Objective: No single serology test on its own can confirm a syphilis diagnosis. We assessed the utility of a reactive treponemal test (TT) against a non-treponemal test (NTT), each as a sole indicator of infection warranting immediate treatment and the performances of the traditional algorithm (TA) vs. reverse algorithm (RA) in the serodiagnosis of syphilis in a high human immunodeficiency virus (HIV) prevalence setting.

Patients and Methods: A rapid plasma reagin (RPR)-NTT and an immunochromatographic (IC)-TT were tested in parallel for the presence of syphilis antibodies in fresh baseline sera of pregnant women living with/without HIV (WLWH/WLWoH) ≥20 weeks gestational age. Reactive samples were confirmed using either the TA or RA. Indeterminate results were resolved using a second but different IC-TT, which detected different spirochaete antigen preparations from those present in the first one-step syphilis screening (1-SSS)-IC-TT.

Results: 1,208 pregnant women were enrolled in the study. 600 were WLWoH, whilst 608 were WLWH. 1-SSS-RPR-NTT detected early syphilis at 28 (2.3%) seroprevalence, while 25 (2.1%) were further confirmed positive by the IC-TT. Re-testing of the three indeterminate samples using the second but different IC-TT confirmed two more cases, slightly increasing the seroprevalence to 2.2%. One WLWoH result remained indeterminate. Using 1-SSS-IC-TT first, early/past syphilis cases were 53 (4.4%), and 25 (2.1%) were further confirmed active infections using the RPR-NTT. Of the 28/53 samples reactive on 1-SSS-IC-TT but negative on RPR-NTT, 20 (1.7%) tested positive on the second but different IC-TT, indicating potentially past infections. However, 8 cases remained inconclusive, with 5 being from WLWoH. Twenty-eight (28) of the 53 1-SSS-IC-TT positives presenting with past syphilis infections or inconclusive test results were immediately treated together with the 25 active cases appropriately requiring treatment, resulting in these 28 women receiving unnecessary interventions (overtreatment). This was in contrast with the 1-SSS-RPR-NTT first approach, which resulted in only 1 case of overtreatment. Potential false-positivity rates were 8/1,208 (0.7%) and 1/1,208 (0.08%), respectively. Sero-diagnosis of syphilis, as determined by the RA and TA, was both at 25/1,208 (2.1%), showing a 100% agreement.

Conclusions: 1-SSS-RPR-NTT detected more active cases at a quarter of the cost of 1-SSS-IC-TT. 1-SSS-IC-TT often resulted in antibiotic over-treatments in patients presenting with previous infections, posing a public health threat of antimicrobial resistance. Sero-diagnostic performance was comparable between TA and RA. Interestingly, inconclusive results were relatively common in WLWoH.