Background Neuromyelitis optica (NMO, Devic syndrome) is associated with antibodies to aquaporin-4 (NMO-IgG/AQP4-Ab) in the majority of cases. NMO- IgG/AQP4-Ab seropositivity in patients with NMO and its spectrum disorders has important differential diagnostic, prognostic and therapeutic implications. So-called cell-based assays (CBA) are thought to provide the best AQP4-Ab detection rates. Objective To compare directly the AQP4-IgG detection rates of the currently most widely used commercial CBA, which employs cells transfected with a full-length (M1)-human AQP4 DNA in a fashion that allows leaky scanning (LS) and thus expression of M23-AQP4 in addition to M1-AQP, to that of a newly developed CBA from the same manufacturer employing cells transfected with human M23-AQP4-DNA. Methods Results from 368 serum samples that had been referred for routine AQP4-IgG determination and had been tested in parallel in the two assays were compared. Results Seventy-seven out of 368 samples (20.9%) were positive for NMO-IgG/AQP4-Ab in at least one assay. Of these, 73 (94.8%) were positive in both assays. A single sample (1.3%) was exclusively positive in the novel assay; three samples (3.9%) were unequivocally positive only in the ‘classic’ assay due to high background intensity in the novel assay. Both median fluorescence intensity and background intensity were higher in the new assay. Conclusions This large study did not reveal significant differences in AQP4-IgG detection rates between the ‘classic’ CBA and a new M23-DNA-based CBA. Importantly, our results largely re-affirm the validity of previous studies that had used the ‘classic’ AQP4-CBA to establish NMO-IgG/AQP4-Ab seropositivity rates in NMO and in a variety of NMO spectrum disorders.