[tapusen]

Dear friends,
                  I was doing some calibration with different concentration solution of D-glucose in water. I'm using one of the portable glucometers available in the market. But to my utter surprise I found that on successive measurement using the same glucose solution the meter gives significantly different results. Like an example I made a 400mg/dL solution of D-glucose in water. I took 5 readings, and the meter gave 226mg/dL, 254mg/dL, 272mg/dL, and keeping that solution in open overnight, on the next I found the reading 554mg/dL, 562mg/dL! Now these readings are observed in the order I've written. Now I'm not from chemistry background,an electronics engineer, Can someone tell me some reasons why this is happening. I understand this cannot be a meter or strip problem. I'm absolutely clueless. Someone please provide me some solution.

With Regards,
Tapabrata Sen

[scwilson]

Well, I don't do a whole lot of biochem or bio analyses, but I do work extensively with instrumentation of every sort. Accordingly, anyone with more relevant experience is welcome to correct me, but I think I have a good idea of what the primary problem might be.

When performing analyses, we need to be aware of what analytical chemists term as "the matrix." The matrix is the environment in which your analyte is present. In this case, your analyte is D-glucose. Your matrix is water--tap water, I assume. In other words, you're trying to measure D-glucose in an aqueous matrix. Well, the problem with this is that these meters are designed to measure glucose in a matrix VERY different from water. In fact, as a chemist, blood is one of the most complex matrices I can readily think of. That means that these strips have been specifically engineered to selectively analyze glucose IN A MATRIX CHARACTERISTIC OF BLOOD. However, your test solution is glucose in water. I don't know the specifics of these test strips, but I'm sure that they're designed and calibrated to accommodate salt concentrations that are typical of blood (which would effect the meter's ability to properly measure the voltage and produce a reading). These are concentrations of salts that are probably quite dissimilar to that present in tap water. Additionally, I'm sure that these test strips and the meter have been calibrated and designed to handle the presence of various proteins in blood. Undoubtedly, there are probably dozens of proteins which could and would interfere with the meter's ability to selectively measure glucose, which means teams of engineers and billions of dollars went into designing a test strip which could "correct" for these interferences or mitigate their impact in some way, shape, or form. And, I'm sure that the meter had to be calibrated upon its manufacture, because the test strips STILL weren't perfect, but a good calibration was able to get things about right.

On top of ALL of that, I'm sure the meter was calibrated for and the test strips were designed for the pH typical of blood, for the temperature typical of a blood sample, for the viscosity typical of a blood sample, for the base voltage of blood in the absence of glucose, etc. The point I'm trying to make is that there are countless variables that these devices and strips need to accommodate, and very few of these compounding variables will be present in a matrix as simple as tap water. These matrices are far too different.

I will say that it's a little peculiar that the meter gave different values for the same solution on different days. But, even so, it's not unlikely, given the matrix.

Now, if you are determined to test the accuracy of your device, it is possible, but it would require you to obtain at least a few milliliters of blood. For the purposes of using convenient and round numbers, let's say you got a hold of 1 dL of blood and you measure it and the meter reads 100 mg/dL. If you added in another 100 mg of D-glucose to this dL of blood and mixed it thoroughly, you could expect the new concentration to be 200 mg/dL. You could test this spiked blood sample to see if the meter, indeed, reads out the predicted concentration. However, obtaining enough blood to make this practical would be difficult, especially since clotting could change the matrix enough to make the readings inaccurate. I'm not really sure that this is realistic, but it would be the only "home-brewed" calibration method that I can think of.

Lastly, you could measure your D-glucose in water solution. Then, add in some glucose and measure it again. The readings won't be accurate--again, the matrix is too far removed from the design specifications of the meter and strips--but I WOULD expect the meter to report higher readings to reflect the increase in glucose that you added. That's probably the only "proof" you could get that the meter is doing its job. Your meter probably reported different concentrations on different days due to any number of problems associated with the matrix, but notice that measurements made in the same sitting were fairly stable, so measuring your glucose solution and then spiking in glucose should probably at least confirm that the meter is capable of detecting changes in glucose concentration and is doing its job. Just try and do your "test" in the same sitting, so that we can try and control any time-dependent variables.

[tapusen]

Thanks Mr. scwilson for your excellent well explained reply.
Actually the first 3 readings I've given were take on same sitting and on the next day observed the last two readings. As you can also see that every reading is giving a value greater than the previous reading. This is a striking observation. Yes I understand that the fact that these meters are calibrated for blood glucose, yet if it had given more or less the same reading it could have been calibrated accordingly. And yes I'll try out your suggestion for testing the accuracy of the meter. It's an excellent thought. Thanking you again sir.