Optimising treatments for sexually transmitted infections: surveillance, pharmacokinetics and pharmacodynamics, therapeutic strategies, and molecular resistance prediction

Abstract

Progressive antimicrobial resistance in Neisseria gonorrhoeae, Mycoplasma genitalium, and Trichomonas vaginalis has created a pressing need for treatment optimisations for sexually transmitted infections (STIs). In this Review, we aim to highlight urgent needs in global STI management, including: (1) improved surveillance to monitor antimicrobial resistance and clinical outcomes; (2) systematic pharmacokinetic and pharmacodynamic evaluations to ensure resistance suppression and bacterial eradication at all sites of infection; (3) development of novel, affordable antimicrobials; and (4) advancements in new molecular and point-of-care tests to detect antimicrobial resistance determinants. Antimicrobial resistance among STIs is a global public health crisis. Continuous efforts to develop novel antimicrobials will be essential, in addition to other public health interventions to reduce the global STI burden. Apart from prevention through safer sexual practices, the development of STI vaccines to prevent transmission is a crucial research priority.

Introduction

Antimicrobial resistance among sexually transmitted infections (STIs) has amplified the importance of optimising treatment strategies. Neisseria gonorrhoeae and Mycoplasma genitalium have developed clinical resistance to all available therapeutic antimicrobials, and Trichomonas vaginalis has shown clinical resistance to recommended regimens.^1-5 The impact of progressive antimicrobial resistance in STIs is considerable given an estimated incidence of 87 million gonorrhoea cases and 156 million trichomoniasis cases in 2016,⁶ and a substantial number of M genitalium infections worldwide, for which no global estimates exist. Bacterial resistance to recommended regimens is uncommon in Chlamydia trachomatis.⁷ However, clinical treatment failures are occurring more frequently with azithromycin than with doxycycline in patients with chlamydial infection, particularly rectal infections.⁸ Treponema pallidum has also developed resistance to azithromycin and results in clinical treatment failure,⁹ making this antimicrobial suboptimal as an alternative therapy.

Evidence-based STI treatment guidelines have been developed by WHO, International Union against STIs (IUSTI)-European Branch, US Centers for Disease Control and Prevention (CDC), and other national organisations.^10-12 However, there are concerns regarding the recommended therapy for gonorrhoea with ceftriaxone, even when combined with azithromycin, in light of the reported progressive antimicrobial resistance to both drugs.¹² N gonorrhoeae has an extraordinary capacity to develop and acquire novel antimicrobial resistance determinants to different antimicrobials including extended-spectrum cephalosporins (ESCs). In 2016, the first global treatment failure with dual therapy (ceftriaxone plus azithromycin) was reported in the UK.³ In 2018, the first three isolates with combined ceftriaxone resistance and high-level azithromycin resistance (ie, a minimum inhibitory concentration [MIC] of >256 mg/L) were identified in the UK and Australia.^13,14

Azithromycin is recommended for M genitalium internationally.¹¹ However, high degrees of azithromycin resistance in M genitalium reported in Europe, the USA, Japan, and Australia question azithromycin’s sustained activity against this STI.⁴ M genitalium resistance to alternative regimens (eg, moxifloxacin) has also been reported internationally.^4,11 Finally, resistance to the recommended regimens for trichomoniasis has been verified in the USA.⁵

Given the heightened awareness of STI antimicrobial resistance, an expert consultation to discuss approaches to STI treatment was held in Washington, DC, USA, in April, 2018. We explored key issues, including surveillance for antimicrobial resistance in STIs, drug pharmacokinetics and pharmacodynamics, effectiveness of single-dose monotherapies, dual or combination regimens, and multidose therapies, molecular tests including point-of-care (POC) tests for antimicrobial resistance prediction, and novel antimicrobials for STIs. The meeting objectives were to inform the public health and research communities regarding crucial gaps and next steps needed to improve STI management, ensure effective treatment at all anatomical sites of infection, conserve existing treatment options, and develop strategies to prevent the spread of antimicrobial resistance among STI pathogens. In this Review, we summarise key findings from this meeting and next steps for the control of STIs globally.

The burden of antimicrobial resistant STIs

Global surveillance for STI antimicrobial resistance

In 2016, WHO clearly defined N gonorrhoeae as a priority STI because of the risk of antimicrobial resistance leading to untreatable infections.¹⁵ N gonorrhoeae antimicrobial susceptibility has been monitored through the WHO Gonococcal Antimicrobial Surveillance Programme (GASP), representing a laboratory network of 70–80 countries worldwide.¹ According to GASP data, 32 countries reported resistance or decreased susceptibility of N gonorrhoeae to ceftriaxone or cefixime, or both, in 2015–16, which was a decrease from 2009–14. The proportion of countries reporting gonococcal resistance to azithromycin (81–83%) and ciprofloxacin (97–100%) has been high from 2014 to 2016.¹ However, the full magnitude of the gonococcal antimicrobial resistance problem is unclear because of the absence of data from many countries. The gaps are notable in areas with the highest gonorrhoea incidence, which could be due to insufficient aetiological diagnosis, capacity, and financial resources. Furthermore, the absence of universally accepted and quality-assured antimicrobial resistance testing methodologies and clinical resistance breakpoints for ESCs and azithromycin remains a challenge in gonococcal surveillance internationally.¹

There are no international surveillance systems in place for antimicrobial resistance monitoring in M genitalium or T vaginalis. However, M genitalium azithromycin resistance exceeds 40% in many countries with available data, and increasing moxifloxacin resistance has been reported worldwide.⁴

European surveillance for STI antimicrobial resistance

The European Gonococcal Antimicrobial Surveillance Programme (Euro-GASP) performs annual sentinel surveillance of gonococcal antimicrobial resistance across the European Union (EU) and European Economic Area (EEA). In 2016, no isolates with ceftriaxone resistance were detected from 25 countries. However, the proportion of gonococcal isolates with decreased susceptibility to ceftriaxone increased from 15·0% in 2015 to 17·7% in 2016. Both resistance to cefixime and resistance to azithromycin have been stable in the EU and EEA since 2014. In 2016, cefixime resistance was detected in 2·1% of N gonorrhoeae isolates compared with 1·7% in 2015, and azithromycin resistance was detected in 7·5% compared with 7·1%. Resistance to ciprofloxacin was at 46·5% in 2016, compared with 49·4% in 2015.¹⁶

US surveillance for STI antimicrobial resistance

The US Gonococcal Isolate Surveillance Project (GISP)¹⁷ has found a stable proportion of isolates with elevated ceftriaxone MICs or reduced ceftriaxone susceptibility (0·3% in 2016 and 0·4% in 2017) and elevated cefixime MICs (0·3% in 2016 and 0·2% in 2017). There was an increase in the proportion of isolates with azithromycin resistance from 2·5% in 2014 to 4·4% in 2017.¹⁷

The Sexually Transmitted Diseases Surveillance Network, which includes six sexual transmitted disease clinic sites throughout the USA, reported a 4·3% prevalence of low-level metronidazole resistance in T vaginalis in 2009–10.¹⁸ Both high-level metronidazole and high-level tinidazole resistance have been identified among T vaginalis isolates.⁵ Although high-level nitro-imidazole resistance is rare, it has been associated with clinical treatment failure.⁵

Priorities in antimicrobial resistance surveillance for STIs

WHO has identified a range of strategies for the global health sector (panel), which include information for focused action, interventions for impact in the areas of prevention and care, package interventions for impact, delivering for equity, financing, and innovation for acceleration. Given the increasingly numerous reports of gonococcal resistance or decreased susceptibility to ESCs, azithromycin, fluoroquinolones, and other antimicrobials, key priorities are to increase the number of representative gonococcal isolates, link epidemiological and clinical data (including treatment outcomes), and expand antimicrobial resistance surveillance globally—particularly in regions with scarce data such as Africa, Central America and the Caribbean, and central and southeast Asia.^1,2 This expansion will require continuous international support, and political and financial commitment. Incorporating M genitalium antimicrobial resistance surveillance in an expanded system of laboratories that already provide specimens for gonococcal surveillance would be valuable. Accordingly, enhanced antimicrobial resistance surveillance for STIs, including N gonorrhoeae, M genitalium, and T vaginalis, with the ability to monitor for treatment failures, is essential. Guidance on the surveillance threshold levels that should prompt changes in STI treatment guidelines, in the context of ineffective treatment options, will be crucial to preventing transmission of STIs, including antimicrobial resistant strains.

Pharmacokinetic and pharmacodynamic considerations for STIs

Pharmacological factors affecting antimicrobial efficacy

Dosage optimisations of antimicrobials for STIs require a thorough understanding of their pharmacokinetic and pharmacodynamic (PK–PD) relationships, in light of antimicrobial resistance in N gonorrhoeae and M genitalium, and the variation in treatment efficacy between anatomical sites of infection.^19,20 The key PK–PD parameters important to predicting the clinical efficacy of antimicrobials can vary by antimicrobial class. Such parameters include the time of so-called free or unbound drug concentration above the MIC (fT>MIC), the ratio of maximum drug concentration to MIC (C_max/MIC), and the ratio of the area under the concentration curve at 24 h to MIC (AUC_0–24/MIC; figure 1).²¹ For example, β-lactam antimicrobial activity is predominantly time dependent (fT>MIC), whereas the activity of macrolides, tetracyclines, fluoroquinolones, and nitroimidazoles depends on concentration (C_max/MIC and AUC_0–24/MIC).^20,21 Although macrolides and tetracyclines share a common PK–PD driver with the fluoroquinolones (AUC/MIC), the reasons for the association between drug exposure relative to the MIC and bacterial kill rate are distinct. For macrolides, there is a prolonged post-antibiotic or persistent effect, in which a constant rate of kill occurs while the drug concentration at the effect site remains above a critical level. As it declines below the critical level, bacterial death ceases but the bacterium does not immediately go back into normal growth. By contrast, fluoroquinolones have a more modest post-antibiotic effect, but increasing the magnitude of the exposure produces an increase in the bacterial kill rate.

Figure 1: Pharmacokinetic parameters for predicting the clinical efficacy of antimicrobial agents.

AUC_0–24/MIC=ratio of area under the concentration curve at 24 h to MIC. C_max=maximum drug concentration. ESC=extended-spectrum cephalosporin. fT>MIC=free-drug concentration above MIC. MIC=minimum inhibitory concentration. MTC=maximum tolerated concentration. t_max=time to C_max.

C_max and AUC can be affected by antimicrobial dose, volume of distribution of the antimicrobial tissue concentration relative to blood concentration, and bioavailability. For azithromycin, single oral dose regimens can achieve a higher C_max and similar AUCs than longer courses of the same dose. Furthermore, single-dose azithromycin regimens have been shown to achieve bacterial clearance more rapidly than longer courses of the same total dose in preclinical infection models.²² However, diarrhoea and vomiting can affect a drug’s bioavailability and reduce C_max and AUC.²³

Several other factors can have an effect on antimicrobial efficacy. First, the drug must be able to penetrate the tissue at the site of infection and achieve sufficient concentrations in the relevant compartments. For example, C trachomatis only replicates intracellularly so it is crucial that drugs achieve the necessary intracellular concentrations to eliminate the organism.²⁴ Intracellular concentrations are particularly relevant for azithromycin, which has a long half-life (68 h) due to ion trapping within the cell resulting in subinhibitory concentrations in the extracellular space weeks after infection.²⁴ Dual therapy with azithromycin and ceftriaxone, an extracellular drug with a 6–8 h half-life, can thus target N gonorrhoeae and M genitalium in both spaces. However, the drugs’ different characteristics at sites of infection not only affect antimicrobial efficacy but can also result in selective pressure for the development of resistant organisms. Additionally, the degree of protein binding to the drug is important, as only free drug is pharmacologically active and able to penetrate cells, unless there is a transporter protein present. Unlike most other antimicrobials, protein binding for azithromycin is dose dependent and reaches 50% at low serum concentrations (<500 mg dose), decreasing to 12% at higher doses (>1000 mg).²⁵ This suggests that, at high concentrations, protein binding could become saturated, resulting in more free drug, depending on the dose and the state of plasma proteins. By contrast, ceftriaxone protein binding reaches over 90%, with a smaller degree of variation in the amount of plasma protein binding that is dependent on concentration.^26,27

Other factors that can affect antimicrobial efficacy include: drug solubility; pH at site of infection; inflammatory response, particularly the uptake of antimicrobials such as azithromycin by phagocytes; site-specific behaviours such as douching, which can damage the epithelium and remove drug-laden cells; and organism burden at the site of infection.^24,28

Antimicrobial efficacy for oropharyngeal STIs

For many antimicrobials, efficacy is lower for oropharyngeal gonorrhoea than at other sites of infection.¹⁹ This fact is particularly concerning because the oropharyngeal niche is an ideal environment for generating antimicrobial resistance through horizontal transfer of genetic material. However, data regarding PK–PD measures of antimicrobials in the oropharyngeal sites for STIs are scarce. Infection of the oropharynx is usually asymptomatic with minimal inflammation.²⁹ Some antimicrobials are transported via phagocytes, which could also have an effect on antimicrobial penetration and efficacy at this site.

N gonorrhoeae has been identified in the tonsils, tonsillar crypts, tonsillar exudate, and saliva;³⁰ similar gonococcal burdens have been measured in the pharynx and saliva.³¹ One study suggests that a substantial population-attributable fraction of rectal gonorrhoea cases might be due to the use of partners’ saliva for lubrication during anal sex.³¹ Therefore, intracellular drug concentrations, antimicrobial tonsillar-to-plasma ratios, and potentially saliva-to-plasma ratios might be relevant measures for drug penetration into the oropharynx. In addition to azithromycin, ciprofloxacin reaches high intracellular concentrations, but ceftriaxone and gentamicin achieve relatively low intracellular concentrations.^32-35 Azithromycin and gentamicin have low protein-binding properties and good saliva drug concentrations (saliva-to-plasma ratios of 6·0 for azithromycin and 1·0 for gentamicin), whereas ciprofloxacin has moderate saliva drug concentrations (saliva-to-plasma ratio of 0· 5).^25,36 Azithromycin also has an excellent tonsillar-to-plasma ratio of 150.²⁵ By contrast, ceftriaxone has a poor saliva-to-plasma ratio (<0·004), and a poor tonsillar-to-plasma ratio (0·2).^27,37

Priorities in PK–PD factors and oropharyngeal STIs

Many research gaps exist regarding PK–PD parameters and other factors that affect antimicrobial efficacy for STIs. For example, the effects of douching on drug removal, disruption of mucosal surfaces, and the microbiome at sites of infection are largely unknown. Given the potential contribution of oropharyngeal infections to gonococcal antimicrobial resistance development, we need an in-depth understanding of the microbiology and natural history of gonorrhoea and other STIs at this site. This understanding could be facilitated by the development of an oropharyngeal gonorrhoea human challenge model. Furthermore, we need a rationale for the efficacy of ceftriaxone despite poor saliva-to-plasma and tonsillar-to-plasma ratios. Studies to explore drug dosage, duration of therapies, and suppression of emerging antimicrobial resistance with drugs such as ESCs and azithromycin would assist in optimising treatment regimens. Comprehensive PK–PD studies and optimised simulations with current and future antimicrobials, especially for oropharyngeal STIs, are essential. For these PK–PD studies, appropriate animal models, epithelial cell models to evaluate intracellular activity of drugs, and in-vitro dynamic hollow-fibre bioreactor models would be valuable.

Relationship between drug exposure and response

Despite the idea that gonorrhoea dual therapy (ceftriaxone and azithromycin) might hinder the development or spread of antimicrobial resistance, no effective resistance suppressive therapies are available for STIs. In traditional exposure–response relationships, there is a direct link between drug exposure and bacterial kill rate. With antimicrobial resistance suppression, the shape of the relationship between drug exposure and response resembles an inverted U (figure 2).^39,40 Low exposures cause little selective pressure and resistance amplification, whereas higher exposures result in increased selective pressure by causing more damage to fully susceptible bacterial populations than less-susceptible populations. When a maximal size of the less-susceptible bacterial population is achieved, higher drug exposures will begin to affect the antimicrobial resistant subpopulations.

Figure 2: Inverted U plot illustrating the relationship between drug exposure and response in resistance suppression.

CFU=colony forming units. AUC_0–24/MIC=ratio of area under the concentration curve at 24 h to MIC. Reproduced from Drusano and colleagues,³⁸ by permission of the American Society for Microbiology.

Challenges to antimicrobial resistance suppression in STIs include the presence of multiple infections and sites of infection that can restrict the ability to treat them with single agents, and different selective pressures for antimicrobial resistance on specific pathogens. For example, the azithromycin MIC for N gonorrhoeae is much higher than the azithromycin MIC for M genitalium, and N gonorrhoeae contains four alleles of the target 23S rRNA gene compared with only one 23S rRNA gene allele in M genitalium.^{24, 41-43} Therefore, the antimicrobial resistance selection pressure on N gonorrhoeae could be weaker than for M genitalium after azithromycin treatment.

STI antimicrobial monotherapies

Antimicrobials need to generate efficient bacterial kill rates at effect sites, which might change substantially over time into concentration-dependent or concentration-independent killing. When single-dose monotherapies are used for STIs, bacterial burdens at sites of infection must be low enough so that eradication endpoints can be achieved. This point is illustrated by a study that treated group A streptococcal pharyngitis with 500 mg of ceftriaxone, in which free tonsillar ceftriaxone concentrations above the MIC would have to be present for 36 h to achieve optimal bacterial eradication rates in the oropharynx.³⁷ Unfortunately, no single dose of ceftriaxone could achieve this effect because of variability in drug clearance and volume (half-life) between patients.

Additional concerns with single-dose monotherapies are error-prone replication and pharmacokinetic mis-matching.⁴⁴ Fluoroquinolones were once extremely potent against gonococci but are no longer recommended because of the emergence of antimicrobial resistance. Fluoroquinolones inhibit DNA topoisomerases required for replication, but they also result in error-prone replication by inducing the stress (SOS) response involved in DNA repair, recombination, and mutagenesis.^38,44 Low-fidelity polymerases induced during the SOS response lack error-checking abilities, and enhance the emergence of fluoroquinolone resistance mutations. Other drugs like some β-lactam antimicrobials can also induce error-prone replication. For fluoroquinolones, resistance emergence has occurred because of differences in bacterial burdens and variance in pharmacokinetics between patients.^38,44 The shorter half-life of ciprofloxacin (4 h) probably results in drug concentrations that are low enough at the effect site with smaller kill rates but still higher than the error-prone induction concentration. These factors should be taken into consideration when recommending single-dose therapies for bacterial STIs.

Dual antimicrobial therapy for STIs

Combination therapies can be advantageous when there is a synergistic or, at least, an additive interaction, so that both the bacterial kill rate is improved and the likelihood of suppressing antimicrobial resistance is increased.^38,45 Mathematical modelling using murine models or in-vitro pharmacokinetic models with dynamic hollow-fibre bioreactors has been beneficial in evaluating drug interactions in combination therapies for non-STIs. For example, hollow-fibre bioreactor models have shown that moxifloxacin and rifampicin administered daily over 7 days are antagonistic with respect to kill rates for Mycobacterium tuberculosis, but can suppress amplification of less-susceptible organisms.⁴⁴ However, there is a major mismatch in drug half-lives when both drugs are coadministered for only 5 out of 7 days, resulting in sufficient moxifloxacin to continue induction of error-prone replication but insufficient rifampicin to suppress antimicrobial resistance.⁴⁴

Parametric approaches can assist in analysing drug combination therapies, using an interaction term (α) that determines the drug interaction. If α is positive and the lower 95% confidence boundary does not overlap with zero, the interaction represents significant synergy. In the case of dual therapy for multiple STIs, α values would need to be positive for all targeted bacterial populations to maximise kill rates and effectively suppress the amplification of antimicrobial resistance. Because this approach is fully parametric, Monte Carlo simulations can then be used to identify drug effects for a population of patients.^40,45

Priorities for antimicrobial resistance suppression approaches for STIs

The emergence of antimicrobial resistance among bacterial pathogens is fully stochastic. Therefore, more data are needed on pathogen-specific bacterial burdens, antimicrobial resistance mutational frequencies, exposures that optimise bacterial kill rate and extent, antimicrobial resistance suppression exposures, and effect site penetration, especially for the oropharynx, to model optimal therapies for STIs. Optimising single or combination therapies to achieve rapid kill rates and suppression of the emergence of antimicrobial resistance among bacterial pathogens is extremely challenging because these outcomes represent different goals of therapy. In particular, there can be multiple STI pathogens at various sites, and multiple pathways by which the pathogens can become resistant to the antimicrobials being used. To design dual therapies, two different concentration–time profiles at all infection sites need to be tracked, while monitoring the effect of the antimicrobials on both susceptible (wild-type) bacterial populations and subpopulations that have acquired an elevated MIC value.

Molecular tests for prediction of antimicrobial resistance in STIs

Point-of-care tests

WHO is supporting the development and validation of novel POC tests for STIs that are accurate, rapid, cost-effective, and can facilitate same-day visits for evaluation, testing and treatment.⁴⁶ Increased availability of these tests can improve screening for susceptible patients while reducing over-treatment, incorrect or unnecessary antimicrobial prescription, and the resulting emergence of antimicrobial resistance associated with syndromic STI management. POC tests based on antigen detection by lateral flow or optical immunoassay formats for N gonorrhoeae, C trachomatis, or M genitalium lack sufficient sensitivity.⁴⁶ However, rapid POC tests with high sensitivities and specificities for these STIs, based on the nucleic acid amplifications test (NAAT), might soon be available.⁴⁶ Rapid and sensitive POC tests that can also detect antimicrobial resistance determinants to predict treatment susceptibility in STIs would be invaluable. Based on mathematical transmission modelling predictions, POC tests including antimicrobial resistance prediction for N gonorrhoeae could more effectively avert cases and slow the spread of antimicrobial resistance than N gonorrhoeae NAATs or cultures could alone.^47,48

Antimicrobial resistance determinants for N gonorrhoeae and M genitalium

Many antimicrobial resistance determinants have been identified in N gonorrhoeae that confer decreased susceptibility or complete resistance to different antimicrobials.⁴⁹ The main mutation associated with ciprofloxacin resistance is in the gyrA gene encoding a Ser91Phe alteration in the GyrA subunit of DNA gyrase, which correlates with low to intermediate resistance.^49,50 Concomitant mutations in GyrA amino acid position Asp95 and in the parC gene, which encodes the ParC subunit of topoisomerase IV, are associated with high ciprofloxacin resistance.^49,50 For azithromycin, moderate to high resistance has been associated with the C2611T (MIC>1–2 mg/L) and the A2059G (MIC≥256 mg/L) target mutations in several of the four 23S rRNA alleles. However, mutations causing an overexpression of the MtrCDE efflux pump can also cause azithromycin resistance.^49-51 Gonococcal antimicrobial resistance determinants for ESCs are more complex and can involve many genes and mutations. Decreased susceptibility and complete resistance to ESCs are associated mainly with mosaic penA alleles encoding mosaic penicillin-binding protein 2 (PBP2; lethal target for β-lactams). However, there are various mosaic penA alleles, causing very different ESC MICs.^49,50 Mutations that cause an overexpression of the MtrCDE efflux pump, resulting in increased drug efflux, as well as mutations in the porB1b allele, encoding alterations in amino acid 120 (eg, Gly120Lys in the porin PorB1b) and 121 (eg, Ala121Asp) that decrease drug influx, further increase the MICs of antimicrobials (including ESCs and azithromycin).^49-51

In M genitalium, azithromycin resistance is primarily caused by target mutations in nucleotide A2058 or A2059 (E coli numbering) in region V of the 23S rRNA gene.^4,52 Moxifloxacin resistance is caused by target mutations in the quinolone-resistance-determining region (QRDR) of the parC gene, primarily in ParC amino acid positions Ser83 and Asp87 (M genitalium numbering). However, the correlations between several reported mutations in the parC QRDR and increased moxifloxacin MICs and clinical treatment failure have not yet been established.^4,53,54

Molecular antimicrobial resistance prediction assays for N gonorrhoeae and M genitalium

Many real-time PCR assays have been developed for detection of antimicrobial resistance determinants for ciprofloxacin, azithromycin, or ESCs as single agents in N gonorrhoeae.^49,50 The first commercially available assay with a European CE marking for in-vitro diagnostics (CE–IVD), the ResistancePlus NG Cipro assay (SpeeDx, London, UK), targets opaA and porA for detection of N gonorrhoeae, the gyrA wild-type (Ser91), and resistance mutation (Ser91Phe).⁵⁰ This test has high sensitivities and specificities for gonococcal detection and prediction of ciprofloxacin susceptibility or resistance.⁵⁰ A few molecular assays that detect antimicrobial resistance determinants for multiple antimicrobials have also been developed, with promising results for examining gonococcal isolates and gonococcal-positive NAAT specimens.⁵⁰ Unfortunately, many molecular methods for gonococcal antimicrobial resistance prediction have shown suboptimal specificities due to cross-reaction with commensal Neisseria species inhabiting extragenital sites, or have not undergone validation for clinical NAAT specimens, or both. Furthermore, test sensitivities and specificities for the prediction of resistance to azithromycin and particularly to ESCs have been suboptimal because of the large number of antimicrobial resistance determinants involved.^49,50

For M genitalium, it is recommended that molecular testing for macrolide resistance is done on all positive samples to guide therapy, with subsequent use of azithromycin for macrolide-susceptible infections and alternative treatment (eg, moxifloxacin) for macrolide-resistant M genitalium infections.^11,54,55 The CE–IVD-licensed M genitalium ResistancePlus (SpeeDx) is available in Europe and Australia.⁵⁵ The MGres kit (Diagenode, Liege, Belgium) and RealAccurate TVMGres PCR kit (Pathofinder, Maastricht, Netherlands) have also received CE–IVD marking for use in the European market.⁵⁶ Mutations mediating moxifloxacin resistance can also be detected; however, the correlations between several parC mutations and moxifloxacin resistance are not completely understood and no commercial assay currently exists.⁵⁴

Newer technologies such as whole genome sequencing have the potential to identify all known and novel antimicrobial resistance determinants and improve the ability to predict antimicrobial resistance. Whole genome sequencing of gonococcal isolates has been shown to predict susceptibilities within plus or minus one MIC dilution for five antimicrobials approximately 93% of the time.⁵⁷ However, the error rate widely differed for antimicrobials and was much higher for ESCs because of their multiple antimicrobial resistance determinants.⁵⁷ Rapid, real-time sequencing using the small hand-held sequencer MinION (Oxford Nanopore Technologies, Oxford, UK) has been shown to accurately predict decreased susceptibility to ciprofloxacin, azithromycin, and ESCs in gonococcal isolates.⁵⁸ The MinION could have future application for whole genome sequencing of gonococcal antimicrobial resistance determinants directly in clinical non-viable NAAT samples if phenotypic antimicrobial resistance determination is not feasible, because of the scarcity of N gonorrhoeae cultures in many settings.

Priorities for STI antimicrobial resistance prediction

Rapid, sensitive, and specific molecular tests, including POC tests, that predict resistance or susceptibility to multiple antimicrobials in clinical NAAT samples are imperative to guiding individualised therapy for both N gonorrhoeae and M genitalium infections. However, we need to continuously identify novel antimicrobial resistance determinants for existing therapies and newer antimicrobials (eg, gepotidacin and zoliflodacin).^46,59 Additional research to explore the induction and selection of antimicrobial resistance determinants, their spread, biological fitness, effect on MICs, and correlation with clinical outcomes is a high priority. Newer technologies such as whole genome sequencing, nanotechology, and microfluidics for prediction of STI antimicrobial resistance in clinical NAAT specimens deserve further attention.

Additional antimicrobial strategies for treatment optimisations

Multidose regimens

Administration of multidose antimicrobial regimens could be a rational strategy to improve clinical efficacy for M genitalium infections, trichomoniasis, and possibly gonorrhoea. A meta-analysis of studies involving a 5-day course of azithromycin reported that extended dosing might be more effective and less likely to cause resistance, compared with azithromycin monotherapy for treatment of M genitalium urethritis.^60,61 However, another study found no significant difference in cure rates between extended and single-dose azithromycin therapy, nor in the selection of post-treatment macrolide resistance.⁶² Resistance-guided sequential M genitalium therapy has had favourable outcomes, in which initial empirical therapy with doxycycline is followed by multidose azithromycin (if testing confirms susceptibility to macrolides) or sitafloxacin (if testing suggests macrolide resistance).⁶³ The investigators reported high cure rates (92%), which could have been due, in part, to a reduction in the bacterial burden of M genitalium with doxycycline.⁶³

For T vaginalis, multidose metronidazole therapy has been shown to result in lower repeat infection rates than a single dose among women with or without HIV infection.^64,65 Although the exact reasons for this finding are unclear, concerns regarding metronidazole resistant T vaginalis (~4–9%) or high re-infection rates from untreated partners support the use of multidose therapy for women, coupled with partner therapy.

Novel antimicrobials

Several novel antimicrobials are under clinical investigation as potential therapeutic options for gonorrhoea, and some also seem to be active against M genitalium infections and trichomoniasis (table).^{4,46,59,66-79} A phase 3 randomised controlled clinical trial (RCT) with the first fluoroketolide, solithromycin, for gonorrhoea was published in 2019.⁷¹ In the microbiological intention-to-treat (mITT) population, solithromycin 1·0 g monotherapy showed microbiological cure in only 80·5% of urogenital gonorrhoea infections.⁷¹ Gepotidacin, a triazaacenaphthylene topoisomerase II inhibitor, was more than 95% effective as monotherapy for uncomplicated urogenital gonorrhoea in a phase 2 RCT.⁶⁹ However, three of 69 participants showed clinical treatment failure, because of gonococcal strains with a pre-existing ParC Asp86Asn mutation;⁶⁹ post-treatment isolates from two treatment failures had an additional Ala92Thr resistance mutation in GyrA, which had been selected for during treatment.⁷⁰ Zoliflodacin, the first-in-class spiropyrimidinetrione (a topoisomerase II inhibitor), was shown to have greater than 96% microbiological cure rates in the mITT population in a phase 2 trial, with zoliflodacin 3·0 g monotherapy for urogenital and rectal gonorrhoea, but a lower cure rate (nine [82%] of 11) for pharyngeal infections.⁶⁶ A phase 3 RCT with zoliflodacin was initiated in October, 2019, in Europe, South Africa, Thailand, and the USA.⁸⁰ High in-vitro activities of solithromycin,⁷² gepotidacin,⁶⁸ and zoliflodacin⁶⁷ against M genitalium have also been reported. Additionally, lefamulin (a pleuromutilin) has been shown to have high in-vitro activity against multiple STIs, including N gonorrhoeae⁷³ and M genitalium (table).⁷⁴ Finally, several drugs in known antimicrobial classes look promising for the treatment of M genitalium infections and trichomoniasis.^75-79 For example, pristinamycin has been used for treatment of M genitalium infections in Australia and Europe, and sitafloxacin has been used in Japan; pristinamycin is well tolerated but sitafloxacin should be used with caution because of potential side-effects that include phototoxicity.^75-78

Table:

Potential future antimicrobials for Neisseria gonorrhoeae, Mycoplasma genitalium, or Trichomonas vaginalis

	Class	Validated activity	Resistance mutations	Clinical study	Availability
Zoliflodacin66,67	Spiropyrimidinetrione (topoisomerase II inhibitor)	N gonorrhoeae, M genitalium, C trachomatis	GyrB Asp429Ala*, Asp429Asn*, Asp450Asn*, Asp 450Thr*	Phase 3 clinical trial initiated in 2019 (gonorrhoea)	Not yet available
Gepotidacin68-70	Triazaacenaphthylene (topoisomerase II inhibitor)	N gonorrhoeae, M genitalium	ParC Asp86Asn† GyrA Ala92Thr†	Phase 2 clinical trial finished (gonorrhoea)	Not yet available
Solithromycin71,72	Fluoroketolide	N gonorrhoeae, M genitalium, C trachomatis	Ala2058Gly†, Ala2059Gly† in 23S rRNA	Phase 3 clinical trial finished (gonorrhoea)	Not yet available
Lefamulin73,74	Pleuromutilin	N gonorrhoeae, M genitalium, C trachomatis	None reported	In vitro only	Not yet available
Pristinamycin75,76	Streptogramin	M genitalium	None reported	Prospective cohort	Europe, Australia (special access)
Sitafloxacin77,78	Fourth-generation fluoroquinolone	M genitalium, N gonorrhoeae, C trachomatis	Affected by conventional fluoroquinolone resistance mutations	Prospective cohort	Japan
Secnidazole79	Nitroimidazole	T vaginalis	None reported	Phase 3 clinical trial ongoing (trichomoniasis)	Not yet available

^*Induced or selected in laboratory strains.

^†Identified in clinical strains.

Research and public health implications

The salient points essential to optimising current and future antimicrobial therapies for STIs are summarised in figure 3. First, improved global and national antimicrobial susceptibility surveillance programmes for N gonorrhoeae and M genitalium are needed, ideally linked to epidemiological data of patients and whole genome sequencing. The resulting information can be correlated with clinical outcomes and used to detect antimicrobial resistance trends and inform treatment guidelines and public health programmes. Secondly, current STI therapies need systematic PK–PD evaluations to ensure effective bacterial killing and antimicrobial resistance suppression at all sites of infection, particularly in the oropharynx. Azithromycin in a single high dose can achieve a high C_max and AUC with excellent saliva-to-plasma and tonsillar-to-plasma ratios. However, because of azithromycin’s long half-life, any new infections with N gonorrhoeae or M genitalium up to 4 weeks after therapy, especially in the oropharyngeal site, might be exposed to subinhibitory concentrations that could potentially promote induction or selection of macrolide resistance.^81-84 Consequently, patients treated with azithromycin should avoid unprotected sexual intercourse for a minimum of 2 weeks after treatment. A high dose of ceftriaxone (1·0 g) has been estimated in Monte Carlo simulations to achieve a median fT>MIC of 24·3 h for a gonococcal strain with an MIC of 0·5 mg/L, which was the basis for recommending this dose in the UK.⁸⁵ However, the lower boundary of the 95% CI in this scenario is an fT>MIC for ceftriaxone of only 11·1 h, illustrating that ceftriaxone 1·0 g monotherapy is not sufficient to cure a substantial proportion of cases caused by gonococcal strains with ceftriaxone MICs of 0·5 mg/L. Additionally, ceftriaxone has poor saliva-to-plasma and tonsillar-to-plasma ratios, despite its reported high cure rates for pharyngeal gonorrhoea. Experimental studies and mathematical modelling experiments are important to understand the interactions and dynamic effects of ceftriaxone and azithromycin as dual therapy and in monotherapies, their gonococcal and M genitalium kill rates, antimicrobial resistance suppression, and optimal dosing. Third, development of novel antimicrobials with activity against STIs, and conservation strategies (eg, specific indications and use of antimicrobials only when there are no pre-existing resistance mutations) are crucial. Optimal doses and appropriate administration regimens of these new agents as single-dose or combination therapies should be determined using PK-PD data and Monte Carlo simulations. Lastly, advancements in rapid, sensitive, and specific POC tests and molecular assays that are cost-effective and can identify the STI causal agents are needed, as well as multiple antimicrobial resistance determinants to guide individualised therapy and minimise syndromic management.

Figure 3: Key priorities for STI treatment optimisations.

POC=point-of-care. PD=pharmacodynamics. PK=pharmacokinetics. STI=sexually transmitted infection.

Antimicrobial resistance among STIs is a global public health crisis. In 2015, WHO released a global action plan on antimicrobial resistance that included five fundamental objectives: (1) improve awareness and understanding of antimicrobial resistance; (2) strengthen knowledge through surveillance and research; (3) reduce the incidence of infection; (4) optimise the use of antimicrobial agents; and (5) ensure sustainable investment in countering antimicrobial resistance.⁵⁹ The development of antimicrobial stewardship for STIs would support these objectives by coordinating programmes to promote accurate diagnosis and appropriate use of antimicrobials, improving patient outcomes, reducing antimicrobial resistance, and decreasing the spread of infections caused by multidrug-resistant organisms. Using similar approaches to antimicrobial stewardship already applied to other infectious diseases would be a judicious strategy to facilitate multidisciplinary collaborations to optimise STI therapies.

Undoubtedly, because pathogens such as N gonorrhoeae and M genitalium can develop or acquire more antimicrobial resistance mutations, clinical resistance will probably also develop to new therapies. Therefore, continuous efforts to develop novel antimicrobials will be essential, in addition to other public health interventions to reduce the global STI burden. Apart from prevention through safer sexual practices, the development of STI vaccines to prevent transmission is a crucial research priority and might be the only sustainable solution for effective control of STIs including antimicrobial resistant pathogens.

Panel: Modified WHO Global Health Sector Strategies for STIs¹⁵*.

Information for focused action

• STI surveillance globally

• STI estimates for countries

• Gonococcal antimicrobial surveillance programmes globally, ideally also for Mycoplasma genitalium

Interventions for impact

Prevention and care

• Sexual health education, human papillomavirus vaccine, condoms

• Early diagnosis and treatment

• Partner notification, testing, and treatment

Package interventions for impact

• Elimination of mother-to-child transmission of syphilis and HIV

• Human papillomavirus vaccines

• Control of gonococcal and M genitalium antimicrobial resistance

Delivering for equity

• Integration of STI services with primary health care, sexual and reproductive health, maternal and child health, adolescent health, and HIV care

• Extension of reach to key and specific populations

• Community involvement

• Accessibility and quality of care

Financing

• Development of investment case

• Support for financial protection schemes

• Reduction of commodity prices

• Increase in procurement efficiency

Innovation for acceleration

• STI vaccine development

• Point-of-care tests

• New STI treatment options

STI=sexually transmitted infection. *We have extended this list to include M genitalium.

Search strategy and selection criteria.

We based our Review on international and domestic reports, official documents, and published work. We searched PubMed for articles published from Jan 1, 2000, to Nov 3, 2019, to identify relevant manuscripts by use of the terms “sexually transmitted infections”, “Neisseria gonorrhoeae”, “Mycoplasma genitalium”, “surveillance”, “pharmacokinetics and pharmacodynamics”, “azithromycin”, “antimicrobial resistance”, “antimicrobial resistance suppression”, “antimicrobial resistance determinants”, “molecular assays for antimicrobial resistance”, “novel antimicrobials”, and combinations of these terms. We also included cross-references, landmark articles, and references suggested by peer reviewers. We restricted our search to works published in the English language.