Analysis of accuracy for the Freestyle Libre glucose monitor
The purpose of the study was to evaluate the performance and usability of the FreeStyle® Libre™ Flash glucose monitoring system (Abbott Diabetes Care, Alameda, CA) for interstitial glucose results compared with capillary blood glucose results.
The takeaway from this study is that the sensor provides a larger number of measurements, including day and night readings, which can be used to evaluate glucose patterns and trends. In comparison, BG readings via fingerstick provide single, intermittent measurements, which may not capture intervals of extreme variability or those overnight high and low points. In a recent study, it was demonstrated that the use of continuous glucose monitoring with or without alarms reduces time spent outside glucose targets compared with only self-monitoring of BG via fingerstick. The Libre system provides the user with the current glucose result, glucose pattern, and trend information on the display of the hand-held reader when the sensor is scanned. This type of monitoring system may benefit individuals who have ceased sensor use due to alarm fatigue, becoming overwhelmed by alarms or for those who have not been able to use a CGM type system due to cost. Further research is needed to evaluate whether the Libre System could provide additional benefits such as improvement in glycemic outcomes with continued use over time, as well as improved compliance with sensor wear regardless of alarms for low and high values.
See full research report here: http://online.liebertpub.com/doi/full/10.1089/dia.2014.0378
Libre II just isn’t accurate at all. Accucheck shows 8.3, Libre II shows 11.4. Huge variance. It’s a waste of time to use this product. I can go on and on with my readings over time and how they differ. I kept a chart but bottom line, I got rid of it.
This is the worst ‘instrumentation’ I have seen over my engineering career.
I have a strong love/hate relationship with Libre II. My A1c has drifted down to 5.1 (really) from 7.4 as a T1D using the Libre II. Lows are my main concern mid day.
It seems over half the sensors fail or die very early. The inaccuracy can read 50 points low when the BG is about 100 and can read 100 points high when the BG is 100. Sensor calibration appears to be the culprit if you ignore those compression lows. Typically the reader is 20-40 points high or 15 points low depending in which side of 100 per finger stick.
Abbott seems to have NO interest in improving the product over the 15 months I have been using it. They always send replacement sensors and readers.
I am waiting for the Dexcom G7 in the US market and look at how it does.
The same thing happens to me. I’m a new diabetic and because of all the lows my Insulin was discontinued. I started using a traditional glucometer and the BG has been high. I was just about to have my Trulicity decreased when I started doing finger pricks to compare. This could lead to be very dangerous. I’m a retired Nurse.
I agree with others re accuracy of Libre 2 sensors Out of 5 only 2 lasted after 5 five days. Support replaced said problem sensors . with xtra highs and lows I rat sugar or shoot up insulin. I quit. Abbott is messing up my AIC. Ketoacidosis is serious
Using the libre with my smartphone. Has been fine, until last three days it was reading very low BS under 50, and when I wouldeat, it would shoot up fast, then plummet below 50 again within 2 hours. Today, third day of this I blood tested and my bs read 115, but the libre reads 40! (tested both 3 times in 15 minutes) Not sure what to do!
Hi Mary, I’d replace a libre showing that kind of repeated inaccuracy, but be sure to reach out to Abbott for a free replacement on any under-performing sensor or any that does not last the full 14 days.
Yes, I have the same problem. Today my the libre tested at 77 and a finger prick tested at 128. Don’t worry to much if the sensor readings jump down. It’s readings seem to be pretty inaccurate.
My mother in laws number was 60 on the libre and 200 with the finger prick test on her accu check… explain that? Scared me half to death. This is not the first time there have been very low numbers, now we know enough to check it the old fashioned way.
I made her eat peanut butter toast and drink milk to try to raise the numbers before. Now I think I know the problem. It is your Libre system
Hi Elaine,
there are lots of causes of false low readings
the most common is a compression low from reduced circulation when laying or resting on the CGM. Another is poor circulation to the CGM site area.
one of the worst problems with CGM is the potential for false readings. avoiding wearing on areas that are lean, have poor circulation or where we are likely to lay sleep or sit is helpful and maintaining good hydration also helps them be more accurate.
Just started using the freestyle libre (4th day), and unlike other posts, the comparisons I’ve done, the finger prick tests have all shown 30 to 40 points lower than the libre. I see that most other posts show the opposite. Could it be the location of the sensor?
could be location
could be the meter you’re using to compare
also it depend on the range, Libre tend to read lower in low numbers and higher in high numbers.
I just switched from the Libre 14 Day to the Libre 2 fo the alarms. I was looking for some comparisons of accuracy comparing the Libre 14 Day to the Libre 2. According to all studies the Libre is the most accurate CGM. Libre 2 especially in the low range. Reading down the comments I noted that you did eventually address the issue of compression lows, something I dealt with and corrected 99% of by speaking with my Endo who recommended a slight adjustment in how I position the sensor. Many people do not seem to be aware that they are suppose to check their BG by finger stick if it looks like a treatment decision is needed, or. how they feel doesn’t match the reading. Also, that they should use the strip function on their Libre Reader to do so. These devices have variability and going between devises and brands can result in a wide and confusing variation.
I am confused as to why this sight is not educating people regarding the difference between IFG. and BG readings and how to use CGM systems such as the Libre. Also the difference between A1C and GMI values, the latter of which is showing to be a better indicator of long term control and risk of complications than A1C. Further people with the same A1C can have much greater variability and average BG numbers than people with the same GMI.
So my question as a medical practitioner who also happens to have diabetes: Why are you not taking the opportunity to educate people so they can use these systems most effectively? And you seem to discredit Libre and recommend DexCom a lot. What is your relationship to DexCom Inc.?
Hello Dr Koretsky.
We educate our community both individually through patient appointments and consultations. Our articles reviewing different devices technologies and products is shared on our site and we have many articles and resources on most diabetes related topics including CGM use. We work with all the device companies as trainers and consultants but are not in any way commissioned for sales of devices. We make simply share our experiences and work with patients to give them support and education to make the appropriate choices in technologies and products to best fit their needs.
Due to slightly better accuracy, real time alerts on smart phones and interactivity with other devices and technologies the Dexcom G6 typically meets our population needs best, but the Libre 2 has shown some improvement in these areas and we hope the companion app will improve these areas further. We have no desire to see One strong CGM on the market, we hope to see many robust manufacturers challenging and pushing one another forward to serve the Diabetes community bette and better.
I encourage you to continue to explore our site and keep an eye out for our updated return of Type 1 university which includes indepth education opportunities into many areas of diabetes management.
Hello. I’ve been using the 14 day Libre, and continue to use the Freestyle Lite BG. Recently, I’ve been getting low readings (65). I follow with a blood test and the difference is huge! Today, my fasting numbers with Libre: 80 My blood strip showed 145. That is disturbing. Also, the Libre showed that my sugar was super low at around 3 a.m. so I got very worried. I think I will have to set the alarm for 3 a.m. and test with blood strip, cause it’s scary to see the low numbers on my Libre first thing when waking u.
To provide further information: I do not use Insulin. Have been using Bydureon but stopped injecting 2 weeks ago.
Hi Edi, definitely follow up with your prescriber regarding those low blood sugars, your medications appear to need further adjustments.
and I recommend reaching out to freestyle regarding the accuracy issue you are seeing. we know the system is far less accurate in the lower and upper BG ranges.
I have been using the FreeStyle Libre 14 and prior to that I was using the 10. I take multiple readings each day but when I do a finger prick test it is always higher than the Libre. Prior to using the FreeStyle I had A1C’s in the high sixes and the low sevens, Now I seem to be in the mid to high sevens over a three month period. As I prepare for my Diabetic Quarterly exam I was OK with a 7.4 according to my monitor. I was surprised when my Lab A1C came back at 8.1. I will be discussing with my Doctor if I should continue with the FreeSyle or just go back to the finger prick method.
That’s an unfortunately accuracy issue there Terry, you might want to look into the Dexcom system for improved accuracy, real time alerts and it’s also covered by medicare if needed.
I have the same problem! My A1c tests are always higher then the Freestyle readings. According to my last test it was a 7.9 from the Lab. My 90 day average from the Freestyle was 146 so I thought it should be below 7.0. Also, I use One Touch Ultra to spot check and the Freestyle readings and always lower then the finger prick. I have discussed this with my Doctor and he doesn’t seem too concerned about the accuracy of the Freestyle. I guess I will try for a lower average using the Freestyle only as an estimate.
HI Terry, there are other issues that can also cause A1C to run higher than CGM data would indicate. The most common are vitamin C supplement use and Anemia. I recommend talking to your doctor about ruling out other potential issues since this is a discrepancy wider than we would anticipate from CGM inaccuracy issues.
My wife recently came out of the hospital, she had been in there for very low Glucose and was unconscious when the ambulance arrived. After 8 days her glucose level was stabilized. I obtained a test unit of this system for a 14 day trial. Her glucose readings were basically going up and down to excessive levels during the last week of use and we were not using the finger stick method. That was a mistake.
I just checked her reading with the 14 day system and the meter showed “LO”. I woke her and started feeding her things to get her glucose levels up. She appeared to be normal, not the drunk appearance I was used to when her levels are low. I scanned the device on the back of her arm again after 15 minutes and eating. She appeared perfectly normal. I do not know what the LO reading indicates but presume it is below 50. So, I then used her old meter to check her finger. “145” This discrepancy of approximately 100 points from finger stick to the 14 day meter just seems to be a bit much. We will continue to use this meter but it really does show a need to follow instructions and check the accuracy… often. I do not know if it will be worth the cost to continue with this system if it is still not accurate. My concern is that what if the scan were showing 145 but the actual reading was LO on the finger stick but we did not check until she was unable to remain unconscious. How would I know that she was passed out and not just asleep? That was the main point of getting this device to keep from waking her to check her glucose and to present an ACCURATE chart for the doctor to see.
HI Charles, the accuracy issues of the Libre system concern us as well, particularly below 100 and above 250.
when we are sleeping any CGM system is prone to what are called compression lows. this is when laying on the sensor reduces fluid circulation, causing an artificially low reading. Wearing them in an area that is less likely to be laid on can help, making sure to maintain good hydration can also reduce false low readings
I would also recommend checking out the Gvoke Hypopen by Xeris medical. this is an autoinjector (like an epi pen) for glucagon, a hormone that will force liver glucose release and raise blood sugars in an emergency. this could be a literal life saver for your wife, and should be carried by anyone using insulin.
https://www.gvokeglucagon.com/
i just started using the libre – i love it but I still check my finger sticks and its off by 20 pts after dinner and 15 after lunch… which one do I believe? I want to believe i’m doing great..But if there is a huge difference in readings i might as well keep checking it with the finger sticks.. Please help..
HI Amy,
First whenever blood sugars are moving(like after a meal) we expect to see a larger than typical difference between finger stick and CGM values since the glucose getting to the interstitial fluid (Fat layer) takes longer than it does to the blood.
when blood sugars are stable (over night or between meals) we anticipate a 15% variability between a CGM reading and a blood sugar. This is because meters are FDA approved with up to a 15% variability for 85% of their checks (The other 15% can be even farther off!) so if the meter is spot on then the cgm might read 15% off, and if the CGM is spot on the meter might be reading 15% off.
diabetes is a disease we tend to think of in terms of “math” but the numbers are really never quite as exact as we’re often led to believe.
After 3 weeks of use, I find it completely unuseable. It’s always off by anywhere between 20 and 50 points against a finger prick
We find accuracy to be our number one problem with the libre as well. Hopefully the Libre 2 will have improved accuracy. We find the Dexcom G6 to be a more accurate product for most users.
The Libre CGM has been a Godsend for me. I started with the 10 Day, migrated to the 14 Day and now, after finishing the last sensor for the 14 Day, I’ll be using the Libre 2. As you’ve mentioned, both the last two Libres had readings too high when over 145 and too low when less than 60-65. However, Dexcom G6 is not really much better. In using the Libre as a RELATIVE indicator of real BG status, I’ve been able to keep my A1c below 6.5 with my last one at 6.1. My Enco says she’d like it a bit higher to prevent any real bad lows, but I’ve gotten the routine down so well that this isn’t likely to be an issue anymore. Here’s what a distributor of CGMs had to say:First, the FreeStyle Libre 2 is an integrated Continuous Glucose Monitor (iCGM). The FDA classifies an iCGM as a CGM that can be used as part of an integrated system with other compatible medical devices, software, and electronic interfaces. Examples include automated insulin dosing systems, insulin pumps, blood glucose meters or other electronic devices used for diabetes management. It’s important to note that Abbott states “this system must not be used with an automated insulin dosing system, such as a closed-loop insulin delivery system, and insulin suspend systems.” You’re probably asking yourself, “What does all of this mean?” It means that for now, the FreeStyle Libre 2 can integrate with software and other medical electronic interfaces but cannot control delivery of insulin through an insulin pump.
Second, the FreeStyle Libre 2 has a customizable audio alarm which alerts you within a minute if you are too high or too low without having to scan. This feature brings the FreeStyle Libre 2 on a level playing field with the Dexcom G6. This alarm is optional, giving you the freedom to turn it on or off. The FreeStyle Libre does not have any alarms and requires you to check your reader to see if you’re approaching hypoglycemia or hyperglycemia.
The FreeStyle Libre 2 system has a combined mean absolute relative difference (MARD) of 9.3% (9.2% for adults and 9.7% for pediatrics), making it the only iCGM sensor sustaining a high level of accuracy over 14 days. For those that are not familiar with MARD, it is a commonly used measurement of performance for continuous glucose monitoring devices. According to the American Diabetes Association, “the mean absolute relative difference (MARD) is currently the most common metric used to assess the performance of CGM systems. MARD is the average of the absolute error between all CGM values and matched reference values. A small percentage indicates that the CGM readings are close to the reference glucose value, whereas a larger MARD percentage indicates greater discrepancies between the CGM and reference glucose values.”
Lastly, the FreeStyle Libre can only be used by adults 18 years old or older while the FreeStyle Libre 2 has been approved for children ages 4 and up, as well as adults. The FreeStyle Libre 2 also has one-third fewer false alarms in children when compared to competitor CGM brands.
I just downloaded the Librre 2 phone app and in a couple of weeks I’ll know how well it works. The app for the 14 Day worked perfectly. It’s been stated that the algorithm for the Libre 2 has been improved over the 14 Day too. At this point, I don’t think it’s possible to say that Dexcom is any better than the Libre 2. If the patient is a Medicare beneficiary, there’s no charge for the reader or sensors from either manufacturer.
I HAVE USED THE LIBRE FREESTYLE FO 3 MONTHS. IT IS ALWAYS BETWEEN 40 AND 50 MG OFF. IT IS A WASTE OF MONEY AND TIME. DO THE DEXCOM G6.
I’ve just started using the sensor and am still pricking my finger . The Freestyle Libre indicates my sugar is 145 while my finger prick says 245….Really the readings are not accurate with the freestyle libre.
we expect a 15% variation from finger stick to libre at the best of times, which is wider than we’d like. But that is far outside of acceptable variation, remember if you are trending up or down that variance gets wider. The libre is also quite sensitive to cirulation changes so keeping hydrated and wearing it in a location with good circulation can help too.
i have just started using the freestyle meter, and though i was told there may be a 15 minute delay in up to the minute readings, i’ve found to my horror it is way off on actual reading as compared to my BG teat. Not for me… i have issues with hypoglycaemia and that degree of difference is dangerous….just have to stick to the old system.. not happy..
HI Judith, we don’t like tha accuracy issues of the libre either. We hope the new Libre2 will be more accurate but have not had the opportunity for a trial yet.
The Dexcom G6 is typically far more accurate at lower and higher blood sugars.
I use The FreeStyle Libre 14 days. In my experience, the main draw back is its reliability. In comparison with the results of contemporaneous blood sugar tests, a disparity of up to 20% could be expected(according to reps of the manufacturer with whom I talked). However I found that the rate of the disparity is not consistent. Moreover in the case of two of the sensors I have used, a disparity of 25% and up to 35% happened on more than one occasion. The agents at the support service were helpful in sending me alternative sensors, but until now the problem has persisted. I need to prick my fingers even with the FreeStyle. Was not the whole point to minimize the need to do just that?
I started using Libre 14 six months ago and have the same disparity you cited. This really blows me away because I thought the reader was giving me a sign of progress. The A1C shows that I am not making progress. I’m still looking for an answer.
I see it’s been approved in Europe. Sensors good for 14 days. No transmitter battery to die, with $600 cost for battery replacement.
Was not able to find indication of what the system cost is, might be when approved in US? Initial cost and ongoing (sensor) cost please?