Evaluation of Subcutaneous Glucose Monitoring Systems under Routine Environmental Conditions in Patients with Type 1 Diabetes: Glucose Monitoring evaluation in type 1 diabetes

Publication from Health

Felix Aberer, Martin Hajnsek , Markus Rumpler , Sabine Zenz, Petra Baumann , Hesham Elsayed, Adelheid Puffing, Gerlies Treiber, Thomas R. Pieber, Harald Sourij and Julia K. Mader

Diabetes Technology and Therapeutics , 2/2017


Continuous and flash glucose monitoring (GM) systems have been established in diabetes care. We compared the sensor performance of 3 commercially available GM systems. A total of 12 patients with type 1 diabetes were included in a single-centre, open-label study in which the sensor performance of the Abbott FreeStyle libre (Abbott), Dexcom G4 Platinum (Dexcom) and Medtronic MiniMed 640G (Medtronic) systems over 12hours was compared during mimicked real-life conditions (meals, exercise, hypo- and hyperglycaemia). Sensor performance was determined by fulfilment of ISO 15197:2013 criteria, calculating mean absolute relative difference (MARD), and was also illustrated using Parkes error grid and Bland-Altman plots. Sensor performance during changes in metabolic variables (lactate, betahydroxybutyrate, glucagon, non-esterified-fatty-acids) was determined by Spearman's rank correlation coefficient testing. The systems fulfilled ISO 15197:2013 criteria by 73.2% (Abbott), 56.1% (Dexcom) and 52.0% (Medtronic). The MARDs±standard deviation in the entire glycaemic range were 13.2%±10.9% (Abbott), 16.8%±12.3% (Dexcom) and 21.4%±17.6% (Medtronic), respectively. All sensors performed less accurately during hypoglycaemia and best during hyperglycaemia. We did not observe an influence of metabolic variables on sensor performance.