E D I T O R I A L C O M M E N TA R Y

Serodiagnosis of Syphilis in the Recombinant Era: Reversal of

Fortune
Karen W. Hoover1 and Justin D. Radolf2
1Divisionof Sexually Transmitted Disease Prevention, National Center for HIV/AIDS, Viral Hepatitis, Sexually Transmitted Disease, and Tuberculosis Prevention, Centers for
Disease Control and Prevention, Atlanta, Georgia; and 2Departments of Medicine, Pediatrics, Immunology, and Genetics and Developmental Biology, University of
Connecticut Health Center, Farmington

(See the article by Park et al, on pages 1297304.)

Treponema pallidum, the spirochete that various medical conditions unrelated to chemiluminescence immunoassays (EIAs
causes syphilis, cannot be cultured. As syphilis; nontreponemal test reactivity, and CIAs, respectively) have become
a result, syphilis is usually diagnosed by therefore, must be confirmed by trepo- commercially available [1]. In addition to
tracking the immunologic footprints nemal testing. Conventional treponemal their analytic sensitivity, these assays have
of its etiologic agent. Serodiagnosis of tests use whole organisms and may be the additional advantages of being auto-
syphilis requires the detection of 2 dis- falsely reactive because of cross-reacting matable and generating a spectrophoto-
tinct types of antibodies, nontreponemal serum antibodies that in most cases are metric or luminescent signal that can be
and treponemal [1]. Nontreponemal an- thought to derive from commensal mi- stored electronically. To reduce the time
tibodies, measured by the reactive rapid croorganisms [2]. In addition, a reactive and labor required for syphilis screening,
plasma reagin (RPR) and Venereal Dis- treponemal test cannot distinguish active many laboratories have adopted reverse-
ease Research Laboratory (VDRL) tests, from inactive infection. Traditionally, sequence screening in which sera are
are directed against lipoidal antigens of serodiagnosis of syphilis has been per- tested first by a treponemal EIA/CIA fol-
the host and probably the organism; they formed using an algorithm in which sera lowed by reflexive nontreponemal testing
rise during active infection and often are screened for nontreponemal anti- of specimens with reactivity above a de-
decline following treatment. Their pri- bodies and reactivity is confirmed by fined cutoff [5].
mary usefulness is as a biomarker of testing for treponemal antibodies [1, 3]. Using the reverse-sequence algorithm,
disease activity. Treponemal antibodies, The traditional sequence, long recom- discordant (ie, EIA/CIA-reactive, RPR-
detected by the fluorescent treponemal mended by the Centers for Disease nonreactive) results would be expected
antibody absorbed (FTA-ABS) and Control and Prevention (CDC) [3], has in patients with early primary, latent, or
Treponema. pallidum particle agglutina- performed well in identifying syphilis treated syphilis, many of whom do not
tion (TP-PA) tests, are directed against T. patients with active disease and who are have nontreponemal antibodies. How-
pallidum proteins; they rise early in the most infectious. Along with serologic test ever, 2 recent studies published in Mor-
course of infection and usually remain results, a patients clinical history and bidity and Mortality Weekly Reports
detectable for life, even after successful physical examination are used to confirm (MMWR) provide strong evidence that
treatment. Neither test should be used the diagnosis and guide management [3]. EIAs/CIAs used as screening tests have
alone. Biologic false positive non- The advent of recombinant DNA additional unforeseen performance
treponemal tests are associated with technology in the early 1980s fostered problems [6, 7]. A 2008 study that as-
optimism that serodiagnostic assays using sessed reverse-sequence testing in 4 New
Received 20 July 2011; accepted 20 July 2011. cloned T. pallidum antigens would over- York City laboratories found that 56%
Correspondence: Karen W. Hoover, MD, MPH, Centers for come the shortcomings of conventional of 6548 EIA-reactive serum specimens
Disease Control and Prevention, 1600 Clifton Rd NE, MS
E80, Atlanta, GA 30333 ([email protected]). treponemal tests [4]. The antigens ulti- were discordant [7]. Approximately 17%
The Journal of Infectious Diseases 2011;204:12956 mately selected for diagnostic use were of the discordant sera that underwent
The Author 2011. Published by Oxford University Press on thought to be specific for T. pallidum and confirmatory treponemal testing with
behalf of the Infectious Diseases Society of America. All
rights reserved. For Permissions, please e-mail: journals. often were formatted as solid-phase im- either a TP-PA or FTA-ABS test were
[email protected] munoassays, a newly available platform. nonreactive, suggesting that the EIA re-
0022-1899 (print)/1537-6613 (online)/2011/2049-0001$14.00
DOI: 10.1093/infdis/jir528 Over the years, a number of enzyme and sults were false-positives. This study also

EDITORIAL COMMENTARY d JID 2011:204 (1 November) d 1295

made apparent the burden created for reverse-sequence testing. In this study for patient management. Although the
health departments and clinicians who of 255 discordant sera (CIA1/RPR-) in CDC continues to recommend the tradi-
must assess large numbers of patients a low-prevalence population screened tional algorithm [3, 6], it recognizes that
who would not have been identified us- for syphilis, patients whose sera had re- reverse-sequence testing is here to stay.
ing the traditional sequence. A follow-up active TP-PA results were more likely
study published in MMWR earlier this to have high-risk behaviors, a history Notes
year assessed reverse-sequence testing in 6 of treated syphilis, or higher optical Financial Support. This work was supported
laboratories in California, Illinois, and density cutoff index (ODI) values. The by the Centers for Disease Control and Prevention
(K. W. H.); and the National Institutes of Health
New York and included populations at correlation between ODI value and TP-PA
(grant AI-26756 to J. D. R.)
low and high risk for syphilis [6]. The test reactivity suggests that reporting ODI
Potential conflicts of interest. All authors:
57% rate of discordance among 4834 values might provide useful information No reported conflicts.
EIA/CIA-reactive serum specimens in to guide clinical management and war- All authors have submitted the ICMJE Form
this analysis was similar to that in the rants further study. Another key finding of for Disclosure of Potential Conflicts of Interest.
Conflicts that the editors consider relevant to the
2008 MMWR. Of importance, the rate of the Park et al study is that 28% (7/25) of
content of the manuscript have been disclosed.
unconfirmed EIA/CIA reactivity was untreated patients with CIA1/RPR-/
higher in the low-risk population than in TP-PA- sera had nonreactive CIAs within References
the high-risk population (41% vs 14%, 12 months of the initial test. Their data 1. Radolf JD, Pillay A, Cox DL. Treponema and
respectively), further suggesting problems add to the growing body of evidence Brachyspira, human hostassociated
with specificity when EIAs/CIAs are used suggesting caution when interpreting dis- spirochetes. In: Versalovic J, Carroll KC, Funke
G, Jorgensen JH, Landry ML, Warnock DW,
for initial screening. Although neither cordant syphilis serologies, particularly in
eds. Manual of clinical microbiology. 10th ed.
MMWR report endorsed reverse-sequence low-risk persons, and underscore the need Washington, DC: ASM Press, 2011:94163.
testing, they made 2 recommendations for confirmatory TP-PA testing. 2. Miller JN. Value and limitations of non-
regarding its use: (1) EIA/CIA-reactive The study by Park and colleagues serves treponemal and treponemal tests in the
laboratory diagnosis of syphilis. Clin Obstet
sera should undergo reflex testing with as a valuable reminder that the problem of Gynecol 1975; 18:191203.
a nontreponemal test for confirmation discordance is not just one of analytic 3. Workowski KA, Berman S. Sexually trans-
and identification of active disease, and sensitivity, but also of specificity [10]. In mitted diseases treatment guidelines, 2010.
MMWR Recomm Rep 2010; 59:1110.
(2) discordant sera should be tested re- the absence of a gold standard serodiag-
4. Radolf JD, Lernhardt EB, Fehniger TE,
flexively with a conventional treponemal nostic test, clinical and demographic Lovett MA. Serodiagnosis of syphilis by
test to confirm EIA/CIA reactivity. The information is essential in interpreting enzyme-linked immunosorbent assay with
purified recombinant Treponema pallidum
more recent MMWR report recom- syphilis serologies and in comparing in-
antigen 4D. J Infect Dis 1986; 153:10237.
mended that the TP-PA test be used ex- dividual assays. Along with the MMWR 5. Sena AC, White BL, Sparling PF. Novel
clusively as the confirmatory treponemal reports [6, 7], the Park et al study further Treponema pallidum serologic tests: A par-
test because the FTA-ABS is less sensitive emphasizes the need for a research agen- adigm shift in syphilis screening for the 21st
century. Clin Infect Dis 2010; 51:7008.
and specific, is inherently subjective, and da to clarify the utility and caveats of 6. Centers for Disease Control and Prevention.
requires expensive instrumentation [8, 9]. reverse-sequence testing. The source of Discordant results from reverse sequence
It is ironic that the new-generation false-positive EIA/CIA results must be syphilis screeningfive laboratories, United
States, 20062010. MMWR Morb Mortal
serodiagnostic assays resulting from re- determined, with characterization of anti- Wkly Rep 2011; 60:1337.
combinant DNA technology require con- gen-binding patterns in discordant sera 7. Centers for Disease Control and Prevention.
firmation by the conventional treponemal that are unconfirmed with TP-PA testing. Syphilis testing algorithms using treponemal
tests for initial screeningfour laboratories,
tests they were developed to supplant. The performance of commercially avail-
New York City, 20052006. MMWR Morb
Consistent with the 2010 Sexually Trans- able treponemal tests should be compared Mortal Wkly Rep 2008; 57:8725.
mitted Disease Treatment Guidelines, the head-to-head using sera from patients 8. Cole MJ, Perry KR, Parry JV. Comparative
MMWR authors concluded that patients whose risk behaviors, clinical histories, evaluation of 15 serological assays for the
detection of syphilis infection. Eur J Clin
with EIA/CIA-reactive, RPR-nonreactive, and outcomes are known; in these studies, Microbiol Infect Dis 2007; 26:70513.
and TP-PA nonreactive sera are unlikely it is essential to include sera from 9. Young H, Aktas G, Moyes A. Enzywell re-
to have syphilis and that further evalua- pregnant and human immunodeficiency combinant enzyme immunoassay for the
serological diagnosis of syphilis. Int J STD
tion or treatment is not indicated unless virusinfected patients. In addition, an
AIDS 2000; 11:28891.
primary syphilis is suspected [3, 6]. assessment of long-term outcomes of 10. Park IU, Chow JM, Bolan G, Stanley M,
A report by Park and colleagues [10] untreated individuals with discordant Shieh J, Schapiro JM. Screening for syphilis
with the treponemal immunoassay: analysis
in this issue of the Journal provides serologies and nonreactive confirmatory
of discordant serolog results and implications
additional insight into the interpretation treponemal tests will provide data to for clinical management. J Infect Dis 2011;
of discordant serologies identified with support evidence-based recommendations 204:1297304.

1296 d JID 2011:204 (1 November) d EDITORIAL COMMENTARY