Factors Associated With Relapse in Patients With Neurosarcoidosis.

BACKGROUND AND OBJECTIVES Five to 10 percent of patients with sarcoidosis show involvement of the nervous system. Given the serious nature of neurosarcoidosis and the high risk of irreversible neurologic damage, rapid diagnosis, accurate prognostication, and individualized treatment are crucial. The aim of this study was to identify factors associated with relapses in neurosarcoidosis. METHODS In this multicenter retrospective cohort study, patients diagnosed with possible, probable, or definite neurosarcoidosis between January 1, 2010, and June 30, 2024, at 6 tertiary neuroimmunology centers across Europe were included. Patients were identified from the respective hospital-based databases by the participating centers. Clinical presentation, imaging results, CSF analysis, and immunosuppressive therapies were evaluated. Predictors of relapse, defined as clinical relapse or progression, or new MRI lesions on follow-up ≥2 months after initial manifestation, were analyzed with log-rank tests and Cox regression models, and therapeutic strategies in clinical practice were assessed. The association between identified risk factors (RFs) and therapeutic strategies was explored. RESULTS A total of 174 patients with neurosarcoidosis were included with a median follow-up of 24 months (interquartile range 12-48.8). The mean age was 48.1 years, and 57.5% were female. CNS parenchymal lesions including encephalitis, myelitis, and optic neuritis (hazard ratio [HR] 2.3, 95% CI 1.3-4.2); CSF-specific oligoclonal bands (OCBs) (HR 2.1, 95% CI 1.3-3.6); CSF glucose <40 mg/dL (HR 2, 95% CI 1.1-3.4); and CSF protein ≥1,000 mg/dL (HR 2.1, 95% CI 1.3-3.5) were identified as RFs of relapse. Patients with 3-4 RFs had a median relapse-free survival of 12 months (95% CI 4.3-19.7), compared with 36 months (95% CI 24.3-47.7) in patients with 1-2 RFs and 120 months (95% CI 2.1-237.9) in patients without RFs (p < 0.01). The likelihood of treatment escalation increased with the number of RFs, from 14.3% in patients without RFs to 28.2% and 50% in patients with 1-2 or 3-4 RFs, respectively (p < 0.01). DISCUSSION The identification of specific RFs, including parenchymal lesions, OCBs, CSF glucose <40 mg/dL, and CSF protein ≥1,000 mg/dL, enables better prognostication and might inform individualized treatment approaches. Patients with multiple RFs are at greater risk of relapse, possibly suggesting the need for more aggressive therapies.