Syphilis detection has traditionally relied on non-treponemal tests for screening followed by confirmation with a treponemal test, an algorithm implemented worldwide.To evaluate the performance of treponemal and non-treponemal tests in the effective detection of syphilis among blood donors.A cross-sectional study was conducted with 384 samples, initially reactive for syphilis by non-treponemal tests (VDRL). All samples were retested by VDRL, ELISA, CLIA, and FTA-ABS, the latter as a confirmatory test. We calculated kappa coefficients to establish agreement between tests and determine performance indices for each assay.We identified agreement between non-treponemal and treponemal tests. Concordance between VDRL and ELISA was 78.8% and between CLIA and VDRL was of 76% (p < 0.005). The concordance between ELISA and CLIA was 83%. When comparing the results of treponemal and non-treponemal tests with FTA-ABS, concordance rates ranged from 44.2% to 61.9%. Performance indices showed that non-treponemal and treponemal tests exhibit sensitivity values between 89.70% and 99.39%. The positive predictive value was higher for the CLIA test (95.27%), whereas the highest false-positive rate was observed for the VDRL test (94.52%).The samples analyzed were reactive from the outset, which could interfere with the measurement parameters. However, they provided valuable information to evaluate the implemented algorithms.