(P390) Performance evaluation of syphilis testing algorithms in a high prevalence population of men and transgender women who have sex with men

Background: Despite a global syphilis resurgence, data on performance of syphilis testing algorithms in high prevalence populations are sparse. We conducted a sub-study of stored specimens from the Pre-exposure Prophylaxis Initiative (iPrEx) study, which enrolled men and transgender women who have sex with men from 2007-2009 from six countries to compare the traditional versus reverse sequence syphilis (RSS) testing algorithms.

Methods: The CDC STD Laboratory Reference and Research Branch received frozen sera from iPrEx participants’ baseline visit (n=1198). Specimens were tested for syphilis within 5 days after thawing using the Rapid Plasma Reagin (RPR, Arlington Scientific) for nontreponemal antibodies, and for treponemal-specific antibodies using the T. pallidum particle agglutination (Serodia TP-PA, Fujirebio) and TrepSure Enzyme immunoassay (EIA, Trinity Biotech), following respective manufacturer’s instructions. Specimens that were either TP-PA or EIA reactive were tested in INNO-LIA syphilis score (LIA, Line immunoassay, Fujirebio). The EIA signal was evaluated in context to TP-PA and LIA test results.

Results: Of 1198 baseline sera, 133 (11.1%) were RPR reactive, of which 111 (83.5%) were confirmed with TP-PA in the traditional algorithm. When EIA replaced TP-PA in the same algorithm, 127 (95.5%) cases were confirmed. In the RSS algorithm, initial EIA testing was positive in 283 (23.6%) specimens, of which 127 (45%) were reactive in RPR and 156 (55%) as nonreactive. Inclusion of TP-PA as a confirmatory test in RSS for EIA+/RPR- gave 90 (57.7%) sera as reactive and 66 (42.3%) as nonreactive. When the TP-PA was replaced with LIA for confirmation of EIA+/RPR- specimens, 118 (75.6%) sera were reactive and 38 (24.4%) were nonreactive. Comparison of EIA signal strength to TP-PA and EIA reactivity showed concordance of 66-99% for TP-PA and 95-100% for LIA at an antibody index of 10-15 and above.

Conclusion: The RSS detected more syphilis cases compared to the traditional algorithm in this high sero-prevalence population. Increasing the EIA signal strength could minimize the proportion of specimens requiring confirmatory testing in the RSS; however, a stronger correlation between screening and confirmatory treponemal tests are required along with validated antibody index cutoffs for implementation.