Using Essen Stroke Risk Score Combined With Cerebrovascular Imaging To Predict The Risk Of Recurrent Vascular Events In Patients With Non-cardiac Ischemic Stroke

[Background]"China Cardiovascular Report2011" indicates that the morbidity of stroke patients in China increased at an annual average rate of12.12%.40%of patients surviving from cerebral infarction experience recurrence, and the disability and mortality are proportional to the frequency of recurrent stroke. These findings state that both of the risk of stroke recurrence and stroke mortality are extremely high. Recent researches have demonstrated that the medication and therapy in secondary prevention of stroke could efficiently reduce the incidence of recurrence. However, due to the fact that each individual shows distinct conditions and types of strokes, the risk of recurrence reveals differences between stroke patients. Thus, in order to favor the decentralization of stroke care and conduct an effective secondary prevention in clinical practice, the risk of recurrence should be predicted based on various assessment scales in stroke. The Essen Stroke Risk Score which has been widely used in stroke population in western countries is designed to predict the risk of recurrence in non-cardiogenic ischemic stroke patients, while the ischemic events still occur in the patients assessed as low risk. Therefore, it is essential to elevate the prediction of recurrent non-cardiogenic ischemic stroke by further improving ESRS.[Methods] We included the admitted patients with non-cardiogenic ischemic stroke in the Department of Neurology at Huashan Hospital affiliated to Fudan University from June1st,2007to May2nd,2009. We collected all the general medical history, past medical history, biochemical blood tests as well as extracranial intracranial vascular imaging of enrolled patients. By using vascular imaging in sequence of DSA, carotid ultrasonography, and CTA or MRA, blood vessels could be classified into4categories in terms of degree of stenosis, that are: no stenosis, mild stenosis (blood vessels narrow<50%), intermediate stenosis (blood vessels narrow≥50%without occlusion), and vascular occlusion. We further combined ESER with the vessel assessment and then come up with2new models. The first model is10scoring model (ESRS10): we add additional vessel assessments to existing Essen Stroke Risk Scoring system, and calculating the scores with the following method:blood vessels narrow≥50%=1point, blood vessels narrow<50%=0point, total score is10points. The second model is11scoring model (ESRS11). In this model, the vessel assessment of ESRS10was modified with the other scoring method, that is, blood vessels narrow≥50%=2point, blood vessels narrow<50%=0point, total score is11points. We followed-up the endpoints of all enrolled patients, including recurrent ischemic events and all-cause mortality. Then we analyzed the association between various degrees of blood vessel stenosis and endpoints, the relation between different ESRS and endpoints, and the correlation between new scoring models (ESRS10/ESRS11) and endpoints, respectively. SPSS17.0software package was used for statistical data analysis. Differences are considered statistically significant at p<0.05.[Results]1.After ruling out cardiogenic cerebral embolism, we totally enrolled323patients with acute ischemic stroke. Because60cases were lost to follow up, we ultimately included263effective cases. By following up the endpoints in enrolled patients occurring within3year after their first onset, we find that the occurrence of endpoints within the first year is7.9%, the second year is6.9%, and the third year is5.4%.2. During the first year post-stroke,25recurrent ischemic events had occurred. We used ROC curve to evaluate the performance of ESRS in predicting the recurrent risk in our department. The area under the ROC curve and its corresponding95%confidence interval are:AUC=0.667,95%CI:0.55-0.77, P<0.001, which demonstrates that ESRS could effectively predict the risk of recurrent ischemic stroke in our department.3. The comparison of the relation between endpoints and different degree of blood vessel stenosis unveils that the baseline levels of stroke risk factor between each groups are basically the same. However, compared with the patient with stenosis<50%, the patients with stenosis≥50%are more likely to meet the endpoints. x2=4.856, P=0.028. The area under the ROC curve of new ESRS10model and its95%confidence intervals are: AUC=0.722, P<0.001,95%(CI):0.62-0.82. The area under the ROC curve of new ESRS11model and its95%confidence intervals are:0.749, P<0.001,95%(CD:0.65-0.84. These statistics indicate that new ESRS model can effectively predict the risk of recurrent ischemic event in our department.4. Statistically comparing the area under the ROC curve of ESRS11model and ESRS10model. Z=3.15>1.96, P<0.05. The differences between these2models are considered statistically significant.[Conclusions] ESRS can predict the endpoints in patients with non-cardiogenic ischemic stroke in our department. Blood vessel stenosis is an independent risk factor for recurrent ischemic stroke. The degree of stenosis can effectively reflect the short-term (within1year) risk of recurrent non-cardiogenic ischemic stroke, while it cannot better predict the long-term (within3year) risk of recurrence. The new model combined with vessel assessments could elevate the prediction of original ESRS. Thus, it is better than existing ESRS.