第一部分 神经精神狼疮临床预后分析

目的：对北京协和医院神经精神狼疮患者进行临床预后分析，包括长期生存、中间结局、亚型预后及影响因素相关研究。建立北京协和医院单中心神经精神狼疮预后研究队列，为神经精神狼疮多中心、前瞻性队列建立及长期预后研究提供框架及基础，并指导后续治疗及预后相关研究。

方法：对北京协和医院2012年6月至2016年6月期间风湿免疫科住院的神经精神狼疮（NPSLE）患者进行回顾性分析。根据入选标准及排除标准筛选出明确诊断并在此次住院病程中出现NPSLE的患者。对NPSLE患者进行人口学、病史、检查结果、治疗等数据进行录入，通过门诊病历、住院病历结合电话问卷资料进行随访，研究终点包括生存情况、神经精神症状转归及器官损伤。以北京协和医院单中心的中国系统性红斑狼疮研究协作组（CSTAR）注册队列一期患者中无神经精神症状的系统性红斑狼疮(non-NPSLE)患者222例为对照组。采用卡方检验、T检验、秩和检验、Kaplan-Meier生存分析及Logistic回归分析等方法对临床特征、预后结局及其相关因素进行分析。

结果：188例NPSLE患者，女性:男性12.4:1，平均年龄29.8±10.9岁（12-63岁）。NP起病时SLE病程中位数9月（0-354月），94例（50.3%）患者神经精神症状发生于SLE起病1年内，69例（36.7%）NPSLE患者同时存在多种神经精神症状。共265例神经精神事件包括16种NPSLE亚型，以中枢神经系统受累为主，常见的NPSLE亚型为癫痫、急性意识障碍(ACS)、脑血管病变(CVD)、头痛及精神病(>10%总NPSLE)。188例NPSLE患者平均SLEDAI-2K评分20.4±9.2 分（0-44 分），41例（21.8%）患者ACR/SLICC损伤指数（SDI）评分≥1分。实验室检查方面，101例（53.7%）患者抗ds-DNA抗体阳性，78例（41.5%）患者抗rRNP抗体阳性，148例（78.4%）患者补体减低。将NPSLE组患者与non-NPSLE组进行比较，患者性别、年龄无显著差异。NPSLE组患者除外神经症状后SLEDAI-2K评分显著高于non-NPSLE组，提示NPSLE与SLE高活动度相关；NPSLE组SDI评分显著高于non-NPSLE组。NPSLE组患者抗rRNP抗体阳性率高于non-NPSLE组。188例NPSLE患者中，141例（75.0%）患者接受了激素冲击治疗，185例（98.4%）接受了免疫抑制剂治疗，其中163例（86.7%）免疫抑制剂为CTX。134例（71.3%）患者进行了地塞米松或地塞米松联合甲氨蝶呤鞘内注射治疗。住院期间死亡病例3例（1.6%），死因方面感染2例或心源性事件1例。188例NPSLE患者共随访172例，死亡16例，NPSLE患者整体1、2、3、4年生存率为94.2%、92.7%、91.1%、82.8%。对NPSLE死亡患者进行死因分析，其中感染为主要死亡原因，占37.5%。将NPSLE组与non-NPSLE组生存曲线进行比较，NPSLE组患者较non-NPSLE患者生存率显著降低（P<0.01），风险比（HR）为2.771（95%可信区间 1.935–11.76）。Kaplan-Meier单因素分析结果显示癫痫（p=0.024）、SLEDAI-2K评分≥15分（p=0.043）显著降低生存率。随访时长>6月的患者中，8例患者出现NP症状反复，NPSLE复发可出现于同亚型间，也可见于不同亚型。整体1、2年NP症状复发率为3.1%、5.4%。 Kaplan-Meier单因素分析显示头痛（p=0.049）、脑脊液细胞学结果异常（p=0.039）显著影响复发率。对156例NPSLE生存患者进行器官损伤分析，NP起病时SDI评分平均值0.22，末次随诊SDI评分平均值0.44，末次随诊SDI评分显著升高（p＜0.01）。二分类Logistic回归分析结果显示与癫痫[OR 7.232, 95%CI 2.472-21.147，p=0.000]、脊髓炎[OR 20.990, 95%CI 1.756-250.859，p=0.016]与SDI评分升高相关。67例NPSLE-癫痫患者共随访60例,死亡9例，11例癫痫患者末次随访时仍服用抗癫痫药，3例患者末次随访6月内出现癫痫症状反复。二分类Logistic回归分析结果未显示与癫痫反复或抗癫痫药物依赖相关因素。NPSLE-CVD患者29例，其中静脉窦血栓9例，蛛网膜下腔出血（SAH）3例，脑梗死19例。29例SLE-CVD患者共随访27例，死亡3例（2例静脉窦血栓及1例SAH），根据比较起病与随访时改良Rankin量表评分结果，17例脑梗死患者功能恢复显著（p<0.05）。21例NPSLE-精神病患者，共随访19例，根据比较起病与随访时个人和社会功能量表（PSP）评分结果19例患者个人和社会功能恢复显著（p<0.05）。

结论：NPSLE多出现于SLE起病1年内，疾病活动期，抗rRNP抗体高阳性率，癫痫为最常见亚型。NPSLE较non-NPSLE组患者生存率降低。NPSLE1、2、3、4年生存率为94.2%、92.7%、91.1%及82.8%；感染是最常见死亡原因；SLE高活动度、癫痫显著降低NPSLE患者生存率。NP症状复发多出现在2年内，1、2年复发率分别为3.1%和5.4%，头痛、CSF细胞学异常是患者复发危险因素。NPSLE患者SDI评分显著升高，SDI评分升高与癫痫、脊髓炎亚型相关。NPSLE-癫痫、CVD及精神病患者功能恢复良好。

 

关键词：系统性红斑狼疮；神经精神狼疮；生存；复发；器官损伤

 

第二部分 神经精神狼疮神经影像分析

目的：研究神经精神狼疮患者头颅核磁共振成像结果与临床特征特征及预后关系。

方法：由2位神经影像学专家根据统一标准，对北京协和医院神经精神狼疮患者头颅核磁共振成像（MRI）结果，进行重新读取及报告。采用卡方检验、Kaplan-Meier生存分析等方法对MRI结果与临床特征及预后进行分析。

结果：84例NPSLE患者头颅MRI中，57例（67.9%）患者MRI表现异常，其中炎性病变7例（8.3%），血管病变53例（61.9%）。血管病变中，大血管病变3例（3.6%），小血管病变52例（61.9%）。小血管病变中，白质高信号（WMH）40例（47.6%），为最常见表现，额叶为最常见受累部位。头颅MRI结果与临床特征分析结果显示，SLE病程≥1年的患者与WMH改变相关（p=0.048）；存在多种神经精神症状的患者与脑萎缩相关（p=0.030），脑脊液蛋白升高与多发皮下小梗死灶存在相关性（p=0.039）。头颅MRI结果与NPSLE亚型分析结果显示，NPSLE-头痛患者头颅MRI正常率高（p=0.036），NPSLE-认知障碍患者脑萎缩存在相关性（p=0.017）。MR结果与抗体分析未发现相关性。84例患者共随访82例，死亡7例。Kaplan-Meier分析MRI结果与预后关系，炎性病变患者生存率显著减低（p=0.046）。

结论：多数NPSLE患者头颅MRI结果表现异常，其中最常见MRI异常表现为小血管病变中WMH。NPSLE患者头颅MRI与临床特征分析结果提示，SLE病程与WMH、多种NP症状与脑萎缩、脑脊液蛋白水平升高与多发皮下小梗死灶相关。MRI结果与NP亚型分析提示，NPSLE-头痛组患者MRI异常率低，NPSLE-认知障碍与脑萎缩相关。MR结果与抗体分析未发现相关性。MRI结果预后分析显示，头颅MRI出现炎性病变表现的患者生存率降低。

关键词：系统性红斑狼疮；神经精神狼疮；核磁共振成像

 

第三部分 神经精神狼疮脑脊液生物标志物分析

目的：自身免疫/炎症损伤是弥漫性NP症状住院病理机制，前期研究通过对NPSLE-精神病患者与non-NPSLE患者脑脊液进行同位素标记的相对和决定定量（Isobaric tags for relative and absolute quantitation，iTRAQ）蛋白质组学进行分析，结果中出现的蛋白表达异常。从中选取胶质细胞激活相关蛋白，CHI3L1 具有神经系统炎症调节作用，PRSS2促进神经再生进行验证，分析其与NPSLE关系及在NPSLE 中的临床意义。

方法：本研究纳入新发或症状加重的NPSLE患者共18例，分为三组患者：NPSLE-头痛组、NPSLE-精神病组、NPSLE-急性意识障碍组。收集3组患者的临床资料及脑脊液，部分患者收集同期治疗后脑脊液。ELISA方法测定脑脊液中CHI3L1及PRSS-2水平。比较各组结果差异。

结果：三组患者18例NPSLE患者，女性17例，年龄27.7±10.1岁，SLE病程中位时间9月，SLEDAI-2K评分20.6±9.5分。三组患者除外NP症状后SLEDAI-2K评分无统计学差异。CHI3L1在NPSLE-急性意识障碍组患者CSF中水平显著高于NPSLE-头痛或NPSLE-精神病组（p<0.05），NPSLE-头痛组与NPSLE-精神病组间无显著差异，治疗中前后期NPSLE患者CSF中CHI3L1水平无显著差异。PRSS-2在三组患者CSF中水平无显著差异，治疗前后NPSLE患者CSF中PRSS2水平无显著差异。

结论：CHI3L1作为CSF中神经炎性标志物，在NPSLE患者脑脊液中表达异常。CHI3L1在NPSLE-急性意识障碍组患者中较NPSLE-头痛或精神病组患者升高显著，提示NPSLE-急性意识障碍发病机制与中枢神经系统炎症相关，可考虑加强局部针对性治疗。

关键词：系统性红斑狼疮；神经精神狼疮；神经炎症

PART ONE Clinical analysis and long-term outcomes of patients with neuropsychiatric systemic lupus erythematosus Objective：This study aims to analyze the prognosis, both mortality and morbidity, for patients with neuropsychiatric systemic lupus erythematosus (NPSLE) in a large single-center of Peking Union Medical College of Hospital (PUMCH). Methods：Patients with NPSLE in Peking union medical college hospital(PUMCH) were collected retrospectively from June 2012 to June 2016，only neuropsychiatric events attributed to lupus was included. These patients were followed up by outpatient medical records and telephone follow-up. Data for baseline, follow-up and survival were collected; including demography, manifestations, activity (SLEDAI-2K), SLE International Collaborating Clinics Damage Index (SDI), and medications. SLE patients without any neuropsychiatric (non-NPSLE) syndromes from single-center CSTAR database were taken as control group. Clinical features of NPSLE patients compared with non-NPSLE patients were analysed by chi-square test or T test. Kaplan-Meier method was adopted for survival analysis and relapse analysis. Associated factors for increased SDI score were analyzed by Logistic Regression model. Result：In this study 188 patients were included, female to male is 12.4:1, the average age was 29.8±10.9 years(12 to 63 years). The median of SLE duration was 9 months (0-354months). In 94 patients (50.3%), NP syndrome occurred within the first year after SLE onset. 69 patients (36.7%) had more than 1 NP syndromes. The average SLEDAI-2K scores was 20.4±9.2 (0-44). 41 (21.8%) patients had SDI score more than one. Totally 265 NP events were manifested in these patients. Sixteen subtypes of NPSLE were identified, and the most frequent manifestations were seizure, acute confusion state, cerebral vascular disease, headache and psychosis. Compared with non-NPSLE patients, NPSLE patients had a significantly higher SLEDAI-2K score when the scores of NP syndrome were excluded. 141 patients (75.0) accepted therapy of methylprednisolone pulse，and 134 patients (71.3%) accepted therapy of intrathecal injection. 172 NPSLE patients were followed and 16 patients died during these years. The 1, 2, 3 and 4 year survival rates were 94.2%, 92.7%, 91.1% and 82.8%, respectively. The most common reason for death was infection. Kaplan-Meier analysis indicated that NPSLE significantly decrease the survival rate of SLE patients. Single factor analysis for survival rate of NPSLE patients indicated that SLEDAI-2K score >=15 or seizures significantly decrease the survival rate. The 1, 2, year relapse rates were 3.1%, 5.4%, respectively. Single factor analysis for relapse rate of NPSLE patients indicated that headache or abnormal cerebrospinal fluid cytology significantly increase the relapse rate. SDI score increased significantly and was associated with seizures and myelitis. 67 NPSLE-seizures patients and 60 patients were followed. 16 patients died, 11 patients took antiepileptic drug at the time of the last follow-up, and 3 patients still had epileptic seizures. Logistic regression analysis indicated no relative factors with these results. 29 NPSLE-CVD patients and 27 patients were followed. Patients with stroke were evaluated by modified Rankin Scale, and at the time of the last follow-up the function of stroke patients improved significantly.21 NPSLE-psychosis patients and 19 patients were followed. Patients with psychosis were evaluated by personal and social performance scale (PSP), and at the time of the last follow-up the function of these patients improved significantly. Conclusion: Most NPSLE events occurred in the presence of generalized disease activity. The most common subtype of NPSLE is seizure. NPSLE decrease the survival rate of SLE patients. Survival rates in our cohort are comparable to previous reported one. For NPSLE patients, SLEDAI-2K scores and seizures are prognostic factors and deserve more attention in the future. Most patients of seizures, CVD and psychosis have a favorable prognosis. Key words：neuropsychiatric systemic lupus erythematosus，clinical characteristics，survival, organ damage, relapse PART TWO Magnetic resonance imaging result analysis and long-term outcomes of patients with neuropsychiatric systemic lupus erythematosus Objectives: To describe brain magnetic resonance imaging (MRI) abnormalities in neuropsychiatric systemic erythematosus lupus (NPSLE) and correlate them with clinical manifestation, laboratory data and prognosis. Methods: This retrospective cross-sectional study included patients presenting NPSLE undergoing brain MRI between 2012.6 and 2016.6. Clinical features, laboratory data and outcomes were recorded. MRI findings were defined as inflammatory-like, large-vessel disease (LVD), and small-vessel disease (SVD); SVD was classified as white-matter hyperintensities (WMH), recent small subcortical infarcts, lacunes, microbleeds, and brain atrophy. Results: We included 84 patients (mean 20.8 ± 10.2 years), 97.6% women. The most frequent syndromes were seizures (41.7%), acute confusional state (20.2%), and headache (14.3%). Brain abnormalities were found in 67.9%. SVD was the most common (63.1%), followed by inflammatory-like lesions (8.3%) and LVD (3.6%). The most frequent SVD findings were WMH (47.6%), atrophy (22.6%), recent small subcortical infarcts (11.9%). SLE disease course more than one year correlated with WMH (p = 0.048), patients with more than one NP symptoms with atrophy (0.030), elevated protein in CSF with recent small subcortical infarcts (0.039). Headache correlated with normal MRI (0.036), cognitive disorder with atrophy. No relationship was found between MRI and autoimmune antibody. 82 patients were followed and 7 patients died. Kaplan-Meier analysis indicated that inflammatory-like lesions significantly decreased the survival rate of NPSLE patients. Conclusions: Vascular disease is the hallmark of NPSLE. Certain syndromes and outcomes are prone to some brain damage. MRI could provide significant clinical information and insights into the pathological substrate. Key words：neuropsychiatric systemic lupus erythematosus，magnetic resonance imaging PART THREE Cerebrospinal fluid biomarkers in patients with neuropsychiatric systemic lupus erythematosus Background and objective：Pathogenesis of NPSLE has not been well understood and different type of NPSLE may involve different pathophysiology. Earlier study had explored biomarkers in cerebrospinal fluid (CSF) through CSF proteomics applying Isobaric tags for relative and absolute quantization (iTRAQ) technology. Proteins associated with activation of gliocyte, neuroinflammation and structure or function of synapses is found modified in the CSF of NPSLE patients as compared to control groups. The objective of this study is to check the result. We chose 2 kinds of proteins associated with activation of gliocyte, chitinase 3-like protein 1(CHI3L1)and protease serine 2(PRSS2),and this study is to discuss the role of them in cerebrospinal fluid(CSF) in NPSLE and its clinical significance. Methods ：18 NPSLE patients were included in this study and were divided into three groups according to NP subtype: the NPSLE-headache group, the NPSLE-psychosis group and the NPSLE-acute confusional state (ACS) group. Each group consisted of 6 patients. The clinical data and CSF of these patients was collected，and CSF after therapy was collected in some of these patients. The level of CHI3L1 and PRSS2 in CSF and serum was detected by ELISA. Result：Levels of CHI3L1 in CSF significantly elevated in NPSLE-ACS group compared to NPSLE-headache group (186.6±23.2 vs 104.1±67.3ng/ml，P=0.029) and NPSLE-psychosis group (186.6±23.2 vs 107.1±41.5ng/ml，p=0.004). There is no significant difference between NPSLE-psychosis group and NPSLE-headache group. No significant decline level of CHI3L1 in CSF was found after therapy. Levels of PRSS2 in CSF had no significant difference between each two group, yet between before and after therapy. Conclusion：Levels of CHI3L1 in CSF are significantly elevated in NPSLE-ACS group, indicating that the pathogenesis behind ACS is inflammation of central nervous system. Key words：Neuropsychiatric systemic lupus erythematosus, cerebrospinal fluid, CHI3L1, PRSS2