Article Text

PDF

Validation of a laboratory-developed real-time PCR protocol for detection of Chlamydia trachomatis and Neisseria gonorrhoeae in urine
  1. M J Hopkins1,
  2. L J Ashton1,
  3. F Alloba2,
  4. A Alawattegama2,
  5. I J Hart1
  1. 1Liverpool Specialist Virology Centre, Royal Liverpool University Hospital, Liverpool, UK;
  2. 2Department of Genitourinary Medicine, Royal Liverpool University Hospital, Liverpool, UK
  1. Correspondence to Dr MJ Hopkins, Liverpool Specialist Virology Centre, Royal Liverpool University Hospital, Liverpool L7 8XP, UK; m.hopkins{at}liv.ac.uk

Abstract

Objective To evaluate a sensitive and specific, real-time PCR assay with internal control for Chlamydia trachomatis and Neisseria gonorrhoeae DNA detection in urine specimens.

Methods The diagnostic performance of a laboratory-developed quadruplex assay (LDQA) targeting the cryptic plasmid and MOMP genes of C trachomatis, the porA pseudogene of N gonorrhoeae and a synthetic internal control was assessed using 1028 urine specimens. The LDQA was compared with the Roche COBAS Taqman CT test and the COBAS Amplicor NG assay with supplemental confirmation tests. The subsequent performance of the LDQA in detecting N gonorrhoeae was monitored in comparison with bacterial culture from swabs.

Results 88 (8.6%) urines were determined as C trachomatis positive in the diagnostic evaluation. LDQA sensitivity and specificity were calculated to be 100% and 99.9%, respectively, for C trachomatis. The LDQA showed high specificity with isolates of other Neisseria species and gave complete concordance with resolved data for N gonorrhoeae detection. However, the incidence of N gonorrhoeae infection was low, with 17 (1.7%) positive patients. A post-implementation audit of 14 316 patients gave the LDQA N gonorrhoeae urine PCR protocol (porA, OPA, 16s rDNA) a sensitivity of 96.9% and specificity of 99.8% in comparison with bacterial culture from swabs.

Conclusions The LDQA was found to be an effective method for the detection of C trachomatis and N gonorrhoeae DNA in urine samples, and the PCR protocol has replaced bacterial culture for the screening of N gonorrhoeae in asymptomatic men and women in the laboratory.

  • Chlamydia trachomatis
  • laboratory diagnosis
  • Neisseria gonorrhoeae
  • real-time PCR
  • urine

Statistics from Altmetric.com

Nucleic acid amplification test (NAAT) methodologies are currently recommended as the standard of care for the diagnosis of genital chlamydia infection in England.1 The gold standard method for the diagnosis of Neisseria gonorrhoeae infection is bacterial culture, although NAAT are also advocated to increase the sensitivity of detection in non-invasive sample types.2 A number of NAAT for the detection of Chlamydia trachomatis employ, or are pursuing, multiplex formats to reduce the risk of underreporting infections, and companies are increasingly offering combined tests capable of detecting both C trachomatis and N gonorrhoeae in the same analysis.3 4 Previous reports have identified problems with the sensitivity of N gonorrhoeae detection in female urine by PCR, and also highlighted the absence of antimicrobial sensitivity data from these tests.5 However, the relatively small increase in laboratory reagent and labour costs associated with the multiplex format and the ease of non-invasive sexual specimen collection makes this an attractive option for clinics.

Different genetic targets are often preferred for screening and confirmation assays to increase the overall specificity of NAAT diagnostic procedures.1 6 Laboratory-developed assays (LDA) can offer useful alternatives to the range of commercial tests, particularly when existing equipment can be utilised and the new protocols aligned to established workflows. However, the laboratory must undertake significant validation of LDA to ensure the assay and method of sample preparation are suitable for diagnostic use. A small number of laboratories in the UK employ LDA for routine screening of C trachomatis but only limited data are available in the literature for real-time LDA for the multiplex detection of C trachomatis and N gonorrhoeae.7–9 Here we describe a real-time multiplex LDA protocol for the detection of C trachomatis and N gonorrhoeae and validate its use with first-void urine samples.

Methods

Bacterial strains and EQA samples

Two external quality assurance (EQA) panels for C trachomatis and N gonorrhoeae were supplied by Quality Control for Molecular Diagnostics (Glasgow, UK). PCR specificity for N gonorrhoeae was also assessed using isolates of Neisseria cinerea (n=16), Neisseria flavescens (n=2), N gonorrhoeae (n=8), Neisseria lactamica (n=7), Neisseria meningitidis (n=4), Neisseria mucosa (n=1), Neisseria perflava (n=1), Neisseria polysaccharea (n=1), Neisseria sicca (n=3), Neisseria subflava (n=3) and Neisseria spp. (n=7).

Patients and specimens

A total of 1028 urine specimens was obtained from 1015 patients attending the Genitourinary Medicine Clinic at the Royal Liverpool University Hospital during the period August to October 2007. The median age of the patients was 25 years (range 15 to 73 years) with 556 being male (55%) and 459 female (45%). Samples were collected as part of routine testing for sexually transmitted infections, with patient consent obtained to test for both chlamydia and gonorrhoea infection. Patients continued to be swabbed for routine microscopy and N gonorrhoeae culture for comparison with the N gonorrhoeae PCR results.

Results from the laboratory-developed quadruplex assay (LDQA) N gonorrhoeae PCR protocol were audited in comparison with swabs for bacterial culture over the period February 2008 to February 2009. A total of 14 319 patients was screened during this time. Of these, 3393 were asymptomatic women, and the remaining 10 926 were symptomatic men and women.

DNA isolation

For the LDQA and the N gonorrhoeae confirmation assays (NsppID and NG-OPA/16s rDNA), genomic DNA was extracted from 500 μl urine or bacterial suspension and eluted in 100 μl buffer using the Magnapure Classic automated extraction system with the DNA Isolation Large Volume Kit (Roche Diagnostics, Burgess Hill, UK) according to the manufacturer's instructions.

Urine samples (1 ml) were prepared for the C trachomatis confirmation test (Artus CT Plus assay; Qiagen, Crawley, UK) by centrifugation for 5 minutes at 12 000g and resuspension of the cell pellet in 400 μl water. The entire volume was then processed using the Qiagen M48 Biorobot and Magattract Virus Mini M48 Kits (Qiagen), eluting in 50 μl buffer. If necessary, extracts were stored at +4°C for up to 24 h or longer at −70°C until required.

Laboratory-developed quadruplex assay

N gonorrhoeae assay components were introduced to an established C trachomatis assay with the best performing quadruplex taken forward for the LDQA.9–11 Details of the oligonucleotides used in the final assay are given in table 1. The assay used LC480 probe master mix (Roche Diagnostics) in accordance with the manufacturer's instructions to a final volume of 25 μl, with each reaction containing 10 μl purified DNA. Amplification and detection were carried out using a LightCycler 480 PCR machine (Roche Diagnostics) with the following cycling parameters; 5 minutes at 95°C, followed by 50 cycles of 10 s at 95°C, 45 s at 60°C, 1 s at 72°C, and a final step of 30 s at 40°C. The amplification signal for the N gonorrhoeae porA pseudogene fragment or either of the two C trachomatis PCR targets was taken as a positive result. The internal control was required to amplify in order to record a negative result.

Table 1

Oligonucleotide primers and probes used in this evaluation

COBAS Amplicor/Taqman assays

Samples (500 μl) were processed with 50 μl of the resultant lysate being used in the Roche COBAS Amplicor NG or Taqman CT assays according to the manufacturer's instructions (Roche Diagnostics). Results from either assay were deemed inhibitory if the internal control failed to amplify in otherwise negative samples.

Confirmation PCR assays

All urine samples identified as positive or inhibitory for C trachomatis by the COBAS Taqman CT assay or LDQA were repeat tested with the Artus CT Plus PCR kit (Qiagen) on a LightCycler 1.2 PCR machine (Roche Diagnostics) according to the manufacturer's instructions. Published PCR directed towards the Neisseria 16S rRNA gene (NSppID), and the opacity protein gene (NG-OPA) were used to confirm the presence of N gonorrhoeae DNA in any sample identified as positive, inhibitory or grey zone by the COBAS Amplicor NG test or LDQA.12 13 The NSppID assay was run as described previously but was later replaced with the NG-OPA PCR multiplexed with a 16s rDNA PCR for post-implementation confirmation of the LDQA (table 1). The NG-OPA/16s rDNA duplex PCR was run on the LC480 under the same conditions as the LDQA, but the extension time at 72°C was increased to 10 s.

Bacterial culture

N gonorrhoeae was isolated using GC agar base with VCAT selective supplement (Oxoid, Basingstoke, UK). Suspect colonies were identified to species level on the basis of Gram stain, oxidase test, Phadebact GC kit (Launch Diagnostics, Longfield, UK) and API NH test (BioMerieux, Basingstoke, UK).

Statistical analysis

Kappa analysis was used to calculate the level of agreement between LDQA for detection of C trachomatis and/or N gonorrhoeae in comparison with resolved data from both screening assays and all confirmation tests. The level of agreement was also calculated for confirmed N gonorrhoeae PCR results against bacterial culture.

Results

Assay sensitivity and specificity

The LDQA scored full marks on the C trachomatis and N gonorrhoeae EQA panels, correctly identifying all positive samples and giving no false-positive results. The COBAS tests achieved a score of 10 out of 10 for the C trachomatis EQA panel, but falsely identified a sample containing N lactamica as N gonorrhoeae positive in the N gonorrhoeae panel. However, the NsppID confirmation assay differentiated this as a non-gonococcal Neisseria-positive sample on the basis of melt temperature (66°C). Neither the COBAS NG Amplicor test nor the N gonorrhoeae porA component of the LDQA amplified any non-gonococcal isolates from the laboratory panel of 53 Neisseria species but correctly identified the eight isolates of N gonorrhoeae.

Diagnostic evaluation

Positive urines were identified on the basis of at least two positive results—either two PCR or PCR with microscopy or culture. Two samples were excluded from the C trachomatis kappa analysis due to repeatedly inhibitory results with both the COBAS and Artus PCR tests. Table 2 shows that resolved data identified 88 of 1026 (8.6%) C trachomatis-positive samples and gave good concordance with LDQA results (kappa value 0.99). Both the COBAS CT Taqman and the LDQA screen each identified one false-positive C trachomatis sample. The LDQA correctly detected C trachomatis DNA in two samples that gave negative results from the COBAS Taqman CT screen (table 3). Overall, the sensitivity and specificity of the LDQA C trachomatis screen were calculated to be 88 of 88 (100%) and 937 of 938 (99.9%), respectively.

Table 2

LDQA results in comparison with resolved data from 1028 urines for C trachomatis and N gonorrhoeae

Table 3

Details of C trachomatis test results for the 13 discordant specimens

The LDQA identified 17 (1.7%) of the 1028 urine samples as porA pseudogene PCR positive, which gave 100% concordance with resolved N gonorrhoeae data (table 2). All 17 samples were confirmed as NG-OPA PCR positive but four could not be amplified using the NsppID 16s rDNA assay (table 4). The COBAS Amplicor NG PCR identified a larger number of urine samples that required N gonorrhoeae confirmation tests, with a total of 45 of 1028 giving a positive (n=19), grey zone (n=6), or inhibitory (n=20) result with this screening assay. Seventeen were the resolved N gonorrhoeae positives, but 24 could not be confirmed by any of the other tests employed (table 4). The remaining four were identified as non-gonoccocal Neisseria spp on the basis of NsppID PCR melt curve analysis. The sensitivity and specificity of the LDQA N gonorrhoeae screen were calculated as 17 of 17 (100%) and as 1011 of 1011 (100%), respectively, using this dataset.

Table 4

Details of N gonorrhoeae test results for the 34 discordant specimens

Three (0.3%) of the 1028 samples were confirmed positive for dual infection of C trachomatis and N gonorrhoeae.

N gonorrhoeae follow-up audit

The LDQA was introduced into routine service in addition to the existing bacterial culture, and results from the two techniques were reviewed after 12 months. Kappa analysis of the N gonorrhoeae urine PCR protocol results showed a strong correlation with bacterial culture (kappa value 0.99) identifying 210 patients positive by all three PCR targets (porA, OPA, 16s rDNA). Twenty-five of these were negative by bacterial culture, whereas there were an additional six false-negative PCR results (table 5). Sensitivity and specificity of the N gonorrhoeae urine PCR protocol were calculated as 96.9% and 99.8%, respectively, using this dataset. Gonorrhoea infection was diagnosed in 11 of 3393 asymptomatic women. One case was culture positive/PCR negative, but two cases that were initially negative by bacterial culture from cervical swabs were PCR positive. The latter two cases were culture positive on patient recall.

Table 5

Neisseria gonorrhoeae urine PCR protocol validation results in comparison with bacterial culture data from 14 319 sexual health screens

Discussion

The use of NAAT for the detection of C trachomatis is well established and is increasingly utilised for the detection of N gonorrhoeae, particularly as emphasis shifts towards the use of non-invasive specimen types.1 14 15 Several manufacturers now supply NAAT platforms that can screen urine samples for both C trachomatis and N gonorrhoeae in the same reaction. Results presented here show that the LDQA is a sensitive and specific test for the detection of C trachomatis and N gonorrhoeae in first-void urine samples. Dean et al16 reported the COBAS Amplicor CT test to have a sensitivity and specificity of 96.9% and 98.2%, respectively, but there are no published data available comparing the COBAS TaqMan CT test (used in our evaluation) with the Aptima Combo 2, which was cited as the most sensitive assay in that study. Both Roche and Abbott have included secondary chlamydial targets in their assays to facilitate the detection of new variant strains.8 The LDQA detects the cryptic plasmid and MOMP PCR products in different channels of the real-time PCR machine, which facilitates local monitoring of this issue.

Numerous genetic targets have been assessed for the detection of N gonorrhoeae, although bacterial culture is often considered the diagnostic standard due to genetic heterogeneity in this organism.9 17–20 If requested, the LDQA allows flexible detection of C trachomatis and/or N gonorrhoeae through the addition of a porA pseudogene target to the assay. This PCR was chosen because of the high sensitivity and specificity previously demonstrated with the pseudogene target.10 21 The porA PCR was found to be more sensitive and specific than the COBAS Amplicor NG test, which has previously been shown to generate a considerable number of inhibitory or reactive results that could not be confirmed, and our findings support this (table 4). The NSppID assay has been reported as a useful confirmation test to improve the specificity of the COBAS NG assay.12 However, NSppID assay sensitivity was limited in comparison with the NG-OPA confirmation PCR and an NG-OPA/16s rDNA duplex assay was subsequently adopted as the confirmation assay to complement the LDQA (porA) screen.

Insufficient data for N gonorrhoeae infection was generated from the initial diagnostic evaluation due to the low prevalence of this organism in our patient population. The LDQA screen was implemented alongside the existing bacterial culture service and a 12-month review of data showed the urine PCR protocol was comparable to N gonorrhoeae culture from swabs. Although culture was taken as the standard for comparison, the authors feel the 25 culture negative/urine PCR positives (table 5) could equally be ‘true’ positives because amplification of all three N gonorrhoeae targets in the screen and confirmation PCR (porA, OPA, 16s rDNA) was required for a N gonorrhoeae PCR-positive result to be issued. This assumption would improve the performance characteristics of the N gonorrhoeae urine PCR protocol, although further comparison with additional testing protocols would be needed to establish this. No obvious problems were observed identifying gonorrhoea infection in the 3393 asymptomatic women tested over this time period. This may be related partly to the use of a nucleic acid purification technique in this protocol to concentrate DNA and efficiently remove inhibitors, unlike the original COBAS method, which relied on a simple lysis procedure for sample preparation.

Review of these data led to a change of policy in our clinic, with the urine PCR protocol becoming the favoured screen over bacterial culture from swabs for asymptomatic patients. However, antibiotic resistance profiling remains an important consideration, and N gonorrhoeae culture is maintained as part of the confirmation protocol. The authors do not advocate LDA above validated assays from commercial companies that provide support and a degree of accountability, but firmly believe they warrant consideration by specialist molecular diagnostic laboratories when in vitro diagnostics (IVD) approved tests are not appropriate for existing workflows. Data presented here may be useful to inform laboratory decisions on screening or confirmation protocols, or the clinical choice of sample.

Key messages

  • This study has validated the C trachomatis assay previously described by Jalal et al,7 for use with first-void urine specimens.

  • LDQA followed by NG-OPA/16s rDNA PCR confirmation allowed for the specific and sensitive multiplex detection of gonorrhoea infection from a non-invasive sample.

  • Laboratories may find these NAAT data useful when reviewing their existing C trachomatis and N gonorrhoeae workflows.

Acknowledgments

The authors are grateful to Dr Helen Palmer, Scottish Bacterial Sexually Transmitted Infections Reference Laboratory, Royal Infirmary of Edinburgh, UK, for the provision of Neisseria strains, and to Dr Steven Lane, Centre for Medical Statistics and Health Evaluation, University of Liverpool, for statistical advice throughout the study. The authors thank Mrs Jean Howell in the Department of Genitourinary Medicine, Royal Liverpool University Hospital, for assistance compiling the N gonorrhoeae data.

References

View Abstract

Footnotes

  • Funding This study was supported by the Royal Liverpool University Hospital.

  • Competing interests None.

  • Patient consent Obtained.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.