RESUMO
BACKGROUND AND PURPOSE: To clarify the relevance of titres of IgG antibodies against contactin-associated protein-2 (CASPR2) in diagnosing anti-CASPR2 encephalitis and to describe features and outcomes. METHODS: This was a retrospective analysis of 64 patients with CASPR2 antibodies, categorized independently as 'autoimmune encephalitis' or 'other disease'. Logistic regression methods were performed to identify potential predictors of 'autoimmune encephalitis' in addition to CASPR2 antibodies. RESULTS: An upfront CASPR2 antibody serum titre cut-off at ≥1:200 had a diagnostic sensitivity of 85% and a specificity of 81%. Logistic regression analyses indicated that, in addition to titre, encephalitic magnetic resonance imaging (MRI) was a significant predictor of 'autoimmune encephalitis' (Nagelkerke's R2 = 0.81, P < 0.001) with high sensitivity (84%) and very high specificity (100%). Patients with CASPR2 antibodies and an estimated probability of >70% of having anti-CASPR2 encephalitis (n = 22) had limbic encephalitis (n = 18, one patient plus ataxia), Morvan syndrome (n = 2) or a hyperkinetic movement disorder (n = 2). Median modified Rankin score (mRS) at diagnosis was 3 (range 1-4). Twenty patients were male; median age was 64 (range 54-75) years; 5/15 patients with cerebrospinal fluid data had intrathecal CASPR2 antibody synthesis, and 12/19 with follow-ups >3 months (median 12 months, range 4-43 months) improved by ≥1 mRS point resulting in a median mRS of 2 (range 0-6; one death; all but one having received immunotherapy); and 2/15 patients with follow-up MRI developed hippocampal atrophy. CONCLUSIONS: Only higher CASPR2 serum antibody titres indicate anti-CASPR2 encephalitis, and diagnostic accuracy increases if MRI findings are considered. Anti-CASPR2 encephalitis has characteristic features and a favourable outcome with immunotherapy.