Complications, outcomes, and implications of a prolonged vegetative state in anti-NMDA receptor encephalitis: a...

In anti-NMDA receptor (NMDAR) encephalitis, delayed recovery and slow improvement makes it difficult to assess treatment refractoriness, leaving a knowledge gap about prolonged impaired consciousness. We aimed to assess treatment response, long-term outcomes, and their predictors in patients with NMDAR encephalitis and a prolonged vegetative state.

In this international retrospective cohort study, we included patients with NMDAR encephalitis from Jan 1, 2007 to Sept 30, 2024 who remained in a vegetative state (unresponsive wakefulness) for 9 months or longer in 21 hospitals across Austria, Brazil, Chile, China, Germany, Hong Kong, Japan, the Netherlands, Spain, Switzerland, and the USA. Data on disease presentation and long-term outcomes were collected via medical record review and a structured questionnaire sent to referring physicians. Outcomes were death, ability to follow commands, reaching a modified Rankin Scale (mRS) score of 2, or complete recovery (mRS 0 with return to premorbid activities). Competing risks analysis was used to assess survival and predictors of death, mRS score 2, and recovery.

45 patients were identified (38 female and seven male; median age 22 [IQR 19-31] years). The patients had impaired consciousness a median of 9 days after symptom onset (IQR 5-16). All 45 patients received first-line immunotherapy; 41 (91%) received second-line, and 18 (40%) received third-line immunotherapies. 21 (47%) of 45 patients had ovarian teratomas, which were removed in 20 patients. Median durations were: vegetative state 399 days (IQR 307-698); intensive care unit stay 275 days (189-354); mechanical ventilation 270 days (195-394); and hospitalisation 474 days (349-715). 13 (28%) patients were resuscitated from dysautonomic cardiac arrest. After a median of 5 years (IQR 3-7), 15 (33%) of 45 patients completely recovered (mRS score ≤1 and returned to all previous activities), 13 (29%) substantially improved (mRS score 2), 11 (24%) had mRS score 3-5, and six (13%) died. Five patients began command-following a median of 11 months (IQR 2-21) after last immunotherapy. Estimated cumulative incidence for reaching an mRS score of 2 was 66% at 5 years and 76% at 10 years, and for recovery was 32% at 5 years and 54% at 10 years. Teratomas were associated with a lower probability of reaching mRS score 2 (sub-distribution hazard ratio 0·39 [95% CI 0·18-0·84], p=0·0160), whereas older age (1·10 per year [1·04-1·23], p=0·0052) and higher NMDAR encephalitis 1-Year Functional Status (NEOS)2 score (1·51 [1·12-2·04], p=0·0072) were associated with increased mortality.

In people with NMDAR encephalitis and prolonged impairment of consciousness, full or substantial recovery was reached in approximately two-thirds of cases. Consequently, early assessment of treatment response or refractoriness might underestimate delayed improvement, prolonged therapy needs, or spontaneous recovery. Futility decisions should therefore be individualised with multidisciplinary input and based on extended follow-up.

Instituto de Salud Carlos III and Fundació La Caixa.