I have noticed a phenomenon....

[Oldvatr]

Yes, I AM talking Dawn Phenomenon, but not as I used to know it.

In the past, here on the forum. I have explained it as being a liver dump that the body triggers to get the blood sugar levels up to face the morning. But I am coming to the conclusion that there may be something else going on and I may be oversimplifying it, Indeed there may be room for some more experimentation, and it may lead to better understanding as to why some bgl meters are appearing to be better than others

For 4 years now I have been parallel testing using 2 bgl meters on the same drop of blood. I do between 6 and 8 tests a day, and more if re-testing if a misread is suspected. I log all my results in a spreadsheet that I use to determine trends and use to adjust my medications and diet. Simples, no?

No! I never get the two meters to agree but I do get a measure of any underlying difference in calibration between them exists. For example I use an SD Codefree meter as I self fund it and compare against the GP supplied meters. I have found (and reported on the forum) that I get a constant shift between them in that the SD seems on average to read 1,2 mmol/l higher than any other meter I have used, be it XCEED, NEO, ACCUCHEK, or the GP surgery one, I have thus assumed it to be due to a calibration problem since others report a similar offset in their SD. Whatever the cause, I now rely on this offset being there, and so I declare a misred if either my SD reading difference is less than this average or greater than 3 mmol/l from the other meter, and do a retest on both - generally this gives a clue to misread if one or other retests with a changed value, but the other repeats the original reading. If both differ significantly on the retest, then maybe the retest is wrong.

Recently I was in hospital.and they messed around with my meds, which led to significant changes in blood composition, such as RBC and Heamatocrit and cholesterol levels. In fact I got quite jumbled up and had organ failures also contributing to the mash up. Proper car crash time. While they played God with me, I continued trying to maintain my bgl control and failed miserably. Hospital food, diuretics, kidneys refusing to cooperate whatever, I had terrible glucose readings in the range 12 to 32+ mmol/l and higher. But the thing I noted was that the two meters no longer tracked each other like they used to. They would differ by up to 5 or 6 mmol.l. Not only that, but a retest also confirmed this wild offset, and on one occasion I used both hands and did about 4 extra retests. Nope, the SD was closer to Pluto than my Carefree, and both of these were wildly out compared to the Hospital meter (a new Abbott one with bells and whistles attached special to the NHS) So something in my blood was changing, and each meter responded to it differently, Now I know that meters are sensitive to haematocrit levels, but not by much according to the ISO standard which makes provision for this, But maybe there is something else going on in my blood that is also affecting their accuracy.

Since being discharged from Hospital, I have resumed my quest to control my bgl, as any self respecting T2D should aim to do, and my spreadsheet is now showing a constant 1.9 mmol/l offset between my two meters, whereas before it was 1.2 Now I do not have 4 years of data yet, but it does seem to be a new phenomenon that is repeatable. Now my meters have not changed, but my blood composition has, and I am currently suffering low sodium levels due to the blood thinners I am on.

What has this got to do with DP? Well, yesterday I had a full set of daily readings that were all in the 5's on one meter and in the 6's and 7's on the other meter. This morning when I awoke, both my readings were raised by 2 mmol/l by presumably DP. But the difference between the meters was 3 mmol/l which is unusual. After 2 retests, the two meters gave repeat results that confirmed this offset was real, and that the meters were reacting to my bgl differently. So I have to ask, is the DP also having different blood composition in the morning that skews the results as well as just added sugars? is this a common problem that anyone else has noticed? An offset of 3 at the notional levels means that at least one of the meters is not meeting the ISO spec that allows +/- 0.8 ,mmol.l difference at that range so they should be tracking within 1.6 of each other, not 3.

 

[Mr_Pot]

I have noticed, just by looking at the specs, not by owning them, that meters have different ranges of haemocrit levels. How well they cope with extremes, or indeed how they do it, I don't know, but this might be a clue to your variation.

 

[Oldvatr]

I have noticed, just by looking at the specs, not by owning them, that meters have different ranges of haemocrit levels. How well they cope with extremes, or indeed how they do it, I don't know, but this might be a clue to your variation.

Possible, but from the research I did on the ISO standards seemed to only show a slight inaccuracy at the extremes due to haematocrit. I think the sensitivity to maltose and maltodextrin was more significant, and this forced meter manufacturers to take steps to reduce this effect - it was mainly a chemistry problem with the sensor technology since not all meters suffered from it. I think haematocrit has to exceed the 80% and 20% limits to affect meters to not meet the ISO and this is not a very common problem unless in extreme anemia, which I think would have been noticed this morning if it had been the only cause.

 

[becca59]

Very interesting! Have been saying for ages that discrepancies in meters is quite high. It is why I stopped worrying about Libre readings /unless very high/low) which tend to be in the middle of the 2 meters I possess. It always makes me roll my eyes when people get hung up over .2 of a mmol Re the dawn phenomenon, it can be very visual on a Libre when you have an upwards trending arrow.

 

[Oldvatr]

I think I have found a couple of clues to assist in resolving this observation of mine.

BGL meters that use test strips actually measure electrical conductivity in the blood sample. The test strip uses an enzyme film that reacts to the presence of glucose by reducing its resistivity, and this is what the meter is calibrated to measure. With no strip inserted, the meter sees very high impedance so switches off. Insertion of the strip into the meter will present a 'known' impedance that the meter can measure and recognise as a valid strip insertion. The meter can also detect a used strip since that has a different resistance, as does a partial misread due to inadequate sample size.

So the meter measures the drop in resistance as the enzyme gets converted by the blood sample.

But other things in the blood also alter conductivity. We know pure water is a non conductor, but adding impurities such as salt (sodium chloride) will affect the result. This is why we need to wash our fingers before testing since our fingertips do exude sweat (aka salty water) and natural oils too, but can also be contaminated by contact with other impurities such as food.

Haematocrit is a measure of how much haemoglobin is present in the blood. Now haemoglobin contains heme iron in small quantities and this is also a conductor. This may explain the sensitivity to haematocrit.

But there is another conductor or two in the blood drop that may also be affecting the readings. One of these is sodium, and others are magnesium and potassium, all classed as electrolytes because they conduct charge. The one I noticed recently is sodium since my blood composition recently suffered a severe drop in level to below the danger mark.

Unfortunately I was not in nerdy scientist mode. being somewhat ill at the time, so I did not keep records. But I can trace some erratic bgl levels as possibly being correlated to said sodium levels.
My sodium dropped to 121 mmol/l and my bgl at the time was winging around in the 20's. I was also seeing more than 3 mmol/l difference between my meter readings, where under more sane conditions the difference is usually around 1.2 quite consistently (for me at least).

Obviously I took umbrage at this loss of vital substances, and took steps to rectify the situation. I drank Marmite, ate Bovril on Toast, and upped my salt sprinkles. I suffered a dilemma where the GP said stop the diuretics, and the hospital consultant said put them back in. I am on half measures now which is a compromise, but seems to be working.

Now as my sodium levels improved, my bgl levels also dropped, Not only that but the difference between them narrowed as well. Then I noticed that if I drank a Marmite before my evening blood sample, then the difference was consistently lower than if I drank a coffee instead. Initially I blamed the lactose in the milk, until I noticed that Marmite is actually quite high in carbs, and should have been giving me higher readings, not lower. Recently my sodium levels tested at 136, which is Normal. My sugars have dropped back down to the 4's and 5's that I had been used to pre op, and the diffference between the meters is less than 0,9 mmol/l so they are tracking much better now,

So my theory is that sodium intake may be affecting our meter accuracy, and it may be strip technology dependant too, so different meters will be affected differently. This is an effect that so far is not covered by the ISO tests, and may explain why the SD Codefree is often reported as being reading high.

I also wonder about if this plays into the Dawn Phenomenonon that I mentioned in my OP. What if sodium gets filtered out during he night in readiness for the morning pee session, then maybe this is artificially raising the resistance (ie, less conduction) of the blood so we get a morning bump? Just a thought.

I need some engineering/ scientific support in this, so I am tagging @Jim Lahey and @Prem51 in the hope of extending the research a bit further. Please invite anyone else you know who may be ble to join in this discussion.

 

[Member496333]

Thanks for the tag but this isn’t something I spend much time thinking about. As far as testing goes I just wash my wands under plain running water (no soap), dry with paper kitchen towel (not a potentially soiled hand towel) and prick away

I’m very sure there are numerous factors that can potentially affect test results, but personally I just use a meter system that a believe to be reliable, and use the same hand washing prep each time. I only care about fasting trends these days, so absolute accuracy isn’t paramount to me.

 

[TriciaWs]

Some interesting info here:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2769960/

I have wondered how much my morning dehydration affects my readings.

 

[Rachox]

Ha, @TriciaWs , beat me to it! Reading down the whole thread I was thinking that maybe not drinking over night leading to mild dehydration might affect the fasting reading and bam Tricia had already mentioned it!

 

[Member496333]

I would think that your level of hydration affects your actual blood glucose concentration rather than the accuracy of the measurement per se, but that is a guess.

 

[Solaire]

Yes, I AM talking Dawn Phenomenon, but not as I used to know it.

In the past, here on the forum. I have explained it as being a liver dump that the body triggers to get the blood sugar levels up to face the morning. But I am coming to the conclusion that there may be something else going on and I may be oversimplifying it, Indeed there may be room for some more experimentation, and it may lead to better understanding as to why some bgl meters are appearing to be better than others

For 4 years now I have been parallel testing using 2 bgl meters on the same drop of blood. I do between 6 and 8 tests a day, and more if re-testing if a misread is suspected. I log all my results in a spreadsheet that I use to determine trends and use to adjust my medications and diet. Simples, no?

No! I never get the two meters to agree but I do get a measure of any underlying difference in calibration between them exists. For example I use an SD Codefree meter as I self fund it and compare against the GP supplied meters. I have found (and reported on the forum) that I get a constant shift between them in that the SD seems on average to read 1,2 mmol/l higher than any other meter I have used, be it XCEED, NEO, ACCUCHEK, or the GP surgery one, I have thus assumed it to be due to a calibration problem since others report a similar offset in their SD. Whatever the cause, I now rely on this offset being there, and so I declare a misred if either my SD reading difference is less than this average or greater than 3 mmol/l from the other meter, and do a retest on both - generally this gives a clue to misread if one or other retests with a changed value, but the other repeats the original reading. If both differ significantly on the retest, then maybe the retest is wrong.

Recently I was in hospital.and they messed around with my meds, which led to significant changes in blood composition, such as RBC and Heamatocrit and cholesterol levels. In fact I got quite jumbled up and had organ failures also contributing to the mash up. Proper car crash time. While they played God with me, I continued trying to maintain my bgl control and failed miserably. Hospital food, diuretics, kidneys refusing to cooperate whatever, I had terrible glucose readings in the range 12 to 32+ mmol/l and higher. But the thing I noted was that the two meters no longer tracked each other like they used to. They would differ by up to 5 or 6 mmol.l. Not only that, but a retest also confirmed this wild offset, and on one occasion I used both hands and did about 4 extra retests. Nope, the SD was closer to Pluto than my Carefree, and both of these were wildly out compared to the Hospital meter (a new Abbott one with bells and whistles attached special to the NHS) So something in my blood was changing, and each meter responded to it differently, Now I know that meters are sensitive to haematocrit levels, but not by much according to the ISO standard which makes provision for this, But maybe there is something else going on in my blood that is also affecting their accuracy.

Since being discharged from Hospital, I have resumed my quest to control my bgl, as any self respecting T2D should aim to do, and my spreadsheet is now showing a constant 1.9 mmol/l offset between my two meters, whereas before it was 1.2 Now I do not have 4 years of data yet, but it does seem to be a new phenomenon that is repeatable. Now my meters have not changed, but my blood composition has, and I am currently suffering low sodium levels due to the blood thinners I am on.

What has this got to do with DP? Well, yesterday I had a full set of daily readings that were all in the 5's on one meter and in the 6's and 7's on the other meter. This morning when I awoke, both my readings were raised by 2 mmol/l by presumably DP. But the difference between the meters was 3 mmol/l which is unusual. After 2 retests, the two meters gave repeat results that confirmed this offset was real, and that the meters were reacting to my bgl differently. So I have to ask, is the DP also having different blood composition in the morning that skews the results as well as just added sugars? is this a common problem that anyone else has noticed? An offset of 3 at the notional levels means that at least one of the meters is not meeting the ISO spec that allows +/- 0.8 ,mmol.l difference at that range so they should be tracking within 1.6 of each other, not 3.

 

[Bluetit1802]

I would think that your level of hydration affects your actual blood glucose concentration rather than the accuracy of the measurement per se, but that is a guess.

I am guessing the same thing.

 

[Mr_Pot]

I seem to remember that the strips have electrodes, in addition to the main ones that measure the glucose, to enable the meter to compensate for some other factors. I can't find a reference at the moment so I don't know what factors. I did find this if you are interested and have plenty of time to spare.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2769960/

 

[Bluetit1802]

I seem to remember that the strips have electrodes, in addition to the main ones that measure the glucose, to enable the meter to compensate for some other factors. I can't find a reference at the moment so I don't know what factors. I did find this if you are interested and have plenty of time to spare.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2769960/

One of the factors is it changes the whole blood (capillary) result to a plasma (venous) equivalent

 

[Oldvatr]

I seem to remember that the strips have electrodes, in addition to the main ones that measure the glucose, to enable the meter to compensate for some other factors. I can't find a reference at the moment so I don't know what factors. I did find this if you are interested and have plenty of time to spare.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2769960/

My SD Codefree has 3 terminals on the strip, and my Caresense Dual has 4. My Freestyle meters had 3 .

Because the effect of the resistance change is small and analogue, I suspect that the three terminals are used to form two arms in a wheatstone bridge so one arm is made up of electrolyte that does not see blood drop to act as the reference and provide the temperature compensation and electrolyte variations over each lot (autocalibration) needed for accuracy, There may be some other use for these terminals using frequency signals for impedance checks, not just d.c. resistance and the two methods could be used together.