Risk Factors Associated With Early Neurological Deterioration In Ischemic Stroke

Objective: We performed a prospective cohort study to explore the risk factors associated with early neurological deterioration(END)in stroke and to explore the possible pathogenesis of early prevention.Methods: According to the inclusion and exclusion criteria,274 patients were admitted with acute ischemic stroke in the acute phase of our hospital(January 2017-December 2017)prospectively assessed.All patients were assessed the degree of neurological deficit by the National Institutes of Health Stroke Scale(NIHSS).Within 7 days of onset of disease assess NIHSS scale again,compared with two scores before and after the increase,the scores of ?4points was admitted in END group(n=60)and others in non-END group(n=214).(1)To compare the general data(sex,age,smoking history,hypertension,diabetes mellitus,history of recent infection)and laboratory indicators within 24 hours of admission(random blood glucose,homocysteine,low density lipoprotein,urea nitrogen / creatinine,urine specific gravity),TOAST classification and the incidence of post-stroke infection.(2)Application of the ROC curves analysis on the threshold of all measurement data.(3)In addition,we will analyze the correlation strength between different risk factors and END through the multivariate Logistic regression analysis,and set up a regression model.Results: 1.The participants included 274 patients with AIS,including END group of 60 cases,the non-END group of 214 case.The incidence of END is 21.90%.2.Comparison tow groups of general data,in age,gender,smoking history,history of diabetes and hypertension showed no statistically significant(P> 0.05).3.The comparison of two groups of laboratory indexes,Random blood glucose(RBG),blood urea nitrogen / creatinine(BUN/Cr),urine specific gravity(SG),homocysteine(HCY)in END group are all higher than non-END group(P <0.05),especially RBG and BUN/Cr(P <0.01),but the results showed that the LDL levels compared the differences between the two groups was no statistical significance.4.By plotting the ROC curve to predict the likelihood of END occurrence,it was found that the predicted value of END was moderate for RBG(AUC = 0.705 95% CI: 0.636-0.774)and SG(AUC = 0.704 95% CI:0.623-0.785);The predicted value of BUN/Cr for END was lower(AUC = 0.634 95% CI:0.555-0.712),and neither Sen nor Spe was high(Sen 63.3%,Spe 58.9%);The value of HCY for END was predicted to be low(AUC = 0.662 95% CI: 0.586-0.739),sensitivity was moderate but specificity was lower(Sen 75%,Spe 58.9%).5.END group were higher than non-END group in infection,particular recent infection.END group compared with the recent infection centered on1 W before stroke,the recent infection of non-END group was centered on 2W-4W.6.The effect of TOAST on the occurrence of END was analyzed.It showed that there was a significant difference in TOAST between the two groups(P=0.006).Further comparison results showed that CE type is more prone to END(P<0.05).7.Logistic regression analysis showed that SG,RBG,HCY,infection were the risk factors for END,but BUN/Cr and TOAST did not enter the model.At the same time,we get the regression model: Y=-3.197+3.622×infection+0.12×RBG+0.047×HCY+1.278×SG.Conclusion: 1.The incidence of END is as high as 21.9%,which deserves attention.2.Infection,SG,HCY,and RBG can predict END independently.3.the regression model: Y=-3.197+3.622×infection+0.12×RBG+0.047×HCY+1.278×SG.