Background: During 19972008, early syphilis cases in the County of San Diego (COSD) increased over 1200%, from 39 to 525, with 412/525 cases (78%) in 2008 occurring among men who have sex with men (MSM). In other jurisdictions, MSM with syphilis, particularly if co-infected with HIV, have been shown to be at high risk of repeat syphilis, and, therefore, might represent important targets for enhanced syphilis prevention and control efforts.
Objectives: To describe the proportion of MSM in COSD diagnosed with early syphilis during January 2004June 2007 who subsequently had repeat syphilis, and to identify predictors of repeat syphilis.
Methods: We defined repeat syphilis as a case of syphilis occurring within two years in a previously appropriately treated person with a four-fold titer decrease and then signs of syphilis or a four-fold titer increase. We used univariable logistic regression to identify patient characteristics, obtained from syphilis interview records, that significantly predicted repeat syphilis. We then entered characteristics that significantly predicted repeat syphilis into a multivariable prediction model, retaining those with p <0.05, and calculated the area under the curve (AUC) for the final model.
Results: Among 634 MSM in COSD diagnosed with syphilis, 74 (11.7%, 95% confidence interval [CI], 9.3%14.4%) had repeat syphilis. HIV-infected MSM comprised 60% of the sample and 74% of those with repeat syphilis. Among MSM with known HIV status (n=614), repeat syphilis was significantly associated only with HIV infection (odds ratio 1.95, 95% CI 1.12, 3.37), with an AUC for the final model of 0.57.
Conclusions: HIV-infected MSM in COSD diagnosed with syphilis are at high risk of repeat syphilis within two years.
Implications for Programs, Policy, and/or Research: COSD should consider developing interventions to prevent and control repeat syphilis among HIV-infected MSM. This prediction model for repeat syphilis should be validated.