Background: Antibodies against aquaporin-4 (AQP4), a water channel particularly expressed on perivascular astrocytic podocytes, are proposed as a marker for the diagnosis of neuromyelitis optica (NMO). However, a consensus on seroprevalence and optimal detection method has not yet been reached. Objectives: To investigate the performance of different assays to detect anti-AQP4 antibodies. Methods: We set up five different assays. Two of them were capable to detect perivascular IgG reactivity on brain tissue by immunofluorescence (NMO-IgG). Other three assays have been set to detect anti-AQP4 antibodies: immunofluorescence and flow cytometry on AQP4-expressing cells, and a radioimmunoprecipitation assay. We assessed sensitivity and specificity of these assays by interrogating sera of 33 NMO patients, 13 patients at high risk to develop NMO (hrNMO), 6 patients affected by acute partial transverse myelitis (APTM), 20 patients with multiple sclerosis (MS), and 67 age- and sex-matched healthy controls. Results: We found that the presence of serum NMO-IgG and anti-AQP4 reactivity is almost exclusively restricted to patients with NMO and hrNMO. Seroprevalence and sensitivity ranged from 30 to 47%, depending on the assay. Specificity ranged from 95 to 100%. Comparing results obtained in the five assays, we noticed lack of concordance in some samples. Conclusions: Detection of NMO-IgG or anti-AQP4 antibodies may represent a valuable tool to assist neurologists in the differential diagnosis between patients with NMO, hrNMO, APTM, or MS. The current lack of a gold standard to detect anti-AQP4 antibodies implies the necessity to standardize the detection of these antibodies.