Addition Of LDL-C To ABCD3Score To Predict Patients At Aarly Risk Of Stroke After Transient Ischaemic Attack

Objective: We aimed to predict ABCD3-L score for early risk of stroke aftertransient ischaemic attack （TIA）. To evaluate the predictive value of three predictivemodels: ABCD2，ABCD3and ABCD3-L, for early stroke risk after TIA and seclectthe best one.Methods: We analysed pooled data from patients with clinically defined TIAwho were investigated while in secondary care in our hospital neurology department.Items that contribute to all clinical variables were recorded. We score patientsaccording to the standard of ABCD2score, ABCD3score and ABCD3-L score,respectively. We divided the150patients into three group-risk stratification: low（score0–3），moderate （score4–5） and high （score≥6）. Stroke status within2and7days was recorded by in-person assessment，and observe the incidence ofcerebral infarction. We draw Receiver operating characteristic curve and view the areaunder the curve and then evaluate the predictive value of three predictive models forearly stroke risk after TIA. Analyse the relationship among risk factors, different riskstratification and incidence rate of early stroke occurrence after TIA based on χ2test.Multivariate logistic regression analysis was done with stroke occurrence at early timeintervals after onset of TIA as the dependent variable, with risk factors of independentclinical variables. Statistical analysis was done with SPSS version18.0. The p<0.05is referred to as a significance level.Results:1、150patients were included in the study. Patients who had a “dualtransient ischaemic attack” was25（16.67%）and with the level of LDL-C≥2.6mmol/L was37（24.67%）. Incidence rate of stroke occurrence within7days aftera TIA was30（20%），and20（13.33%）in2days.2、In the ABCD2、ABCD3and ABCD3-L three predictive models, the area underthe curve and95%Confidence Interval for predicting stroke risk within2days after a TIA were0.703（0.581-0.824）、0.786（0.676-0.896）、0.845（0.756-0.934），respectively, and the largest area under the curve was ABCD3-L score. The threepredictive models did differ for prediction of2days stroke risk after a TIA, p<0.05.3、In the ABCD2、ABCD3and ABCD3-L three predictive models, the area underthe curve and95%Confidence Interval for predicting stroke risk within7days after aTIA were0.758（0.663-0.852）、0.834（0.752-0.917）、0.878（0.808-0.948），respectively, and the largest area under the curve was ABCD3-L score. The threepredictive models did differ for prediction of7days stroke risk after a TIA, p<0.05.4、The difference of the incidence of cerebral infarction was significant amongthe low，moderate and high risk stratification, using ABCD2、ABCD3and ABCD3-Lthree predictive models to assess early stroke risk within2and7days after a TIA,respectively，p<0.05.5、The single factor analysis between the group with a stroke occurrence and thegroup not showed that difference being significant in the two groups, aboutBP≥140/90mmHg、 dual transient ischaemic attack、 the level of LDL-C≥2.6mmol/L、Hypertension、Diabetes mellitus, p<0.05.6、In the multivariate logistic regression analysis, we finded that there wasclosely related to stroke at early time intervals after onset of TIA, which was the levelof LDL-C≥2.6mmol/L、dual transient ischaemic attack、Hypertension、Diabetesmellitus、BP≥140/90mmHg，（OR=3.38-10.43，P＜0.05），and both95%CI＞1. There were each independent predictors of stroke risk after TIA.Conclusion: The modified version of the ABCD3-L score appeared to besuperior to the original version （ABCD3and ABCD2score） for predicting earlystroke risk within2and7days after a TIA.