Abstract Background Use of treponemal tests to screen for syphilis (caused by Treponema pallidum pallidum ) poses challenges with infection status classification, especially in settings where other treponemal infections are endemic. This study aimed to determine the validity of the syphilis surveillance testing strategy implemented since 2004 using two treponemal tests, and estimate the seroprevalence of active syphilis infection in Ghana where yaws (caused by Treponema pallidum pertenue) is endemic. Methods We retested sera from the 2007 HIV sentinel survey (HSS) using a traditional algorithm, quantitative rapid plasma reagin test followed by qualitative Treponema pallidum haemagglutination assay. The adjusted seroprevalence of active syphilis was calculated by applying the proportions of active syphilis within identified categories of HSS samples during the retesting, to the entire population of HSS samples. The 95% confidence intervals (CIs) were calculated for each proportion, and the t -test was used to assess differences in proportions. Results Of 2,214 samples that were retested, 203 (9.2%) had active syphilis infection, 21 (0.9%) were biological false reactions, 640 (28.9%) were past or treated syphilis infections, and 1,350 (61%) were uninfected. The current syphilis testing strategy overestimated the seroprevalence of active syphilis infection by a third (HSS versus traditional algorithm: 6.0% (95% CI: 5.6–6.3) vs. 4.5% (95% CI: 4.2–4.8); p