Introduction Rapid tests (RTs) improve screening and treatment in people without ready access to laboratory testing. Newer RTs linking syphilis and HIV tests on a single device could simplify testing and lower costs. Previous laboratory evaluations of dual HIV/Syphilis RTs suggest high accuracy using trained technicians and characterised sera. However, data from field evaluations would more likely represent “real world” performance.

Methods We searched the medical literature for studies evaluating performance of dual syphilis/HIV RTs against laboratory-based reference tests for syphilis and HIV, and compared performance across studies. For the syphilis component of the RTs, we compared results using laboratory-based treponemal tests (TPPA or TPHA) as the reference and (when available) TPPA+/RPR+ as the reference, considering RPR titers>1:4 to represent active syphilis (vs. previously treated infections).

Results We found 19 studies evaluating dual syphilis/HIV RT performance, of which 7 (37%) were field evaluations studying at least one of three diagnostics: SD Bioline HIV/Syphilis Duo Test (n=4); Chembio Dual Path Platform HIV-Syphilis Assay (n=2); or Medmira Multiplo Rapid TP/HIV Antibody Test (n=1). All used HIV EIA and TPPA or TPHA tests as reference standards; 6 also reported RPR titers. Study populations were pregnant women (n=3), female sex workers (n=1), high-risk men (n=2) and STD clients (n=1), representing a total of 13 915 persons (median study size, 415 participants; range 175 – 9983). Across studies, prevalence of HIV ranged from <1% to 78% (median, 25.3%), and of T. pallidum (TP) from <1% to 40.2% (median, 8.2%). RT sensitivity for HIV against EIA ranged from 93.8% to 100% (median, 99.1%), and specificity from 97% to 100% (median, 99.4%). RT sensitivity for TP against TPPA or TPHA ranged from 52.7% to 96.5% (median, 81%), and specificity from 89% to 100% (median, 98.8%), with better performance in study populations with higher TPPA/RPR+ prevalence. Using TPPA+/RPR+>1:4 as the standard, RT sensitivity ranged from 88.5% to 100% (median, 94.3%).

Conclusion In the few published field evaluations of dual syphilis/HIV RTs, performance of the HIV component was high for all tests studied. Sensitivity of the syphilis component against TPPA was poorer, but was more accurate using probable active syphilis infection as the standard.