RESUMEN
Objective To explore the correlation between the glucose variability and the death of critically ill patients in neurology department.Methods Clinical characteristics of 231 patients admitted to neurointensive care unit (N-ICU) were analyzed retrospectively.The patients were divided into two groups:survival group(n =190)and death group (n =41).The baseline data,such as gender,age,mechanical ventilation,acute physiology and chronic health evaluation Ⅱ (APACHE Ⅱ) score were recorded.All data of blood glucose measured during hospitalization in N-ICU were collected.Standard deviation (SD),coefficient of variability (CV),maximum successive change and mean amplitude of glycemic excursions (MAGE) were used as glucose variability index.Logistic regression analysis was used to study the association between the glucose variability and the death.Receiver operating characteristic (ROC) curve was adopted to evaluate the application value of each glucose variability index on predicting mortality.Results Glucose variability,APACHE Ⅱ score (deleting the age component) and mechanical ventilation were significantly correlated with death (OR =4.959,1.444,5.472 ; 95％ CI 2.015 ～ 12.202,1.240 ～ 1.681,1.455 ～20.574;P =0.000,0.000,0.012).The predictive value was high and same when using SD and APACHE Ⅱ score (deleting the age component) as predictor (area under the ROC curve =0.910,0.957 ;Z =1.396,P =0.163).The application value of CV,maximum successive change and MAGE was moderate (area under the curve =0.847,0.856,0.872).Conclusion There is significant association between the glucose variability and the death of critically ill patients in neurology department.