks, compared to an increase by 30 mg/dL. Sorry, I can’t show you any data here; there just are none that I am aware of.
Some people may argue that it’s clear that lower blood sugar levels are always better. For example, they may argue that having HbA1c levels at the lower end of normal is associated with a lower risk of chronic disease than having HbA1c levels at the high end of normal. And there is indeed some evidence to suggest that the risk of cardiovascular disease, for example, increases with HbA1c even within the normal range. Now, HbA1c is a measure of our average blood sugar response over roughly the last 3 months, so the obvious assumption would be that lower blood glucose responses to our typical meals should lead to lower HbA1c levels. However, I think it’s important to understand that you do not have to keep your blood sugar responses to your meal to under 30 mg/dL to have a low HbA1c. There is nothing magical about that 30 mg/dL figure. I, for example, regularly have blood sugar increases by 40 or even 50 mg/dL over my baseline after a high-carb meal, and my most recent HbA1c is 4.8%. That’s quite low. Similarly, HbA1c was on the low end of the normal range in the study we discussed earlier by Freckmann and colleagues, which showed average increases in blood sugar by 40-50 mg/dL after each meal. That is because if you have a low baseline glucose level, and your blood glucose concentrations only rise briefly and don’t stay elevated chronically, your average blood sugar level could still be quite low. And that would be reflected in a low HbA1c. The idea that you need a flat blood glucose curve to achieve a low HbA1c is simply not correct. You need to have good glucose tolerance to have a low HbA1c, and that way, no matter what you eat, your average glucose, as reflected by your HbA1c, will always be on the low end.
In conclusion, the claim that increases in blood sugar levels after a meal by more than 30 mg/dL are abnormal is factually incorrect. In fact, increases by 40-50 mg/dL and even 60 mg/dL occur regularly, even in the most glucose-tolerant individuals. Regarding the claim that increases by more than 30 mg/dL are dangerous or linked to adverse health outcomes, I am not aware of even a single paper supporting that claim.
For these reasons, I suggest you ignore this claim that an increase by 30 mg/dL is a spike that should be avoided. And I say this as someone who strongly believes that blood sugar concentrations and related biomarkers, such as our insulin sensitivity, are among the most important chronic disease risk factors. But I also feel that it is not helpful to make even fully glucose-tolerant people think that they have a problem when in fact, their blood sugar regulation is just fine.
OK, so if small bumps in blood sugar are not abnormal and dangerous, what kind of glucose response to a meal is abnormal and associated with an increased risk of chronic disease?
How Do I Define Blood Sugar Spikes?
Well, one thing we need to be clear about is that we currently don’t have a clinical definition of blood sugar spikes. In other words, you could not be diagnosed as having blood sugar spikes.
We also need to be very clear that the thresholds used to diagnose diabetes and prediabetes only apply to the fasting state and the 2-hour time point in an OGTT. These thresholds do NOT apply to any random time point after a meal. In other words, if your blood glucose concentration rises above 140 mg/dL about one hour after a meal, this does not mean that you have prediabetes.
So, then, how do we interpret our glucose levels in that 2-hour window following a meal?
I mentioned earlier that several research teams have studied the normal blood sugar levels in people with normal glucose tolerance, as determined by standardized clinical testing. I have summarized all of the data from people who did not have prediabetes or diabetes in the table below.
One thing we can see quite clearly is that in all of these study populations, sugar levels are mostly in the 70-140 mg/dL range, 93-99% of the time. But in all populations, sugar levels do spend some time in the 140-180 mg/dL range. That percentage varies from study to study, between 0.4 and 4% of the time. This means that, on average, people with normal glucose tolerance do occasionally have sugar levels in the 140-180 mg/dL range, but only briefly, for a total of up to an hour a day or so.
Now, here is something interesting, though: in three of the studies (JDRF CGM Study Group, Freckmann and colleagues, and Shah and colleagues), glucose levels never, not in a single participant, reached 180 mg/dL or more. In the other three studies in which this metric has been reported (Mazze et al., Rodriguez-Segade et al., Chan et al.), some participants had sugar levels exceeding 180 mg/dL occasionally. In the paper by Chan and colleagues, the authors studied obese youth: so these were young participants, but they carried excess body fat mass, a major risk factor for insulin resistance and glucose intolerance. And they described explicitly that some of the obese youth occasionally had excessively high sugar levels reaching as high as 239 mg/dL. That is definitely not normal. The other two studies that reported occasional sugar levels exceeding 180 mg/dL had enrolled older individuals, and again, older age is a major risk factor for insulin resistance and glucose intolerance. This suggests to me that those few individuals in these studies who had sugar levels exceeding 180 mg/dL occasionally were already slightly glucose intolerant, possibly due to insulin resistance related to higher body fat mass or higher age.
Therefore, I interpret these data to cumulatively suggest that in fully glucose-tolerant populations, ideal glucose levels are mostly in the 70-140 mg/dL range, with very rare and only brief excursions into the 140-180 mg/dL range, whereas levels of 180 mg/dL or higher are never seen. That is what I define as normal, and what you should aim for.
This also means that I would interpret sugar levels reaching 180 mg/dL or higher as an early sign of glucose intolerance, which we should probably avoid.
And that is why I define any increase to 180 mg/dL as a blood sugar spike.
There is another rationale for considering an increase to 180 mg/dL or higher a spike, and that is that we have evidence suggesting that regular increases in blood sugar to 180 mg/dL or higher are associated with an increased risk of chronic disease. So let’s discuss this next.
Do We Have Evidence that Blood Sugar Spikes to 180 mg/dL or Higher are a Health Risk?
We know for sure that having chronically elevated blood sugar levels, as in diabetes, is associated with substantially increased risks of a number of chronic diseases, such as cardiovascular disease and chronic kidney disease. But what about people who mostly have normal glucose levels, but spike to 180 mg/dL or higher regularly? Well, it would be ideal if we had data from tens of thousands of people who had worn a CGM for an extended period of time so that we could actually see whether those with a spiking pattern have an increased risk of disease. Unfortunately, we don’t have any such data.
However, we have some indirect evidence based on a biomarker of blood sugar spikes called 1,5-anhydro-glucitol, or 1,5-AG. I mentioned this biomarker briefly in a prior blog post, but I guess most of you have not heard about it, so let me briefly explain what this is about.
In short, 1,5-AG is a sugar that all of us have in our blood. And under normal circumstances, the blood concentrations of 1,5-AG are pretty stable. But, as soon as our blood glucose levels exceed about 180 mg/dL (the renal threshold of glucose), we start to excrete some glucose in the urine, and 1,5-AG is excreted along with that glucose. This means that the more our blood glucose levels rise above 180 mg/dL, and the longer they remain elevated above 180 mg/dL, the lower our 1,5-AG levels in blood will be.
This means that 1,5-AG levels in the blood of someone who has diabetes and who constantly has blood sugar levels over 180 mg/dL will be very, very low. But it also means that for anyone who does not have diabetes or even pre-diabetes, blood 1,5-AG levels will be reduced if they experience frequent blood sugar spikes to 180 mg/dL or more.
This is of interest because lower blood levels of 1,5-AG are consistently and strongly associated with an increased risk of cardiovascular disease, chronic kidney disease, and cancer mortality, and that is even in people without diabetes or prediabetes, and even when adjusted for traditional markers of blood sugar regulation, such as HbA1c or fasting plasma glucose.
Now, even though the literature on 1,5-AG is compelling, we need to consider that 1,5-AG is an indirect biomarker of blood sugar spikes. It is certainly possible that other conditions also lower 1,5-AG concentrations in blood, and maybe these conditions predispose to chronic disease. For example, if an early, pre-clinical stage of kidney disease were to affect the re-absorption capacity for 1,5-AG, then that by itself may explain at least some of the observed associations between low blood 1,5-AG levels and chronic disease risk. Therefore, we must be clear that the evidence linking blood sugar spikes, as I define them here, and chronic disease risk is certainly not conclusive at this point.
Reactive Hypoglycemia
A CGM can also be useful to identify regular episodes of low blood sugar. Hypoglycemia is defined as a blood sugar level below 70 mg/dL. Often, particularly if concentrations drop below 60 mg/dL, people do feel the body responding to these low blood sugar levels. Common symptoms include feeling hungry, particularly for something sweet, feeling dizzy, anxious, tired or weak, sweating, shaking, or heart palpitations.
A CGM can be a great tool to detect if you frequently suffer from hypoglycemia. If you observe your glucose levels dropping below 70 mg/dL regularly, I suggest you discuss this with a doctor. Hypoglycemia can be dangerous, and if you experience this regularly, you should not take it lightly.
One word of caution, however: if you lie on your CGM sensor at night, it can show very low glucose levels, not because you actually have hypoglycemia, but because of the pressure on your sensor. And, CGM sensors can also be faulty and show incorrectly low glucose levels. It may be prudent to have a hand-held glucose meter available, the one that requires finger pricking to get a drop of blood, so that you can check your actual blood glucose concentrations if you ever have any weird or extreme glucose readings on your CGM.
Now, hypoglycemia can develop for many reasons, but a common reason is in response to a blood sugar spike. One thing that can happen is that someone eats a high-carb meal, and that leads to a rapid increase in blood sugar levels. Then the body senses that blood sugar levels are too high, and secretes a lot of insulin, which then drives glucose levels down too fast, and too much. If you observe this pattern, as illustrated in the figure below, try to take steps to avoid the blood sugar spike from happening in the first place; that will dramatically reduce your risk of reactive hypoglycemia. A good place to start would be this blog post on how to avoid blood sugar spikes, and this one on reactive hypoglycemia.
Summary
In summary, taking these considerations together, I conclude that any increase to 180 mg/dL or higher is probably a first sign of mild glucose intolerance because it doesn’t normally occur in the most glucose-tolerant populations, ever, no matter what they eat. And such an increase to 180 mg/dL or higher is probably associated with an increased risk of several chronic diseases, at least if it occurs regularly.
That’s why I define an increase to 180 mg/dL as a spike that we want to avoid.
Further, we should keep our blood sugar in the 70-140 mg/dL range most of the time, with only occasional and brief excursions into the 140-180 mg/dL range. If you spend more than an hour or so per day in that 140-180 mg/dL range, that is likely also a sign of early-stage, preclinical glucose intolerance.
My strong suggestion is that if you have diabetes or prediabetes, you follow suggestions by your doctor to control your blood sugar levels. I also recommend that you try to identify the root cause or causes of your glucose intolerance, such that you can address these so that you can prevent your condition from getting worse and, ideally, even reverse it. To really understand glucose tolerance and the factors determining glucose tolerance, I recommend the blog post on the regulation of blood sugar. Next, it can be helpful to assess whether you suffer from insulin resistance to some degree, and then to look into possible root causes of that insulin resistance.
Similarly, if you do not have diabetes or prediabetes, but show early signs of glucose intolerance, try to identify and address the root cause or causes of your glucose intolerance to prevent it from progressing to prediabetes or manifest diabetes. Check the blog posts linked in the prior paragraph. Signs of such early-stage glucose intolerance that are apparent from your CGM data would be regular blood sugar spikes to 180 mg/dL or higher, if your sugar level is in the 140-180 mg/dL range for extended periods of time after meals such that it adds up to more than an hour per day, or if your sugar levels remain substantially elevated above your baseline two hours after the start of a meal.
I have summarized how I would identify early signs of mild glucose intolerance that has not yet progressed to prediabetes or diabetes in this figure below.
To me, this is actually the more impactful use of a CGM: to identify any issues in your glucose tolerance early so that you can address them, not to totally flatten an already normal glucose curve.
If this blog post was helpful to you, I’d appreciate it if you shared this blog post with anyone you know who you think may be interested.
Take care.
References
- Juvenile Diabetes Research Foundation Continuous Glucose Monitoring Study Group. Variation of interstitial glucose measurements assessed by continuous glucose monitors in healthy, nondiabetic individuals. Diabetes Care 2010; 33: 1297-9.
- Rodriguez-Segade et al.; Continuous glucose monitoring is more sensitive than HbA1c and fasting glucose in detecting dysglycemia in a Spanish population without diabetes. Diabetes Research and Clinical Practice 2018; 142: 100-9.
- Freckmann et al.; Continuous glucose profiles in healthy subjects under everyday life conditions and after different meals. Journal of Diabetes Science and Technology 2007; 1: 695-703.
- Hoffmann and Honigsberg. Glycated hemoglobin as an integrator of cardiovascular risk in individuals without diabetes: lessons from recent epidemiological studies. Current Atherosclerosis Reports 2022; 24: 435-42.
- Shah et al.; Continuous glucose monitoring profiles in healthy nondiabetic participants: a multicenter prospective study. Journal of Clinical Endocrinology and Metabolism 2019; 104: 4356-64.
- Mazze et al.; Characterizing glucose exposure for individuals with normal glucose tolerance using continuous glucose monitoring and ambulatory glucose profile analysis. Diabetes Technology & Therapeutics 2008; 10: 149-159.
- Chan et al.; Continuous glucose monitoring and its relationship to hemoglobin A1c and oral glucose tolerance testing in obese and prediabetic youth. Journal of Clinical Endocrinology and Metabolism 2015; 100: 902-10.
- Zhou et al.; Reference values for continuous glucose monitoring in Chinese subjects. Diabetes Care 2009; 32: 1188-93.
- Dungan et al.; 1,5-anhydroglucitol and postprandial hyperglycemia as measured by continuous glucose monitoring system in moderately controlled patients with diabetes. Diabetes Care 2006; 29: 1214-9.
- Ikeda and Hiroi. Cardiovascular disease and 1,5-anhydro-d-glucitol. Global Health Medicine 2019; 1: 83-7.
- Rebholz et al.; Serum levels of 1,5-anhydroglucitol and risk of end-stage renal disease. American Journal of Epidemiology 2017; 186: 952-60.
- Kira et al.; Association between a biomarker of glucose spikes, 1,5-anhydroglucitol, and cancer mortality. BMJ Open Diabetes Research and Care 2020; 8: e001607.
21 Responses
Thank you so much for this very informative article. I recently purchased a CGM even though not diagnosed pre diabetic I do have a family history of diabetes and at 72 years old and slightly overweight I am keen to keep healthy. I appear to be in the normal range. I usually eat low carb-ish but can have a binge on high carb. My glucose response to these binges appears to be okay… returns to normal within 2hours.
Really appreciate the informative article!
I have one question. In “Do We Have Evidence that Blood Sugar Spikes to 180 mg/dL or Higher are a Health Risk?”, you say there’s indirect evidence. One part you mention is that low 1,5-AG is associated with an increased risk of various diseases. However, the cited papers seem all use subjects with abnormally low 1,5-AG. Based on my understanding, a healthy person whose blood sugar spikes to 180 mg/dL doesn’t normally have such a low 1,5-AG concentration. Am I missing something?
Bob,
I think it’s important to recognize that 1,5-AG is not a well-standardized lab test, and that
Bob,
Thanks for commenting. Some papers report detailed data on the distribution of 1,5-AG among people with and without diabetes. I have linked one below, see Figure 1. So, there clearly is some overlap, and the data on 1,5-AG in relation to CGM data would tell us that those with more frequent spikes to 180 mg/dL or higher would be among those with 1,5-AG in the lower third or quarter. And the epidemiological data tell us that such lower 1,5-AG concentrations are linked to an increased risk of CVD, CKD, and cancer, even among patients without diabetes and even when adjusted for fasting glucose and HbA1c.
Your point is still well taken that it remains somewhat unclear whether those participants without diabetes/prediabetes and very low 1,5-AG levels of <10 ug/mL have such low levels indeed because of frequent spikes, or whether something else may be going on these (early stage kidney disease, for example). And whatever that may be could certainly be confounding the relationship between 1,5-AG and disease outcomes. That's why I try to be cautious in my statement on this and typically conclude that these data 'suggest' a link between frequent blood sugar spikes and an increased risk of dieases.
Best,
Mario
https://pubmed.ncbi.nlm.nih.gov/28535187/
Can you please comment on why a persons CGM and meter would show an 89 average glucose level but everytime you have your A!C tested it is 5.7. According to a chart an 89 average glucose corresponds with an A1C of around 4.8. Why are these way off and I don’t have any of the issues that cause your A1C to be off like being anemic. I have never seen anyone address this issue. Also which of these is the most accurate. Thanks
Hi Victoria,
Have you seen this with several different CGM sensors, or just with one? Please note that any CGM sensor can be bad, and read consistently too low or too high. It may be worthwhile to confirm your CGM readings every once in a while against a finger prick-glucose monitor measurement. Note, however, that there is always a 15 min delay between changes in blood glucose and interstitial fluid glucose (which is what the CGM measures), so in the 2 hours after a meal, there could be some expected deviation between the readings.
Other than that, my best guess would be that you have genetically rather longer-lived red blood cells!? Hba1c is an estimate based on averages, and it is clear that some people will deviate substantially in either direction. These people with more substantial deviations in the lifespan of their red blood cells that would affect the reliability of HbA1c to estimate average glucose would be expected to be around 5% of the population, so not all that rare.
Cheers
Mario
Thanks! Super interesting.
I was wondering about the arbitrariness of the cutoff lines, speaking from personal experience: I’m in my early 40s and consider myself rather healthy. Work out more than daily, eat well, good weight, not smoking etc. I got a CGM as a little experiment after my kid was diagnosed with T1D and I was curious to see my own BG curves.
I was surprised to see that after carb-rich meals I easily spike to e.g. 8.3 mmol/L, which scared me a bit. At the same time, the shape of the peaks is super spikey: they immediately collapse and I don’t stay over 7.8 for more than maybe 15-20 min, sometimes dropping all the way back to fasting (<5.6) within 30 min.
How would one read this? Are the heighest numbers more important, indicating an issue? Or the clearance rate (peak shape/width)? Could it be that the baseline values and hence peak heights vary among individuals and are less improtant?
The fact that the glucose levels come down rapidly suggest that you are very insulin sensitive. That’s good.
Also, your glucose levels don’t go all that high, so I would discourage you from thinking about such increases as a ‘spike’. Seems to me that your response is fully within what we would call normal glucose tolerance.
If you are curious about the very rapid increase in glucose, particularly if this ever rises above 10 mmol/L, this could be due to a slightly diminished first-phase insulin response. Do you observe this in response to a high-carb meal where the prior meal(s) were lower in carbs? If so, then that is very common, and a result of eating more carbs that your body has had to deal with in recent meals.
Thanks Mario for your quick and insightful reply.
I need some more days to experiment but I feel that it’s rather regular for a carb-rich meal. My breakfast is not that heavy but include fresh fruit (maybe 20g of carbs), and doesn’t generate any peak. But it is also usually preceded by a workout. My post-lunch peak sounds like the first-phase response; but I do get them after dinner as well, although I would normally have dinner after a 5-6h of fasting and a mild workout. Do you know if this can also lead to a kind of a first-phase response?
What do you mean with ‘first-phase response’? When we talk about ‘first-phase response’, we mean the first-phase insulin response. That plays a major role in preventing an early rise in blood glucose, and therefore, a diminished first-phase insulin response can lead to a rapid rise in blood glucose can be an indication.
Thanks Mario. In the meantime I had the time to realize this is not a single blog post and read a lot of your other pieces. Super interesting, and I also understand “first-phase response” a bit better. In my case I don’t think I see much difference between lunch (after about 18h with very little carbs) or dinner (maybe 5-6h with no carbs). But I’m also much more relaxed about the trends I see based on your other posts. I see consistent clearance of peaks within 20-40 min.
Thanks again for your content!
Thanks Mario! In the meantime I had the time to see all your other content and get a lot of context. I now understand what you mean by first-phase response. In my case it is hard for me to say if that’s what it is because on days in which the first carb-rich meal (lunch) was followed by a spike, the second one (dinner) was similar. In any event, I am much more relaxed as I see the very rapid decline practically always.
I do seee other interesting patterns: of a second hump (usually much lower), which I tend to attribute to protein or more complex carbs.
Anyway, thanks for your very insightful balanced content. It’s rare to find this quality and dedication without the bias of an ideological conviction so common in these spaces.
Hi Mario,
Thanks again for all the excellent content.
I’d like to ask something about the 2h mark. Your post mentiones two different benchmarks: staying in the so-called pre-diabetic range (>140 mg/dL) 2h post meal; or staying above baseline (>100 mg/dL) 2h post-meal, which you also highlight in the video.
The former is very clearly abnormal in light of the results you share later. But I’m wondering about the latter, which I also saw mentioned in many other places. First, in many of the real-life studies, it seems like a return to baseline 2h post meal is not very common. For example, in Freckman et al. 2007 all meals show higher AUC 2-3h post meal, and elevated blood glucose is higher than baseline 3-4h or even 4-5h in the others. This is very similar to Song et al. 2023 (Nutrients).
It also matches my own experience, that some meals (particularly “healthier” ones, with more complex and less refined carbs) take 2-3h to clear, while pancakes and syrup are cleared very fast. Further, what I noticed is that if I e.g. go on a bike ride while at a post-meal peak, blood sugar drops, but then sometimes climbes again without eating anything additional. To me, this indicates that the lingering elevated level is not a set amount that is in the blood and takes time to clear, but rather new carbs being titrated into the bloodstream from the digestive system. Otherwise, clearing by physical activity would just get rid of the peak.
In light of this, how hard do you think the 2h clearance border actually is? Could it be that it’s a theoretical value based on OGTTs that are challegning but at the same time as naked as it gets? Or is it based on empirical observations I’m not aware of?
Thanks!
Thank you for your comment.
We don’t really have conclusive evidence on this, but in my own experience from running 100s of glucose tolerance tests, I’d say that if the 2-hour concentration is >140 mg/dL (7.8 mmol/L), that is strongly indicative of prediabetes, i.e., clinical-grade glucose intolerance. And any meaningful residual elevation above baseline or at least 100 mg/dL at the 2-hour mark is indicative of earlier-stage, more mild glucose intolerance, often due to insulin resistance.
I don’t know the Song paper, but the Freckman data are very consistent with my take on this. Look at figure 2: the mean interstitial glucose level two hours after the three meals was about 90 mg/dL, or at least under 100 mg/dL. The mean +- 2 SD (i.e., the 95th percentile) was higher, but note that due to the small sample size, this is just one person. Not very meaningful, because that one person may well have been slightly and pre-clinically glucose intolerant.
One point you make is a very good one though. If someone eats slowly, in several courses, and uses strategies to slow down carbohydrate digestion (low-GI complex carbs later in the meal after a pre-load of non-starchy vegetables/salad and protein), glucose will enter the blood stream much more slowly. However, even in that case, if you start the 2-hour clock when you eat the bulk of your carbs, most of that will be cleared in most people by the 2-hour mark, or shortly thereafter. Also, in that scenario, your glucose level may rise quite minimally, in which case I wouldn’t be concerned about this anyways.
Taken together, my 2-hour rule is not intended as a strong clinical diagnostic tool, but to give people one additional datapoint to detect shifts in their glucose tolerance. If you have basal and 2-hour values mostly in the 80s when you are 30, and then that shifts to mostly around 95 in your 40s, and then the 2-hour value lingers around 120 in your 50s, then this is nothing that a doctor would diagnose, but it can give you a sense that you are slowly becoming glucose intolerant. And understanding this and being able to act sooner by identifying and addressing the root cause(s) of that glucose intolerance can be very valuable.
Cheers
Mario
Thanks again Mario.
I have a small followup, looking closer at Fig. 2 in Freckman et al:
The mean curve we see there is very smooth: flat, up when you eat, and down again. I imagine that this is to a degree an artefact of presenting an average, in which many of the ups and downs of personal physiology and CGM reading errors cancel each other out.
From your experience, how often do real-life CGM outputs look like that? vs how often do you see minor fluctuations that are not easily attributed to food/excercise?
Again, I’m asking from personal experience: I see quite a lot of “messy” peaks: for example, I eat a carb-rich meal; BG goes up (not crazy – let’s say around 130 mg/dL); then it goes down a bit, and up a bit, and back to baseline. Sometimes it would even bring me back to baseline very fast (e.g. 1h post meal), but then moderatly climb again. The latter is similar to what I described following an excercise, in which BG suddenly “comes back”, which to me seems more like continues titration into the blood stream.
Bottom line: how normal is it to have rather messy ups and downs? Is it common, or should a healthy person aspire to look like the mean curve in Freckman et al.?
Thanks!
Sorry, just noticed that I never responded …
You are correct that the average blood glucose responses of a group of people look a lot smoother than those of any individual. That is because when we are wearing a CGM, we tend to move around a bit, then sit down, then be a bit stressed, etc., all of which will affect the blood glucose level enough to cause the seemingly random ups and downs you describe. Add to that the inacuracies in the measurement of glucose in interstitial fluid, which is what the CGM measures, and this means that no one has the smooth curves shown in papers that use averages of groups of people.
Thank you Mario. I’m using the Vively App & keep being told I’m having ‘spikes’ whenever my glucose goes over 6mmol/L. Very unsettling. By your criteria Ive only had one spike, when I ate 8 pieces of brown rice sushi, it went to 10.2 very briefly. So useful to see that as sushi was until now my go-to healthy lunch. Not even the brown rice saved me! So I think I’ll ditch vivley & keep reading your stuff. Fear mongering is not helpful.
Spartano came out in The Journal of Clinical Endocrinology & Metabolism, 2024, 1, 1–7, where those defined as normoglycemic, (presceened with A1c or FBS and wearing a Dexcom G6 Pro spent “1.3% of time (>15 minutes/day) at greater than 180 mg/dL.” Just more support for not taking these numbers with the same rigor as a lab derived number.
Yes, but I would suggest that any spike above 180 mg/dL may already be a sign of slightly impaired glucose tolerance. We don’t have research on this yet, but my best bet would be that those people have a measurably reduced disposition index (the product of insulin sensitivity and insulin secretion in response to a glucose challenge), and as a result lower oral glucose tolerance.
Hi Mario! Thanks so much for putting so much useful information out into the world! Do you know why a young (33), healthy, active woman would have consistently high glucose overnight? Since my doctor started measuring my fasting glucose in my late 20’s, it has always been on the high-normal, low-prediabetic range (99-100 for example). I have been using a CGM for a few months and it consistently shows that my glucose doesn’t really go below 100 (on average) during the night. I stop eating ~2-3 hours before bed and try not to eat a lot of carbs at night. Any ideas?
Hi Caroline,
Could it be that you have excess visceral and ectopic (liver) fat? You can complete the poster I share in this video here to get a sense of how likely that is:
https://youtu.be/Fg3n-vi2t3Y?si=8KRHUOXI0g-EZs6U
Other than that, have you ever measured your insulin sensitivity and beta-cell function? That may give you a sense of what may be going on:
https://youtu.be/gdqp-j9bxMI?si=qlamhwEZJ0T40p0e
If you find that you are insulin resistant, it may be worth looking through the potential root causes of that insulin resistance:
https://youtu.be/HYtnlRCq83s?si=nIr85gwPk6daleEP
Cheers
Mario