Background Mycoplasma genitalium (Mg) has rapidly increased its resistance to azithromycin, which has been first-line therapy for non-chlamydial-non-gonococcal urethritis (NCNGU), a proportion of which is due to Mg, and for treating Mg specifically. New commercial nucleic acid amplification tests (NAATs) are likely to greatly increase diagnosis and treatment of Mg, potentially promoting resistance. We previously developed the first transmission-dynamic model of Mg, which we now use to examine alternative approaches to NAAT testing and treatment.

Methods Our model synthesises evidence from surveillance data, and epidemiological and behavioural studies, and accounts for parameter uncertainty, including the fitness-costs and benefits of drug resistance. The model incorporates resistance due to de novo mutation and transmission. We examined scenarios regarding (i)targeting of NAAT (only testing symptomatic patients and partners vs testing all patients); (ii)using NAATs detecting resistance vs only detecting Mg; (iii)choice of first-line therapy (including continuing using azithromycin except where resistance has been detected vs alternative first-line regimen).

Results If azithromycin continues to be first-line therapy then resistance (and incidence of sequelae) will continue to rise, exacerbated by increased NAATs-based diagnoses. If asymptomatic screening occurs then resistance will increase 3.9(95%CI:2.1–5.7) times as rapidly as if only symptomatic patients and partners are tested. Pre-treatment resistance testing mitigates but does not prevent increases in resistance, due to resistance arising from frequent de novo mutation. Long-term outcomes of alternative regimens are highly uncertain.

Conclusion This work supports recommendations not to screen for Mg until a better treatment regimen has been determined. NAATs should include resistance-testing but this is not a panacea. Improved understanding of Mg’s natural history is urgently required, along with better surveillance of testing, diagnosis, and treatment, to monitor clinical adherence to guidelines, quantify drug-resistance fitness costs and benefits, and reduce uncertainty in decision-making. Internet-based testing and prescribing is a grave concern and needs to be controlled.

Disclosure No significant relationships.