Background

Syphilis is concentrated among high-risk groups, but the epidemiology of syphilis reinfection is poorly understood. We characterized factors associated with syphilis incidence, including reinfection, in a high-risk cohort in Peru.

Methods

Participants in the NIMH CPOL trial were assessed at baseline and 2 annual visits with HIV/STI testing and behavioral surveys. Participants diagnosed with syphilis also attended 4- and 9-month visits. All participants underwent syphilis testing with RPR screening and TPPA confirmation. Antibiotic treatment was provided according to CDC guidelines. Reinfection was defined as a 4-fold titer increase or recurrence of seroreactivity after successful treatment with subsequent negative RPR titers. The longitudinal analysis used a Possion generalized estimating equations model with backward selection of variables in the final model (criteria P <0.02).

Results

Of 2,709 participants, 191 (7.05%) were RPR-reactive (median 1:8, range 1:1-1:1024) with TPPA confirmation. There were 119 total cases of incident syphilis, which included both reinfection and first-time incident cases. In the bivariate analysis, the oldest 2 quartiles of age (incidence ratio (IR) 3.84; P <0.001 and IR 8.15; P <0.001) and being MSM/TW (IR 6.48; P <0.001) were associated with higher risk of incident syphilis infection. Of the sexual risk behaviors, older age of sexual debut (IR 12.53; P <0.001), not being in a stable partnership (IR 1.56, P = 0.035), higher number of sex partners (IR 3.01; P <0.001), unprotected sex in the past 3 months (IR 0.56; P = 0.003), HIV infection at baseline (IR 3.98; P <0.001) and incident HIV infection during the study period (IR 6.26; P = 0.003) were all associated with incident syphilis. In the multivariable analysis, older age group (adjusted incidence ratio (aIR) 6.18; P <0.001), men reporting having sex with a man (aIR 4.63; P <0.001), and incident HIV infection (aIR 4.48; P = 0.008) were significantly associated.

Conclusions

We report a high rate of syphilis reinfection among high-risk men who have evidence of previous syphilis infection. Our findings highlight the close relationship between HIV incidence with both incident syphilis and syphilis reinfection. Further studies on syphilis reinfection are needed to understand patterns of syphilis reinfection and new strategies beyond periodic testing of high-risk individuals based on HIV status are needed.