Background Serological testing continues to be a crucial tool for syphilis diagnosis and control. The commonly accepted syphilis screening algorithm is screening with non-treponemal tests such as RPR or VDRL, and confirming with treponemal tests such as TP-PA. Recently, automation has been introduced whereby serological screening using treponemal tests has resulted in reduced labour time and removal of the subjectivity associated with the traditional testing algorithm. The objective of this study was to compare the performance characteristics of two FDA approved automated tests, the BioRad BioPlex 2200 Syphilis IgG and the DiaSorin LIAISON treponemal assays, with known predicate tests. The BioPlex 2200 syphilis IgG is a multiplex test that utilises three analytes (15-, 17-, & 47-kDa) to detect specific IgG antibodies, whereas the LIASION treponemal assay uses only one analyte (17 kDa) in a single step sandwich method to detect both syphilis IgG and IgM antibodies.
Methods A total of 1086 commercially obtained sera tested in this study consisted of: 430 from pregnant women, 409 from HIV positive individuals, and 111 from known syphilis patients of various disease stages. Characterised syphilis samples (n=140) were also obtained from the CDC serum repository. All samples were screened by the Bioplex IgG, Liaison, RPR and TP-PA tests. Any indeterminate results were repeated at least once.
Results Of the 1086 samples tested, the syphilis reactivities were the following: 551 (50.7%) by BioPlex IgG, 528 (48.6%) by LIAISON, and 509 (46.9%) by TP-PA. The sensitivity and specificity when compared to TP-PA for LIASION was 98.8% and 90.5% respectively. The BioPlex IgG sensitivity and specificity when compared to TP-PA was 85.1% and 80% respectively. Overall, 443 (40.8%) samples were found to be reactive and 450 (41.4%) non reactive to both LIAISON and BioPlex IgG. All three tests agreed on 877 (81%) samples. On the 209 discordant samples TP-PA agreed with LIAISON 85.2% (n=178), BioPlex 7.2% (n=15), but disagreed with both tests 7.7% (n=16).
Conclusion Both tests have high throughput, walk-away capability, and would be useful in low prevalence settings. There was good agreement between the LIAISON and the BioPlex IgG in 893 (82%) samples (Cohen's κ=0.64). The LIAISON had higher sensitivity most likely due to its detection of both IgG and IgM, while the BioPlex detected only IgG antibodies. Both tests show significant promise in the future of syphilis serology.