新型冠状病毒肺炎患者入院血糖水平对其预后评估的临床价值研究

目的　探讨新型冠状病毒肺炎患者入院血糖水平对其预后评估的临床价值。方法　选取 2020年 1月 18日～ 2月 26日序贯入住华中科技大学同济医学院附属同济医院明确诊断为新型冠状病毒肺炎（ COVID-19）的 420例患者为研究对象，记录是否并发糖尿病、入院首次血糖水平和入院时临床分型及院内死亡事件等资料，依据是否并发糖尿病分为糖尿病组与非糖尿病组，入组后根据院内死亡事件分为生存亚组与死亡亚组，根据临床分型分为普通型亚组、重型亚组与危重型亚组，分析比较组内各亚组间入院血糖水平的差异，另根据入院血糖水平（ GLU）分为 GLU 3.9~7.8mmol/L亚组、

Clinical Value of Admission Blood Glucose Level on Prognosis of COVID-19 Patients

Objective　To explore the clinical value of admission blood glucose level on prognosis of COVID- 19 patients.Methods　A total of 420 novel coronavirus pneumonia (COVID-19) patients admitted to Tongji Hospital of Tongji MedicalCollege from January 18, 2020 to February 26, 2020 were selected as the subjects of study. The data of diabetes or not,admissionblood glucose level(GLU), clinical severity grade were collected through the electronic medical record system,and the outcome,which defined as in-hospital motality, was also monitored. The patients were divided into diabetes group and non-diabetes groupin terms of the complication of diabetes, and then, firstly, stratified these two groups into survival subgroup and non-survivalsubgroup in according to the event of in-hospital motality, GLU between these two subgroups were compared.Secondly,according to the clinical severity grade, these two groups were stratified into moderate subgroup, severe subgroup and criticalsubgroup, and GLU among these subgroups were also compared. Thirdly, according to the admission blood glucose level,stratified these two groups into GLU 3.9~7.8 mmol /L subgroup,GLU 7.8~10.0mmol/L subgroup and GLU>10.0mmol/Lsubgroup, the in-hospital motality rates among these subgroups were compared. Finally,the multivariate logistic regression wasused to explore whether increased GLU were independent risk factor for in-hospital motality in diabetes group and non-diabetesgroup when adjusted for sex, age and underlying disease. Results　In non-diabetes group,compared with Survival subgroup,GLUwas significantly elevated in non-Survival subgroup[6.96（5.95, 8.23）mmol/L vs 5.96 (5.32, 6.92) mmol/L],the difference wasstatistically significant（U=6047.0, P<0.001）, but in diabetes group,there was no significant difference between non-survivalsubgroup and Survival subgroup [12.42(8.41, 18.17) mmol/L vs 9.88 (7.79, 14.02) mmol/L], the difference was statisticallysignificant（U=1 200.5, P=0.059）.In Non-diabetes group, GLU elevated remarkably along with the clinical severity gradeincreased, moderate subgroup, severe subgroup, critical subgroup GLU were 5.87(5.24, 6.69) mmol/L, 6.94(5.95, 7.90) mmol/L,9.73 (6.22, 11.64) mmol/L, the difference were statistically significant, respectively(U=723.0~4978.0,all P<0.01). However indiabetes group,there was no significant difference on GLU when the clinical severity grade increased, moderate subgroup, severesubgroup, critical subgroup GLU were 9.88（7.81, 11.93）mmol/L, 12.42（8.43, 16.94）mmol/L, 11.43(7.89, 18.76)mmol/L,the difference were statistically significant ，respectively (U=262.0~946.5, all P>0.05).In non-diabetes group, GLU> 10.0 mmol/L subgroup had the hightest in-hospital motality rate (72.0%) among all three subgroups, the differences were statisticallysignificant（χ2=24.607, 9.625, all P<0.01）, when compared between GLU 3.9~7.8mmol/L subgroup (in-hospital motality rate24.8%) and GLU 7.8~10.0mmol/L subgroup (in-hospital motality rate 30.0%), there was no significant difference on in-hospitalmotality rate (χ2=0.383, P > 0.05).However, in diabetes group, along with GLU increased, it had no significant difference on inhospitalmotality rate, GLU 3.9~7.8mmol/L subgroup, GLU 7.8~10.0mmol/L subgroup, GLU> 10.0mmol/L subgroup, the inhospitalmotality rate were 34.8%, 41.4%, 49.2%, respectively(χ2=0.236~1.380, all P> 0.05). Multivariate logistic regressionshowed, in non-diabets group, GLU>10.0 mmol/L was the independent risk factor when adjusted for sex, age and underlyingdisease,odds ratio was 7.969, and 95% confidence interval was 3.022~21.013,but in diabets group.It seemed that GLU>10.0mmol/L was not the independent risk factor. Conclusion　Admission blood glucose is a good predictor for disease severity andoutcome in non-diabetes patients with COVID-19. When admission hyperglycemia occurs, it tends to predict a poor prognosis.