China has experienced dramatic syphilis resurgence in recent years. According to the Chinese Health Statistical Digest by the Chinese Ministry of Health, the reported incidence of syphilis in the general population increased from 0.54 per 100,000 people in 1995 to 23.07 and 26.86 per 100,000 people in 2009 and 2010, respectively.1 The number of reported syphilis cases ranked third after virus hepatitis and tuberculosis cases and ranked first among sexually transmitted diseases.2 Thus, the Department of Public Health is experiencing increased pressure to diagnose and treat syphilis.
Treponema pallidum (TP), the bacterium that causes syphilis, cannot be cultured. Furthermore, because the natural course of syphilis infection is characterized by periods without clinical manifestations, the serological detection of specific antibodies to TP remains the mainstay of the laboratory diagnosis of syphilis.3,4 A traditional stepwise testing algorithm for the serological diagnosis of syphilis involves a 2-step approach of first screening with nontreponemal tests that measure anticardiolipin antibodies that are produced during active infection, followed by treponemal tests that detect antibodies to TP as a confirmation of positive screening test results.3 However, there are several potential problems associated with nontreponemal tests, such as the prozone phenomenon, biological false positives, and low sensitivity in the detection of very early and latent syphilis.5,6 Recently, the most widely used nontreponemal tests such as the rapid plasma reagin test and the venereal disease research laboratory carbon antigen test were not recommended as initial screening tests in several articles.7–10
Although the US Centers for Disease Control and Prevention (CDC) still recommends screening for syphilis with a nontreponemal test,11 in both North America and Europe, the testing paradigm is shifting toward increased use of treponemal enzyme immunoassays (EIAs) and chemiluminescence immunoassays (CLIAs) for initial syphilis screening, followed by a nontreponemal test (rapid plasma reagin or venereal disease research laboratory) for specimens with positive results.12–16 This emerging method is called as reverse screening algorithm, in contrast to the traditional stepwise screening approach. European syphilis guidelines in 2008 recommended an EIA or a TP particle agglutination (TPPA) assay as an initial screening test.15 If the initial screening test is positive, only then is a different type of treponemal antigen test recommended as a confirmatory assay. If the confirmatory test is positive, a quantitative nontreponemal test is recommended to test for the serological activity of syphilis and to monitor the effect of treatment. The reverse screening algorithm offers several significant advantages including an objective reported values of screening results obtaining from the ratio of specimen signals to the cutoff value, the potential to automate testing, and enhanced sensitivity for the identification of patients with late/latent or early syphilis.17 Nevertheless, the expense of the newer algorithm is apparently higher because of the use of both screening and confirmatory treponemal antigen tests.
The newer CLIA was reported to have a same role as EIAs in screening.18,19 Recently, the CLIA has been incorporated into German national maternity-care guidelines as an alternative to TPPA screening in pregnant women.20 According to the medical regulation of China, syphilis testing is necessary before surgery operation, invasive procedure, blood donation, and transfusion. In addition, benefiting from automated enzyme or CLIA analyzers, approximately two-thirds of grade A tertiary hospitals in Guangzhou select reverse screening algorithm to test numerous blood specimens. Considering the volume of serum samples for syphilis testing per day and labor costs, the ARCHITECT Syphilis TP assay has been adopted as an initial screening test, and TPPA is used as a confirmatory test in our laboratory since 2010. However, to date, no one has explored the value of signal-to-cutoff (S/CO) ratios in ARCHITECT Syphilis TP using CLIA technology, which is characterized by random accession. In the present study, we used receiver operating characteristic (ROC) curve analysis to evaluate the S/CO ratios in ARCHITECT Syphilis TP screening tests, and we propose a cost-effective reverse screening algorithm based on S/CO ratios for our hospital population.
MATERIALS AND METHODS
Patients and Specimens
Sun Yat-sen University Cancer Center was the largest integrated cancer center in southern China for cancer care and prevention. The clinical laboratory was equipped with laboratory information management system and COBAS P612 (Roche, Waiblingen, Germany) for fully automatic aliquoting and sorting of barcoded and centrifuged samples.
Serum samples (n = 8980) sent to our clinical laboratory for routine patient testing for syphilis upon hospital admission, from December 2011 to June 2012, were analyzed for the study. The patients were predominantly cancer patients, of which the serum prevalence of syphilis (refers to a patient with positive antibody to TP by ARCHITECT Syphilis TP, despite the toluidine red unheated serum test [TRUST] status and clinical manifestation) was 3.3%. The median age of the patients was 52 years (range, 1–89 years), and the male/female ratio is 1.49. Ethical approval was obtained from Cancer Center research ethics committee. We obtained informed consent from all patients involved.
The initial screening test (ARCHITECT Syphilis TP), TRUST, and TPPA were performed on the day on which the sample was collected.
Initial Screening by ARCHITECT Syphilis TP
Each sample was first tested by ARCHITECT Syphilis TP. High-throughput ARCHITECT Syphilis TP (Abbott, Denka Seiken Company Limited, Tokyo, Japan) was chosen as a screening test and was run on an ARCHITECT i2000SR (Abbott, Chicago, IL). This test is a 2-step immunoassay for the qualitative detection of antibody to TP (anti-TP) in human serum or plasma using CLIA technology. In the first step, the sample, microparticles coated with recombinant TP antigens (TpN15, TpN17, and TpN47), and diluent are combined. Anti-TP present in the sample then binds to the coated microparticles. In the second step, after washing, an acridinium-labeled antihuman IgG and IgM conjugate is added. After another wash cycle, pretrigger and trigger solutions are added to the reaction mixture. The resultant chemiluminescent reaction is measured by the ARCHITECT immunoassay optical system. The presence or absence of anti-TP in the specimen is determined by comparing the chemiluminescent signal in the reaction to the cutoff signal defined in a previous calibration. If the chemiluminescent signal in the specimen is greater than or equal to the cutoff signal, meaning that the S/CO ratio is 1.0 or greater, the specimen is considered reactive for anti-TP. The immunoassay’s S/CO ratio was recorded directly from the automated ARCHITECT i2000SR.
Confirmation Assays With TPPA and TRUST
Samples reactive in the screening tests were reflexively confirmed by a manual TPPA test, which is the gold standard for assaying anti-TP. The TPPA assay (Fujirebio Incorporation, Tokyo, Japan) is based on the agglutination of colored gelatine particle carriers sensitized with sonicated TP (Nichols Strain) antigen. Considering that the amount of treponemal antibody was proportional to the S/CO ratio in ARCHITECT Syphilis TP assay or the titer in TPPA, we hoped to explore the internal connection between the 2 treponemal antigen tests. To evaluate the titers in TPPA, the vertical 8 wells in the “U”-shaped microplate were used to quantitatively determine antibody titers with additional 4 titers: 1:160, 1:320, 1:640, and 1:1,280. The pattern of agglutination in each well was observed visually. The results of the TPPA assay were reported as either negative or positive with an end point titer.
All samples reactive in the screening tests were next subjected to a TRUST to ascertain the serological activity of syphilis. The TRUST (Rongsheng Biotech Company Limited, Shanghai, China) is a nontreponemal antibody test based on the macroscopic flocculation reaction, which can detect reagins, that uses sized toluidine-red particles coated with lipoidal antigens as the visualization agent. To avoid prozone effects and false-negative results, we tested undiluted serum samples and serial 2-fold dilutions of the serum samples, which were prepared in 0.9% saline. The diluent ratios were 1:2, 1:4, and 1:8. The assays were performed according to the instructions provided by the manufacturers. The results were reported as either negative or positive for an end point titer.
The S/CO ratios without normal distribution were expressed as a median and quartiles. The Mann-Whitney U test was used to compare the S/CO ratios of TPPA-negative and TPPA-positive samples, and the χ2 test was used to compare proportions unless Fisher exact test was indicated due to small counts. The correlation between S/CO ratios and TPPA titers was analyzed by the Spearman test. An ROC analysis was used to study the relationship between the S/CO ratios of the reactive samples and the TPPA confirmatory test results. More specifically, the assay sensitivity was plotted against the false-positive rate (1 − the specificity value). The diagnostic sensitivity, diagnostic specificity, positive predictive value, and optimal cutoff point of the S/CO ratio were identified from analyses of the ROC curves and associated data. A comparison of the area under the curve (AUC) was performed using a 2-tailed P test, which compared the AUC to the diagonal line of no information (AUC, 0.5). In addition, P values less than 0.05 were considered statistically significant. All statistical analyses were performed using SPSS Statistics 17.0 (SPSS, Incorporation, Chicago, IL).
The S/CO Ratios Were Positively Correlated With the TPPA Titers
A total of 8980 serum samples were screened by ARCHITECT Syphilis TP during a consecutive 6-month period, of which 319 samples (3.55%) were reactive. The 25th, 50th (median), and 75th percentile values for the reactive samples’ S/CO ratios were 4.51, 14.87, and 24.42, respectively, and the minimum and maximum values were 1.02 and 41.26. Of the 319 CLIA-reactive samples, 272 (85.3%) were confirmed to be positive by TPPA and 47 (14.7%) were negative. The S/CO ratios of 47 TPPA-negative samples ranged between 1.02 and 9.8, whereas the ratios of the 272 TPPA-positive samples were between 1.21 and 41.26. Moreover, TPPA-positive samples had higher median S/CO ratios than did TPPA-negative samples (19.0 vs. 1.7). Substantial differences were evident in the distribution of the S/CO ratios between the TPPA-positive samples and the TPPA-negative samples (Mann-Whitney U = 875.5, P < 0.001), suggesting a relationship between the S/CO ratio and the TPPA status. We further explored the relationship between the S/CO ratios and the TPPA titers in the 272 TPPA-positive samples (Fig. 1). The S/CO ratios were positively correlated with the serum titers of antibody determined by TPPA (Spearman correlation coefficient: r = 0.809, P < 0.001).
The S/CO Ratio of 9.9 Was the Optimal Cut Point Providing 100% Specificity
Receiver operating characteristic curve analysis was used to determine the optimal S/CO cutoff value for the prediction of positive results of ARCHITECT Syphilis TP (Fig. 2). From the ROC curve and associated data (see Supplemental Table 1, https://links.lww.com/OLQ/A80, which demonstrates the ROC curve data), an S/CO ratio of 9.9 was determined as the optimal cut point that could provide 100% specificity. This S/CO ratio of 9.9 corresponded to a diagnostic sensitivity, diagnostic specificity, and positive predictive value for anti-TP in the serum of 71.3%, 100.0%, and 100.0%, respectively. All samples with screening S/CO ratios above 9.9 (194/194) were confirmed by TPPA testing, compared with only 62.4% (78/125) of the samples with screening S/CO ratios between 1.0 and 9.9 (2-sided Fisher exact test, P < 0.001). The false-positive rate was zero above this cutoff level, indicating that when the S/CO ratio is 9.9 or greater, there is no need for a confirmatory TPPA assay.
TPPA Confirmation Was Necessary for TRUST-Negative Samples With S/CO Ratios Between 1.0 and 9.9
In contrast to the fact that the S/CO ratios of TPPA-negative samples were all between 1.0 and 9.9, the S/CO ratios of the TRUST-negative samples were distributed between 1.0 and 41.26, implying that the S/CO ratios could not forecast the TRUST status.
We found that TRUST-positive samples with S/CO ratios of 1.0 or greater (n = 121) and the TRUST-negative samples with S/CO ratios of 9.9 or greater (n = 87) were all positive for treponemal antibody confirmed by TPPA. Of the 111 TRUST-negative samples with S/CO ratios between 1.0 and 9.9, 64 samples were confirmed to be positive for treponemal antibody by TPPA and 47 samples were confirmed to be negative (Fig. 3).
According to 2008 European guidelines on the management of syphilis,15 in which the screening-reactive samples by one treponemal antigen test should be confirmed by the second treponemal antigen test, all the 319 screening-reactive samples (100%; 319/319) need TPPA confirmation. The US CDC recommends that only the discrepant results (reactive treponemal antigen test, nonreactive nontreponemal antigen test) from the reverse screening algorithm need confirmation with a second treponemal antigen test.21 TRUST-positive samples (n = 121) do not need further confirmation for treponemal antibody according to CDC, as verified by our data. Therefore, a total of 198 patients (62.1%; 198/319) who had discrepant results need TPPA confirmation, as suggested by the US CDC. However, we found 87 TRUST-negative samples with an S/CO ratio of 9.9 or greater were all positive for TPPA by ROC analysis. These samples were unnecessarily further subjected to TPPA confirmation. Therefore, only the remaining TRUST-negative samples (n = 111) with the S/CO ratios between 1.0 and 9.9 needed to be subjected to TPPA confirmation. These samples occupied only 34.8% (111/319) of the screening-reactive samples. In another words, the TPPA confirmation was unnecessary in 65.2% of the screening-reactive patients (Fig. 4). In short, the overall costs on reverse syphilis screening algorithm are the most expensive according to 2008 European guidelines, moderate according to the US CDC, the cheapest in our laboratory when introducing the optimal cut point of the S/CO ratio of 9.9.
Owing to the increasing incidence of syphilis, more and more medical laboratories have begun to use high-throughput treponemal antigen tests in initial syphilis screening. In China, a diagnosis of syphilis may cause greater mental stress for a patient than nonsexually transmitted diseases because of the country’s reserved sexual culture. Therefore, it is very important to eliminate false-positive results. As in the Kaiser Permanente Northern California Regional Laboratory,22 almost every Chinese medical laboratory using reverse screening algorithms performs 2 different treponemal antigen tests and 1 nontreponemal antigen test on the same screening-reactive specimen to minimize the false positivity. The results from all 3 tests are reported simultaneously. In contrast to the traditional screening algorithm, which applies only 1 treponemal antigen test and 1 nontreponemal antigen test, the reverse screening algorithm definitely boosts laboratory costs. One study reported that this cost increase is largely caused by the substantially higher number of confirmatory treponemal tests required for the reverse screening algorithm.23 Thus, our concern should be ensuring 100% specificity while reducing the need for confirmatory testing in reverse screening algorithm.
The ARCHITECT Syphilis TP is run on random-access analyzers, facilitating syphilis screening within 1 hour and with minimal hands-on time, allowing for the rapid diagnosis of syphilis infection using a CLIA. The interpretation of the results is based on the S/CO ratios. Specimens with S/CO values less than 1.0 are considered nonreactive for syphilis antibody, whereas specimens with S/CO values of 1.0 or greater are considered reactive. A recent study by Wong et al.24 demonstrated a strong correlation between the S/CO ratios (an index value) of the TREP-SURE EIA (Trep-Sure; Phoenix Biotech, Mississauga, Ontario, Canada) and TPPA-positive results. In addition, both Yen-Lieberman et al.25 and Loeffelholz et al.26 reported that the strength of the signal (antibody index) of the quantitative BioPlex Syphilis IgG assay (Bio-Rad Laboratories, Hercules, CA) could be used to identify likely false-positive results, thereby reducing the need for confirmatory testing. Yen-Lieberman et al.25 demonstrated that specimens with BioPlex AIs greater than 6.0 were always positive when tested with a supplemental EIA and therefore proposed an algorithm in which only specimens with a BioPlex Syphilis IgG AI less than 6.0 are subjected to a confirmatory EIA.
Based on the present findings, we propose the following syphilis screening algorithm in our laboratory (Fig. 5). Each sample that is reactive in the initial ARCHITECT Syphilis TP screening is further tested by the quantitative TRUST as recommended by the US CDC, if the sample has positive TRUST result or the TRUST-negative sample has an S/CO ratio of 9.9 or greater, the specificity is 100%, so there is no need to perform TPPA confirmation, and a positive anti-TP result is directly reported along with the TRUST status. If the sample has a negative TRUST result, along with an S/CO ratio between 1.0 and 9.9, the next TPPA confirmation is required. If the result is negative for TPPA, the negative anti-TP result is finally reported. If the result is positive for TPPA, the positive anti-TP result is finally reported along with the negative TRUST status.
Compared with ARCHITECT Syphilis TP, TPPA is time-consuming because it is operated by manual and the labor cost is more expensive. In our study, 65.2% of syphilis screening-reactive patients could profit from the improved reverse syphilis screening algorithm because they could not only save the cost of TPPA confirmation but also receive reporter earlier. Furthermore, laboratory technician can save more time because there is no need to test for TPPA in most of the initial screening-reactive samples.
It is particularly difficult to interpret conflicting treponemal test results, such as an initial screening-reactive treponemal test and a negative confirmatory test. For reverse syphilis algorithm, the analytical sensitivity of the confirmatory treponemal test must be at least equivalent to the screening assay to confirm the weak reactive results in initial screening tests. Recently, the TPPA test was confirmed to have higher analytical sensitivity than LIAISON CIA.27 Based on published sensitivity and specificity data, the TPPA test is currently considered to be the most suitable confirmatory treponemal test.28 Most opinions about conflicting treponemal tests tend to interpret the outcomes of the screening treponemal test as likely false positives.11,13,22,28–30
Marangoni et al.30 reported that 129 serum samples scored reactive by ARCHITECT Syphilis TP were analyzed by both T. pallidum hemagglutination test and Western blot and presented negative results. However, they considered that the reasons for the false positivity of ARCHITECT Syphilis TP were still unknown at present. In our study, the seroprevalence of syphilis is 3.35%; the proportion of conflicting treponemal results (reactive ARCHITECT Syphilis TP and negative TPPA) was 14.7% (47/319). Similarly, a study in general population screened with the use of treponemal CIA (LIAISON; DiaSorin Incorporation, Stillwater, MN) showed that the syphilis seroprevalence was 2% (439/21623) and the conflicting treponemal results (reactive LIAISON CIA and negative TPPA) was 16.2%.22 Although the population in our study was predominately composed of cancer patients with underlying inflammatory conditions, their false-positive rate was not higher than a general screening population. Therefore, false-positive result may not be attributed to cancer and inflammation. A study coming from the US CDC showed that the false-positive rate of reverse syphilis algorithm was 2.9 times greater in the low-prevalence population than in the high-prevalence population.31 Therefore, the higher false-positive rate may be related to the low prevalence. The use of a confirmatory test remains a must to avoid false-positive results from CLIA.
The current study possessed several limitations. The samples used in this study were all collected from patients with cancer. In addition, the results may be biased toward a low-risk population. The S/CO cutoff value of 9.9 needs to be further evaluated by different laboratories in populations with varying prevalence of syphilis. Conflicting treponemal results should either be followed up or further confirmed by a third treponemal test.
Despite these limitations, we evaluated the S/CO ratios from ARCHITECT Syphilis TP in initial syphilis screening and determined the ideal cutoff value to identify specimens that did not require an additional treponemal assay for confirmation. We also proposed a potentially cost-effective algorithm for reverse syphilis screening based on the optimal cutoff S/CO ratio for our hospital population. Considering the high efficiency and high specificity, the present improved reverse screening algorithm for TP antibody using S/CO ratios from CLIA is strongly recommended to each clinical laboratory using reverse syphilis algorithm.
1. Ministry of Health of People’s Republic of China. Chinese Health Statistical Digest,1996–2010. Available at: http://www.moh.gov.cn/publicfiles//business/htmlfiles/zwgkzt/ptjty/
. Accessed April 2011.
2. Chinese Center for Disease Control and Prevention. Available at: http://www.chinacdc.cn
3. Larsen SA, Steiner BM, Rudolph AH. Laboratory diagnosis and interpretation of tests for syphilis. Clin Microbiol Rev 1995; 8: 1–21.
4. Singh AE, Romanowski B. Syphilis: Review with emphasis on clinical, epidemiological and some biologic features. Clin Microbiol Rev 1999; 12: 187–209.
5. Calonge N. Screening for syphilis infection: Recommendation statement. Ann Fam Med 2004; 2: 362–365.
6. Peter CR, Thompson MA, Wilson DL. False-positive reactions in the rapid plasma reagin-card, fluorescent treponemal antibody-absorbed, and hemagglutination treponemal syphilis serology tests. J Clin Microbiol 1979; 9: 369–372.
7. Young H. Standard Serologic Testing for Syphilis in Individual Patients: The European View. IUSTI/WHO Conference on STI Europe Syphilis Guideline Expert Workshop, Mykonos, Greece, 2004.
8. Manavi K, Young H, McMillan A. The sensitivity of syphilis assays in detecting different stages of early syphilis. Int J STD AIDS 2006; 17: 768–771.
9. Geusau A, Kittler H, Hein U, et al. Biological false-positive tests comprise a high proportion of venereal disease research laboratory reactions in an analysis of 300,000 sera. Int J STD AIDS 2005; 16: 722–726.
10. Creegan L, Bauer HM, Samuel MC, et al. An evaluation of the relative sensitivities of the venereal disease research laboratory test and the Treponema pallidum
particle agglutination test among patients diagnosed with primary syphilis. Sex Transm Dis 2007; 34: 1016–1018.
11. CDC. 2011. Discordant results from reverse sequence syphilis screening—five laboratories, United States, 2006 –2010. MMWR Morb Mortal Wkly Rep 2011; 60: 133–137.
12. Pohl D, Hotton A, Gratzer B. Discordant Syphilis EIA Test Results: Are Newer Tests Better? Atlanta, GA: Centers for Disease Control and Prevention National STD Prevention Conference, 2010.
13. Peterman T, Schillinger J, Blank S, et al. Syphilis testing algorithms using treponemal tests for initial screening–four laboratories, NewYork City, 2005–2006. MMWR Morb Mortal Wkly Rep 2008; 57: 872–875.
14. Philips-Rodriguez D PD, Schillinger J. Past and Current Syphilis Diagnoses Among Treponema pallidum EIA 1/RPR—Patients With a High Rate of HIV Infection; Findings From Medical Record Review, New York City 2008–2009. Atlanta, GA: Centers for Disease Control and Prevention National STD Prevention Conference, 2010.
15. French P, Gomberg M, Janier M, et al. IUSTI: 2008 European guidelines on the management of syphilis. Int J STD AIDS 2009; 20: 300–309.
16. Pope V. Use of treponemal tests to screen for syphilis. Infect Med 2004; 21: 399–404.
17. Mishra S, Boily MC, Ng V, et al. 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 2011; 38: 190–196.
18. Marangoni A, Sambri V, Accardo S, et al. Evaluation of LIAISON Treponema Screen, a novel recombinant antigen-based chemiluminescence immunoassay for laboratory diagnosis of syphilis. Clin Diagn Lab Immunol 2005; 12: 1231–1234.
19. Hagedorn HJ, Kraminer-Hagedorn A, De Bosschere K, et al. Evaluation of INNO-LIA syphilis assay as a confirmatory test for syphilis. J Clin Microbiol 2002; 40: 973–978.
20. Wellinghausen N, Dietenberger H. Evaluation of two automated chemiluminescence immunoassays, the LIAISON Treponema Screen and the ARCHITECT Syphilis TP, and the Treponema pallidum
particle agglutination test for laboratory diagnosis of syphilis. Clin Chem Lab Med 2011; 49: 1375–1377.
21. 2010 STD Treatment Guidelines - Syphilis - new information regarding the emergence of azithromycin-resistant Treponema pallidum and the criteria for spinal fluid examination to evaluate for neurosyphilis. Available at: http://www.cdc.gov/std/treatment/2010
. Accessed December 16, 2010.
22. Park IU, Chow JM, Bolan G, et al. Screening for syphilis with the treponemal immunoassay: Analysis of discordant serology results and implications for clinical management. J Infect Dis 2011; 204: 1297–1304.
23. Owusu-Edusei K Jr, Koski KA, Ballard RC. The tale of two serologic tests to screen for syphilis–treponemal and nontreponemal: does the order matter? Sex Transm Dis 2011; 38: 448–456.
24. Wong EH, Klausner JD, Caguin-Grygiel G, et al. Evaluation of an IgM/IgG sensitive enzyme immunoassay and the utility of index values for the screening of syphilis infection in a high-risk population. Sex Transm Dis 2011; 38: 528–532.
25. Yen-Lieberman B, Daniel J, Means C, et al. Identification of false-positive syphilis antibody results using a semiquantitative algorithm. Clin Vaccine Immunol 2011; 18: 1038–1040.
26. Loeffelholz MJ, Wen T, Patel JA. Analysis of bioplex syphilis IgG quantitative results in different patient populations. Clin Vaccine Immunol 2011; 18: 2005–2006.
27. Zhang W, Yen-Lieberman B, Means C, et al. The impact of analytical sensitivity on screening algorithms for syphilis. Clin Chem 2012; 58: 1065–1066.
28. Cole MJ, Perry KR, Parry JV. Comparative evaluation of 15 serological assays for the detection of syphilis infection. Eur J Clin Microbiol Infect Dis 2007; 26: 705–713.
29. Woznicova V, Valisova Z. Performance of CAPTIA SelectSyph-G enzyme-linked immunosorbent assay in syphilis testing of a high-risk population: analysis of discordant results. J Clin Microbiol 2007; 45: 1794–1797.
30. Marangoni A, Moroni A, Accardo S, et al. Laboratory diagnosis of syphilis with automated immunoassays. J Clin Lab Anal 2009; 23: 1–6.
31. CDC. Discordant results from reverse sequence syphilis screening—five laboratories, United States, 2006 –2010. Morb Mortal Wkly Rep 2011; 60: 133–137.