OBJECTIVE:To 1) determine, using contemporary recombinant antigen–based assays, the aquaporin-4 (AQP4)–immunoglobulin G (IgG) detection rate in sequential sera of patients assigned a clinical diagnosis of neuromyelitis optica (NMO) but initially scored negative by tissue-based indirect immunofluorescence (IIF) assay; and 2) evaluate the impact of serostatus on phenotype and outcome. METHODS:From Mayo Clinic records (2005–2011), we identified 163 patients with NMO; 110 (67%) were seropositive by IIF and 53 (33%) were scored seronegative. Available stored sera from 49 “seronegative” patients were tested by ELISA, AQP4-transfected cell-based assay, and in-house fluorescence-activated cell sorting assay. Clinical characteristics were compared based on final serostatus. RESULTS:Thirty of the 49 IIF-negative patients (61%) were reclassified as seropositive, yielding an overall AQP4-IgG seropositivity rate of 88% (i.e., 12% seronegative). The fluorescence-activated cell sorting assay improved the detection rate to 87%, cell-based assay to 84%, and ELISA to 79%. The sex ratio (female to male) was 1:1 for seronegatives and 9:1 for seropositives (p < 0.0001). Simultaneous optic neuritis and transverse myelitis as onset attack type (i.e., within 30 days of each other) occurred in 32% of seronegatives and in 3.6% of seropositives (p < 0.0001). Relapse rate, disability outcome, and other clinical characteristics did not differ significantly. CONCLUSIONS:Serological tests using recombinant AQP4 antigen are significantly more sensitive than tissue-based IIF for detecting AQP4-IgG. Testing should precede immunotherapy; if negative, later-drawn specimens should be tested. AQP4-IgG–seronegative NMO is less frequent than previously reported and is clinically similar to AQP4-IgG–seropositive NMO.