To ascertain the reliability of serological diagnosis of neurocysticercosis in the everyday a priori situation of neurological consultation, the enzyme-linked immunosorbent assay test was used to predict the eventual diagnosis of neurocysticercosis in an unselected sample of 1064 consecutive neurological cases. Results showed 69% sensitivity and 71% specificity of the enzyme-linked immunosorbent assay test for the diagnosis of neurocysticercosis. In sharp contrast with publications that have proclaimed the excellent diagnostic performance of immunodiagnostic tests, our results suggest that identification of serum antibodies with standard enzyme-linked immunosorbent assay techniques is of little value when applied to a large and heterogeneous group of neurological patients in an endemic area of cysticercosis, and our results urge a reevaluation of currently used immunodiagnostic tests that are practiced in the serum of suspected cases.