At present, the measurement of IgG avidity appears to be the best method for differentiating primary from nonprimary cytomegalovirus (CMV) infection. This study compared the performances of three denaturation assays for the measurement of CMV IgG avidity: an in-house method and the commercially available assays Enzygnost (Dade-Behring, Germany) and Vidas (bioMérieux, France). The ability of these assays to exclude or to detect a recent CMV infection were calculated according to the results obtained in two control groups of pregnant women: 49 who had seroconverted and 80 with past infections. All three assays demonstrated a good ability to detect a recent infection (98-100%). The Dade-Behring test, in its present form, appears to be ineffective in excluding a recent CMV infection (exclusion ability: 30%), while the in-house method (exclusion ability: 96%) and the bioMérieux method (exclusion ability: 82%) performed better. The practical use of the in-house and the bioMérieux assays was evaluated in 80 women with CMV-specific IgG and a positive IgM result but without documented seroconversion. At the recommended diagnostic thresholds, the concordance between these two tests was 70%. Larger studies will allow more precise determination of the capacities of both assays and specification of the diagnostic thresholds or grey areas to be used.