Recommendation for CGM

[tanuki]

I am thinking about buying continous glucose monitor, has anyone made some research on it and can recommend me the most reliable one?

[mccoy]

We've discussed about the topic, presently there are 2 main commercial choices, the dexcom and the abbot Freestyle libre, the dexcom is probably more reliable but far more expensive, the Abbot freestyle libre 2 or 3 is very convenient and appears to be reasonably accurate, although it must be checked by finger pricks, notwithstanding the manufacturer declaring that's not necessary.

The value of such monitors anyway lies not in the accuracy, rather in the behaviour of the glucose control system after ingestion of determined foods or meals.

I would always rely on regular fasting blood glucose and HbA1c measurements, but  CGMs give you an immediate glimpse of how foods are managed by the homeostatic system. 

[tanuki]

It seems both options are pretty expensive
I thought it's just one time purchase but actually they have to be constantly repurchased so for example libre cost 110 Pounds for 2 sensors that can last max 28 days
Dexcom costs around 160 pounds per month. It's few times more expensive than using finger pricks though obviously it's only option if I want to monitor glucose during sleep or exercise. According to your post they aren't any more accurate than finger pricks, correct?

[IgorF]

Few own cents on topic

They are rather an interesting complement to pricks but their price together with their subtle nuances makes it e.g. for me "too nice to have". I did not tried libre but I used to have 2 dexcoms thus 4x10 days measurements and I derived from them that I have to drop grains at all and that walking (I hate intensive excercising) for approx 40minutes at higher speed or an hour at normal is always a good idea for me after the dinner. There are more interesting things that could be learned with them (e.g. there could be a wisdom behind bigger morning meal than to shift the biggest to the afternoon etc.) all by all they are not always easy to interpret, they are not a 1 time tool, they have their limits and because of these limits they are "improving algorithmically" and this is tuned for their main goal, not for healthy people's curiosity and so on, the list of things to be taken into account is surpsisingly longer.

Also there were many times raised concerns about peaks in healthy people (if they worse to be addressed), about glucose homeostasis in healthy people on unusual regimens (e.g. CR) and so on.

I personally would like CGMs to drop 3x in price to use them maybe more often. In addition I am doing prick check every morning for 2+ years and this is also useful despite there are also nuances (I am using butterfly pricker and palm in many cases, last months I have to do a second measure from finger because my blood physical parameters seems changed a bit and prick method seems is also tuned to more narrow normal parameters). 

My conclusions from my experience - the longer the things are being used - the more interesting things are to be guessed from observations. Both pricks and CGMs are not as good as people expectes them, their primary goal is to support type 1 diabetics and for this goal they are good enough (and CGMs are a really good invention for them). For those who are eager to investigate and tune own lifestyle they are useful but it is better to expect a longtime "calibration" of them to own situation and to align their +-15% inaccuracy with other tests/data.

Br,

Igor

[IgorF]

one more thing, not connected to CGMs core functionality

 

Last year my old samsung galaxy s7 died during traveling and I bought galaxy s20fe in the nearest shop. After a month with probably the latest update my aktiia bracelet started to have connectivity issues. After a search I discovered that there were changes in btle used by android stack and people with CGMs often also complained and since it is not samsung's but rather main android team's decision there were issues with other android phones and with several CGM types and non-medical bluetooth periphery as well.

Then things seems fixed but a month ago they came back in a bit other flavor. I have no idea if dexcom that I think to buy in the coming months will work for me.

So those using android 13 devices should take into account possible unexpected surprises while planning.

 

Br,

Igor

[tanuki]

thanks for sharing your insights, I'm using finger pricks but their accuracy is pretty dismal, I can get relatively high 5.4 on one finger while other finger shows me nice and low 4.8. I think only affordable option will be buying cgm pack for 1-2 months, observer patterns and adjust lifestyle and diet to minimize peaks.
I'm intrigued what concerns you can have about addressing peaks in healthy people?

[IgorF]

The same 4.8-5.6 is observed with pricks by me. Last months however I have sometimes up to 6 and in these cases I am sometimes do the second check. I also know the context - my blood looks different and recent check confirmed I dropped into CR regimen (in addition to changed sleep, urination time pattern, some hunger "bites" in the stomach late-night and so on). I observed this in the past and also was concerned about these higher figures. Since doing pricks 2am is inconvenient I decided to play with CGMs in the past

After a lot of reading, thinking etc, I decided to monitor regularly with pricks but do not pay so much attention to the particular values, since my pattern looks expected somehow - I do have higher glucose in the morning because I am synthesizing it from early night due to calories limit. I also have worse insulin response (less and delayed) and because of this the average reported by CGM is also higher as I would want it. The reason is probably protein - I have it limited to the (safe?) minimum - 1.5g/kg from plants, getting higher will increase all the hormonal things including insulin and igf1.

No idea if my regimen is not making me harm however. On the other hand I will probably tune the calories up after some time and things will change again.

Regarding handling peaks in healthy people. I read this forum and thought a lot about it, people gave valuable own insights here in the past. I came to unpleasant conclusion that in my case it could be as tough as almost switch my second meal to fats completely and I am not comfortable with this.

Also despite a lot of opinions on the internet that it would be highly wanted to have them low (perhaps it would) there is not much evidence they do real harm if not paired with decades of malnutrition, or unfortunate genes. Also the dynamics of glu-ins changes requires to monitor it with tools not available for the home users and later interpret the data with all the individual context and within math models we probably do not have good enough.

And another angle of view. There is an advice to have complex carbs and while purified sugar seems is a real killer that can turn on an unwanted patterns in many people, the complex ones are not working if expected to be slow. They are working with a delay, but as long as they are recognized the enzymes are produced very fast and they will chunk the chains also very fast. Maybe delaying this process with artificially mixed food so there will be additional physical barriers (e.g. like fat in the icecreams) could help but I am eating almost unprepared stuff so can't choose this way.

Also my biggest hope is that our species are evolved to have carbs peaks, so that is why I am not looking to change my eating patterns fundamentally. I decided to drop grains, have more carbs as fructose (I eat a lot fruits) and this way I have two times less peaks, trying to cope others with walking after meals. Here CGMs definitely helped me to adjust my pattern from "healthy in theory" to more reasonable.

However I have no idea what I will do if I will observe increase in morning glucose paired with lab test for it and bigger figures for insulin, thus - lab controls are needed in any case, the error rates of pricks and CGMs are shifting as long as they are operating not in an expected value range for their designed targets.

 

Last but not least - there are also SNPs (I do have one) associated with higher morning glucose, so the topic has to be assessed from all possible angles, there are people that just do good and dont care, others need to investigate))

Br,

Igor

[tanuki]

As I understand there are two problems with elevated blood glucose - spikes in insulin that hypothetically can lead to diabetes later in life and higher chance for protein to react with glucose to create more AGEs. The first claim hasn't been proved, I'm more worrying about the second issue as we know that is happening among diabetics. So it would make sense to minimize level of glucose just to minimize rate of AGE formation during lifetime as we still don't have any method to undo damage they make.

I was surprised by some readings I took, noticed that after eating sweets and ice cream the spike is rather normal and comes down pretty quickly while eating complex carbs like potatoes gives me a big long lasting spike.  The worst thing to me seems to be drinking alcohol while eating - elevation in glucose can persist until the next day.

Another questions, I looked at dexcom and they offer several products - one, G6, G7 but it seems for us not having diabetes it doesn't matter which we choose, is that right?
 

 

[mccoy]

On 6/27/2023 at 11:58 AM, tanuki said:

I'm intrigued what concerns you can have about addressing peaks in healthy people?

I have not been studying the literature, the rationale here is that 3 parameters are of concern:

1. Fasting glucose, easily monitored by finger pricks, should not go over 100 mg/dL
2. HBa1c, the average BG concentration in 3 months (unreliable if red blood cells have not an average dimension, unreliable if bilirubin is not in the average range), should not be over 5.7%
3. Postprandial glucose, measured 2 hours after a meal, should not be > 140 mg/dL

The above are the ADA guidelines, point #3 is rigorously valid for the OGTT but extended arbitrarily to any meals.

In addition to the above, there is a general 4th point  underlined by some authors that high spikes should be avoided and should be as few as possible, constituting a sustained risk of CVD. (besides being a potential source of hypoglycemic troughs as an adjusting reaction of the homeostatic control system).

[IgorF]

2 hours ago, tanuki said:

I was surprised by some readings I took, noticed that after eating sweets and ice cream the spike is rather normal and comes down pretty quickly while eating complex carbs like potatoes gives me a big long lasting spike.

Yes, that was also a surpise for me and later I discovered even, hm, a kind of science based on this, and then commercial projects based on this approach (e.g. reports on gut microbiota "optimality"). But reading and thinking allowed me to cut out unwanted noise and focus on the way things works and our possibilities for any kind of introspection to get some useful conclusions for a personal life choices.

First of all - the homeostatic regulation for glucose is extremely complicated and unfortunately requires to think about it "dynamically". We are not good at it. Some of us are better in thinking in simple algebraic functions but it is still not enough in the case of homeostatic system, we use models to communicate the observations but those models should always be thought in their contexts and that returns back the complexity and brings additional.

The second important thing - instrumental capacities. We do not have tools to do a good realtime measurement of what happens after a meal that will capture a granular enough data to build a model that could be used as a base to study if the peaks are harmful and in what context. In other words - if it is an absolute value that damage something or if it is just AUC. Is it possible to shift the pattern to a shorter-higher or slower-longer somehow. Even if shifted in time pattern of the response to complex carbs is bad (there are some people telling it) or not.

And also - contexts. It is very unlikely that in the absence of artificially engineered edible substances our longterm evolved body can have such a big issue - it will manifest itself with a lot of unwanted consequences and miriads of generations should eliminate it to the resonable degree. At least - to align with many other limitations we do have.

For simple models that act as "good enough proxy" we can use a simple tool that measures snapshots and based on wide datasets we can predict something and this is a case with fasting glucose and fasting insulin taken from blood. But to understand the things better there is much more unconvenient tool that measures things in portal vein and while continously, still not a realtime or close to the fluctuations known, understanding of which is required to research the topic of peaks.

So either we can have a let's say cheap tool with e.g. isotope marker and constant body scanning for it and mass out the technology to accumulate large enough dataset to cover all known things to be adjusted or we can try to just guess about peaks based on something. This guessing is what we are trying  to do, both scientists with portal vein levels monitoring and curious home users with CGMs.

And here comes the fishy thing for as curious. CGMs are "a proxy on a proxy on a proxy...". They just can't be used that way. Like phone cameras in astrophotography or something like that.

They are measuring too far from the place of interest, with too high delay, just one side of the pair, calculating a value on something dependent a bit on other things and then rearranging the data algorithmically to cut the noise and to fit into the model used by their designers. In some words - they are constructed to report "something expected" instead of physical world event reflexion. And they are also do have traditional instrumental errors, like all the measurement tools.

So, while I still think they are useful for a healthy people but rather in just showing the direction and maybe do some small tunes, but the whole context is so big that it makes virtually impossible to answer even if some small unoptimalities are worse to react. I personally more curious if decreasing the insulin with CR and protein restriction is not making more harm and couldnt' it be better to excercise out the glucose asap by making the sceletal muscles hungry enough. But to answer on this question we also do have to understand the relationship between the glu-ins in portal vein and in the periphery blood and if we are all the same (e.g. do have the same proportion portal/periphery) or not.

And to complicate it a bit more - the longer is fasting time - the longer is uncontrolled glucose release/synthesys. There are limits for the unsulin to react. So individual optimal strategy could be derived only from reliable AUC calculation but without a 24h glu+ins monitoring tooling we can't do it.

Taking this way of thinking I decided to just relax on the topic and do what is possible with the tools and data we have.

Br,

Igor

 

[mccoy]

Postprandial Glucose Spikes, an Important Contributor to Cardiovascular Disease in Diabetes?

Edited July 1, 2023 by mccoy

[mccoy]

Postprandial hyperglycemia and postprandial hypertriglyceridemia in type 2 diabetes

Toru Hiyoshi,¹ Mutsunori Fujiwara,^1,2 and Zemin Yao³

Author information Article notes Copyright and License information Disclaimer

 

Go to:

Abstract

Postprandial glucose level is an independent risk factor for cardiovascular disease that exerts effects greater than glucose levels at fasting state, whereas increase in serum triglyceride level, under both fasting and postprandial conditions, contributes to the development of arteriosclerosis. Insulin resistance is a prevailing cause of abnormalities in postabsorptive excursion of blood glucose and postprandial lipid profile. Excess fat deposition renders a vicious cycle of hyperglycemia and hypertriglyceridemia in the postprandial state, and both of which are contributors to atherosclerotic change of vessels especially in patients with type 2 diabetes mellitus. Several therapeutic approaches for ameliorating each of these abnormalities have been attempted, including various antidiabetic agents or new compounds targeting lipid metabolism.

Keywords: postprandial hyperglycemia, postprandial hypertriglyceridemia, Type 2 diabetes mellitus, atherosclerosis

Go to:

 

[mccoy]

Postprandial glycemic response in a non-diabetic adult population: the effect of nutrients is different between men and women

 

[mccoy]

The graph below illustrates the inter-individual variability in postprandial blood glucose, first very evident thing is the huge variability even inside the men-women groups

 

[tanuki]

Thanks for expanding on it, really informative
I decided to buy one of CGM but seems that reviews are overwhelmingly negative for both systems:
https://uk.trustpilot.com/review/dexcom.com
https://uk.trustpilot.com/review/freestylelibre.co.uk

They mentioned constant errors and readings way off compared with finger pricks to the point it seems unusable
Have you guys tried any of those two systems?

[IgorF]

From four of my dexcom runs I observed an unexplained data loss for half a day in one of them, one run was constantly trying to runaway and I had to calibrate it twice a day with a trick - intentional gradual change of what was really reported by a prick measure if the difference was very high (just read about it from longterm user, e.g. if cgm shows 130 and prick gives 90 - "calibrate" with fake 110 and in half an hour with 90 to make it calm and do not confuse the algorythm). Two runs were ok with data close to the prick measurements and with my intuition.  I think for some of these timewindows I did also a lab test and it was traditionally ~5 points lower than the prick but 1-2h difference in time could easily change the values as it seen from CGMs captures))).

 

ps: I also have 2 prick devices, to crosscalibrate them and corresponding papers about their instrumenal errors

 

pps: my next device will still be dexcom if there will not be a better option

 

UPDATE:

just to give some details on my experimental data

I did 4 runs - 2 on early fall and 2 on spring-summer. Detailed reports are too excessive to be shared, here are the aggregates

num  |  avg/day glu  |  standard deviation  |  prick data morning avg/period - used to calibrate

1       | 96                 |  19                            |  92.4

2       | 98                 |  13                            |  90

3       | 102               |  11                            |  93.2

4       | 97                 |  14                            |  89.9

These deviations were calculated by the software itself, no idea if properly.

During this period (9 months approx) I had 4 glu tests in lab with 89-92 and 3 for insulin with 2.49-2.55. Not the exactly days but relatively consistent and correlates with all other things observed in the timeframe. So at a first glance all seems aligned. But looking on the report timelines themselves the variability of the reported figures, data from other timeframes, curious prick shots after the meal that I do not describe here all make the picture more dynamic and harder to convert into a simple and easy to follow actionable item like those copypasted everywhere (do this and dont do that etc.).

I even used the export to feed into a Snyder's model and got some additional graphs with some colours but to be honest I do not know what they could mean if something at all))

To make things more complicated few more things.

Insulin response is described in the literature as 2 phase process - with a burst of preformed and then fast synthesis as many as needed to cover the requirements.

Glucose in the interstitial fluid where CGM measures it has obviously some "hysteresys" comparing to the values in the serum. So to really draw a useful conclusion of what happens in the particular individuum it needs to be understood what formes the peak - inadequate phase 1? inadequate phase 2? Is what visible in the interstitial fluid what is corresponding to phase 1 or 2? Is it possible to react somehow? Could it work better with more available carbs (e.g. glucose gradient will be more visible and insulin reaction will be faster and stronger == an opposite of well-known recommendations)?

I used to cut out grains at all after some combining of them in a hope that maybe different sources of carbs will not arrive to the "enzymatic reactor" at the same time. Now I am trying to do the same but w/o grains and often w/o pulses, just few different sources of carbs in small amounts in different vegs (e.g. a bit of pumpkin/beet root/carrot/sweet potato/lentils and so on - 2-3 of them in small amounts) but I am not sure it is the way we evolved to eat.

 

Edited July 2, 2023 by IgorF
added some experimental data

[IgorF]

For Snyder's model perhaps I used this https://adaychen.shinyapps.io/shinyspecclust/

Also dexcom's official stagement that +-20 points are not even an instrumental error should be taken into account before thinking about usability of the model results. I think it is created to assess how unwanted the peaks are and mostly reacts on a very high jump up but in reality the portal vein concentrations are several times higher and that is where the most things happen. Then there is a question I have no answer - how representative is peripheral blood for the portal vein data, and adding CGM as "a proxy on a proxy". If I do not remember it wrong - pre-/diabetics do have higher amplitude in the portal vein and then they have higher levels in the peripheral blood. How all these nuances are translated algorithmically to the way CGM operates - no idea, but this is crucial to play with this data seriously. That is why I use CGM as a nice to have, direction showing tool but I do not really care about the absolute values, rather about their concordance with all the other checks.

 

Br,

Igor

Edited July 2, 2023 by IgorF

[mccoy]

I used freestyle libre 3 or 4 times. I posted here the results of my first experiment and it was reasonably accurate, except the upward drift i exhibited from about day 10 to day 14.

Then I used it on my autistic son to see if his explosive behaviours were related to hypoglicemic troughs and apparently they weren't.

y wife also wore it but that time the sensor was evidently faulty since it started showing peaks and troughs which could not be there (and weren't, as verified by the finger prick).

Freestyle libre 2 and 3 are maybe better, #3 has automatic data acquisition like dexcom, but 5 times more frequently. Our recent experience is that our accounts disappeared and were not able to order another sensor, I had to create a new one but I'm waiting to decide which sensor to order, I'll watch the two reviews I posted below, type one talks is a T1D patient who has been using all types of sensors for a few years now.

The ones on trust pilot I think are the customers who received faulty sensors and I agree that customer service is often useless.

 

 

 

Edited July 2, 2023 by mccoy

[mccoy]

Today I had the time to listen to the reviews and I would definitely choose the freestyle libre 3. I'm going to order it soon. Accuracy seems to be satisfactory.

The Dexcom 6 here is a lot more expensive if paying by one's pocket.

Also, we don't need to use the sensor all time, unless we are type one diabetics. Once we know what our response to different foods (or exercise, or sleep) is, we can navigate by sight. What I would do if really interested is to wear it once every two months for example. 

I was able to go from a prediabetic > 100 mg/dL FG two years ago to a 80-90 mg/dL only wearing it a couple of times, it is easy to understand how one's own homeostatic glucose system works.

My next step will be to reintroduce some carbs (fruit, beans, whole grain cereals) and see how the system reacts and behave accordingly. I'm not going to reintroduce honey, dried fruit, watermelons or limitless watery fruits.

[tanuki]

Thanks, I am going to try freestyle libre 2, they are currently offering one sensor for free to new customers - at least in UK.

[IgorF]

CGM for healthy person idea criticized:

https://thedietwars.com/peter-attia-cgm/

I think it worse reading from arguments perspective.

I personally will still play with these things periodically, because I am curious but I strongly agree with the point made by the author:

Quote

This does suggest that classification of patients’ glycemic excursions by CGM might provide more information than traditional metrics–or it might not.

I personally think the peaks could be a bad thing, based on a "flux thinking", in other words - if there is something in the complicated feedback looping that serves mechanically as the rate limiter and it is breakable then we would like to avoid such things but so far I suspect that the only wrong thing for us is purified sugar (in large amounts, few grams seems handled ok), just because we never faced it in the previous adaptation cycles thus we could be vulnerable to it. If this is true - we do not have a widely available tool to investigate it, except radioactively marked molecules and a scanner, this costs a huge amount of money and is probably an unwanted type of study. Other methods are too sluggish and hysteresys-prone to derive any conclusion in such an area of investigation

Br,

Igor

Edited September 5, 2023 by IgorF

[Ron Put]

On 9/5/2023 at 6:04 AM, IgorF said:

CGM for healthy person idea criticized:

https://thedietwars.com/peter-attia-cgm/

I think it worse reading from arguments perspective.

I personally will still play with these things periodically, because I am curious but I strongly agree with the point made by the author:

Yep, I am of the same mind.

I wore a CGM for a month and it was an interesting experience, but basically, it told me that I have moderate, normal spikes that subside as expected. Nothing that A1C hadn't told me, at the end. Plus, without measuring insulin, CGMs tell little of value to healthy people. The keto industry loves them because fat folks can eat their eggs and bacon and feel healthy because their glucose stays level. Especially if they don't go running...

Attia is a salesman, not a doctor or a researcher. Yeah, he has the credentials, but not the practice, since he has been raising money since he left McKinsey. The longer I listened to him and the more I learned about him, the less respect as a reliable source I had for him. To his credit, he mixes a large dose of real information with his promotions and knows how to tell a story, but on the whole, I have come to believe that he does more harm than good.

[IgorF]

Running almost 6 days of libre3 I could share some observations.

First and foremost - oficial sensor availability in a country. If it is being sold not via some vendor-orchestrated channel - there will be an unpleasant journey to get the app installed. It was an unexpected surprise for me due to laziness getting to know it before buying.

People in the internet report that the first 24-36-48 hours the differences could be 15 points usually and up to 50 points during rapid changes.

I observed it that way, except I feel it is still worse with rapid changes but with stable states it is pretty good from the second day. After some reading in the articles and books (there are surprisingly a lot of them already published) and also internet sources my intuitive understanding is that this newer generation of sensors are using trained ML models in addition to the previous generation approaches and the result is assessed as good as calibration can't improve it, thus it is not required anymore. (As I understand the previous generations worked like this: first examine the physical world parameters of some things in the ISF, then based on proved math model calculate the expected difference on the remote area (blood), the same way actiia bracelet is doing with cuff calibration. Now there is no need to align with helper tool, the model is so good that most cases can be derived just from ISF parameters observation).

 

Since I have it higher novadays than in the past I used this run to also check for postprandial triglies as a "poor man's" way to assess wider what fuels are circulating (I assume I do not have a lot of FFAs due to body composition and meals), here is my "calibration" of tool usability:

the first day, 10h from installing and so on

time        libre3(glu)      accu-check(glu)    multicarein(tg)     energy      macros(c/f/p)        comment
05:00      112                106.3                     n/a                       n/a           n/a                         fasting
09:14      148                115.3                     63                       1050         63/25/12               1.5h postprandial, 40min walk
19:02      155                140.5                     75                       1200         70/21/9                 1.5h postprandial, 80min walk

the third day

05:29      104                104.5                     n/a                        n/a           n/a                         fasting
19:06      110                126                        86                         1100         67/23/10               1.5h postprandial, 80min walk

then both morning fasting test and random stable state test are almost the same with finger pricks, libre3 is doing even better that palm pricking, I decided not to test triglies further as I expect no changes, for their comparison this could be used as a guidance https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3693638/

The only thing I am not sure is the peaking rises, I am coping them with going for a walk and have no idea to be honest about their amplitude, intuitively they are similar to those I observed in the past with dexcom6 and prick checks, they are significant only after the evening meal and libre3 reports them higher than dexcom6 reported (e.g. 211 and then once "sensor error" after which it moved to 180 to 140 to 90 and another "sensor error" despite I was sitting on a concert after 20 minutes walk which itself can't completely cope the peak, it is definitely algorithmic failure) but that is the reality we live in now with this data-powered tooling - it tries to report expectations without explanations because they are not easily interpretable in the productized instruments.

 

If my phone will not loose the app or if libre3 will be oficially available in my location I think I could select it as the next tool for the similar experiment (which is unlikely in the coming time since I think I know enough about my way of work and don't care anymore about absolute values).

Br,

Igor

 

 

EDITED TO ADD: since my test is coming to an end I must say - I overestimated the accuracy of the tool previously.

I have a feeling that my libre3 while formally perhaps stays within accuracy test ranges (or almost stays, I am lazy to dig very deep) the problem raises with non-uniformity of distribution of wrong reads, e.g. it stays within 20% or 20mg/l accuracy 85% of the time (or what percentage is required to pass the certification) but most of this time is about stable steady state which is of less interest for non-diabetic single-case researches. When the things are starting to move it will be less accurate and probably will fail to report sensible data in 1/3 of the case or more often. Thus it is enough to guess about the directions but that is almost all I can see in the reports, they are too much spiced with algorithms comparing to dexcom6's more sharp movements.

Generally speaking - I expected something similar but relatively high steady state reported continuously the last few nights when my morning checks moved 10mg/dl down forced me to do some more random tests in a steady state - 25 mg/dl higher than already known as overreporting prick seems not expected.

I had seen many sources mentioning that libre3 interferes with ascorbic acid and AFAIR I must have it at the peak levels due to its mechanics despite I am not supplementing with it I am eating approx 1g of it in the food, thus dumping at least half of it unchanged, so I could have known 10mg/dl points of wrong readings because of it.

But on the other hand I saw that my pricking device (accu-chek) is prone of the same ascorbic acid interference and I saw a paper that other antioxidans are also guilty for increased readings of many pricking meters (https://www.sciencedirect.com/science/article/abs/pii/S000991202100076X)

So with the new set of data and graphs I assume my testing brought me not much useful data, I would prefer things that report more physical world events and interpret them myself than get an algorithmic prediction that leaves me in a constant question - is it inacurate right now?

 

 

Edited September 26, 2023 by IgorF

[IgorF]

calibration study on a 12k cohort, for those who wants to match own data against others'

https://www.mdpi.com/1424-8220/24/3/744