Background: The limited literature available on syphilis re-infection suggests that it is occurring to a high degree in locations experiencing syphilis outbreaks.
Objectives: To characterize individuals who are repeatedly infected with syphilis in Philadelphia, and compare the characteristics of repeaters to those infected one time.
Methods: Among individuals over 14 years old with any stage of syphilis reported during 2002-2008, those with subsequent primary, secondary, or early latent (with supportive testing, treatment, symptom, or sexual histories) diagnoses were considered recidivists. A multivariable logistic model was fit to identify factors associated with recidivism.
Results: Of the 3475 individuals reported with syphilis, 117 (3%) had at least one repeat infection. Most recidivists had one subsequent infection (92%), eight individuals (7%) had two, and one person had three subsequent infections. The largest group of recidivists was African American men who have sex with men (MSM) (32%) followed by white MSM (27%); whereas, for non-recidivists the largest groups were heterosexual African American males (27%) and females (15%). The median time between first and second infection for recidivists was 21 months (range: 2-85 months). The median age at first infection for recidivists was slightly younger than for those with single infections – 35 years (range: 16-60 years) versus 39 years (range: 15-97 years). In the male only model (females had no associations), recidivism was significantly associated with reporting as MSM or bisexual and being under 30 years old with interactions between age and sexual orientation.
Conclusions: A growing proportion of Philadelphia’s syphilis cases are among individuals infected repeatedly. The small but significant number of male recidivists identifying as bisexual could have implications in the recent increase in female syphilis cases in Philadelphia.
Implications for Programs, Policy, and/or Research: Prevention messages should include messages about repeat infections. Sexual network models could help assess the impact of recidivism.