INTRODUCTION

Top Abstract Introduction Materials and methods Results Discussion References

Asymptomatic gonococcal infections contribute to the persistence and transmission of gonorrhea in a community (6). Asymptomatic infection occurs in both males (3) and females (2). One approach to eliminating sexually transmitted diseases (STDs) in a community is to screen high-risk persons, followed by the treatment and education of people who test positive. This has been difficult to achieve in the past because of the low sensitivity of diagnostic tests for asymptomatic infections and the need for invasive specimen collection procedures, which results in reduced patient compliance with testing. Nucleic acid amplification tests, which have increased sensitivities over those of conventional tests, and the option to use urine as a specimen provide valuable tools in the screening of asymptomatic persons for STDs.

Recently, in Australia, as part of the National Indigenous Persons Sexual Health Strategy 1997-1999, funds have been made available to encourage voluntary testing for STDs by noninvasive diagnostic techniques. In Queensland, screening of indigenous populations for Chlamydia trachomatis and Neisseria gonorrhoeae using the COBAS AMPLICOR C. trachomatis/N. gonorrhoeae PCR with urine specimens has begun. This duplex PCR has the advantage of simultaneously screening for both pathogens in one specimen in an automated system. Although the reliability of the AMPLICOR C. trachomatis PCR has been evaluated (both locally [4] and internationally) and it has been found to be highly suitable for screening of asymptomatic subjects, few data on the reliability of the AMPLICOR N. gonorrhoeae PCR exist.

Of greatest concern is that specimens infected with wild-type strains of Neisseria subflava and Neisseria cinerea have been reported to have positive results in the AMPLICOR N. gonorrhoeae PCR (9). Although these organisms are considered commensal organisms of the upper respiratory tract, the possibility of transient carriage in the genital tract cannot be overruled. In addition, the normal flora of the local indigenous population has never been investigated.

In this study, the sensitivity and specificity of the COBAS AMPLICOR N. gonorrhoeae PCR test are compared to those of two other PCR tests with three geographically diverse strains of N. gonorrhoeae, Australian strains of other Neisseria spp., and AMPLICOR N. gonorrhoeae PCR-positive and N. gonorrhoeae-negative specimens selected from population with a high prevalence (~9%) of N. gonorrhoeae infection. The study shows that some local strains of N. subflava cross-react with this test and that some clinical specimens exhibit false-positive reactions. In addition, although the AMPLICOR N. gonorrhoeae PCR is shown to be an effective screening test, confirmation of positive results by an alternate PCR method is worthy of serious consideration.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and methods Results Discussion References

Patient specimens. Specimens from patients under investigation for N. gonorrhoeae infection (symptomatic or asymptomatic screening) in a population with a high prevalence (~9%) of N. gonorrhoeae infection were used. Urethral, cervical, and urine (first-pass) specimens were transported, stored, and treated as described in the AMPLICOR kit insert. High-vaginal and vaginal specimens were also collected and were handled identically to cervical specimens because the kit insert does not provide specific instructions for vaginal specimens.

Bacterial strains. Tables 1 and 2 list the strains used in this study. The three N. gonorrhoeae strains used were chosen because they are from different geographical locations and different patients. Crude DNA lysates were prepared as follows: a 1-µl loopful of a colony from an 18- to 24-h culture was placed in 3 ml of sterile distilled water, and the mixture was vortexed until the colonies were dispersed. Crude DNA isolates were sent from other laboratories on dry ice. A 50-µl aliquot was added to 500 µl of AMPLICOR lysis reagent, and the mixture was vortexed for 20 s. After 10 min of incubation at room temperature, 500 µl of AMPLICOR specimen diluent was added, followed by vortexing for 20 s and then a 10-min incubation before further processing. Crude DNA was either amplified immediately or stored at 20°C until amplification. Purified DNA was obtained by a rapid DNA extraction method with the IsoQuick Nucleic Acid Extraction Kit (Progen, Brisbane, Australia), as described in the kit insert. The DNA was quantified with a GeneQuant II calculator (Pharmacia Biotech, Sydney, Australia). All DNA dilutions were performed with sterile, nuclease-free water (Amresco, Solon, Ohio). All tests with bacterial strains were repeated at least twice with samples from new subcultures.

PCR methods. All specimens with positive and equivocal results by PCR were retested immediately by using the unprocessed primary specimen before the result for that specimen was included as a true positive or equivocal. The results obtained by the AMPLICOR PCR were interpreted as described in the kit insert (5). For the cppB nested PCR, a result was interpreted as positive if a distinct 254-bp band was present; no equivocal results were encountered. For the single-tube nested cppB (stncppB) PCR, a cutoff for a positive result was obtained from the mean of five readings (from five separate runs) of a specimen diluted to contain approximately 1 viable CFU/reaction mixture, and an equivocal result was determined from a similar series with 1-in-5 to 1-in-2 dilutions of the specimen used to determine the positive cutoff. A specimen was considered negative for N. gonorrhoeae if the A₄₅₀ was 0.2, positive for N. gonorrhoeae if the A₄₅₀ was 0.8, and equivocal for N. gonorrhoeae if the A₄₅₀ was >0.2 and <0.8. Any run in which a negative control was positive or a positive control was negative was retested.

Commercial PCR methods. The AMPLICOR N. gonorrhoeae PCR was performed on the automated COBAS instrument according to the manufacturer's instructions. The 16S rRNA assay is a prototype method developed by Roche Diagnostic Systems. The assay uses the N. gonorrhoeae 16S rRNA gene described by Rossau et al. (12) (GenBank accession no. X07714). The assay was performed according to the instructions from Roche Diagnostic Systems.

In-house PCR methods. The primers and probes used in the in-house PCR methods are listed in Table 3. Testing of AMPLICOR-treated specimens was performed by adding 50 µl of specimen to 50 µl of the master mixture but without the addition of additional MgCl₂. Testing of the other specimens was performed by adding 50 µl to the master mixture and adjusting the MgCl₂ concentration to 1.5 mM. The initial specificities of all three methods were checked with purified DNAs (100 ng/reaction mixture) from the strains listed in Table 2. Reaction parameters for the three methods used were as follows.

(i) cppB PCR with MspI digestion. The cppB PCR with MspI digestion was performed as described by Ho et al. (7), with some modifications. Reagent concentrations were as follows, primers (HO1 and HO3) at 50 pmol/reaction mixture, deoxynucleoside triphosphates (dNTPs; Pharmacia Biotech) at 200 µM (each), Taq polymerase (Amplitaq; Perkin-Elmer, Melbourne, Australia) at 1 U/reaction mixture, and 1× Amplitaq reaction buffer II (Perkin-Elmer). Cycling parameters were 94°C for 2 min; 40 cycles of 94°C for 30 s, 55°C for 1 min, and 74°C for 30 s; and 74°C for 5 min. A 10-µl aliquot of the amplified PCR product was analyzed by agarose (1.6% [wt/vol] prepared in Tris-borate-EDTA [TBE] buffer [pH 8.3]) gel electrophoresis and ethidium bromide staining (0.5 µg/ml of gel), and the gel was photographed under UV light. The presence of a 390-bp band was interpreted as a positive result. For added specificity, 5 µl of the PCR-amplified material was digested with MspI (Life Technologies) for 1 h and was electrophoresed. The presence of the expected 250- and 140-bp bands confirmed the positive result.

(ii) Nested cppB PCR. The first-round master mixture contained the following: primers (HO1 and HO3) at 50 pmol/reaction mixture, dNTPs (Pharmacia Biotech) at 200 µM (each), Taq polymerase (Amplitaq; Perkin-Elmer) at 1.0 U/reaction mixture, and 1× Amplitaq reaction buffer II (Perkin-Elmer). Cycling parameters were 94°C for 2 min; 30 cycles of 94°C for 30 s, 55°C for 1 min, and 74°C for 30 s; and 74°C for 5 min. A 5-µl aliquot of the amplified PCR product was transferred to a new master mixture containing primers (cppBN1 and cppBN2) at 50 pmol/reaction mixture, dNTPs (Pharmacia Biotech) at 200 µM (each), Taq polymerase (Amplitaq; Perkin-Elmer), 1× Amplitaq reaction buffer II (Perkin-Elmer), and 1.5 mM MgCl₂. Cycling parameters were 94°C for 2 min; 38 cycles of 94°C for 30 s, 55°C for 1 min, and 74°C for 30 s; and 74°C for 5 min. A 10-µl aliquot of the amplified PCR product was analyzed by agarose (1.6% [wt/vol] prepared in TBE buffer [pH 8.3]) gel electrophoresis and ethidium bromide staining (0.5 µg/ml of gel), and the gel was photographed under UV light. The presence of a 254-bp band was interpreted as a positive result. This method achieved a sensitivity of <1 CFU per reaction tube.

(iii) stncppB PCR with liquid-phase hybridization and microwell plate detection. The master mixture contained the following: primers (HO1 and HO3) at 1 pmol/reaction mixture, primers (cppBN3 and cppBN4) at 50 pmol/reaction mixture, dNTPs (Pharmacia Biotech) at 200 µM (each), Taq polymerase (Amplitaq; Perkin-Elmer) at 2.0 U/reaction mixture, and 1× Amplitaq reaction buffer II (Perkin-Elmer). Cycling parameters were 94°C for 2 min; 30 cycles of 94°C for 30 s, 55°C for 30 s, and 74°C for 30 s; 38 cycles of 94°C for 30 s, 40°C for 30 s, and 74°C for 30 s; and 72°C for 5 min. A polyadenylate-digoxigenin tail was attached to the 3' end of cppBPB1 with a digoxigenin oligonucleotide 3' end labelling kit (Boehringer Mannheim, Sydney, Australia) according to the manufacturer's instructions. Liquid-phase hybridization with the probe cppBPB1 was achieved by adding 5 pmol of the labelled probe directly to the amplified specimen and heating the mixture (in the cycler) to 99°C for 10 min (denaturation), followed by exposure to 37°C for 20 min (annealing). One hundred microliters of the posthybridization specimen was added to streptavidin-coated microwell plates (Pathtec, Melbourne, Australia), and the plates were incubated at 37°C for 15 min. The microwell plate was washed four times with TBST buffer (Tris-HCl, 2.54 g/liter; Tris base, 0.47 g/liter; NaCl, 8.76 g/liter; Tween 20, 5 ml/liter). One hundred microliters of conjugate (1:2,000 dilution of anti-digoxigenin-peroxidase [Boehringer-Mannheim] in TBSSM [Tris-buffered saline plus 3% {wt/vol} skim milk powder]) was added to each well, and the plate was incubated at 37°C for 30 min. After four washes with TBST buffer, 100 µl of substrate (four parts 0.01% H₂O₂ and 1 part 0.1% 3,3',5,5'-tetramethylbenzidine in 40% dimethylformamide [Roche Diagnostic Systems]) was added. After incubation (for 10 min at room temperature in the dark), 100 µl of 10% H₂SO₄ was added and the A₄₅₀ (with a reference at 620 nm) was read at 10 min. A specimen was considered negative for N. gonorrhoeae if the A₄₅₀ was 0.2, positive for N. gonorrhoeae if the A₄₅₀ was 0.8, and equivocal for N. gonorrhoeae if the A₄₅₀ was >0.2 and <0.8. This method also achieved a sensitivity of <1 CFU per reaction tube.

	RESULTS

Top Abstract Introduction Materials and methods Results Discussion References

Table 2 lists the bacterial strains used to assess the specificity of the in-house PCR methods. Purified DNA preparations (100 ng/reaction mixture) of these strains gave negative results by all methods. Table 1 shows the sensitivities and specificities of the assays for three different strains of N. gonorrhoeae and for Australian and American Type Culture Collection (ATCC) strains of other Neisseria species. Both the stncppB and the 16S rRNA PCR methods were more sensitive than the AMPLICOR N. gonorrhoeae PCR for all three N. gonorrhoeae strains. Six strains of N. subflava were positive by the AMPLICOR N. gonorrhoeae PCR but were negative by the other two assays. All other strains were negative by all three assays. One of the cross-reacting N. subflava strains was isolated from a vaginal swab.

Table 4 compares the results of the AMPLICOR N. gonorrhoeae PCR and the cppB PCR performed with 207 nonsequential clinical specimens. The results for 29 of 113 (25.7%) of the AMPLICOR N. gonorrhoeae PCR-positive or -equivocal specimens were not confirmed by the more sensitive method, suggesting that these results were false-positive and unconfirmed equivocal results. The 16S rRNA kit became available to us during this study, and 13 of 52 AMPLICOR-positive specimens tested negative by this method, with the results of the 16S rRNA and stncppB PCR methods having 100% correlation. Interestingly, the results for 14 of 17 (82.3%) of the samples with AMPLICOR-equivocal results were not confirmed.

Of the 94 AMPLICOR-negative specimens from a population with a high prevalence of infection, only 2 (2.1%) specimens were positive by the more sensitive nested cppB method. The decision to change to a nested method was made early in the study, when several AMPLICOR-positive specimens were not confirmed to be positive by the cppB method with MspI digestion. The nested method was developed, and the specimens negative by the cppB method with MspI digestion were positive. In our hands, the cppB method with MspI digestion did not obtain sufficient sensitivity to be used as a confirmatory test. All subsequent tests were performed first by the nested cppB method and then by the more refined stncppB method.

	DISCUSSION

Top Abstract Introduction Materials and methods Results Discussion References

The ability to detect both C. trachomatis and N. gonorrhoeae in a single reaction with an automated instrument with sufficient sensitivity to detect low numbers of organisms in a specimen obtained from asymptomatic individuals by noninvasive means has great potential in the fight against STDs. Theoretically, if the core group in a population can be detected, strategic points for intervention can be identified to help break the cycle of transmission (13).

The data presented in this report support the use of the AMPLICOR C. trachomatis/N. gonorrhoeae PCR for the purpose of screening individuals for infection with these two organisms. On the other hand, the specificity of the method for N. gonorrhoeae was shown to be unsuitable for a reliable diagnosis and suggests that confirmatory testing by a more sensitive and specific PCR targeting a different gene is mandatory. Two confirmatory methods were presented in this study. Both tests were shown to be more sensitive and specific than the AMPLICOR N. gonorrhoeae PCR, and therefore, either would be suitable as a confirmatory test for the population used for the present study. Importantly, the lower analytical sensitivity of the AMPLICOR N. gonorrhoeae PCR did not translate into a significant decrease in sensitivity when applied to clinical specimens obtained from an asymptomatic local population with a high prevalence of infection.

The high false-positivity rate in this study differs from that for prospectively tested specimens in the United States, Canada, and Europe (10). Generally, 1 to 2% of uninfected specimens were false positive in these studies, and a positive predictive value of >90% was experienced in areas of high prevalence (15 to 20%) (11). A positive predictive value of 84% (81 of 96) in the present study differs from the value of >90% found by other investigators. One hypothesis is that the proportion of persons carrying cross-reacting Neisseria species is higher for the high-risk Australian populations. In a previous study (1), a false-positivity rate of <1% was experienced in two low-risk populations (for 1,905 patients [815 specimens were processed by my laboratory], the prevalence of N. gonorrhoeae infection was <1.0%). These data would argue against a laboratory source for the cross-reacting organisms because one would expect a similar false-positivity rate regardless of the study group tested. All positive and equivocal specimens in the present study were retested (from a reprocessed primary specimen) by the AMPLICOR PCR on the same day that the confirmatory test was run. To be included in the analysis as a positive or an equivocal result, therefore, the result must have been reproducible. Because of this process, the rate of technical errors can be considered to be low, and degradation of N. gonorrhoeae DNA between primary and secondary testing can be excluded as the cause of false positivity.

In my laboratory the in-house stncppB PCR is used as the confirmatory test because it is believed that the laboratory has sufficient analytical and clinical data to support its use as a confirmatory test in its location and circumstances. It is common practice in my laboratory, in the absence of local data on new commercial PCR kits, to evaluate specimens with positive results by a second PCR method aimed at a target gene different from that used in the kit. The method of Ho et al. (7) was modified so that the sensitivity was greater than that of the AMPLICOR N. gonorrhoeae kit and the same specimens were able to be used. This step was undertaken because the confirmatory 16S rRNA kit was not available in my laboratory at that time. I believe that the analytical data presented here is evidence that the 16S rRNA PCR is highly suited as a confirmatory test. Although not as complete as the stncppB PCR comparison, the evaluation of the 16S rRNA PCR with clinical specimens indicates that this test is as suitable (13 of 52 specimens with positive results by the AMPLICOR PCR were shown to be false positive, with a 100% correlation of the results of the 16S rRNA PCR to those of the stncppB PCR).

The choice of a confirmatory test should be made with caution. First, strains of N. gonorrhoeae which lack the cryptic plasmid have been described (14), although it has been shown that a copy of this gene was present in the chromosomes of all strains tested (5). It is possible that this target number variability between strains may manifest in variable sensitivity between strains, the end result of which could be a false-negative confirmatory test. Second, homology between the sequence of the N. gonorrhoeae cryptic plasmid and the sequences of N. meningitidis and N. lactamica plasmids has been reported (8). This homology was not observed in the data presented in the present study, however. Because of the potential for false-positive and false-negative results when the cppB gene is used as a target for detection, a careful local evaluation is recommended. The 100% correlation between the 16S rRNA and stncppB PCRs in both the analytical and the clinical comparisons presented here provides strong evidence that false-positive AMPLICOR N. gonorrhoeae PCR results are being obtained for the clinical specimens described here at a rate higher than that reported previously. The data also indicate that either of the confirmatory tests described here is suitable for the population described here.

A testing algorithm was developed for this laboratory by using the data obtained coupled with resource availability. All specimens sent for investigation for C. trachomatis and N. gonorrhoeae infections are tested by the AMPLICOR C. trachomatis/N. gonorrhoeae PCR. For specimens which test negative for both organisms or positive for C. trachomatis, a report is issued without further testing. If a specimen is positive for N. gonorrhoeae, a confirmatory test is performed before a report is issued. If the confirmatory test for N. gonorrhoeae is positive, the report states that the specimen is positive for N. gonorrhoeae DNA. If the confirmatory test is negative, an equivocal result is issued and retesting of the patient is encouraged.

Multiplex PCR with the AMPLICOR system is a cost-effective approach to the screening of asymptomatic populations for C. trachomatis and N. gonorrhoeae infections. It is hoped that the data provided here will encourage others to consider confirmation of positive PCR results for N. gonorrhoeae, especially in geographical regions and under circumstances in which the test is newly introduced. When considering the medical, legal, social, psychological, and epidemiological impacts of a false-positive diagnosis of gonorrhea, the added expense of performing a confirmatory test is well justified.