- Email: [email protected]
Microbiological aspects of the diagnosisof Chlamydia trachomatis
Best Practice & Research Clinical Obstetrics and Gynaecology Vol. 16, No. 6, pp. 789±799, 2002
doi:10.1053/beog.2002.0322, available online at http://www.idealibrary.com on
3 Microbiological aspects of the diagnosis of Chlamydia trachomatis Lars éstergaard
MD, PhD, DMSc
Head Research Unit Q Department of Infectious Diseases, Skejby Sygehus, Aarhus University Hospital, DK-8200 Aarhus N, Denmark
The available diagnostic methods for Chlamydia trachomatis infection comprise serology (indirect detection) and culture, antigen detection and nucleic acid ampli®cation (direct detection). The rationale, applications, advantages and disadvantages of the methods and diagnostic targets are discussed. Compared to conventional methods, nucleic acid ampli®cation tests have increased sensitivity. This allows samples to be taken at home by the patient herself and mailed directly to the laboratory. Public health strategies implying home sampling for asymptomatic men and women result in a lower prevalence and a lower risk of short-term complications in terms of pelvic in¯ammatory disease (PID). The importance of predictive values and the association with prevalence are highlighted. Key words: Chlamydia trachomatis; diagnosis; screening; nucleic acid ampli®cation test.
CHLAMYDIA ANTIGENS USED IN DIAGNOSTIC TESTS Chlamydia trachomatis is a bacterium that belongs to the order Chlamydiales and the Chlamydiaceae family.1 The family Chlamydiaceae also includes two other human pathogens, Chlamydia psittaci and Chlamydia pneumoniae that do not cause urogenital infections in humans. All three human pathogens in the family Chlamydiaceae have a common lipopolysaccharide (LPS) antigen expressed on the surface so that diagnostic tests based on the detection of Chlamydia LPS cannot dierentiate between Chlamydia trachomatis and the other two species. In addition to LPS there are also proteins expressed on the surface of C. trachomatis. The major outer membrane protein (MOMP, also designated OMP1) constitutes 60% and is the predominant antigen.2,3 Other surface proteins include OMP2 and the recently discovered polymorphic membrane proteins (Pmps).4 MOMP consists of four variable domains (I, II, III and IV)5,6 which is the antigenic basis for further subdivision of C. trachomatis into the serotypes A through L7,8, of which serotypes D through K can cause urogenital infections. Serological tests based on the MOMP antigen will allow dierentiation between serotypes of C. trachomatis and the other two human pathogens of Chlamydia, and tests based on this antigen are therefore speci®c for C. trachomatis. c 2002 Elsevier Science Ltd. All rights reserved. 1521±6934/02/$ - see front matter *
790 L. éstergaard
Serological tests based on LPS comprise complement ®xation (CF), microimmuno¯uorescence assays (MIF) and enzyme immunoassays (EIA); MIF and EIA are available for antibodies against MOMP. INDIRECT DETECTION OF ONGOING OR PAST INFECTION WITH C. Trachomatis Use of serology for the diagnosis of C. trachomatis infection The value of serology is discussed in Chapter 4. However, serology is not of much help in making clinical decisions in the ®eld of gynaecology and obstetrics because the positive and negative predictive values of serology are either low or not yet assessed for clinically relevant outcomes.9±12 DIRECT DETECTION OF C. Trachomatis Light microscopy and cell culture Chlamydiae were ®rst detected by light microscopy in conjunctival scrapings from orangutangs inoculated with material from trachoma patients in 1907. The diagnostic sensitivity and speci®city of light microscopy, however, are not satisfactory.13 Because Chlamydia depends on ATP and other nutritional factors from a host cell it can reproduce only in other cells. In 1957 Chlamydia was successfully cultured in a chick embryo14, and in 1977 Ripa and MaÊrdh15 developed a method for the culture of C. trachomatis in McCoy cells by adding cycloheximide at the time of inoculation. Cycloheximide inhibits host cell protein synthesis, which leaves more nutrients for the growth and replication of Chlamydia. The inclusion body, containing thousands of C. trachomatis organisms, can be visualized either after staining with iodine, which reacts with the glycogen accumulated in the inclusion body, or by staining with a ¯uorescein-conjugated antibody directed against one of the organism's surface antigens.16 The latter increases the sensitivity of cell culture compared with iodine staining.17 Because the inclusion body is highly characteristic, cell culture is considered to have a speci®city of 100%. However, even with the use of ¯uorescein-conjugated antibodies, the sensitivity is not optimal, partly
Table 1. Diagnostic tests. Indirect detection of C. trachomatis (serology) LPS antigen Complement ®xation test (CFT) Microimmuno¯uorescence (MIF) Enzyme immuno assay (EIA) MOMP antigen Microimmuno¯uorescence (MIF) Enzyme immuno assay (EIA) Direct detection of C. trachomatis (culture, antigen or molecular detection) Microscopy and cell culture Antigen detection DNA/RNA detection
Microbiological aspects of the diagnosis of C. trachomatis 791
because organisms may lose infectivity during transportation and storage which will reduce the likelihood of propagation. In addition, the surface area of the cell culture layer and/or the amount of sample material added to the cell culture in¯uence the sensitivity; a second blind passage of culture may increase the sensitivity. Cell culture, however, is time-consuming and laborious and can therefore be provided by only a few central laboratories. Cell culture is the only method that detects live C. trachomatis and achieves a speci®city high enough to be used for forensic issues. Furthermore, due to the high speci®city, cell culture should be included in the reference standard to be used in comparative studies assessing the validity of new diagnostic tests.
Antigen detection Antigen detection methods comprise enzyme-linked immunosorbent assays (EIA) and direct immuno¯uorescence assays (DFA). Furthermore, rapid in-oce tests (Clearview, Surecell, Testpack, Quick Vue, etc.), which are also based on immunological reactions, have been increasingly applied.18 The diagnostic ecacy of these latter methods, however, is not high enough to warrant clinical use unless the need for a fast test result outweighs the lower diagnostic accuracy. The currently commercially available EIAs all use the LPS as antigen. The LPS part of Chlamydia binds to immobilized anti-LPS antibodies (MicroTrak, Chlamydiazyme, etc.), and the EIA tests are therefore genus-speci®c and detect all Chlamydia species. A secondary antibody that is bound to the Chlamydia is linked to an enzyme which generates a colour change, measured as optical density, on addition of a substrate. The reading of optical density values often does not allow a clear dierentiation between positive and negative samples, i.e. there is a `grey zone' of optical density values, and the sensitivity and speci®city are highly dependent on the chosen cut-o level between negative and positive samples.19 The diagnostic performance can be increased by testing samples within the grey zone by another test (con®rmation).19 However, even when this is done, the sensitivity of EIA cannot exceed that of an optimal cell culture.20 Also, the EIA tests may reveal positive results in the presence of other organisms such as Escherichia coli and Bacteroides sp, and Staphylococcus aureus may be captured instead of Chlamydia due to binding to the Fc region of the antibodies, thereby causing falsepositive reactions.21 A change of cross-reacting vaginal ¯ora may be the reason for the decreased speci®city of EIAs with increasing age.19 Cross-reacting organisms also make the EIA tests unsuitable for samples taken from anatomical regions other than the urethra and endocervix, for example, the pharynx and rectum.22 The cost of EIA is, however, considerably lower than that of molecular ampli®cation tests (see below), and EIAs could be considered in settings where cost is a major issue. A comprehensive review of the performance of EIAs has previously been published.23 In DFA, ¯uorescein-conjugated antibodies directed against either the LPS or the MOMP component react with the Chlamydia surface. The ¯uorescein can subsequently be visualized by ¯uorescence microscopy. Because the size of the extracellular elementary bodies is close to the resolving power of the microscope, DFA requires skilled personnel in order to dierentiate C. trachomatis organisms from non-speci®c ¯uorescent particles. The diagnostic performance of DFA is therefore highly dependent on the number of organisms that should be seen in order to obtain a positive result (the cut-o level).24,25 DFAs can be considered in settings where costs are a major issue.
792 L. éstergaard
DNA/RNA detection DNA/RNA detection assays can be divided into probe assays and ampli®cation assays. In probe assays a synthesized single-stranded oligonucleotide (a probe) hybridizes to a part of the C. trachomatis DNA or RNA. The most widely used probe technique is the Gen-Probe assay, in which a probe reacts with ribosomal RNA (rRNA) of C. trachomatis, which is present in hundreds of copies in each organism. The diagnostic performance of non-ampli®ed probe techniques is not substantially dierent from that of the best EIAs.20 Nucleic acid ampli®cation tests (NAATs) In nucleic acid ampli®cation tests (NAATs) speci®c probes hybridize to C. trachomatis DNA or RNA and the DNA/RNA ¯anked by the primers (target DNA) is exponentially copied. Due to the ampli®cation step, the diagnostic sensitivity of DNA/ RNA ampli®cation tests is theoretically considerably higher than that of antigen detection and probe detection tests. Because DNA is stable in the absence of DNAdegrading enzymes, C. trachomatis organisms that have lost their infectivity during transportation and storage may be detected by ampli®cation tests, which therefore also have a sensitivity higher than that of routine culture. If the regions at which primers anneal are unique for C. trachomatis, the DNA ampli®cation tests have a high speci®city, although their high sensitivity may also increase the risk of contaminating the samples with either native or ampli®ed target DNA. Target genes for NAATs Many nucleic acid regions have been used as target DNA molecules for amplifying C. trachomatis DNA: the major target genes have been the plasmid26±31, the MOMP gene32,33 and the gene encoding rRNA.34 In addition, the rRNA molecules have been used as target in RNA- dependent DNA amplication methods, TMA.35,36 The plasmid The plasmid37,38 is unique for C. trachomatis, is well conserved within the species, and is present in approximately 10 copies in each C. trachomatis organism.37,38 Using the plasmid as target DNA should therefore theoretically lower the detection limit by a factor of 10 compared with a single chromosomal gene, for example the MOMP gene. This has been documented in `in vitro' studies, in which the detection limit was lowered by a factor of four to 100 by using the plasmid as target DNA compared with the MOMP gene.39±41 Increased sensitivity by use of the plasmid, compared with the MOMP gene, has also been shown in clinical samples.42 However, some studies give evidence or suggest that plasmid-free variants are present in clinical samples43,44, and although it may seem that the plasmid is involved in DNA replication45,46 it has been possible to culture a plasmid-free variant.47 Thus, infections caused by plasmid-free variants will be undetected if the plasmid is used as target gene. This may result in diagnostic selection pressure if only infections caused by C. trachomatis containing the cryptic plasmid are detected and treated. A decline in prevalence of plasmid-containing variants and an increase in prevalence of plasmid-free variants might therefore be expected if diagnostic systems targeting only the plasmid are used in the future. An evaluation of the prevalence of plasmid-free variants should therefore be implemented in future survey programmes.
Microbiological aspects of the diagnosis of C. trachomatis 793
The MOMP gene Serotyping of C. trachomatis is performed by serological reactions between arti®cially produced antibodies against MOMP and cultured clinical isolates of C. trachomatis. Because a culture is necessary for serotyping, the proportion of typeable organisms is dependent on viability and the sensitivity of culture. By use of a PCR using MOMPdirected primers ¯anking variable regions of MOMP, it is possible to distinguish between the various types of C. trachomatis on non-cultured clinical samples. This is done by analysing the ampli®ed variable MOMP DNA fragment (genotyping) and can be done either by restriction fragment length polymorphism (RFLP)48±57 or by direct sequencing.58±62 MOMP-based genotyping has correlated well with serotyping, although limitations of genotyping may be observed in cases of concurrent infections with more than one genotype. Because genotyping using MOMP-directed PCRs can be performed directly on crude samples63 it is a powerful tool for epidemiological surveys and studies of genotype-related pathogenesis. In such studies it was found that complaints of urethral discharge and dysuria were most commonly associated with genotypes H and J and that lower abdominal pain was more often associated with F and G group genotypes than with B-complex or C-complex genotypes.64 PCR and ligase chain reaction (LCR) tests using the MOMP as target gene have also been widely used as a con®rmatory test on samples testing positive by a DNA ampli®cation test and negative by the comparative test in order to resolve such discrepant results. In these cases the result of the MOMP-directed PCR or LCR had been decisive for infection. The 16S-rRNA gene By using the 16S rRNA gene, which is present in all bacteria and is the most conserved gene known, all Chlamydia species can be detected by just one primer set. This is done by constructing the primers to anneal at the genus-speci®c regions of the 16S rRNA gene. The genus-speci®c regions ¯ank variable regions that are speci®c for each species. The ampli®ed products thus comprise the variable region ¯anked by the two genusspeci®c regions. The species can be determined by subsequent analyses of the variable region, either by hybridization with speci®c probes, by restriction fragment length polymorphism (RFLP), or by DNA sequencing. Due to the high homology of the 16S RNA gene with other organisms, optimal reaction conditions are crucial in order to avoid annealing of primers to 16S-rRNA genes of other organisms that are present in all non-sterile clinical samples. The ribosomal RNA The AMP-ct (TMA transcription-mediated ampli®cation) uses the rRNA as target molecule. The rRNA is present in hundreds of copies in each Chlamydia organism; this theoretically should contribute to an even higher sensitivity than ampli®cation tests based on detection of the plasmid, which is present in only 10 copies. However, rRNA may be more susceptible than DNA to disintegration and this may aect the clinical sensitivity; the clinical sensitivity of AMP-ct, however, does not seem to dier signi®cantly from that of DNA ampli®cation tests. Choice of target Which target molecule is selected for ampli®cation thus depends on the purpose of detection. If high sensitivity is needed the plasmid or rRNA should be chosen; however,
794 L. éstergaard Table 2. Diagnostic performance of NAATs.
Endocervical swabs Urine women Vaginal secretions (swabs, ¯ush, tampons) Urethral swabs (males) Urine men
Sensitivity (range %)
Speci®city (range %)
Positive predictive value (range %)
Negative predictive value (range %)
64±100 49±100 90±97
96±100 98±100 99±100
73±100 52±100 93±100
99 96±100 98±100
92±96 64±98
98±99 98±100
82±99 80±100
97±99 97±100
plasmid-free strains have been found and this may reduce the clinical sensitivity of plasmid-based ampli®cation tests. If the objective is genotyping, the MOMP gene may be chosen. This system, however, has a lower analytical and clinical sensitivity than plasmidbased systems. Diagnostic validity of NAATs When assessing the diagnostic validity of any test for C. trachomatis it is important to emphasize that no reference standard (gold standard) exists. Therefore, the validity of a new diagnostic test will always be a re¯ection of what has been chosen as the comparator. Table 2 shows ranges of the diagnostic performance of NAATs from various studies according to sampling site. Further data can be found in two comprehensive reviews.23,65 Sampling site For women in whom a pelvic examination is indicated there is no reason not to obtain an endocervical sample by use of a swab. If a urethral swab is also taken an additional 15% of infections will be found. Although it is generally not recommended, NAATs allow testing of throat and rectal swab samples; this will increase the sensitivty even more. Proper collection and transportation of samples are needed in order to achieve a high diagnostic yield. In one study66 only 25% of swab samples were microscopically determined to be adequate, and signi®cantly more infections were detected by PCR in microscopically adequate specimens.67 The increased analytical sensitivity of NAATs allows the use of sample material that contains fewer organisms than normal swab samples, i.e. they can be applied on urine and vaginal secretions. These types of sample do not require invasive procedures and can be obtained without medical assistance. Furthermore, NAATs allow the samples to be taken at home and sent through the mail to the diagnostic laboratory.68 The home sampling strategy has a considerable impact on case ®nding.69 Randomized studies have shown that, with a strategy of home sampling compared to oce testing, more asymptomatic cases can be found and more noti®ed partners will be tested.70,71 This will lead to a 50% lower rate of pelvic in¯ammatory disease (PID) and a 50% lower community-wide prevalence within 1 year.72 THE IMPORTANCE OF CONSIDERING THE PREDICTIVE VALUES WHEN REQUESTING A TEST The sensitivity (the ability of the test to detect positive samples) and the speci®city (the ability of the test to detect negative samples) are used to describe
Microbiological aspects of the diagnosis of C. trachomatis 795 Table 3. Predictive values in a population with a prevalence of 10%. Infected
Not infected
Total
Test positive Test negative
190 10
18 1782
208 1792
Total
200
1800
2000
Sensitivity: 190/200 95%; speci®city: 1782/1800 99%; positive predictive value: 190/208 91%; negative predictive value: 1782/1792 99%.
Table 4. Predictive values in a population with a prevalence of 1%. Infected
Not infected
Total
Test positive Test negative
19 1
20 1960
39 1961
Total
20
1980
2000
Sensitivity: 19/20 95%; speci®city; 1960/1980 99%; positive predictive value: 19/39 49%; negative predictive value: 1960/1961 99%.
the performance of the test. However, the positive predictive value (the likelihood that a positive test result is true) and the negative predictive value (the likelihood that a negative test result is true) are more relevant to the tested individual. This is because both an overlooked infection and a false-positive result for a sexually transmitted infection may have personal consequences. In contrast to the terms sensitivity and speci®city, the predictive values depend on the prevalence of infection in the tested population73, and despite rather high test validity in terms of sensitivity and speci®city, the clinician should consider the predictive values when requesting a test. An example is given for a test having a sensitivity of 95% and a speci®city of 99% applied on 2000 young symptomatic women with a prevalence of infection of 10% (Table 3) and on 2000 asymptomatic pregnant women aged 35 with a prevalence of 1% (Table 4). As seen, the positive predictive value decreases dramatically with decreasing prevalence, and when testing patients with a 1% prevalence, the chance that a positive test result is incorrect is higher than it being correct. Acknowledgement This study was supported by a grant from the Danish Centre for Evaluation and Health Technology Assessment.
SUMMARY The value of serology in the diagnosis of ongoing C. trachomatis infections is limited. More patients with late complications, such as infertility and ectopic pregnancy, have antibodies against C. trachomatis. However, a substantial fraction of healthy controls also have antibodies against C. trachomatis, and up to now the predictive value of serology has not been established for clinically relevant outcomes; serology is therefore
796 L. éstergaard
Practice points . aims. To obtain a diagnosis of Chlamydia trachomatis infection . management. Be aware of the predictive value of a test result in your setting before requesting the test . investigations. The increased sensitivity of NAATs allows samples to be taken by the patient him/herself. This has major public health implications but has no diagnostic advantage over endocervical and urethral swab samples if the patient is having a pelvic examination anyway. Do not use NAATs for test of cure within 4 weeks after initiation of therapy
Research agenda . predictive values for serology testing should be assessed for clinically relevant outcomes . suciently powered comparative studies of NAATs applied on low prevalence populations are needed . with the increased use of NAATs using the plasmid as target, the number of plasmid-free strains should be surveyed . the impact in terms of infertility and ectopic pregnancy of screening programmes based on NAATs should be assessed
of limited use in the clinical decision procedure. Cell culture is still the method with the highest speci®city and it should be used in clinical cases with potential legal signi®cance. Also, cell culture should be included in extended reference standards for assessing the diagnostic validity of new tests. Diagnostic methods based on antigen detection, such as direct ¯uorescence assays and enzyme immunoassays, are less expensive than nucleic acid ampli®cation tests. However, the sensitivity of nucleic acid ampli®cation tests is higher and this allows urine and vaginal swabs to be used as samples. These samples can be obtained by the patient him/herself and mailed directly to the laboratory. Public health measures using home sampling for screening asymptomatic men and women have been shown to reduce the risk of pelvic in¯ammatory disease and the community-wide prevalence. Partner noti®cation is still important, and `test of cure' can be considered ± but only after the ®rst 4 weeks following initiation of therapy. It is crucial for the clinician to know the positive and negative predictive values of a C. trachomatis test in his/her setting. REFERENCES 1. The National Institute of Health. NIH Taxonomy Browser (http://www.ncbi.nlm.nih.gov/htbin-post/ Taxonomy). 2. Zhong GM, Reid RE & Brunham RC. Mapping antigenic sites on the major outer membrane protein of Chlamydia trachomatis with synthetic peptides. Infection and Immunity 1990; 58: 1450±1455. * 3. Campbell LA, Kuo CC & Grayston JT. Structural and antigenic analysis of Chlamydia pneumoniae. Infection and Immunity 1990; 58: 93±97. * 4. Grimwood J & Stephens RS. Computational analysis of the polymorphic membrane protein superfamily of Chlamydia trachomatis and Chlamydia pneumoniae. Microbial Comparative Genomics 1999; 4: 187±201.
Microbiological aspects of the diagnosis of C. trachomatis 797 5. Zhang YX, Morrison SG, Caldwell HD & Baehr W. Cloning and sequence analysis of the major outer membrane protein genes of two Chlamydia psittaci strains. Infection and Immunity 1989; 57: 1621±1625. 6. Yuan Y, Zhang YX, Watkins NG & Caldwell HD. Nucleotide and deduced amino acid sequences for the four variable domains of the major outer membrane proteins of the 15 Chlamydia trachomatis serovars. Infection and Immunity 1989; 57: 1040±1049. 7. Toye B, Zhong GM, Peeling R & Brunham RC. Immunologic characterization of a cloned fragment containing the species-speci®c epitope from the major outer membrane protein of Chlamydia trachomatis. Infection and Immunity 1990; 58: 3909±3913. * 8. Stephens RS, Wagar EA & Schoolnik GK. High-resolution mapping of serovar-speci®c and common antigenic determinants of the major outer membrane protein of Chlamydia trachomatis. Journal of Experimental Medicine 1988; 167: 817±831. 9. Keay SD, Barlow R, Eley A et al. The relation between immunoglobulin G antibodies to Chlamydia trachomatis and poor ovarian response to gonadotropin stimulation before in vitro fertilization. Fertility and Sterility 1998; 70: 214±218. 10. Spandorfer SD, Neuer A, LaVerda D et al. Previously undetected Chlamydia trachomatis infection, immunity to heat shock proteins and tubal occlusion in women undergoing in-vitro fertilization. Human Reproduction 1999; 14: 60±64. 11. Sharara FI, Queenan JT Jr, Springer RS et al. Elevated serum Chlamydia trachomatis IgG antibodies. What do they mean for IVF pregnancy rates and loss? Journal of Reproductive Medicine 1997; 42: 281±286. *12. Claman P, Amimi MN, Peeling RW et al. Does serologic evidence of remote Chlamydia trachomatis infection and its heat shock protein (CHSP 60) aect in vitro fertilization-embryo transfer outcome? Fertility and Sterility 1996; 65: 146±149. 13. Vinette-Leduc D, Yazdi HM, Jessamine P & Peeling RW. Reliability of cytology to detect chlamydial infection in asymptomatic women. Diagnostic Cytopathology 1997; 17: 258±261. 14. T'ang FF, Chang HL, Huang YT & Wang KC. Studies on the etiology of trachoma with special reference to isolation of the virus in chick embryo. Chinese Medical Journal 1957; 75: 429±447. *15. Ripa KT & MaÊrdh PA. Cultivation of Chlamydia trachomatis in cycloheximide-treated Mccoy cells. Journal of Clinical Microbiology 1977; 6: 328±331. 16. Stamm WE, Tam M, Koester M & Cles L. Detection of Chlamydia trachomatis inclusions in Mccoy cell cultures with ¯uorescein-conjugated monoclonal antibodies. Journal of Clinical Microbiology 1983; 17: 666±668. 17. Phillips LE, Smith PB, Riddle GD et al. Ortho enzyme immunoassay versus McCoy cell monolayers stained by iodine or ¯uorescent antibody for detection of Chlamydia trachomatis. Journal of Clinical Microbiology 1990; 28: 1647±1648. 18. Suchland KL, Counts JM & Stamm WE. Laboratory methods for detection of Chlamydia trachomatis: survey of laboratories in Washington State. Journal of Clinical Microbiology 1997; 35: 3210±3214. 19. éstergaard L & Mùller JK. Use of PCR and direct immuno¯uorescence microscopy for con®rmation of results obtained by Syva MicroTrak Chlamydia enzyme immunoassay. Journal of Clinical Microbiology 1995; 33: 2620±2623. 20. Mùller JK, éstergaard L & Hansen JT. Clinical evaluation of four non-related techniques for detection of Chlamydia trachomatis in endocervical specimens. Immunology and Infectious Diseases 1994; 4: 191±196. 21. Kellogg JA, Seiple JW & Hick ME. Cross-reaction of clinical isolates of bacteria and yeasts with the chlamydiazyme test for chlamydial antigen, before and after use of a blocking reagent. American Journal of Clinical Pathology 1992; 97: 309±312. 22. Hammerschlag MR, Rettig PJ & Shields ME. False positive results with the use of chlamydial antigen detection tests in the evaluation of suspected sexual abuse in children. Pediatric Infectious Disease Journal 1988; 7: 11±14. *23. Black CM. Current methods of laboratory diagnosis of Chlamydia trachomatis infections. Clinical Microbiological Review 1997; 10: 160±184. 24. Thejls H, Gnarpe J, Gnarpe H et al. Expanded gold standard in the diagnosis of Chlamydia trachomatis in a low prevalence population: diagnostic ecacy of tissue culture, direct immuno¯uorescence, enzyme immunoassay, PCR and serology. Genitourinary Medicine 1994; 70: 300±303. 25. Quinn TC, War®eld P, Kappus E et al. Screening for Chlamydia trachomatis infection in an inner-city population: a comparison of diagnostic methods. Journal of Infectious Diseases 1985; 152: 419±423. 26. éstergaard L, Birkelund S & Christiansen G. Use of polymerase chain reaction for detection of Chlamydia trachomatis. Journal of Clinical Microbiology 1990; 28: 1254±1260. 27. Claas HC, Melchers WJ, de Bruijn IH et al. Detection of Chlamydia trachomatis in clinical specimens by the polymerase chain reaction. European Journal of Clinical Microbiology and Infectious Diseases 1990; 9: 864±868.
798 L. éstergaard 28. Mahony JB, Luinstra KE, Jang D et al. Chlamydia trachomatis con®rmatory testing of PCR-positive genitourinary specimens using a second set of plasmid primers. Molecular and Cellular Probes 1992; 6: 381±388. 29. NaÈher H, Drzonek H, Wolf J et al. Detection of C. trachomatis in urogenital specimens by polymerase chain reaction. Genitourinary Medicine 1991; 67: 211±214. 30. Ossewaarde JM, Riee M, Rozenberg Arska M et al. Development and clinical evaluation of a polymerase chain reaction test for detection of Chlamydia trachomatis. Journal of Clinical Microbiology 1992; 30: 2122±2128. 31. Ratti G, Moroni A & Cevenini R. Detection of Chlamydia trachomatis DNA in patients with nongonococcal urethritis using the polymerase chain reaction. Journal of Clinical Pathology 1991; 44: 564±568. 32. Holland SM, Gaydos CA & Quinn TC. Detection and dierentiation of Chlamydia trachomatis, Chlamydia psittaci, and Chlamydia pneumoniae by DNA ampli®cation. Journal of Infectious Diseases 1990; 162: 984±987. 33. Bobo L, Coutlee F, Yolken RH et al. Diagnosis of Chlamydia trachomatis cervical infection by detection of ampli®ed DNA with an enzyme immunoassay. Journal of Clinical Microbiology 1990; 28: 1968±1973. 34. Claas HC, Melchers WJ, de Bruijn IH et al. Detection of Chlamydia trachomatis in clinical specimens by the polymerase chain reaction. European Journal of Clinical Microbiology and Infectious Diseases 1990; 9: 864±868. 35. Mouton JW, Verkooyen R, van der Meijden WI et al. Detection of Chlamydia trachomatis in male and female urine specimens by using the ampli®ed Chlamydia trachomatis test. Journal of Clinical Microbiology 1997; 35: 1369±1372. 36. Mahony J, Chong S, Jang D et al. Urine specimens from pregnant and nonpregnant women inhibitory to ampli®cation of Chlamydia trachomatis nucleic acid by PCR, ligase chain reaction, and transcriptionmediated ampli®cation: identi®cation of urinary substances associated with inhibition and removal of inhibitory activity. Journal of Clinical Microbiology 1998; 36: 3122±3126. 37. Sriprakash KS & Macavoy ES. Characterization and sequence of a plasmid from the trachoma biovar of Chlamydia trachomatis. Plasmid 1987; 18: 205±214. 38. Palmer L & Falkow S. A common plasmid of Chlamydia trachomatis. Plasmid 1986; 16: 52±62. 39. Roosendaal R, Walboomers JM, Veltman OR et al. Comparison of dierent primer sets for detection of Chlamydia trachomatis by the polymerase chain reaction. Journal of Medical Microbiology 1993; 38: 426±433. 40. Ossewaarde JM, Riee M, Rozenberg-Arska M et al. Development and clinical evaluation of a polymerase chain reaction test for detection of Chlamydia trachomatis. Journal of Clinical Microbiology 1992; 30: 2122±2128. 41. Martin JL, Alexander SY, Selwood TS & Cross GF. Use of the polymerase chain reaction for the detection of Chlamydia trachomatis in clinical specimens and its comparison to commercially available tests. Genitourinary Medicine 1995; 71: 169±171. 42. Mahony JB, Luinstra KE, Sellors JW & Chernesky MA. Comparison of plasmid- and chromosome-based polymerase chain reaction assays for detecting Chlamydia trachomatis nucleic acids. Journal of Clinical Microbiology 1993; 31: 1753±1758. *43. An Q, Radclie G, Vassallo R et al. Infection with a plasmid-free variant Chlamydia related to Chlamydia trachomatis identi®ed by using multiple assays for nucleic acid detection. Journal of Clinical Microbiology 1992; 30: 2814±2821. 44. Schachter J, Stamm WE & Quinn TC. Discrepant analysis and screening for Chlamydia trachomatis. Lancet 1996; 348: 1308±1309. 45. Fahr MJ, Sriprakash KS & Hatch TP. Convergent and overlapping transcripts of the Chlamydia trachomatis 7.5-kb plasmid. Plasmid 1992; 28: 247±257. 46. Hatt C, Ward ME & Clarke IN. Analysis of the entire nucleotide sequence of the cryptic plasmid of Chlamydia trachomatis serovar L1. Evidence for involvement in DNA replication. Nucleic Acids Research 1988; 16: 4053±4067. 47. Peterson EM, Marko BA, Schachter J & de la Maza LM. The 7.5-kb plasmid present in Chlamydia trachomatis is not essential for the growth of this microorganism. Plasmid 1990; 23: 144±148. 48. Lan J, Walboomers JM, Roosendaal R et al. Direct detection and genotyping of Chlamydia trachomatis in cervical scrapes by using polymerase chain reaction and restriction fragment length polymorphism analysis. Journal of Clinical Microbiology 1993; 31: 1060±1065. 49. Frost EH, Deslandes S & Bourgaux Ramoisy D. Sensitive detection and typing of Chlamydia trachomatis using nested polymerase chain reaction. Genitourinary Medicine 1993; 69: 290±294. 50. Rodriguez P, Vekris A, de Barbeyrac B et al. Typing of Chlamydia trachomatis by restriction endonuclease analysis of the ampli®ed major outer membrane protein gene. Journal of Clinical Microbiology 1991; 29: 1132±1136.
Microbiological aspects of the diagnosis of C. trachomatis 799 51. van de Laar MJ, van Duynhoven YT, Fennema JS et al. Dierences in clinical manifestations of genital chlamydial infections related to serovars. Genitourinary Medicine 1996; 72: 261±265. 52. Yang CL, Maclean I & Brunham RC. DNA sequence polymorphism of the Chlamydia trachomatis omp1 gene. Journal of Infectious Diseases 1993; 168: 1225±1230. 53. Lan J, Melgers I, Meijer CJ et al. Prevalence and serovar distribution of asymptomatic cervical Chlamydia trachomatis infections as determined by highly sensitive PCR. Journal of Clinical Microbiology 1995; 33: 3194±3197. 54. Sayada C, Denamur E, Or®la J et al. Rapid genotyping of the Chlamydia trachomatis major outer membrane protein by the polymerase chain reaction. FEMS Microbiology Letters 1991; 67: 73±78. 55. Black CM, Tharpe JA & Russell H. Distinguishing Chlamydia species by restriction analysis of the major outer membrane protein gene. Molecular and Cellular Probes 1992; 6: 395±400. 56. Gaydos CA, Bobo L, Welsh L et al. Gene typing of Chlamydia trachomatis by polymerase chain reaction and restriction endonuclease digestion. Sexually Transmitted Diseases 1992; 19: 303±308. *57. Frost EH, Deslandes S, Veilleux S et al. Typing Chlamydia trachomatis by detection of restriction fragment length polymorphism in the gene encoding the major outer membrane protein. Journal of Infectious Diseases 1991; 163: 1103±1107. 58. Poole E & Lamont I. Chlamydia trachomatis serovar dierentiation by direct sequence analysis of the variable segment 4 region of the major outer membrane protein gene. Infection and Immunity 1992; 60: 1089±1094. 59. Viscidi RP, Bobo L, Hook EW & Quinn TC. Transmission of Chlamydia trachomatis among sex partners assessed by polymerase chain reaction. Journal of Infectious Diseases 1993; 168: 488±492. 60. Frost EH, Deslandes S & Bourgaux Ramoisy D. Chlamydia trachomatis serovars in 435 urogenital specimens typed by restriction endonuclease analysis of ampli®ed DNA. Journal of Infectious Diseases 1993; 168: 497±501. 61. Hayes LJ, Pecharatana S, Bailey RL et al. Extent and kinetics of genetic change in the omp1 gene of Chlamydia trachomatis in two villages with endemic trachoma. Journal of Infectious Diseases 1995; 172: 268±272. 62. Quinn TC, Gaydos C, Shepherd M et al. Epidemiologic and microbiologic correlates of Chlamydia trachomatis infection in sexual partnerships. Journal of the American Medical Association 1996; 276: 1737±1742. 63. Pedersen LN, Kjaer HO, Mùller JK et al. High-resolution genotyping of Chlamydia trachomatis from recurrent urogenital infections. Journal of Clinical Microbiology 2000; 38: 3068±3071. 64. van Duynhoven YT, Ossewaarde JM, Derksen-Nawrocki RP et al. Chlamydia trachomatis genotypes: correlation with clinical manifestations of infection and patients' characteristics. Clinical Infectious Diseases 1998; 26: 314±322. *65. éstergaard L. Diagnosis of urogenital Chlamydia trachomatis infection by use of DNA ampli®cation. APMIS 1999; 89 (supplement): 5±36. 66. Kellogg JA, Seiple JW, Klinedinst JL et al. Improved PCR detection of Chlamydia trachomatis by using an altered method of specimen transport and high-quality endocervical specimens. Journal of Clinical Microbiology 1995; 33: 2765±2767. 67. Kellogg JA, Seiple JW, Klinedinst JL & Stroll E. Di-Quik stain as a simpli®ed alternative to Papanicolaou stain for determination of quality of endocervical specimens submitted for PCR detection of Chlamydia trachomatis. Journal of Clinical Microbiology 1996; 34: 2590±2592. 68. éstergaard L, Mùller JK, Andersen B & Olesen F. Diagnosis of urogenital Chlamydia trachomatis infection in women based on mailed samples obtained at home: multipractice comparative study. British Medical Journal 1996; 313: 1186±1189. 69. Andersen B, Olesen F, Mùller JK & éstergaard L. Population-based strategies for outreach screening of urogenital Chlamydia trachomatis infections: a randomized, controlled trial. Journal of Infectious Diseases 2002; 185: 252±258. 70. Andersen B, éstergaard L, Mùller JK & Olesen F. Home sampling versus conventional contact tracing for detecting Chlamydia trachomatis in male partners of infected women: randomised study. British Medical Journal 1998; 316: 350±351. 71. éstergaard L, Andersen B, Olesen F & Mùller JK. Ecacy of home sampling for screening of Chlamydia trachomatis: randomised study. British Medical Journal 1998; 317: 26±27. *72. éstergaard L, Andersen B, Mùller JK & Olesen F. Home sampling versus conventional swab sampling for screening of Chlamydia trachomatis in women: a cluster-randomized 1-year follow-up study. Clinical Infectious Diseases 2000; 31: 951±957. 73. éstergaard L. Detection of Chlamydia trachomatis in a low prevalence population. European Journal of Clinical Microbiology and Infectious Diseases 1995; 14: 471±472.
doi:10.1053/beog.2002.0322, available online at http://www.idealibrary.com on
3 Microbiological aspects of the diagnosis of Chlamydia trachomatis Lars éstergaard
MD, PhD, DMSc
Head Research Unit Q Department of Infectious Diseases, Skejby Sygehus, Aarhus University Hospital, DK-8200 Aarhus N, Denmark
The available diagnostic methods for Chlamydia trachomatis infection comprise serology (indirect detection) and culture, antigen detection and nucleic acid ampli®cation (direct detection). The rationale, applications, advantages and disadvantages of the methods and diagnostic targets are discussed. Compared to conventional methods, nucleic acid ampli®cation tests have increased sensitivity. This allows samples to be taken at home by the patient herself and mailed directly to the laboratory. Public health strategies implying home sampling for asymptomatic men and women result in a lower prevalence and a lower risk of short-term complications in terms of pelvic in¯ammatory disease (PID). The importance of predictive values and the association with prevalence are highlighted. Key words: Chlamydia trachomatis; diagnosis; screening; nucleic acid ampli®cation test.
CHLAMYDIA ANTIGENS USED IN DIAGNOSTIC TESTS Chlamydia trachomatis is a bacterium that belongs to the order Chlamydiales and the Chlamydiaceae family.1 The family Chlamydiaceae also includes two other human pathogens, Chlamydia psittaci and Chlamydia pneumoniae that do not cause urogenital infections in humans. All three human pathogens in the family Chlamydiaceae have a common lipopolysaccharide (LPS) antigen expressed on the surface so that diagnostic tests based on the detection of Chlamydia LPS cannot dierentiate between Chlamydia trachomatis and the other two species. In addition to LPS there are also proteins expressed on the surface of C. trachomatis. The major outer membrane protein (MOMP, also designated OMP1) constitutes 60% and is the predominant antigen.2,3 Other surface proteins include OMP2 and the recently discovered polymorphic membrane proteins (Pmps).4 MOMP consists of four variable domains (I, II, III and IV)5,6 which is the antigenic basis for further subdivision of C. trachomatis into the serotypes A through L7,8, of which serotypes D through K can cause urogenital infections. Serological tests based on the MOMP antigen will allow dierentiation between serotypes of C. trachomatis and the other two human pathogens of Chlamydia, and tests based on this antigen are therefore speci®c for C. trachomatis. c 2002 Elsevier Science Ltd. All rights reserved. 1521±6934/02/$ - see front matter *
790 L. éstergaard
Serological tests based on LPS comprise complement ®xation (CF), microimmuno¯uorescence assays (MIF) and enzyme immunoassays (EIA); MIF and EIA are available for antibodies against MOMP. INDIRECT DETECTION OF ONGOING OR PAST INFECTION WITH C. Trachomatis Use of serology for the diagnosis of C. trachomatis infection The value of serology is discussed in Chapter 4. However, serology is not of much help in making clinical decisions in the ®eld of gynaecology and obstetrics because the positive and negative predictive values of serology are either low or not yet assessed for clinically relevant outcomes.9±12 DIRECT DETECTION OF C. Trachomatis Light microscopy and cell culture Chlamydiae were ®rst detected by light microscopy in conjunctival scrapings from orangutangs inoculated with material from trachoma patients in 1907. The diagnostic sensitivity and speci®city of light microscopy, however, are not satisfactory.13 Because Chlamydia depends on ATP and other nutritional factors from a host cell it can reproduce only in other cells. In 1957 Chlamydia was successfully cultured in a chick embryo14, and in 1977 Ripa and MaÊrdh15 developed a method for the culture of C. trachomatis in McCoy cells by adding cycloheximide at the time of inoculation. Cycloheximide inhibits host cell protein synthesis, which leaves more nutrients for the growth and replication of Chlamydia. The inclusion body, containing thousands of C. trachomatis organisms, can be visualized either after staining with iodine, which reacts with the glycogen accumulated in the inclusion body, or by staining with a ¯uorescein-conjugated antibody directed against one of the organism's surface antigens.16 The latter increases the sensitivity of cell culture compared with iodine staining.17 Because the inclusion body is highly characteristic, cell culture is considered to have a speci®city of 100%. However, even with the use of ¯uorescein-conjugated antibodies, the sensitivity is not optimal, partly
Table 1. Diagnostic tests. Indirect detection of C. trachomatis (serology) LPS antigen Complement ®xation test (CFT) Microimmuno¯uorescence (MIF) Enzyme immuno assay (EIA) MOMP antigen Microimmuno¯uorescence (MIF) Enzyme immuno assay (EIA) Direct detection of C. trachomatis (culture, antigen or molecular detection) Microscopy and cell culture Antigen detection DNA/RNA detection
Microbiological aspects of the diagnosis of C. trachomatis 791
because organisms may lose infectivity during transportation and storage which will reduce the likelihood of propagation. In addition, the surface area of the cell culture layer and/or the amount of sample material added to the cell culture in¯uence the sensitivity; a second blind passage of culture may increase the sensitivity. Cell culture, however, is time-consuming and laborious and can therefore be provided by only a few central laboratories. Cell culture is the only method that detects live C. trachomatis and achieves a speci®city high enough to be used for forensic issues. Furthermore, due to the high speci®city, cell culture should be included in the reference standard to be used in comparative studies assessing the validity of new diagnostic tests.
Antigen detection Antigen detection methods comprise enzyme-linked immunosorbent assays (EIA) and direct immuno¯uorescence assays (DFA). Furthermore, rapid in-oce tests (Clearview, Surecell, Testpack, Quick Vue, etc.), which are also based on immunological reactions, have been increasingly applied.18 The diagnostic ecacy of these latter methods, however, is not high enough to warrant clinical use unless the need for a fast test result outweighs the lower diagnostic accuracy. The currently commercially available EIAs all use the LPS as antigen. The LPS part of Chlamydia binds to immobilized anti-LPS antibodies (MicroTrak, Chlamydiazyme, etc.), and the EIA tests are therefore genus-speci®c and detect all Chlamydia species. A secondary antibody that is bound to the Chlamydia is linked to an enzyme which generates a colour change, measured as optical density, on addition of a substrate. The reading of optical density values often does not allow a clear dierentiation between positive and negative samples, i.e. there is a `grey zone' of optical density values, and the sensitivity and speci®city are highly dependent on the chosen cut-o level between negative and positive samples.19 The diagnostic performance can be increased by testing samples within the grey zone by another test (con®rmation).19 However, even when this is done, the sensitivity of EIA cannot exceed that of an optimal cell culture.20 Also, the EIA tests may reveal positive results in the presence of other organisms such as Escherichia coli and Bacteroides sp, and Staphylococcus aureus may be captured instead of Chlamydia due to binding to the Fc region of the antibodies, thereby causing falsepositive reactions.21 A change of cross-reacting vaginal ¯ora may be the reason for the decreased speci®city of EIAs with increasing age.19 Cross-reacting organisms also make the EIA tests unsuitable for samples taken from anatomical regions other than the urethra and endocervix, for example, the pharynx and rectum.22 The cost of EIA is, however, considerably lower than that of molecular ampli®cation tests (see below), and EIAs could be considered in settings where cost is a major issue. A comprehensive review of the performance of EIAs has previously been published.23 In DFA, ¯uorescein-conjugated antibodies directed against either the LPS or the MOMP component react with the Chlamydia surface. The ¯uorescein can subsequently be visualized by ¯uorescence microscopy. Because the size of the extracellular elementary bodies is close to the resolving power of the microscope, DFA requires skilled personnel in order to dierentiate C. trachomatis organisms from non-speci®c ¯uorescent particles. The diagnostic performance of DFA is therefore highly dependent on the number of organisms that should be seen in order to obtain a positive result (the cut-o level).24,25 DFAs can be considered in settings where costs are a major issue.
792 L. éstergaard
DNA/RNA detection DNA/RNA detection assays can be divided into probe assays and ampli®cation assays. In probe assays a synthesized single-stranded oligonucleotide (a probe) hybridizes to a part of the C. trachomatis DNA or RNA. The most widely used probe technique is the Gen-Probe assay, in which a probe reacts with ribosomal RNA (rRNA) of C. trachomatis, which is present in hundreds of copies in each organism. The diagnostic performance of non-ampli®ed probe techniques is not substantially dierent from that of the best EIAs.20 Nucleic acid ampli®cation tests (NAATs) In nucleic acid ampli®cation tests (NAATs) speci®c probes hybridize to C. trachomatis DNA or RNA and the DNA/RNA ¯anked by the primers (target DNA) is exponentially copied. Due to the ampli®cation step, the diagnostic sensitivity of DNA/ RNA ampli®cation tests is theoretically considerably higher than that of antigen detection and probe detection tests. Because DNA is stable in the absence of DNAdegrading enzymes, C. trachomatis organisms that have lost their infectivity during transportation and storage may be detected by ampli®cation tests, which therefore also have a sensitivity higher than that of routine culture. If the regions at which primers anneal are unique for C. trachomatis, the DNA ampli®cation tests have a high speci®city, although their high sensitivity may also increase the risk of contaminating the samples with either native or ampli®ed target DNA. Target genes for NAATs Many nucleic acid regions have been used as target DNA molecules for amplifying C. trachomatis DNA: the major target genes have been the plasmid26±31, the MOMP gene32,33 and the gene encoding rRNA.34 In addition, the rRNA molecules have been used as target in RNA- dependent DNA amplication methods, TMA.35,36 The plasmid The plasmid37,38 is unique for C. trachomatis, is well conserved within the species, and is present in approximately 10 copies in each C. trachomatis organism.37,38 Using the plasmid as target DNA should therefore theoretically lower the detection limit by a factor of 10 compared with a single chromosomal gene, for example the MOMP gene. This has been documented in `in vitro' studies, in which the detection limit was lowered by a factor of four to 100 by using the plasmid as target DNA compared with the MOMP gene.39±41 Increased sensitivity by use of the plasmid, compared with the MOMP gene, has also been shown in clinical samples.42 However, some studies give evidence or suggest that plasmid-free variants are present in clinical samples43,44, and although it may seem that the plasmid is involved in DNA replication45,46 it has been possible to culture a plasmid-free variant.47 Thus, infections caused by plasmid-free variants will be undetected if the plasmid is used as target gene. This may result in diagnostic selection pressure if only infections caused by C. trachomatis containing the cryptic plasmid are detected and treated. A decline in prevalence of plasmid-containing variants and an increase in prevalence of plasmid-free variants might therefore be expected if diagnostic systems targeting only the plasmid are used in the future. An evaluation of the prevalence of plasmid-free variants should therefore be implemented in future survey programmes.
Microbiological aspects of the diagnosis of C. trachomatis 793
The MOMP gene Serotyping of C. trachomatis is performed by serological reactions between arti®cially produced antibodies against MOMP and cultured clinical isolates of C. trachomatis. Because a culture is necessary for serotyping, the proportion of typeable organisms is dependent on viability and the sensitivity of culture. By use of a PCR using MOMPdirected primers ¯anking variable regions of MOMP, it is possible to distinguish between the various types of C. trachomatis on non-cultured clinical samples. This is done by analysing the ampli®ed variable MOMP DNA fragment (genotyping) and can be done either by restriction fragment length polymorphism (RFLP)48±57 or by direct sequencing.58±62 MOMP-based genotyping has correlated well with serotyping, although limitations of genotyping may be observed in cases of concurrent infections with more than one genotype. Because genotyping using MOMP-directed PCRs can be performed directly on crude samples63 it is a powerful tool for epidemiological surveys and studies of genotype-related pathogenesis. In such studies it was found that complaints of urethral discharge and dysuria were most commonly associated with genotypes H and J and that lower abdominal pain was more often associated with F and G group genotypes than with B-complex or C-complex genotypes.64 PCR and ligase chain reaction (LCR) tests using the MOMP as target gene have also been widely used as a con®rmatory test on samples testing positive by a DNA ampli®cation test and negative by the comparative test in order to resolve such discrepant results. In these cases the result of the MOMP-directed PCR or LCR had been decisive for infection. The 16S-rRNA gene By using the 16S rRNA gene, which is present in all bacteria and is the most conserved gene known, all Chlamydia species can be detected by just one primer set. This is done by constructing the primers to anneal at the genus-speci®c regions of the 16S rRNA gene. The genus-speci®c regions ¯ank variable regions that are speci®c for each species. The ampli®ed products thus comprise the variable region ¯anked by the two genusspeci®c regions. The species can be determined by subsequent analyses of the variable region, either by hybridization with speci®c probes, by restriction fragment length polymorphism (RFLP), or by DNA sequencing. Due to the high homology of the 16S RNA gene with other organisms, optimal reaction conditions are crucial in order to avoid annealing of primers to 16S-rRNA genes of other organisms that are present in all non-sterile clinical samples. The ribosomal RNA The AMP-ct (TMA transcription-mediated ampli®cation) uses the rRNA as target molecule. The rRNA is present in hundreds of copies in each Chlamydia organism; this theoretically should contribute to an even higher sensitivity than ampli®cation tests based on detection of the plasmid, which is present in only 10 copies. However, rRNA may be more susceptible than DNA to disintegration and this may aect the clinical sensitivity; the clinical sensitivity of AMP-ct, however, does not seem to dier signi®cantly from that of DNA ampli®cation tests. Choice of target Which target molecule is selected for ampli®cation thus depends on the purpose of detection. If high sensitivity is needed the plasmid or rRNA should be chosen; however,
794 L. éstergaard Table 2. Diagnostic performance of NAATs.
Endocervical swabs Urine women Vaginal secretions (swabs, ¯ush, tampons) Urethral swabs (males) Urine men
Sensitivity (range %)
Speci®city (range %)
Positive predictive value (range %)
Negative predictive value (range %)
64±100 49±100 90±97
96±100 98±100 99±100
73±100 52±100 93±100
99 96±100 98±100
92±96 64±98
98±99 98±100
82±99 80±100
97±99 97±100
plasmid-free strains have been found and this may reduce the clinical sensitivity of plasmid-based ampli®cation tests. If the objective is genotyping, the MOMP gene may be chosen. This system, however, has a lower analytical and clinical sensitivity than plasmidbased systems. Diagnostic validity of NAATs When assessing the diagnostic validity of any test for C. trachomatis it is important to emphasize that no reference standard (gold standard) exists. Therefore, the validity of a new diagnostic test will always be a re¯ection of what has been chosen as the comparator. Table 2 shows ranges of the diagnostic performance of NAATs from various studies according to sampling site. Further data can be found in two comprehensive reviews.23,65 Sampling site For women in whom a pelvic examination is indicated there is no reason not to obtain an endocervical sample by use of a swab. If a urethral swab is also taken an additional 15% of infections will be found. Although it is generally not recommended, NAATs allow testing of throat and rectal swab samples; this will increase the sensitivty even more. Proper collection and transportation of samples are needed in order to achieve a high diagnostic yield. In one study66 only 25% of swab samples were microscopically determined to be adequate, and signi®cantly more infections were detected by PCR in microscopically adequate specimens.67 The increased analytical sensitivity of NAATs allows the use of sample material that contains fewer organisms than normal swab samples, i.e. they can be applied on urine and vaginal secretions. These types of sample do not require invasive procedures and can be obtained without medical assistance. Furthermore, NAATs allow the samples to be taken at home and sent through the mail to the diagnostic laboratory.68 The home sampling strategy has a considerable impact on case ®nding.69 Randomized studies have shown that, with a strategy of home sampling compared to oce testing, more asymptomatic cases can be found and more noti®ed partners will be tested.70,71 This will lead to a 50% lower rate of pelvic in¯ammatory disease (PID) and a 50% lower community-wide prevalence within 1 year.72 THE IMPORTANCE OF CONSIDERING THE PREDICTIVE VALUES WHEN REQUESTING A TEST The sensitivity (the ability of the test to detect positive samples) and the speci®city (the ability of the test to detect negative samples) are used to describe
Microbiological aspects of the diagnosis of C. trachomatis 795 Table 3. Predictive values in a population with a prevalence of 10%. Infected
Not infected
Total
Test positive Test negative
190 10
18 1782
208 1792
Total
200
1800
2000
Sensitivity: 190/200 95%; speci®city: 1782/1800 99%; positive predictive value: 190/208 91%; negative predictive value: 1782/1792 99%.
Table 4. Predictive values in a population with a prevalence of 1%. Infected
Not infected
Total
Test positive Test negative
19 1
20 1960
39 1961
Total
20
1980
2000
Sensitivity: 19/20 95%; speci®city; 1960/1980 99%; positive predictive value: 19/39 49%; negative predictive value: 1960/1961 99%.
the performance of the test. However, the positive predictive value (the likelihood that a positive test result is true) and the negative predictive value (the likelihood that a negative test result is true) are more relevant to the tested individual. This is because both an overlooked infection and a false-positive result for a sexually transmitted infection may have personal consequences. In contrast to the terms sensitivity and speci®city, the predictive values depend on the prevalence of infection in the tested population73, and despite rather high test validity in terms of sensitivity and speci®city, the clinician should consider the predictive values when requesting a test. An example is given for a test having a sensitivity of 95% and a speci®city of 99% applied on 2000 young symptomatic women with a prevalence of infection of 10% (Table 3) and on 2000 asymptomatic pregnant women aged 35 with a prevalence of 1% (Table 4). As seen, the positive predictive value decreases dramatically with decreasing prevalence, and when testing patients with a 1% prevalence, the chance that a positive test result is incorrect is higher than it being correct. Acknowledgement This study was supported by a grant from the Danish Centre for Evaluation and Health Technology Assessment.
SUMMARY The value of serology in the diagnosis of ongoing C. trachomatis infections is limited. More patients with late complications, such as infertility and ectopic pregnancy, have antibodies against C. trachomatis. However, a substantial fraction of healthy controls also have antibodies against C. trachomatis, and up to now the predictive value of serology has not been established for clinically relevant outcomes; serology is therefore
796 L. éstergaard
Practice points . aims. To obtain a diagnosis of Chlamydia trachomatis infection . management. Be aware of the predictive value of a test result in your setting before requesting the test . investigations. The increased sensitivity of NAATs allows samples to be taken by the patient him/herself. This has major public health implications but has no diagnostic advantage over endocervical and urethral swab samples if the patient is having a pelvic examination anyway. Do not use NAATs for test of cure within 4 weeks after initiation of therapy
Research agenda . predictive values for serology testing should be assessed for clinically relevant outcomes . suciently powered comparative studies of NAATs applied on low prevalence populations are needed . with the increased use of NAATs using the plasmid as target, the number of plasmid-free strains should be surveyed . the impact in terms of infertility and ectopic pregnancy of screening programmes based on NAATs should be assessed
of limited use in the clinical decision procedure. Cell culture is still the method with the highest speci®city and it should be used in clinical cases with potential legal signi®cance. Also, cell culture should be included in extended reference standards for assessing the diagnostic validity of new tests. Diagnostic methods based on antigen detection, such as direct ¯uorescence assays and enzyme immunoassays, are less expensive than nucleic acid ampli®cation tests. However, the sensitivity of nucleic acid ampli®cation tests is higher and this allows urine and vaginal swabs to be used as samples. These samples can be obtained by the patient him/herself and mailed directly to the laboratory. Public health measures using home sampling for screening asymptomatic men and women have been shown to reduce the risk of pelvic in¯ammatory disease and the community-wide prevalence. Partner noti®cation is still important, and `test of cure' can be considered ± but only after the ®rst 4 weeks following initiation of therapy. It is crucial for the clinician to know the positive and negative predictive values of a C. trachomatis test in his/her setting. REFERENCES 1. The National Institute of Health. NIH Taxonomy Browser (http://www.ncbi.nlm.nih.gov/htbin-post/ Taxonomy). 2. Zhong GM, Reid RE & Brunham RC. Mapping antigenic sites on the major outer membrane protein of Chlamydia trachomatis with synthetic peptides. Infection and Immunity 1990; 58: 1450±1455. * 3. Campbell LA, Kuo CC & Grayston JT. Structural and antigenic analysis of Chlamydia pneumoniae. Infection and Immunity 1990; 58: 93±97. * 4. Grimwood J & Stephens RS. Computational analysis of the polymorphic membrane protein superfamily of Chlamydia trachomatis and Chlamydia pneumoniae. Microbial Comparative Genomics 1999; 4: 187±201.
Microbiological aspects of the diagnosis of C. trachomatis 797 5. Zhang YX, Morrison SG, Caldwell HD & Baehr W. Cloning and sequence analysis of the major outer membrane protein genes of two Chlamydia psittaci strains. Infection and Immunity 1989; 57: 1621±1625. 6. Yuan Y, Zhang YX, Watkins NG & Caldwell HD. Nucleotide and deduced amino acid sequences for the four variable domains of the major outer membrane proteins of the 15 Chlamydia trachomatis serovars. Infection and Immunity 1989; 57: 1040±1049. 7. Toye B, Zhong GM, Peeling R & Brunham RC. Immunologic characterization of a cloned fragment containing the species-speci®c epitope from the major outer membrane protein of Chlamydia trachomatis. Infection and Immunity 1990; 58: 3909±3913. * 8. Stephens RS, Wagar EA & Schoolnik GK. High-resolution mapping of serovar-speci®c and common antigenic determinants of the major outer membrane protein of Chlamydia trachomatis. Journal of Experimental Medicine 1988; 167: 817±831. 9. Keay SD, Barlow R, Eley A et al. The relation between immunoglobulin G antibodies to Chlamydia trachomatis and poor ovarian response to gonadotropin stimulation before in vitro fertilization. Fertility and Sterility 1998; 70: 214±218. 10. Spandorfer SD, Neuer A, LaVerda D et al. Previously undetected Chlamydia trachomatis infection, immunity to heat shock proteins and tubal occlusion in women undergoing in-vitro fertilization. Human Reproduction 1999; 14: 60±64. 11. Sharara FI, Queenan JT Jr, Springer RS et al. Elevated serum Chlamydia trachomatis IgG antibodies. What do they mean for IVF pregnancy rates and loss? Journal of Reproductive Medicine 1997; 42: 281±286. *12. Claman P, Amimi MN, Peeling RW et al. Does serologic evidence of remote Chlamydia trachomatis infection and its heat shock protein (CHSP 60) aect in vitro fertilization-embryo transfer outcome? Fertility and Sterility 1996; 65: 146±149. 13. Vinette-Leduc D, Yazdi HM, Jessamine P & Peeling RW. Reliability of cytology to detect chlamydial infection in asymptomatic women. Diagnostic Cytopathology 1997; 17: 258±261. 14. T'ang FF, Chang HL, Huang YT & Wang KC. Studies on the etiology of trachoma with special reference to isolation of the virus in chick embryo. Chinese Medical Journal 1957; 75: 429±447. *15. Ripa KT & MaÊrdh PA. Cultivation of Chlamydia trachomatis in cycloheximide-treated Mccoy cells. Journal of Clinical Microbiology 1977; 6: 328±331. 16. Stamm WE, Tam M, Koester M & Cles L. Detection of Chlamydia trachomatis inclusions in Mccoy cell cultures with ¯uorescein-conjugated monoclonal antibodies. Journal of Clinical Microbiology 1983; 17: 666±668. 17. Phillips LE, Smith PB, Riddle GD et al. Ortho enzyme immunoassay versus McCoy cell monolayers stained by iodine or ¯uorescent antibody for detection of Chlamydia trachomatis. Journal of Clinical Microbiology 1990; 28: 1647±1648. 18. Suchland KL, Counts JM & Stamm WE. Laboratory methods for detection of Chlamydia trachomatis: survey of laboratories in Washington State. Journal of Clinical Microbiology 1997; 35: 3210±3214. 19. éstergaard L & Mùller JK. Use of PCR and direct immuno¯uorescence microscopy for con®rmation of results obtained by Syva MicroTrak Chlamydia enzyme immunoassay. Journal of Clinical Microbiology 1995; 33: 2620±2623. 20. Mùller JK, éstergaard L & Hansen JT. Clinical evaluation of four non-related techniques for detection of Chlamydia trachomatis in endocervical specimens. Immunology and Infectious Diseases 1994; 4: 191±196. 21. Kellogg JA, Seiple JW & Hick ME. Cross-reaction of clinical isolates of bacteria and yeasts with the chlamydiazyme test for chlamydial antigen, before and after use of a blocking reagent. American Journal of Clinical Pathology 1992; 97: 309±312. 22. Hammerschlag MR, Rettig PJ & Shields ME. False positive results with the use of chlamydial antigen detection tests in the evaluation of suspected sexual abuse in children. Pediatric Infectious Disease Journal 1988; 7: 11±14. *23. Black CM. Current methods of laboratory diagnosis of Chlamydia trachomatis infections. Clinical Microbiological Review 1997; 10: 160±184. 24. Thejls H, Gnarpe J, Gnarpe H et al. Expanded gold standard in the diagnosis of Chlamydia trachomatis in a low prevalence population: diagnostic ecacy of tissue culture, direct immuno¯uorescence, enzyme immunoassay, PCR and serology. Genitourinary Medicine 1994; 70: 300±303. 25. Quinn TC, War®eld P, Kappus E et al. Screening for Chlamydia trachomatis infection in an inner-city population: a comparison of diagnostic methods. Journal of Infectious Diseases 1985; 152: 419±423. 26. éstergaard L, Birkelund S & Christiansen G. Use of polymerase chain reaction for detection of Chlamydia trachomatis. Journal of Clinical Microbiology 1990; 28: 1254±1260. 27. Claas HC, Melchers WJ, de Bruijn IH et al. Detection of Chlamydia trachomatis in clinical specimens by the polymerase chain reaction. European Journal of Clinical Microbiology and Infectious Diseases 1990; 9: 864±868.
798 L. éstergaard 28. Mahony JB, Luinstra KE, Jang D et al. Chlamydia trachomatis con®rmatory testing of PCR-positive genitourinary specimens using a second set of plasmid primers. Molecular and Cellular Probes 1992; 6: 381±388. 29. NaÈher H, Drzonek H, Wolf J et al. Detection of C. trachomatis in urogenital specimens by polymerase chain reaction. Genitourinary Medicine 1991; 67: 211±214. 30. Ossewaarde JM, Riee M, Rozenberg Arska M et al. Development and clinical evaluation of a polymerase chain reaction test for detection of Chlamydia trachomatis. Journal of Clinical Microbiology 1992; 30: 2122±2128. 31. Ratti G, Moroni A & Cevenini R. Detection of Chlamydia trachomatis DNA in patients with nongonococcal urethritis using the polymerase chain reaction. Journal of Clinical Pathology 1991; 44: 564±568. 32. Holland SM, Gaydos CA & Quinn TC. Detection and dierentiation of Chlamydia trachomatis, Chlamydia psittaci, and Chlamydia pneumoniae by DNA ampli®cation. Journal of Infectious Diseases 1990; 162: 984±987. 33. Bobo L, Coutlee F, Yolken RH et al. Diagnosis of Chlamydia trachomatis cervical infection by detection of ampli®ed DNA with an enzyme immunoassay. Journal of Clinical Microbiology 1990; 28: 1968±1973. 34. Claas HC, Melchers WJ, de Bruijn IH et al. Detection of Chlamydia trachomatis in clinical specimens by the polymerase chain reaction. European Journal of Clinical Microbiology and Infectious Diseases 1990; 9: 864±868. 35. Mouton JW, Verkooyen R, van der Meijden WI et al. Detection of Chlamydia trachomatis in male and female urine specimens by using the ampli®ed Chlamydia trachomatis test. Journal of Clinical Microbiology 1997; 35: 1369±1372. 36. Mahony J, Chong S, Jang D et al. Urine specimens from pregnant and nonpregnant women inhibitory to ampli®cation of Chlamydia trachomatis nucleic acid by PCR, ligase chain reaction, and transcriptionmediated ampli®cation: identi®cation of urinary substances associated with inhibition and removal of inhibitory activity. Journal of Clinical Microbiology 1998; 36: 3122±3126. 37. Sriprakash KS & Macavoy ES. Characterization and sequence of a plasmid from the trachoma biovar of Chlamydia trachomatis. Plasmid 1987; 18: 205±214. 38. Palmer L & Falkow S. A common plasmid of Chlamydia trachomatis. Plasmid 1986; 16: 52±62. 39. Roosendaal R, Walboomers JM, Veltman OR et al. Comparison of dierent primer sets for detection of Chlamydia trachomatis by the polymerase chain reaction. Journal of Medical Microbiology 1993; 38: 426±433. 40. Ossewaarde JM, Riee M, Rozenberg-Arska M et al. Development and clinical evaluation of a polymerase chain reaction test for detection of Chlamydia trachomatis. Journal of Clinical Microbiology 1992; 30: 2122±2128. 41. Martin JL, Alexander SY, Selwood TS & Cross GF. Use of the polymerase chain reaction for the detection of Chlamydia trachomatis in clinical specimens and its comparison to commercially available tests. Genitourinary Medicine 1995; 71: 169±171. 42. Mahony JB, Luinstra KE, Sellors JW & Chernesky MA. Comparison of plasmid- and chromosome-based polymerase chain reaction assays for detecting Chlamydia trachomatis nucleic acids. Journal of Clinical Microbiology 1993; 31: 1753±1758. *43. An Q, Radclie G, Vassallo R et al. Infection with a plasmid-free variant Chlamydia related to Chlamydia trachomatis identi®ed by using multiple assays for nucleic acid detection. Journal of Clinical Microbiology 1992; 30: 2814±2821. 44. Schachter J, Stamm WE & Quinn TC. Discrepant analysis and screening for Chlamydia trachomatis. Lancet 1996; 348: 1308±1309. 45. Fahr MJ, Sriprakash KS & Hatch TP. Convergent and overlapping transcripts of the Chlamydia trachomatis 7.5-kb plasmid. Plasmid 1992; 28: 247±257. 46. Hatt C, Ward ME & Clarke IN. Analysis of the entire nucleotide sequence of the cryptic plasmid of Chlamydia trachomatis serovar L1. Evidence for involvement in DNA replication. Nucleic Acids Research 1988; 16: 4053±4067. 47. Peterson EM, Marko BA, Schachter J & de la Maza LM. The 7.5-kb plasmid present in Chlamydia trachomatis is not essential for the growth of this microorganism. Plasmid 1990; 23: 144±148. 48. Lan J, Walboomers JM, Roosendaal R et al. Direct detection and genotyping of Chlamydia trachomatis in cervical scrapes by using polymerase chain reaction and restriction fragment length polymorphism analysis. Journal of Clinical Microbiology 1993; 31: 1060±1065. 49. Frost EH, Deslandes S & Bourgaux Ramoisy D. Sensitive detection and typing of Chlamydia trachomatis using nested polymerase chain reaction. Genitourinary Medicine 1993; 69: 290±294. 50. Rodriguez P, Vekris A, de Barbeyrac B et al. Typing of Chlamydia trachomatis by restriction endonuclease analysis of the ampli®ed major outer membrane protein gene. Journal of Clinical Microbiology 1991; 29: 1132±1136.
Microbiological aspects of the diagnosis of C. trachomatis 799 51. van de Laar MJ, van Duynhoven YT, Fennema JS et al. Dierences in clinical manifestations of genital chlamydial infections related to serovars. Genitourinary Medicine 1996; 72: 261±265. 52. Yang CL, Maclean I & Brunham RC. DNA sequence polymorphism of the Chlamydia trachomatis omp1 gene. Journal of Infectious Diseases 1993; 168: 1225±1230. 53. Lan J, Melgers I, Meijer CJ et al. Prevalence and serovar distribution of asymptomatic cervical Chlamydia trachomatis infections as determined by highly sensitive PCR. Journal of Clinical Microbiology 1995; 33: 3194±3197. 54. Sayada C, Denamur E, Or®la J et al. Rapid genotyping of the Chlamydia trachomatis major outer membrane protein by the polymerase chain reaction. FEMS Microbiology Letters 1991; 67: 73±78. 55. Black CM, Tharpe JA & Russell H. Distinguishing Chlamydia species by restriction analysis of the major outer membrane protein gene. Molecular and Cellular Probes 1992; 6: 395±400. 56. Gaydos CA, Bobo L, Welsh L et al. Gene typing of Chlamydia trachomatis by polymerase chain reaction and restriction endonuclease digestion. Sexually Transmitted Diseases 1992; 19: 303±308. *57. Frost EH, Deslandes S, Veilleux S et al. Typing Chlamydia trachomatis by detection of restriction fragment length polymorphism in the gene encoding the major outer membrane protein. Journal of Infectious Diseases 1991; 163: 1103±1107. 58. Poole E & Lamont I. Chlamydia trachomatis serovar dierentiation by direct sequence analysis of the variable segment 4 region of the major outer membrane protein gene. Infection and Immunity 1992; 60: 1089±1094. 59. Viscidi RP, Bobo L, Hook EW & Quinn TC. Transmission of Chlamydia trachomatis among sex partners assessed by polymerase chain reaction. Journal of Infectious Diseases 1993; 168: 488±492. 60. Frost EH, Deslandes S & Bourgaux Ramoisy D. Chlamydia trachomatis serovars in 435 urogenital specimens typed by restriction endonuclease analysis of ampli®ed DNA. Journal of Infectious Diseases 1993; 168: 497±501. 61. Hayes LJ, Pecharatana S, Bailey RL et al. Extent and kinetics of genetic change in the omp1 gene of Chlamydia trachomatis in two villages with endemic trachoma. Journal of Infectious Diseases 1995; 172: 268±272. 62. Quinn TC, Gaydos C, Shepherd M et al. Epidemiologic and microbiologic correlates of Chlamydia trachomatis infection in sexual partnerships. Journal of the American Medical Association 1996; 276: 1737±1742. 63. Pedersen LN, Kjaer HO, Mùller JK et al. High-resolution genotyping of Chlamydia trachomatis from recurrent urogenital infections. Journal of Clinical Microbiology 2000; 38: 3068±3071. 64. van Duynhoven YT, Ossewaarde JM, Derksen-Nawrocki RP et al. Chlamydia trachomatis genotypes: correlation with clinical manifestations of infection and patients' characteristics. Clinical Infectious Diseases 1998; 26: 314±322. *65. éstergaard L. Diagnosis of urogenital Chlamydia trachomatis infection by use of DNA ampli®cation. APMIS 1999; 89 (supplement): 5±36. 66. Kellogg JA, Seiple JW, Klinedinst JL et al. Improved PCR detection of Chlamydia trachomatis by using an altered method of specimen transport and high-quality endocervical specimens. Journal of Clinical Microbiology 1995; 33: 2765±2767. 67. Kellogg JA, Seiple JW, Klinedinst JL & Stroll E. Di-Quik stain as a simpli®ed alternative to Papanicolaou stain for determination of quality of endocervical specimens submitted for PCR detection of Chlamydia trachomatis. Journal of Clinical Microbiology 1996; 34: 2590±2592. 68. éstergaard L, Mùller JK, Andersen B & Olesen F. Diagnosis of urogenital Chlamydia trachomatis infection in women based on mailed samples obtained at home: multipractice comparative study. British Medical Journal 1996; 313: 1186±1189. 69. Andersen B, Olesen F, Mùller JK & éstergaard L. Population-based strategies for outreach screening of urogenital Chlamydia trachomatis infections: a randomized, controlled trial. Journal of Infectious Diseases 2002; 185: 252±258. 70. Andersen B, éstergaard L, Mùller JK & Olesen F. Home sampling versus conventional contact tracing for detecting Chlamydia trachomatis in male partners of infected women: randomised study. British Medical Journal 1998; 316: 350±351. 71. éstergaard L, Andersen B, Olesen F & Mùller JK. Ecacy of home sampling for screening of Chlamydia trachomatis: randomised study. British Medical Journal 1998; 317: 26±27. *72. éstergaard L, Andersen B, Mùller JK & Olesen F. Home sampling versus conventional swab sampling for screening of Chlamydia trachomatis in women: a cluster-randomized 1-year follow-up study. Clinical Infectious Diseases 2000; 31: 951±957. 73. éstergaard L. Detection of Chlamydia trachomatis in a low prevalence population. European Journal of Clinical Microbiology and Infectious Diseases 1995; 14: 471±472.