Aims

Subcutaneous continuous glucose monitoring (CGMS) has transformed management of Type 1 diabetes (T1DM). The FreeStyle Libre flash glucose monitoring system (FSL-CGM) is increasingly used but limited by patient concerns with sensor location. We investigated the accuracy and acceptability of the FSL-CGM at alternative sites during free-living and under experimental conditions.

Materials and methods

Participants with T1DM were provided with three FSL-CGM sensors applied to the upper arm, the lower back, and the anterior chest. On day 2-3 FSL-CGM sensor glucose was compared to venous glucose following a standard meal; during and after an exercise test; and after skin cooling. Participants completed 14-day use of the sensors with concomitant sensor scanning at all sites and capillary glucose tests. The primary outcome was accuracy between sensor sites of 14-day mean glucose. Clarke’s error grids, precision absolute relative deviation (PARD) and mean absolute relative deviation (MARD) were calculated.

Results

In the twenty participants accuracy of the back sensor was 97.9% and chest sensor 98% compared to the arm. During experimental conditions the arm sensor was more accurate than the back and chest. All sensors recorded higher glucose concentration than venous samples during exercise. The arm and chest sites were most preferred with greatest sensor failures from the back.

Conclusions

FSL-CGM is clinically accurate when sensors are applied to alternate chest or back sites. Greater variability occurs during rapid change in glucose concentration with all sensor sites compared to venous glucose. Understanding these variabilities allows appropriate use of an economically viable CGM.

This article is protected by copyright. All rights reserved.