Control solution...who knew? Meter vs Labs

[zauberflote]

TL;DR “is it best to get and use bg meter control solution to better predict lab results?”

Hello all - my first ever HbA1c lab work August 2018 came back US 6.5, with estimated average glucose 140/7.78. In US, if you get the 6.5 twice at any time, you are diabetic. Bought a store brand meter and soon joined up on these boards. Chose LCHF as WOE; lost 20 lbs by not doing anything else, and have watched my BGs creep steadily downward on the meter.
That meter never matched the lab numbers but is always consistent with itself, so, apples to apples. When I started in early Sept, it told me nothing useful until I had a few weeks of data points to calculate the average of (I use an app called Glucose to record results and do my math for me). By that time it was telling me av est BG was around 125 or 130.
New lab work last week. It tells me my HbA1c is US 6.4, est av bg 137/7.6. My meter has been giving results which average out as of today to Est Average Glucose of 112.3/6.23; estimated HbA1c of US 5.54 (the app uses the same conversion chart my dr uses, I’ve looked it up. )
Very frustrated, although I have heard over on the Low Carb forum that this discrepancy is not that uncommon. Bought a different store’s store brand meter and strips today. Tested it together with my existing one before supper. Same drop of blood. The two were within 4/0.222 points of each other. In my mind that difference is statistically insignificant.
Then I ran across “Control Solution, testing meter, for use of” in the instructions. “Consult your place of purchase to find out how to obtain Control Solution”.
So, now I’m curious as to why the pharmacists didn’t try to upsell me to purchase the control solution. I ask the experts here- is my not having tested the meter with control solution (in order to what...calibrate the meter? Teach it my blood?) apt to lie at the bottom of the huge discrepancy between what the lab says and what I’d hoped (from my evidence) it would say? Not to mention that my meter would have to have been reading progressively wrongly lower and lower compared to itself in each previous week or month over 5 months, new batteries and all, which doesn’t make any sense to me at all.
Thanks for any and all comments chris

 

[EllieM]

Remember that there are plenty of blood conditions (eg anaemia) which can skew your hba1c one way or the other. Plus, there's a possible error in any individual hba1c test. If there's a significant difference, you could try asking for a fructosamine test. Personally, I'd believe the meter over the hba1c....

https://diatribe.org/whats-normal-a1c-when-it-misleading

 

[zauberflote]

Hi @EllieM and thanks for that! It was a pretty full lab workup so I know I’m not anemic, but it’s good to have more info to work with. Fructosamine! Wonder if Medicare covers that. I felt the difference between what I got and what I’d expected was pretty significant - the difference between just barely not diabetic, and not even pre-diabetic.
I love that Diatribe! The guy Adam is the most thoughtful person, and so young as well! It was the first info I found, and that’s what started me on knowing my foods.
Thanks again!

 

[zauberflote]

Just an eye-rolling update. I bought another, different store’s brand meter, which consistently reads within the low 15% margin of error of my main meter. So to me that says they agree, broadly. Which makes that unexpected high HbA1c even weirder. I have ordered one of those A1c test kits from Amazon...and plan to have the dr rx me the fancy-schmancy meter he was told to rx to patients by a rep who fed him lunch and promised him the moon if he hit a goal of xxx meters prescribed this year. (My speculation but quite close to truth I expect). This is the dr who told me I could eat more than 2 pieces of fruit in a day. He hasn’t figured out yet that many of us got where we are by happily eating our “9 a day” fruit and veg, only whole grains, not much sugary sweet stuff, and very little fat of any kind. He wasn’t even born yet when fat miraculously became the villain. What DO they teach them in med school these days?! Sigh.

 

[LornaFarrell]

When are you testing? If you test before meals, and say an hour or two afterwards, you may have missed a spike. Depending on how low carb your meal was. For example, yesterday’s lunch put me up to 10 but back down to 7 within an hour. The data you are collecting probably won’t include those spikes, so the average you calculate is lower than the real average.

 

[Listlad]

I always wondered about this. And yes, I maybe am ignorant on such matters. But I see the raised blood sugars after food as being a beautiful Gaussian curve with a bell shaped peak. That being the case, given the way our bodies change and the food we eat changes and the blend changes, how do we/ you know if you have hit the peak right on the nose when you test. Or are we saying every peak arrives at the same time every time regardless? Or is that an inherent weakness in testing with a meter?

Go easy on the response as I am asking the dumb question to try and solve a query in my own mind.

Picking the peak on a moving target has been a task of mine in an earlier life.

 

[LornaFarrell]

I always wondered about this. And yes, I maybe am ignorant on such matters. But I see the raised blood sugars after food as being a beautiful Gaussian curve with a bell shaped peak. That being the case, given the way our bodies change and the food we eat changes and the blend changes, how do we/ you know if you have hit the peak right on the Jose when you test. Or are we saying every peak arrives at the same time every time regardless?

Go easy on the response as I am asking the dumb question to try and solve a query in my own mind.

Picking the peak on a moving target has been a task of mine in an earlier life.

You can’t really know unless you have something like a libre or test every 15 minutes (which isn’t exactly practical!) So many factors make a difference, especially the GI of the foods and if you take insulin when you took it and what type you have. Once or twice I’ve tested at an hour and been in target, only to discover a peak later on!

 

[britishpub]

I must say I can never understand why people think that their meter is incredibly accurate, and a lab tested HbA1c inaccurate.

A blood glucose meter is a cheap piece of kit, using even cheaper bits of cardboard/plastic with a tiny amount of chemical on the end, shoved into them.

Even if the meter left the factory in perfect condition, and accurate to the +/- 15%, the chances of it being transported in optimal conditions, then stored in optimal conditions by the retailer, then kept at home in optimal conditions by the user, with strips manufactured, transported, stored, sold and then stored again in perfect conditions are miniscule.

Add to that a test taken in someones kitchen, from a finger prick on a finger that might or might not be perfectly clean, and yet despite all of that it is considered to be more accurate than a lab test.

In fact, it appears that a few tests a day taken in non-lab conditions outweigh a HbA1c test.

Bonkers.

 

[Mr_Pot]

I must say I can never understand why people think that their meter is incredibly accurate, and a lab tested HbA1c inaccurate.

A blood glucose meter is a cheap piece of kit, using even cheaper bits of cardboard/plastic with a tiny amount of chemical on the end, shoved into them.

Even if the meter left the factory in perfect condition, and accurate to the +/- 15%, the chances of it being transported in optimal conditions, then stored in optimal conditions by the retailer, then kept at home in optimal conditions by the user, with strips manufactured, transported, stored, sold and then stored again in perfect conditions are miniscule.

Add to that a test taken in someones kitchen, from a finger prick on a finger that might or might not be perfectly clean, and yet despite all of that it is considered to be more accurate than a lab test.

In fact, it appears that a few tests a day taken in non-lab conditions outweigh a HbA1c test.

Bonkers.

Agreed, and also people need to remember that HbA1c is not an average but an indication of what the average level has been. Measuring with a blood glucose meter gives a snapshot of glucose level but does not take into account duration. A moderately high level maintained for a long time would probably contribute to more glycated haemoglobin than a very high short spike for example. On the other hand, a prick test during a spike would have a big contribution to a calculated average.

 

[Listlad]

You can’t really know unless you have something like a libre or test every 15 minutes (which isn’t exactly practical!) So many factors make a difference, especially the GI of the foods and if you take insulin when you took it and what type you have. Once or twice I’ve tested at an hour and been in target, only to discover a peak later on!

This is what I suspected. Unless you test with a meter several times over a predetermined working interval.

Yes, that’s why I was thinking of the Libre, which some thought might be overkill.

 

[Listlad]

Agreed, and also people need to remember that HbA1c is not an average but an indication of what the average level has been. Measuring with a blood glucose meter gives a snapshot of glucose level but does not take into account duration. A moderately high level maintained for a long time would probably contribute to more glycated haemoglobin than a very high short spike for example. On the other hand, a prick test during a spike would have a big contribution to a calculated average.

It does depend on the duration of the curve and not just the peak height, I would surmise. AKA “ the area under the curve”.

 

[bulkbiker]

This is what I suspected. Unless you test with a meter several times over a predetermined working interval.

Yes, that’s why I was thinking of the Libre, which some thought might be overkill.

When using a home meter you are looking for trends not absolute values.
If your bloods are usually in the 9's say and they start to go down to the 7's over time then you can see progress..
If pre meal you are 5 and it spikes to 10 after eating you know you've had an excess of carbs..
The Libre is not the be all and end all either as it can seem to react very differently according to the wearer, from my reading of others' feedback. HbA1c's can also be inaccurate and are mostly done at time periods that can be unhelpful so that's why I use as much data as I can get.
I take readings every day (usually 3) and watch out for trends but don't get too hung up on individual readings. Works well for me.

 

[Listlad]

When using a home meter you are looking for trends not absolute values.
If your bloods are usually in the 9's say and they start to go down to the 7's over time then you can see progress..
If pre meal you are 5 and it spikes to 10 after eating you know you've had an excess of carbs..
The Libre is not the be all and end all either as it can seem to react very differently according to the wearer, from my reading of others' feedback. HbA1c's can also be inaccurate and are mostly done at time periods that can be unhelpful so that's why I use as much data as I can get.
I take readings every day (usually 3) and watch out for trends but don't get too hung up on individual readings. Works well for me.

With a test meter. How do you know you have picked out the peak? Or whether you have missed it (too early or too late) ?

What if there are several peaks?

 

[bulkbiker]

With a test meter. How do you know you have picked out the peak? Or whether you have missed it (too early or too late) ?

What if there are several peaks?

Don't really care.. if something has raised my blood sugar enough that I can see a spike I don't care about the absolute value just that I have caught a significant rise.. several peaks is pretty unlikely unless you are eating every hour..

 

[Scott-C]

Here's a couple of my cgm graphs which illustrate some of the points in this thread.

The first one was a nice tight 100% in range day, minimal SD of 0.3, average 4.9.

The next one, though, got of to a really bad start with a drop to 3 and then a bounce up to 12 through a combination of Foot on Floor, underbolusing and shortening pre-bolus timing too much on account of the low, and overtreating.

So that day was a mess, reflected in the rubbish SD of 2.5 and an off the scale PGS of 59.26, but the interesting thing is that the average for the day was 5.1, practically the same as the earlier 4.9!

It just goes to show how misleading averages can be when looked at in isolation, and why other metrics like time in range, standard deviation and the newer PGS, personal glycemic state, need to be looked at for a fuller picture.

 

[Mr_Pot]

Here's a couple of my cgm graphs which illustrate some of the points in this thread.

The first one was a nice tight 100% in range day, minimal SD of 0.3, average 4.9.

The next one, though, got of to a really bad start with a drop to 3 and then a bounce up to 12 through a combination of Foot on Floor, underbolusing and shortening pre-bolus timing too much on account of the low, and overtreating.

So that day was a mess, reflected in the rubbish SD of 2.5 and an off the scale PGS of 59.26, but the interesting thing is that the average for the day was 5.1, practically the same as the earlier 4.9!

It just goes to show how misleading averages can be when looked at in isolation, and why other metrics like time in range, standard deviation and the newer PGS, personal glycemic state, need to be looked at for a fuller picture.

View attachment 31793

View attachment 31794

Excellent illustration of the problem.
I assume the HbA1c shown is calculated - how does it compare with your measured result?

 

[Scott-C]

Excellent illustration of the problem.
I assume the HbA1c shown is calculated - how does it compare with your measured result?

The cgm app I use, xDrip+, has a statistics page. When I choose the 90 day option, which I suspect most closely emulates a blood a1c test, it's usually reasonably close.

My last measured a1c was 28, can't recall now what xDrip+ was predicting but it was in the same general ball-park, around 31 or so.

Of course, being T1, docs used to get on my case about that being too low, but, fortunately, I'm in an NHS area, Lothian, where libre is freely prescribed to any T1 who wants it, so they are now more used to T1s showing safely dropping numbers.

The good thing about the app is that while the docs simply have the a1c to go on, I can just show them the 90 day AGP graph, which overlays 90 days worth of graphs, and provided that's showing a fairly tight band, especially in the 25 to 75 percentiles, it's easy enough to satisfy them that a lower a1c is do-able not by having hypos all the times but just by not going above 8 or 9 too much. I've honestly not had any remotely serious hypos at all for several years now and that's because I can tail them off just by watching the cgm trace.

Having that slew of numbers like TIR, SD etc. makes fending off concerns about a low a1c much easier.

 

[Mr_Pot]

If I have understood @Scott-C 's graph correctly he is getting BG readings at 5 min intervals 24 hours a day and his estimated HbA1c is "reasonably close" to his measured one. The OP and others who think that estimated and actual HbA1c should be close, based on a few finger prick tests a day, are asking a lot.

 

[Listlad]

Here's a couple of my cgm graphs which illustrate some of the points in this thread.

The first one was a nice tight 100% in range day, minimal SD of 0.3, average 4.9.

The next one, though, got of to a really bad start with a drop to 3 and then a bounce up to 12 through a combination of Foot on Floor, underbolusing and shortening pre-bolus timing too much on account of the low, and overtreating.

So that day was a mess, reflected in the rubbish SD of 2.5 and an off the scale PGS of 59.26, but the interesting thing is that the average for the day was 5.1, practically the same as the earlier 4.9!

It just goes to show how misleading averages can be when looked at in isolation, and why other metrics like time in range, standard deviation and the newer PGS, personal glycemic state, need to be looked at for a fuller picture.

View attachment 31793

View attachment 31794

The orange peak in the second curve represents what variable, Scott? Glucose level as a consequence of?

 

[Listlad]

The cgm app I use, xDrip+, has a statistics page. When I choose the 90 day option, which I suspect most closely emulates a blood a1c test, it's usually reasonably close.

My last measured a1c was 28, can't recall now what xDrip+ was predicting but it was in the same general ball-park, around 31 or so.

Of course, being T1, docs used to get on my case about that being too low, but, fortunately, I'm in an NHS area, Lothian, where libre is freely prescribed to any T1 who wants it, so they are now more used to T1s showing safely dropping numbers.

The good thing about the app is that while the docs simply have the a1c to go on, I can just show them the 90 day AGP graph, which overlays 90 days worth of graphs, and provided that's showing a fairly tight band, especially in the 25 to 75 percentiles, it's easy enough to satisfy them that a lower a1c is do-able not by having hypos all the times but just by not going above 8 or 9 too much. I've honestly not had any remotely serious hypos at all for several years now and that's because I can tail them off just by watching the cgm trace.

Having that slew of numbers like TIR, SD etc. makes fending off concerns about a low a1c much easier.

View attachment 31795

View attachment 31796

Okay. Nice resource for you.