Assessment of the Accuracy,Feasibility,Safety of Continuous Glucose Monitoring System

Acute glucose metabolic disorder (hyperglycemia, hypoglycemia and high glucose variability) is common during critical illness and associated with poor prognosis.Continuous glucose monitoring system (CGMS), which providing continuous glucose values, is promising to overcome these problems mentioned above and finally optimize glucose control.But its feasibility, accuracy and confounding factors are controversial.The aims of this study is to assess the accuracy, feasibility and safety of CGMS in daily routine in a general ICU in comparison with simultaneous laboratory arterial blood glucose (ABG) values

• Real-time continuous blood glucose monitoring system(CGMS) (Device): A total of 20 patients were enrolled for the place of CGMS.CGMS were placed on the right chest wall of each patients and recorded the interstitial glucose. Arterial blood glucose (ABG) were recorded every four hours. The duration of monitoring set was 5 days.CGMS were compared with ABG at the same time point. A total of 600 pairs of glucose level were collected

Primary outcomes

• The Spearman correlation analysis [CGMS will be used up to 5 days]: correlation coefficient (between 0.3 and 0.5):low correlation;(between 0.5 and 0.8);moderate correlation;(above 0.8):high correlation
• Bland-altman analysis [CGMS will be used up to 5 days]: more than 95% CGMS-arterial blood glucose(ABG) data pairs should be within the agreement range.
• The International Standardization Organization (ISO) criteria [CGMS will be used up to 5 days]: more than 95% of CGMS readings should be within 0.83mmol/L( of the reference values when the reference value is ≤5.56mmol/L and within 15.0% of the reference values when the reference value is >5.56mmol/L
• Mean absolute relative difference(MARD) and possible confounding factors [CGMS will be used up to 5 days]: The MARD in critical ill patients should not be more than 14%.Stratified MARD according to 7 factors(history of diabetes, blood glucose variability （determined using standard deviation of blood glucose）, glucose ranges(<4.4mmol/L,4.4-10mmol/L,>10mmol/L）, vasoactive drugs,continuous renal replacement therapy(CRRT), glucocorticoids, enteral nutrition(EN) and parenteral nutrition(PN) will be calculated and made a comparison.
• Clarke error grid analysis [CGMS will be used up to 5 days]: More than 98% data pairs should locate in "accuracy zone(Zone A)" and "clinically acceptable zone(Zone B)"
• Surveillance error grid [CGMS will be used up to 5 days]: There is no definite data field.The results will be analyzed comprehensively

Other outcomes

• Continuous data display/Total monitoring time [CGMS will be used up to 5 days]: CGM should continuously measure glucose and display in real time >95% of the time for the duration of total study period.
• device-related data gaps [CGMS will be used up to 5 days]: Skips in data acquisition should not exceed 30 minutes at a time
• Reasons for early disconnection of CGMS [CGMS will be used up to 5 days]: the number of each disconnection reason(3 categories:Accidentally,poor signal and early discharge)
• Number of Participants With Local complications, such as infection, bleeding, bruises and redness [CGMS will be used up to 5 days]: Number of Participants With Local complications, such as infection, bleeding, bruises and redness

Sex: All, ages: 18 Years to any

Inclusion Criteria:

* Inclusion criteria: Critically ill patients who were older than 18 years old and expected length of ICU stay longer than 48 hours

Exclusion Criteria:

* 1\) they were pregnant.2) had broken skin. 3) had a platelet count of less than 30 × 109/L 4) had participated in another trial.5) diabetic ketoacidosis or hyperosmotic coma.6) judged to be improper to complete this trial by the investigator

• Yan Kang, MD, Study director, West China Hospital