Background: Traditionally, nontreponemal tests that detect serological evidence of T. pallidum infection are used for syphilis screening. Reactive tests are confirmed with treponemal tests that detect antibodies to T. pallidum We examined the feasibility and performance of a reverse sequence diagnostic algorithm, screening with a treponemal test (e.g. IA) followed by a nontreponemal test for confirmation, in a high prevalence setting.
Methods: 2,350 serum specimens were prospectively collected from STD clinics serving at-risk clients. The traditional algorithm of VDRL followed by TPPA was performed on all specimens. Specimens were then blinded and tested by the TrepSure EIA and TPPA. Results of all three testing methodologies were analyzed to evaluate the performance of both algorithms.
Results: 198 (8.4%) out of 2,350 specimens were VDRL-reactive, of which 189 were reactive on TPPA thus resulting in the detection of 189 infections. When the same specimens were screened by EIA, 478 (20.3%) were reactive, 186 (38.9%) of which confirmed reactive by VDRL. Of the 292 (61.1%) discordant EIA-reactive and VDRL non-reactive specimens, only 249 (85.3%) were reactive by TPPA. Signal-to-cutoff values of all EIA results revealed very strong correlation with the probability of reactivity on a TPPA.
Conclusions: Screening with an EIA for syphilis infection in a high prevalence setting resulted in the detection of more seropositive individuals than screening with VDRL, but did not result in the detection of more syphilis cases. An algorithm that begins with an EIA screen may require additional tests for the resolution of discordant cases. When EIA-reactive, VDRL non-reactive specimens are detected, the use of TPPA as a reflex test may not provide clear guidance regarding treatment. The signal-to-cutoff ratio on the initial EIA could potentially be used to predict TPPA reactivity as in many cases, it could eliminate, the need to perform the TPPA test.