A new capture ELISA kit (ETI-Rubek-M) for rubella IgM antibody determination was compared with a commercial reverse (Rubenostika IgM) and two indirect ELISAs (Rubelisa M and Rubazyme-M). Of the 89 sera from 38 patients with acute rubella infection, 70 gave identical results with the four assays (78% concordance), 74 were in agreement in the two indirect ELISAs, and 83 concorded in the reverse assays (83% and 93% concordance, respectively). On the whole, ETI-Rubek-M and Rubenostika IgM proved equally sensitive; sensitivity of the indirect ELISA kits was lower, Rubazyme-M being slightly more sensitive than Rubelisa M. Specificity of the two reverse assays was 100%, whereas Rubelisa M and Rubazyme-M showed a specificity of 88.8% and 93.3%, respectively. In indirect ELISAs, false positive IgM results were mainly observed in cytomegalovirus IgM-positive sera from patients recovering from a primary cytomegalovirus infection. Two sequential sera from one of these patients with IgM to cytomegalovirus were found to be true positive for rubella-specific IgM antibody with ETI-Rubek-M, Rubelisa M and Rubazyme-M.