CGM graphs to view?

[aamcle]

Hi.

I'm T2 and trying out a CGM which has proved to be fascinating, I can see which meals and what foods send me out of the green zone.

I would very much like to see some example graphs, for instance what does a health graph look like or an insulin resistant graph?

So far I have not found published examples, if you know were I might find some I would be grateful if you would post the links.


Many Thanks.

 

[Anna DUK]

Hi @aamcle, welcome to the forum. Everyone's graph is going to look different and will be unique to them. There are so many variables that will make sugar levels rise and fall throughout the day, but typically you will notice different foods impact you more than others. This is based on how your body copes and how much insulin resistance you have.

For people with type 2 diabetes, usually it is considered a healthy range for levels to be between 4-7mmol/L before eating, and less than 8.5mmol/L two hours after eating. Some people are given different figures to aim for based on their personal circumstances.

What sort of results have you been getting? You are able to use the CGM to set your ideal range and from there it will tell you your time in range (TIR) as a percentage.

 

[PerSpinasAdAstra]

Hi.

I'm T2 and trying out a CGM which has proved to be fascinating, I can see which meals and what foods send me out of the green zone.

I would very much like to see some example graphs, for instance what does a health graph look like or an insulin resistant graph?

So far I have not found published examples, if you know were I might find some I would be grateful if you would post the links.


Many Thanks.

It's not actually known what exactly healthy looks like. I did find some sample graphs once in an opinion piece written by a cardiologist where he used graphs to illustrate what he believes to be entirely healthy patterns up to prediabetic, but it was just an opinion piece. It is known that most non-diabetics spend most of their time below 7.8 mmol/L - somewhere between 95% and 99% of the time, but some go well above that level much more often and for longer than average without any proven health issues. No long-term studies have been completed (that I know of) where health outcomes were followed for years after CGM data was gathered, so it's unknown whether those who go well above 7.8 mmol/L regularly and for extended periods of time are on the path to diabetes and/or health complications. Aiming for 8.5 mmol/L 2 hour after eating is a reasonable compromise for a Type 2, fairly close to the patterns associated with a healthy person.

Be aware though that CGMs can be quite inaccurate, particularly at higher blood glucose levels. Right now I'm wearing a Freestyle Libre 2 and a Dexcom One+ at the same time as I'm experimenting to compare the accuracy of the two CGMs along with a set of blood glucose meters. Right now the Dexcom is reading 5.7 mmol/L and the Dexcom graph says I never went above 6.7 after dinner (I went for a walk to keep levels down). The Libre 2 reading right now is 7.2 mmol/L and the graph from that says I went above 8 mmol/L for a while after dinner. Five different finger stick meters testing the same drop of blood tell me the Dexcom is much more accurate for me, probably because I can calibrate that CGM model with a finger stick meter whereas the Libre 2 can't be calibrated, but that isn't to say every Libre 2 will read higher than every Dexcom One+. Putting the graphs into context requires some testing with a finger-stick meter, particularly when levels are stable such as before meals, to give you an idea of how high or low your CGM might be reading. The accuracy can even vary from day to day - I believe I've observed significantly increased inaccuracy due to dehydration for example. CGMs do have to meet basic standards of accuracy but those standards leave a lot of wiggle room. It's probably best not to make long-term decisions about diet based on one or two CGMs by themselves, but to verify them with a good finger-stick meter.

 

[mashedupmatt]

I could share mine but it would give you nightmares - using a CGM is absolutely very useful for recognising what affects your BG - don't get to hung up on the numbers though 🙂

 

[Lisa65]

Here's mine today. Rise in the morning between 6am and 9am which I always get on work days, it started to come back down but then I decided to have some fish and chips at work, about 100g of chips and one Asda battered cod fillet, total about 50-55g carbs. You can see that it shot my levels up briefly but after two hours they were back to normal and continued to ease down through the afternoon. I have long acting insulin but no short acting to cover carbs, and I don't usually eat that many carbs in one go.

 

[Tdm]

Heres my libre graph from when i was new to insulin...much better now, but as now on dexcom its hard to pull off graphs and when you do they helpfully have my name on them...

 

[rebrascora]

Obviously not a normal non diabetic person but here is one of my Libre graphs from a very good day on insulin....

I wish they were all like that!

 

[Woodywoodpecker]

Obviously not a normal non diabetic person but here is one of my Libre graphs from a very good day on insulin....View attachment 33711

I wish they were all like that!

That is good, wish mine was like that

 

[Lisa65]

Here's mine today, without the addition of fish and chips!

 

[rebrascora]

That is good, wish mine was like that

It was a bit of a cheat day as it was during my experiment last October where I stopped using my basal insulin (Levemir) and just used my Fiasp to keep me in range day and night as a very crude sort of approximation of what a pump would do or indeed a healthy pancreas I suppose.
It was a really interesting experiment and one I enjoyed and surprisingly it was not as challenging or burdensome as I expected. Just a couple of units of Fiasp jabbed every couple of hours and a correction before bed to get my levels down to the bottom of range ie about 4 and one during the night to keep me topped up with insulin. I don't think it would have been so successful if I didn't use Fiasp as a couple of units are more or less gone in 2hours, unlike Novo(not so)Rapid which would have been overlapping a lot and just not as effective.
Plus of course I follow a low carb way of eating, so less severe spikes anyway.

 

[guybrush]

It's not actually known what exactly healthy looks like. I did find some sample graphs once in an opinion piece written by a cardiologist where he used graphs to illustrate what he believes to be entirely healthy patterns up to prediabetic, but it was just an opinion piece. It is known that most non-diabetics spend most of their time below 7.8 mmol/L - somewhere between 95% and 99% of the time, but some go well above that level much more often and for longer than average without any proven health issues. No long-term studies have been completed (that I know of) where health outcomes were followed for years after CGM data was gathered, so it's unknown whether those who go well above 7.8 mmol/L regularly and for extended periods of time are on the path to diabetes and/or health complications. Aiming for 8.5 mmol/L 2 hour after eating is a reasonable compromise for a Type 2, fairly close to the patterns associated with a healthy person.

Be aware though that CGMs can be quite inaccurate, particularly at higher blood glucose levels. Right now I'm wearing a Freestyle Libre 2 and a Dexcom One+ at the same time as I'm experimenting to compare the accuracy of the two CGMs along with a set of blood glucose meters. Right now the Dexcom is reading 5.7 mmol/L and the Dexcom graph says I never went above 6.7 after dinner (I went for a walk to keep levels down). The Libre 2 reading right now is 7.2 mmol/L and the graph from that says I went above 8 mmol/L for a while after dinner. Five different finger stick meters testing the same drop of blood tell me the Dexcom is much more accurate for me, probably because I can calibrate that CGM model with a finger stick meter whereas the Libre 2 can't be calibrated, but that isn't to say every Libre 2 will read higher than every Dexcom One+. Putting the graphs into context requires some testing with a finger-stick meter, particularly when levels are stable such as before meals, to give you an idea of how high or low your CGM might be reading. The accuracy can even vary from day to day - I believe I've observed significantly increased inaccuracy due to dehydration for example. CGMs do have to meet basic standards of accuracy but those standards leave a lot of wiggle room. It's probably best not to make long-term decisions about diet based on one or two CGMs by themselves, but to verify them with a good finger-stick meter.

This is why I just couldn't use a CGM. I know you calibrate them but knowing how nearly everyone I've spoken to about them has says how inaccurate they are compared to a figure prick test.

I'd love looking at all the data provides but that would be useless knowing it could be inaccurate

 

[rebrascora]

This is why I just couldn't use a CGM. I know you calibrate them but knowing how nearly everyone I've spoken to about them has says how inaccurate they are compared to a figure prick test.

I'd love looking at all the data provides but that would be useless knowing it could be inaccurate

I find Libre sufficiently accurate to calculate insulin doses and corrections from and Libre does not have a calibration function like Dexcom does. Over the 14 day period of a Libre 2 I might only do 2 or 3 finger pricks. I did one earlier today just out of curiosity when my levels were really stable and in range and my finger prick was 6.5 and the Libre 6.7 Usually I do a couple of checks on the first few days of a new sensor, but after that I go by Libre apart from when hypo and then I go by how I feel and my own personal experience of my recovery from hypos. This is the main time when CGM will be unreliable and we are advised to finger prick, but I am happy not to and just wait for Libre to catch on that I have recovered half and hour to 40 mins later. The other times when you have to be wary of it is if you lie on it during the night and get a "compression low" or if levels are rising fast if I haven't timed my insulin very well and it usually over extrapolates and shows higher that I actually went, but settles back down to the right level if you give it time and the graph usually doesn't show you going as high as the readings at the time actually were. So it might show me at 10.8 and rising fast but later the graph will show I didn't actually go above 10.

I think some Type 2s who are not on insulin get a bit too hung up on the precision and seem to think the decimal place is relevant when finger prick devices are not actually accurate to that decimal place either and as long as you are within a couple of mmols, CGM is an incredibly useful bit of kit and has taken a lot of work and stress out of my diabetes management and many others who are using insulin.

 

[Bruce Stephens]

I'd love looking at all the data provides but that would be useless knowing it could be inaccurate

I find them (usually) good enough. Not exactly matching test strips, but does that mean the CGM is inaccurate or the test strip is? Or (much more likely) a bit of both. Nowadays "inaccurate" ones tend to be obviously broken: they head down and give up.

 

[Proud to be erratic]

This is why I just couldn't use a CGM. I know you calibrate them but knowing how nearly everyone I've spoken to about them has says how inaccurate they are compared to a figure prick test.

I'd love looking at all the data provides but that would be useless knowing it could be inaccurate

I think @guybrush, you are at risk of misunderstanding the discussions about accuracy along with misunderstanding the utility and benefits of each method.

Neither CGM nor finger prick tests are particularly accurate. Each gives an answer to one decimal point but neither is truly that accurate as a moment in time. However when a CGM reading "feels" dubious then we do use fps to get a better confirmation for a potentially suspect individual number.

Even if one had the time or capacity to take, say, 24 fps per day - such readings would still only provide a fairly crude view of that day's overall behaviour. Someone could get a sound understanding of how their body managed the metabolism of 3 meals a day - and still have many fp tests to spare! Whereas a day's CGM graph recording someone's interstitial blood glucose, for every 5 minutes in that day, would reveal far more about that person's metabolic behaviour; day after day.

It all comes down to what someone needs to know - and thus the purpose and benefit of these 2 different testing methods. A person with diabetes who is NOT insulin dependent depends on making wise food choices to assist them in their BG management, alongside using exercise and activity to further help. Using fp testing can, for most people, provide ample visibility of their metabolic behaviour. Structured fp testing, possibly only 8 x daily, can be extremely informative and supportive of someone who is not insulin dependent for their BG management.

Before CGM became more readily available, fp testing was the only way for an insulin dependent individual to manage their BG, but frequently needed a lot more than 8 fps in practice (and that was without needing to comply with DVLA requirements). However this process also needed a strong hypo awareness and forecasting BG trends involved a lot of guesswork. Today, even without really accurate CGM readings BG management hugely benefits from the trend indicators and alerts, that CGM makes possible.

I am extremely confident that the data from a CGM would certainly not be useless to you. But CGM is expensive in relation to fps and for the most part fps would provide ample data to help you manage your D, far more cost effectively. I'm assuming you are not insulin dependent.

Horses for Courses!

 

[nonethewiser]

This is why I just couldn't use a CGM. I know you calibrate them but knowing how nearly everyone I've spoken to about them has says how inaccurate they are compared to a figure prick test.

I'd love looking at all the data provides but that would be useless knowing it could be inaccurate

Remember being told by consultant when Libre was first rolled out that it just doesn't work for everyone, for whatever reason I've no idea but admittedly there has been members on here who have had major issues with device.

Apart from a software issue with libre always found it to be a accurate enough, so like @rebrascora would happily go off libre readings to bolus & such.

Now use Dexcom G6, do odd bg test to check accuracy & if necessary calibrate but tbh it's hardly ever out, even when bg levels are high or low.

Don't mind admitting that these monitors have been life changing, do feel for those people who don't fair well on them though.

 

[everydayupsanddowns]

You might find this study of CGM profiles of people without diabetes interesting @aamcle

Thread 'BG Profiles in Healthy Non-Diabetics'

Apr 11, 2022

Continuous Glucose Monitoring Profiles in Healthy Nondiabetic Participants: A Multicenter Prospective Study

This study provides normative sensor glucose data in a healthy, nondiabetic population of children and adults.

academic.oup.com

A fairly large study from 2019 seeking to map out what "normal" BG profiles look like during the day and overnight, using CGM measures; often see it cited. Perhaps useful resource for people looking to benchmark themselves against "normal" etc.

Units are mainly in American for BG (mg/dl: divide by 18 for mmol/l) and for HbA1c (%; easy to find tools on-line for conversion to mmol/mol).

There is some circularity: the main thing for deciding whether somebody is "non-diabetic" is their HbA1c, so no surprise if BG profiles are...

• Eddy Edson
• Replies: 2
• Forum: Links

• 

 

[aamcle]

Thank you all for taking time to reply, sorry I was slow to get back to the forum.

Aamcle

 

[littlevoice359]

I just found the attached graphic comparing AGP graphs for diabetic and non-diabetic individuals. This graph rolls up 90 days of data to give a graph of blood sugar variability during a 24hr period.

 

[aamcle]

So far so good I was at 52, my estimated a1c is now 41 and that's with me experimenting.

How high does a Belgium Bun send me, what about some pasta, toast and so on.

Aamcle


Ps toast is BAD!

 

[grovesy]

My HBA1C estimate is never right it is always alot lower than my Laboratory Result.