Study On Prognostic Evaluation Value Of Early Glycemic Lability Index In Critically Ill Patients

Until now, studies have shown that stress-induced hyperglycemia is closely related topoor prognosis of critically ill patients in intensive care unit. However, Whether intensiveinsulin therapy should be conducted to control blood sugar is still controversial. Besides,there are different outcomes despite of the same level of blood sugar in critically ill patients.Some studies suggest that poor clinical outcomes not only interrelate with the level ofblood sugar, but also relate to glycemic variability (GV). Therefore, reasons of GV,mechanism that GV causes poor outcomes, how to choose the best indicator to assess GVand the ability of GV to assess the prognosis of patients, these questions above arebecoming the new hot spots of research. This study selected critically ill patients inintensive care unit as object, GLI as criterion of GV,28-day outcome as prognosticindicator, to explore the value of early GLI in prognostic evaluation of critically ill patientsand to provide some basis for clinical practice in GV control.Objective: To explore the value of early GLI in prognostic evaluation of critically illpatients in intensive care unit. To look for the appropriate range of glycemic variabilityusing GLI as quantitative indicator, and to provide a reference for the monitoring andregulation of blood glucose in the clinical practice.Methods: We retrospectively analyzed68critically ill patients in intensive care unit(ICU). All patients were treated with a unified program of blood glucose monitoring andregulation. The patients were divided into Survival group(46cases) and Death group(22cases) according to28-day outcome. By comparing the age, gender, APACHEⅡscore,blood glucose concentration at admission (BGadm), mean blood glucose level(BGmean),incidence of hyperglycemia, incidence of hypoglycemia, GLI, we sought whether therewere statistical differences among the2groups，and analyzed the correlation between earlyGLI(24h)and APACHE Ⅱscore. ROC(Receiver operating curve) was drawn to evaluatepredictive values of GLI and APACHE Ⅱscore assessed during the early period after admission in critically ill patients. The cases were stratified according to the GLI or GLIcombining with BGmean. The prognosis was then compared between the subgroups.Results:①The age, gender, BGadm, BGmean, incidence of hyperglycemia showedno statistical differences between Survival group and Death group (P>0.05),but APACHEⅡ score and early GLI were significantly higher in Death group than Survival group, andthere was a significantly statistical difference (APACHE Ⅱ score:26.5vs21.2;earlyGLI:64.5vs35.1;P<0.01)(Figure4.1).②Early GLI and APACHE Ⅱscore had significantlypositive correlation(correlation coefficient=0.449,P<0.01). The predictive values betweenearly GLI and APACHE Ⅱscore in the prognosis of critically ill patients had no statisticaldifference (0.741vs0.714,P=0.64).③With the early GLI increasing, the28-day mortalityand ICU length of stay increased. The28-day mortality between subgroup Q2(GLI=12.94~26.56(mmol/L)2/h/d) and subgroup Q3(GLI=27.36~76.58(mmol/L)2/h/d) showedstatistical difference (11.8%vs47.1%,P<0.05).④The cases were divided into4subgroupsaccording to the median of GLI and BGmean. Whether in the hypoglycemia subgroup(<8.715mmol/L) or in the hyperglycemia subgroup (>8.715mmol/L),the higher GLIduring the first24hrs was correlate with the increasing of the28-day mortality(hypoglycemia subgroup:50%vs10%, hyperglycemia subgroup:55%vs14.3%,P<0.05).When early GLI were the same, the28-day mortality between hyperglycemia subgroupand hypoglycemia subgroup showed statistical no difference (lower early GLIsubgroup:10.0%vs14.3%, higher early GLI subgroup:55.0%vs50.0%,P<0.05). The28-daymortality of high BGmean+high GLI subgroup was the highest(55%).Conclusion: The early GLI is significantly associated with the severity of disease andprognosis, possessing the same prognostic evaluation value with APACHE Ⅱ score incritically ill patients. The higher early GLI is an risk factor for death of critically ill patients,and controling the early GLI might improve the patients outcome. To control the early GLIbelow26.56(mmol/L)2/h/d may be acceptable in critically ill patients.