Some assumptions are so obvious no one bothers to state them, or even remember them.
Assumption: Medical professionals are trained to use Evidence Based Medicine (EBM).
One might assume that EBM means something specific, and I'm sure it does, but that specific thing is different for different people.
One thing that EBM sometimes means is: common sense is no substitute for evidence. There are uncountable times that common sense has been wrong in the medical context.
So, you have a lot of people commenting here that this is obvious common sense, but many medical professionals will pull out a reference chart of caloric content and glycemic index and say "look at the evidence".
So, it is very useful to do studies that bring evidence to common sense.
---
My other pet peeve with EBM is that it does a poor job of understanding that different interventions work for different people - it is time consuming and expensive to do that kind of investigation, thus some/many medical professionals do not understand or believe it.
This comment is intended as a critique, not a dismissal.
Your take on Evidence Based Medicine (EBM) is wrong.
At the top of the evidence hierarchy is N-of-1 trials (and below that are high quality meta-analyses of trials). Nothing is more informative about treatment response in a person than testing it in that person. This is the heart of personalized medicine, and exactly for the reason you stated: Different interventions work differently for different people.
And any practitioner worth their salt is unsurprised by this headline. A great example is that illness and inflammation increase insulin resistance via counter-regulatory hormones.
You got one thing right, intuition often turns out to be wrong. That is why the vast majority experimental therapeutics built on great ideas never get passed initial testing.
It is wonderful that we have things like CGM's and patients can establish their own baselines and their own evidence profiles on a high(ish) frequency basis.
Such tools also enable studies that can be done at a scale and cost level that is reasonable and can push forward the communal knowledge base.
1. You wear a CGM for 2 weeks, and log everything you eat.
2. At the end, you get a personalized report about how different foods that you ate affect you, personally. And they can extrapolate to other foods that you didn't eat during the testing period.
3. Zoe's model gets better and better the more people sign up, and the more data they get.
This is the type of thing that the linked studies refute - there are either too many external variables to control, or our bodies glucose response is far more complex than we currently understand. Or both - both of the above seem possibly true at the same time.
Either way, if something as simple as eating a few bites of bacon before you eat your toast can change your glucose response, manually logging your meals for Zoe isn’t gonna provide enough data for any reliable extrapolations.
> Either way, if something as simple as eating a few bites of bacon before you eat your toast can change your glucose response, manually logging your meals for Zoe isn’t gonna provide enough data for any reliable extrapolations
I'm not sure - the fact that it does vary from person to person, doesn't mean there aren't groups of people for whom it behaves more predictably, which could be clinically useful.
A CGM is not easy to use reproducibly. The quality control of the manufacturers is not that good, the placement of the sensor (particularly if it is near a place where insulin injections have caused lumps under the skin, or if it is on the back of the arm, where there may not be a lot of flesh) may cause some troubles, the calibration process is dubious, since CGM readings lag blood glucose by around half an hour, rapidly changing blood sugar levels seem to confuse the monitors, and very high or very low blood sugars seem to cause the accuracy of the sensors to deteriorate noticeably a few days prior to their respective expiration dates.
And how do you know that the logs are accurate? The best subjects are incarcerated subjects.
All that said, it's a long row to hoe, and the monitors are definitely worth all the annoyances; they are way better than peeing on a strip of testape like we did back in the old two-holer. But science is very hard.
> And they can extrapolate to other foods that you didn't eat during the testing period.
I would need to see an independent researcher confirm they can do this. As of now, I don't believe they can, and I'm not even sure the theory is reasonable.
> So, you have a lot of people commenting here that this is obvious common sense
There's a lot of reductionist commentary that happens under these links primarily because people aren't reading past the headline or a brief skim of the article.
The topic might be common sense, but measuring and quantifying it with hard data is valuable.
The PR-massaged headlines usually omit the interesting parts and focus on something basic because they want to appeal to a wide audience, not because the study itself was so simplistic that it could be summarized by a headline.
> My other pet peeve with EBM is that it does a poor job of understanding that different interventions work for different people
This has not been my experience with doctors throughout the years. In fact, I'd say the opposite has been true.
In my experience, particularly when diagnoses are trickier, doctors are more likely check the efficacy of the current treatment and change things up if it's not working out.
> "So, you have a lot of people commenting here that this is obvious common sense"
You have this about everything, everywhere. It's a pet peeve how much stuff people will attribute to "common sense" so they can do the internet "I'm superior" thing.
"Wear sunscreen, it's just common sense". No it isn't. We evolved on Earth under the sun, we feel good when going out in the sun, it's bright and beautiful. Rubbing petrochemical distillate or industrially processed plant extract on your skin so the invisible light discovered in 1801 doesn't denature the invisible DNA discovered in 1869 is not common sense it's learned behaviour. Nothing much about Science is common sense, it took thousands of years from the dawn of Civilization until the Enlightenment era and still people can go through years of education and then choose to believe what we want to believe instead of what the evidence shows.
'Common Sense' is that the world is simple, designed for a purpose by a human-like mind one or two levels up from us on the power scale, and inhabited by life-like energies and spirits, some of them malevolent. Common Sense is that things which didn't happen today or yesterday will probably never happen. Common Sense is that things which happen together cause each other; if the relative comes to town and the crop fails then they are bad luck, if the relative comes to town and the baby is born healthy then they are good luck.
Why would it be any kind of 'common sense' - 'sound judgement not based on any specialised knowledge' - that glucose (1747) response differs for the same meal if you need a continuous glucose monitor (FDA approved in 1999) to find that out?!
You are using one particular "definition" of the phrase common sense. It does appear when I search as well. But so do others.
knowledge, judgement, and taste which is more or less universal and which is held more or less without reflection or argument. As such, it is often considered to represent the basic level of sound practical judgement or knowledge of basic facts that any adult human being ought to possess.
Now with that definition some of what you said very definitely would no longer be the case as the above definition would seem to automatically "adjust with the times" so to speak in that what we might expect most adults to know and understand changes over the years, decades and definitely centuries.
To be fair, what I quoted said "any", so I guess they mean closer to 99.99% but I'm with you on "unfortunately common sense seems to not be quite as common as one might wish for" ;)
And this isn’t even universally true. It’s tied to skin color. Sunscreen won’t do much for someone with extremely dark skin. In medical lit, “common” often refers to “white persons.”
Because evidence based medicine is often about broader problems, medicine 3.0 is about solving problems for a specific person. It could be a rare type of cancer that a specific treatment has been crafted for that person - in this case theres no evidence to say that exact treatment works, maybe the over arching therapy is backed by evidence but it feels like a trend away from needing strong evidence to back the application of the therapy.
I agree 100% but sometimes you see studies that confirm not only common sense but also very well researched science that is frankly beyond the need for replication, and that definitely triggers a "wouldn't that money be spent more efficiently by trying to do something new instead?? feeling anytime I see one of those studies.
A counter-point, in a certain sense: when the conclusions of scientific papers (in these softer science fields), contradict common sense, they tend to be unreproducible; the ones which don't, are.
The problem with studying humans is, roughly, the central limit theorem doesnt work: properties of biological and social systems do not have well-behaved statistics. So all this t-test pseudoscience can be a great misdirection, and common sense more reliable.
In the case where effect sizes are small and the data generating process "chaotic", assumptions of the opposite can be more dangerous than giving up on science and adopting "circumstantial humility". (Consider eg., that common sense is very weakly correlated across its practicioners, but "science" forces often pathological correlations on how people are treated -- which can signficantly mangify the harm).
> when the conclusions of scientific papers (in these softer science fields), contradict common sense, they tend to be unreproducible; the ones which don't, are.
Citation needed?
I don't know what would lead to that conclusion. And it would seem to run counter to the entire history of the field of psychology, for example.
Can't find the citation, but remember gwern mentioning a study in one of his posts on replication that found that unintuitive findings tend to be both less replicable and more cited than intuitive ones.
Psychology is the field that is most hit with replication failures and has a slew of unintuitive results that turn out to be malpractice.
Psychology is also the field with a slew of unintuitive results that have been repeatedly replicated as correct. And what is "intuitive" anyways? What was extremely non-intuitive a century ago is common sense today.
So that's why I question the assertion. You're right that there are tons of replication failures, but whether intuition correlates with replicability way doesn't seem relevant. Especially when the point of so much research is to look for currently "non-intuitive" things, so of course that's where more replication issues might exist. It doesn't mean you should stop researching in that direction.
Blood glucose is dependent on more variables than just what you eat.
Decreased sleep slow-wave activity (not just sleep time, but the actual restorative function of sleep) significantly decreases next day insulin response [1].
This is my area (I work in neurotech/sleeptech), but other things that come to mind are changes in changes in gut biome, which can be altered by previous meals, and I assume is always in flux (not my area of expertise), hormonal changes, I'm sure there are others.
However, the conclusion that it undermines the CGM measurements. In fact, I think this makes CGMs more valuable, not less. For those without diabetes, I always thought you'd use the CGM for a few weeks, figure out what your body responds to, and then sort out your diet.
This shows that it isn't that simple, and that we likely need to be monitoring more regularly. If my breakfast spiked my glucose unexpectedly, that may be a signal that I should change my lunch in order to reduce the likelihood of another spike. It becomes about constant management, rather than a 1 time look under the hood.
> For those without diabetes, I always thought you'd use the CGM for a few weeks, figure out what your body responds to, and then sort out your diet.
This was my plan sometime in the next few months.
I think it's still a valid plan, just with some caveats. Anyway due to cost and annoyance I'm unlikely to wear a CGM for more than a couple of months. But that should be enough time to get usable data - like all body tracking data, I'll end up using it as guidance rather than rule.
The caveats are that I'll also need to track my sleep and workouts during that time and carefully look for patterns in the data related to all three.
I already know sleep and workouts are strongly correlated and not always in the way you'd expect. I did a strong HIIT class at 6pm last night, and due to a bit of crunch on personal projects it was my first proper workout in a week. My sleep tracker (Galaxy Watch 6) gave me a very poor sleep and energy score. 61/100 for sleep, 69/100 for energy. I normally score high eighties to low 90s in both.
These scores usually but not always match with my actual feeling. But today I also feel tired (it's 8am here now).
If I had a CGM, I'd also be keeping a strong eye out for unusual glucose response today.
This is more complex than I'd like it to be - I wish my body was as simple to read as just getting a single number like glucose response and making adjustments from that! But as a lifelong migraine sufferer who now has my migraines almost entirely under control by making lifestyle changes, I'm well aware that how my body responds is always a combination of many things. Stormy weather + ate cheese + slightly too strong coffee + stress from work = boom, migraine. Take away any one of those and maybe I wouldn't have got one. I fully expect my body's glucose response to be just as complex.
Well, yes, but it's also pretty dependent on what you eat. From the article: "Additionally, the summarized study did not record the timing of snack and water consumption."
A carb-heavy snack before a meal is going to result in a radically different response than eating the same meal on an empty stomach. I'm glad they acknowledged it, but it feels absurd to publish with this headline if they didn't record snacks.
If a normal-weight person undergoes significant long-term sleep deprivation -- like >50% of nights on 4-6 hours of sleep over the course of decades -- what would the expected effect to be? And if the person's bloodwork nevertheless shows no particular signs of poor insulin response?
Well, you're getting into measuring sleep by time, which is an antiquated idea that we need to move on from. I write about this on the affectable sleep blog [1]. We shouldn't be looking at time, but the restorative function of sleep, and we need to take each individuals restorative function as independent. The average diet recommends 2400 calories (I think), but you wouldn't say that if someone got only 2000 calories a day, they'd starve.
However, on a population level, it is known that shift workers have increased incidence of diabetes, heart disease, obesity, etc. Some of that likely comes down to socio-economic factors, but we can't ignore that daytime sleep or potentially even shifting schedules, is not as restorative as consistent sleep, regardless of sleep time.
I feel I haven't answered your question, but I think this is the best I've got.
What about the famous breakfast? One prevalent theory among nutritionists is that the breakfast has huge influence on the insulin responses during the day: a low carb/no sugar, high protein and fat breakfast will not spike your blood glucose level and therefore will lead to a more moderate response to meal intake during the day as opposite to a high carb breakfast. I think the study should control this variable as well.
Yeah they bought veri.co last year and announced the integration of metabolic into their platform last week.
As the authors mention in the end, personalized nutrition based on CGM only makes less sense than also integrating sleep, stress, movement,... - which Oura is really good at.
I've been diagnosed with Type 2 diabetes and wear a CGM. I find that my blood glucose spikes much later than what is suggested. For example, I might eat a meal and my blood glucose doesn't spike until after 2 hours. When it spikes, I see the body react with insulin and it drives my blood sugar levels down, so I'm not sure if I actually do have T2D or if this is just how my body works.
Another thing I've noticed is that if I eat a very rice-heavy meal, my blood sugar levels may rise throughout the night. I don't think this is insulin-resistance but rather my body digesting the rice. You can't expect the body to digest all the rice in 2 hours, can you, there certainly must be parts that are protected from the stomach acid until much later in the digestion process. So that feeds into the high blood sugar levels overnight in my opinion.
I think if anything, CGMs have opened up the idea of what diabetes really is and how different bodies handle blood sugar. I think I'm borderline T2D, not full-on T2D despite what my doctor says, and I've started wondering if my blood sugar has always been high, but normal for me. On average it's about 120 mg/dL, but I do see my body react properly to new sources of blood sugar and drive it back to "normal" levels, so the idea that I have insulin resistance doesn't make sense to me.
I’m diabetic and manage it with a CGM and insulin pump (HbA1c at 6.8 after many years of struggles).
For convenience and my experiment, I eat practically the same meals at lunch every day, precisely weighed, always starting from very similar morning glucose levels, and strictly respecting timing and consistency.
I NEVER get the same response.
Never.
It’s an experiment I’ve been running on myself for a year.
It’s useful for me, but for the diabetes team following me, “that’s not possible, there must be other factors, it doesn’t show, it’s the ‘CGM algorithm’” (a mystical object no one knows anything about, except that it’s supposedly intelligent).
This study is interesting.
I hope this kind of information, this doubt, trickles down into the medical community. Even though I don’t have much hope. Maybe in years and years.
I am a developer and i work from home. In my experiment i ruled out “any kind of movement” :-D (no, i am not a bag of fat, only a lotto e “not in perfect shape”).
But. I can confirm that lot of stress and lot of coffe have a big and misurable effect.
I definitely see effects on my glucose the day after when I change my exercise routine, such as doing an extra LISS session. Anticipation of something stressful or exciting does also have an effect. I haven't noticed any effect from simply expecting more exercise later (without emotions attached).
I believe a lot of this comes down to liver and muscle glycogen storage and release. (I'm T1)
I think the “experts” thought they knew but CGM is showing they really don’t know.
Sometimes my blood glucose will be high-ish like maybe 135. Then if I drink some coke, ie pure sugar, it causes my blood glucose to spike and then my insulin kicks in and it drives it back down to 110-120. It’s as if the switch to turn on the insulin wasn’t turned on because there was a slow steady creep up of blood glucose and then it needed to be woken up.
The body sure is strange but one thing I do know is that CGMs are changing things for sure.
I do _not_ have any sort of insulin resistance, and have worn a CGM for a while, a few different times.
(Purchased sorta 'just for fun', to evaluate my own metabolic responsiveness. I wrote about it once: [0])
if I moved around much after eating, my muscles would "soak" lots of glucose from my blood and the insulin change would be relatively low. If it was something sugary and I then went to sleep, it would be a big slow rise for a few hours of blood sugar, then insulin would seem to be released, and it would decline a bit, then flatten, then decline, until it went back to a regular level. When waking up, it would often by around 70 mg/DL, and even if I eat zero carbs (or am full-on fasting!) it would bounce up/around during the day. I appreciated that my body seemed quite capable of doing whatever it needed to do to raise my blood glucose levels. (They always go up when exercising, for instance, even if no food is consumed)
I think everyone would likely find the data interesting, it's so freaky customized to yourself, it cannot help but be interesting. They're expensive, so I think buying even one sensor and wearing it for 2 weeks or however long is worth it.
I'm a fellow type 2, be sure to talk more with your doctor about these points as such comments/questions are of things they should have explained very clearly the first time they said your sugar seemed high.
Type 2 "resistance" is about the quality of the response to high glucose levels, not the complete lack of a response to them. There shouldn't be a long buildup overnight in that scenario for a person without insulin resistance/deficiency and still having a measurable insulin response is normal/expected of all but the worst Type 2 diabetics. Yes, you're still digesting, but in individuals without diabetes the blood sugar peak occurs (and ends) well before digestion is finished because influxes of carbs can still be effectively managed by the insulin alone rather than by the lack of additional carbs to digest. If it were just that one's digestion were a lot slower than a normal persons then it should still result in a lower, but still quickly managed to baseline, peak.
You may well actually be prediabetic though, it just depends on the specific numbers for A1C/average & peaks combined over time and not the presence of a response itself. The recommendations between higher side prediabetic and lower side type II diabetic shouldn't be all that different in the end anyways though.
> You may well actually be prediabetic though, it just depends on the specific numbers for A1C/average & peaks combined over time and not the presence of a response itself.
Just to clarify - A1C is itself an easily-measured proxy for diabetes mellitus, but it's itself a heuristic. There are groups for whom it is known that the "standard" A1C range is actually incorrect, because of confounding factors that affect the A1C measurement but are unrelated to the metabolic dysfunction or general sugar levels.
Your point is correct, though, that what OP is describing is consistent with diabetes, and the actual clinical recommendations for prediabetes and Type II diabetes are often the same, at least in the early stages.
> I find that my blood glucose spikes much later than what is suggested.
I'm a type 1 and my blood glucose can response can vary wildly. Sometimes it can spike quickly (15-30 minutes), other times it can take 2, 3, or even 4 hours. The reverse is also true, in that insulin can sometimes effect me quickly, and sometimes it can take effect 3+ hours later. In general, when having a big meal, I'll take a "fairly large" amount of insulin, but not enough for the entire meal; then I'll take more[1] when my blood starts increasing into 200+ range. Otherwise, I risk it dropping because the insulin was having a "fast" day but the food was "slow".
My endo finds it weird, but we've gone over specific cases of it, including exactly what I ate, when, and what the CGM history for the day looked like. I'm just defective :)
[1] Note I'm not talking about reactively taking more insulin after the "right" amount turned out not to be enough (which is generally a bad idea). Rather, I'm talking about splitting what I know to be about the right amount... into 2 different doses.
Physical activity, mental activity, and sickness can also massively effect dose.
On days I'm physically active I feel like I need less than half the dose of insulin then days I'm caught behind a desk. This said, high impact mental stress that requires lots of thinking can rapidly drop my glucose level.
And that's not counting being ill where my body will pretend like it's insulin resistant for hours then suddenly try to process it all at once.
2 other factors other than physical activity and stress that seem to impact insuline sensitivity:
1. Temperature (room / outside)
2. How well you’re hydrated
Regarding the *time* it takes for eaten Carbs/glucose to enter the bloodstream, it seems that simply standing/being postural often slows down digestion, while sitting/laying down (think of a couch potato watching tv) seems to increase digestion speed.
Typically the glucose level in the interstitial fluid takes longer to respond to your food intake, and the GCMs measure that instead of the blood glucose level.
I just found out a friend was diagnosed with diabetes due to a sugar test while their a1c was still in the normal range. We’ve got a long way to go yet with diabetes research.
I mean, it is well known that A1c has a 3 month (usually) lag time assuming no hemolytic states. Could have been just caught early depending on what you mean by sugar test
I guess I’m confused, A1c is a measurement used for diabetes, not fatty liver/nafld. Can absolutely be correlated with diabetes and elevated A1c but those are two independent things we seem to be talking about
There is some genetic variability in the correlation between blood glucose versus HbA1c. The HbA1c test is essentially an integrator of blood glucose level over time but different people have different average red blood cell lifespans.
Usually T2D is diagnosed based on a fasting glucose test. I also recently worse a CGM because I was borderline T2D and wanted to know how various foods impacted my levels. My peak was about 2 hours after the meal, so about 1.5 hours and change, given the lag of the CGM.
I also found that sweets not only spiked my values but did so over a 6-12 hour period! That was shocking. I have not had dessert for 107 days because of it. I plan to do a fasting test (and A1C) in August to see how many values are. I hope they are much lower, but we'll see.
EDIT: also, popcorn with nothing on it but salt gave me a huge spike.
Popcorn (even plain air popped) is more carbs as a percentage of mass (~70%) than eating a bag of Hershey Kisses (60%) so it's no surprise. I've always found starches to be the bigger culprit of my spikes than sweets since it's relatively easy to overlook and much easier to consume a large quantity without thinking about it.
Odd, I would have expected this to be somewhat the case? Specifically, I would expect your activity leading into the meals, along with your expectations of what you will be doing after, would have some impact? Probably more impact from how hungry you were going into a meal. If you were already sated, for example, I'd expect your body to largely try and push the meal through as fast as it can.
Neat to see what other people's priors are, on this.
> Odd, I would have expected this to be somewhat the case?
I don't think anybody was expecting to be surprised by this study. In practice, most science is pretty boring and rarely breaks expectations. But being unsurprising does not mean it's not worth doing. A lot of studies are simply validating expected outcomes and providing foundational data points for future studies to refer back to. For example, a future study might use this study to justify funding ("as shown in Study 2025.abcd, glucose is highly variable... we propose to further study this by controlling for ... which will help us understand the influence of ..." etc etc).
Ah, great point! I definitely got the impression that this was a surprise, but I can't say why I got that impression. That said, I did not intend this as a criticism of running the study! Quite the contrary, glad they did it. I was only meaning my comment to be contra the sense of surprise I had.
I'm very sensitive to sugar and starch when "at rest". If I wake up first thing in the morning and drink an orange juice or have a bowl of white rice before bed, my blood sugar and mood are out of control. But the same foods mid-day before/after/during intense mental or physical work are very well tolerated if not beneficial. So my priors agree with yours, timing and context absolutely matters.
Non expert here ... My understanding was sleep, stress and many other variables all impact these things significantly (before we even consider food). Having a different context when you then add the same food thus should not result in the same outcome.
The study used 7-day rotating meals, so one would expect some consistency on average in a weekly rhythm, but it also only covered 14 days, so only two data points per meal and participant, if I read this correctly. In that case, I’m not surprised either that there is a lot of variation, also considering that this is a new meal regimen for the participants.
I think I agree with the idea you are saying. That people think you can formulate our body to ignore a lot of the state that it holds.
I'll note that even basic physics has that problem. Try explaining to a 5th grader why a feather would fly in the same arc that a rock will take, if there was no air.
> Odd, I would have expected this to be somewhat the case?
Well of course. Studies like this help collect data and quantify the variations.
There's more to the study than the headline or even the simple summary. Knowing the range of variations is important, as well as starting to build a foundation to understand some of the factors that lead to the variations.
I've been working in this space for seven years, this isn't new. There have been big studies from the UK in 2020 [1] and Israel in 2015 [2] about this.
Some of my learnings:
- Don't start your day with a large amount carbs. Have some insulin in your blood before eating that big bowl of oatmeal. Or just go for some yougurt with nuts and seeds.
- The classical order of a three course meal (salad first, then main dish, then dessert) is pretty good in terms of preventing glucose spikes.
- Going for a walk after a meal is great for bringing glucose levels down.
- Eat at least 2h before going to sleep. Having high glucose levels disrupts sleep.
- Alcohol lowers the glucose response of a meal, but is still bad unfortunately.
- Diet Coke works. No spike vs loads of sugar with a real coke.
- Stress can spike glucose like crazy, e.g. being in an interview or during takeoff.
- If you really want to know how you react to some food, keep the circumstances (time of day, sleep, physical activity, stress) similar. There's too much influence beyond just the meal.
I bet that everyone who is wearing a premium smartwatch or an Oura ring now will be using a CGM now and then in the next years.
"Additionally, the summarized study did not record the timing of snack and water consumption. This is relevant because the sequence and timing of subsequent meals (or snacks), as well as the food processing and variability, can influence postmeal glucose responses." - probably a factor, hydration status is huge. This result is kind of like 'duh' though.
Hydration, activity and stress levels. Adrenal responses can pause the digestive process to conserve energy for fight or flight. Muscle use also diverts resources, stressed or not. And being dehydrated essentially concentrates everything in your blood, complicating kidney and liver function.
Shouldn't come as a surprise - there are so many factors involved in glucose response to food that it's almost impossible to replicate a glucose response even with a controlled intake. Sleep, activity levels (before, after, during, even days before, etc), stress, hormones, all of these are major factors involved in how glucose is metabolized.
Jessie Inchauspe [1] created a business (she's the "glucose goddess") and wrote books around the different blood sugar responses she and others have to meals, exercise, etc... I don't recall all the ideas, but here are the two I remember: Exercise before and/or after eating smooths out the blood sugar response to food. When eating a meal, it's better to eat high-fiber stuff first, then complex carbs, then protein, then fats, and eat simple carbs last (Or something like that order).
GLUT-4 activation in the muscles primes them to take up glucose and other nutrients. So air squats or going for a walk before/after a meal can significantly blunt the insulin response, because there's less glucose in your blood stream, because more of it has been ported into muscles.
Yeah, I think so, though my recollection is that a baked potato's glycemic index is kinda high. For more authoritative list of details, you could listen to the following podcast with Inchauspé: https://overcast.fm/+AA4G32XuwoA
It would be unfortunate if this article were to discourage folks from trying CGMs out. I am not diabetic but have found a lot of value in using a Stelo CGM sensor. At minimum, it gives me awareness of my current state. I routinely check it before considering desserts for example. Also, I was shocked at the spikes I would get from some food. Maybe an "obvious in retrospect" feedback was drinking a bottle of (non-diet) Coke. It consistently gave me 50 unit spikes (mg/dL). Nothing came close to it, including large meals or meals + desserts. On the other hand, drinking diet coke had absolutely no impact, which was contrary to what I've read before (the body is confused and produces insulin). Another surprising observation is the effect of consuming Feta (a particular kind of Greek cheese) : no matter what else I may have eaten, a piece of feta always leads to a reduction of 10 to 20 points for at least half an hour after consuming it(!) I doubt this is generally true, which speaks to each body reacting differently.
She’ll think that she’s cracked it, and then the next day, with the same meal at the same time, her sugar levels go high.
Her words: there is nothing else where you have studied for over 30 years and STILL feel like you know nothing.
It is incredibly demoralising for her sometimes - especially when she’s suffering also from a high/low sugar level. I have the upmost respect for anyone having to do the amount of work, to get to zero (sometimes).
I'm also T1D and yep. That's how it goes. Despite having one of the same two breakfasts pretty much every workday for a decade, it's a crapshoot whether I'll be 300 (very high) when I get to work or 50 (very low) or 110 (good). I just have to adjust when I get in to the office. Most days I deal with it fine, but every once in a while I get fed up with it and want to throw something.
Still pissed my insurance stopped covering the Omnipod Dash last year, especially since the 5 is just an artificially locked down dash. Omnipod 5 automated mode is garbage compared to aaps
I like to say that the CGM is, far an away, the biggest life improvement for diabetics since the invention of insulin. And this plays into one of the reasons
- You able to learn better about how your body responds to different foods
- You are more secure about your blood sugar dropping when you don't expect/realize it (sleeping, driving, etc)
- You better able to detect when your body isn't behaving the way you would normally expect it to (the point we're discussing here)
She’s on closed loop and it has certainly changed her life - perimenopause was really difficult but the loop has meant she has much better control. The fact it cuts out when it drops has given her confidence to sleep better etc
Keto works well, but I'm concerned about the extreme fat intake. I did manage to lower my A1C from 9s to mid 5s using Keto as a T1D, but eventually settled into a high protein, moderate fat, low-ish carb diet and that has worked pretty well without being hyper-restrictive. Heavily inspired by the late Dr. Richard K. Bernstein.
every endocrinologist i've ever had says you need at least some carbohydrates. that said, they're usually not nutritionists, and there's historically been so little focus on food as medicine within western medicine that i think we "officially" know as much about that as we do about diabetes at all (which is a lot, but still surprisingly little).
there are plenty of good reasons we eat carbs (especially complex carbs, plus trying to avoid processed stuff); as a diabetic you just have to prepare ahead of time. it sucks, but honestly, as a 20+ year T1D, not eating cake sucks more than having to plan to eat cake :)
> not eating cake sucks more than having to plan to eat cake :)
But you have never felt the alternative. It's not just eating cake. You have to see the whole video that I linked to understand the differences (they might not all apply). Think a smoker who've never known what's it like without smoking.
dad of an 8 y/o T1D here, and tbh, "how about just ..." is about the worst thing you can say to anyone dealing with this, especially given the subject of this post. things are just never the same. maybe they will be later in adulthood? i feel like i'm in a position to beta test this with my kiddo and i don't like it at all.
I understand that T1D is special, but this is said for every disorder.
There are many levels to keto diet, and believe me, you can't do the epilepsy version even if you'd wanted, it takes time and hard effort, for children it's done inside a hospital.
So assuming you watch everything they eat, you can keep track of GCM, and try very slowly. Like going from 60-30-10 carb-protein-fat to 50-30-20 ratio, do they have better consistent blood sugar ?
There should be protocols to do this for T1D the safest way.
What about the famous breakfast? One prevalent theory among nutritionists is that the breakfast has huge influence on the insulin responses during the day: a low carb/no sugar, high protein and fat breakfast will not spike your blood glucose level and therefore will lead to a more moderate response to meal intake during the day as opposite to a high carb breakfast. I think the study should control this variable as well.
As a long time diabetic (t1d) with direct relatives also afflicted with this horrible disease, it’s frustrating that we often miss the key factors in insulin resistance and fail to see how they fit in.
The level of serotonin and its interplay with our current level of immune response/inflammation in our body is the single best predictor of insulin resistance. That’s simply because every good thing you can do to lower resistance increases serotonin. Serotonin then decreases inflammation. When you exercise for a long period of time you are not only increasing your immediate availability of serotonin, you are also increasing your daily availability for serotonin.
Obviously it’s a little more complicated but from a big picture standpoint, either something is increasing inflammation in you today and increasing your insulin resistance or (hopefully) the opposite is happening and serotonin is moving your insulin through your blood stream like a mag lev train.
> if an individual eats the same meal on two occasions but starts with the vegetables on one occasion but not the other, this may lead to different glucose responses to the same meal within the same individual.
> Other behavioral and individual factors are known to influence CGM responses.[22][6] Postmeal physical activity — even as minor as leg fidgeting[23] and walking for a few minutes every hour — reduces postmeal glucose responses.[24] Sleep quality has also been associated with changes in postmeal CGM glucose responses.[25] Emerging studies are also uncovering the relationship between an individual’s gut microbiome and their postmeal glycemic responses.[26][27][28]
So basically it's the impact of the parasympathetic/sympathetic nervous system and the order of food ingestion that could induce different glucose response.
It's interesting because I assumed that beside the nervous system, whatever order we eat our food, it all mixes in the stomach and then start to process. Negating the importance of order but studies implies it does.
It is interesting, but it has been long well "known" that if you eat the more fibrous foods first this effectively slows the digestion of what comes next.
Sounds expected? If you are glycogen depleted, it will cause less of an insulin spike. This sounds like if you fill up the same type of car with 5 gallons of gas. Sometimes it reports full, and sometimes it doesn't. That sounds odd, until you accept that you are never fully sure the gas level you started at.
Leaving aside the science, this paper draws the wrong conclusion. Just because many factors influence BG response does not mean CGMs are somehow invalid as a tool for personalized nutrition. Quite the opposite. A good nutritionist does not work in a vacuum. They know the holistic nature of BG and will tailor their recommendations to the patient based on lifestyle. (I worked on a holistic diabetes management platform)
I like the idea that the proliferation of LLMs become frameworks for understanding that can help advance a field or help curious thinkers come up with novel hypotheses. If the body is something like a trained model, it’s not crazy to think that it will respond to something like fake sugar (eg aspartame) with biological processes that would only be expected for real sugar.
There are problems in the US healthcare system with slow adoption of clinical practice guidelines and lack of research funding for non-pharmaceutical treatments. But on an objective scientific basis the recent studies funded by pharma tend to be some of the highest quality work. They have huge budgets and are held to fairly high standards now by the FDA so we tend to see large subject populations, multiple sites, proper controls, and rigorous statistical analysis.
Is this already pretty expected? We already knew that glucose response was highly dependent on time of day, pre or post meal movement, and whether the meal is moderated by other components like fats that can slow down the immediate impact.
i'm a type one diabetic. this is uh, not news to me. i can eat the same thing every single day at the same time and see different numbers after every meal. stress, hydration, heat/cold, fatigue, if the wind is blowing, so many other things affect blood sugar.
My mother and I recently did an oral glucose tolerance test while wearing CGMs and discovered that in us, the CGM measurements (Dexcom G7) and the reference instrument measurements (whatever Quest Diagnostics uses) virtually _never_ lined up, even when including the G7's advertised error margins. The blood glucometer readings in me didn't line up either. Admittedly this was on the first day using the G7, when it's apparently less accurate, but due to this I've been doubtful of CGM readings in me/her. I'm sure CGMs work on a population level, but for us, on that day? Nope.
As the parent of a type 1 diabetic...all I can say is "duh, no shit." Activity levels, emotional state, how much time between eating & bedtime, agem among other things affect glucose response.
Sorry to complain, but what the hell is this horrible AI summary link?
The actual study is not open, but surely enough something a bit more precise could be given.
One factor that scientific food experiments don't seem to include is gluttony. Presumably because it is subjective (though nonetheless real for that). Yet if people are motivated by the amount of comfort and/or pleasure they obtain from eating their favourite foods this may have an effect on their physiological response.
I realise the word 'gluttony' is off-putting (for atheists for instance) but I don't know a better one. Perhaps 'over-excitement'?
The symptoms of excitement are of course similar to the fight-or-flight response. Raised heart rate for one. Butterflies in the stomach indicate release of adrenaline, which in turn raises blood glucose -- not from the food being eaten but from the body's stores. So there's a link right there.
How excited you are about a given food will depend not only upon the foodstuff itself but also on your personal taste and personal history. If you have taken pleasure in its consumption many times in the past this may affect your present level of excitement.
'Arousal' is another word that just sprang to mind.
Here's a (Gemini) AI summary that seems to be capturing the main point that the article wants to convey.
```
This page summarizes a study about continuous glucose monitors (CGMs) and how individuals can have highly variable glucose responses even when eating the same meals. The study, which involved 30 participants without diabetes, found a weak-to-moderate correlation between glucose responses to identical meals, with about 80% of the variation attributed to within-participant differences or measurement error.
The article notes that high glucose variability is linked to increased risk of death, greater hunger, and poorer mental health and sleep. While CGMs are beneficial for people with diabetes to prevent hypoglycemic episodes, this study suggests that developing personalized nutrition recommendations for glycemic control may be more complex than previously thought due to inconsistent individual responses. The authors emphasize the need for more reliable dietary assessment and a deeper understanding of the behavioral, dietary, and individual factors that influence glucose responses.
```
Some assumptions are so obvious no one bothers to state them, or even remember them.
Assumption: Medical professionals are trained to use Evidence Based Medicine (EBM).
One might assume that EBM means something specific, and I'm sure it does, but that specific thing is different for different people.
One thing that EBM sometimes means is: common sense is no substitute for evidence. There are uncountable times that common sense has been wrong in the medical context.
So, you have a lot of people commenting here that this is obvious common sense, but many medical professionals will pull out a reference chart of caloric content and glycemic index and say "look at the evidence".
So, it is very useful to do studies that bring evidence to common sense.
---
My other pet peeve with EBM is that it does a poor job of understanding that different interventions work for different people - it is time consuming and expensive to do that kind of investigation, thus some/many medical professionals do not understand or believe it.
This comment is intended as a critique, not a dismissal.
Your take on Evidence Based Medicine (EBM) is wrong.
At the top of the evidence hierarchy is N-of-1 trials (and below that are high quality meta-analyses of trials). Nothing is more informative about treatment response in a person than testing it in that person. This is the heart of personalized medicine, and exactly for the reason you stated: Different interventions work differently for different people.
And any practitioner worth their salt is unsurprised by this headline. A great example is that illness and inflammation increase insulin resistance via counter-regulatory hormones.
You got one thing right, intuition often turns out to be wrong. That is why the vast majority experimental therapeutics built on great ideas never get passed initial testing.
It is wonderful that we have things like CGM's and patients can establish their own baselines and their own evidence profiles on a high(ish) frequency basis.
Such tools also enable studies that can be done at a scale and cost level that is reasonable and can push forward the communal knowledge base.
This is the idea behind the "Zoe" service: https://zoe.com/en-us
1. You wear a CGM for 2 weeks, and log everything you eat.
2. At the end, you get a personalized report about how different foods that you ate affect you, personally. And they can extrapolate to other foods that you didn't eat during the testing period.
3. Zoe's model gets better and better the more people sign up, and the more data they get.
This is the type of thing that the linked studies refute - there are either too many external variables to control, or our bodies glucose response is far more complex than we currently understand. Or both - both of the above seem possibly true at the same time.
Either way, if something as simple as eating a few bites of bacon before you eat your toast can change your glucose response, manually logging your meals for Zoe isn’t gonna provide enough data for any reliable extrapolations.
> Either way, if something as simple as eating a few bites of bacon before you eat your toast can change your glucose response, manually logging your meals for Zoe isn’t gonna provide enough data for any reliable extrapolations
I'm not sure - the fact that it does vary from person to person, doesn't mean there aren't groups of people for whom it behaves more predictably, which could be clinically useful.
A CGM is not easy to use reproducibly. The quality control of the manufacturers is not that good, the placement of the sensor (particularly if it is near a place where insulin injections have caused lumps under the skin, or if it is on the back of the arm, where there may not be a lot of flesh) may cause some troubles, the calibration process is dubious, since CGM readings lag blood glucose by around half an hour, rapidly changing blood sugar levels seem to confuse the monitors, and very high or very low blood sugars seem to cause the accuracy of the sensors to deteriorate noticeably a few days prior to their respective expiration dates.
And how do you know that the logs are accurate? The best subjects are incarcerated subjects.
All that said, it's a long row to hoe, and the monitors are definitely worth all the annoyances; they are way better than peeing on a strip of testape like we did back in the old two-holer. But science is very hard.
> And they can extrapolate to other foods that you didn't eat during the testing period.
I would need to see an independent researcher confirm they can do this. As of now, I don't believe they can, and I'm not even sure the theory is reasonable.
Yup, and this was the original dream of Theranos - lots of measurements, lots of data. It's a nice dream, shame about the lies though.
> So, you have a lot of people commenting here that this is obvious common sense
There's a lot of reductionist commentary that happens under these links primarily because people aren't reading past the headline or a brief skim of the article.
The topic might be common sense, but measuring and quantifying it with hard data is valuable.
The PR-massaged headlines usually omit the interesting parts and focus on something basic because they want to appeal to a wide audience, not because the study itself was so simplistic that it could be summarized by a headline.
The map is not the territory, if only more professionals and doctors read some Alfred Korzybski, the world would have been a better place.
> My other pet peeve with EBM is that it does a poor job of understanding that different interventions work for different people
This has not been my experience with doctors throughout the years. In fact, I'd say the opposite has been true.
In my experience, particularly when diagnoses are trickier, doctors are more likely check the efficacy of the current treatment and change things up if it's not working out.
> "So, you have a lot of people commenting here that this is obvious common sense"
You have this about everything, everywhere. It's a pet peeve how much stuff people will attribute to "common sense" so they can do the internet "I'm superior" thing.
"Wear sunscreen, it's just common sense". No it isn't. We evolved on Earth under the sun, we feel good when going out in the sun, it's bright and beautiful. Rubbing petrochemical distillate or industrially processed plant extract on your skin so the invisible light discovered in 1801 doesn't denature the invisible DNA discovered in 1869 is not common sense it's learned behaviour. Nothing much about Science is common sense, it took thousands of years from the dawn of Civilization until the Enlightenment era and still people can go through years of education and then choose to believe what we want to believe instead of what the evidence shows.
'Common Sense' is that the world is simple, designed for a purpose by a human-like mind one or two levels up from us on the power scale, and inhabited by life-like energies and spirits, some of them malevolent. Common Sense is that things which didn't happen today or yesterday will probably never happen. Common Sense is that things which happen together cause each other; if the relative comes to town and the crop fails then they are bad luck, if the relative comes to town and the baby is born healthy then they are good luck.
Why would it be any kind of 'common sense' - 'sound judgement not based on any specialised knowledge' - that glucose (1747) response differs for the same meal if you need a continuous glucose monitor (FDA approved in 1999) to find that out?!
You are using one particular "definition" of the phrase common sense. It does appear when I search as well. But so do others.
Now with that definition some of what you said very definitely would no longer be the case as the above definition would seem to automatically "adjust with the times" so to speak in that what we might expect most adults to know and understand changes over the years, decades and definitely centuries.Most? 50.1%?
To be fair, what I quoted said "any", so I guess they mean closer to 99.99% but I'm with you on "unfortunately common sense seems to not be quite as common as one might wish for" ;)
> "Wear sunscreen, it's just common sense"
And this isn’t even universally true. It’s tied to skin color. Sunscreen won’t do much for someone with extremely dark skin. In medical lit, “common” often refers to “white persons.”
If you think about it, Evidence Based Medicine is very much against Medicine 3.0 whereby individualized treatments/therapies are used.
Why?
Because evidence based medicine is often about broader problems, medicine 3.0 is about solving problems for a specific person. It could be a rare type of cancer that a specific treatment has been crafted for that person - in this case theres no evidence to say that exact treatment works, maybe the over arching therapy is backed by evidence but it feels like a trend away from needing strong evidence to back the application of the therapy.
How is treatment crafted? Will waving a crystal stick be a part of it?
Literally all medical practitioners are trained to take individual people circumstances into account.
Taking individuals circumstances into account isn't evidence based medicine though...
I agree 100% but sometimes you see studies that confirm not only common sense but also very well researched science that is frankly beyond the need for replication, and that definitely triggers a "wouldn't that money be spent more efficiently by trying to do something new instead?? feeling anytime I see one of those studies.
A counter-point, in a certain sense: when the conclusions of scientific papers (in these softer science fields), contradict common sense, they tend to be unreproducible; the ones which don't, are.
The problem with studying humans is, roughly, the central limit theorem doesnt work: properties of biological and social systems do not have well-behaved statistics. So all this t-test pseudoscience can be a great misdirection, and common sense more reliable.
In the case where effect sizes are small and the data generating process "chaotic", assumptions of the opposite can be more dangerous than giving up on science and adopting "circumstantial humility". (Consider eg., that common sense is very weakly correlated across its practicioners, but "science" forces often pathological correlations on how people are treated -- which can signficantly mangify the harm).
> when the conclusions of scientific papers (in these softer science fields), contradict common sense, they tend to be unreproducible; the ones which don't, are.
Citation needed?
I don't know what would lead to that conclusion. And it would seem to run counter to the entire history of the field of psychology, for example.
Can't find the citation, but remember gwern mentioning a study in one of his posts on replication that found that unintuitive findings tend to be both less replicable and more cited than intuitive ones.
Psychology is the field that is most hit with replication failures and has a slew of unintuitive results that turn out to be malpractice.
Psychology is also the field with a slew of unintuitive results that have been repeatedly replicated as correct. And what is "intuitive" anyways? What was extremely non-intuitive a century ago is common sense today.
So that's why I question the assertion. You're right that there are tons of replication failures, but whether intuition correlates with replicability way doesn't seem relevant. Especially when the point of so much research is to look for currently "non-intuitive" things, so of course that's where more replication issues might exist. It doesn't mean you should stop researching in that direction.
Blood glucose is dependent on more variables than just what you eat.
Decreased sleep slow-wave activity (not just sleep time, but the actual restorative function of sleep) significantly decreases next day insulin response [1].
This is my area (I work in neurotech/sleeptech), but other things that come to mind are changes in changes in gut biome, which can be altered by previous meals, and I assume is always in flux (not my area of expertise), hormonal changes, I'm sure there are others.
However, the conclusion that it undermines the CGM measurements. In fact, I think this makes CGMs more valuable, not less. For those without diabetes, I always thought you'd use the CGM for a few weeks, figure out what your body responds to, and then sort out your diet.
This shows that it isn't that simple, and that we likely need to be monitoring more regularly. If my breakfast spiked my glucose unexpectedly, that may be a signal that I should change my lunch in order to reduce the likelihood of another spike. It becomes about constant management, rather than a 1 time look under the hood.
https://doi.org/10.1016/j.sleep.2022.03.005
> For those without diabetes, I always thought you'd use the CGM for a few weeks, figure out what your body responds to, and then sort out your diet.
This was my plan sometime in the next few months.
I think it's still a valid plan, just with some caveats. Anyway due to cost and annoyance I'm unlikely to wear a CGM for more than a couple of months. But that should be enough time to get usable data - like all body tracking data, I'll end up using it as guidance rather than rule.
The caveats are that I'll also need to track my sleep and workouts during that time and carefully look for patterns in the data related to all three.
I already know sleep and workouts are strongly correlated and not always in the way you'd expect. I did a strong HIIT class at 6pm last night, and due to a bit of crunch on personal projects it was my first proper workout in a week. My sleep tracker (Galaxy Watch 6) gave me a very poor sleep and energy score. 61/100 for sleep, 69/100 for energy. I normally score high eighties to low 90s in both.
These scores usually but not always match with my actual feeling. But today I also feel tired (it's 8am here now).
If I had a CGM, I'd also be keeping a strong eye out for unusual glucose response today.
This is more complex than I'd like it to be - I wish my body was as simple to read as just getting a single number like glucose response and making adjustments from that! But as a lifelong migraine sufferer who now has my migraines almost entirely under control by making lifestyle changes, I'm well aware that how my body responds is always a combination of many things. Stormy weather + ate cheese + slightly too strong coffee + stress from work = boom, migraine. Take away any one of those and maybe I wouldn't have got one. I fully expect my body's glucose response to be just as complex.
Well, yes, but it's also pretty dependent on what you eat. From the article: "Additionally, the summarized study did not record the timing of snack and water consumption."
A carb-heavy snack before a meal is going to result in a radically different response than eating the same meal on an empty stomach. I'm glad they acknowledged it, but it feels absurd to publish with this headline if they didn't record snacks.
If a normal-weight person undergoes significant long-term sleep deprivation -- like >50% of nights on 4-6 hours of sleep over the course of decades -- what would the expected effect to be? And if the person's bloodwork nevertheless shows no particular signs of poor insulin response?
Asking for a friend.
Well, you're getting into measuring sleep by time, which is an antiquated idea that we need to move on from. I write about this on the affectable sleep blog [1]. We shouldn't be looking at time, but the restorative function of sleep, and we need to take each individuals restorative function as independent. The average diet recommends 2400 calories (I think), but you wouldn't say that if someone got only 2000 calories a day, they'd starve.
However, on a population level, it is known that shift workers have increased incidence of diabetes, heart disease, obesity, etc. Some of that likely comes down to socio-economic factors, but we can't ignore that daytime sleep or potentially even shifting schedules, is not as restorative as consistent sleep, regardless of sleep time.
I feel I haven't answered your question, but I think this is the best I've got.
Thanks for the effort!
What about the famous breakfast? One prevalent theory among nutritionists is that the breakfast has huge influence on the insulin responses during the day: a low carb/no sugar, high protein and fat breakfast will not spike your blood glucose level and therefore will lead to a more moderate response to meal intake during the day as opposite to a high carb breakfast. I think the study should control this variable as well.
> One prevalent theory among nutritionists
why don't they test it. should be easy create a study?
i think oura ring wanted to combine cgm and sleep values. Not sure if they ever went ahead with cgm partnership.
Yeah they bought veri.co last year and announced the integration of metabolic into their platform last week.
As the authors mention in the end, personalized nutrition based on CGM only makes less sense than also integrating sleep, stress, movement,... - which Oura is really good at.
They just announced CGM integration last week.
I've been diagnosed with Type 2 diabetes and wear a CGM. I find that my blood glucose spikes much later than what is suggested. For example, I might eat a meal and my blood glucose doesn't spike until after 2 hours. When it spikes, I see the body react with insulin and it drives my blood sugar levels down, so I'm not sure if I actually do have T2D or if this is just how my body works.
Another thing I've noticed is that if I eat a very rice-heavy meal, my blood sugar levels may rise throughout the night. I don't think this is insulin-resistance but rather my body digesting the rice. You can't expect the body to digest all the rice in 2 hours, can you, there certainly must be parts that are protected from the stomach acid until much later in the digestion process. So that feeds into the high blood sugar levels overnight in my opinion.
I think if anything, CGMs have opened up the idea of what diabetes really is and how different bodies handle blood sugar. I think I'm borderline T2D, not full-on T2D despite what my doctor says, and I've started wondering if my blood sugar has always been high, but normal for me. On average it's about 120 mg/dL, but I do see my body react properly to new sources of blood sugar and drive it back to "normal" levels, so the idea that I have insulin resistance doesn't make sense to me.
I’m diabetic and manage it with a CGM and insulin pump (HbA1c at 6.8 after many years of struggles).
For convenience and my experiment, I eat practically the same meals at lunch every day, precisely weighed, always starting from very similar morning glucose levels, and strictly respecting timing and consistency.
I NEVER get the same response. Never. It’s an experiment I’ve been running on myself for a year. It’s useful for me, but for the diabetes team following me, “that’s not possible, there must be other factors, it doesn’t show, it’s the ‘CGM algorithm’” (a mystical object no one knows anything about, except that it’s supposedly intelligent).
This study is interesting. I hope this kind of information, this doubt, trickles down into the medical community. Even though I don’t have much hope. Maybe in years and years.
Would be interesting if a non-diabetic person does the same trial, as a "control".
Note: This is pure speculation, and I'm not claiming to know anything more than passing knowledge about diabetes.
Do you see any relation to physical exertion of the day before/after? What about mental signals, like "I'm going to be walking a bunch tomorrow"?
I am a developer and i work from home. In my experiment i ruled out “any kind of movement” :-D (no, i am not a bag of fat, only a lotto e “not in perfect shape”). But. I can confirm that lot of stress and lot of coffe have a big and misurable effect.
thanks for "misurable". i'll steal this wonderful combination of miserable and measurable, i see lots of use in devops for it... ;-)
I definitely see effects on my glucose the day after when I change my exercise routine, such as doing an extra LISS session. Anticipation of something stressful or exciting does also have an effect. I haven't noticed any effect from simply expecting more exercise later (without emotions attached). I believe a lot of this comes down to liver and muscle glycogen storage and release. (I'm T1)
I think the “experts” thought they knew but CGM is showing they really don’t know.
Sometimes my blood glucose will be high-ish like maybe 135. Then if I drink some coke, ie pure sugar, it causes my blood glucose to spike and then my insulin kicks in and it drives it back down to 110-120. It’s as if the switch to turn on the insulin wasn’t turned on because there was a slow steady creep up of blood glucose and then it needed to be woken up.
The body sure is strange but one thing I do know is that CGMs are changing things for sure.
I do _not_ have any sort of insulin resistance, and have worn a CGM for a while, a few different times.
(Purchased sorta 'just for fun', to evaluate my own metabolic responsiveness. I wrote about it once: [0])
if I moved around much after eating, my muscles would "soak" lots of glucose from my blood and the insulin change would be relatively low. If it was something sugary and I then went to sleep, it would be a big slow rise for a few hours of blood sugar, then insulin would seem to be released, and it would decline a bit, then flatten, then decline, until it went back to a regular level. When waking up, it would often by around 70 mg/DL, and even if I eat zero carbs (or am full-on fasting!) it would bounce up/around during the day. I appreciated that my body seemed quite capable of doing whatever it needed to do to raise my blood glucose levels. (They always go up when exercising, for instance, even if no food is consumed)
I think everyone would likely find the data interesting, it's so freaky customized to yourself, it cannot help but be interesting. They're expensive, so I think buying even one sensor and wearing it for 2 weeks or however long is worth it.
[0]: https://josh.works/cgm
I'm a fellow type 2, be sure to talk more with your doctor about these points as such comments/questions are of things they should have explained very clearly the first time they said your sugar seemed high.
Type 2 "resistance" is about the quality of the response to high glucose levels, not the complete lack of a response to them. There shouldn't be a long buildup overnight in that scenario for a person without insulin resistance/deficiency and still having a measurable insulin response is normal/expected of all but the worst Type 2 diabetics. Yes, you're still digesting, but in individuals without diabetes the blood sugar peak occurs (and ends) well before digestion is finished because influxes of carbs can still be effectively managed by the insulin alone rather than by the lack of additional carbs to digest. If it were just that one's digestion were a lot slower than a normal persons then it should still result in a lower, but still quickly managed to baseline, peak.
You may well actually be prediabetic though, it just depends on the specific numbers for A1C/average & peaks combined over time and not the presence of a response itself. The recommendations between higher side prediabetic and lower side type II diabetic shouldn't be all that different in the end anyways though.
> You may well actually be prediabetic though, it just depends on the specific numbers for A1C/average & peaks combined over time and not the presence of a response itself.
Just to clarify - A1C is itself an easily-measured proxy for diabetes mellitus, but it's itself a heuristic. There are groups for whom it is known that the "standard" A1C range is actually incorrect, because of confounding factors that affect the A1C measurement but are unrelated to the metabolic dysfunction or general sugar levels.
Your point is correct, though, that what OP is describing is consistent with diabetes, and the actual clinical recommendations for prediabetes and Type II diabetes are often the same, at least in the early stages.
> I find that my blood glucose spikes much later than what is suggested.
I'm a type 1 and my blood glucose can response can vary wildly. Sometimes it can spike quickly (15-30 minutes), other times it can take 2, 3, or even 4 hours. The reverse is also true, in that insulin can sometimes effect me quickly, and sometimes it can take effect 3+ hours later. In general, when having a big meal, I'll take a "fairly large" amount of insulin, but not enough for the entire meal; then I'll take more[1] when my blood starts increasing into 200+ range. Otherwise, I risk it dropping because the insulin was having a "fast" day but the food was "slow".
My endo finds it weird, but we've gone over specific cases of it, including exactly what I ate, when, and what the CGM history for the day looked like. I'm just defective :)
[1] Note I'm not talking about reactively taking more insulin after the "right" amount turned out not to be enough (which is generally a bad idea). Rather, I'm talking about splitting what I know to be about the right amount... into 2 different doses.
Also a T1D and yea, I know what you mean.
Physical activity, mental activity, and sickness can also massively effect dose.
On days I'm physically active I feel like I need less than half the dose of insulin then days I'm caught behind a desk. This said, high impact mental stress that requires lots of thinking can rapidly drop my glucose level.
And that's not counting being ill where my body will pretend like it's insulin resistant for hours then suddenly try to process it all at once.
2 other factors other than physical activity and stress that seem to impact insuline sensitivity:
1. Temperature (room / outside)
2. How well you’re hydrated
Regarding the *time* it takes for eaten Carbs/glucose to enter the bloodstream, it seems that simply standing/being postural often slows down digestion, while sitting/laying down (think of a couch potato watching tv) seems to increase digestion speed.
Have you measured the lag time for your CGM?
Typically the glucose level in the interstitial fluid takes longer to respond to your food intake, and the GCMs measure that instead of the blood glucose level.
Interesting. What’s your a1c level ?
I think you could have a point at around 100-105 baseline but 120 seems too high.
I just found out a friend was diagnosed with diabetes due to a sugar test while their a1c was still in the normal range. We’ve got a long way to go yet with diabetes research.
I mean, it is well known that A1c has a 3 month (usually) lag time assuming no hemolytic states. Could have been just caught early depending on what you mean by sugar test
No they already are in early stage fatty liver. This has been a long time coming and surprises no one. The only surprise is not failing the a1c.
I guess I’m confused, A1c is a measurement used for diabetes, not fatty liver/nafld. Can absolutely be correlated with diabetes and elevated A1c but those are two independent things we seem to be talking about
It's high, like 6.7-6.9. It doesn't make sense to me because that would suggest average blood glucose of over 150, but it's objectively around 120.
There is some genetic variability in the correlation between blood glucose versus HbA1c. The HbA1c test is essentially an integrator of blood glucose level over time but different people have different average red blood cell lifespans.
RBCs life span can vary between humans. That's why hba1c is not always the correct surrogate.
CGMs react slower and can be quite wrong compared to finger pricking.
Slower yes - there is typically a 15-20 min lag time - not two hours.
Do you know if you do(not) have delayed gastric emptying?
Usually T2D is diagnosed based on a fasting glucose test. I also recently worse a CGM because I was borderline T2D and wanted to know how various foods impacted my levels. My peak was about 2 hours after the meal, so about 1.5 hours and change, given the lag of the CGM.
I also found that sweets not only spiked my values but did so over a 6-12 hour period! That was shocking. I have not had dessert for 107 days because of it. I plan to do a fasting test (and A1C) in August to see how many values are. I hope they are much lower, but we'll see.
EDIT: also, popcorn with nothing on it but salt gave me a huge spike.
Popcorn (even plain air popped) is more carbs as a percentage of mass (~70%) than eating a bag of Hershey Kisses (60%) so it's no surprise. I've always found starches to be the bigger culprit of my spikes than sweets since it's relatively easy to overlook and much easier to consume a large quantity without thinking about it.
Odd, I would have expected this to be somewhat the case? Specifically, I would expect your activity leading into the meals, along with your expectations of what you will be doing after, would have some impact? Probably more impact from how hungry you were going into a meal. If you were already sated, for example, I'd expect your body to largely try and push the meal through as fast as it can.
Neat to see what other people's priors are, on this.
> Odd, I would have expected this to be somewhat the case?
I don't think anybody was expecting to be surprised by this study. In practice, most science is pretty boring and rarely breaks expectations. But being unsurprising does not mean it's not worth doing. A lot of studies are simply validating expected outcomes and providing foundational data points for future studies to refer back to. For example, a future study might use this study to justify funding ("as shown in Study 2025.abcd, glucose is highly variable... we propose to further study this by controlling for ... which will help us understand the influence of ..." etc etc).
Ah, great point! I definitely got the impression that this was a surprise, but I can't say why I got that impression. That said, I did not intend this as a criticism of running the study! Quite the contrary, glad they did it. I was only meaning my comment to be contra the sense of surprise I had.
It's a pretty common reaction :)
https://news.ycombinator.com/item?id=43964238
https://news.ycombinator.com/item?id=43964122
https://news.ycombinator.com/item?id=43964428
https://news.ycombinator.com/item?id=43963196
I'm very sensitive to sugar and starch when "at rest". If I wake up first thing in the morning and drink an orange juice or have a bowl of white rice before bed, my blood sugar and mood are out of control. But the same foods mid-day before/after/during intense mental or physical work are very well tolerated if not beneficial. So my priors agree with yours, timing and context absolutely matters.
Non expert here ... My understanding was sleep, stress and many other variables all impact these things significantly (before we even consider food). Having a different context when you then add the same food thus should not result in the same outcome.
The study used 7-day rotating meals, so one would expect some consistency on average in a weekly rhythm, but it also only covered 14 days, so only two data points per meal and participant, if I read this correctly. In that case, I’m not surprised either that there is a lot of variation, also considering that this is a new meal regimen for the participants.
People think medicine is Physics. They really do.
I think I agree with the idea you are saying. That people think you can formulate our body to ignore a lot of the state that it holds.
I'll note that even basic physics has that problem. Try explaining to a 5th grader why a feather would fly in the same arc that a rock will take, if there was no air.
Well, I did not want to say "Mathematics" because it sounded a bit exaggerated. Of course you are right.
> Odd, I would have expected this to be somewhat the case?
Well of course. Studies like this help collect data and quantify the variations.
There's more to the study than the headline or even the simple summary. Knowing the range of variations is important, as well as starting to build a foundation to understand some of the factors that lead to the variations.
I've been working in this space for seven years, this isn't new. There have been big studies from the UK in 2020 [1] and Israel in 2015 [2] about this.
Some of my learnings:
- Don't start your day with a large amount carbs. Have some insulin in your blood before eating that big bowl of oatmeal. Or just go for some yougurt with nuts and seeds.
- The classical order of a three course meal (salad first, then main dish, then dessert) is pretty good in terms of preventing glucose spikes.
- Going for a walk after a meal is great for bringing glucose levels down.
- Eat at least 2h before going to sleep. Having high glucose levels disrupts sleep.
- Alcohol lowers the glucose response of a meal, but is still bad unfortunately.
- Diet Coke works. No spike vs loads of sugar with a real coke.
- Stress can spike glucose like crazy, e.g. being in an interview or during takeoff.
- If you really want to know how you react to some food, keep the circumstances (time of day, sleep, physical activity, stress) similar. There's too much influence beyond just the meal.
I bet that everyone who is wearing a premium smartwatch or an Oura ring now will be using a CGM now and then in the next years.
[1]: https://pubmed.ncbi.nlm.nih.gov/32528151/ [2]: https://pubmed.ncbi.nlm.nih.gov/26590418/
"Additionally, the summarized study did not record the timing of snack and water consumption. This is relevant because the sequence and timing of subsequent meals (or snacks), as well as the food processing and variability, can influence postmeal glucose responses." - probably a factor, hydration status is huge. This result is kind of like 'duh' though.
https://pubmed.ncbi.nlm.nih.gov/28739050/
Hydration, activity and stress levels. Adrenal responses can pause the digestive process to conserve energy for fight or flight. Muscle use also diverts resources, stressed or not. And being dehydrated essentially concentrates everything in your blood, complicating kidney and liver function.
Shouldn't come as a surprise - there are so many factors involved in glucose response to food that it's almost impossible to replicate a glucose response even with a controlled intake. Sleep, activity levels (before, after, during, even days before, etc), stress, hormones, all of these are major factors involved in how glucose is metabolized.
Chewing, acid production, gut microbiome, and possibly more.
I think it will/would take a lot of data to uncover the most important factor(s).
Jessie Inchauspe [1] created a business (she's the "glucose goddess") and wrote books around the different blood sugar responses she and others have to meals, exercise, etc... I don't recall all the ideas, but here are the two I remember: Exercise before and/or after eating smooths out the blood sugar response to food. When eating a meal, it's better to eat high-fiber stuff first, then complex carbs, then protein, then fats, and eat simple carbs last (Or something like that order).
[1] https://en.wikipedia.org/wiki/Jessie_Inchauspé
GLUT-4 activation in the muscles primes them to take up glucose and other nutrients. So air squats or going for a walk before/after a meal can significantly blunt the insulin response, because there's less glucose in your blood stream, because more of it has been ported into muscles.
Is the order salad, baked potato, steak, dessert?
Yeah, I think so, though my recollection is that a baked potato's glycemic index is kinda high. For more authoritative list of details, you could listen to the following podcast with Inchauspé: https://overcast.fm/+AA4G32XuwoA
It would be unfortunate if this article were to discourage folks from trying CGMs out. I am not diabetic but have found a lot of value in using a Stelo CGM sensor. At minimum, it gives me awareness of my current state. I routinely check it before considering desserts for example. Also, I was shocked at the spikes I would get from some food. Maybe an "obvious in retrospect" feedback was drinking a bottle of (non-diet) Coke. It consistently gave me 50 unit spikes (mg/dL). Nothing came close to it, including large meals or meals + desserts. On the other hand, drinking diet coke had absolutely no impact, which was contrary to what I've read before (the body is confused and produces insulin). Another surprising observation is the effect of consuming Feta (a particular kind of Greek cheese) : no matter what else I may have eaten, a piece of feta always leads to a reduction of 10 to 20 points for at least half an hour after consuming it(!) I doubt this is generally true, which speaks to each body reacting differently.
Yep.
My wife is T1D and this is infuriating for her.
She’ll think that she’s cracked it, and then the next day, with the same meal at the same time, her sugar levels go high.
Her words: there is nothing else where you have studied for over 30 years and STILL feel like you know nothing.
It is incredibly demoralising for her sometimes - especially when she’s suffering also from a high/low sugar level. I have the upmost respect for anyone having to do the amount of work, to get to zero (sometimes).
I'm also T1D and yep. That's how it goes. Despite having one of the same two breakfasts pretty much every workday for a decade, it's a crapshoot whether I'll be 300 (very high) when I get to work or 50 (very low) or 110 (good). I just have to adjust when I get in to the office. Most days I deal with it fine, but every once in a while I get fed up with it and want to throw something.
Right there with ya. The most frustrating aspect of the beetus for me is how inconsistent glucose/insulin responses are, drives me absolutely mad
More data can help inform semi closed loop systems to deal with this variability.
For example, AAPS has since version 3.2 dynamic IFS. ( https://androidaps.readthedocs.io/en/latest/DailyLifeWithAap... )
For me this works quite well
Still pissed my insurance stopped covering the Omnipod Dash last year, especially since the 5 is just an artificially locked down dash. Omnipod 5 automated mode is garbage compared to aaps
*ISF (Insulin Sensitivity Factor)
Thank you, I’ll check it out
I like to say that the CGM is, far an away, the biggest life improvement for diabetics since the invention of insulin. And this plays into one of the reasons
- You able to learn better about how your body responds to different foods
- You are more secure about your blood sugar dropping when you don't expect/realize it (sleeping, driving, etc)
- You better able to detect when your body isn't behaving the way you would normally expect it to (the point we're discussing here)
She’s on closed loop and it has certainly changed her life - perimenopause was really difficult but the loop has meant she has much better control. The fact it cuts out when it drops has given her confidence to sleep better etc
How about just not eat carbs at all and have more consistent & lower blood sugar? (this was an interesting case https://www.youtube.com/watch?v=CG8UU7P8FBU)
I do keto diet long term but for other reasons, often the epilepsy version where it's more strict and higher fat.
Keto works well, but I'm concerned about the extreme fat intake. I did manage to lower my A1C from 9s to mid 5s using Keto as a T1D, but eventually settled into a high protein, moderate fat, low-ish carb diet and that has worked pretty well without being hyper-restrictive. Heavily inspired by the late Dr. Richard K. Bernstein.
I target for high ketones (1.5+) & low glucose 4-5mmol (blood test 1 hour after wake up) and higher protein interferes with both of them.
Curious - what's the reason for targeting ketone levels?
Better mental health.
every endocrinologist i've ever had says you need at least some carbohydrates. that said, they're usually not nutritionists, and there's historically been so little focus on food as medicine within western medicine that i think we "officially" know as much about that as we do about diabetes at all (which is a lot, but still surprisingly little).
there are plenty of good reasons we eat carbs (especially complex carbs, plus trying to avoid processed stuff); as a diabetic you just have to prepare ahead of time. it sucks, but honestly, as a 20+ year T1D, not eating cake sucks more than having to plan to eat cake :)
> not eating cake sucks more than having to plan to eat cake :)
But you have never felt the alternative. It's not just eating cake. You have to see the whole video that I linked to understand the differences (they might not all apply). Think a smoker who've never known what's it like without smoking.
dad of an 8 y/o T1D here, and tbh, "how about just ..." is about the worst thing you can say to anyone dealing with this, especially given the subject of this post. things are just never the same. maybe they will be later in adulthood? i feel like i'm in a position to beta test this with my kiddo and i don't like it at all.
I understand that T1D is special, but this is said for every disorder.
There are many levels to keto diet, and believe me, you can't do the epilepsy version even if you'd wanted, it takes time and hard effort, for children it's done inside a hospital.
So assuming you watch everything they eat, you can keep track of GCM, and try very slowly. Like going from 60-30-10 carb-protein-fat to 50-30-20 ratio, do they have better consistent blood sugar ?
There should be protocols to do this for T1D the safest way.
Watch the video.
What about the famous breakfast? One prevalent theory among nutritionists is that the breakfast has huge influence on the insulin responses during the day: a low carb/no sugar, high protein and fat breakfast will not spike your blood glucose level and therefore will lead to a more moderate response to meal intake during the day as opposite to a high carb breakfast. I think the study should control this variable as well.
As a long time diabetic (t1d) with direct relatives also afflicted with this horrible disease, it’s frustrating that we often miss the key factors in insulin resistance and fail to see how they fit in.
The level of serotonin and its interplay with our current level of immune response/inflammation in our body is the single best predictor of insulin resistance. That’s simply because every good thing you can do to lower resistance increases serotonin. Serotonin then decreases inflammation. When you exercise for a long period of time you are not only increasing your immediate availability of serotonin, you are also increasing your daily availability for serotonin.
Obviously it’s a little more complicated but from a big picture standpoint, either something is increasing inflammation in you today and increasing your insulin resistance or (hopefully) the opposite is happening and serotonin is moving your insulin through your blood stream like a mag lev train.
On a side note, I do love cgms.
> if an individual eats the same meal on two occasions but starts with the vegetables on one occasion but not the other, this may lead to different glucose responses to the same meal within the same individual.
> Other behavioral and individual factors are known to influence CGM responses.[22][6] Postmeal physical activity — even as minor as leg fidgeting[23] and walking for a few minutes every hour — reduces postmeal glucose responses.[24] Sleep quality has also been associated with changes in postmeal CGM glucose responses.[25] Emerging studies are also uncovering the relationship between an individual’s gut microbiome and their postmeal glycemic responses.[26][27][28]
So basically it's the impact of the parasympathetic/sympathetic nervous system and the order of food ingestion that could induce different glucose response.
It's interesting because I assumed that beside the nervous system, whatever order we eat our food, it all mixes in the stomach and then start to process. Negating the importance of order but studies implies it does.
It is interesting, but it has been long well "known" that if you eat the more fibrous foods first this effectively slows the digestion of what comes next.
Sounds expected? If you are glycogen depleted, it will cause less of an insulin spike. This sounds like if you fill up the same type of car with 5 gallons of gas. Sometimes it reports full, and sometimes it doesn't. That sounds odd, until you accept that you are never fully sure the gas level you started at.
Leaving aside the science, this paper draws the wrong conclusion. Just because many factors influence BG response does not mean CGMs are somehow invalid as a tool for personalized nutrition. Quite the opposite. A good nutritionist does not work in a vacuum. They know the holistic nature of BG and will tailor their recommendations to the patient based on lifestyle. (I worked on a holistic diabetes management platform)
I like the idea that the proliferation of LLMs become frameworks for understanding that can help advance a field or help curious thinkers come up with novel hypotheses. If the body is something like a trained model, it’s not crazy to think that it will respond to something like fake sugar (eg aspartame) with biological processes that would only be expected for real sugar.
Fascinating. I've always felt that this was true with my own body, yet every medical professional I've spoken to said that I was wrong.
Hate to sound like "that guy" but a majority of medical professionals are running their practice on outdated data or false studies funded by pharma.
There are problems in the US healthcare system with slow adoption of clinical practice guidelines and lack of research funding for non-pharmaceutical treatments. But on an objective scientific basis the recent studies funded by pharma tend to be some of the highest quality work. They have huge budgets and are held to fairly high standards now by the FDA so we tend to see large subject populations, multiple sites, proper controls, and rigorous statistical analysis.
Is this already pretty expected? We already knew that glucose response was highly dependent on time of day, pre or post meal movement, and whether the meal is moderated by other components like fats that can slow down the immediate impact.
80% of the variation due to individual differences OR measurement error.
i'm a type one diabetic. this is uh, not news to me. i can eat the same thing every single day at the same time and see different numbers after every meal. stress, hydration, heat/cold, fatigue, if the wind is blowing, so many other things affect blood sugar.
I wish they would’ve checked insulin response along with the glucose response.
My mother and I recently did an oral glucose tolerance test while wearing CGMs and discovered that in us, the CGM measurements (Dexcom G7) and the reference instrument measurements (whatever Quest Diagnostics uses) virtually _never_ lined up, even when including the G7's advertised error margins. The blood glucometer readings in me didn't line up either. Admittedly this was on the first day using the G7, when it's apparently less accurate, but due to this I've been doubtful of CGM readings in me/her. I'm sure CGMs work on a population level, but for us, on that day? Nope.
My graph: https://i.imgur.com/FzPdH1g.png
Mom's graph: https://i.imgur.com/5DR1G30.png
Discussion: https://reddit.com/r/PeterAttia/comments/1k301o4/my_ogtt_exp...
As the parent of a type 1 diabetic...all I can say is "duh, no shit." Activity levels, emotional state, how much time between eating & bedtime, agem among other things affect glucose response.
Sorry to complain, but what the hell is this horrible AI summary link? The actual study is not open, but surely enough something a bit more precise could be given.
[dead]
[dead]
[dead]
One factor that scientific food experiments don't seem to include is gluttony. Presumably because it is subjective (though nonetheless real for that). Yet if people are motivated by the amount of comfort and/or pleasure they obtain from eating their favourite foods this may have an effect on their physiological response.
I realise the word 'gluttony' is off-putting (for atheists for instance) but I don't know a better one. Perhaps 'over-excitement'?
The symptoms of excitement are of course similar to the fight-or-flight response. Raised heart rate for one. Butterflies in the stomach indicate release of adrenaline, which in turn raises blood glucose -- not from the food being eaten but from the body's stores. So there's a link right there.
How excited you are about a given food will depend not only upon the foodstuff itself but also on your personal taste and personal history. If you have taken pleasure in its consumption many times in the past this may affect your present level of excitement.
'Arousal' is another word that just sprang to mind.
Here's a (Gemini) AI summary that seems to be capturing the main point that the article wants to convey.
``` This page summarizes a study about continuous glucose monitors (CGMs) and how individuals can have highly variable glucose responses even when eating the same meals. The study, which involved 30 participants without diabetes, found a weak-to-moderate correlation between glucose responses to identical meals, with about 80% of the variation attributed to within-participant differences or measurement error.
The article notes that high glucose variability is linked to increased risk of death, greater hunger, and poorer mental health and sleep. While CGMs are beneficial for people with diabetes to prevent hypoglycemic episodes, this study suggests that developing personalized nutrition recommendations for glycemic control may be more complex than previously thought due to inconsistent individual responses. The authors emphasize the need for more reliable dietary assessment and a deeper understanding of the behavioral, dietary, and individual factors that influence glucose responses. ```