Article Text

PDF

Rapid accurate point-of-care tests combining diagnostics and antimicrobial resistance prediction for Neisseria gonorrhoeae and Mycoplasma genitalium
  1. Syed Tariq Sadiq1,
  2. Fulvia Mazzaferri2,
  3. Magnus Unemo3
  1. 1Applied Diagnostic Research and Evaluation Unit, St George's, University of London, London, UK
  2. 2Diagnostic and Public Health Department, Infectious Diseases and Tropical Medicine Section, University of Verona, Verona, Italy
  3. 3World Health Organization Collaborating Centre for Gonorrhoea and other STIs, Örebro University, Örebro, Sweden
  1. Correspondence to Dr Tariq Sadiq, Institute for Infection and Immunity, St George's, University of London, London SW17 0RE, UK; ssadiq{at}sgul.ac.uk

Abstract

In addition to inadequate access to early diagnosis and treatment with antimicrobial agents for patients and sexual contacts, management and control of STIs is significantly challenged by emergence and spread of antimicrobial resistance (AMR), particularly for STIs such as Neisseria gonorrhoeae and Mycoplasma genitalium. This is further compounded by use of nucleic acid amplification techniques for diagnosis, resulting in reduced phenotypic AMR testing for N. gonorrhoeae and absence or suboptimal AMR surveillance for guiding treatment of both STIs in many settings. Rapid accurate point-of-care (POC) tests for diagnosis of all STIs would be valuable but to significantly impact treatment precision and management of N. gonorrhoeae and M. genitalium infections, combinations of rapid POC diagnostic and AMR testing (POC-AMR) will likely be required. This strategy would combat STI burden and AMR emergence and spread by enabling diagnosis and individualised treatment at the first healthcare visit, potentially reducing selection pressure on recommended antimicrobials, reducing transmission of resistant strains and providing means for AMR surveillance. Microfluidic and nanotechnology platforms under development for rapid detection of STIs provide a basis to also develop molecular rapid POC-AMR prediction. A number of prototypic devices are in the pipeline but none as yet approved for routine use. However, particularly for N. gonorrhoeae, more knowledge is required to assess which antimicrobials lend themselves to a genotypic POC-AMR approach, in relation to genotypic–phenotypic associations and potential impact clinically and epidemiologically. Key for successful deployment will include also understanding cost-effectiveness, cost-consequences and acceptability for key stakeholders.

  • NEISSERIA GONORRHOEA
  • ANTIMICROBIAL RESISTANCE
  • M GENITALIUM
  • DNA AMPLIFICATION
  • DIAGNOSIS

This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/

View Full Text

Statistics from Altmetric.com

Introduction

The WHO estimated that there were over 350 million new cases of the curable STIs Chlamydia trachomatis infections (131 million cases), gonorrhoea (Neisseria gonorrhoeae; 78 million), syphilis (Treponema pallidum subspecies pallidum; 6 million) and trichomoniasis (Trichomonas vaginalis; 143 million) among adults in 2012.1 Mycoplasma genitalium infections, which are not estimated by the WHO, are also prevalent worldwide.2 ,3 Although the major burden of these STIs affects low and middle income countries, severely impacting maternal and child health in particular,1 ,4 there are also high rates of transmission in more resourced settings, particularly among men who have sex with men and young heterosexual men and women.1–3

Use of rapid point-of-care (POC) tests for STIs could enable prompt aetiological diagnosis and more specific antimicrobial therapy at the first healthcare visit, immediate appropriate health promotion, general counselling and sexual contact notification, potentiating reduction of sequelae and further transmission of infection, particularly for populations that cannot or do not access sexual healthcare services. The introduction of such POC tests has also potential to improve service efficiencies and improve cost-effectiveness.5 ,6 Effective and validated rapid immunochromatographic lateral flow POC tests are commercially available for diagnosis of syphilis.7 However, for effective use in clinical management, no sufficiently sensitive rapid POC tests for diagnosis of other curable bacterial STIs such as C. trachomatis, N gonorrhoeae and M. genitalium have existed.8–11 Commercial introduction of highly accurate laboratory-based nucleic acid amplification techniques (NAATs) over recent decades has significantly improved detection of C. trachomatis and N. gonorrhoeae in many resourced settings and raised expectations of increased accuracy of diagnostics in general for these infections. Thus, further development of non-viral STI POC tests has mainly focused on combining NAATs with microfluidic and nanotechnology platforms to link high sensitivity with rapidity. To date, the only commercially available NAAT-based POC platform for STIs has been the GeneXpert System for simultaneous detection of C. trachomatis and N. gonorrhoeae,11 and recently a separate test for T. vaginalis became available.12 However, the GeneXpert platform fails to fulfil the ASSURED criteria (affordable, sensitive, specific, user-friendly, rapid, equipment-free, delivered to those who need it),13 stated by WHO, for rapid POC tests, for example, because the assays take around 90 min from sample to result, needs a source of electricity and is costly. Most recently, a rapid molecular POC test for C. trachomatis was approved in Europe (Atlas Genetics io),14 with a 30 min turnaround time, and many additional rapid molecular POC tests for non-viral STIs are in the pipeline.

Molecular antimicrobial resistance determinants as a basis for antimicrobial resistance prediction in laboratory and at POC

The emergence and spread of antimicrobial resistance (AMR), compromising particularly the treatment of N. gonorrhoeae and M. genitalium infections, provide significant clinical and public health challenges and threaten to undermine several of the potential diagnostic and control approaches highlighted above.

Regarding gonorrhoea, sequential use of empirical monotherapy with different antimicrobial classes over many decades has resulted in high prevalence of resistance to most introduced antimicrobials.15 This is further compounded by reduced use of culture in favour of NAATs for diagnosis, which in turn reduces availability of phenotypic AMR testing and surveillance data. National and international management guidelines have responded to this challenge by recommending combination therapies (mainly ceftriaxone 250–500 mg×1 intramuscularly plus azithromycin 1–2 g×1 orally), to try to minimise further emergence and/or spread of resistance.16–19

Concerning M. genitalium, a common cause of non-gonococcal genital discharge syndromes, which often require empirical therapy, resistance to recommended first-line treatment with macrolides (azithromycin or josamycin in some countries) has rapidly increased during the last decade; moreover, resistance to second-line therapy with fluoroquinolones (moxifloxacin) has emerged and is spreading internationally.2 ,10 ,20 The 2016 European guideline on M. genitalium infections recommends that all M. genitalium NAAT-positive samples should ideally be examined for macrolide resistance before treatment is initiated.2

Accordingly, accurate molecular tests for prediction of AMR, both in laboratory and as part of POC test development, are important for both these infections. For M. genitalium, this is more straightforward because both macrolide and fluoroquinolone resistance are mediated by restricted sets of single nucleotide polymorphisms (SNPs) in single copy genes encoding their respective antimicrobial targets, namely domain V of 23S rRNA and ParC, respectively.2 ,3 ,10 ,20 These SNPs are also easily detected by NAATs directly from clinical samples. However, the full extent of the relationship between several mutations in the 23S rRNA gene and, particularly, the parC gene and potential phenotypic resistance and treatment failure still remains unknown.10 In N. gonorrhoeae, this type of restricted SNP set causing resistance is mainly limited to fluoroquinolones, for which resistance and susceptibility can be predicted with relatively high sensitivity and specificity.10 ,15 ,21–23 Conversely, accurate prediction of resistance to antimicrobials such as penicillins, cephalosporins, tetracyclines and macrolides is significantly more difficult and complex because AMR additionally can be the result of multiple mechanisms beyond alterations of the antimicrobial target, such as plasmid-born determinants (β-lactamases and TetM resulting in resistance to penicillin and tetracyclines, respectively), efflux pump systems (increasing efflux of several antimicrobials out from cells) and the porin protein PorB (reducing influx of several antimicrobials into cells). In addition, many different mutations in the antimicrobial target genes, for example, penA (encoding the penicillin-binding protein 2 (PBP2), which is the lethal target for penicillins and cephalosporins), can exist in circulating N. gonorrhoeae strains and new mutations have continuously been identified with different effects on phenotypic susceptibility. Finally, some antimicrobial target genes are present in multiple copies, for example, the four gene alleles encoding 23S rRNA, which is the target for macrolides. Consequently, for most relevant antimicrobials, the presence of many AMR determinants has poor positive predictive value (PPV) for phenotypic resistance in N. gonorrhoeae because, although often essential for AMR, their presence may not be sufficient.10 ,15 ,22 ,24 However, interestingly, by turning this around, absence of known ‘essential’ AMR determinants, that is, detection of ‘wild-type’ genetic determinants, is frequently associated with high PPV for susceptibility, particularly but not exclusively for fluoroquinolones. Indeed, previous work has demonstrated high accuracy for predicting phenotypic susceptibility to ciprofloxacin by identifying absence of the amino acid alteration S91F in the quinolone resistance determining region of GyrA (encoded by the gyrA gene, and a common and dominant AMR determinant for fluoroquinolones,), directly from genital, rectal and pharyngeal clinical samples, previously diagnosed with N. gonorrhoeae by NAAT in the laboratory.21 However, because of significant gene similarities as well as lateral gene transfer between gonococcal and non-gonococcal Neisseria spp., particularly in the pharynx, where gonococci and commensal Neisseria spp. can co-exist for a long time without the onset of symptoms, accurate extragenital detection of gonococcal-specific AMR determinants will be continuously challenging for all assays.15 ,22

The high PPV of detecting absence of essential AMR genotypic determinants for the prediction of antimicrobial susceptibility, in situations such as fluoroquinolones resistance, is important to grasp, as it is a more accurate and a safer strategy than attempting to predict resistance in N. gonorrhoeae by detecting the same AMR genotypic determinants. Accordingly, a false positive result in the former strategy, which means detecting a known essential AMR determinant in a phenotypically susceptible strain, whether this determinant was really there or whether this was due to a poorly specific assay, would result in a safe treatment decision of falling back to standard empirical therapy, although in this case the opportunity of using an active but older drug would be lost. On the contrary, if detection of known essential AMR determinants was used to predict phenotypic resistance, a false negative result, perhaps because of poor sensitivity of the assay, would provide false reassurance of antimicrobial susceptibility in a patient with antimicrobial-resistant N. gonorrhoeae. Clearly, in situations where there were still many unknown essential genotypic determinants of phenotypic resistance, using absence of known determinants as a predictor of phenotypic susceptibility would not also be safe.

A further and vital key to understanding utility of molecular prediction of resistance or, preferably, susceptibility to antimicrobials is the prevalence of circulating phenotypically resistant or susceptible strains in the specific settings where the assays are used. For example, molecular detection of susceptibility to fluoroquinolones might currently be useful to spare use of cephalosporins and/or macrolides in USA and Europe, where ciprofloxacin susceptibility is at around 20%25 and 50%,26 respectively, but completely pointless in, for example, China, Korea and Vietnam, where resistance rates are mainly 100%.27–29

Future perspectives

Rapid accurate POC tests for diagnosis of all STIs would be valuable. However, to significantly affect the STI burden in many settings, particularly for N. gonorrhoeae and M. genitalium infections, a combination of rapid POC diagnostic and AMR testing (POC-AMR) will likely be required. This strategy could combat both STI burden and AMR emergence and spread by enabling diagnosis and individualised treatment at the first healthcare visit, potentially reducing selection pressure on currently recommended antimicrobials, reducing transmission of AMR strains and providing means for AMR surveillance.

Until rapid accurate POC tests for simultaneous detection of pathogen and prediction of AMR are available, validated, accurate and quality-assured laboratory-based molecular assays for prediction of AMR should become commercially available and implemented. These assays should be used to significantly strengthen AMR surveillance in N. gonorrhoeae globally by predicting AMR or susceptibility in NAAT samples or non-viable culture samples. Where applicable, these assays can also be used to guide individualised treatment with at least fluoroquinolones, to spare the use of cephalosporins and/or macrolides.

For M. genitalium, similar molecular assays should be used for surveillance of resistance to macrolides as well as fluoroquinolones internationally. At a minimum, assays predicting macrolide resistance should be used on all M. genitalium-positive clinical samples to spare the use of azithromycin, which is in accordance with the 2016 European guideline on M. genitalium infections.2 Nevertheless, as for N. gonorrhoeae, now is also the time to move from laboratory-based molecular assays towards developing appropriate POC-AMR target product profiles that define users, patient populations and stipulate requirements of performance characteristics and the nature of technologies to be developed. Appropriate validation, quality assurance and confidence in the result of such assays are mandatory and may be achieved through the joint commitments of industry, healthcare and academia to ensure robust premarketing and postmarketing evaluations of assays as well as through national and international expert advisory groups using such evidence to give guidance on appropriate use. The WHO STI POC diagnostic initiative (http://www.who.int/reproductivehealth/ topics/rtis/pocts/en/) aims to facilitate and support access to affordable quality-assured STI POC tests within national STI programmes through generating evidence, developing recommendations and providing technical assistance to WHO Member States and other relevant public health institutions on STI POC testing and POC tests. This WHO initiative also includes the essential component of analytical as well as clinical evaluations of promising POC tests, including the type of tests discussed in this paper, and the initiative is advised by a group of international experts.

A number of currently available NAAT-based POC tests lend themselves to AMR prediction including GeneXpert (despite its limitations in regard to current assay time of around 90 min and cost) and more recently the Atlas Genetics io. Both platforms have real-time multiplex PCR modular capacity, which can be used for detection of pathogen and a limited number of AMR determinants in the same test. Currently, POC tests are being developed on the Atlas Genetics io platform to simultaneously diagnose (1) N. gonorrhoeae and fluoroquinolone susceptibility and (2) M. genitalium and macrolide resistance (personal communication Sadiq: PRECISE study preciseresearch.co.uk). Early results of prototypes will be available soon. However, because new AMR determinants are continuously evolving, molecular AMR prediction is unlikely to completely replace the value of phenotypic AMR detection and surveillance. Regular, quality-assured national and international AMR surveillance programmes for both N. gonorrhoeae and M. genitalium as well as further research regarding correlates between AMR determinants (novel and previously known), in vitro resistance and clinical treatment outcome are imperative. For the extremely slow-growing and fastidious M. genitalium, which may require 5–6 weeks culture and/or co-culture in a Vero cell culture system from clinical samples,10 wider access to quality-assured culturing facilities and further research to improve transportation media and culture methodologies would be valuable. Culturing of contemporary clinical M. genitalium isolates is important to monitor development of new AMR as well as to understand genetic mechanisms behind AMR.10 Also for N. gonorrhoeae, further research to optimise collection and transportation of specimens, as well as culture methodologies, particularly from pharyngeal and rectal sites, is key. This research needs to be prioritised to continue vigilance for novel genotypic–phenotypic AMR associations and associations with treatment failures. Due to evolution of pathogens as well as their AMR in the face of antimicrobial and diagnostic selection pressures,30 for at least some antimicrobials any accurate molecular POC-AMR may need continual updating of the molecular target identity to remain effective, which will be informed by the AMR surveillance and research mentioned above. This will require significant flexibility in industry and in diagnostic regulatory pathways that allow for new targets to be used on platforms without need for extensive, costly and importantly time-consuming clinical validation, as epidemiological changes in prevalence of AMR as well as mutations resulting in AMR happen rapidly. Whole or partial genome sequencing and other novel molecular technologies will most likely provide effective tools for updating of molecular AMR prediction assays as well as revolutionise the molecular AMR prediction for particularly N. gonorrhoeae by themselves. Rapid POC whole genome sequencing technologies, enabling detection of full-length sequences of all genes involved in AMR, may not be so far from being realised given advances in the nanopore-based technologies.31 However, although accuracy for diagnosis and AMR prediction has been demonstrated using conventional next-generation sequencing technologies,32 ,33 nanopore-based technologies may still lack the required accuracy to detect all AMR mutations, confidently enough for a complex resistome such as in N. gonorrhoeae, for the purposes of clinical management.

Finally, costs and cost-effectiveness of using POC-AMR assays will need to be robustly modelled and evaluated taking into account the added value, above the direct costs per patient benefit, that such technologies may bring to bear on antibiotic stewardship and prolonging use of existing antimicrobials.34 Priority needs to be given to affordable POC solutions that can be used in a variety of settings, in high, low and middle income countries as part of the general fight against AMR.

Key messages

  • The development of rapid point-of-care (POC) tests that accurately predict antimicrobial resistance (POC-AMR) would significantly benefit management of Neisseria gonorrhoeae and Mycoplasma genitalium infections.

  • Current POC tests may support the development of POC-AMR particularly for fluoroquinolones in N. gonorrhoeae and macrolides in M. genitalium infections.

  • Further phenotypic–genotypic validation and ongoing phenotypic surveillance are required to establish the current and future molecular determinants of AMR in these infections.

  • Industry and diagnostic regulatory pathways will need to be flexible enough to include current and newly evolved determinants on POC test platforms in a timely fashion.

References

View Abstract

Footnotes

  • Contributors All authors contributed to the manuscript.

  • Funding National Institute for Health Research, 10.13039/501100000272, grant number II-LB-0214-20005.

  • Competing interests STS is grant holder for the National Institute for Health Research (NIHR) i4i Programme (grant number II-LB-0214-20005). The views expressed are those of the authors and not necessarily those of the NIHR, the NHS or the UK Department of Health. STS has also received funding from Atlas Genetics to conduct performance evaluations of its io POC system. None for MU and FM.

  • Provenance and peer review 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.