脑静脉和静脉窦血栓形成的临床特点、近期预后及多因素预测模型

Cerebral Vein and Dural Sinus Thrombosis: Clinical Features, Prognosis and Predictive Models

目的描述脑静脉和静脉窦血栓形成（cerebral vein and sinus thrombosis，CVST）的临床表现、影像学特点、诊治方法和预后，评价预后的影响因素。方法在单中心观察性研究中，连续性入选我院神经科诊断为CVST患者，进行回顾性分析。收集患者的人口学信息、症状和体征、辅助检查结果（脑电图、磁共振和全脑血管造影）和诊断治疗方法。随访在患者出院前一天完成。神经功能结局通过改良Rankin量表（modified Rankin Scale，mRS）评定， mRS≤2定义为神经功能结局良好。对神经功能预后进行单因素分析和多因素分析。应用受试者工作曲线（receiver operating curve，ROC）的曲线下面积（area under curve，AUC）评价多因素模型，并与CVST预后量表比较。结果2000年5月至2010年6月，共有171例CVST住院患者纳入研究。头痛是常见的症状，占91.8%，癫痫占36.5%，意识障碍占33.3%。89例（50.2%）患者出现脑梗死，54例（31.6%）出现脑出血。45例（26.3%）患者直窦受累。126例（73.7%）患者得到抗凝治疗，17例患者（10%）进行了溶栓治疗。随访时，137例（80.1%）患者预后良好（mRS≤2）。多因素分析显示年龄≥37岁[相对危险（relative risk，RR）3.25，95%可信区间（confidence interval，CI）1.29～8.17]、性别（女性，RR 4.06，95%CI 1.43～11.5）、意识障碍（RR 3.92，95%CI 1.48～10.4）、直窦血栓形成（RR 7.78，95%CI 1.66～36.4）、抗凝治疗（RR 0.13，95%CI 0.05～0.35）与预后独立相关。本研究预测模型（不包括治疗因素）和CVST预后量表（取值范围0～9）的预测准确性由AUC表示，分别为0.85（95%CI 0.77～0.92，P＜0.001）和0.65（95%CI 0.53～0.77，P＜0.01）。两模型的预测价值差异有显著性（P＜0.01）。结论 CVST患者预后相对良好，部分患者预后不良。多元统计分析显示，年龄（≥37岁）、女性、意识障碍、直窦血栓形成和预后独立相关。在本研究中，CVST的神经功能预后可以较准确地预测，需要进一步前瞻性多中心研究进行验证。

Objective To describe the clinical features of cerebral vein and sinus thrombosis （CVST）, to evaluate the short-term prognosis of CVST and to determine the predictive factors for CVST. Methods In a single center observational study, consecutive patients who were diagnosed as having CVST were retrospectively investigated. Demographic data, symptoms and signs from onset to diagnosis, clinical and imaging features, location of the thrombus and therapy were collected. Patients were followed up on the day before discharge. Primary outcome as independence was assessed by modiifed Rankin Scale （mRS） score 2 at the end of follow-up. Univariate analysis and multivariate Logistic regression analysis were performed to predict the outcome. Receiver operating characteristic （ROC） curves were used to analyze the accuracy. The comparison of area under the curve （AUC） was calculated to test the superiority of the predicting model to the previous clinical score. Results From May 2000 to Jun 2010, we recruited 171 adult patients with CVST. Headache （91.8%） was the most common symptom, followed by seizure （36.5%） and disturbance of consciousness （33.3%）. 50.2% （n=89） of the patients suffered from cerebral infarction, 31.6% （n=54） from cerebral hemorrhage. A straight sinus thrombosis was conifrmed in 45 （26.3%） patients in the cohort. Anticoagulation was initiated in 126 （73.7%） patients and thrombolysis was needed in 17 （10.0%） patients. At the end of follow-up, 137 patients （80.1%） were independent by mRS 2. Multivariate Logistic regression analysis showed independent factors were age （ 37, relative risk [RR] 3.25, 95%conifdence interval [CI] 1.29-8.17）, sex （female, RR 4.06, 95%CI 1.43-11.5 ）, conscious disturbance （RR 3.92, 95%CI 1.48-10.4）, straight sinus thrombosis （RR 7.78, 95%CI 1.66-36.4） and coagulation （RR 0.13, 95%CI 0.05-0.35）. The AUC of the predicting model in this study （ with the exception of anticoagulation ） and CVST score （range from 0 to 9） were 0.85 （95%CI 0.77-0.92, P〈0.001） and 0.65 （95%CI 0.53-0.77, P〈0.01） respectively. There was a signiifcant difference between the two models （P〈0.01）. Conclusion The prognosis of CVST is relatively good and a subgroup （19.9%） of CVST patients is at increased risk of bad outcome. The short-term outcome of CVST patients can be predicted well in our study and need to be evaluated prospectively.