BACKGROUND: In 2005, syphilis screening in the Greater Toronto Area of Canada moved from the rapid plasma reagin (RPR) to a treponemal enzyme immunoassay (EIA). We sought to understand the consequences of this change on laboratory results and testing patterns with a population-based retrospective study of laboratory-based diagnoses of syphilis. METHODS: Samples positive under RPR (1998-2005) and EIA (2005-2008) screening were confirmed with an alternate treponemal test, and during the latter period underwent RPR testing. We compared monthly rates and the forecasting relationship between positives and future submissions with time-series methods, and assessed risk factors for EIA(+)/RPR(-) results using Poisson regression. RESULTS: A total of 3,092,938 submissions were included. Following EIA implementation, confirmed positive rates increased by 10.3 per 100,000 population (P<0.001). 0.59% of EIA(+)/RPR(-) individuals converted to RPR(+) within 2 months. EIA(+)/RPR(-) patients were more likely to be male (incidence rate ratio [IRR]: 2.3, 95% confidence interval [CI]: 1.6-2.5), asymptomatic (IRR: 1.8, 95% CI: 1.3-2.8), and aged>50 years (IRR: 2.4, 95% CI: 1.6-3.5) than those with EIA(+)/RPR(+) results. We detected a significant positive feedback loop between positive tests and subsequent submissions. This relationship was only transiently evident for EIA(+)/RPR(-) results up to 1 year following the changeover. CONCLUSIONS: EIA screening facilitates identification of probable latent syphilis and earlier serological detection of infectious syphilis, but may transiently cause increases in testing and indirectly suggests that physicians' interpretation of RPR(-) serology may lead to partner testing. In the absence of a true gold standard, implementation of EIA screening warrants careful communication regarding serological interpretation.
BACKGROUND: In 2005, syphilis screening in the Greater Toronto Area of Canada moved from the rapid plasma reagin (RPR) to a treponemal enzyme immunoassay (EIA). We sought to understand the consequences of this change on laboratory results and testing patterns with a population-based retrospective study of laboratory-based diagnoses of syphilis. METHODS: Samples positive under RPR (1998-2005) and EIA (2005-2008) screening were confirmed with an alternate treponemal test, and during the latter period underwent RPR testing. We compared monthly rates and the forecasting relationship between positives and future submissions with time-series methods, and assessed risk factors for EIA(+)/RPR(-) results using Poisson regression. RESULTS: A total of 3,092,938 submissions were included. Following EIA implementation, confirmed positive rates increased by 10.3 per 100,000 population (P<0.001). 0.59% of EIA(+)/RPR(-) individuals converted to RPR(+) within 2 months. EIA(+)/RPR(-) patients were more likely to be male (incidence rate ratio [IRR]: 2.3, 95% confidence interval [CI]: 1.6-2.5), asymptomatic (IRR: 1.8, 95% CI: 1.3-2.8), and aged>50 years (IRR: 2.4, 95% CI: 1.6-3.5) than those with EIA(+)/RPR(+) results. We detected a significant positive feedback loop between positive tests and subsequent submissions. This relationship was only transiently evident for EIA(+)/RPR(-) results up to 1 year following the changeover. CONCLUSIONS: EIA screening facilitates identification of probable latent syphilis and earlier serological detection of infectious syphilis, but may transiently cause increases in testing and indirectly suggests that physicians' interpretation of RPR(-) serology may lead to partner testing. In the absence of a true gold standard, implementation of EIA screening warrants careful communication regarding serological interpretation.