Comparison of Serological and Molecular Treponema pallidum Tests in HIV Patients in Kenya

Abstract

Syphilis is a common co-infection among people living with Human Immunodeficiency Virus (HIV), where it has a significant impact. Syphilis is caused by Treponema pallidum and is typically diagnosed using two main types of tests: non treponemal tests, such as rapid plasma reagin (RPR) and venereal disease research laboratory (VDRL), and treponemal tests, like the Treponema pallidum hemagglutination assay (TPHA)- considered the gold standard in Kenya. The sensitivity of these serological tests for primary syphilis generally ranges between 70 80%, Currently molecular methods such as Polymerase chain reaction (PCR) offers higher sensitivity and specificity than serological tests. This study, therefore, compared the performance of serological tests (VDRL, RPR, TPHA) against PCR in detecting syphilis among HIV patients attending the Comprehensive Care Clinic (CCC) at Nyeri County Referral Hospital (NCRH). This cross-sectional study consented and recruited a random sample of 177 HIV- positive patients, who were tested with three serological tests (RPR, VDRL, and TPHA) and one molecular method (PCR). A sociodemographic questionnaire was also administered. Key outcomes were syphilis prevalence and test accuracy. Participants had a mean (Standard deviation – SD) age of 48.3 years (SD 11. 07); 40. 1% were 51 or older. Females comprised 60.5%, 58.8% were married, 48.6% had been HIV positive for 1 8 years, 88.1% were on a lamivudine/ tenofovir disoproxil fumarate / dolutegravir / adefovir-diphosphate (3TC/TDF/DTG/AFV) ART regimen, and 7.9% experienced virological failure. Only 1.1% previously tested positive for syphilis. PCR detected a syphilis prevalence of 18.6%. Test performance metrics in this high-prevalence setting (18.6% by PCR) were: RPR—100% sensitivity, 76.4% specificity, and a kappa of 0. 0.546 (moderate agreement); VDRL—100% sensitivity, 55.6% specificity, and a kappa of 0. 0.317 (fair agreement); TPHA—100% sensitivity, 94.4% specificity, and a kappa of 0.864 (near- perfect agreement); RPR/VDRL combined with TPHA— 100% sensitivity, 54.2% specificity, and a kappa of 0. 0.306 (fair agreement). Using TPHA as the gold standard, the sensitivity, specificity, and kappa for RPR and VDRL were 97%, 97%, 0.63%, and 97.6%, 58. 1%, 0.377, respectively. Combining RPR/VDRL yielded similar results. Switching the gold standard to TPHA slightly reduced the sensitivity of RPR and VDRL and increased their specificity. Considering weakly reactive TPHA samples as negative improved the sensitivity and specificity of TPHA to 100% against PCR. In summary, neither RPR nor VDRL can be used alone or as confirmatory tests. TPHA, with weakly reactive samples considered negative, had the highest agreement with PCR. Nonetheless, low- positive TPHA results should be interpreted with caution, as they may represent early seroconversion or false positives.