Is Interstitial glucose measured by CGMs different from the finger prick-measured glucose?

[mysorian]

Which is more stable in its instantaneous value?
Isn't CGM by far the more useful measurement?
Is there a good correlation between the two measurements?
How does the needle in the CGM sample the fluid? Does it not clog or get blocked with time?

 

[KennyA]

Short answer - yes it's different.

I don't know what you mean by "more stable in its instantaneous value". The CGM return seems to lag blood glucose by about 15-20 minutes. Glucose gets into the bloodstream fairly quickly but then takes longer to get to interstitial fluids.

My experience with CGM is that it showed patterns well (when the sensors worked) but wasn't reliable at individual point measurements - it usually read one or two mmol/l off from the fingerprick, even allowing for lag. The best I got was consistent error/deviation. Where the CGM scored for me was showing what happened when I wouldn't normally be testing blood - when asleep, for example. Having got that info I don't need to repeat the experiment until/unless something changes.

From that, it's obvious that my experience shows that the two systems do not always correlate well. The CGM shows patterns and flows: the fingerprick gives you a more accurate (hopefully?) BG snapshot.

As I'm a diet only T2 predominantly interested in my blood glucose values, with no current hypo/hyper issues, I pay more attention to post-eating fingerprick values and haven't used a CGM for 18 months or so, maybe longer. I don't actually need a CGM at all, and if offered them via the NHS wouldn't take it up.

I think there are others who would benefit more from a CGM, and for example I would think there is value to any T2 newly diagnosed using one for a while for the "ups and downs" information.

I have no idea how the tech side of the sensor works. The microfilament goes through the skin but shouldn't draw blood. Of the three periods of use, one failed after a few days and one never worked at all. Both were replaced by Abbot, but I don't know why they failed.

 

[Deleted member 527103]

The CGM return seems to lag blood glucose by about 15-20 minutes.

Interstitial fluid which is what CGMs read lags 15 minutes or so behind blood glucose.
However, the view that the values reported on CGMs is behind is rather old fashioned (although, sadly, it is repeated again and again on fora and, sometimes by HCP).
The algorithm used by CGMs to convert interstitial fluid readings to blood goose readings, takes this lag into consideration and “predicts“ the current reading.
Therefore, unless the current trend changes direction in the last 15 minutes (e.g. recovering from a hypo), the reported CGM value is very close to the calculated value now.

This has been mentioned many times on this forum.
Unfortunately, it looks as if it will need to be repeated many more times.

 

[KennyA]

Interstitial fluid which is what CGMs read lags 15 minutes or so behind blood glucose.
However, the view that the values reported on CGMs is behind is rather old fashioned (although, sadly, it is repeated again and again on fora and, sometimes by HCP).
The algorithm used by CGMs to convert interstitial fluid readings to blood goose readings, takes this lag into consideration and “predicts“ the current reading.
Therefore, unless the current trend changes direction in the last 15 minutes (e.g. recovering from a hypo), the reported CGM value is very close to the calculated value now.

This has been mentioned many times on this forum.
Unfortunately, it looks as if it will need to be repeated many more times.

Thank you, obviously my experience was a few years ago, before the algorithms were tweaked. My readings did lag.

 

[LittleGreyCat]

Accuracy:
Venous blood draw>finger prick peripheral vascular>interstitial fluid.
Anyway, that is my understanding.

In my personal experience Libre 2 sensors are variable.
Most read high compared to a finger prick.
A few read low.
The predicted HbA1c is generally significantly lower than the venous blood draw results.

An interesting thought, though.
Is the interstitial fluid the medium through which the cells take up glucose from the blood?
In which case, is it more realistic at a cellular level than the glucose circulating in the blood stream?
That is, it shows what the individual cell sees when it takes up glucose.

Answers on a postcard.

 

[Deleted member 527103]

It is important to remember the limitations of all CGMs (not just Libre).
These have been mentioned many times but in brief
- bedding in time. Some bodies react to the alien object in their arm so readings can take a couple of days to settle.
- compression lows. False low readings if you apply pressure to the sensors.
- unbendable. Seems obvious but if you place the sensor on a part of the arm that bends (e.g. edge of muscle), it will peel,off.
- inaccurate at extreme values. CGMs are designed to be most accurate at “normal” levels which are between about 4 and 8mmol/l. Outside of this range, they will be inaccurate. For example, it is common for the readings to be out by 2 or more mmol/l when in double figures.

If these limitations are understood, CGMs can be accurate enough for insulin dosing.
The advice is to test with finger pricks if the CGM does not match how you feel.
Yes, finger pricks are considered most accurate. If you are using a CGM to replace finger pricks, you are missing the point. The value is seeing what happens between pricks. Yes, they can significantly reduce the number of finger pricks but not all.

 

[Bill_St]

Consider the individual cells in your brain. Where do they get their glucose? They do not have blood capillaries. They are bathed in interstitial fluid. How do you “Feel” your glucose level?
So often we feel a hypo before it shows in BGM results, whether from blood or CGM. Because we detect he change in interstitial.

Consider the tested levels of Lag for Libre. They used to be 7 minutes +-3 minutes but the new algorithms have cut this. They are now 2.4 minutes +-4 minutes. I’m just going from memory here so these may be just rough figure.
The point is the 2.4 minutes -4 minutes becomes a negative value. Libre is ahead of blood at certain times! Lag becomes negative!
While the average Lag may be 2.4minutes the actual reading can vary between 6.4 minutes and -2 minutes. And all of these values, BGM and Libre are only correct for 95% of the time.

So most people end up confused. I even heard a DSN telling a group that the lag for Libre is 2 minutes 4 seconds!
The issue really - Does it matter?
Precise measurement of these values is actually impossible - even in a hospital laboratory. The British Medical Journal recently published a strong warning to medical professionals that all medical measurement is imprecise and taking it as absolutely correct could lead to incorrect diagnosis. Measurements change and comparisons between two values can completely mislead.

We mistakenly think that we need ”accurate” glucose values to compute our insulin needs. But what about that insulin itself. It is Not of constant efficacy. It might be measured and tested at the factory gate but never before it is pushed into your arm. It can vary considerably depending upon the distribution and storage and indeed how we keep it. Some US testing gave truly shocking results - but it simply is not tested here.

In short, far too many people get hung up over numbers. Treatment is based upon individual estimation - guesswork.

 

[Bill_St]

Consider the individual cells in your brain. Where do they get their glucose? They do not have blood capillaries. They are bathed in interstitial fluid. How do you “Feel” your glucose level?
So often we feel a hypo before it shows in BGM results, whether from blood or CGM. Because we detect he change in interstitial.

Consider the tested levels of Lag for Libre. They used to be 7 minutes +-3 minutes but the new algorithms have cut this. They are now 2.4 minutes +-4 minutes. I’m just going from memory here so these may be just rough figure.
The point is the 2.4 minutes -4 minutes becomes a negative value. Libre is ahead of blood at certain times! Lag becomes negative!
While the average Lag may be 2.4minutes the actual reading can vary between 6.4 minutes and -2 minutes. And all of these values, BGM and Libre are only correct for 95% of the time.

So most people end up confused. I even heard a DSN telling a group that the lag for Libre is 2 minutes 4 seconds!
The issue really - Does it matter?
Precise measurement of these values is actually impossible - even in a hospital laboratory. The British Medical Journal recently published a strong warning to medical professionals that all medical measurement is imprecise and taking it as absolutely correct could lead to incorrect diagnosis. Measurements change and comparisons between two values can completely mislead.

We mistakenly think that we need ”accurate” glucose values to compute our insulin needs. But what about that insulin itself. It is Not of constant efficacy. It might be measured and tested at the factory gate but never before it is pushed into your arm. It can vary considerably depending upon the distribution and storage and indeed how we keep it. Some US testing gave truly shocking results - but it simply is not tested here.

In short, far too many people get hung up over numbers. Treatment is based upon individual estimation - guesswork.

If you wish to read further on this see :

Your results may vary: the imprecision of medical measurements

### What you need to know Clinicians and patients need to interpret a multitude of medical measurements. These are often central to monitoring health and informed decision making. Has the serum cholesterol concentration come down since starting a statin? Have vitamin D levels gone up? Is the...

www.bmj.com

 

[Bill_St]

Consider the individual cells in your brain. Where do they get their glucose? They do not have blood capillaries. They are bathed in interstitial fluid. How do you “Feel” your glucose level?
So often we feel a hypo before it shows in BGM results, whether from blood or CGM. Because we detect he change in interstitial.

Consider the tested levels of Lag for Libre. They used to be 7 minutes +-3 minutes but the new algorithms have cut this. They are now 2.4 minutes +-4 minutes. I’m just going from memory here so these may be just rough figure.
The point is the 2.4 minutes -4 minutes becomes a negative value. Libre is ahead of blood at certain times! Lag becomes negative!
While the average Lag may be 2.4minutes the actual reading can vary between 6.4 minutes and -2 minutes. And all of these values, BGM and Libre are only correct for 95% of the time.

So most people end up confused. I even heard a DSN telling a group that the lag for Libre is 2 minutes 4 seconds!
The issue really - Does it matter?
Precise measurement of these values is actually impossible - even in a hospital laboratory. The British Medical Journal recently published a strong warning to medical professionals that all medical measurement is imprecise and taking it as absolutely correct could lead to incorrect diagnosis. Measurements change and comparisons between two values can completely mislead.

We mistakenly think that we need ”accurate” glucose values to compute our insulin needs. But what about that insulin itself. It is Not of constant efficacy. It might be measured and tested at the factory gate but never before it is pushed into your arm. It can vary considerably depending upon the distribution and storage and indeed how we keep it. Some US testing gave truly shocking results - but it simply is not tested here.

In short, far too many people get hung up over numbers. Treatment is based upon individual estimation - guesswork.

Perhaps the best way to compare CGM ”accuracy” is to compare two sensors on the same person. Or one sensor to the best blood measurement and plot them on a chart. If perfect, they will plot onto a single 45 degree line. Differences show up as scatter to each side of that line. Here is the line for Libre. Note a few dots well off the line but most are on it or nearby. This is a chart of hundreds of measurement and not odd results from a single comparison. (Remember the 95% sample above )

 

[Bill_St]

Consider the individual cells in your brain. Where do they get their glucose? They do not have blood capillaries. They are bathed in interstitial fluid. How do you “Feel” your glucose level?
So often we feel a hypo before it shows in BGM results, whether from blood or CGM. Because we detect he change in interstitial.

Consider the tested levels of Lag for Libre. They used to be 7 minutes +-3 minutes but the new algorithms have cut this. They are now 2.4 minutes +-4 minutes. I’m just going from memory here so these may be just rough figure.
The point is the 2.4 minutes -4 minutes becomes a negative value. Libre is ahead of blood at certain times! Lag becomes negative!
While the average Lag may be 2.4minutes the actual reading can vary between 6.4 minutes and -2 minutes. And all of these values, BGM and Libre are only correct for 95% of the time.

So most people end up confused. I even heard a DSN telling a group that the lag for Libre is 2 minutes 4 seconds!
The issue really - Does it matter?
Precise measurement of these values is actually impossible - even in a hospital laboratory. The British Medical Journal recently published a strong warning to medical professionals that all medical measurement is imprecise and taking it as absolutely correct could lead to incorrect diagnosis. Measurements change and comparisons between two values can completely mislead.

We mistakenly think that we need ”accurate” glucose values to compute our insulin needs. But what about that insulin itself. It is Not of constant efficacy. It might be measured and tested at the factory gate but never before it is pushed into your arm. It can vary considerably depending upon the distribution and storage and indeed how we keep it. Some US testing gave truly shocking results - but it simply is not tested here.

In short, far too many people get hung up over numbers. Treatment is based upon individual estimation - guesswork.

Finally Insulin concentration - look at the graphs in this medical article - USA but would ours be better? https://journals.sagepub.com/doi/10.1177/1932296817747292?icid=int.sj-full-text.similar-articles.2

 

[Jaylee]

Meeeh, personal experience.. using a 3rd party app with L2. It’s spot on with the meter within appropriate range & even when high.
Hypos, however.. but there is a long lag it could read up to 45 minutes low.? Where as the meter will put me back in the “zone” after treatment. & that can be a few minutes after I feel OK.

My theory is the brain’s cognitive takes priority as it waves the “white flag” first, the interstitial fluid level representation in my arm has to wait in line… Hence the “lag.”

My meter is my guide after a low. But then I’m still fully hypo aware… I find the algorithm of the L2 pretty good at giving me a “heads up” with the prediction. This lowering the severity of a hypo, even when treated..

 

[EllieM]

bedding in time. Some bodies react to the alien object in their arm so readings can take a couple of days to settle.

And some people can react badly to a sensor and never get a good reading from it (the reason why I had to stop using the original libre 1).

 

[Jaylee]

And some people can react badly to a sensor and never get a good reading from it (the reason why I had to stop using the original libre 1).

Interestingly, due to a bodge on my script, despite ordering a week in advance to process at my local & the chemist remembering me order the previous Friday.
Then trying to “score” a funded one at my local Asda. None in stock in readiness for when I go away for 7 days.
& weighing up after sorting it out the script won’t be ready till midweek… (whilst I’m out of town..)

I’m going to slap on a Dexy G6 for 10 days still boxed from a trial I never got round to, & give it a “road trip.”