研究目的
Assessing how reliably glucose trend indicators in continuous interstitial glucose monitoring (CGM) systems match future glucose changes, as such information is missing and these indicators are used in therapeutic decisions.
研究成果
Trend indicators in CGM systems do not always accurately predict future glucose changes, with approximately 60% matching and over 10% differing by at least two categories. Accuracy is lower around times of carbohydrate intake and insulin delivery, where only about half match. Manufacturers should update labeling to reflect these limitations, and users should be educated to interpret trends cautiously, especially after meals or insulin administration, to avoid potential hypoglycemia or hyperglycemia risks.
研究不足
The study is limited to two specific CGM systems (Dexcom G5 and FreeStyle Libre) and may not generalize to other systems. The analysis did not account for physical exercise, which could affect glucose trends. The clinical relevance might be limited if patients adapt based on experience, and the systems did not include predictive management features at the time of the study. The sample size of 20 participants might not capture all variabilities in diabetes management.
1：Experimental Design and Method Selection:
An investigator-initiated, open-label, single-center, single-arm, interventional clinical trial was conducted to compare trend indicators from two CGM systems (Dexcom G5 Mobile and FreeStyle Libre) with glucose changes calculated from tissue glucose (TG) readings. The design involved participants using the systems for 14 days with three 48-hour study site visits. Trend indicators were recorded manually in a diary during site visits, and TG data were downloaded from devices for analysis. Linear interpolation was used to calculate glucose changes over 30 minutes after trend indicator recording.
2：Sample Selection and Data Sources:
20 adult participants with type 1 diabetes were selected from a subject database. Eligibility criteria included age at least 18 years, use of multiple daily injections or continuous subcutaneous insulin infusion, and exclusion of severe illnesses or other risk factors. Data sources included manual diary entries of trend indicators and downloaded TG data from CGM sensors.
3：List of Experimental Equipment and Materials:
Two CGM systems: Dexcom G5 Mobile (DG5) and FreeStyle Libre (FL), each with two sensors per participant. Blood glucose meters for reference measurements. Diaries for manual recording. Computers for data download and analysis.
4：Experimental Procedures and Operational Workflow:
Sensors were applied to participants (FL on upper arms, DG5 on abdomen). During study site visits, blood glucose measurements were taken at least hourly from 0600 to 2400 and once at night at 0300, with parallel scanning of FL and recording of trend indicators from all sensors. Dynamic phases with induced glucose changes were included. Data were analyzed by comparing recorded trend indicators with glucose changes calculated from TG readings 30 minutes later.
5：Data Analysis Methods:
Trend indicators were categorized based on manufacturer definitions. Contingency tables and descriptive statistics were used to compare indicated and calculated trends. Stratification by study visit and analysis excluding times around carbohydrate intake and insulin delivery were performed.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
