目的：系统性红斑狼疮（Systemic Lupus Erythematosus, SLE）是一种累及多系统的自身免疫性疾病，神经精神狼疮（Neuropsychiatric Systemic Lupus Erythematosus, NPSLE）是其严重并发症之一。本研究旨在探讨SLE患者癫痫发作的病因学构成、临床特征及预后差异，为临床诊断和治疗提供依据。

方法：本研究采用回顾性队列研究方法，收集2015年1月至2023年12月北京协和医院住院诊治的130例SLE伴癫痫发作患者的临床资料，进行病因分类、分组，比较不同病因组的人口学特征、临床表现、实验室检查、影像学表现、治疗方案及预后情况。采用SPSS 27.0进行统计学分析。若整体检验P＜0.05，采用校正多重检验（Bonferroni法），进行组间两两比较。

结果：130例患者中，女性占82.3%，发病年龄中位数为29岁（四分位距[Interquartile Range, IQR]：22.0-36.0）。病因分布为：狼疮脑炎32.3%（42例）、脑血管病11.5%（15例）、可逆性后部脑病综合征（Posterior Reversible Encephalopathy Syndrome, PRES） 12.3%（16例）、感染7.7%（10例）、药物/代谢因素各0.8%（1例），不可归因组34.6%（45例）。癫痫发作以全面性强直-阵挛发作为主，感染组癫痫持续状态发生率最高（20.0%）。PRES组血液系统（93.8%）受累率显著高于感染组（P = 0.004）、肾脏（100%）系统受累率显著高于狼疮脑炎组（52.4%，P = 0.001）及不可归因组（28.9%，P＜0.001），浆膜腔积液（87.5%）受累率显著高于狼疮脑炎组（40.5%，P = 0.001）。脑血管病组的抗磷脂抗体阳性率显著高于狼疮脑炎组、PRES组、不可归因组（P均＜0.001）。感染组的补体C3、C4水平最高（C3中位数0.94g/L，IQR 0.77-1.07，C4中位数0.19g/L，IQR 0.14-0.23），入院时疾病活动度以轻中度活动为主（轻中度共占80%），而其他各组重度活动率均在80%以上，差异具有统计学意义（P均＜0.001）。仅在感染组、狼疮脑炎组出现脑脊液白细胞升高、细胞学炎性率、白细胞介素-6升高，且与其他各组存在统计学意义（P＜0.005）。PRES组头磁共振成像的异常率最高（100%），额颞顶枕叶均易出现病灶，以顶叶（81.3%）、枕叶（87.5%）为著。脑电图背景活动（P = 0.812）和癫痫样放电（P = 0.392）在各组间无统计学差异。狼疮脑炎组、PRES、不可归因组常选用甲泼尼龙冲击治疗及环磷酰胺治疗。狼疮脑炎组及不可归因组更多采用鞘内注射甲氨蝶呤治疗。预后方面，PRES组末次SLEDAI-2K评分最高（中位数5分，IQR 2-9），但未达到统计学差异（P = 0.102）。感染组随访时药物难治性癫痫的比例最高（30.0%），狼疮脑炎组次之（16.7%），脑血管病组最低（0.0%），各组间P值未达到统计学差异（P = 0.285）。

结论：全面性强直-阵挛发作是SLE癫痫发作的主要发作类型。SLE癫痫发作的具体病因包括狼疮脑炎、脑血管病、可逆性后部脑病综合征与感染性疾病等。由自身免疫与炎症介导的狼疮脑炎是SLE癫痫发作的主要病因。脑脊液白细胞升高、脑脊液细胞学提示炎症反应，脑脊液IL-6升高以及抗神经抗体阳性有助于诊断狼疮脑炎，但要与感染性疾病等鉴别。病因诊断对治疗方案选择具有重要意义。

Objective: Systemic lupus erythematosus (SLE) is a multisystem autoimmune disease, and neuropsychiatric lupus (NPSLE) is one of its severe complications. This study aimed to investigate the etiological composition, clinical characteristics, and prognostic differences of seizures in SLE patients, providing evidence for clinical diagnosis and treatment.

Methods: A retrospective cohort study was conducted, analyzing clinical data from 130 SLE patients with seizures hospitalized at Peking Union Medical College Hospital between January 2015 and December 2023. Patients were classified by etiology and compared across groups regarding demographic characteristics, clinical manifestations, laboratory findings, imaging features, treatment regimens, and outcomes. Statistical analysis was performed using SPSS 27.0. If the overall test yielded P＜0.05, corrected multiple comparisons (Bonferroni method) were performed for pairwise comparisons between groups.

Results: Among 130 patients (82.3% female), the median age of onset was 29 years (Interquartile Range, IQR: 22.0-36.0). Etiological distribution included: lupus encephalitis (32.3%, 42 cases), cerebrovascular disease (11.5%, 15 cases), posterior reversible encephalopathy syndrome (PRES, 12.3%, 16 cases), infection (7.7%, 10 cases), drug/metabolic factors (0.8% each, 1 case), and unclassified group (34.6%, 45 cases). Generalized tonic-clonic seizures predominated, with the highest status epilepticus incidence in the infection group (20.0%). The PRES group exhibited a significantly higher involvement rate of the hematologic system (93.8%) compared to the infection group (P = 0.004), as well as a higher renal system involvement rate (100%) than both the lupus encephalitis group (52.4%, P = 0.001) and the non-attributable group (28.9%, P＜0.001). Additionally, the PRES group had a significantly greater prevalence of serous cavity effusion (87.5%) than the lupus encephalitis group (40.5%, P = 0.001). The cerebrovascular group showed significantly higher antiphospholipid antibody positivity than lupus encephalitis, PRES, and unclassified groups (P＜0.001). The infection group had the highest complement C3 (median 0.94 g/L, IQR 0.77-1.07) and C4 levels (median 0.19 g/L, IQR 0.14-0.23), with predominantly mild-to-moderate disease activity (80%), while other groups had >80% severe activity (P＜0.001). Cerebrospinal fluid (CSF) leukocytosis, cytological inflammation, elevated interleukin-6 (IL-6) occurred only in infection and lupus encephalitis groups (P＜0.005). PRES had the highest MRI abnormality rate (100%), with lesions predominantly in parietal (81.3%) and occipital lobes (87.5%). No intergroup differences were observed in EEG background activity (P = 0.812) or epileptiform discharges (P = 0.392). Methylprednisolone pulse therapy and cyclophosphamide were commonly used in lupus encephalitis, PRES, and unclassified groups. Intrathecal methotrexate was more frequent in lupus encephalitis and unclassified groups. The PRES group showed the highest final Systemic Lupus Erythematosus Disease Activity Index 2000 scores (median 5, IQR 2-9), though not statistically significant (P = 0.102). At last follow-up, the infection group exhibited the highest proportion of drug-resistant epilepsy (30.0%), followed by the lupus encephalitis group (16.7%), while the cerebrovascular disease group showed the lowest rate (0.0%). However, the differences among groups did not reach statistical significance (P = 0.285).

Conclusion: Generalized tonic-clonic seizures represent the predominant seizure type in SLE. The etiological spectrum of SLE-associated seizures encompasses lupus encephalitis, cerebrovascular disease, PRES, and infectious disorders. Among these, lupus encephalitis mediated by autoimmune and inflammatory mechanisms constitutes the primary etiology of seizure activity in SLE. Diagnostic indicators for lupus encephalitis include CSF pleocytosis, inflammatory cytological findings, elevated CSF IL-6 levels, and the presence of neural autoantibodies, though differentiation from infectious etiologies remains essential. Accurate etiological diagnosis carries significant therapeutic implications for treatment strategy selection.