- The April 21 issue of the Journal of Diabetes Science and Technology, has a paper from Damiano et al. entitled, “A Comparative Effectiveness Analysis of Three Continuous Glucose Monitors: The Navigator, G4 Platinum, and Enlite.”
- Dexcom’s G4 Platinum was the most accurate sensor (MARD: 10.8%), followed by Abbott’s FreeStyle Navigator (12.3%); both were significantly more accurate than Medtronic’s Enlite with the Veo/MiniMed 530G algorithm (17.9%; p<0.005).
In the April 21 issue of the Journal of Diabetes Science and Technology, Drs. Ed Damiano, Steven Russell, and colleagues have published a paper entitled, “A Comparative Effectiveness Analysis of Three Continuous Glucose Monitors: The Navigator, G4 Platinum, and Enlite.” This represents the full data set and analysis following Dr. Russell’s ADA 2013 oral (#171) of the same name. As a reminder, 24 patients (12 adults, 12 adolescents) with type 1 diabetes each participated in one 48-hour closed-loop experiment. Reference plasma glucose measurements (GlucoScout) were obtained every 15 minutes (n=4,657 values) and compared to corresponding CGM values from three sensors (Abbott’s FreeStyle Navigator; Dexcom’s G4 Platinum; Medtronic’s Enlite) worn simultaneously by each participant. The GlucoScout was used for CGM calibration and to assess CGM accuracy.
Dexcom’s G4 Platinum was the most accurate sensor (MARD: 10.8%; 85% in Zone A of the Clarke Error Grid), followed closely by Abbott’s FreeStyle Navigator (12.3%; 84% in Zone A); both were significantly more accurate than Medtronic’s Enlite with the Veo/MiniMed 530G algorithm (17.9%; 69% in Zone A; p<0.005 for MARD). The Dexcom G4 Platinum data (10.8%) came in far better than its labeled accuracy of 13.2%, while the Medtronic Enlite data (17.9%) fared worse than its labeled accuracy of 13.6% (Abbott’s 12.3% was nearly identical to the 12.8% in its label). The authors acknowledge that calibration and study population differences (use of reference quality glucose, patients <18 years) may have played a role in the better-than-labeled accuracy seen with the G4 Platinum; they do not speculate on why the Medtronic Enlite came out worse. Directionally speaking, the results are consistent with other head-to-head studies we’ve heard, including the SPACE 2 trial shared at ATTD 2014 (see below).
The paper concludes, “This head-to-head-to-head comparative effectiveness study reveals the G4 Platinum as the most accurate and precise of the current generation of CGMs, followed closely by the Navigator, with both devices markedly more accurate and precise than the Enlite sensor with the Veo/530G algorithm.” The NIH, Helmsley Charitable Trust, and the Charlton Fund for Innovative Research in Diabetes funded this study. Overall, we are very glad to see independent, head-to-head CGM studies like this one for two main reasons. First, manufacturers self-report data, which sometimes casts doubt on the robustness of the study and results. Second, it’s near impossible to confidently compare accuracy across different single-product CGM studies, as there are a vast number of variables that factor in: calibration procedures, insertion sites and technique, reference glucose values, time in different glucose ranges, data pairing and statistical analyses, study population differences, etc.
We note that the publication includes a plethora of other CGM accuracy and reliability metrics (including many interesting graphs and charts), which are not included here for brevity sake – see the full paper here.
- Very large errors (MARD >50%) were statistically significantly much less common with the G4 Platinum (0.5%) than with the Enlite (4.3%; p = 0.0001); the Navigator (1.4%) was intermediate between the G4 and Enlite, but not statistically different from either (p = 0.1 and p = 0.06, respectively).
- An “unexpected result” was that CGM accuracy was worse in the adult cohort than for the adolescent cohort for all three CGMs. These differences between the adult and adolescent cohorts were found to be statistically significant for the Navigator (14.8% vs. 9.7%; p=0.001) and the G4 Platinum (11.9% vs. 9.6%, p=0.02), but not for the Enlite (20.1% vs. 15.6%; p =0.08).
- There were minimal differences in accuracy during the night vs. daytime hours. The authors note that all sensors were placed on the abdomen, which may have reduced the risk of compression artifacts at night. As a result, this finding may not translate to other sensor sites.
- CGM data were initially collected using the Medtronic Guardian receiver, but the authors worked with Medtronic to postprocess the CGM data using the Veo/530G calibration algorithm. The authors provided Medtronic with the raw Enlite sensor data and calibration plasma glucose values that were entered into the Guardian receiver. Medtronic postprocessed the data with the Veo/530G algorithm. If the calibration algorithm requested additional calibrations, the authors provided Medtronic with the first plasma glucose measurement after the calibration request. “Therefore, calibrations were provided prospectively to the calibration algorithm just as they would have been if the Veo or 530G systems had been used. Thus, the results presented here are representative of the performance of the Enlite sensor with the Veo or 530G device, not with the Guardian or Revel.”
- Abbott Diabetes Care declined to similarly postprocess the Navigator data with their latest calibration algorithm. This algorithm is used in the Navigator II CGM currently available outside of the United States. This was unfortunate to hear, since comparative head-to-head accuracy studies like these are so rare.
- In Dexcom's workshop at ATTD 2014, Dr. Jort Kropff (Academic Medical Center, Amsterdam, Netherlands) presented on the SPACE 2 trial, a very detailed and thorough head-to-head comparison of the Dexcom G4 Platinum and Medtronic Enlite sensors. The four-center, 24-patient study gathered CGM accuracy data over six days in the home setting (vs. fingersticks on the Accu-Chek Aviva meter) and over six hours in an in-clinic assessment (vs. YSI; included breakfast + a 180% insulin bolus to induce hypoglycemia). The 24 patients wore two sensors each. Devices were calibrated per manufacturer's instructions (it was not specified how many calibrations this entailed for the Enlite). Consistent with independent data previously presented from Drs. Ed Damiano and Steven Russell, the Dexcom G4 Platinum was more accurate than the Enlite – overall MARD of 13.6% with the G4 Platinum “significantly outperformed” the 16.6% MARD seen with the Enlite (p=0.0002). The G4 Platinum was also significantly more accurate for CGM values <70 mg/dl: 17.6% vs. 24.6% for the Enlite (p=0.005). Dr. Kropff emphasized that both devices were much less accurate in the hypoglycemia region, leaving room for future generations to improve. The G4 Platinum was significantly more accurate than the Enlite on all days of the study. The Clarke Error Grid also supported the MARD findings: 83% of G4 Platinum points fell in Zone A compared to 65% for the Enlite. Looking at individual sensor performance, 15% of Enlite sensors had a MARD <10% vs. 40% of G4 Platinum sensors. In concluding, Dr. Kropff emphasized that collecting both in-clinic and home CGM accuracy data is critical, as the latter adds “essential information” (e.g., a wide range of glycemia and duration of usage).
Close Concerns Questions
Q: What is the lower limit of accuracy for subcutaneous CGM? In their recent paper in the Annals of the New York Academy of Sciences, Drs. Peyser et al. suggested 8-10% is the lower limit.
Q: How negatively are fingerstick calibrations affecting real-world CGM accuracy?
Q: Will factory calibration result in better real-world CGM accuracy? (i.e., because it will eliminate the variability inherent in fingerstick calibration)
Q: How accurate will Abbott’s factory calibrated Flash Glucose Monitoring system be? Data thus far has suggested accuracy slightly better than the FreeStyle Navigator (~9-11%)
--by Adam Brown and Kelly Close