This chapter covers a range of bacteria and bacterial diseases in which immunoassays play a role. The bacteria included are: Streptococcus pyogenes, Treponema pallidum, Borrelia burgdorferi, Helicobacter pylori, Shiga-toxin producing Escherichia coli, Legionella, Bartonella henselae, Leptospira, Coxiella, Brucella and Francisella tularensis. For each species, the etiologic agent and pathogenesis are described, followed by the principles involved in diagnosis and the typical assay technology involved. The following diseases are included in the bacteria sections: pharyngitis, Strep throat, necrotizing fasciitis, impetigo, cellulitis, otitis media, mastoiditis, pneumonia, acute rheumatic fever, glomerulonephritis, streptococcal toxic shock syndrome, syphilis, Lyme disease, gastritis, duodenal ulcer, gastric ulcer, hemolytic uremic syndrome, microangiopathic hemolytic anemia, thrombocytopenia, acute renal dysfunction, Legionnaire’s disease, Pontiac fever, cat scratch disease, leptospirosis, Q fever, brucellosis, and tularemia.

Chapter 9.20 of The Immunoassay Handbook [4th ed.]describes a recent strategy to manage laboratory screening for syphilis infection using antibody detection by immunoassay, followed by confirmation with a traditional agglutination test. If a reader of this chapter had just arrived at a serology laboratory, he/she might be surprised to hear that a test with suboptimal sensitivity and specificity would be used to confirm an extremely sensitive and specific test.  This reader might also be surprised to hear that the initial test was fully automated whereas the confirmatory test required manual interpretation and is widely recognized to be imprecise and subjective.

To fully understand how syphilis screening is performed today, the discussion must extend past a consideration of the performance issues regarding the proteins/non-proteins that are measured and fully consider the social and ethical environment that surrounds syphilis testing.  In this short review, these topics will be presented, with particular reference to recent papers from outside the USA that illustrate the multiple approaches used for detecting this challenging sexually transmitted infection.

Before the modern era, a primary syphilis infection would be identified by a characteristic sore (chancre) that was typically seen on the male sex organ.  While lacking in sensitivity (especially for women), this visual diagnosis did offer good specificity, because syphilis was a common infection in many parts of the world, well known to physicians, with a prevalence approaching 5-10% in some countries in the era before antibiotics.  Further, treatments (of dubious values) were available and eagerly sought out by those with evidence of infection.  Today syphilis is less commonly encountered and it is not commonly identified in the primary phase, because the sore marking the bacteria’s entry point into the body may not be visible, for both men and women.

Secondary infection produces symptoms such as rash and muscle pain that are not specific.  Again, due to the low prevalence of syphilis in some cultures, and the lack of medical care in other cultures, a secondary infection may not be diagnosed and treated.

Following secondary syphilis (several months to a year after infection), patients enter a latent (symptomless) stage.  Many years later, 15-30% of infected individuals may develop tertiary syphilis, which can present with a variety of serious health problems, including blindness or brain damage.

Tests which measure the lipoidal products produced by a syphilis infection (non-treponemal tests) have been shown to lack sensitivity for primary syphilis and for latent disease (especially several years post-infection).  Tests that measure antibodies to syphilis antigens (treponemal tests) may not detect some cases of primary syphilis when only IgG antibodies are measured.   Because no test for measuring syphilis IgM antibodies has been cleared by the FDA since 1991, many US labs use a “total antibody test” (detecting both IgM and IgG antibodies) for immunoassay measurements.  These tests are sensitive during primary syphilis. But unlike the non-treponemal tests, they remain positive after treatment.  In the 21st century, using a manual non-treponemal test for screening is a challenge in laboratories that may perform dozens or hundreds of tests daily.  No automated RPR test has been cleared by the FDA, and the interpretation of RPR slides is a challenge for modern technologists, who often have not heard the word “titer” since their training days.  But a non-treponemal screening test will be negative for people who have been treated in the past for syphilis.  So using the traditional algorithm, a confirmed result means that the patient is likely to have active syphilis or recently treated syphilis.

Screening with fully automated treponemal tests would appear to solve the major workflow issue created by RPR screening. When the treponemal (antibody) screen is run first in an environment where prevalence of syphilis is low, the number of positive screen results produced by an antibody test is often 1% or less (a combination of true and false positives), and the number of RPR confirmatory tests will be limited.   This approach to screening/confirmation is known as “the reverse screening algorithm” or reverse sequence syphilis screening” (RSSS).

Using the traditional algorithm (non-treponemal test followed by treponemal test), the number of false positive initial screens will depend upon the presence of conditions that can yield a biological false positive (pregnancy, infection) but will not depend on the amount of seropositivity in the community. A confirmatory treponemal test will help identify specimens that are biological false positives because the conditions that cause such results do not affect the immunological method.

When samples are positive by a treponemal screening test, or by more than one treponemal test, but are then negative by a non-treponemal confirmatory test, the next step may be a tense conversation between doctor and patient.  While the screening result may be wrong, there are no identifiable biological causes and hence no simple way to identify questionable results.

Because RPR tests generally become negative in late syphilis, it would be expected that patients with untreated disease might occasionally be misclassified by RSSS. In one of the first published studies of this algorithm [1], the CDC proposed that “If they have not been previously treated, patients with reactive results from treponemal tests and nonreactive results from nontreponemal tests should be treated for late latent syphilis.” This document also predicted increasing use of RSSS due to increased testing requirements and economic pressures in larger laboratories.

This 2008 guidance presents two challenges to implementation. First, like any serology test, the syphilis immunoassays will have a measurable false positive rate. While that rate may be below 1%, in areas with very low syphilis prevalence, many or most of the positive results may be due to these false positives. A physician may go forward with a treatment that is unnecessary and possibly disturbing to the patient and the patient’s family. Second, determining if the patient has been treated in the past is no longer straightforward in an era when exposure to B-lactam antibiotics, tetracyclines, and macrolides for indications other than syphilis is widespread [2]. There may be no practical way to identify a history “non-observed treated” syphilis.

To further define the number of patients for whom treatment would be appropriate, some experts have proposed a third round of testing, following RSSS [3]. One would expect that a second treponemal test should show concordance with the first treponemal test; the absence of concordance would suggest that there was an initial false positive. But this approach has not been universally accepted because the ability of different treponemal tests to detect late untreated syphilis has not been evaluated for most commercial methods. Also, many different tests are being used as the “third test”, including TP-PA, FTA-ABS, and multi-parameter line immunoassays. For a physician, given discordant syphilis results, treatment with an inexpensive and highly effective drug may be a more straightforward option.

The 2008 CDC report also noted that the recommendations for reverse screening “might not be appropriate in countries with different patterns of seroreactivity, systems of health care, and epidemiology of disease.”

In 2008 the European Guidelines on the Management of Syphilis were updated [4]. A treponemal test was recommended for screening, with an IgM test when primary syphilis was suspected, and the use of a (different) second antigen based test (on second specimen) was proposed for confirmation. An EIA or TPPA could be used, depending on the initial screening method.  IgG immunoblot was recommended as a supplementary confirmatory test when the first two antigen-based tests did not give consistent results.   This 2008 guidance document recommends the use of RPR testing to monitor the success of treatment, but greatly limits the situations where it may be appropriate for screening.

In developing nations, the laboratory confirmation of syphilis infection poses many challenges, due to the lack of sophisticated instrumentation. The 2006 WHO guidance can be consulted for further details [5].

Establishing whether the traditional or reverse screening algorithm will perform better requires 1) the ability to run a very large comparative study and 2) the ability to determine the likelihood of a “missed diagnosis”.  While many laboratories can run many samples, not many laboratories can justify running two methods in parallel for an extended period.  Only a few studies completely meet these ideal criteria. Finally, the cost of “over diagnosis” and “overtreatment” can be roughly quantified; it is much harder to determine the cost of “missed diagnosis” and “under treatment”.

In the 2008 report published by the CDC, there was no information available regarding patient history and treatment.  It was noted that in New York, with an overall seroreactivity of 6% by an immunoassay screen, only 44%  of the samples tested were also RPR positive, meaning that 56% of the patients had results that were inconclusive. When these inconclusive specimens were retested by TP-PA, 98% were positive, suggesting that methodological false positives played only a minor role in the discrepancies.  

In 2010, Maple et al reported from the UK that 12% of hospital specimens that were reactive using two commercial EIA’s were negative when tested by TP-PA [6].   Further testing by a line blot assay was positive for 81% of these samples.   The majority of the discrepant samples were from HIV patients and were unlikely to represent early acute cases. This small study (N=226) was the first to demonstrate that confirmatory immunoassay methods might give different results, although no explanation for these results was offered.

A 2011 report organized by the CDC [7] considered the financial impact of a one-step approach as compared to the two-step algorithm. The authors found that original and the “reverse” algorithm both led to the detection and treatment of the same number of individuals.  However, the two-step testing approach led to treatment of fewer individuals, since there was a low return rate when patients were recalled for treatment.

In 2011 Mishra et al in Canada reported a large comparison involving 1,037,025 blood bank samples run by RSSS, compared to over two million samples analyzed by a traditional screen [8].  Using the traditional approach  0.46% of their samples were confirmed positive for syphilis; with RSSS, the positivity rate increased to 0.60%, whereas 1.64% of their samples EIA positive/RPR negative. It would appear that the number of confirmed active cases increased by 0.14%, but some of these sera may represent post-treatment collections, so without more clinical information the data cannot support increased sensitivity.  

In 2012, Park et al reported a study of 21, 623 patients in California, who were screened in a regional clinical laboratory by immunoassay [9].  58% of the samples positive in the screening test could not be confirmed by RPR; when TPPA was used as a third test, 78%  of this group were found to be positive by TPPA.  This left a group of samples that were screen positive but not confirmed by either of two tests. When the initial screen was then repeated, 23% of these screen positive/RPR negative/TPPA negative samples generated a negative screening result. The authors concluded that samples with this positive/negative/negative profile were likely false positive, in a low-prevalence population. Based on the results of this study and similar studies at other California locations, the CDC summarized results on 140,176 sera and concluded that specimens that were non-reactive in a third test using TPPA were likely to be false positive [3] . 

In the same year, a small comparison of the “traditional” and “reverse” screening algorithm was reported by Binnicker et al at the Mayo Clinic [10]. Among 1000 patients tested, there were six samples that were considered to be false positives by the “reverse” algorithm.  However, two other patients were diagnosed with possible latent syphilis and treated appropriately. 

In 2012, Gratix et al from Alberta Canada described their experience using RSSS with 96,509 sera. Results were compared to the testing performed by the traditional algorithm in previous years [11].  Three cases of RPR negative primary syphilis were identified, and the detection of late latent syphilis (LLS) doubled; as a result 58% of the syphilis cases detected were LLS.  The majority of the patients who received this diagnosis were born outside of Canada, with a median of 5 years since immigration to Canada.  The authors pointed out that the some of the LLS patients might suffer from a non-venereal treponemal disease or they may have been unable to recall previous treatment.  However these two factors should be the same during the previous “traditional algorithm” period and would not account for the increased prevalence of LLS.

A small study was presented in 2012 in a letter to the Journal of Clinical Microbiology from Lipinsky et al. in Israel [12]. The authors analyzed data from 12,235 patients tested by both immunoassay and RPR.  Results that were positive by either method or both were reflexed to the TP-PA test.  While 99% of specimens positive by both methods were TP-PA positive, only 59% of 334 specimens positive by immunoassay only could be confirmed by TP-PA and none of 65 specimens positive by RPR only could be confirmed by TP-PA.  The authors concluded that RSSS was effective because no syphilis cases were missed, and the throughput was 10-fold higher.

In 2013 Baläo et al reported their experience in Brazil [13].  Results for 76,350 blood donor candidates tested by RSSS were compared to results previously obtained using the traditional algorithm using VDRL and FTA-ABS confirmation.  For the reverse algorithm, samples were screened by immunoassay and confirmed by VDRL and FTA-ABS.  In this setting, the reverse algorithm led to a significant reduction of false positives, but led to a large increase in the number of positive results, as well as a small number of indeterminate results.  The authors favored the use of RSSS due to “its capacity to detect different stages of the syphilis infection, include the cases unlikely to be diagnosed by nonspecific tests”.

A report from Lee et al (2013) in Korea describes the use of TPPA, FTA-ABS and a multi-parameter line immunoassay to further characterize discordant results from 15,713 samples analyzed by RSSS [14]. They demonstrated that FTA-ABS and line immunoassays added little if any additional information after TP-PA. However the sera were not clinically characterized so the behavior of LSS samples could not be determined.

A 2013 report from Hunter et al in Australia describes their experience with RSSS on 28,261 clinical specimens in Australia [15].  The patients were clinically referred, which may explain why the EIA positivity rate was 4.1%.  Of these 133 samples, 11.3% could not be confirmed by RPR.  Most subjects in this group were from high prevalence populations. Further testing was performed by TPPA and FTA; for the RPR negative samples the TPPA and FTA results demonstrated 94% agreement. Patients were not clinically characterized so the behavior of LSS samples could not be determined.

In 2014 Tong et al from China described their experience with a series of 24,124 subjects received in a hospital laboratory [16].  They performed RPR, TP-PA and chemiluminescence immunoassay on each sample and subsequently determined the outcomes of the traditional analysis, the RSSS, and the 2008 European guideline recommendation (immunoassay followed by TP-PA) (see reference #). All patients were clinically classified based on a combination of serodiagnosis and personal history.  The authors reported that 2749 patients in this cohort were diagnosed with syphilis (11.1%) and among these patients, 665 had not been previously diagnosed.  While only 6 patients were classified as suffering from primary syphilis and 3 with secondary syphilis, 43 were diagnosed as early latent cases, 425 as late latent cases, and 188 as tertiary cases.  For this cohort, the sensitivity of the traditional testing algorithm was only 76%, compared to greater than 99% for RSSS and the European algorithm. All three algorithms had specificity greater than 99.9%, which is not surprising since the serology results were used to make the clinical classification. This recent report differs from other published work because of the high disease prevalence and the high percentage of samples with advanced disease, which may reflect lack of medical care and lack of access to antibiotics (used to treat other conditions) that are effective against syphilis. 

The CDC’s 2008 discussion of RSSS pointed out that “the best algorithm would depend on the population studied” [1].  In a patient cohort with little syphilis positivity, there is little difference between the traditional algorithm and RSSS.  While the reagent costs of an immunoassay are generally higher, the semi-automated ELISA methods used in the previous decade have been largely replaced by fully automated systems, where there is a labor savings that offsets the difference in reagent costs. 

Published studies have reported results for patients run in a hospital setting (often samples from clinics) as well as from blood banks. In the hospital setting  where there can be a higher rate of syphilis positivity (4% or more), or in a setting in a developing country where access to healthcare thru community facilities like STD clinics is limited, the likelihood of encountering late latent syphilis (LSS) increases. The detection of undiagnosed LSS in the studies described in this review ranged from 0.1% of all cases in a US study to 1.9% in a Chinese study. While a blood bank can do a comparison of very large sample sets, these institutions are unlikely to have clinical data or personal history for those subjects, so these reports cannot determine whether there is undiagnosed latent syphilis in the studied population.

These observations summarized here offer an interesting perspective on the meaning of discrepant results from the reverse algorithm.  It appears that in practice one should consider any patient with a confirmed positive immunoassay screen and a discrepant RPR result as a possible case of latent syphilis.  Further, one must consider the possibility of an undiagnosed syphilis infection that has been successfully (but unintentionally) treated. Determining the need for additional treatment will remain a decision to be made by the physician, on a case by case basis.

1. Peterson, T., Schillinger, J., Blank, S., Berman, S., Ballard, R., Cox, D., Johnson, R., Hariri, S., Selvam, N. "Syphilis testing algorithms using treponemal tests for initial screening--four laboratories, New York City, 2005-2006." MMWR Weekly 57 (32), 872-875 (2008).
2. Oboho, I.K., Ghanem, G.K. "Blissful ignorance when managing pregnant women with syphilis and nonreactive nontreponemal tests?" Sex Transm. Dis. 40 (4), 316-317 (2013).
3. Radof, J.D., Bolan, G., Park, I.U., Chow, J.M., Schillinger, J.A., Pathela, P., Blank, S., Zanto, S.N., Hoover, K.W., Workowski, K.A., Cox, D.L., Ballard, R.C. "Discordant results from reverse sequence syphilis screening--five laboratories, United States, 2006-2010." MMWR Weekly 60 (5), 133-137 (2011).
4. French, P., Gomberg, M., Janier, M., Schmidt, B., van Voorst Vader, P., Young, H. "IUSTI: 2008 European Guidelines on the Management of Syphilis." Int. J. STD AIDS 20 (5), 300-309 (2009).
5. WHO. The Use of Rapid Syphilis Tests. WHO/TDR: 1-17 (2006).
6. Maple, P.A.C., Ratcliffe, D., Smit, E. "Characterization of particle agglutination assay-negative sera following screening by treponemal total antibody enzyme immunoassays." Clin. Vaccine Immunol. 17 (11), 1718-1722 (2010).
7. Owusu-Edusei, Koski, K.A., Ballard, R.C. "The tale of two serologic tests to screen for syphilis - treponemal and nontreponemal: does the order matter?" Sex Transm. Dis. 38 (5), 448-456 (2011).
8. Mishra, S., Boily, M-C., Ng, V., Gold, W.L., Okura, T., Shaw, M., Mazzulli, T., Fisman, D.N. "The laboratory impact of changing syphilis screening from the rapid-plasma reagin to a treponemal enzyme immunoassay: a case-study from the Greater Toronto Area." Sex Transm. Dis. 38 (3), 190-196 (2011).
9. Park, I.U., Chow, J.M., Bolan, G., Stanley, M., Shieh, J., Schapiro, J.M. "Screening for syphilis with the treponemal immunoassay: analysis of discordant serology results and implications for clinical management." J. Infect. Dis. 204 (9), 1297-1304 (2011).
10. Binnicker, M.J., Jeperson, D.J., Rollins, L.O. "Direct comparison of the traditional and reverse syphilis screening algorithms in a population with a low prevalence of syphilis." J. Clin. Microbiol. 50 (1), 148-150 (2012).
11. Gratrix, J., Plitt, S., Lee, B.E., Ferron, L., Anderson, B., Verity, B., Prasad, E., Bunyan, R., Zahariadis, G., Singh, A.E. "Impact of reverse sequence syphilis screening on new diagnoses of late latent syphilis in Edmonton, Canada." Sex Transm. Dis. 39 (7), 528-530 (2012)
12. Lipinsky, D., Schreiber, L., Kopel, V., Shainberg, B. "Validation of reverse sequence screening for syphilis." J. Clin. Microbiol. 50 (4), 1501 (2012).
13. Baião, A.M., Kupek, E., Petry, A. "Reverse algorithm for syphilis screening more than halved false positive test results in Brazilian blood donors." Transf. Med. 24 (1), 64-66 (2014).
14. Lee, K., Park, H., Roh, E.Y., Shin, S., Park, K.U., Park, M.H., Song, E.Y. Characterization of sera with discordant results from reverse sequence screening for syphilis. Biomed. Res. Int.  269347 (2013).
15. Hunter, M.G., Robertson, P.W., Post, J.J. "Significance of isolated reactive treponemal chemiluminescence immunoassay results." J. Infect. Dis. 207 (9), 1416-1423 (2013).
16. Tong, M-L., Lin, L-R., Liu, L-L., Zhang, H-L., Huang, S-J., Chen, Y-Y., Guo, X-J., Xi, Y., Liu, L., Chen, F-Y., Zhang, Y-F., Zhang, Q., Yang, T-C. "Analysis of 3 algorithms for syphilis serodiagnosis and implications for clinical management." Clin. Infect. Dis. 58 (8), 1116-1124 (2014).

Carey-Ann Burnham is an Assistant Professor of Pathology & Immunology at Washington University School of Medicine in St. Louis and the Medical Director of Clinical Microbiology for Barnes Jewish Hospital, St. Louis, MO. She is the program director for the CPEP Fellowship in Clinical Microbiology at Washington University. Her research interests are improved methods for diagnostic microbiology, the transmission and epidemiology of community-associated Staphylococcus aureus, and Clostridium difficile. She is on the editorial board for the Journal of Clinical Microbiology and is the 2013 recipient of the American Society for Microbiology Siemens Healthcare Diagnostics Young Investigator Award.

Christopher Doern, PhD D(ABMM) is a board certified medical microbiologist who directs the clinical microbiology laboratory at Children's Medical Center of Dallas. He is an assistant professor of pathology at the University of Texas Southwestern Medical Center Dallas. Prior to taking his position as medical director of clinical microbiology, Dr. Doern completed a fellowship in medical and public health microbiology at the Washington University School of Medicine in St. Louis.

Steven Binder is the Senior Director of Technical Development for the Clinical Diagnostic Group at Bio-Rad Laboratories. He joined Bio-Rad in 1983, where he initially developed and introduced clinical chromatography methods used for catecholamine measurement and hemoglobinopathy screening. He served Bio-Rad as R&D manager for clinical chromatography from 1988 to 1998 and was closely involved in the development of new methods for diabetes monitoring and clinical toxicology. From 1998-2005, he led the development of a fully automated platform for multiplex immunoassay, with an emphasis on autoimmune and infectious diseases. His current work involves the evaluation of novel multiplex and digital technologies, as well as validation and commercial development of new biomarkers.
Steve received a B.A. in History and Science, magna cum laude, from Harvard University. He has authored over 25 papers in peer-reviewed journals and has received 11 US patents.

Rapid plasma reagin test, chancre, microscopic agglutination, pharyngitis, Strep throat, necrotizing fasciitis, impetigo, cellulitis, otitis media, mastoiditis, pneumonia, acute rheumatic fever, glomerulonephritis, streptococcal toxic shock syndrome, syphilis, Lyme disease, gastritis, duodenal ulcer, gastric ulcer, hemolytic uremic syndrome, microangiopathic hemolytic anemia, thrombocytopenia, acute renal dysfunction, Legionnaire’s disease, Pontiac fever, cat scratch disease, leptospirosis, Q fever, brucellosis, tularemia, Streptococcus pyogenes, Treponema pallidum, Borrelia burgdorferi, Helicobacter pylori, Shiga-toxin producing Escherichia coli, Legionella, Bartonella henselae, Leptospira, Coxiella, Brucella, Francisella tularensis.