Cal your meter with a lab standard?

1

The longer I have diabetes the more frustrated I get with the loose standards for BG meters. I think it’s something like +/- 20% for BGs >= 75 mg/dl and +/- 15 mg/dl for BGs <= 70 mg/dl. And we’re making everyday dosing decisions with a medication than can and does kill!



We all calculate insulin correction doses based on these finger-sticks and I’m sure I’ve inadvertantly added a unit or two of unneeded insulin and then had to cram down some carbs to head off a low when things went south. No wonder I have a hard time time losing weight!



Now I use my meter to calibrate my DexCom CGMS. And if the meter is off then the Dex will also be off. I regularly upload my CGM to my computer and then look at the numbers. My last Dex 90 day report predicted an A1c of 6.1% but the number came in at 6.7%. Wtf??!



So I start poking around with Google and find this FDA meeting report on a conference that took place last year during March. The conference topic was about the blood glucose meters that we’re all friendly with. One of the docs said something that caught my eye:


"And what we've noticed is that a CGM calibrated with a YSI, for example, is tremendously more accurate than a CGM that's calibrated even with a good meter value. And by -- I mean, you're always have the problem of the delay, of course, and it's still slightly different in plasma or glucose, what you're measuring. But a perfect calibration makes all the difference with the use of a CGM in a closed loop system. And so I really want to emphasize the need for accuracy when we talk about calibration, and not necessarily insulin dosing."

So this started me thinking. In a perfect world, wouldn't it be nice to sit down with a few dozen meters and test them against a reliable lab standard, like the YSI (Yellow Springs Instruments 2300 Stat Plus), and then take the "pick of the litter." I haven't confirmed this notion yet, but I think that the meters are better at keeping their results in a tight set of data-points (precision) than they are at hitting the target (accuracy).

So what I'm trying to say is that I believe that many meters may be, say 10% high but will generally always be about 10% high. I'm reaching here. Do you think this is true?

I'm wondering if it might be worth pursuing trying to identify one meter, out of a crowd of its siblings, that would actually give +/- 10% results for values >= 75 mg/dl and +/- 5 mg/dl for <= 75 mg/dl. You know how one shirt or pair of pants in the same size group will just fit better than the rest?

Now I know that there are many other variables that can alter an accurate reading such as temperature, humidity, dirty hands, dehydration, and other interferants.

But if I could get a better number to calibrate my CGM, it would make my metabolic life so much easier. I wondered about having a YSI machine at home but they look like a high-maintenance item. They're fairly old technology and they contain membranes and various containers of fluids. Unless you insure a good environment and timely service you will compromise the accuracy. I haven't checked on the price but I wouldn't be surprised if it cost $10k; and then you'd have to purchase a list of supplies. Not a realistic option.

It might be nice, though, to use a lab standard like the YSI to identify one sweet meter that you could rely on.

Anyone else in search of a more accurate meter?







2

Always in search of a better accuracy meter. The problem is most “studies” done are by the meter companies themselves. You can guess how much I trust those studies. My “Spidey Sense” tells me the Accu-Chek Aviva but I use One Touch products because I am usually wearing the Ping or Medtronic pump and both have meters that use One Touch and make my life just a tad bit easier. I think all in all they ALL give poor results (some may be poor results with highs and some may be poor results with lows). My understanding is the FDA is going to insist on better accuracy. Why couldn’t they take this fight up BEFORE they started messing with CGM and Pump companies that by and large were repackaging the same technology that has been around for years? To me it seems like you should fix the problem that EVERY PWD is using to base treatment decisions on not a category that a relatively small percentage of PWD are using.

3

MossDog - The FDA doc that I linked to contains some discussion of different levels of need for accuracy. They talked about hospital settings and T1 diabetics that sought tight control as two groups that would benefit from tighter accuracy requirements. While some T2 diabetics that don’t use insulin and still have significant pancreatic function may not need as great an accuracy.

4

I guess my point is there are plenty of Type 1’s who use MDI who could also benefit from greater accuracy vs a less needful (as I see it) group, as far as FDA intervention into the devices, like those using pump and CGMS. I also see a good number of Type 2’s on this site who do not use insulin that still use there glucometers to judge what they can or cannot eat and/or base their success vs failure on the numbers that, for the most part, can’t be relied upon to be accurate like even those on oral medications would like them to be. Regardless, if the FDA demands greater accuracy it will benefit all PWD. If the FDA dictated that generic medications were only 15-20% the same as name brand medications you can bet there would be a very public outrage!

BTW- I have looked into A1c devices similar to the YSI device as above and you are right- they are not cheap.

5

Some here compare brand X meter with brand Y meter and despair at the differences.

I work at it a little differently: when I get my regular bloodwork done at the lab, I make sure to test my bg with my meter(s) before the blood draw, and after the blood draw, then I compare when the lab results come back.

It has never been more than +/-10 away, and is usually +/-5.

That’s remarkably good and more than good enough for my purposes.

I get the feeling, those who despair that they got “92” and “99” on two different meters and don’t know which number is right and think it’s important, have bg’s that are far easier to control than mine. Or far far better bg control than mine. I personally regard it as a complete success if I have a day where all my numbers are between 70 and 140 (I have such days but they do not happen every day!)

I’ve had T1 long enough that I overlap with the urine testing days. One of my fears is that some congressperson will hear about some diabetic who isn’t happy with bg meter accuracy, and write some law that will get the FDA to ban the current bg meters or raise the accuracy requirements so much that bg testing becomes unaffordable (not as if it’s cheap today but imagine if it cost $10 per test?). I’m never ever gonna go back to not having bg testing. Some might complain about the inaccuracy but its a zillion times better than urine testing and - very literally - it’s something I need for my everyday life.

6

I do the same thing Tim whenever I get a blood draw. It’s never more than +/- 5 - I use the Accu-chek Aviva.

7

Tim - You’re absolutely right that current meter accuracy is way better than urine testing. It looks like your meters have compared well with your lab results. I agree, to be within +/- 10 is reasonable accuracy to me. I also agree that increasing the accuracy with a giant leap in cost would not be a good trade-off. I for on, however, would not mind paying a bit more for a meter that I could use to calibrate my CGM.



What prompted my writing here was an experience I had last week. I took part in a research study that sampled my blood glucose every fifteen minutes for two hours. I used the opportunity to check the accuracy of my meter. I was stunned to find that my meter was 20% to 49% high!



In the worst case the YSI standard measured my blood at 102 and my meter measured 152. Since 152 is not an unusual number, I would often add insulin to bring it down. I would, of course, consider my insulin-on-board before taking a correction dose.



The numbers I saw were alarming and unacceptable to me. I’m working with the meter manufacturer to replace this meter. They’re very skeptical about my report.



There’s not much to protect meter users aside from using the control solution to test. How often do most users use the control solution? Your practice of checking your meters against the lab is good. I have done that at times but not on a regular basis. I must do this every time now. I wonder how many diabetic’s actually do this?

8

Marie - Thanks for your info on this. Your meter compares very favorably to your lab results. I plan to do this check every time now.

9

I looked up the YSI gizmo and you have to email them for the price but it also weighs 25 lbs!

10

I wonder how many diabetic’s actually do this?

In my case, doing “fasting bloodwork” is an unusual challenge. If I wake up and don’t eat breakfast my bg will begin steadily rising… often going up 200 in just a few hours due to dawn phenomenon. So the morning of my labwork is a time that I’m unusually focused on my bg’s trying to balance rising bg with insulin all at the same time making sure I don’t go hypo.

I’ve been doing pretty good with this balancing act the past couple of years and usually my bg is rather stable around the time of the blood draw. I have had a few wacky things happen… bg’s in the 40’s, or bg’s in the 200’s, when my balancing act wasn’t quite so good, but I’ve avoided that for a while now.

In other words, the fasting bloodwork is a major hassle for me! But I end up with a plethora of bg meter measurements before and after the lab work.

11

Yeah - I don’t think that the YSI is practical for us. But I’m still interested in the possibility of screening a batch of meters and picking out the one that’s most accurate. I don’t think this will ever happen but I’d sure like to get my hands on a meter that’s dependably and verifiably accurate, one that’s periodically checked against a reliable standard. I guess the only way that we can approximate that scenario is to do what Marie and Tim suggested above: check your BG at the same time that the lab draws blood.

12

Tim - Have ever considered adjusting your morning basal rates in an attempt to get flatline in-target numbers when no breakfast is eaten? Perhaps your current breakfast dose is actually doing some of the work intended for basal insulin.

13

This month is diatribe they describe a new meter from Lifescan, the Verio that is supposed to be more accurate

http://www.diatribe.us/home.php

It’s in the featured article.

14

10 years ago my 60 second meter was always 5% less that the lab value. No the 5 second meters are totally unpredictable as to higher or lower that the lab. If you get within + or - 10% it is a miracle. This is something were we have really gone backwards.

15

I use the AccuChek Aviva and it is usually pretty close to the lab results. I just got back my results from the last test and per the lab, I was 166 and my meter had 163 so I was pretty happy with that.
Last fall, I needed to get a new meter & asked on the Dexcom group what were good meters. Several people suggested the Wavesense Presto so I got one of those. The Presto seemed to be running way higher than my Aviva (like 20-30 points). One time when I tested the Aviva had me really low and the Presto in the 70s – I had sweat running off me so guess which one I believed. I don’t think the Presto picked up lows at all. When I went to the lab, the lab had me at 84 and the Presto had 110 – it wasn’t even within the 20% range. I decided to stick with my Aviva.

16

Thanks for the link to diaTribe, Annabella. I checked out the new Lifescan meter. It says that it can get 95% of the readings >= 75 mg/dl within +/- 15% and within 15 mg/dl for readings less than 75 mg/dl. This is better than the FDA requirement but I’m hoping for a jump to +/- 10% and +/- 10 mg/dl soon!

17

Kelly - Thanks for reporting your experience on the Aviva. I like the close correlation with your lab numbers. I’ll have to check into the Aviva.

I do know that many of the AccuCheck BG testing products were susceptible to reporting false high numbers when someone was taking certain drugs, like dialysis fluids, that contained non-glucose sugars such as maltose. The FDA reported that some people died based on bad insulin dosing decisions from testing with some meters. All of the reported deaths occurred in hospitals.

18

Anthony - I think we users should be able to periodically check our meters against some lab standard. We depend on them to make critical dosing decisions about insulin, a medication that can be fatal. We don’t even know if our meters conform to the loose +/- 20% standard!

19

Which is a very good point. You are unlikely to be getting these meds unless you are in the hospital. I also called my local dialysis unit and, at least in my area, they do not use that peritoneal dialysate that contains Maltose.

I really like the MultiClix lancing device that comes with the Aviva. The Avivia is one of the last major brand coding meters on the market though.

20

I have always had pretty good luck with AccuChek meters. They do seem to pick up lows better than some other meters. The only thing I don’t like is after I got a new computer with Windows 7, I had to upgrade to their 360 software. There is a backup & restore feature and you think you are backing up, but you don’t! You can export your data but the import feature doesn’t work. I think that is pretty stupid. I don’t like losing my data and have had severl computer crashes that I had to restore data in the past.

I also read about the Maltose problems. I know Freestyle strips also had that problem because there was a doctor that used to post here that was concerned about that with the Navigator. After learning that, I now try to pay more attention to those little inserts - who would ever think you need to read warnings on test strips!