Background: The control of syphilis depends on screening of the population at risk and is usually performed using the Treponema pallidum particle agglutination test (TPPA). Outside Europe the rapid plasma reagin test (RPR) or venereal disease research laboratory test is most often used for screening purposes. Because of the drawbacks in current diagnostic procedures, ie, long turnaround time, the need is felt for a rapid and simple test that can potentially be performed on whole blood.
Objective and study design: In this study a one-step immunochromatographic test (Biorapid Syphilis) and two ELISA, the Bioelisa Syphilis 3.0 and ETI-Treponema Plus, were evaluated.
Methods: Serum samples were collected between February 2000 and May 2006 at the University Hospital in Maastricht, The Netherlands. 145 TPPA-positive sera, confirmed by fluorescent treponemal antibody absorption (FTA-Abs, treponemal test) and/or RPR (non-treponemal) were included. Furthermore, 41 sera from healthy controls and 144 TPPA-negative sera from controls with underlying conditions that might interfere with T pallidum serology were collected.
Results: The sensitivity and specificity of the Biorapid Syphilis, Bioelisa Syphilis 3.0 and ETI-Treponema Plus were 92% and 79%, 100% and 100% and 100% and 100%, respectively, with our selected sera.
Conclusions: The performance of both ELISA was excellent in our study and is favoured over the TPPA because of its ability to be run on an automated system. The sensitivity and specificity of the Biorapid Syphilis were considered too low to implement the test in a hospital laboratory in a developed country but it might be useful in primary healthcare settings in developing countries.
Statistics from Altmetric.com
Syphilis, a sexually transmitted disease caused by the spirochete Treponema pallidum, constitutes a major public health problem. According to the World Health Organisation estimate of 1999, approximately 12 million new cases of syphilis occur worldwide every year, with a wide variation in prevalence between countries.1 2 In Europe the incidence of syphilis infections declined in the early 1990s as a result of the global health campaigns related to the HIV pandemic. By the end of the 1990s, however, the incidence had started to rise again because of an increase in unsafe sex.3 In developing countries, in sub-Saharan Africa for example, syphilis in pregnant women is of particular concern because congenital syphilis causes 26% of all stillbirths and 11% of neonatal deaths.4 In The Netherlands, most of the new syphilis cases were attributed to men having unsafe sex with men, with an increase of 340% between 2000 and 2004.5
Serological testing for treponemal antibodies is the cornerstone for the diagnosis and control of syphilis. In general, the diagnosis of syphilis can be made by different types of diagnostics: (1) direct microscopic examination, which has many logistical disadvantages; (2) non-treponemal serological tests, which lack sensitivity in some stages of syphilis and specificity; (3) treponemal serological tests, which are highly sensitive and specific and can remain positive for life and (4) PCR, under evaluation for T pallidum.6 In The Netherlands, the T pallidum particle agglutination test (TPPA; also known as TPHA: T pallidum haemagglutinin agglutination test), a treponemal serological test, is most frequently used for screening purposes. Positive screening results are confirmed by fluorescent treponemal antibody absorption (FTA-Abs; a treponemal serological test) and the rapid plasma reagin test (RPR; a non-treponemal serological test). The TPHA screening test is a manually performed test with a turnaround time of two hours and requires laboratory facilities. Its performance, including interpretation, depends on the skills and experience of the laboratory technician. The sensitivity and specificity of the TPHA are 76–100% and 98–100%, respectively, depending on disease stage.6 Considering the drawbacks of a long turnaround time, the need for human serum instead of whole blood and the availability of a laboratory with experienced personnel, new tests without one or more of these drawbacks in current diagnostic procedures are needed.
Several rapid syphilis tests have been developed that may enhance active case finding, provided the test is easy to perform, robust and affordable. Used alone, they would be unable to distinguish active from non-active disease because only antibodies against T pallidum are detected, similar to the TPPA. Nevertheless, these tests may prove to be an effective tool in the control of syphilis in difficult-to-reach risk groups in field settings where they can facilitate a crucial intervention. Several rapid tests have been evaluated, resulting in sensitivities ranging between 40% and 100% in different settings.7–12 An overview of the evaluations of various ELISA-based detection methods for T pallidum antibodies is given in table 1. In a recent article by Cole et al,13 10 different ELISA were tested on a panel of 114 serum and plasma samples from syphilis patients and 249 samples from blood donors. No significant differences in sensitivity or specificity were detected between these ELISA. To process large numbers of samples, ELISA-based methods are still preferred to rapid tests.
The aim of this study was to assess the validity and reproducibility of a one-step immunochromatographic rapid test, the Biorapid Syphilis (Biokit SA, Barcelona, Spain) and two commercial ELISA, the Bioelisa Syphilis 3.0 (Biokit) and the ETI-Treponema Plus (DiaSorin SpA, Saluggia, Italy), which can be run on an automated system. All three tests detect specific T pallidum antibodies. To the best of our knowledge, an evaluation of the Biorapid Syphilis and current versions of both ELISA has not been published so far. Furthermore, we have included a broad range of sera from patients with underlying pathological and physiological conditions, which are known to interfere with syphilis serology.14–20
MATERIALS AND METHODS
All TPPA-positive samples collected between February 2000 and May 2006 in the laboratory of Medical Microbiology at the University Hospital in Maastricht, The Netherlands, were selected for this study. The gold standard for a “true TPPA-positive sample” was defined as a sample with a TPPA titre of 1 : 80 or above, confirmed by a positive FTA-Abs and/or a positive RPR and/or a positive immunoblot at a reference centre (National Institute of Public Health and the Environment, Bilthoven, The Netherlands). Ultimately 145 true TPPA-positive serum samples were included, with no more than one sample per patient. RPR was positive in 80 of these samples (55%), ranging from 1 : 1 to 1 : 256; all other samples were negative. A positive RPR result, especially values above 1 : 8, suggests active disease. The FTA-Abs was positive in 141 of the 145 TPPA-positive samples; two samples were negative and two had a borderline result. The immunoblot was performed and found positive in all four of these samples. The exact clinical status and possible drug usage of the patients was not known because most samples were collected at an anonymous venereal disease clinic. We know from clinical experience that in the TPPA-positive population approximately 70% is in the latent phase; primary and tertiary syphilis are rare and occasionally a rash is seen (secondary syphilis; data not published). In our study the control group consisted of 41 healthy TPPA-negative controls and 144 TPPA-negative controls with known underlying conditions that may interfere with T pallidum serology: pregnancy (n = 21); high antistreptodornase titre (n = 10); high anti-cardiolipin antibodies (n = 10); systemic lupus erythematosus (n = 10); diabetes mellitus (n = 7); rheumatoid factor IgM positive (n = 10); leptospirosis (n = 10); borreliosis (n = 10); recent Epstein–Barr virus infection (n = 10); recent cytomegalovirus infection (n = 10); current hepatitis B virus infection (n = 8); current hepatitis C virus infection (n = 9) or current HIV infection (n = 19). Samples were unlinked from any possible patient identifiers and kept at −20°C until evaluation.
T pallidum particle agglutination test, RPR and FTA-Abs
The TPPA (MHA-TP; Fujirebio, Tokyo, Japan) utilises gelatin particle carriers sensitised with purified pathogenic T pallidum (Nichols strain), which agglutinate with antibodies against T pallidum, if present, in serum. The RPR (Syfacard-R*; Abbott Murex, Dartford, UK) utilises tissue lipid cardiolipin (antigen) to detect “reagin”, ie, antibodies, directed against tissue components, which appear in serum as a reaction to tissue damage caused by T pallidum. Finally, in the FTA-Abs (Trepo Spt IF; BioMerieux SA, Marcy l’Etoile, France), human anti-T pallidum immunoglobulins bind to the T pallidum on the slide, which in turn binds fluorescein-labelled goat anti-human immunoglobulins, which can be seen with a fluorescence microscope. All tests were performed according to the manufacturers’ instructions.
One-step immunochromatographic test
The one-step immunochromatographic test Biorapid Syphilis (Biokit) combines anti-human immunoglobulin dyed conjugate and p15 and p17 T pallidum recombinant antigens to detect anti-T pallidum IgG, IgA and IgM antibodies in plasma, serum or whole blood. Briefly, after inserting 25 μl human serum in zone A, the T pallidum-specific antibodies, if present, will bind to the anti-human immunoglobulin dyed conjugate to form an antigen–antibody complex. This complex will fix the recombinant protein on the band in zone B (test zone). When no human T pallidum antibodies are present, no antigen–antibody complexes will form and no fixation will take place in zone B. The superfluous conjugate will flow to zone C and bind to the reagents (control zone). The test result is read visually. All reactions in zone B should be considered positive, according to the manufacturer. The test is invalid without a reaction in zone C.
Enzyme-linked immunosorbent assays
Two ELISA were evaluated in this study: the Bioelisa Syphilis 3.0 (Biokit) and the ETI-Treponema Plus (DiaSorin). Both ELISA are able to detect IgG and IgM antibodies, separately or together, against T pallidum in serum or plasma. In both assays, human antibodies against T pallidum, if present, will bind to p15, p17 and p47 T pallidum recombinant antigens in the microtitre plate coating. Next, the conjugate containing the enzyme peroxidase is added, which will bind to the T pallidum recombinant antigen–human antibody complex. Subsequently, the enzyme substrate, chromogen and after incubation sulphuric (stop) reagents are added. While performing the ETI-Treponema Plus the chromogen can be inserted into the automated system at the beginning of the assay, whereas for the Bioelisa Syphilis 3.0 it should be prepared and inserted 5–10 minutes before usage. The absorbance value for the sample tested was divided by the mean absorbance value of the low positive control (cut-off), if this ratio equalled or was above 1.0, the result was considered positive in both ELISA. A result below 0.9 was considered negative. Both ELISA were performed according to the manufacturers’ instructions, using an automated system (DSX automated system; Dynex Technologies, Inc, USA).
The laboratory characteristics, reported by the manufacturers, of the TPPA and the evaluated tests, are compared in table 2.
A true positive result in either the immunochromatographic test or one of both ELISA was defined as a positive result in a true TPPA-positive sample. A false-positive result was defined as a positive result in a TPPA-negative sample and a false-negative result as a negative result in a true TPPA-positive sample. A true negative result was defined as a negative result in a TPPA-negative sample. The sensitivity of a test was defined as the number of true positives divided by the number of true positives plus false negatives. Specificity was defined as the number of true negatives divided by the number of true negatives plus false positives.
One-step immunochromatographic test
Out of the 145 true TPPA-positive samples included, 12 samples were false negative in the Biorapid Syphilis, resulting in an overall sensitivity of 92% in our sample collection (table 3). The TPPA titre was below or equal to 1 : 320 in eight out of 12 false-negative samples (67%), two out of 12 had an RPR of 1 : 1 and one an RPR of 1 : 16. In selected samples with a TPPA titre above or equal to 1 : 2560 (n = 47), sensitivity increased to 100% but decreased to 83% in selected samples with a TPPA titre below or equal to 1 : 320 (n = 46). The specificity of the Biorapid Syphilis was 79%; all false-positive results are categorised in table 4. Although the Biorapid Syphilis uses a one-step procedure, interpretation was difficult in some cases (fig 1).
The sensitivity of the Biorapid Syphilis is considered too low to be used in a hospital laboratory in a developed country but might be useful in primary healthcare settings in developing countries.
The Bioelisa Syphilis 3.0 and ETI-Treponema Plus show excellent sensitivities and specificities when compared with the TPPA.
When using the Bioelisa Syphilis 3.0 and ETI-Treponema Plus on an automated system, the amount of laboratory and administrative errors in syphilis screening will be reduced.
Enzyme-linked immunosorbent assays
The sensitivity of the Bioelisa Syphilis 3.0 was 100%. As a result of too low a volume of some samples, the ETI-Treponema Plus was performed on samples from 143 out of 145 patients and 171 out of 185 controls. The missing control samples include 10 samples from leptospirosis patients. The sensitivity of the ETI-Treponema Plus was 100%, if two borderline results (absorbance/cut-off ratio 0.988 and 0.936, respectively) are considered positive. Specificity was 100% for both ELISA.
In this study, we report on the sensitivity and specificity of a one-step immunochromatographic rapid test, the Biorapid Syphilis and two ELISA, the Bioelisa Syphilis 3.0 and ETI-Treponema Plus, when compared with the TPPA. Regarding the Biorapid Syphilis, the 12 false-negative reactions resulted in a sensitivity of 92%, which we considered insufficient in our setting. In particular, having a negative Biorapid Syphilis when the RPR is 16 (0.7%), which suggests active disease, is worrisome. Although very easy to perform, the Biorapid Syphilis was often difficult to interpret, which renders the test less suitable for use by non-laboratory personnel at, for example, clinics specialising in sexually transmitted diseases. According to the manufacturer, every visible reaction should be considered positive. As can be seen in fig 1, differentiating between a positive, a weak positive and a negative result is subjective.
In this study, the specificity of the Biorapid Syphilis was determined to be 79%, based on the selection of samples used for the negative control test panel in this study. This specificity may, however, actually be higher when tested on a population at risk. Six out of 19 HIV-positive sera were false positive in the Biorapid Syphilis, which is unfortunate because most syphilis cases occur in areas with the highest HIV prevalence.1 27 Such false-positive reactions in syphilis serology in HIV-infected patients have been described before.15 16 18 28 Five out of 10 cytomegalovirus-positive control samples showed false-positive results with the Biorapid Syphilis, which has not been described before in syphilis serological studies. Furthermore, four out of 10 leptospirosis samples also gave false-positive reactions. Surprisingly, no samples from patients with borreliosis gave false-positive reactions, although these have been described in the literature.29 30 False-positive reactions in syphilis serology have also been described in several autoimmune disorders and indeed were found in our study, especially in the rheumatoid factor IgM-positive group.6 31 32
The largest multicentre evaluation of point-of-care tests with archived sera so far has recently been published by Herring et al.12 Nine rapid point-of-care syphilis tests were evaluated at eight laboratories on different continents. Sensitivities and specificities in that trial ranged between 84.5–97.7% and 92.8–98%, respectively. The sensitivity confidence interval of the Biorapid Syphilis was comparable with the rapid tests used by Herring et al but the specificity confidence interval was somewhat lower. This could be explained by the difference in sample collection. Although the Biorapid Syphilis is not sufficiently sensitive to be used in our hospital laboratory, it might be useful in primary healthcare settings in developing countries, as suggested by Herring et al, although further testing (using whole blood) is needed.
The ELISA tested in our study have excellent sensitivities and the results were equal or better when compared with the results of other ELISA (table 1).13 The specificity of both tests (100%) is especially good, considering the fact that serum samples from patients with most known possible cross-reacting pathological and physiological conditions were included in our population.6 15–18 20 28 29 33 The turnaround time of the ELISA was comparable with that of the TPPA. Both ELISA can be performed on an automated system, which may reduce the handling time and because of the electronic transfer of information, fewer administrative errors will be made. One small disadvantage of the Bioelisa Syphilis 3.0 is that the substrate-chromogen solution must be prepared and inserted in the automated system 5–10 minutes before usage, requiring additional handling by the laboratory technician. This is not the case with the ETI-Treponema Plus substrate-chromogen, which is ready to use and can be inserted at the beginning of the assay.
In conclusion, the sensitivity and specificity of the Biorapid Syphilis are considered too low to implement the test for screening purposes in a hospital laboratory but it might be useful in primary healthcare settings in developing countries. In contrast, both the Bioelisa Syphilis 3.0 and the ETI-Treponema Plus were found to have excellent sensitivities in this study and no false-positive reactions were detected in the control groups. As the ELISA can be run on an automated system, these have a clear advantage over the TPPA as a screening test.
The authors would like to thank Otto Bekers from the Clinical Chemistry Department of the University Hospital Maastricht for providing the sera from patients with diabetes and the Dutch Royal Tropical Institute for the sera from patients with leptospirosis. Furthermore, they wish to thank Peter Terporten and Foekje Stelma for statistical support.
Competing interests: None declared.
Ethics approval: This study was approved by the Medical Ethics Committee of the University Hospital, Maastricht, The Netherlands.
Contributors: LvD wrote the initial draft, participated in the study design with AS and VJG and did the statistical analysis. CJPAH played a key role in collecting materials and drafting the manuscript. FHvT and CAB were responsible for obtaining the free sample kits and contributed equally to writing the article. JD was important in collecting samples and reviewed the article. All authors participated in the set up of the study and the evaluation of the results. They all gave comments and suggestions.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.