Screening for syphilis with dual algorithms: analysis of discordant and concordant serology results in a population with a low prevalence of syphilis.

ABSTRACT

BACKGROUND: Currently, many laboratories have switched the traditional screening algorithm (TSA) to reverse screening algorithm (RSA) for the efficiencies in high-volume syphilis screening. However, confusions have been arisen regarding this paradigm shift. OBJECTIVE: To compare the performance of two algorithms with head-to-head mode. METHODS: Sera screening for syphilis were tested in parallel with chemiluminescence immunoassay (CIA) and toluidine red unheated serum test (TRUST). CIA-reactive sera from the RSA were reflexively tested with TRUST and confirmed with Treponema pallidum particle agglutination assay (TPPA), while the TRUST-reactive serology from the TSA were afterwards tested with TPPA. RESULTS: A total of 110 663 serum samples were screened. The RSA identified 2259 (2.0%) CIA-reactive results, of which 377 (16.7%) showed TPPA nonreactive results, while the TSA identified 934 (0.8%) TRUST-reactive results, of which 67 (7.2%) showed TPPA-nonreactive results. Among the 2259 CIA-reactive results, 1392 (61.6%) were TRUST-nonreactive, of which 350 (25.1%) were TPPA-nonreactive. A total of 182 sera from the 350 TPPA-nonreactive sera were further tested by a second CIA (VITROS Syphilis TPA, VITROS TPA), of which 155 (85.2%) were nonreactive and 27 (14.8%) were reactive. The 27 VITROS TPA-reactive sera were further tested with a treponemal Western blot assay (Euroimmun IgG Western Blot, EuroWB), of which 11 (41%) were indeterminate, 6 (22%) were nonreactive and 10 (37%) were reactive. Among the 10 EuroWB-reactive sera, two seroconverted to TPPA 180+/- after 1-year follow-up. Of 867 CIA-reactive/TRUST-reactive results, 27 (3.1%) were TPPA-nonreactive. CONCLUSIONS: The RSA identified more patients with reactive treponemal serology. However, it also yielded an increased likely false-reactive rate compared with the TSA, especially those results with low index values and TRUST-nonreactive serology, were necessary to retest with a second treponemal test. Further testing results with TPPA, VITROS TPA and EuroWB suggested the false-reactive CIA screening results and the likely false-nonreactive TPPA results when the reactive treponemal results screened with RSA were to be identified.