By Christine Craig, MS, RD, CDCES

As more and more people are using glucose sensors as a tool to improve their blood sugars, they may feel frustrated or confused about the mismatch between their sensor and glucose meter readings. Some many even ditch this new fangled sensor, thinking that their old friend, the meter, provides more reliable information.

AW called the clinic to report an issue with the accuracy of their continuous glucose monitor (CGM). They stated that their blood glucose (BG) and sensor glucose (SG) have been 60-100 points different for the past 24 hours. They are concerned about this difference and want help troubleshooting what could have gone wrong. 

In this article, Christine Craig, MS, RD, CDCES outlines strategies to explore this glucose mismatch and provides practical problem solving actions to clear up the confusion. In addition, we invite you to download a CGM Troubleshooting Cheat Sheet that provides specific information on different interfering substances and other factors that can affect the accuracy of continuous glucose monitor (CGM).

Assessing CGM Accuracy

Assessment of accuracy within CGM or flash glucose monitors in studies uses mean average relative difference or MARD. (For simplification, we will refer to both as CGM or sensor for the remainder of this article). CGM values, also called sensor glucose values, are compared with a standard reference, often the lab-measured Yellow Springs Instrument (YSI) analyzer, and are reported as a percent of the mean or median absolute error between CGM and reference values.

Almost 20 years ago, the MARD values for CGM were about 20%, and now most CGMs have MARD values near or under 10%.¹

As CGM accuracy improves, we now see non-adjunctive indications for many CGMs on the market. This term means the FDA has approved the use of CGM for treatment decisions without BG confirmation. In studies, we can see MARD differences within specific populations, in different CGM site placements, or between different sensor glucose values. For example, a sensor glucose range between 70 and 180 may have a different MARD than a sensor glucose range of less than 54 mg/dl. 

How do we share about sensor accuracy and highlight when differences are most likely to occur? 

As an individual living with diabetes in the real world, MARD can be impacted by factors specific to the sensor or factors of daily living. Day of sensor wear, sensor-to-sensor variation, or insertion factors such as body site selection, skin integrity (thin or loose skin), scar tissue, or body movements can impact sensor glucose accuracy. ²

If the sensor is compressed due to positioning, lack of perfusion to the sensor site can cause the sensor glucose to be temporally lower than actual glucose values.³ In this case, during CGM data interpretation, you may see a fall and then a return of trend within the sensor glucose readings. Often, compression lows occur overnight due to sleeping directly on the sensor, but during the day, compression can occur with other situations such as placement around a beltline, other very tight clothing restrictions, or if positioned within the interior of the upper arm.

In studies, body site selection between arm, abdomen, and buttocks has shown differences in accuracy, and in the real world, for each individual, this actual difference can be more significant than for others. Using recommended site selection can improve sensor accuracy, but it is essential to consider each individual living with diabetes to determine the best placement for optimal sensor accuracy. 

We often hear that BG does not equal SG, but what does this mean? 

A blood glucose meter measures capillary blood glucose, and CGM sensors measure interstitial fluid, a thin layer of fluid surrounding the cells just below the skin. These values correlate well, but they measure two different types of fluids.

This difference in BG vs. SG at one point in time can be alarming at first and is a commonly reported concern related to sensor accuracy. Education about the difference in measure and understanding that this difference is most significant during rates of change (a rise or fall in glucose values) can help address this concern. Blood glucose values change first, followed by sensor glucose.

The lag time between BG and interstitial glucose (what the sensor measures) can be from 2 to 20 minutes.

We may notice this difference most during increased rates of change (such as >2 mg/dl per min) and during circumstances such as after a meal, after treating a low glucose event, during and immediately after exercise, or after taking diabetes medications that have a more rapid effect on glucose such as meal/correction insulin or sulphonylurea.¹

Educating about the difference between BG and SG and how to interpret glucose trends, sensor glucose trend arrows, and monitoring change over time can improve understanding.  

Interfering Substances

The 2024 ADA Standards of Care recommends that for ALL people who wear CGM devices, education and assessment of potential interfering substances occur. Many of these interfering substances can be a common over-the-counter supplement or pain reliever. The effect for all reported interfering substances is a false elevation in sensor glucose versus actual blood glucose values. ⁴ It is cautioned that this could cause missing a severe glucose event.

Each manufacturer has different identified interfering substances.

• For the Libre 2 and 3 systems, more than 500 mg/day of supplemental Vitamin C can interfere. More than 1000 mg/day can interfere with the Libre 2 Plus. ⁵
• Dexcom G4, G5, G6 ⁶, and G7 ⁷ are all affected by hydroxyurea and acetaminophen,
• However for Dexcom G6 and G7 devices acetaminophen dose is specified as greater than 4 grams per day or 1 gram every 6 hours.
• Medtronic Guardian 3 and Guardian 4 devices are affected by hydroxyurea and any acetaminophen dose⁸
• For more information, refer to our cheat sheet: CGM Interfering Substances.

Studies and reports of interfering substances are not comprehensive, so it is recommended that if suspicion occurs that a medication or supplement is contributing to erroneous CGM values, they should contact the manufacturer’s technical support team.⁹

Calibration

For sensors requiring calibration, poor quality of the blood glucose test can also create inaccuracies with the glucose sensor readings. Most manufacturers with CGM calibration requirements recommend ensuring a “clean calibration,” having individuals wash their hands, taking the second drop of blood when hand washing is unavailable, and calibrating when glucose values are more stable, such as before a meal, insulin, or exercise. Luckily, most devices no longer require calibration, but it is important to review technique when applicable. 

Getting Back to AW

AW reported no recent changes with medications, supplement intake, or fortified food sources, and no changes were noted at the sensor site. However, they did report that the CGM was on the last day of manufacturer recommended sensor wear. Sensor integrity variation based on day of sensor was determined to be the main consideration of cause. After changing to a new sensor, AW’s general range of BG to SB differences were observed. This event served as an important reminder to AW to check BG when SG values were not reading as expected and different than any symptoms.

Education When Starting Sensor Important

As educators, we can support individuals concerned with CGM accuracy by providing education about the differences between SG and BG values. Assessing individual factors such as medical conditions, movement and activity, site integrity and placement, issues of compression, and medication and supplement use (assessing for potential inferring substances). We can educate individuals to check BG anytime symptoms do not match SG readings and report any concern of a faulty sensor to the manufacturer. Reporting can often lead to the replacement of faulty sensor. 

Since the arrival of CGM, benefits including improved quality of life factors, A1c reduction, increased Time in Range, and prevention of hypoglycemia have been found.⁴  We know from working in diabetes care and through real-world studies the overall value of CGMs. Unlike BG, CGMs provide a complete picture showing trends for direction, personalized alerts, and supporting pattern recognition. 

Although not ideal for all, growth of users will continue to rise with approval of over-the-counter product options and understanding accuracy issues is essential to ensure safety and support user experience.

For more information, refer to our cheat sheet: CGM Interfering Substances.

References:

1. Bailey, Timothy & Alva, Shridhara. (2021). Landscape of Continuous Glucose Monitoring (CGM) and Integrated CGM: Accuracy Considerations. Diabetes Technology & Therapeutics. 23. S-5. 10.1089/dia.2021.0236.
2. Heinemann L, Schoemaker M, Schmelzeisen-Redecker G, Hinzmann R, Kassab A, Freckmann G, Reiterer F, Del Re L. Benefits and Limitations of MARD as a Performance Parameter for Continuous Glucose Monitoring in the Interstitial Space. J Diabetes Sci Technol. 2020 Jan;14(1):135-150.
3. Mensh BD, Wisniewski NA, Neil BM, Burnett DR. Susceptibility of interstitial continuous glucose monitor performance to sleeping position. J Diabetes Sci Technol. 2013 Jul 1;7(4):863-70
4. American Diabetes Association Professional Practice Committee. 7. Diabetes technology:Standards of Care in Diabetes—2024. Diabetes Care 2024;47(Suppl. 1):S126–S144
5. Full Indications and Important Safety Information. Download On April 17^th 2024 From: https://www.freestyle.abbott/us-en/safety-information.html
6. Interfering Substances and Risks. Downloaded on April 17^th 2024 from: https://www.dexcom.com/en-us/interference.
7. Dexcom G7 Safety Information. Downloaded on April 17^th 2024 from: https://dexcompdf.s3.us-west-2.amazonaws.com/en-us/G7-CGM-Users-Guide.pdf#page=12
8. Medtronic Device Safety Information. Downloaded on April 17^th, 2024 from: https://www.medtronicdiabetes.com/important-safety-information
9. Heinemann L. Interferences With CGM Systems: Practical Relevance? J Diabetes Sci Technol. 2022 Mar;16(2):271-274.