| Objective:1.To explore the high-risk factors of Early Neurological Deterioration(END)of acute cerebral infarction with different etiological mechanisms based on TOAST classification.2.To evaluate the clinical value of WORSEN score in predicting END in different subtypes of acute cerebral infarction based on TOAST classification,as well as to optimize the predictive scoring system by combining the high-risk factors.Method:In this study,399 patients with acute cerebral infarction who were admitted to the Department of Neurology of The First Affiliated Hospital of Dalian Medical University Hospital from June 2019 to June 2022 were retrospectively selected,and according to the Trial of Org10172 in Acute Stroke Treatment(TOAST),a heparin-like drug widely used at home and abroad,the subgroups of large atherosclerotic(LAA),small vascular occlusion(SVO)and cardioemblism(CE)were selected(only patients with cerebral infarction caused by nonvalvular atrial fibrillation were included).And other undetermined etiology(OE)and unidentified etiology were excluded.Two neurologists evaluated the scores of National Institutes of Health Stroke Scale(NIHSS)for seven consecutive days during their stay in the hospital.In this study,the NIHSS score increased by≥2 points,the limb dyskinesia score increased by≥1point,or the new neurological symptoms appeared were defined as END group,and vice versa.The clinical data collected include:(1)demography:gender and age;(2)past medical history:hypertension,diabetes,old myocardial infarction and cerebral infarction;(3)laboratory examination:blood routine,D-dimer,fibrinogen,blood lipid,fasting blood glucose and glycosylated hemoglobin;(4)score:NIHSS score and WORSEN score;(5)imaging examination:location and diameter of intracranial infarction,as well as the number and severity of atherosclerosis of cerebral arteries.To analyze the clinical risk factors of END in different subtybes of acute cerebral infarction,and calculate the area under the curve(AUC)by using the receiver operating characteristic(ROC)curve to evaluate the clinical efficacy in predicating END in different subtybes of acute cerebral infarction,so as to optimize the END prediction system.SPSS25.0 software was used for statistical analysis of the idea,including independent sample T test,rank sum test,chi-square test and binary Logistic regression analysis.Results:1.There were 153 patients in LAA group,and 37 patients developed END,the incidence rate was 24.2%.The comparison of univariate analysis between the two groups showed that the systolic blood pressure,NIHSS score,fasting blood glucose level,homocysteine level and Internal carotid artery(ICA)occlusion ratio in END group at admission were higher than those in non-END group,the difference was statistically significant(P<0.05).Multivariate Logistic regression analysis showed systolic blood pressure(OR=1.026,95%CI:1.004～1.047,P=0.018),NIHSS score(OR=1.212,95%CI:1.039～1.415,P=0.015),fasting blood glucose level(OR=1.195,95%CI:1.013～1.409,P=0.05),P=0.034)was an independent risk factor for the END of LAA type cerebral infarction.The AUC of WORSEN score in predicting the END of LAA cerebral infarction was 0.777(95%CI:0.691～0.863,P=0.000),the cutoff value was 3,the sensitivity was 78.4%,and the specificity was 69.8%.2.There were 150 patients in SVO group,and 34 patients developed END,the incidence rate was 22.7%.Univariate analysis and comparison between the two groups showed that the proportion of past myocardial infarction and diabetes,NIHSS score,fasting blood glucose level,D-dimer level and glycosylated hemoglobin level in END group were higher than those in non-END group,the difference was statistically significant(P<0.05).Multivariate Logistic regression analysis showed that NIHSS score(OR=1.95%CI:1.183～1.851,P=0.001)and fasting blood glucose level(OR=1.353,95%CI:1.019～1.795,P=0.036)at admission were independent risk factors for the END of SVO subtype.The AUC of WORSEN score in predicting the END of acute cerebral infarction was 0.838(95%CI:0.580～0.813,P=0.005),the cutoff value was 1,the sensitivity was 90.9%,and the specificity was 40.5%.3.Among 96 patients in CE group(only including nonvalvular atrial fibrillation),22 patients developed END,the incidence rate was 22.9%.Comparison of single factor analysis between the two groups showed that the plasma fibrinogen concentration and the number and severity of intracranial artery stenosis in END group were higher than those in non-END group,the difference was statistically significant(P<0.05).Multivariate Logistic regression analysis showed that the plasma fibrinogen concentration(OR=2.549,95%CI:1.163～5.585,P=0.019),the number of intracranial stenotic arteries(OR=1.846,95%CI:1.129～3.018,P=0.015)were independent risk factors for the END of CE subtype.The AUC value of WORSEN score in predicting the END of CE-type was 0.697(95%CI:0.580～0.813,P=0.005),the cutoff value was1,the sensitivity was 90.9%,and the specificity was 40.5%.4.Combining the number of intracranial stenotic arteries and plasma fibrinogen concentration,the prediction model of CE-type cerebral infarction END was established:Logit(END)=-5.217+0.505*WORSEN score+0.897*fibrinogen concentration+0.748*number of intracranial artery stenoses,and the model was statistically significant(χ~2=19.70,P=0.002).Through the goodness-of-fit test of Hosmer-Lemeshow,the results showed that Hosmer-Leme showχ~2=7.891,P=0.444,and it had good calibration ability.The AUC value of predicting the END of CE cerebral infarction was 0.791(95%CI:0.678～0.905,P=0.000),the cutoff value was 0.203,the sensitivity was 77.3%and the specificity was 73.6%.The AUC value of Bootstrap repeated sampling method was 0.797(95%CI:0.657～0.885,P=0.000).Conclusion:1.Independent risk factors of patients with different subtybes of acute cerebral infarction according to the TOAST classification developed END were different.Higher baseline NIHSS score,higher fasting blood glucose level and high systolic blood pressure in LAA cerebral infarction patients were independent risk factors for END occurrence.Higher baseline NIHSS score and higher fasting blood glucose levels were independent risk factors for END of SVO infarction.2.The WORSEN score had different clinical efficacy in predicting END in patients with different subtypes of acute cerebral infarction.In TOAST classification,the WORSEN score was more suitable for predicting the END of LAA and SVO cerebral infarction,but had poor applicability in predicting END for CE subtybe infarction.3.Among patients with CE subtype acute cerebral infarction,the WORSEN score combined with the number of intracranial artery stenosis and fibrinogen concentration were more effective than the WORSEN score alone,had better sensitivity and specificity,it was an effective evaluation tool for clinical workers to judge END. |
PDF Full Text Request