One of my meters is lying

For 5 years now I have been parallel testing using two meters on the same drop of blood. One meter was supplied by my GP and is a Caresense Dual. The other is an SD Codefree meter. I have kept my results in an EXCEL spreadsheet as part of my food and medication log. I no longer test my fasting or breakfast levels and only test my main meal in the evening. This meal is the one I vary, and I follow a fixed eating pattern with minimal snacking during the day. I am a couch potato with zero exercise.

I have always seen that the Codefree persistently reads higher than the dual and it is fairly predictable, so I can accept this as a fixed offset which I average out.

I have just done an analysis of the last 100 days of readings to compare both meters against each other, and looking to see if there is any trending pattern by doing trend analysis.

The first thing I find is that when I examine the differences between the two meters, the SD Code-free and the Dual do track each other with a fixed offset that has zero change over time for all the PRE-Meal readings. This means that at a pseudo fasting level they agree pretty well with a constant difference, confirming my original findings.

When I come to the 2HR PP readings, the difference between the meters increases significantly, by typically 2 mmol/l and up to 2.4 mmol/l extra error on top of the Pre Meal offset. The trend line for this graph shows the error increasing over time up to nearly 3 mmol/l difference.

The 4HR PP readings track about halfway between these two extremes, by also rising but at a lesser rate.

So this indicates that something is affecting the accuracy of the SD Codrfree more than the Dual. It could be just that the values being read are higher and thus eating into the ISO error budget allowance. But normally this higher post-prandial reading is only a couple of mmol/l rise above the PRE so I do not think this is to blame.

I suspect that the different chemistries used in both meters is giving different reactions to something in my blood that changes when I eat a meal. It is not in the PRE readings, so when at homeostasis I find they both track. Something in the blood is changing over time in my body.

I have not had a whole blood panel done for nearly 2 years now due to Covid, so I am missing that vital information. I suspect it may be anemia, or LDL content, or increasing infection that is interfering with my meters. Or maybe the SD is more sensitive to Chylomicrons or SFA. Anemia and infection are probably out of the picture since that would affect the PRE readings too. So I suspect it is something temporary that happens following a meal that subsides gradually over time such that at about 6 hours, homeostasis is regained. And it seems only the SD is sensitive to it,

Anyone else got any ideas about what could be affecting the meter? This is over 100 samples, so is not due to any particular meal. I am generally low carbing and cooking from scratch, but there are some processed meals in the mix.

Richard'63 said:

My first instinct is that it isn't anything in your blood. It is the calibration and gain settings in the meters.

Click to expand...

As an engineer, I disagree. The delta change in error is greater than the delta increase allowed considering the limited subrange of results It would also tend to be a linear effect, and this is not apparent from the data. The Dual is also subject to the same error budget. All the readings at the worst case are less than 9 mmol/l, so the biggest error I should be seeing would be 1.35 mmol/l for a very worst-case meter.

For the record, my average daily reading averaged for the day over that 100 days are as follows
Dual = 6.7 mmol/ equivelent to an HbAic of 40.2
SD Codefree - 8.1 mmol.l equivalent to HbA1c of 49.8.

My last recorded HbA1c was 40.

OK done a bit more research, and found the following reports which are of interest
https://care.diabetesjournals.org/content/26/12/3215
and
https://diabetesed.net/how-accurate-are-glucose-meters/
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2769960/
https://agamatrix.com/blog/blood-glucose-meter-accuracy/


Now I don't think it is blood oxygen levels or medicines. supplements, or temperature or out-of-date strip the issues because these are not associated with eating of a meal. One thing to consider is hydration, but I am fairly consistent in this respect.

The one I am interested in is hematocrit. Now, this is the ratio of red blood cells (RBC) to plasma (the remainder). So an increase in LDL and HDL will affect the accuracy, water content in blood will too, the plasma carries other things like amino acids and mineral compounds so there may be a sensitivity to salt perhaps.

Most meters will accept a range of hematocrit values included n the ISO accuracy budget of 15%, and the deviation has to be quite large to push it outside the tolerance band noted above. Q? would a high fat or high protein diet be enough to temporarily push the boundary? Is the Codefree more susceptible to Hematocrit?

PS to eliminate another possible source of error - I always discard the first drop.

The other thing puzzling me is why there is a time factor showing, with the error growing worse over time. This suggests that something is changing. Again, anemia will affect the RBC, as will B12+Iron+ Vit C+ Vit D3 used to make hemoglobin, Without a full blood test I cannot answer this at the moment, but I am due an annual test in September.

This is an interesting report where a health practice compared a large patient population using both SMBG meters and lab analysis, and then looked at how well they correlated
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6937471/

It seems that most SMBG may or may not be susceptible to Maltose, icodextrin, or galactose. Certainly, I was aware of Maltose being bothersome for some meters, and icodextrin is only available on scrip for use during dialysis, so is not relevant to this thread, but I am surprised about Galactose.
https://pubmed.ncbi.nlm.nih.gov/17723085/

Maltose is a byproduct of starch and definitely does adversely affect SMBG meters some more than others
https://www.myfooddata.com/articles/high-maltose-foods.php
These foods are probably being largely avoided if using a Low Carb Diet. I used to eat sweet potato, but not during the 100 days in my test sample. I know I have had problems with beefburgers and faggots and suspected Maltose poisoning.
The one thing that is important and also has a major effect is MSG.

Galactose is found in milk products and forms part of Lactose. Here is an article on natural sources of Galaxtose which may be a surprise to some here
https://www.myfooddata.com/articles/high-galactose-foods.php

Just found this which has some other foods in this category, such as legumes and seeds
https://pubmed.ncbi.nlm.nih.gov/7671974/

I definitely use dairy products.

Hopeful34 said:

As you haven't had a full blood panel done for nearly 2 years, could you ask for one to be done now, rather than waiting till September? I'd think that was a reasonable request to make to your surgery.

Click to expand...

Tried that - got turned down since it is not a life-threatening situation. Surgery is only doing vaccinations. Actually, I do have a blood test due next week for my heart team, but that is not my GP. Also, they pay for it, not the practice who have washed their hands of my CVD and CVE problems. (I am hooked into the regional hospital for those) My hospital team cannot prescribe normal GP care tests and whole blood tests are classed as diabetes care, not heart or stroke which is kidney check only.

It is all down to Primary Care vs Secondary Care, and budgets. They even come under different CCG's. They cannot step over the line.

OK things getting interesting. Researching Galactose bought up the following study
https://pubmed.ncbi.nlm.nih.gov/11172481/
Seems that galactose gets converted into glucose and glycogen in the liver, and can raise blood sugar levels for up to 8 hours. Up to 20% of ingested galactose seems to convert into blood glucose.

I know from bitter experience that some lentils spike me, and I am totally wiped out by chickpeas. But galactose is found in tomatoes, many fruits, and dairy products such as butter and yogurt.

This is similar to Maltose which I am also sensitive to. It may not be the meter being upset by galactose in the plasma, but direct conversion to sugar. But this would affect both meters in my testing, so might not be the effect I am exploring.

Again this does not explain why I am experiencing an increasing effect over time. That is more likely to be due to changes in my body than changes in the meter, such as anemia, So I will phone the doc tomorrow and try to get an FBC test added to my heart team vanpire session.

This is the one I was looking to find. An official study that varied the hematocrit values and tested many different meters to show how readings of a standard glucose sample were affected the results
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3692232/

It does not cover either of the meters I use, and thus is not representative, but it does demonstrate the significant error that a shift in hematocrit can have on a glucometer.

Next question is - can eating a meal temporarily affect hematocrit, especially if the base value is already being skewed by something like anemia.

I knew someone else would think like me (echo chamber)
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2851212/
They tested people eating a light meal then checked the effects on their blood results. Yes, they did find hematocrit decreased following the meal, and the change is recorded as significant. The change was noted at 2hrPP and was still affected at 4hrPP which is in line with my findings.

When waiting for a bus then two come along at the same time
Another study was done in Poland
https://academic.oup.com/labmed/article/48/4/357/4159450
Similar conclusions, but giving actual results.

Ronancastled said:

My own experience with that over reading SD Codefree

https://www.diabetes.co.uk/forum/th...codefree-wild-variations.175381/#post-2280492

I put it down to bad strips. One batch could be good, next would read 1.5 higher for no reason

Click to expand...

I noted this in my early days of using the SD, and also the XCEED and NEO meters that also suffered batch variations, but recently I have not seen much in the way of variation when changing either SD or Dual tubs. I am still trying to find out why my version of EXCEL on Win 10 no longer supports the different methods for doing trendlines. It seems to only have the linear one. So I will experiment with segmenting the plots to see how it varies over the course of a month or so.

For those curious about Galactose, I found the following article. I cannot vouch for the source but it may help identify a form of ranking based on a 200 calorie serving basis. Not sure how useful it will be, but gives a relative comparison I suppose
https://nutritiondata.self.com/foods-000014000000000000000.html