Just how accurate is the glucose meter?

Just how accurate is the glucose meter?

Testing one’s blood glucose with meters has been a great boon to managing diabetes.  The glucose testing meters are fairly accurate and they let us know what is going on.  But they are by no means perfect (See Research GDM 25th Mar 2011) and if you are a stickler for detail they can drive you to distraction.

In the general diabetes clinic outside of the pregnancy setting, a blood glucose in the 4 – 6 mmol/L range is pretty good, 10 – 12 mmol/L is getting up there and levels of 15 – 18 mmol/L are clearly very high.  The glucose meters can sort these groups out consistently but within ranges things are much more variable.  However, in pregnancy we are looking for much more precision: if your before breakfast glucose is 4.3 mmol/L it is excellent, if it is 5.8 mmol/L it is too high.  Technically the meter is allowed to be up to 15% off when compared to the laboratory reading but this range is too wide in caring for gestational diabetes.  If someone with gestational diabetes has high glucose numbers they need treatment, if their numbers are high because the meter is reading high they need a new meter or some adjustment for it.  There are three ways we handle this issue

Firstly, we routinely do a laboratory – meter comparison.  We ask the person newly started on a glucose meter to go to the laboratory.   Just before and after the blood sample is taken the person checks their meter glucose (within 5 minutes either side).  We then express the difference in each compared to the lab as a percentage.  Usually the meter reads a bit higher than the lab but if it is within 5% of the lab we can consider the meter acceptable.  If the readings are way off, 20% or more then we give the person a new meter, review their testing techniques and have another meter laboratory comparison.  If I am doing this I usually ask the person to check the accuracy of the meter at three different times.

Second thing we do is to look at the numbers the woman is getting.  If they are all well with in target then if the meter is reading a bit high it just means that things are even better than they look.  If the numbers are very high then a minor issue with the meter is not responsible and we start insulin or medications.  It is more challenging if the numbers are clearly a little high and the meter is a bit higher than the laboratory.  In this situation I use a new meter, test it three times and if the meter is consistently reading higher, eg 10%, then I adjust the targets for this person upwards by the 10%.

Thirdly we look at the whole picture.  If most of the readings in a time period are elevated and not just the odd one then it is more likely a real problem prompting therapy.  If the oral glucose test used in diagnosing the gestational diabetes is well above normal then the threshold for starting therapy is a bit more aggressive and finally if the baby is on the bigger side we will be more inclined to start treatment.

Many expect the meter to be exact, they are not, though overall they are reasonable.   Expensive lab equipment costing $15,000 – 20,000 has some variance in the glucose readings, the meters for what they cost do a good job but in gestational diabetes we have to make an allowance for the fact they are not perfect.

ER