Apropos to today’s newsflash…
Diabetes advocate and author Riva Greenberg has been on a “meter accuracy kick” lately — researching the heck out of this controversial topic. Very timely considering I’ve been seeing loads of expensive TV ads for Accu-Chek’s new Nano meter, claiming that it’s “23% more accurate” (!)
Riva recently published a piece at the Huffington Post on why meter accuracy is both less, and more, critical than you might think. Truth is, she tells us, meter accuracy is only one part of a much larger story.
A Guest Post by Riva Greenberg
After being lucky enough to receive an iBGStar meter from Sanofi the day before its launch, I ran a few comparison tests between it 
and the Bayer Contour USB, which I’d been using the past two years, and discovered that the iBGStar consistently gave me a reading 20-25 points higher.
So I took out all my meters. There were several, (Sanofi studies show most people use 4 meters on average) and I even ordered two new free meters from FreeStyle. I checked my blood sugar several times on my collection of 7 meters (some think I was a little obsessed) and saw it was rare when two meters gave me the same number!
Given that I feel like my meter is my lifeline, I wanted to find out how meters work and why different meters give different results.
I talked with a number of Chief Medical Officers, MDs and Medical Safety Officers at several meter manufacturers and I’m going to tell you what I learned in layman’s terms.
To better understand the science behind meter and strip technology, you can google “meter accuracy” for white papers and posts that would delight even the geekiest engineer. To better know how accurate your own meter is (in percentage terms), you can “check the package insert that comes with the strips and look online at prescribing information,” says Shawna Gvazdauskas, Head of Devices at Sanofi Diabetes U.S.
Home vs. Hospital Testing
Most home meters measure glucose in so-called “whole blood” (blood as it comes out of our body). Whole blood consists of a liquid, called plasma, and cells, mainly red cells. The percentage of red cells is called the hematocrit. The standard reference lab test measures glucose in plasma (about half to two thirds of the volume of blood).
Home meters are calibrated to give results as though they are measuring glucose in plasma only (called “plasma-equivalent” results). That said, to some degree we’re already on two different playing fields. Second, laboratory tests eliminate virtually all variation, except for manufacturing variation, from their testing.
What that means, according to Dr. Alan Cariski, Worldwide Medical Affairs & Medical Safety Officer at LifeScan, is that hospital standards are much more exacting than testing at home because in hospitals you have: trained technicians, a controlled environment for temperature and humidity, constant maintenance of the machine that performs the test, with checking and refining of the machine’s calibration several times a day, and a much larger sample of blood (5 ml) that’s analyzed for 60 seconds or more, and at much greater expense.
Cariski says lab tests generally come within about plus/minus 4% of a perfect reading. Andreas Stuhr, Medical Director North America at Roche Diagnostics, more or less confirmed that but added: “Even the lab standard is off 5-7% at times, so plus/minus 20% of the hospital lab test (current ISO standard) isn’t as big a leap as we think.”
So Many Variables
There are many factors that go into producing a blood glucose reading at home, and so there’s a lot of room for variability in the final result.
Contour, Verio, and iBGStar meters
Some factors have to do with the meter, some with the strip and some, with us PWDs.
In fact, the biggest contributor to inaccuracy is the strips. Here’s the process as simply as I can put it: glucose interacts with an enzyme on the strip, releasing electrons. Another agent on the strip, called the “mediator,” turns these electrons into an electrical current. The greater the glucose concentration, the greater the current. That current then speeds through the strip. Finally, an algorithm (formula) in the meter converts the current into a concentration of glucose. And voila! You get a number.
But there’s a long list of factors that affect meter/strip accuracy:
- Meter calibration, coding, enzymes and mathematical algorithms (all different in different meters)
- Variable enzymes in strips
- Mediator oxidation and strip freshness/age
- Strips differ somewhat, lot to lot, with somewhat different precision ranges for each lot
- Strips differ in well size (the space in a strip that holds the blood)
- Interfering substances in one’s blood from medications (something as simple as Tylenol), and every manufacturer’s nightmare, hematocrit — that percentage of red blood cells in blood, which can interfere with the electrical current
- Environmental conditions: temperature, climate, altitude
- Lack of meter maintenance
- User error – forgetting to code the meter or coding it incorrectly; not washing hands before testing (there may be some sugar residue on your fingers or sweat on hands); leaving strips exposed to air too long; or using expired strips
As for testing twice on the same meter a minute apart and getting a different number (yes, I tried that too!), I learned that the first drop of blood you squeeze out of your finger is not the same as the very next drop of blood. It may contain more interstitial fluid (the solution that surrounds our cells), which can give a lower reading.
What’s Realistic
Dr. Cariski says the best we can probably expect a meter to ever produce is plus/minus 8% accuracy, and we may never reach that due to all the variables listed above. Also, if we want portable meters that require no calibration or coding, use very small blood samples, with less than 10 second read-outs, the next standard the FDA is currently considering (within plus/minus 15% of a hospital lab test) will bring us close to the theoretical limits of what’s possible regarding accuracy. Unfortunately, the technology needed to get meters closer to a lovely plus/minus 5% just doesn’t exist — for now.
As mentioned, the FDA is currently evaluating tighter ISO standards, down to plus/minus 15% for glucose concentrations equal to or greater than 100 mg/dL, and plus/minus 15 mg/dL for glucose concentrations less than 100 mg/dL. The ruling is expected late this year or early 2013.
So the bar is being raised (or lowered in this care), and the vendors are racing to meet it. Some have already hit the plus/minus 15% accuracy threshold, with products like OneTouch’s Verio IQ, the new Nano from Roche and the Contour EZ from Bayer.
AgaMatrix Up & Up Meter (Poor choice of name?)
Will they have gone far enough? Dr. Barry Ginsberg, topic expert and President of Diabetes Technology Consultants, says for people with Type 2 diabetes who don’t use blood-lowering medications, and check their blood sugar only occasionally to see how they’re doing, our current meter accuracy is fine. For Type 2s on blood-glucose lowering meds, the new proposed guideline of within 15% is appropriate. And for Type 1s on insulin, plus/minus 10% is the goal to aim for; it’s the point at which only 1% of hypoglycemic events would be missed.
I’ve been told by several people in the industry that Agamatrix’s meter and strips are closer to within 10% of a standard lab test. Indeed, Sanofi said they chose Agamatrix as the developers of their new iBGStar because of the company’s WaveSense technology, which is believed to be more accurate with hematocrits (remember, the volume percentage of red blood cells in blood). If you’re adamant about accuracy, you’ll want to look at the meters Agamatrix sells directly: the Keynote, Presto and Jazz meters.
I know for sure I’m not the only one obsessing about this. Fellow diabetes blogger Bernard Farrell told me he had a similar “awakening” about meters showing different glucose numbers when he saw his CGM and meter weren’t tracking. Switching to an Agamatrix meter himself, he saw that it more closely reflected his CGM results, and he was able to bring his A1C down half a percent. There’s at least some anecdotal proof that more accurate meters = improved patient outcomes.
{Disclaimer: I was not compensated by anyone or any company for writing this.}
Thanks Riva, great info! We can’t wait to hear what hot D-topic you’ll be obsessing over next.
I recently wrote about this too. Even when I get different readings from two meters, how do I know which one is most accurate?
I’ve learned to test side-by-side, old meter vs. new for a while, then recognize the new meter’s numbers as (sorry for the buzzword) The New Normal.
Thanks for the great info.
1) I suggest testing your glucose right before blood is drawn at a lab, and write down the result. When the lab results come back, you can do a comparison.
2) We could all use more accurate strips, but at what cost? If it would double the cost of strips, would it be worth it? I hope competition will keep prices in line.
3) When you test your glucose, ask yourself if the result is believable. Sometimes things go wrong in the testing process. If you are saying to yourself “no way”, then don’t act on the results. Clean your hands and retest.
4) Many of us could make good use of better technology. That does not mean great results cannot be had with the technology we have. I am a type 2 on a pump, and have switched from freestyle to one touch to contour as required by my insurance. Every one of my 22 A1c’s in the last 5 years has been between 4.9 and 5.4. Current meters are good enough to get the job done. Don’t let meter inaccuracy keep you from doing your best.
-Lloyd
Thanks for explaining all this confusing stuff for us, Riva! I’ve always wondered about what the variable causing factors were. I focus wholeheartedly on the variables I can have control over and similar to Lloyd, who commented above, have kept my A1c under 5.7 for the last 6 years. So while meter inaccuracy is an important issue, it’s not someone’s reason for out of control blood sugars. That said, there are MANY good reasons for out of control blood sugar. I do, totally, recognize that.
Is the difference going to kill us? Maybe not, at least in the short term. But as someone who likes to keep really tight control, and is extremely sensitive to insulin (CF=145mg/dL per Unit) the small differences can make a huge difference in my control. Most upsetting to me was seeing my A1C jump from 5.8 or lower up to 6.3 for an entire year, until I realized it was because Freestyle’s new butterfly strips were consistently giving me a number that was 20-30 points too low on my OmniPod PDM, perhaps even further off for higher readings. It made me feel like Abbott’s changes were made in complete disregard to the fact that our lives depend on these numbers. And if my bg reads 200 before I go to bed, I need to know whether it’s actually 160 or 240, because that 20% difference will make a huge difference in whether I plunge extremely low overnight from an overcorrection, or start rising because of increasing insulin resistance to wake up in the 300′s. Yes, those numbers really do mean more than the strip manufacturers (and even the FDA) seem to realize.
Thanks for the information on the Omnipod and the Freestyle butterfly test strips. I have also watched my A1C creep up over the last year on that system, but did not know the cause. I typically run pretty tight control, so was not really knowning how to adjust to correct that creep. I will mention this to my doctor when I see her next, and thanks for posting about your experience with this system!
So, standard disclaimer: T2 controlled with diet and exercise. I’m glad various companies are making some progress. All I could think upon reading the announcement of Bayer’s new meter is, “within 15 points while below 100? That’s no improvement at all.” My target range is 80-85 fasting and 90-100 post-meal. I test twice a day (waking and one post-meal). It’s endlessly frustrating not knowing if a single result accurately reflects my current bg or is a fluke. I often end up taking multiple samples and averaging them out (it’s not uncommon to see 15-20 point differences for samples taken with 60s of each other). I’ve tried several meters, and have ended up using Agamatrix Wavesense meters[1], as they’re generally more consistent and closer to the lab results when I have blood draw.
[1] I have a Jazz & Presto, using the Presto lately, as I was out of work and its test strips are cheaper with no distinguishable decrease in accuracy.
Medical Science has got immense growth in modern world. But still diabetes is treated as a chronic disease. It is established that diabetes can be controlled, but it prevails in the body till the death. Diabetes is a metabolic disease due to the level of blood sugar (glucose). In other words it is the inability of the body to convert glucose in to energy.
actos class action lawsuit
After the digestion food is divided in to fats, protein and carbohydrates. These carbohydrates cause the formation of glucose. Glucose is transferred to blood and it is used as the energy for cells. Patients with diabetes this process i.e. the conversion of glucose in to energy, doesn’t work properly. As a result the level of glucose in the blood becomes high. This phenomena in the human body is happens when the insulin secretion comes under defects. Insulin is a hormone which produced by pancreas that control the glucose level in the human body. In a normal physical condition when the glucose level bumps up in the body, the pancreas release adequate insulin to regulate the glucose level. This insulin converts the glucose in to energy for cell and thus it keeps sugar (glucose) level normally. But in patient with diabetes this type of auto regulation system doesn’t happened. The deficiency of insulin may come as the reason for hyperglycemia. Excessive thirst, excessive urination, extreme hunger, fatigue are some the common symptoms of diabetes. Studies reveal that diabetic people are more prone to heart diseases than other people. That is why people with diabetes need more care in daily life
Great article! A lot of people are actually depending on the results of their home meters when it comes to identifying their current blood sugar level. Knowing now that somehow, meters may vary in significant percentages with the results will be able to let people know that depending fully on what their meters reflect may not totally be useful, though it could be helpful in keeping track of fluctuations on our sugar level. In fact, I still believed that regular checkup would recount a more accurate blood sugar level.
internal medicine
While “experts” suggests that blood glucose meter accuracy be within plus or minus 20% this meter boasts plus or minus 15%. Until meters get within plus or minus 1%, 15% will do for now.
You are right. We would never accept that kind of reliability from our iPods, iPads, cell phones, or other fancy electronic devices. Too bad we can’t put that kind of engineering and research dollars into the medical instruments that help us live and manage our daily lives. I am sure we would be better off for it. Priorities??
[...] Our meters, the ones we literally give out blood to several times a day to help us make decisions on medication, diet and managing Diabetes have an error factor of +/- 20%. That’s huge. Plus of minus 20%! Let me put that into perspective for you. Say you get a reading of 100. Well, it might not really be 100, your true glucose level can be anywhere between 80 (great) and 120 (not so great). That’s a 40 point swing that takes you from ‘normal’ to Diabetes. You can get a good sense of what is going on here. [...]
I am sorry for the cross postings but I just realize this post here is more appropriate for me to relate my experience with the iBGStar. I have used the IBGStar to compare with lab results during two separate Oral Glucose Tolerance Tests, by reading the IBGStar immediately prior to drawing blood sample at the lab. During the first OGTT done in October 31st this year, the IBGStar read 7%, 9% and 13% higher than lab values at start of test, 1h and 2h intervals, respectively. During the second OGTT done in September 20, 2012, the IBGStar read 7%, 6%, 12%, 13%, 32% and 46% higher than lab values at varios time intervals. The number of data may be few but from this data I conclude this device to be consistently higher than lab results, with a couple of readings out of the +/-20% FDA range. I would like to switch to a more accurate reader and would appreciate your recomendation. Thanks again.
I’ve generally found the Wavesense products (Presto & Jazz) to be fairly consistent and accurate (at least in comparison tor others I’ve tried). The Presto has been around for a while, so is noticeably cheaper for test strips, but has been just as consistent as the Jazz.
Thank you.
I recently tested two meteres, the Abbott Freestyle and the Nano with the same drop of blood. I did not re-squeeze my finger as I had enough blood to activate both meters. I was a 1/10 of a point of being 40 points apart, the Freestyle being the lesser of the two numbers. I called Accu-Chek company and the rep began telling that they use a different technologies. That just make any sense to me as we all want to know the most accurate bottom line. Regardless of the different technologies, you want to know the “number”. A forty point spread could mean the difference in doing something about it or not. That seems to be the bottom line; so what do I do? Do I consume some sugar or go for a run?
[...] [...]
Wow this is very insightful.I have noticed sometimes my CGM and meter are 20 points off and sometimes times they are 2 points off. And 90% of the time the meter always reads higher then the CGM.