Recommendations on the Collection of Comparator Measurement Data in the Performance Evaluation of Continuous Glucose Monitoring Systems

Abstract

While current systems for continuous glucose monitoring (CGM) are safe and effective, there is a high degree of variability between readings within and across CGM systems. In current CGM performance studies, device readings are compared with glucose concentrations obtained with a comparator ("reference") measurement procedure (usually capillary or venous glucose). However, glucose concentrations from capillary and venous samples can systematically differ, often by as much as 5 to 10%. Different comparator methods have shown biases of up to 8%, and comparator devices of the same brand can systematically differ by more than 5%. To address these issues, the Working Group on CGM of the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC WG-CGM) recommends standardizing study procedures and the comparator measurement process in CGM performance studies. The majority of IFCC WG-CGM members recommend the use of capillary samples as reference, mainly because CGM readings will then be aligned better with results from self-monitoring of blood glucose (SMBG). Even with factory-calibrated CGM systems, manufacturers require CGM users to perform SMBG in some situations, eg, manual calibration, confirmation of extreme readings, discordance between CGM readings and symptoms of hyper- or hypoglycemia, or intermittent signal loss. Comparator devices should meet defined analytical performance specifications for bias and imprecision. Comparator bias can be reduced by retrospective correction of comparator values based on measurements with a method or materials of higher metrological order. Once manufacturers align CGM readings of their systems with comparator results using standardized procedures, variability across CGM systems will be reduced.

Keywords: clinical performance evaluation; comparator data requirements; continuous glucose monitoring; standardization; traceability.

Figure 1.

Example of a traceability chain when estimating bias. Panel (a) Estimation based on methods of higher metrological order. If participant blood/plasma samples are used to verify the comparator method bias, potential commutability issues can be avoided. Panel (b) Estimation based on materials of higher metrological order (“suitable calibrator”). When using capillary comparator measurements, an additional step (indicated by dashed lines and boxes) by way of, eg, a laboratory analyzer (or other suitable analyzers) may be required if the comparator device is not labeled for use with the sample type of the higher-order materials. The “suitable calibrator” is any material with sufficiently well-characterized assigned target concentrations, eg, secondary certified reference materials as defined by ISO 17511. Light gray boxes and dotted lines indicate parts of the traceability chain that are outside of the purview of bias estimation in a CGM performance study. Device types in the columns “measurement procedure/device” are based on what is often used in CGM performance studies. Note that the step between the comparator and the CGM systems is missing because CGM systems measure glucose concentrations in interstitial fluid, where comparator or reference measurements are currently not feasible. Similar traceability chains can be established for other sample origins.