Role of Confirmatory PCRs in Determining Performance of

Vol. 32, No. 10

JOURNAL OF CLINICAL MICROBIOLOGY, OCt. 1994, p. 2490-2493

0095-1137/94/$04.00+0 Copyright ©) 1994, American Society for Microbiology

Role of Confirmatory PCRs in Determining Performance of Chlamydia Amplicor PCR with Endocervical Specimens from Women with a Low Prevalence of Infection JAMES B. MAHONY,l12* KATHLEEN E. LUINSTRA,1 JOHN W. SELLORS,' 3,4 LAURA PICKARD,' SYLVIA CHONG,' DAN JANG,1 AND MAX A. CHERNESKY" 2'5 McMaster University Regional Virology and Chlamydiology Laboratory, St. Joseph's Hospital,1 and the Departments of Pathology,2 Pediatrics,5SFamily Medicine,3 and Clinical Epidemiology and Biostatistics,4 McMaster University, Hamilton, Ontario, Canada Received 8 April 1994/Returned for modification 17 May 1994/Accepted 8 July 1994

The role of confirmatory PCR assays for determining the performance of Chlamydia Amplicor PCR for endocervical specimens from women with a low prevalence of infection was evaluated. An endocervical swab was collected from 770 women and tested by culture or direct fluorescent antibody (DFA) staining. A second swab was tested by Chlamydia Amplicor PCR (Roche Molecular Systems, Branchburg, NJ.). Discordant results were resolved by three confirmatory PCRs: one targeting the plasmid by using different primers and two directed to the major outer membrane protein (MOMP) gene. Of the 30 swabs that were positive by culture or DFA (3.9%), 27 were positive by Amplicor PCR. An additional five swabs were positive by Amplicor PCR but negative by culture or DFA. Both plasmid and MOMP confirmatory PCRs identified the five culture-DFA negatives and the three Amplicor negatives as true positives. The three specimens originally classified as negative by Amplicor PCR were positive on repeat Amplicor testing. After resolution of the discordant results by confirmatory PCR testing, the sensitivity of the initial Amplicor PCR was 91.4% (32 of 35 specimens), changing to 100% after storage and repeat testing. The specificity of Amplicor PCR was 100% (735 of 735 specimens). Our results demonstrated that plasmid and MOMP confirmatory PCRs worked equally well in resolving false-positive and false-negative Amplicor PCR results. Some specimens may contain inhibitors of Amplicor PCR which may disappear with time.

Chlamydia trachomatis infections have historically been diagnosed by isolation in cell culture, enzyme immunoassay (EIA), direct fluorescent antibody (DFA), and, more recently, direct DNA hybridization and PCR. PCR has been developed by several laboratories (4-8, 10-19, 21-23) and has been shown to have sensitivity equal to or greater than that of culture and EIA for a variety of clinical specimens including male urethral swabs, female endocervical swabs, and first-void urine specimens (2, 4, 8, 11-13, 15-19, 23). A commercially available PCR, Amplicor Chlamydia (Roche Molecular Systems, Branchburg, N.J.), has been developed (10) and recently approved by the Food and Drug Administration. Two reports have appeared evaluating the performance of Amplicor PCR with endocervical specimens from women in populations with a 8-10% prevalence of infection (1, 3). In a previous study evaluating PCR for the detection of C. trachomatis in first-void urine specimens from men we introduced the use of confirmatory PCR testing with a second pair of plasmid primers to resolve discordant specimens (11, 13). Use of a plasmid-based confirmatory PCR improved the specificity of PCR and the agreement of results obtained between laboratories in an interlaboratory agreement study (11, 13, 14). We now extend these studies and evaluate the role of plasmid and major outer membrane protein (MOMP) confirmatory PCRs in determining the performance of Amplicor PCR on endocervical specimens from women with a low prevalence of infection.

* Corresponding author. Mailing address: Division of Virology, St. Joseph's Hospital, 50 Charlton Avenue East, Hamilton, Ontario L8N 4A6, Canada. Phone: (905) 521-6021. Fax: (905) 521-6083.

2490

MATERIALS AND METHODS Specimens. From January to December 1993, a total of 770 women were enrolled in the study from two hospital gynecology clinics (n = 676), a college student health clinic (n = 61), and a sexually transmitted disease clinic (n = 33). After giving informed consent (approval for the study was obtained by the McMaster University Ethics Review Committee), the patient was given a standardized questionnaire for the collection of information on demographics, sexual behavior, history, and symptoms. Two endocervical swabs were collected in randomized order; one was collected with a Dacron swab and put into chlamydia transport medium (9) for culture, while a second was put into the Amplicor collection tube for PCR. Specimens were held at 4°C until they were transported to the laboratory usually within 3 h of collection. A small number of specimens collected at the end of the day were held at 4°C overnight or frozen at -70°C until they could be processed the next working day. Culture or DFA. Endocervical swab specimens were inoculated onto McCoy cell monolayers in 96-well microculture plates with a blind passage as described previously (9). Five original specimens that were toxic in culture were stained for elementary bodies by centrifuging the specimens (0.2 to 0.4 ml) for 15 min at 14,000 rpm in a microcentrifuge, resuspending the pellet in 50 to 100 ,ul of phosphate-buffered saline, and staining with a species-specific fluorescein isothiocyanate-labelled anti-MOMP monoclonal antibody (Syva; Microtrak). A specimen was considered positive if two or more elementary bodies were observed. PCR. Amplicor Chlamydia is a plasmid-based PCR assay for the detection of C. trachomatis DNA in genital tract specimens (10). The test employs biotinylated plasmid primers CP24-

AMPLICOR PCR PERFORMANCE IN LOW-PREVALENCE POPULATION

VOL. 32, 1994

TABLE 1. Confirmatory PCR results obtained for discordant endocervical specimens from eight women PCR result Patient

Culture-DFA result

H064 B50 B200 B205 3999R H071 H077 B212

+ +

+

Amplicor

Confirmatory

(initial)

Plasmid KL1-KL2

MOMP

MOMP

C`T1-C1T2

CT0005-06

+ + + + +

+

NDa

+

+b

+ +

+b

-c -c -c

+ + + + + +

+ ND ND ND +

+ + + + + +

a ND, not done because specimen was insufficient. b Specimen B50 required dilution of 1:8 to become positive by both confirmatory PCRs. c Specimens H071, H077, and B212, initially tested negative by Amplicor PCR, were positive in a repeat Amplicor test after storage for 2 weeks at 4°C.

CP27 and was performed according to the manufacturer's instructions. Specimens with an A450 above 0.5 were considered positive, and those with an A450 below 0.2 were considered negative. Specimens with an A450 between 0.2 and 0.5 were called equivocal, and the readings were repeated in duplicate. If two out of three absorbance readings were below the adjusted A450 cutoff of 0.25, the specimen was considered negative: Confirmatory PCR. Discordant specimens were tested by three different PCRs using KL1-KL2 plasmid primers (13) and MOMP primers CT1-CT2 described by Dutilh et al. (7) and CT0005-06 described by Bobo et al. (4). RESULTS The mean age of the 770 women was 24.9 years, and 324 (42.1%) had signs or symptoms of endocervicitis (frequency; dysuria; yellow, green, or opaque discharge; or intermenstrual bleeding), while the rest were without symptoms or signs of endocervicitis. Specimens from 30 women were positive by culture or DFA, for an overall prevalence of 3.9%. Twenty-seven of these specimens were also positive by initial Amplicor PCR testing. There were a total of eight specimens with discordant results (Table 1). The five which were negative by culture or DFA were confirmed as positive by confirmatory plasmid and MOMP PCRs. The three specimens that were missed on initial Amplicor testing were positive in both confirmatory PCRs and on repeat Amplicor PCR testing after storage at 4°C for 14 days. Table 2 summarizes the performance of the initial Amplicor test and compares it to diagnosis using culture or DFA after expanding the standard of positivity. This was done by incorporating into the denominator results for specimens positive in

2491

the confirmatory PCR tests. Initial Amplicor PCR testing raised the Chlamydia prevalence from 3.9% (determined by culture or DFA) to 4.2%. After discordant results were resolved with confirmatory PCRs, the sensitivity of the initial Amplicor test was 91.4% (32 of 35 specimens) and the specificity was 100% (735 of 735 specimens). The positive and negative predictive values for the initial Amplicor PCR test were 100% (32 of 32 specimens) and 99.6% (735 of 738 specimens), respectively. Repeat Amplicor testing after storage identified the original three negative specimens as positive. The sensitivity and specificity of the Amplicor PCR were similar when the data were analyzed according to symptoms or signs recorded on the questionnaires (data not shown). DISCUSSION The performance of the Amplicor PCR on endocervical specimens from women in a low-prevalence population was determined by testing one swab by culture or DFA and a second swab by Amplicor and then resolving discordants with a testing algorithm that included more than one confirmatory PCR. One of the confirmatory tests used a different pair of plasmid primers, while the others amplified a MOMP gene fragment. In this study of 770 Canadian women with a prevalence of C. trachomatis infection of 3.9% by culture or DFA, Amplicor PCR had a good sensitivity (91.4%) and excellent specificity (100%). The prevalence rates were unaffected by recording of symptoms or signs of endocervicitis (data not shown). Bass et al. (1) prospectively examined 645 women with a prevalence of infection of 7.8% by culture and reported an Amplicor sensitivity of 96% and specificity of 100%. Bauwens et al. (3) evaluated 587 women (culture prevalence of 10%) with a prototype kit and found a sensitivity of 89% for Amplicor versus 93% for culture. In the Bauwens study the swabs were left in the tubes during transportation to the laboratory, which may have contributed to the lower sensitivity, since Loeffelholz et al. (10) showed that leaving the swab in the tube for an extended period of time decreases the sensitivity of Amplicor, possibly because of leaching of inhibitors of Taq polymerase from the swabs. In a second group of 362 women, Bauwens et al. (3) used a revised Amplicor kit and found its sensitivity to be 60%, compared with that of culture. The specificity was 99%. They attributed the low sensitivity of Amplicor to the following possibilities: the order of swabs was not randomized; they collected a cytobrush for culture first, followed by a second swab for Amplicor; and the manufacturer's swab was not used to collect the specimen for PCR. In the only other published evaluation of Amplicor in women, Smith et al. (20) tested 157 stored endocervical specimens (19.8% positive by culture) and reported a sensitivity of 98.4% for Amplicor and 79% for culture. The specificities were 99.5 and 100%, respectively. Four of the five Amplicor-positive, discordant specimens in our study were positive by both plasmid and MOMP PCR confirmatory tests when tested undiluted. Patient specimen

TABLE 2. Performance of culture or DFA compared with Amplicor PCR for detection of C. trachomatis in endocervical specimens from 770 womena

PPV (%)b Specificity (%) Sensitivity (%) 100 (735/735) 100 (30/30) 85.7 (30/35) 3.9 Culture or DFA 100 (32/32) 91.4 (32/35) 100 (735/735) 4.2 Amplicor PCR confirmed and a Calculations based on resolved test results using an expanded standard of positivity by culture, DFA, Amplicor PCR. b PPV, positive predictive value; NPV, negative predictive value. Assay

Prevalence (%)

NPV (%)b 99.3 (735/740) 99.6 (735/738)

2492

MAHONY ET AL.

B50 (Table 1) required a dilution to 1:8 to be positive in both confirmatory PCRs, suggesting the presence of inhibitors for both confirmatory assays in these specimens. Repeat Amplicor PCR testing also achieved a negative result, but the specimen became positive after dilution of the specimen. This specimen was toxic in cell culture and negative by DFA and would have been scored as an Amplicor false positive if the confirmatory PCRs were not performed with various dilutions of the specimen. The reason why the initial Amplicor PCR test was positive is unclear. One explanation could be that inhibitors were absent or present at low levels the first time the Amplicor test was performed and that during storage or freeze-thawing the inhibitor levels increased, possibly as a result of autolysis or apoptosis of leukocytes or endothelial cells releasing degradative enzymes. Bauwens et al. (3) have reported specimens changing from Amplicor negative to positive after a few days of storage and suggested that some specimens may contain thermolabile inhibitors that are destroyed with storage. Three of our specimens (H071, H077, and B212), initially negative by Amplicor PCR, were positive in a repeat test 2 weeks later. The effects of storage on PCR results for clinical specimens are not well understood. It is possible that inhibitors in the form of degradative enzymes such as proteases could accumulate with storage or freeze-thawing, resulting in negative PCR results, and then over time become inactivated, resulting in positive PCR results. The occurrence and kinetics of these two processes could be dependent on the nature of the cells present in specific specimens. In an Amplicor PCR evaluation performed by Bass et al. (1), three Amplicor-positive, culture-negative specimens could be confirmed only after phenol-chloroform extraction of DNA. Both Bass et al. (1) and Bauwens et al. (3) reported Amplicor PCR negative results that became positive after phenol-chloroform extraction of DNA. Future studies will be required to elucidate the chemical nature and stability of PCR inhibitors and to better understand what effect they might have on the performance of amplification technologies. Supplementary PCR testing using different plasmid primers or MOMP gene fragment amplification performed equally well in our small series. Bauwens et al. (3) had one discordant specimen that was plasmid PCR negative but MOMP PCR positive on extracted DNA. This difference may have been due to the 40 cycles of amplification used with the confirmatory MOMP primers versus the 30 cycles used for Amplicor PCR. Plasmid-based PCRs are usually considered to be more sensitive because of multiple copies of the plasmid. The increased sensitivity of plasmid PCRs over MOMP or rRNA PCRs has recently been demonstrated (12). Primer differences may also explain why Bass et al. (1) were unable to confirm one of their Amplicor-positive specimens. The inclusion by the manufacturer of either a pretreatment step to remove inhibitors or an internal PCR control to identify specimens that may be falsely negative should improve the performance of Amplicor PCR. Until this issue is resolved, evaluations of nucleic acid amplification assays, because they are more sensitive than existing methods such as culture, EIA, and DFA, will identify a small number of specimens that are positive only by amplification. These positive results should be confirmed in the case of PCR with primers for a different fragment of the same gene or a different gene. Future studies with larger numbers of discordants may show differences in confirmatory PCR results for initial apparent false positives or false negatives which may be due to selective effects of inhibitors present in clinical specimens.

J. CLIN. MICROBIOL.

ACKNOWLEDGMENTS We gratefully acknowledge the physicians (C. Devlin, R. Stopps, and L. Grant) and nursing staff at McMaster University Medical Centre 4B and 4F Clinics; the physicians (J. Lamont, K. Lamont, D. Pond, and P. Roth) and nursing staff at the Surgical Center, Henderson Division, Hamilton Civics Hospital; the physicians (M. Bridge and J. MacDonald) and staff at Mohawk College Student Health Clinic; and the staff at the Hamilton Civics Hospital General Division STD Clinic. The secretarial assistance of Lisa Rizzo and Luci Weatherley is gratefully acknowledged.

REFERENCES 1. Bass, C. A., D. L. Jungkind, N. S. Silverman, and J. M. Bond. 1993. Clinical evaluation of a new polymerase chain reaction assay for detection of Chlamydia trachomatis in endocervical specimens. J. Clin. Microbiol. 31:2648-2653. 2. Bauwens, J. E., A. M. Clark, M. J. Loeffelholz, S. A. Herman, and W. E. Stamm. 1993. Diagnosis of Chlamydia trachomatis urethritis in men by polymerase chain reaction assay of first-catch urine. J. Clin. Microbiol. 31:3013-3016. 3. Bauwens, J. E., A. M. Clark, and W. E. Stamm. 1993. Diagnosis of Chlamydia trachomatis endocervical infections by a commercial polymerase chain reaction assay. J. Clin. Microbiol. 31:3023-3027. 4. Bobo, L., F. Coultee, R. Yolken, T. Quinn, and R. P. Viscidi. 1990. Diagnosis of Chlamydia trachomatis cervical infection by detection of amplified DNA with an enzyme immunoassay. J. Clin. Microbiol. 28:1968-1973. 5. Claas, H. C., W. J. Melchers, I. H. de Bruijn, M. de Graaf, W. C. van Dijk, J. Lindeman, and W. G. Quint. 1990. Detection of Chlamydia trachomatis in clinical specimens by the polymerase chain reaction. Eur. J. Clin. Microbiol. Infect. Dis. 9:864-868. 6. Claas, H. C. J., J. H. T. Wagenvoort, H. G. M. Niesters, T. T. Tio, J. H. van Risoort-Vos, and W. G. V. Quint. 1991. Diagnostic value of the polymerase chain reaction for chlamydia detection as determined in a follow-up study. J. Clin. Microbiol. 29:42-45. 7. Dutilh, B., C. Bebear, D. Taylor-Robinson, P. Rodriquez, A. Vekris, J. Bonnet, and M. Garret. 1989. Specific amplification of a DNA sequence common to all Chlamydia trachomatis serovars using the polymerase chain reaction. Res. Microbiol. 140:7-16. 8. Griffais, R., and M. Thibon. 1989. Detection of Chlamydia trachomatis by the polymerase chain reaction. Res. Microbiol. 140:139141. 9. Jang, D., J. W. Sellors, J. B. Mahony, L. Pickard, and M. A. Chernesky. 1992. Effects of broadening the gold standard on the performance of a chemiluminometric immunoassay to detect Chlamydia trachomatis antigens in centrifuged first void urine and urethral swab samples from men. Sex. Transm. Dis. 19:315-319. 10. Loeffelholz, M. J., C. A. Lewinski, S. R. Silver, A. P. Purohit, S. A. Herman, D. A. Buonagurio, and E. A. Dragon. 1992. Detection of Chlamydia trachomatis in endocervical specimens by polymerase chain reaction. J. Clin. Microbiol. 30:2847-2851. 11. Mahony, J. B., K. E. Luinstra, D. Jang, J. W. Sellors, and M. A. Chernesky. 1992. Chlamydia trachomatis confirmatory testing of PCR-positive genitourinary specimens using a second set of plasmid primers. Mol. Cell. Probes 6:381-388. 12. Mahony, J. B., K. E. Luinstra, J. W. Sellors, and M. A. Chernesky. 1993. Comparison of plasmid- and chromosome-based polymerase chain reaction assays for detecting Chlamydia trachomatis nucleic acid. J. Clin. Microbiol. 31:1753-1758. 13. Mahony, J. B., K. E. Luinstra, J. W. Sellors, D. Jang, and M. A. Chernesky. 1992. Confirmatory polymerase chain reaction testing for Chlamydia trachomatis in first-void urine from asymptomatic and symptomatic men. J. Clin. Microbiol. 30:2241-2245. 14. Mahony, J. B., K. E. Luinstra, J. Waner, G. McNab, H. Hobranzska, D. Gregson, J. W. Sellors, and M. A. Chernesky. 1994. Interlaboratory agreement study of a double set of PCR plasmid primers for detection of Chlamydia trachomatis in a variety of genitourinary specimens. J. Clin. Microbiol. 32:87-91. 15. Naher, H., H. Drzonek, J. Wolf, M. von Knebel Doeberitz, and D. Petzoldt. 1991. Detection of C. trachomatis in urogenital specimens by polymerase chain reaction. Genitourin. Med. 657:211214. 16. Ossewaarde, J. M., M. Rieffe, M. Rozenberg-Arska, P. M. Ossen-

VOL. 32, 1994

17.

18.

19.

20.

AMPLICOR PCR PERFORMANCE IN LOW-PREVALENCE POPULATION

koppele, R. P. Nawrocki, and A. M. van Loon. 1992. Development and clinical evaluation of a polymerase chain reaction test for detection of Chlamydia trachomatis. J. Clin. Microbiol. 30:21222128. Ostergaard, L., J. Traulsen, S. Birkelund, and G. Christiansen. 1991. Evaluation of urogenital Chlamydia trachomatis infections by cell culture and the polymerase chain reaction using a closed system. Eur. J. Clin. Microbiol. Infect. Dis. 10:1057-1061. Palmer, H. M., C. B. Gilroy, B. J. Thomas, P. E. Hay, C. Gilchrist, and D. Taylor-Robinson. 1991. Detection of Chlamydia trachomatis by the polymerase chain reaction in swabs and urine from men with non-gonococcal urethritis. J. Clin. Pathol. 44:321-325. Ratti, G., A. Moroni, and R. Cevenini. 1991. Detection of Chlamydia trachomatis DNA in patients with non-gonococcal urethritis using polymerase chain reaction. J. Clin. Pathol. 44:564-568. Smith, L. W., C. L. Morrison, C. Patrizio, and A. McMillan. 1993.

2493

Use of a commercial PCR kit for detecting Chlamydia trachomatis. J. Clin. Pathol. 46:822-825. 21. Watson, M. W., P. R. Lambden, and L N. Clarke. 1991. Genetic diversity and identification of human infection by amplification of the chlamydial 60-kilodalton cysteine-rich outer membrane protein gene. J. Clin. Microbiol. 29:1188-1193. 22. Williams, T. W., S. D. Tyler, S. Giercke, D. R Pollard, P. McNicol, and K. R. Rozee. 1992. Comparison of polymerase chain reaction and Chlamydiazyme for the detection of Chlamydia trachomatis in clinical specimens. Eur. J. Clin. Microbiol. Infect. Dis. 11:233-236. 23. Wu, C., M. Lee, S. Yin, D. Yang, and S. F. Cheng. 1992. Comparison of polymerase chain reaction, monoclonal antibody based enzyme immunoassay, and cell culture for detection of Chiamydia trachomatis in genital specimens. Sex. Transm. Dis. 19:193-197.