| Objective:To evaluate the impact of continuous glucose monitoring system (CGMS) on glycemic control in critical illness.Methods:The patients of this observation study were collected in an Emergency Intensive Care Unit (EICU) department in a university tertiary hospital from July 1,2015 to December 31,2015. According to the inclusion criteria, CGMS and selective glucose monitoring techniques (Arterial blood gas analyzer glucose, ABG & Capillary Blood Glucose Monitoring device, CBG) were applied to the eligible patients. The general clinical data were collected, including demographic information (gender, age, etc.), admission reason, history of diabetes, Baseline glucose value, Baseline blood pH, Baseline lactate, Baseline blood pressure, Glasgow coma scale (GCS) on admission day, Acute Physiology and Chronic Health Evaluation II (APACHE II) on admission day, injury severity score (ISS), nutrition risk screening 2002 (NRS2002), ABG blood values, CBG blood values, CGMS blood values. Blood glucose variability (GV) data were also collected, including the average blood glucose, maximum blood glucose, minimum blood glucose, standard deviation (SD), glucose lability index, (GLI), mean amplitude of glycemic excursions (MAGE), Coefficient of variation (CV). The length of hospital stay, length of ICU stay, mechanical ventilation time and outcomes were also collected. The Pearson correlation coefficient and a Bland-Altman plot were applied to evaluate the correlation between CGMS and ABG The proportions of glucose values in different ranges monitored by three techniques were compared. CGMS values were also compared according to the history of diabetes and between trauma and non-trauma patients. The difference between ICU survivors and ICU non-survivors were analyzed, and receiver operating characteristic curve (ROC) was used to evaluate the prospective values.Results:1. A total of 52 critical ill patients were enrolled in the study. Of these patients,38 were males (73.1%) and 14 were females. The average age was 52±18; the median and interquartile range of APACHE II score was 18 (14-30).36 (69.2%) of the enrolled patients were severe trauma patients. The length of ICU stay was 9.0 (6.0-19.0) days and the mortality was 28.8%(15/52).2.456 CGMS readings and 456 ABG readings, in a 1:1 ratio, were obtained from the patients. CGMS and ABG had a Pearson correlation coefficient of 0.951 (R2=0.904, P＜0.001). A Bland-Altman plot showed that the mean of difference between 456 pairs glucose values was-0.04mmol/L,5.26% of the readings were out of the 95% confidence interval.3. Glucose readings of CGMS were taken 42,460 times. Glucose measurements of CBG were performed 763 times and ABG were 456 times. The glucose readings lower than 4.44mmol/L were 784 (1.85%)、5 (0.65%)、7 (1.54%), respectively (P=0.046). The glucose readings lower than 4.44mmol/L were 60 (0.14%),0 (0.00%),2 (0.44%) (P=0.141). And the number of patients with severe hypoglycemia (<2.22mmol/L) were 3(5.77%),1,0, respectively.4. There were 18 (34.6%) diabetes patients in the study. The average blood glucose, maximum blood glucose and minimum blood glucose were higher in diabetes patients. Meanwhile, glucose SD, GLI, CV and MAGE were all higher in diabetes patients compared to the non-diabetes patients (P＜0.001, P=0.017, P=0.013, P＜0.001).5. The enrolled patients were composed of 36 trauma patients and 16 non-trauma patients. The median and interquartile range of APACHE Ⅱ score were 17 (13-26),26 (17-32) (P=0.035). The mortality was 22.2% (8/36),43.8% (7/16) (P=0.114). The ABG levels and CBG levels were significantly higher in non-trauma patients compared to the trauma patients (P=0.035、P=0.006).6. The overall mortality was 28.8%(15/52). APACHE Ⅱ scores and SOFA scores were significantly higher in ICU non-survivors (P=0.005,0.007). Baseline glucose values were significantly higher in ICU non-survivors compared to survivors (15.20 (7.40-16.70) mmol/L versus 8.90 (7.86-12.00) mmol/L, P=0.018). The GLI and MAGE were all higher in non-survivors (P＜0.001,P=0.031). A significant decrease tread of GLI was observed in survivors compared to the non-survivors. The ROC about prospective value of GLI showed an area under the receiver operating characteristic curve (AUC) of 0.843, higher than APACHE Ⅱ score (0.749) and SOFA score (0.739).7.36 (69.2%) severe trauma patients were included in the study. The mortality of was 22.2% (8/36). The ISS scores, APACHE Ⅱ scores and SOFA scores were all higher in trauma non-survivors (P=0.012,0.002,0.001). Compared to trauma survivors, the GLI was significantly higher in trauma non-survivors (0.89 (0.58-1.16) versus 0.30 (0.23-0.50), P=0.004). In addition, a significant decrease tread of GLI was observed in trauma survivors. The AUC about the prospective value of GLI (0.826) was lower than APACHE Ⅱ scores (0.844) and SOFA scores (0.857).Conclusion:The continuous glucose monitoring system (CGMS) was more advantage mTCU critical illness compared to selective glucose monitoring techniques according to this single center observational study. The results of CGMS were promising and accuracy enough for the use in ICU and the percentage of hypoglycemia detected by CGMS was significantly higher. Clinicians can get continues glucose change curve and GV values, which can help to control the blood glucose more conveniently in critical illness and provide a new method for prognosis diagnosis. |
PDF Full Text Request