Mycoplasma genitalium Macrolide and Fluoroquinolone Resistance Detection and Clinical Implications in a Selected Cohort in New Zealand

ABSTRACT

Mycoplasma genitalium has been associated with infections of the genitourinary tract, and prevalence is secondary to Chlamydia trachomatis. The clinical observation of increasing treatment failure indicating antibiotic resistance, especially in cases of recurrent urethritis, has been confirmed by molecular testing. Mutations in the 23S rRNA gene can cause macrolide resistance, and topoisomerase/gyrase mutations can cause fluoroquinolone resistance. In this study, 115 M. genitalium DNA-positive samples were analyzed. Eighty-nine (77.4%) samples had a 23S rRNA mutation present, and 26 (22.6%) were wild type (no resistance mutation). Fluoroquinolone mutation screening was performed on 86 (74.8%) of the 115 samples, of which 20 (23.3%) samples had a mutation or mutations associated with increased resistance. This study shows the increasing antibiotic resistance in New Zealand and the need for appropriate guidelines to treat at-risk patients.

INTRODUCTION

Mycoplasma genitalium, first isolated in 1980 (1), is a well-established cause of nongonococcal urethritis (NGU) (2). The organism is also associated with pelvic inflammatory disease, endometritis, and cervicitis (2). M. genitalium cannot routinely be cultured in the clinical laboratory and requires molecular techniques such as real-time PCR for detection and DNA sequencing to test for antibiotic resistance. Around 10 years ago, an extended course of azithromycin was reportedly 95% effective at eradicating this organism (3), but macrolide resistance (here represented by azithromycin resistance [AZ^r]) has emerged since (4, 5). Fluoroquinolones provide an alternative treatment option for azithromycin-resistant strains. Recently, an Australian study has reported moxifloxacin treatment failure in 12% of patients, suggesting that monitoring of mutations pretreatment is advisable (6, 7).

The aim of this study is to determine macrolide and fluoroquinolone resistance (FQ^r) rates in clinical specimens referred to Canterbury Health Laboratories (CHL), a tertiary diagnostic laboratory in the South Island of New Zealand, and to evaluate the clinical impact of resistant M. genitalium strains. In 2011, a study performed at CHL first documented the presence of M. genitalium macrolide resistance in New Zealand (8) from samples collected in 2009, but treatment failures were seen prior to this period. The current study will look at a larger patient cohort and confirm the presence of fluoroquinolone resistance in New Zealand. Fluoroquinolone resistance is not routinely tested for but has been suspected based on treatment failures. The results from this study will determine the requirement for routine molecular resistance testing and support the development of clinical guidelines to assist patient management.

RESULTS

In total, 157 M. genitalium DNA-positive samples were collected. Specimen sites included 114 (72.6%) urine samples and 43 (27.4%) swabs (vaginal, cervical, and rectal). Only 115 (73.3%) samples from 104 patients produced sufficient amplicon for macrolide resistance testing as determined by the presence of a 23S rRNA PCR product. The age distribution for this sample group was 15 to 68 yr (median, 27 yr), with 80 (69.5%) patients male. Eighty-nine of these 115 (77.4%) samples had a mutation present, and 26 (22.6%) were wild type (no resistance mutation) (Table 1). Clinical details were obtained for 90 out of 115 (78.3%) of patients, with the most common symptoms discharge, dysuria, and persistent urethritis. Nine patients (10%) were asymptomatic.

TABLE 1.

Frequency of 5 different 23S rRNA gene mutations in 115 samples

Point mutationa	No. (%) of samples
None (wild type)	26 (22.6)
A2058C	2 (1.74)
A2058G	20 (17.4)
A2058T	1 (0.9)
A2059C	2 (1.74)
A2059G	64 (55.7)

^aThe nucleotide numbers shown are based on the Escherichia coli position.

Topoisomerase/gyrase mutation screening in the quinolone resistance-determining region (QRDR) was performed on 86 (74.8%) of the 115 samples, which produced a PCR product suitable for DNA sequencing. Twenty (23.3%) samples had a previously documented mutation or mutations (9–13) associated with increased resistance (Table 2).

TABLE 2.

Number of isolates with mutation/polymorphism(s) in fluoroquinolone resistance-determining genes

Gene	Mutation/polymorphisma	No. of isolates	Reference
gyrA	T447C (silent)	2
	T465G (silent)	1
gyrB	G1392A (R464K)	2	14
	G1398A (silent)	1
	T1485C (silent)	1
parC	A247C (S83R)	5	10
	A259A (D87N)	1	10
	G241T (G81C)	2	11
	G248A (S83N)	1	9
	G248T (S83I)	9	10
	C234T (silent)	36
parE	G1297A (A433T)	1
	C1228T (silent)	1
	G1263A (silent)	2
	G1384A (silent)	1
	G1437A (silent)	1
	T1278C (silent)	3

^aThe nucleotide numbering shown is based on the M. genitalium position.

The patient with a parE mutation, G1297A (A433T) (also macrolide susceptible based on 23S rRNA mutation), had recurrent urethritis, but the mutation is of unknown clinical significance, as no antibiotic or test of cure (TOC) information is available. Two-step mutations were detected in only 2 patients' samples: parC(S83R) gyrB(R464K) and parC(S83I) gyrB(R464K) (Table 3), which can cause high-level resistance to the fluoroquinolones (9, 12, 14). The demographic and clinical details of 4 clinical cases who failed both macrolide and fluoroquinolone treatments are summarized in Table 3. Six patients had single mutations in the parC gene that confer resistance to moxifloxacin but had no 23S rRNA mutations associated with macrolide resistance (M. genitalium AZ^r).

TABLE 3.

Clinical management of 4 patients with M. genitalium FQ^r and AZ^ra

Parameter	Result for patient
	Male, 29 yr old	Male, 52 yr old	Male, 52 yr old	Male, 41 yr old
Symptom(s) on presentation	Urethral discharge	Urethral discharge, penile pain	Dysuria, aching testicles	Urethral discharge
C. trachomatis/Neisseria gonorrhoeae result	ND	ND	ND	ND
Initial treatment	Amoxicillin, 500 mg t.d.s. for 1 wk	Azithromycin, 1 g stat	Azithromycin, 1 g stat	Doxycycline, 100 mg b.i.d. for 7 days
M. genitalium/Trichomonas vaginalis result	M. genitalium detected, T. vaginalis ND	M. genitalium detected, T. vaginalis ND	M. genitalium detected, T. vaginalis ND	M. genitalium detected, T. vaginalis ND
1st treatment for M. genitaliumb	Extended course of azithromycin	Extended course of azithromycin	Extended course of azithromycin	Azithromycin, 1 g stat
TOC	M. genitalium detected	M. genitalium detected	M. genitalium detected	M. genitalium detected
M. genitalium AZr	A2058G	A2059G	A2058G	A2058G
Symptom(s) at 1st follow-up	Continued discharge 2 wk postdiagnosis	Improved, no discharge, but abnormal penile sensation	Asymptomatic	Ongoing symptoms
2nd treatmentc	Moxifloxacin	Moxifloxacin	Moxifloxacin	Moxifloxacin
Symptom(s) at 2nd follow-up	Asymptomatic	Abnormal penile sensation	Dysuria, aching testicles	Urethral discharge
TOC	3 wk posttreatment, M. genitalium detected	1 mo post-1st treatment, M. genitalium detected	4 mo post-1st treatment, M. genitalium detected	1 mo post-1st treatment, M. genitalium detected
M. genitalium FQr	gyrB(R464K) parC(S83R)	gyrB(R464K) parC(S83I)	parC(S83R)	parC(S83I)
TOC	2 wk later, M. genitalium detected	1 mo post-2nd treatment, M. genitalium detected	1 mo post-2nd treatment, M. genitalium detected	1.5 mo post-2nd treatment, M. genitalium detected
3rd treatmentd	Pristinamycin	Pristinamycin	Pristinamycin	Pristinamycin
Final TOC	M. genitalium ND	M. genitalium ND	M. genitalium ND	M. genitalium ND
Duratione	187 days	124 days	295 days	165 days

^aAbbreviations: ND, not detected; t.d.s, 3 times a day; stat, at once; b.i.d., twice a day; TOC, test of cure; AZ^r, macrolide (azithromycin) resistance; FQ^r, fluoroquinolone resistance.

^bThe extended course represents 500 mg azithromycin stat followed by 250 mg daily for 4 days.

^cThe moxifloxacin treatment was 400 mg daily for 7 days.

^dThe pristinamycin treatment was 1 g t.d.s. plus 100 mg doxycycline b.i.d. for 10 days.

^eDuration from first presentation including primary care.

A subset of 33 patients had follow-up treatment with TOC samples and were analyzed for the presence of 23S rRNA mutations and treatment response to extended azithromycin treatment (Table 4).

TABLE 4.

Clinical response to extended azithromycin treatment

Mutation status	No. of patients (n = 33) who:
	Failed TOCa	Passed TOC
23 rRNA mutation	21	2
No mutation	0	10

^aTOC, test of cure.

The 23S rRNA gene mutation has a sensitivity of 100% and specificity of 83.3%, with a 91.3% positive predictive value (PPV) and 100% negative predictive value (NPV) for predicting a macrolide treatment response as measured by a negative TOC.

DISCUSSION

M. genitalium is now accepted as a pathogen with a clinical spectrum similar to Chlamydia trachomatis. Unlike C. trachomatis, it has rapidly developed resistance to first- and second-line antibiotics. Prevalence data indicate a significant burden of disease caused by M. genitalium, with rates second only to C. trachomatis for urogenital infections (15). Other countries have found prevalence rates of 1 to 2.8% among sexually active adolescents and higher rates among various other at-risk groups (16–18).

The data presented here show that 77.4% of specimens referred to CHL from patients attending sexual health services around New Zealand have a mutation that confers resistance to macrolide antibiotics, predominately through mutation A2059G. Similar rates were reported in a recently published Auckland study (19). While M. genitalium infection is common in high-risk populations and M. genitalium resistance has been described as an emerging phenomenon in that population, the true prevalence of M. genitalium disease and resistance mutations in low-risk populations is unknown and should be the target of future research to fully understand the potential impact of these resistance mutations on population health.

This study measures neither prevalence nor incidence as the population is ill defined and the samples tested are not representative of the New Zealand population, nor are they a regional representative sample. However, the reported high resistance rate in this publication demonstrates a worrisome trend in New Zealand.

This is the first report of fluoroquinolone resistance rates in M. genitalium infections in New Zealand, and the documented rate is consistent with those from other studies (10, 11). M. genitalium resistance to fluoroquinolones is increasing overseas and shows the same trend as macrolide resistance. This is a serious challenge for the treating clinician due to lack of alternative treatments. Even though treatments such as pristinamycin can be used successfully in patients with M. genitalium resistance to both macrolides and fluoroquinolones, it is unlikely that pristinamycin will become a sustainable alternative. Furthermore, pristinamycin is not readily available and needs special approval to be used in New Zealand.

This study indicates that fluoroquinolone resistance (FQ^r) can already be present in M. genitalium prior to macrolide treatment and might be selected out by nonspecific use of fluoroquinolones. The 4 patients presented in Table 3 were treated with moxifloxacin without any prior testing for fluoroquinolone mutations. Retrospective testing during this study showed that they contain well-characterized resistance mutations. One patient also had fluoroquinolone resistance mutations in his previous samples; the other 3 patients' samples did not produce suitable PCR product for sequencing. As patients can have resistance to azithromycin, moxifloxacin, or both antibiotics, it is therefore important to include screening for M. genitalium FQ^r in patients with or without M. genitalium AZ^r. An ideal screening assay would incorporate both 23S rRNA and parC genes to guide appropriate antimicrobial treatment.

The current lack of routine molecular resistance testing for M. genitalium disease increases the time from diagnosis of M. genitalium infection to effective treatment for the patients (as outlined in Table 3). This is not only an inconvenience issue as patients will have to be recalled and examined and tested several times: it also contributes to the time the patient continues to be unwell, the number of potential contacts who may contract the infection, and the cost due to the extended time the patients are actively managed in the health care system. This study demonstrated an average duration of 192 days (almost 6.5 months) from first presentation of the patient to the completion of successful treatment. There are other possible reasons for this, outside the lack of rapid methods to detect resistance, including, noncompliance of patients, reinfections, and nonattendance at scheduled follow-up visits. Many of these possible barriers to successful patient management lie outside the control of the treating clinician and the diagnostic laboratory. However, a rapid diagnostic assay can assist the clinician to implement appropriate treatments at the earliest opportunity, which in turn is likely to reduce the number of patients that will be lost to follow-up and remain inappropriately treated for prolonged periods. Early appropriate treatment will also reduce the rate of transmission of M. genitalium strains with resistance to first- and second-line antibiotics.

The issue of potential reinfection posttreatment in patients with no detectable 23S rRNA mutation associated with macrolide resistance would present with clinical treatment failure. It can be difficult to establish the exact cause of the treatment failure due to lack of detailed clinical information. These findings are similar to another study, in which 87% of patients failed azithromycin treatment in association with 23S rRNA point mutations (7).

Patients presenting with nongonococcal urethritis can be treated initially with 100 mg doxycycline twice a day (b.i.d.) for 7 days as the first-line treatment. This will be more than 95% effective in men who are chlamydia positive, and there is no evidence that this would induce macrolide resistance in those who are M. genitalium positive (20). Although azithromycin at 1.0 g can also be used, and its use would favor compliance, there is the risk of developing macrolide resistance. Ideally, M. genitalium testing with resistance genotyping would be done at the time of presentation. If the patient is positive for M. genitalium, treatment should be based on the genotypic resistance profile. In cases where no genotypic resistance profile is available, a regimen of 500 mg azithromycin orally on day 1 and then 250 mg orally on days 2 to 5 is recommended. A test of cure should be collected no earlier than 3 weeks after the start of treatment, and if test of cure fails, then 400 mg moxifloxacin orally once a day for 7 days should be given. If TOC is still not achieved, specialist advice is needed.

In cases where a genotypic resistance profile is available and macrolide and moxifloxacin resistance-mediating mutations are absent, 500 mg azithromycin orally on day 1 and then 250 mg orally on days 2 to 5 is recommended. Where macrolide resistance-mediating mutations are present and moxifloxacin resistance-mediating mutations are absent or unknown, the recommended treatment in our setting is 400 mg moxifloxacin orally once a day for 7 days.

In other cases, a specialist's advice should be sought. A test of cure should be collected no earlier than 3 weeks after the start of treatment and done in all cases, as recommended in the review for the 2016 European guidelines on Mycoplasma genitalium infection (2).

With the availability of newer molecular assays that can detect the infecting organism concurrently with mutations to commonly prescribed treatment regimens, it has become possible to provide the treating clinician with rapid information on the optimal management of their individual patient (21, 22). A new diagnostic paradigm will see multiplex PCR assays screening for syndromes rather than individual pathogens and will provide resistance or susceptibility data in one single step, enabling the clinician to treat optimally at the first or second encounter with the patient. The patient is likely to respond with increased compliance experiencing treatment success, and diagnostic assays will support antimicrobial stewardship by reducing the use of ineffective treatment and by collecting real-time resistance data that can monitor resistance trends.

Conclusion.

The data show very high levels of macrolide resistance and lower, but significant, fluoroquinolone resistance in M. genitalium strains, which is consistent with other studies. Diagnostic assays for M. genitalium, including antibiotic resistance screening, would provide a significant benefit to patients with M. genitalium receiving earlier and correctly targeted treatments and lower the burden on the health care system by avoiding recurrent presentations and reducing the use of inappropriate therapy contributing to rising antimicrobial resistance worldwide.

MATERIALS AND METHODS

Clinical samples suspected of M. genitalium infection were collected from 2010 to 2016 in an Abbott multi-Collect specimen collection kit and sent to CHL; samples were submitted from the local Sexual Health Centre and some referral sites from the South and North Islands of New Zealand. The samples were processed on the M2000 real-time system (Abbott Molecular, Australia). Extracted nucleic acid was stored at −80°C until testing. M. genitalium DNA was screened by real-time PCR targeting the M. genitalium Pa gene (MgPa) gene as described by Jensen et al. (23). Mutations associated with macrolide resistance were detected using primers targeting region V of the 23S rRNA gene (4). Fluoroquinolone resistance mutations in the gyrA, gyrB, parC, and parE genes were screened using primers as described by Shimada et al. (9, 14). DNA sequencing was performed using BigDye Terminator v3.1 (Applied Biosystems, USA) on the ABI Prism 3130xl genetic analyzer (Applied Biosystems, USA). Data were analyzed using the McNemar test for sensitivity and specificity. Ethical approval for this project was granted by University of Otago, Christchurch, Human Ethics Committee (Health).