Saturday 3 May 2008

Calibration of glucose devices

The calibration of metering devices for blood glucose has small side effects that I would like to describe here. Blood can be analysed in its pure or filtered form. This has effects on the blood glucose that can be measured within the blood:

I. whole blood: results from picking into the finger or other parts of the body. It contains different cells and liquid where the glucose is soluted.

II. blood plasma: results from a filtering method where cells and liquid are separated. The liquid is the plasma part that is of interest for analysis with laboratory equipment. The higher concentration of the liquid results in a higher concentration of glucose per mmol/L. Thus the blood glucose measured in plasma will be 15% higher. The reason for this approach is to eleminate the influence of the blood cell concentration. Some people have a higher concentration of blood cells than others (reflected by the haematocrit indicator). Furthermore the concentration changes in dependance of nutrition or menstruation.

The blood glucose from the finger is always whole blood. Now the manufacturer can calibrate its glucose device in two ways:

I. whole blood: its results reflect the blood glucose concentration in the finger. Only Roche and Bayer are calibrating their devices for Germany and Austria this way. For other parts of the world their devices are plasma calibrated.

II. blood plasma: its results are comparable with laboratory equipment that is analysing blood plasma. To reach this the measured value in whole blood is multiplied with 1.15. Most devices in the world are calibrated this way.

This means in conclusion that most manufacturers or countries prefer that the devices are comparable with laboratory figures. This makes sense since the users get a positive feedback that their device is still measuring fine.

In addition the plasma calibration has a positive side effect on the glucose control. The glucose values are always exagerated by 15%. This means that insulin users with correction rules are acting earlier to reduce their glucose level. Because of the percentual nature the effect increases the higher the measured glucose level is. The next table gives a comparision of the figures:
whole blood cal.263544536270798897106
blood plasma cal.30405060708090100110120
whole blood cal.115123132141150159167176185194
blood plasma cal.130140150160170180190200210220
whole blood cal.203211220229238247255264273282
blood plasma cal.230240250260270280290300310320
whole blood cal.291300308317326335344352361368
blood plasma cal.330340350360370380390400410420

A diabetic with a blood plasma calibrated device measures 170 mg/dL. With a target value of 120 mg/dL he will apply 2 units of insulin for correction.

The same diabetic with a whole blood calibrated device would measure 150 mg/dL at the same time. With a target value of 120 mg/dL he will apply only 1 unit of insulin for correction.

Thus the use of plasma calibrated devices can lead to tighter glucose control because the values are 15% exagerated. At least it is feasible to follow this assumption. On the other side there is a small likelyhood that some hypoglycamic episodes will not be identified as that.