This article is more than 5 years old. Information may no longer be current.
CGM accurate, reliable for unsupervised closed-loop system at home
Click Here to Manage Email Alerts
The Freestyle Navigator II continuous glucose monitor was a numerically and clinically accurate tool for a closed-loop insulin delivery system in an unsupervised home setting, according to research in Diabetes Technology & Therapeutics.
In three unsupervised, randomized, open-label, crossover home closed-loop studies assessing sensor accuracy across a wide glycemic range, researchers found that, when assessed by absolute relative difference, the Freestyle Navigator II continuous glucose monitor (CGM; Abbott Diabetes Care) was less accurate in the hypoglycemic range compared with the euglycemic range, but absolute difference was comparable.
“Our analysis suggests that in ambulatory, real-life conditions, the Freestyle Navigator II CGM achieved nearly comparable [mean absolute relative difference] scores with those reported in the manufacturer’s labeling and in controlled research facility settings,” the researchers wrote.
Hood Thabit, MD, of the Wellcome Trust-MRC Institute of Metabolic Science at the University of Cambridge, and colleagues analyzed data from 41 adults (mean age, 39 years; 23 men) and 16 adolescents with type 1 diabetes (mean age, 15.6 years; 10 boys) on insulin pump therapy for at least 3 months. Participants used insulin pump therapy with real-time CGM at home for two overnight periods and one day-and-night period with closed-loop control. There was no supervision provided during the study periods; telephone support was available to participants for technical issues.
Researchers collected both CGM and capillary glucose values (10,597 pairs). To assess numerical accuracy, the researchers used absolute relative difference and International Organization for Standardization (ISO) limits, and they assessed clinical accuracy with Clarke error grid analysis.
Researchers found that, in the capillary glucose range, overall sensor accuracy showed a mean absolute relative difference of 14.2 ± 15.5% and a median absolute relative difference of 10%. Researchers observed the lowest mean absolute relative difference in the hyperglycemic range at 9.8%.
The mean absolute relative difference was consistent at 12.3% and was not different between participants within the euglycemic and hyperglycemic range (P = .06).
Throughout the duration of the study, sensor readings from the Freestyle Navigator II were within ISO accuracy criteria 70% of the time, according to researchers.
“In the hyperglycemic range, the improved numerical accuracy facilitates safe operation of [closed-loop] by avoiding hypoglycemia from CGM sensor over-reading,” the researchers wrote. “Conversely, sensor inaccuracy at the hypoglycemic range is mitigated by insulin delivery suspension during [closed-loop] operation.”
In a post hoc analysis, researchers found that participants who spent the most time in the hypoglycemia range also had a lower mean absolute relative difference overall.
“This is consistent with findings of lower [mean absolute relative difference] in the hypoglycemia range and may potentially contribute to bias in studies reporting primarily on hypoglycemia reduction outcomes,” the researchers wrote. “Further work is needed to determine individual-level factors influencing sensor performance and impact on [closed-loop] performance.” – by Regina Schaffer
Disclosure: Thabit reports no relevant financial disclosures. Please see the full study for a list of the other authors’ relevant financial disclosures.
Click Here to Manage Email Alerts