EFFORTS TO CONTROL SYPHILIS in the United States have broadened in recent years, with launching of the National Plan to Eliminate Syphilis from the United States in 19991 and release of an updated Syphilis Elimination Effort action plan in 2006.2 However, there has not been a national population-based measure of syphilis seroprevalence in 25 years.3 Syphilis surveillance in the United States has relied largely on case reports submitted by state and local health departments and seroprevalence monitoring among specific high-risk populations (e.g., in jails), which have been useful in following trends in early syphilis to gauge the effectiveness of prevention efforts.4 However, case report data may underestimate the true burden of syphilis, due to such factors as lack of screening among many asymptomatic infected persons and underreporting of diagnosed cases. In addition, case reporting and selection biases may contribute to the marked racial disparities seen for syphilis, in which blacks are disproportionately affected.4 Blacks may be more likely to attend public clinics where case reporting is thought to be most complete, and may be overrepresented at the limited sites where screening occurs. We analyzed data from the National Health and Nutrition Examination Surveys (NHANES) 2001–2004 to determine the prevalence of positive serologic tests for syphilis among a representative sample of the US population aged 18 to 49 years.
Study Population and Survey Design
The NHANES 2001–2004 were continuous annual surveys conducted by the National Center for Health Statistics.5 Each year, a complex, stratified, multistage probability sample design was used to select a representative sample of the civilian, noninstitutionalized US population of all ages. Race/ethnicity categories in NHANES were defined by self-report as non-Hispanic (NH) white, NH black, and Mexican-American. Persons who did not fit into one of these categories, including those who reported being “multiracial” were classified as “other” and included in the total population. NH blacks and Mexican-Americans were sampled at a higher rate compared with other racial/ethnic groups to allow sufficient sample sizes for analysis.
Each survey participant was interviewed and underwent a standardized health examination and biologic specimen collection. Data on sexual behavior were collected using audio computer-assisted self-interview, and sex was defined as vaginal, oral, or anal sex. NHANES participants were not asked about a prior clinical diagnosis of syphilis. All participants tested for syphilis were given an opportunity to obtain their test results by telephone using a password-protected system. Reminder letters were sent to adults and phone calls were made to minors to encourage participants to learn their test results. Participants with positive tests received standardized counseling messages and were referred to health care providers for evaluation and possible treatment.
Protocols for the NHANES 2001–2004 were evaluated and approved by the institutional review board of the US Centers for Disease Control and Prevention (CDC). Written, informed consent was obtained from all survey participants.
Sera from 18- to 49-year-old NHANES 2001–2004 participants were evaluated at CDC using serologic tests for syphilis infection according to a standardized algorithm based on screening with a treponemal test, a strategy found to have favorable performance in other settings.6–8 Detailed methods for all laboratory tests in the algorithm, including cut-off values for positive and indeterminate results, are documented on the NHANES Web site.9 All specimens were first tested for syphilis IgG treponemal antibody using an enzyme immunoassay (EIA). During 2001 to November 2002, the Captia Syphilis-G EIA was used.10–12 However, because of availability issues with the Captia test, during November 2002, all EIA testing was switched to the Trep-Chek Syphilis EIA.13,14 Performance of the two EIA tests was comparable on internal quality assurance evaluations (side-by-side testing). Specimens with positive or indeterminate EIAs (of either type) underwent nontreponemal rapid plasma reagin (RPR) testing.15,16 Specimens with nonreactive RPRs or RPR titers <1:8 underwent confirmatory testing with Treponema pallidum particle agglutination (TP-PA).17
The algorithm for defining a final determination of syphilis infection was as follows. Specimens with a positive or indeterminate EIA and an RPR titer ≥1:8 were considered to be positive for syphilis, and more likely to reflect recent infection. Specimens with a positive or indeterminate EIA, a nonreactive RPR or an RPR titer <1:8, and a positive TP-PA were considered to be positive for syphilis, and more likely to reflect remote infection. Specimens with a negative EIA or a negative TP-PA were considered to be negative for recent or remote syphilis infection.
SUDAAN software version 9.0 (Research Triangle Institute, Cary, NC) was used for statistical analyses to incorporate sampling weights and account for the clustered sample design in calculating variance estimates based on a Taylor Series approximation. All seroprevalence estimates were weighted to represent the total US civilian, noninstitutionalized population and to account for oversampling and nonresponse to the household interview and physical examination. We estimated weighted syphilis seroprevalence overall and by selected demographic and behavioral variables. Bivariate significance tests for association were based on a Wald χ2 statistic with 2-sided P-values. The confidence intervals for the prevalence estimates were calculated using the Clopper-Pearson method.18,19 The critical values used for the confidence intervals were based on the t-distribution, with degrees of freedom equal to the number of strata minus the number of primary sampling units. Relative standard errors (RSEs) were computed for each weighted estimate as (SE/Estimate)*100. An RSE >30% is considered unstable and the corresponding estimate should be interpreted cautiously.
We also performed a weighted analysis of the positive predictive value (PPV) of the EIA screening test, alone and in combination with RPR results, for a final determination of syphilis infection as defined by our laboratory testing algorithm. The PPV was defined as the proportion of initially positive tests that were ultimately determined to represent syphilis infection, after accounting for the sampling methods of NHANES 2001–2004.
A total of 25, 917 individuals were selected to participate in NHANES 2001–2004. The overall response rate to the interview was 82%. Of 6540 18- to 49-year-olds interviewed, 95.9% were examined, and 91.9% of those tested for syphilis. Overall, 5767 specimens underwent syphilis EIA testing; testing algorithm results are shown in Figure 1. EIA testing was positive or indeterminate for 126 specimens, of which 10 had RPR titers ≥1:8. Of the remaining 116 specimens, 60 had positive TP-PA tests. All 19 specimens with an RPR titer ≥1:1 had a positive TP-PA, versus 41 of 97 nonreactive specimens.
The prevalence of syphilis seroreactivity among 18- to 49-year-olds in the United States was 0.71% (95% CI: 0.51–0.96; Table 1). Prevalence of syphilis infection with a nonreactive RPR or RPR titer <1:8 (more likely remote infection) was 0.63% (95% CI: 0.44–0.87) and that of higher titer infection (RPR titer ≥1:8, more likely recent infection) was 0.08% (95% CI: 0.03–0.20); however, the latter estimate had a RSE >30% and should be interpreted with caution. Table 1 shows overall syphilis seroprevalence (both recent and remote infection) according to selected characteristics. Seroprevalence was similar among males and females (0.76% vs. 0.67%; P = 0.59) and increased with age, less education, and lower income (P <0.002 for each). NH blacks had the highest seroprevalence (4.3%), followed by Mexican-Americans (0.98%) and NH whites (0.07%; P <0.001), although small numbers of syphilis cases among NH whites made a precise estimate unreliable (RSE >30%).
Syphilis seroprevalence was higher among those who had first sex before the age of 15 years than among those whose first sex was later (1.5% vs. 0.45%, P = 0.006; Table 1). Seroprevalence was also higher among persons reporting >5 lifetime sex partners, compared with those with ≤5 lifetime partners (P <0.001). Small numbers of men reporting a history of a same-sex sex partner and those reporting homosexual, bisexual, or “other” sexual identity limited our ability to make reliable estimates of syphilis seroprevalence among men who have sex with men (MSM). Although syphilis seroprevalence was slightly higher among men who had a history of a same-sex sex partner (1.4%) and among those who self-identified as nonheterosexual (2.2%) compared with those who did not (0.6% for each), these differences were not statistically significant. Among the 28 NHANES participants with a positive HIV test, syphilis seroprevalence was 13.3%.
Table 2 shows the PPV of the EIA screening test, alone and in combination with RPR results, for a final determination of syphilis infection as defined by our laboratory algorithm. Of 112 specimens with positive EIAs, 69 were ultimately determined to be positive for syphilis, yielding a weighted PPV of 48%. In other words, 52% of all positive screening EIAs were ultimately deemed to be false positives. All 28 specimens with a positive EIA and a reactive RPR (titer ≥1:1) were determined to be positive for syphilis (PPV: 100%), whereas only 41 of 84 specimens with positive EIAs and nonreactive RPRs had syphilis (PPV: 37%). Of 14 specimens with indeterminate EIAs, only 1 (with an RPR titer of 1:4) had syphilis.
We found that the prevalence of syphilis seroreactivity was low (0.71%) in the general, noninstitutionalized US population of 18- to 49-year-olds. However, there were marked racial disparities in syphilis infections, with NH blacks disproportionately affected. This nationally representative survey supports surveillance data; case reporting and selection biases are unlikely to play a major role in explaining apparent racial disparities.4 Because positive treponemal serologic tests are a marker for lifetime syphilis infection, the magnitude of disparity among racial/ethnic groups does not necessarily reflect current differences in syphilis infections, but rather differences in cumulative incidence for each group over the previous approximately 15 to 20 years. The magnitude of the differences between NH blacks and NH whites in NHANES 2001–2004 was comparable to that observed in 1992 national surveillance data, where the black:white rate ratio was 62:1. It is encouraging that the black:white rate ratio in surveillance data has subsequently decreased to 5.4:1 in 2005.4
Although progress has been made in reducing racial disparities in syphilis in recent years, reports of outbreaks and increased numbers of primary and secondary syphilis cases among MSM have been on the rise.20–23 In fact, currently more than 60% of new syphilis infections are estimated to occur in MSM.23 In our analysis, small numbers of MSM made precise syphilis seroprevalence estimates unreliable. Syphilis seroprevalence appeared to be higher in MSM, but differences by MSM status were not statistically significant. Since we evaluated lifetime syphilis infections that may have occurred over many years, recent increases may not have been as apparent. In addition, the NHANES surveys are not well-suited for evaluating low prevalence diseases like syphilis in small subpopulations such as MSM that have not been explicitly oversampled. Groups of MSM with high-risk behavior and high syphilis rates that are localized within a few urban centers may be missed with the broad, population-based sampling for NHANES.
In addition to evaluating seroprevalence, we used the NHANES data to evaluate laboratory screening with treponemal tests. Traditionally, testing for syphilis in the United States has relied on screening with a nontreponemal test followed by a confirmatory treponemal test. However, many sites in the United States are adopting initial screening with treponemal tests, instead of nontreponemal tests, because of economic and logistical considerations,24 and several have documented that treponemal-based screening algorithms are a reasonable alternative to algorithms based on nontreponemal screening, with comparable sensitivity, specificity, and predictive values.6–8 Screening with our treponemal testing algorithm identified 41 specimens that were determined to be positive for syphilis (59% of all those found), which would have been missed with nontreponemal RPR screening. Some of these may have represented either very early or very longstanding latent syphilis infection, situations that might still benefit from therapy. However, lack of information on prior diagnosis and treatment makes it unclear whether these had clinical or public health importance. The PPV of screening EIA in our assessment was 48%. Even with very specific treponemal tests, we would predict a substantial number of false positives in a very low prevalence population, so the PPV we observed was not out of the range of what might be expected. When a positive screening EIA was followed by a reactive RPR (of any titer), the PPV was 100% for a final determination of syphilis infection, suggesting that confirmation with another treponemal test would not be necessary in this scenario. However, the PPV of the screening EIA was much lower if the follow-up RPR was nonreactive. In this scenario, confirmatory testing might be helpful, although interpretation of results is not always clear. Although a positive second treponemal test would likely confirm a syphilis infection, a negative second treponemal test might represent a false positive on the first test or a false negative on the second; for purposes of our analysis, we assumed the first test to be falsely positive. Use of treponemal tests for syphilis screening will require careful consideration of clinical and therapeutic histories when treponemal tests are positive in the setting of nonreactive nontreponemal tests.
Our analysis had several limitations. First, low prevalence led to unstable estimates for many subpopulations, including NH whites and MSM. Second, syphilis treponemal seroreactivity indicates infection at any time over a lifetime, and so these estimates do not shed light on new or untreated syphilis infections in recent years. Categorization as “recent” or “remote” infection using RPR titer will undoubtedly result in misclassification, and the number of infections we identified with RPR titers ≥1:8 was too small to make reliable estimates. Third, no information about a reported history of diagnosed or treated syphilis was available; thus, we could not determine the extent of undiagnosed syphilis infections in the United States. Finally, the previous assessment of national syphilis seroprevalence, from NHANES II (1976–1980),3 involved a nontreponemal-based screening algorithm using 3 different tests from those used in the current assessment, including the automated reagin test, which has not been available as a nontreponemal screening test for years.25,26 Because of this, and the fact that race/ethnicity was categorized differently in NHANES II, we felt we could not accurately evaluate trends in syphilis seroreactivity over time.
In conclusion, we found a low prevalence of lifetime syphilis infection in the general population of US 18- to 49-year-olds. However, consistent with surveillance findings, these nationally representative data underscore the disproportionate impact of syphilis on NH blacks in the United States. Although the racial gap in syphilis infections has narrowed substantially since the 1990s,4 continued efforts to minimize these racial disparities remain a major national priority.2 Future efforts will also need to address the burgeoning epidemic of syphilis among MSM.
1. Centers for Disease Control and Prevention. The National Plan to Eliminate Syphilis from the United States. Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for HIV, STD, and TB Prevention; 1999.
2. Centers for Disease Control and Prevention. Together We Can: The National Plan to Eliminate Syphilis from the United States. Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention; 2006.
3. Hahn RA, Magder LS, Aral SO, et al. Race and the prevalence of syphilis seroreactivity in the United States population: a national sero-epidemiologic study. Am J Public Health 1989; 79:467–470.
4. Centers for Disease Control and Prevention. Sexually Transmitted Disease Surveillance 2005 Supplement, Syphilis Surveillance Report. Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention; 2006.
6. Aberle-Grasse J, Orton SL, Notari E, et al. Predictive value of past and current screening tests for syphilis in blood donors: changing from a rapid plasma reagin test to an automated specific treponemal test for screening. Transfusion 1999; 39:206–211.
7. Reisner BS, Mann LM, Tholcken CA, et al. Use of the Treponema pallidum
-specific Captia Syphilis IgG assay in conjunction with the rapid plasma reagin to test for syphilis. J Clin Microbiol 1997; 35:1141–1143.
8. Silletti RP. Comparison of CAPTIA Syphilis G enzyme immunoassay with rapid plasma reagin test for detection of syphilis. J Clin Microbiol 1995; 33:1829–1831.
10. Pope V, Fears MB. Captia Syphilis-G: an enzyme immunoassay for treponemal antibodies. In: Larsen SA, Pope V, Johnson RE, et al., eds. A Manual of Tests for Syphilis. Washington, DC: American Public Health Association; 1998:332–345.
11. Young H, Moyes A, McMillan A, et al. Screening for treponemal infection by a new enzyme immunoassay. Genitourin Med 1989; 65:72–78.
12. Halling VW, Jones MF, Bestrom JE, et al. Clinical comparison of the Treponema pallidum
CAPTIA Syphilis-G enzyme immunoassay with the fluorescent treponemal antibody absorption immunoglobulin G assay for syphilis testing. J Clin Microbiol 1999; 37:3233–3234.
13. Schmidt BL, Edjlalipour M, Luger A. Comparative evaluation of nine different enzyme-linked immunosorbent assays for determination of antibodies against Treponema pallidum
in patients with primary syphilis. J Clin Microbiol 2000; 38:1279–1282.
15. Portnoy J, Brewer JH, Harris A. Rapid plasma reagin card test for syphilis and other treponematoses. Public Health Rep 1962; 77:645–652.
16. Portnoy J. Modifications of the rapid plasma reagin (RPR) card test for syphilis, for use in large scale testing. Am J Clin Pathol 1963; 40:473–479.
17. Pope V, Fears MB. Serodia Treponema pallidum
passive particle agglutination (TP-PA) test. In: Larsen SA, Pope V, Johnson RE, et al., eds. A Manual of Tests for Syphilis, Supplement. Washington, DC: American Public Health Association; 2000:363–378.
18. Clopper C, Pearson S. The use of confidence or fiducial limits illustrated in the case of the binomial. Biometrika 1934; 26:404–413.
19. Korn E, Graubard B. Confidence intervals for proportions with small expected number of positive counts from survey data. Survey Methodol 1998; 24:193–201.
20. Centers for Disease Control and Prevention. Primary and secondary syphilis among men who have sex with men—New York City, 2001. MMWR Morb Mortal Wkly Rep 2002; 51:853–856.
21. Chen JL, Kodagoda D, Lawrence AM, et al. Rapid public health interventions in response to an outbreak of syphilis in Los Angeles. Sex Transm Dis 2002; 29:277–284.
22. D'Souza G, Lee JH, Paffel JM. Outbreak of syphilis among men who have sex with men in Houston, Texas. Sex Transm Dis 2003; 30:872–873.
23. Heffelfinger JD, Swint EB, Berman SM, et al. Trends in primary and secondary syphilis among men who have sex with men in the United States. Am J Public Health 2007; 97:1076–1083.
24. Pope V. Use of treponemal tests to screen for syphilis. Infect Med 2004; 21:399–404.
25. Simon RD, Peacock AM. Comparison between the automated reagin test and reagin screen test methods of VDRL screening tests for syphilis in use in a routine laboratory. J Clin Pathol 1977; 30:626–629.
26. Stevens RW, Stroebel E. The automated reagin test: results compared with VDRL and FTA-ABS tests. Am J Clin Pathol 1970; 53:32–34.