Prediction Quality of Glucose Trend Indicators in Two Continuous Tissue Glucose Monitoring Systems

Background: Continuous interstitial glucose monitoring (CGM) systems often provide glucose trend indicators (e.g., arrows) in addition to current glucose values. These indicators are recommended to be used in therapeutic decisions, because they are ascribed predictive qualities by CGM system manufacturers and expert committees. This study assessed how reliably trend indicators match future glucose change, because such information is missing. Methods: In a clinical trial, two different CGM systems were used by 20 participants, with two sensors of each system per patient. Participants used the systems for 14 days with three study site visits (48 h each). During study site visits, glucose trend indicators, as displayed by the CGM systems, were recorded at least once per hour during daytime and once at night in a diary. In addition, CGM data were downloaded from the devices. Trend indicators were compared with glucose change calculated from CGM data >30 min after recording the trend indicator. Results: Approximately 60% of trend indicators matched the glucose change calculated from CGM data. More than 10% of trend indicators differed by at least two trend indicator categories. Focusing on trend indicators recorded around carbohydrate (CHO) intake and insulin deliveries resulted in approximately half of trend indicators matching the calculated glucose change. Conclusions: Trend indicators do not always match future glucose change, especially within the first few hours after CHO intake and insulin deliveries. Manufacturers' labeling and recommendations should reflect this, so that CGM users can make informed decisions.