We evaluated the performance of two immunoblot assays: the INNO-LIA Syphilis Score (LIA) and the MarDx T. pallidum IgG Marblot Test (TWB), as compared with that of the Murex ICE Syphilis enzyme immunoassay (EIA), the Serodia Treponema pallidum particle agglutination (TPPA) assay and the fluorescent treponemal antibody-absorption (FTA-abs) assay, for the serological diagnosis of syphilis using serum samples of 135 attendees of the social hygiene clinics of the Department of Health in Hong Kong newly diagnosed with syphilis and provided with clinical stages (39 in primary, 20 in secondary, 18 in early latent and 58 in latent of unknown duration) and of 43 normal healthy subjects between October and December 2004. The differences in the overall sensitivities of the LIA assay and the EIA/TPPA/FTA-abs assays were not statistically significant (P > 0.05) whereas the overall sensitivity of the TWB assay was significantly lower (P < 0.05) than the overall sensitivities of the EIA, the TPPA and the FTA-abs assays. The LIA assay had an overall sensitivity of 94.1% (95% CI 88.7-97.0%) whereas the TWB assay 65.2% (95% CI 56.8-72.7%). Both the LIA and the TWB assays have a specificity of 100%. When consensus results were derived from the most predominant results of the EIA, the TPPA and the FTA-abs assays, the LIA assay had a positive agreement with the consensus results of 98.5% (95% CI 94.5-99.6%) whereas the TWB assay 68.2% (95% CI 59.8-75.6%). Therefore, the LIA assay performed significantly better (P < 0.05) than the TWB assay. The LIA assay can be considered to be a valid alternative confirmatory test for the serological diagnosis of syphilis.