Ciano-Petersen Nicolás Lundahl, Villagrán-García Macarena, Muñiz-Castrillo Sergio, Farina Antonio, Vogrig Alberto, Goncalves David, Nicole Fabien, Rogemond Véronique, Picard Geraldine, Aubart Mélodie, Psimaras Dimitri, Oliver Begoña, Serrano-Castro Pedro J, Joubert Bastien, Honnorat Jerome
French Reference Center on Paraneoplastic Neurological Syndromes and Autoimmune Encephalitis, Hospices Civils de Lyon, Hôpital Neurologique, Bron, France.
MeLiS Institute - UCBL-CNRS UMR 5284 - INSERM U1314, Université Claude Bernard Lyon 1, France.
Neurol Neuroimmunol Neuroinflamm. 2025 Jul;12(4):e200421. doi: 10.1212/NXI.0000000000200421. Epub 2025 Jun 18.
During the recovery phase of anti-N-methyl-D-aspartate receptor (NMDAR) encephalitis, up to 25% of relapses have been reported. Herein, we aimed to clinically characterize these relapses, analyze potential clinical predictors during the first episode, and evaluate the impact of immunotherapy in their occurrence.
This was a retrospective observational study of patients diagnosed with anti-NMDAR encephalitis relapses between January 2007 and June 2022 at the French Reference Center for Paraneoplastic Neurological Syndromes and Autoimmune Encephalitis with a follow-up longer than 1 year.
Among 507 patients, 49 (9%) presented relapses after a median time of 720 days (range 149-8,280) and a median follow-up of 1752 days (range 390-9,229 days, interquartile range 1760 days). A total of 36 patients (73%) experienced 1 relapse, 9 (18%) had 2, and 4 (8%) had 3 relapses. Most patients presented an isolated core symptom (25/45, 55%). Relapses were less severe than the first episode, as reflected by a lower maximal modified Rankin Scale (median 5, range 3-5, vs median 3, range 0-6; = 0.0001). At the first episode, patients experiencing relapses had shorter intensive care unit stays (22 days; vs 39 days; = 0.04). In addition, presenting CSF pleocytosis >20 white blood cell decreased the risk of relapse by 71% (HR 0.29; CI 0.13-0.66; = 0.003), and having a paraneoplastic etiology decreased the risk by 68% (HR 0.32; CI 0.12-0.87; = 0.02). Moreover, during the first episode, they were treated less frequently with first-line (39/49, 79%, vs 190/197, 96%; = 0.0001) and second-line immunotherapies (20/49, 40%, vs 142/197, 72%; = 0.0001) and more frequently with delay >30 days (20/38, 52%, vs 58/185, 31%; = 0.01) and >60 days (10/20, 50%, vs 39/138, 28%; = 0.04), respectively. In addition, administering rituximab during the first episode with a delay <60 days decreased the risk of relapse by 60% (HR 0.40; CI 0.19-0.84; = 0.01).
Relapses of anti-NMDAR encephalitis are uncommon, mostly monosymptomatic, and less severe than the first episode. At onset, presenting CSF pleocytosis or an underlying tumor decreases the risk of relapses. In addition, the early administration of first-line and second-line immunotherapies, particularly rituximab, could protect against further relapses.
在抗N-甲基-D-天冬氨酸受体（NMDAR）脑炎的恢复阶段，据报道复发率高达25%。在此，我们旨在对这些复发进行临床特征描述，分析首次发作期间潜在的临床预测因素，并评估免疫治疗对其发生的影响。
这是一项对2007年1月至2022年6月期间在法国副肿瘤性神经综合征和自身免疫性脑炎参考中心被诊断为抗NMDAR脑炎复发且随访时间超过1年的患者进行的回顾性观察研究。
在507例患者中，49例（9%）在中位时间720天（范围149 - 8280天）后出现复发，中位随访时间为1752天（范围390 - 9229天，四分位间距1760天）。共有36例患者（73%）经历1次复发，9例（18%）经历2次复发，4例（8%）经历3次复发。大多数患者表现为单一核心症状（25/45，55%）。复发不如首次发作严重，这体现在改良Rankin量表的最高评分较低（中位值5，范围3 - 5，而首次发作时中位值为3，范围0 - 6；P = 0.0001）。在首次发作时，经历复发的患者在重症监护病房的停留时间较短（22天；对比39天；P = 0.04）。此外，脑脊液淋巴细胞增多>20个白细胞可使复发风险降低71%（风险比0.29；95%置信区间0.13 - 0.66；P = 0.003），而具有副肿瘤病因可使风险降低68%（风险比0.32；95%置信区间0.12 - 0.87；P = 0.02）。此外，在首次发作期间，他们接受一线免疫治疗（39/49，79%，对比190/197，96%；P = 0.0001）和二线免疫治疗（20/49，40%，对比142/197，72%；P = 0.0001）的频率较低，而延迟>30天（20/38，52%，对比58/185，31%；P = 0.01）和>60天（10/20，50%，对比39/138，28%；P = 0.04）治疗的频率较高。此外，在首次发作时延迟<60天给予利妥昔单抗可使复发风险降低60%（风险比0.40；95%置信区间0.19 - 0.84；P = 0.01）。
抗NMDAR脑炎的复发并不常见；大多为单症状性，且不如首次发作严重。发病时，脑脊液淋巴细胞增多或存在潜在肿瘤可降低复发风险。此外，早期给予一线和二线免疫治疗，尤其是利妥昔单抗，可预防进一步复发。
