The past, present and future of molecular testing for Neisseria gonorrhoeae in Australia: still challenging

Introduction

Neisseria gonorrhoeae infections are a major cause of sexually transmitted infections (STIs) worldwide with a World Health Organization estimate of 87 million new infections per year.¹ In Australia, the annual number of N. gonorrhoeae notifications has increased steadily from 15,012 in 2013 to 40,541 in 2023.² Neisseria gonorrhoeae infections and associated complications are well documented.^3,4 In Australia, N. gonorrhoeae infections are disproportionately distributed, with a much higher prevalence in Indigenous Australian populations compared to non-Indigenous Australians.⁵

For those working in the area of sexual health, nucleic-acid amplification tests (NAATs) need no introduction as to why they are used worldwide for the detection of N. gonorrhoeae. NAATs have also developed into highly efficient diagnostic testing strategies for easy-to-collect non-invasive specimens such as urine and self-collected swabs – commonplace for STI investigations. The other important advantage is the ability to simultaneously detect other STIs, such as Chlamydia trachomatis – for example, 22% of all N. gonorrhoeae-positive samples detected by our laboratory in Perth, WA, Australia are also positive for C. trachomatis,⁶ which highlights the importance co-infection detection. Many commercial C. trachomatis and N. gonorrhoeae assays are available to suit different laboratory requirements, with many (if not all) continuing to progress to more efficient workflows. A list of high-throughput sample-to-result systems (>1000 samples per 24 h) available in Australia, highlighting key features, is described in Table 1.

Despite the above advantages, N. gonorrhoeae NAATs have been plagued with specificity problems associated with cross-reaction with commensal Neisseria species. The problem has been most pronounced for extragenital sites, such as oropharyngeal swabs, where commensal neisseriae are ubiquitous.^7–10 Supplementary testing (whereby samples testing positive in a screening N. gonorrhoeae NAAT are confirmed by a second NAAT) has been widely implemented in many jurisdictions to address these issues of non-specificity.^11–14

More recently, the utility of supplemental testing has become more contentious. Supplemental testing represents additional work and cost to the laboratory requiring nucleic acid retrieval or separate extraction, assay set-up, testing, result interpretation, additional quality control and quality assurance testing. This added workload was clinically justified given the poor specificity of earlier tests. Today, this justification has diminished given the improved sensitivity and specificity of modern N. gonorrhoeae NAATs, essentially redesigned to mitigate the potential for false-positive results.^6,10,14,15 These improvements have caused additional problems for supplementary N. gonorrhoeae assays:

Despite these challenges, a new and exciting role for N. gonorrhoeae supplemental testing is emerging. Given heightened concerns over N. gonorrhoeae antimicrobial resistance (AMR),¹⁶ the integration of reliable genotypic AMR markers into N. gonorrhoeae NAATs, allows for individualised therapy and enhanced antimicrobial stewardship.^17,18

N. gonorrhoeae supplemental testing in Australia

The first Public Health Laboratory Network (PHLN) guidelines for the use and interpretation of nucleic acid detection tests for N. gonorrhoeae testing in Australia were published in 2005.¹¹ The primary concern of the PHLN recommendations focussed on reducing the likelihood of the laboratory issuing false-positive results. Supplemental testing was advocated for all N. gonorrhoeae-positive screening results to address unacceptable rates of false-positive results caused by cross-reactivity with non-gonococcal Neisseria, or false-negative results due to target loss, with genetic exchange within the Neisseria genus the cause of some of these issues.^19–22 As a consequence, supplemental testing was embedded into Australian routine molecular diagnostics.²³

As screening assays improved in sensitivity and specificity over the following years, a new issue emerged with one of the recommendations, ‘If a sample is positive in a screening assay but a suitable supplemental assay is negative, then the result should be issued as negative’ (p. 358¹¹).²⁴ Data collated from quality assurance programs showed that supplemental testing may lead to false-negative results for samples with low N. gonorrhoeae load, i.e. supplemental testing was now leading to sensitivity problems – we just could not get it right! A subsequent review of the guidelines was undertaken by the National Neisseria Network in 2015 focusing on false-negative results.⁷ The review concluded that supplementary testing remains best practice but recommended negative supplementary results from N. gonorrhoeae-positive urogenital screening results should not be reported as negative without an appropriate explanatory comment indicating that gonococcal infection cannot be excluded. Furthermore, given the improved sensitivity and specificity of newer NAATs,¹⁹ laboratories may need to review the performance of their supplementary testing methods to ensure they are not reporting false-negative results.

On-going NAAT improvements and our Perth experiences

Although many second-generation assays have shown substantially less cross reactivity with non-gonococcal Neisseria species than the earlier generation assays,¹⁹ reports of non-specificity issues have continued.⁹ It should be noted, there was no commercial test available in Australia prior to 2017 that was validated by the manufacturer for testing oropharyngeal and anorectal samples,¹⁴ yet laboratories continued to test extragenital samples due to clinical need underpinned by local validation data. In fact, given the widespread concerns over the potential for false-positive results, Australian laboratories have been vigilant when validating new screening assays. For example, our Perth laboratory took particular care when transitioning from one C. trachomatis and N. gonorrhoeae screening assay to another, in this case moving from RealTime m2000 (Abbott Molecular) to cobas c4800 (Roche Molecular Systems).¹⁴ In the same study, we also evaluated the cobas c6800 C. trachomatis and N. gonorrhoeae assay as the first screening assay available in Australia with an oropharyngeal and anorectal testing claim (Table 1). Given the c6800 assay has an oropharyngeal claim and is reported to be more sensitive than c4800, we were interested to see how this assay performed on c4800-positive oropharyngeal samples, in light of documented c4800-specificity issues with the report in New Zealand of a false positive result with Neisseria macacae strain.⁹

The overall results from the m2000 and c4800 comparison (n = 344) were highly concordant, with only a few discordant samples; most of which could be explained by low N. gonorrhoeae loads indicated by late cycle of quantitation (C_q) results. However, when we explored a little deeper into the supplemental testing results from our in-house dual-target supplementary assay (opa and porA; i.e. N. gonorrhoeae duplex), the confirmatory rates were significantly higher for m2000 compared to c4800, with oropharyngeal samples the key difference. We subsequently demonstrated that N. gonorrhoeae duplex failed to confirm some true-positive N. gonorrhoeae samples. The observed discrepancies were due to a combination of c4800 producing false-positive results for oropharyngeal samples as well as sensitivity issues related to the N. gonorrhoeae-duplex assay (mostly porA). The overall results from the c4800 and c6800 comparison (n = 400) were also highly concordant, with a few c4800-positive oropharyngeal and anorectal samples (late C_q) that were negative on c6800, despite the 1 colony-forming unit (CFU) mL^–1 sensitivity claim of the c6800 test.

Similarly, another recent deep dive into the c6800 assay (n = 300) using quantitative PCR analysis,⁶ revealed a small number of oropharyngeal samples with solid C_q values of 27.5, 29.0, 30.3 and 31.5 (equivalent to 2460, 890, 380 and 170 CFUs mL^–1 respectively), which did not test positive by any other commercial supplemental test, including Xpert CT/NG, despite an Xpert N. gonorrhoeae oropharyngeal sensitivity claim of 7 CFUs mL^–1. These samples should have readily confirmed if they were truly positive for N. gonorrhoeae – a finding additionally supported our own lower limit-of-detection studies. Hence, based on these data, we showed that c6800 may still be prone to false-positive results for oropharyngeal samples. Given both versions of the cobas test share the same target (DR-9), could these discrepancies be examples of the ‘N. macacae story’? To follow up on this, we tested the above N. macacae isolate (kindly provided by Dr Bromhead in New Zealand) with c6800 and conclusively proved that this particular N. macacae strain does not cross-react with the third-generation c6800 assay, thus confirming improved specificity of the c6800 version for this particular organism.¹⁰

For these reasons we re-affirm

Finally, although supplemental testing may not be required for urogenital samples based on our data, we continue supplemental testing for all sample types by default for another reason: AMR detection (outlined below).

Supplemental testing for AMR

Although culture still remains the definitive AMR test, NAATs have largely replaced culture as the primary method for modern-day gonorrhoea diagnosis. Given the decline in culture and the subsequent reduction of AMR data,¹⁶ including limitations in conducting surveillance in remote settings, the integration of reliable molecular markers into routine supplemental assays has already embedded itself as a useful addition to enhance culture-based surveillance.²⁶ But more recently, these rapid molecular AMR tools provide new potential beyond surveillance, including the ability to rapidly predict successful treatment options and reduce AMR selection pressure. For example, direct detection of penicillinase-producing N. gonorrhoeae using a number of in-house assays have been utilised for quite some time, albeit with some specificity challenges.^27,28 In addition, there are a number of in-house and commercial assays to predict AMR to ciprofloxacin, azithromycin and third-generation cephalosporins, all defined as high priority targets.²⁹ Overall, the lack of convenient and economical commercial AMR assays impinges the uptake AMR testing in clinical laboratories, as in-house tests require significant development, validation, more oversight and considerably more work effort.

When the opportunity was presented to add a reliable and cost-efficient molecular AMR target into our diagnostic workflow, our laboratory switched our N. gonorrhoeae duplex to the opa, porA and gyrA ResistancePlusGC assay (SpeeDx).⁶ In addition to detecting N. gonorrhoeae, this assay also predicts ciprofloxacin susceptibility by N. gonorrhoeae gyrA characterisation.²⁹ Using our existing N. gonorrhoeae supplemental testing workflow, we simply added a like-for-like supplemental assay (opa and porA) with the added bonus of a clinically useful resistance determinate. Although not recommended first-line treatment, ciprofloxacin offers multiple benefits, including oral administration, while also being an effective treatment for urogenital and extragenital infections. However, at the time of initial clinical presentation the clinician does not have laboratory information if ciprofloxacin would be efficacious. The most common scenario leading to ciprofloxacin use would be a patient recalled for treatment from a positive asymptomatic screen – useful but not ideal. Hence, the dilemma of ciprofloxacin use: prescribers may choose to delay ceftriaxone treatment in some patients awaiting resistance results, while balancing this decision with the risk of prolonging the infectious period and the potential risk to the patient and further transmission. Unlike a recent UK study assessing the impact of targeted gyrA testing for the clinical management of gonorrhoea,³⁰ our approach is not targeted (test all positives), utilises the opa and porA as the supplemental test, and focusses on rapid turnaround of results (multiple runs per day) to help our clinicians with this dilemma. Furthermore, ongoing assessment of clinical value related to specific AMR marker is important, especially in the context of changing global and localised epidemiology.

Future approaches

New multiplexing technologies would be required for the high-throughput systems to meet the technical requirements of the N. gonorrhoeae supplemental testing-AMR challenge. The ideal assay characteristics would include:

Quite simply, the laboratory could customise the handling of the supplementary and AMR results according to local guidelines, own laboratory validation data, or AMR prevalence. Ultimately, however, N. gonorrhoeae AMR prediction assays need to progress to point-of-care capability and include all the high-priority AMR targets. Multi-target sample-to-result systems such as BioFire Film Array (bioMerieux, Marcy-l’Etoile, France), or other next-generation multiplexing assays may be up to this challenge.

Conclusion

Despite laboratory evidence to suggest that supplemental testing may not be required for urogenital samples, supplemental testing is here to stay at least for oropharyngeal samples. However, predicting AMR may prove more important than N. gonorrhoeae supplemental testing to support a positive screening result for a urogenital sample, given that urogenital testing represents the majority of samples tested. Although automated and semi-automated STI screening solutions exist in Australia, a substantial need remains for fully automated sample-to-result molecular assays to meet increasing demands for STI screening services. Whatever the testing approach, AMR prediction has a strong future in molecular diagnostics, as we fight the growing state of AMR, not just within N. gonorrhoeae, but other STIs such as Mycoplasma genitalium. Until more rapid molecular AMR tests become available, we continue to evaluate new N. gonorrhoeae screening assays with a high degree of scrutiny using supplemental assays as comparators, given the historical context of screening assay non-specificity in this organism.