| Objective: The objective of the study was to screen out risk factors associated with poor in-hospital neurological function status after intravenous thrombolysis in Chinese patients with acute ischemic stroke,adjust and control confounding factors associated with such poor neurological function status,develop and validate a prediction model to predict the risk of poor neurological function status and deduce a visual individual nomogram.Methods:1.A total of 878 patients with acute ischemic stroke who received intravenous thrombolytic therapy from January 2018 to January 2020 at the two advanced Stroke Centers of grade III,Grade A hospitals in China were included in this study.Baseline data and treatment data of the subjects were collected.The National Institutes of Stroke Scale Health Stroke Scale(NIHSS)was used to evaluate the neurological function status of the patients.The NIHSS score ≥ 16 at day 7 after onset indicated that the patients had poor neurological function status.Univariate analysis and multivariate analysis were used to screen out factors associated with the poor neurological function status.2.The continuous variable of DNT was converted to the rank variable DNT1.After excluding the effects of interaction to the poor neurological function status,the initial crude model and adjustment models were established.In SPSS24.0,two to more risk factors related to both the variable of DNT1 and the outcome were involved for adjustment.The variable of OTT was adjusted and controlled in the same way.3.In the study,we had taken the least absolute shrinkage and selection operator(LASSO)to screen out the predictors and established the risk prediction model.The discrimination,calibration and clinical efficacy of the model in the training set and the validation set was verified,and finally we derived a visual nomogram for individual prediction.Results:1.After univariate analysis and multivariate analysis,variables of age[odds ratio(OR): 1.099,95% confidence interval(CI): 1.052～1.149],NIHSS2(NIHSS scored immediately after thrombolysis,OR: 1.286,95% CI:1.201～1.377,P<0.001),CHOL(OR: 1.614,95%CI: 1.036～2.514,P< 0.05),UREA(OR: 1.205,95% CI: 1.045～1.390,P<0.05),CT2(brain CT examined24 h after thrombolysis,OR: 6.153,95%CI: 2.696～14.045,P<0.001)and LDVT(lower extremity deep vein thrombosis,OR: 4.398,95% CI:1.560～12.398,P<0.05)were risk factors associated with poor neurological function status in hospital after intravenous thrombolysis.Variables of Lipid regulation(OR: 0.065,95% CI: 0.02～0.215,P<0.001),and HDL-C(OR:0.038,95%CI: 0.007～0.202,P<0.001)was found to be protective factors against poor neurological function status.2.Three models were established to adjust and control the confounding factors of variable DNT.Model 1 was the initial Model,and only variable DNT1 were brought into the model,the dummy variable was set with its first rank as reference,the OR value of the second grade was 1.413,95% CI was(0.73～2.737),the OR value of the third grade was 2.003,95% CI was(1.08～3.716),the P value for trend was 0.026.Model 2,the variable OTT and NIHSS1 were controlled at the same time.The OR value of the second grade was 1.024,95% CI was(1.002～1.273),and the OR value of the third grade was 1.063,95% CI was(1.009-1.458),the P value for trend was 0.004.Model3.All the confounding factors were adjusted.The OR value of the second grade of DNT1 was 1.054,95% CI was(1.007～1.436),and the OR value of the third grade was 1.019,95% CI was(1.002～1.231),the P value for trend was 0.088.Three models were established to adjust and control confounding factors of variable OTT.Model 1 was the initial Model,only OTT1 was brought into the model,and dummy variables were set based on its first rank.The OR value of the second rank was 1.368,95% CI(1.186～1.730),and the OR value of the third rank was 1.800.95% CI was(1.459-2.394),the P value for trend was0.004.Model 2,DNT and NIHSS1 were controlled,the OR value of the second rank was 1.538,95% CI was(1.227～2.276),and the OR value of the third rank was 1.935,95% CI was(1.478～2.829),the P value for trend was0.016.Model 3,all confounding factors were controlled,the OR value of the second rank was 1.180,95% CI was(0.265～5.259),the OR value of the third rank was 1.265,95% CI was(0.400～4.002),P value for trend was 0.039.3.By LASSO regression,variables of age,NIHSS1(NIHSS score before thrombolysis);NIHSS3(NIHSS score at 24 h after thrombolytic therapy);HDL-C;antiplatelet(antiplatelet therapy);CT2(brain CT reexamined 24 hours after thrombolysis);Lower extremity Venous color Doppler Ultrasound(the same as LDVT)were used as predictors to establish a risk prediction model.It has been verified in training set and validation set.This risk prediction model showed good discrimination(the areas under receiver operator characteristic curves(ROC)were 0.9626 and 0.9413 respectively),and good calibration(the maximum deviation of training set(Emax)=0.072,the average deviation(Eavg)=0.01,P=0.528,the maximum deviation of validation set(Emax)=0.123,the average deviation(Eavg)=0.019,P=0.594).The decision curve analysis(DCA)in the training set and validation set showed good performance,suggesting that the prediction model obtained had good clinical efficacy.Conclusions:1.Variables of age,NIHSS2,CHOL,UREA,CT2,and LDVT were risk factors for poor neurological function status in patients with acute ischemic stroke after intravenous thrombolysis.Variables of lipid regulation and HDL-C were protective factors to avoid poor neurological function status.2.After adjusting and controlling confounding factors,it was found that there was no trend correlation between DNT and the status of poor in-hospital neurological function after intravenous thrombolysis in acute ischemic stroke.OTT had a trend correlation with the status of poor in-hospital neurological function after intravenous thrombolysis in acute ischemic stroke.3.The prediction model obtained in this study had good discrimination,calibration and clinical efficacy,and the new nomogram derived could provide an reference to predict the risk of poor in-hospital neurological function status after intravenous thrombolysis in Chinese patients with acute ischemic stroke. |
PDF Full Text Request