IntroductionAccurate blood glucose measurements are important in persons with diabetes during normal daily activities (NDA), even more so during exercise. We aimed to investigate the performance of fluorescence sensor-based and glucose oxidase-based interstitial glucose measurement during (intensive) exercise and NDA.Research design and methodsProspective, observational study in 23 persons with type 1 diabetes when mountain biking for 6 days, followed by 6 days of NDA. Readings of the Eversense (fluorescence-based continuous glucose monitoring (CGM); subcutaneously implanted) and of the Free Style Libre (FSL; glucose oxidase-based flash glucose monitoring (FGM); transcutaneously placed) were compared with capillary glucose levels (Free Style Libre Precision NeoPro strip (FSLCstrip)).ResultsMean average differences (MAD) and mean average relative differences (MARD) were significantly different when comparing exercise with NDA (reference FSLCstrip); Eversense MAD 25±19 vs 17±6 mg/dL (p<0.001); MARD 17±6 vs 13%±6% (p<0.01) and FSL MAD 32±17 vs 18±8 mg/dL (p<0.01); MARD 20±7 vs 12%±5% (p<0.001).When analyzing the data according to the Integrated Continuous Glucose Monitoring Approvals (class II–510(K) guidelines), the overall performance of interstitial glucose readings within 20% of the FSLCstrip during exercise compared with NDA was 69% vs 81% for the Eversense and 59% vs 83% for the FSL, respectively. Within 15% of the FSLCstrip was 59% vs 70% for the Eversense and 46% vs 71% for the FSL.ConclusionsDuring exercise, both fluorescence and glucose oxidase-based interstitial glucose measurements (using Eversense and FSL sensors) were less accurate compared with measurements during NDA. Even when acknowledging the beneficial effects of CGM or FGM, users should be aware of the risk of diminished accuracy of interstitial glucose readings during (intensive) exercise.