- βEversense implantable CGM showed the best real-world accuracy at 14.8% MARD
- βHead-to-head comparison used the same participants and meter, making results highly reliable
- βStudy confirms all CGMs are less accurate at home than in clinical trials β important for patient awareness
| Journal | Diabetes technology & therapeutics |
| Year | 2020 |
| Authors | Jafri, Balliro, El-Khatib |
| PMID | 32453604 |
What Was Studied β And Why It Matters
Continuous glucose monitors (CGMs) have transformed diabetes management, giving people real-time insight into their blood sugar levels. But with several devices on the market, a key question remains: which one is most accurate in everyday life?
Researchers at a leading diabetes center set out to answer this by comparing three CGM systems β the Dexcom G5, the Abbott Freestyle Libre Pro, and the Senseonics Eversense β under real home conditions over six weeks. Most accuracy studies happen in controlled clinical settings, so this home-based comparison offers a much more realistic picture of how these devices actually perform.
How the Study Worked
Twenty-three adults with Type 1 diabetes wore all three devices at the same time for six weeks while living their normal daily lives. Twice a day, participants checked their blood sugar with a standard fingerstick meter. Researchers then compared those fingerstick readings to what each CGM reported at the same time.
Accuracy was measured using something called Mean Absolute Relative Difference (MARD) β essentially, how far off the CGM reading was from the fingerstick result, on average. A lower MARD means better accuracy.
What the Study Found
- Eversense (implantable CGM): Most accurate, with a MARD of 14.8%
- Dexcom G5: Second, with a MARD of 16.3%
- Freestyle Libre Pro: Third, with a MARD of 18.0%
The Eversense was statistically more accurate than the Libre Pro, and also outperformed the G5 in a larger head-to-head comparison. All three devices were somewhat less accurate at home than they appear in clinical trials β which is important for patients to understand when interpreting their CGM numbers.
What This Means for Real Patients
No CGM is perfect, and this study confirms that real-world accuracy is lower than what manufacturers often report. The gap between devices may seem small in percentage terms, but when your blood sugar reading guides an insulin decision, even a few points of difference can matter.
Here's what to take away:
- Always confirm with a fingerstick when you feel symptoms that don't match your CGM reading β especially before treating a low.
- CGM choice matters, but so does proper sensor placement, calibration habits, and consistent supplies. Having reliable, high-quality CGM supplies β like those available through MDS Diabetes β helps ensure your device performs as well as possible.
- The Eversense is implanted under the skin by a healthcare provider and lasts up to 90 days, which is a very different experience from a wearable patch sensor. Discuss with your doctor whether this option suits your lifestyle.
- The Dexcom G5, while slightly behind Eversense, remained a strong performer and is widely used and supported.
Bottom Line
In this real-world comparison, the implantable Eversense CGM edged out the Dexcom G5 and Freestyle Libre Pro in accuracy for people with Type 1 diabetes at home. However, all three devices showed lower accuracy than lab studies suggest β a reminder that no CGM fully replaces fingerstick checks in critical moments. Talk to your care team about which device fits your needs, and make sure you're stocked with the right supplies to get the best performance from your CGM every day.