Clinical Infectious Diseases

MAJOR ARTICLE

Azithromycin 1.5g Over 5 Days Compared to 1g Single

Dose in Urethral Mycoplasma genitalium: Impact on
Treatment Outcome and Resistance
Tim R. H. Read,1,2 Christopher K. Fairley,1,2, Sepehr N. Tabrizi,3,4,5 Melanie Bissessor,1,2 Lenka Vodstrcil,2 Eric P. F. Chow,1,2
Mieken Grant,2 Jennifer Danielewski,3,5 Suzanne M. Garland,3,4,5 Jane S. Hocking,6 Marcus Y. Chen,1,2 and Catriona S. Bradshaw 1,2
1
Central Clinical School, Monash University, and 2Melbourne Sexual Health Centre, Alfred Health, Melbourne, Australia; 3Murdoch Childrens Research Institute, Parkville, Victoria, Australia;
4
Department of Obstetrics and Gynaecology, University of Melbourne; 5Department of Microbiology and Infectious Diseases, the Royal Women’s Hospital, and 6Melbourne School of Population and
Global Health, University of Melbourne, Victoria, Australia

Background.  We evaluated the impact of extended azithromycin (1.5g over 5 days) on selection of macrolide resistance and
microbiological cure in men with Mycoplasma genitalium urethritis during 2013–2015 and compared this to cases treated with
azithromycin 1g in 2012–2013.
Methods.  Microbiological cure was determined for men with M. genitalium urethritis treated with azithromycin 1.5g using
quantitative polymerase chain reaction specific for M. genitalium DNA on samples 14–100 days post-treatment. Pre- and post-treat-
ment macrolide resistance mutations were detected by sequencing the 23 S gene.
Results.  There was no difference in proportions with microbiological cure between azithromycin 1.5g and 1g: 62/106 (58%;
95% confidence interval [CI], 49%, 68%) and 56/107 (52%; 95%CI 42–62%), P = .34, respectively. Also, there was no difference in
the proportion of wild-type 23 S rRNA (presumed macrolide sensitive) infections cured after 1.5g and azithromycin 1g: 28/34 (82%;
95%CI 65–92%) and 49/60 (82%; 95%CI 70–90%), P=1.0, respectively. There was no difference between 1.5g and 1g in the propor-
tions of wild-type infections with post-treatment resistance mutations: 4/34 (12%; 95%CI 3–27%) and 11/60 (18%; 95%CI 10–30%),
respectively, P = .40. Pre-treatment resistance was present in 51/98 (52%; 95%CI 42–62%) cases in 2013–2015 compared to 47/107
(44%; 95%CI 34–54%) in 2012–2013, P = .25.
Conclusions.  Extended azithromycin 1.5g was no more effective than a single 1g dose at achieving cure of M. genitalium ure-
thritis and importantly did not reduce the selection of macrolide resistance. Nonmacrolide and new approaches for the treatment
of M. genitalium urethritis are required.
Keywords.  Mycoplasma genitalium; nongonococcal urethritis; azithromycin; macrolide resistance.

Mycoplasma genitalium causes approximately 15%–25% of In Melbourne, Australia, in 2012–2013, azithromycin 1g

nongonococcal urethritis (NGU) cases in men [1]. A  sin-
gle 1g dose of azithromycin has been a common treatment
for NGU. However, since 2006, multiple countries have
reported increasing failure rates of azithromycin 1g for
infections due to M.  genitalium [2–4]. Lau et  al recently
reviewed 21 studies of the efficacy of azithromycin 1g for
M.  genitalium in 1490 people and reported that the aver-
age cure rate has fallen from 85% (95% confidence interval
[CI], 82%–88%) in studies performed from 1999 to 2008 to
67% (95% CI, 57%–77%) for studies performed from 2009
to 2013 [5].
Received 6 July 2016; editorial decision 10 October 2016; accepted 20 October 2016;
published online October 24, 2016.
Correspondence: T. R. H. Read, 580 Swanston St, Carlton, Victoria 3053 Australia (tread@
mshc.org.au).
Clinical Infectious Diseases®  2017;64(3):250–6
© The Author 2016. Published by Oxford University Press for the Infectious Diseases Soci-
ety of America. All rights reserved. For permissions, e-mail journals.permissions@oup.com.
DOI: 10.1093/cid/ciw719
cured 61% of 155 genital M. genitalium infections (69% were
NGU). Macrolide resistance mutations (MRMs) were found in
36% of pretreatment samples and were strongly associated with
azithromycin failure [6]. Importantly, post-treatment MRMs
were detected in 11% of genital M. genitalium cases infected
with initially susceptible strains, suggesting the selection of
resistant strains during treatment with azithromycin 1g. The
selection of resistance in the setting of widespread use of sin-
gle-dose azithromycin 1g for the slow-growing M. genitalium
and its associated syndromes may explain why rising rates of
circulating MRMs are now greater than 30% in many countries
[7]. Recent European guidelines recommend a 5-day course
of azithromycin 1.5g, dosed 500 mg initially, then 250 mg for
another 4 days for M. genitalium [8]. Three nonrandomized
Swedish studies have reported high proportions cured with
this extended 1.5g regimen when compared to azithromycin
1g or when stratified by pretreatment macrolide resistance
[9–11]. When data from the 2 studies reporting pretreatment
macrolide resistance are pooled, extended azithromycin failed

250 • CID 2017:64 (1 February)  •  Read et al

Downloaded from https://academic.oup.com/cid/article-abstract/64/3/250/2736225

by guest
on 26 March 2018
 and MRMs were subsequently detected in 3/71 (4.2%; 95% CI,
0.9%, 11.9%) [9, 11].
Based on these preliminary data, in October 2013 the
Melbourne Sexual Health Centre (MSHC) began treating men
with NGU with a 5-day course of azithromycin 1.5g (500 mg
initially followed by 250 mg daily for 4 days) to determine if
this improved cure rates by reducing the rate of emergence
of MRMs during treatment. Here, we report treatment out-
comes and the proportions of cases with MRMs before and
after treatment with azithromycin 1.5g and compare this with
data from the preceding period in 2012–2013 in which all
men with M.  genitalium–associated NGU received azithro-
mycin 1g.
METHODS
MSHC is the only public clinic that treats sexually transmit-
ted infections (STIs) in Melbourne. The center diagnoses
approximately 1500 cases of NGU in men each year. Clients
provide basic sexual behavioral data at arrival. These data,
along with clinical notes and laboratory results, are stored in
an electronic clinic database. Using this database, we identi-
fied all cases of M. genitalium in men with NGU diagnosed
at MSHC from 1 October 2013 to 30 June 2015. Cases from
1 July 2012 to 30 June 2013 and treated with azithromycin 1g
were used as historical controls and identified in the same way.
Members of the research team extracted the following data
from electronic medical records: date of presentation, pres-
ence of urethral symptoms, date of M. genitalium test, type of
antibiotic treatment, prior antibiotic prescribed for this epi-
sode, result of test of cure, genital symptoms after treatment,
risk of reinfection (sex after treatment with untreated part-
ners), and documentation of poor adherence. Only men with
NGU (defined as up to 1 month of urethral symptoms such
as discharge, discomfort, or dysuria) treated with extended
azithromycin 1.5g (or 1g for controls) and with urethral/
urine samples positive for M. genitalium were included in the
analysis. The Alfred Hospital Ethics Committee approved the
study (no 304/15).
Laboratory Methods
Extracted DNA from pre- and post-treatment specimens with
M. genitalium detected was stored at –30°C. Samples from the
study period were processed as described previously; samples
from April 2014 to September 2014 were not available for anal-
ysis [6]. A  M.  genitalium–specific polymerase chain reaction
(PCR) targeting the 16S ribosomal RNA gene was used to detect
M. genitalium DNA [12]. Mycoplasma genitalium load was sub-
sequently determined on all positive samples as described pre-
viously [6]. Sanger sequencing was used to determine MRMs
at positions 2058 and 2059 (Escherichia coli numbering) in the
23S rRNA gene of M. genitalium on all pretreatment and persis-
tently positive post-treatment samples [4].
Downloaded from https://academic.oup.com/cid/article-abstract/64/3/250/2736225
by guest
on 26 March 2018
Statistical Methods
The analysis of microbiological cure was restricted to men
who were retested (test of cure) for M. genitalium 14–100 days
after the start of treatment. Positive tests of cure were defined
as treatment failures and negative tests as cures. Proportions
cured and proportions with post-treatment resistance were
calculated, and 95% CIs were calculated using the binomial
exact distribution. P values for proportions were calculated
using Fisher exact test or Wilcoxon rank-sum test for medi-
ans. Bacterial load was log transformed, and significance of
comparisons was assessed using t test. Odds ratios (ORs) and
95% CIs for predictors of treatment failure were calculated
from a logistic regression model using the group analyzed for
microbiological cure, adjusted for potential confounders in the
multivariate model.
To determine if the proportion of men with urethritis in
whom resistance was selected during treatment with extended
azithromycin 1.5g (wild-type pre-treatment and MRMs
detected post-treatment) differed from the proportion of men
given a single dose of 1g, we performed a historical compari-
son using controls from a series of patients with M. genitalium
treated at MSHC with azithromycin 1g in 2012 and 2013 [6]. As
the current analysis of azithromycin 1.5g is restricted to men
with NGU, we excluded women and asymptomatic men from
the historical dataset to ensure comparable populations. Using
the 2-sample Z test for proportions, we compared proportions
with pre-treatment and selected resistance and treatment out-
comes between the 2 groups. The 106 cases in the 2013–2015
group gave 82% power to detect a 14% increase in the propor-
tion cured (P  =  .05). Statistical calculations were performed
using Stata, version 13.
RESULTS
From 1 October 2013 to 30 June 2015, 215 men with NGU had
positive tests for M. genitalium. Of these, 169 were treated with
azithromycin 1.5g and 98/169 had sufficient stored sample for
23S MRM genotyping (Figure 1). Of the 169 men, 143 (85%)
returned to the clinic and 106 (63%) had a test of cure 14–100
days after starting treatment (median, 36 days; intraquartile
range, 29–52). Sixty-three had both genotyping of pre- and
post-treatment M. genitalium specimens and a test of cure
14–100 days after treatment.
Characteristics of Participants
Of the 106 men with M.  genitalium urethritis who received
azithromycin 1.5g and returned for test of cure, 41 (38%) were
men having sex with men (MSM) (Table 1). MSM were older,
reported more partners, and were less likely than heterosexual
men to report sexual partners outside Australia (19% compared
to 55%; P  =  .002). MSM were twice as likely to have MRMs
detected in pre-treatment samples (76% compared to 39%;
P = .005).
Azithromycin 1.5g for M. genitalium • CID 2017:64 (1 February) • 251

Figure 1.  Selection for analyses of patients with urethral Mycoplasma genita-
lium from the group treated with azithromycin 1.5g.
Treatment Outcomes Following Extended Azithromycin 1.5g Compared to 1g
In the primary analysis of 106 men with tests of cure at 14–100
days, 62 (58%; 95% CI, 49%, 68%) were microbiologically cured.
For MSM, the proportion cured was 14/41 (34%; 95%CI, 20%,
50%), lower than in heterosexual men for whom the proportion
cured was 48/65 (74%; 95% CI, 61%, 84%; P < .001; Table 2).
Recurrent or persistent urethral symptoms (discharge, burn-
ing, dysuria) at the time of test of cure were more common in
patients who did not achieve microbiological cure (34/44; 77%;
95% CI, 62%, 89%) compared to those who did (10/63; 16%;
95% CI, 8%, 28%; P < .001).
To determine if proportions cured differed between extended
azithromycin 1.5g and single-dose azithromycin 1g, we com-
pared the current dataset (2013–2015) with a dataset of males
with M. genitalium urethritis who received 1g (2012–2013),
97% of whom returned for assessment before our treatment
protocol changed [6]. There was no significant difference in
proportions cured by azithromycin 1.5g (58%; 95% CI, 49%,
68%) compared to azithromycin 1g (52%; 95% CI, 42%, 62%;
P = .34; Table 3). Pretreatment macrolide resistance was detected
in 52% (95% CI, 42%, 62%) of males with M. genitalium urethri-
tis in 2013–2015 and in 44% (95% CI, 34%, 54%) of males with
M. genitalium urethritis in 2012–2013 (P = .25). While the prev-
alence of pretreatment resistance was the same in heterosexual
males in the 2 time periods (39% and 40%, respectively; P = .9),
it was higher in MSM in 2013–2015 (76%) compared to MSM in
2012–2013 (53%; P = .05). Proportions cured were the same for
both regimens in those with pretreatment resistance (14% and
15%; P = .9) and in those without (82% for each regimen; P = 1.0).
Macrolide Resistance
MRMs were detected in 51 of 98 genotyped pre-treatment sam-
ples (52%; 95% CI, 42%, 62%). Treatment outcomes are shown
in Table 4 by pre-treatment genotype. In the 63 cases who were
genotyped and analyzed for microbiological cure (Figure  1),
pre-treatment MRMs were detected in 81% of cases who experi-
enced microbiological failure and in 12% of cases who appeared
microbiologically cured (P < .001).
Six patients with wild-type infections had persistent infec-
tion after treatment; only 4 of these post-treatment samples
were available for genotyping. MRMs were detected in all
4 at test of cure, meaning the proportion of wild-type infec-
tions with post-treatment resistance was 4/34 (12%; 95%
CI, 3%, 27%). Sexual histories of the 6 men with wild-type
treatment failures indicate that reinfection was more likely
in one man who resumed sex with a partner treated with the
same regimen; his post-treatment sample was not sequenced
(Supplementary Table  1). Of wild-type cases treated with 1g

Table 1.  Characteristics of 106 Men With Mycoplasma genitalium Urethritis Who Returned for Test of Cure

Heterosexual Men, Men Who Have Sex With Men,a

Characteristic N = 65 N = 41 P Value Total N = 106b

Age, y (median [IQR]) 29 (24–32) 32 (27–39) .02c 31 (25–35)

N partners in previous 3 months (median [IQR]) 2 (1–4) 3 (2–6) .01c 3 (1–4)
Reported sex outside Australia, n/N (%)d 29/53 (55) 5/27 (19) .002e 34/80 (43))
Prior azithromycin for this nongonococcal urethritis, 4/65 (6) 6/41 (15) .15e 10/106 (9)
n (%)b
Specimen
First-void urine, n (%) 60 (92) 41 (100) 101 (95)
Urethral swab, n (%) 5 (8) 0 5 (5)
Macrolide resistance pre-treatment, n/N (%)f 25/64 (39) 26/34 (76) <.001e 51/98 (52)

Abbreviation: IQR, intraquartile range; N, total number; n, number with characteristic.

a
Includes bisexual men.
b
Proportion of medical records documenting outcome of azithromycin treatment.
c
Rank-sum test.
d
Denominator varies because not all patients provided behavioral data.
e
Fisher exact test.
f
Denominator varies because only 98 stored pre-treatment samples were tested for macrolide resistance mutations (see Figure 1).

252 • CID 2017:64 (1 February)  •  Read et al

Downloaded from https://academic.oup.com/cid/article-abstract/64/3/250/2736225

by guest
on 26 March 2018
 Table 2.  Proportion of Men With Urethral Mycoplasma genitalium Cured test. All 31 post-treatment samples available for genotyping
After Extended Azithromycin 1.5g Overall and Stratified by Pre-Treatment
from azithromycin failures contained MRMs.
Resistance Genotype and Other Risk Groups

Bacterial Load
Proportion Cured,a % (95% Confidence
Patient Group n/N Interval) Bacterial load was examined in patients stratified by pre-treat-
All patients in microbiological 62/106 58 (49, 68) ment MRMs and treatment outcomes. The 6 wild-type infec-
cure analysis tions that were not cured had significantly higher pre-treatment
By pre-treatment resistance mutationb
bacterial loads (mean log10 4.6) compared to the 28 wild-type
  Not detected 28/34 82 (65, 92)
infections (mean log10 3.2) that were cured (P = .02; Figure 2),
 Detected 4/29 14 (8, 35)
Sexual preference indicating organism load may play a role in selected resist-
 Heterosexual 49/66 74 (62, 83) ance. The 4 macrolide-resistant infections that were cured had
  Men who have sex with 14/41 34 (21, 50) lower pre-treatment bacterial loads (mean log10 1.6) than the
menc
remaining 25 macrolide-resistant infections (mean log10 3.1)
Previous azithromycin treatment for this episode
 No 60/96 63 (52, 72)
that persisted (P = .04). The mean pre-treatment bacterial load
 Yes 2/10 20 (3, 56) was lower in resistant (log10 2.9) than in wild-type (log10 3.4)
Sex outside Australiab infections but not significantly so (P = .06; Figure 3a). In cases
 No 28/46 61 (45, 75) who experienced treatment failure, the bacterial load was lower
 Yes 21/34 62 (44, 78) in post-treatment samples (mean log10 2.2) than in pre-treat-
Abbreviations: N, total number; n, number with characteristic.
a
ment samples (log10 3.3; P < .01; Figure 3b). Post-treatment load
Mycoplasma genitalium not detected by polymerase chain reaction 14–100  days after
treatment. was measured in 4 of the 6 wild-type treatment failures, and
the mean was 2.6 log10 lower than the mean pre-treatment load
b
Smaller denominator due to limited number of samples available for sequencing or
limited data.
c
Includes bisexual men. (P = .01). In each of the 4 cases, it was at least 1.5 log10 lower
than the corresponding pre-treatment load.

azithromycin in the historical control group, post-treatment Predictors of Treatment Failure

MRMs were detected in 18% (95% CI, 10%, 30%), which
was not significantly (P  =  .4) greater than the 12% (95% CI,
3%, 27%) of cases with post-treatment MRMs after extended
azithromycin 1.5g. Of 29 cases with pre-treatment MRMs, 4
appeared to be cured (14%; 95% CI, 8%, 35%; Table 2). These
men were asymptomatic when they returned for test of cure
(range, 32–95  days post-treatment), and none returned with
further symptoms that would suggest an initial false-negative
In univariate analysis, failure of extended azithromycin 1.5g
was associated with the following: pre-treatment resistance,
reporting male sexual partners, number of sexual partners,
and prior use of azithromycin for this NGU episode (Table 5).
Treatment failure was not associated with the following: sex
outside Australia, bacterial load, and risk of reinfection. In mul-
tivariate analysis, factors that remained significantly associated
with treatment failure were pre-treatment resistance (adjusted

Table 3.  Pre-treatment Resistance, Selected Resistance, and Outcomes After a Single Dose of Azithromycin 1g in 2012–2013 or Extended Azithromycin
1.5g Over 5 Days in 2013–2015 for Men With Mycoplasma genitalium Urethritis

1.5g Over 5 Days 1g Single Dosea

Patient Group n/N % (95% CI) n/N % (95% CI) P Valueb

Proportion cured 62/106 58 (49, 68) 56/107 52 (42, 62) .34

Pretreatment resistance
 Overall 51/98 52 (42, 62) 47/107 44 (34, 54) .25
 Heterosexual 25/64 39 (27, 52) 31/77 40 (29, 52) .90
  Men who have sex with menc 26/34 76 (59, 89) 16/30 53 (34, 72) .05
Proportion of each pretreatment genotype cured
  Wild type 28/34 82 (65, 92) 49/60 82 (70, 90 1.0
 Resistant 4/29 14 (8, 35) 7/47 15 (6, 28) .90
Selected resistanced 4/34 12 (3, 27)e 11/60 18 (10, 30) .40

Abbreviation: CI, confidence interval, N = total number, n = number with characteristic.

a
Data differ from the series published by Bissessor et al [6] because this analysis was restricted to males with nongonococcal urethritis.
b
Study populations were compared using the 2-sample Z test for proportions.
c
Includes bisexual men.
d
Proportion of initially wild-type infections with positive test of cure and macrolide mutation detected.
e
Treatment failed in 6/34 wild-type infections, but post-treatment samples were only available for genotyping in 4 of these.

Azithromycin 1.5g for M. genitalium • CID 2017:64 (1 February) • 253

Downloaded from https://academic.oup.com/cid/article-abstract/64/3/250/2736225

by guest
on 26 March 2018
 Table 4.  Treatment Outcomes by Pretreatment Genotype M. genitalium infections. Management algorithms for NGU
that account for resistant M. genitalium must now be developed.
Microbiological Cure
Analysis
This is the largest study to report outcomes of treating M.
genitalium with 1.5g extended azithromycin and the only study
Pretreatment Genotype Cureda Failedb Excludedc Total
to evaluate this regimen restricted to a single anatomical site
Wild type, n (%) 28 (88)   6 (19) 13 (37) 47 and incorporating bacterial load and resistance mutations.
Resistant,d n (%)   4 (12) 25 (81) 22 (63) 51
However, it is a retrospective study with incomplete resistance
Total 32 31 35 98
a
data and incomplete follow-up; 15% of otherwise eligible men
Mycoplasma genitalium not detected by polymerase chain reaction 14–100  days after
treatment. did not return for a test of cure. So, it is possible that men who
b
Mycoplasma genitalium detected by polymerase chain reaction 14–100 days after were cured were less likely to return. The pre-treatment gen-
treatment.
c
Excluded because no test of cure between 14 and 100 days after treatment.
otypes in >60% of the men who did not return show MRMs,
d
Macrolide resistance mutation detected by sequencing. implying a likely lower cure rate in this group and that we have
not overestimated the failure rate. The OR for treatment failure
with MRMs was 24, so the failure rate in this analysis is greatly
OR [aOR], 24.2; 95% CI, 5.7, 102.3; P < .001) and being MSM influenced by the high proportion with resistance, meaning
(aOR, 4.9; 95% CI, 1.1, 21.1; P = .04). this finding will not be generalizable to populations in which
resistance is less common. We cannot exclude the possibility
DISCUSSION that some apparently wild-type infections included undetected
Increasing the dose and duration of azithromycin to 1.5g over minority populations with MRMs. Nor can we be sure that some
5 days for the treatment of NGU did not increase the proportion of the treatment failures were not reinfections. The sexual histo-
of M. genitalium infections that were cured and did not reduce ries of those with wild-type infections (Supplementary Table 1)
the rate at which MRMs were detected after treatment. Half whose treatment failed suggest reinfection was unlikely for the
of all M. genitalium infections in Melbourne are now resistant majority. A number of other observations in these cases argue
to macrolides; this is significantly more common in MSM in against reinfection and are more in favor of selected resist-
whom three quarters of infections are resistant. Infections that ance. First, the genotype has changed from wild type to include
persisted despite treatment tended to have a higher pre-treat- MRMs in every case and there are no wild-type infections as we
ment bacterial load, and MRMs were detected in post-treatment might expect with some reinfections. Second, the bacterial load
samples, even if they were initially wild-type infections. These is at least 1.5 log lower in each post-treatment sample compared
findings raise the question whether continued widespread use to pre-treatment samples. Finally, in univariate analysis, there is
of azithromycin for NGU in either dosage regimen will result no association between treatment failure and the reported risk
in a gradual increase in the proportion of macrolide-resistant of reinfection.
Extended azithromycin 1.5g did not significantly increase
the proportion of M.  genitalium urethritis cases cured (58%)
compared to historical controls treated with a single dose of 1g
at the same center 2 years previously (52%). Proportions cured
using each regimen were identical for macrolide-susceptible
and macrolide-resistant infections. This contrasts with findings
from 7 studies of extended azithromycin 1.5g for M. genitalium
in which 332 of 386 individuals (86%) had negative tests of cure
[9–11, 13–16]. The higher proportion of treatment failures in
our study is likely due to the higher proportion with pre-treat-
ment macrolide resistance (52%) compared to most other stud-
ies and is consistent with a global trend of increasing resistance
[7, 17]. The proportion with post-treatment resistance (12%;
95% CI, 3%, 27%) is also somewhat higher than the 4.2% (95%
CI, 0.9%, 11.9%) in the 2 Swedish studies noted earlier [9, 11],
nevertheless this difference is not significant (P = .15).
It was hoped that the extended azithromycin 1.5g regimen
would reduce the selection of MRMs in those wild-type M. gen-
Figure 2.  Mycoplasma genitalium urethritis pre-treatment log10 bacterial load in italium infections in which treatment failed; this regimen is
urine, by pre-treatment genotype and outcome of treatment with azithromycin 1.5 g
now being recommended for NGU in European treatment
over 5 days. P values compare mean loads in cured and failed cases within each
genotype. guidelines [8]. However, in Melbourne, resistance appeared to

254 • CID 2017:64 (1 February)  •  Read et al

Downloaded from https://academic.oup.com/cid/article-abstract/64/3/250/2736225

by guest
on 26 March 2018
Read et al

Figure 3.  Mycoplasma genitalium log10 bacterial load in men with nongonococcal urethritis. A, Pre-treatment log10 bacterial loads in wild-type and macrolide-resistant
infections. B, Log10 bacterial loads before and after extended azithromycin 1.5g in treatment failures.

be selected in similar proportions of wild-type infections (18%
after 1g and 12% after 1.5g), regardless of the use of a single dose
or 5-day regimen.
The wild-type infections that persisted and became mac-
rolide resistant tended to have higher pre-treatment bacter-
ial loads, and the resistant infections that were cured tended
to have lower loads. However, bacterial load was not a sig-
nificant predictor of treatment outcome, possibly due to the
small number of cases where outcome was not already pre-
dicted by pre-treatment genotype. Russian men with ure-
thritis due to M.  genitalium (all wild type) were monitored
for the disappearance of M. genitalium DNA after receiving
the macrolide josamycin [18]. Mycoplasma genitalium DNA
persisted for the longest period of time in the 13% of men
with the highest pre-treatment bacterial load, leading to
treatment failure and selection of MRMs in half of this group

Table 5.  Predictors of Failure of Extended Azithromycin 1.5g to Treat Urethral Mycoplasma genitalium Infection in Patients With Test of Cure 14–100 Days
After Treatment and Pre-Treatment Genotype

Patient Group Unadjusted OR (95% CI) P Value Adjusted ORa (95% CI) P Value

Pre-treatment genotype
Wild type Reference Reference
Resistant 29.2 (7.4, 115.4) <.001 24.2 (5.7, 102.3) <.001
Bacterial loadb 1.1 (.9, 1.3)  .3
Sexual preference
 Heterosexual Reference Reference
  Men who have sex with menc 5.6 (13.4, 13.0) <.001 4.9 (1.1, 21.1)  .04
Previous azithromycin recorded
 No Reference Reference
 Yes 4.4 (1.1, 17.8)  .04 .8 (.1, 6.6)  .9
Sex outside Australia
 No Reference
 Yes 1.0 (.4, 2.5)  1.0
Number of partners in past 3 monthsd
  or per additional partner 1.2 (1.0, 1.4)  .03 1.1 (.8, 1.5)  .4
Risk of reinfectione
 No Reference
 Yes 1.6 (.4, 6.9)  .5

Abbreviations: CI, confidence interval; OR, odds ratio. 

a
OR for treatment failure, adjusted for sexual preference and pre-treatment resistance mutations.
b
OR per log10 increase in bacterial load in pre-treatment sample.
c
Includes bisexual men.
d
Number of sexual partners in the 3 months prior to presentation with nongonococcal urethritis.
e
Sex between treatment and test of cure with new or preexisting partners.

Azithromycin 1.5g for M. genitalium • CID 2017:64 (1 February) • 255

Downloaded from https://academic.oup.com/cid/article-abstract/64/3/250/2736225

by guest
on 26 March 2018
 but not in those with lower bacterial loads. We also observed
a reduction in bacterial load after failed treatment. In other
studies where extended azithromycin 1.5g performed well,
some patients were previously treated with doxycycline,
which usually fails to eradicate M.  genitalium [9, 10]. It is
plausible that prior treatment with doxycycline reduced bac-
terial load and increased the likelihood of success with a sec-
ond antibiotic.
Most MSM with M.  genitalium urethritis in Melbourne
already have a resistant infection, and there seems to be lit-
tle value in treating these men with either dosage regimen of
azithromycin without a negative test for MRMs. It is unclear
why resistance is so high in this group. However, the high
prevalence of other STIs for which azithromycin is used alone
or in combination suggests that M.  genitalium circulating in
this population is more likely to be exposed to this selective
pressure [19]. MSM also appear to have a higher prevalence of
antibiotic-resistant Neisseria gonorrhoeae, perhaps for similar
reasons [20, 21].
The implications of the findings in this study are, first, that
extended azithromycin 1.5g has no advantage over the single
1g dose and is unlikely to stem the rise of macrolide resistance.
Second, a high bacterial load may favor the selection of MRMs
and some resistant infections with low bacterial load may still be
cured. After treatment failure, persistent infections have lower
bacterial loads, so it is possible that an antibiotic sequence or
combination could exploit this, whereby one antibiotic lowers
load and increases the efficacy of the other antibiotic. Alternative
classes and combinations of antibiotics are required for M. geni-
talium, and the development of rapid point of care tests and tests
that identify MRMs will improve our selection and stewardship
of available antibiotics. In the meantime, clinical services need
to decide how to manage NGU and pelvic inflammatory disease
where azithromycin remains effective for Chlamydia trachoma-
tis, but where resistant M. genitalium is increasingly likely to be
present.
Supplementary Data
Supplementary materials are available at Clinical Infectious Diseases
online. Consisting of data provided by the author to benefit the reader,
the posted materials are not copyedited and are the sole responsibility
of the author, so questions or comments should be addressed to the
author.
Notes
Acknowledgments.  We are grateful to Jimmy Twin, Jenny Su, and
Karen Worthington for their invaluable assistance with this project.
Financial support.  This work was supported by the Australian National
Health and Medical Research Council (Early Career Fellowships 1091536 to
T. R. H. R. and 1091226 to E.P. F. C., and program grant 568971).
256 • CID 2017:64 (1 February)  •  Read et al
Downloaded from https://academic.oup.com/cid/article-abstract/64/3/250/2736225
by guest
on 26 March 2018
Potential conflicts of interest.  T. R. H. R., M. G., and C. S. B. declare their
institution received research funding from SpeeDx Pty Ltd, Sydney, Australia.
All other authors report no potential conflicts. The authors have submitted the
ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the
editors consider relevant to the content of the manuscript have been disclosed.
References
1. Taylor-Robinson D, Jensen JS. Mycoplasma genitalium: from Chrysalis to multi-
colored butterfly. Clin Microbiol Rev 2011; 24(3):498–514.
2. Bradshaw CS, Chen MY, Fairley CK. Persistence of Mycoplasma genitalium fol-
lowing azithromycin therapy. PLoS One 2008; 3(11):e3618.
3. Bradshaw CS, Jensen JS, Tabrizi SN, et al. Azithromycin failure in Mycoplasma
genitalium urethritis. Emerg Infect Dis 2006; 12(7):1149–52.
4. Jensen JS, Bradshaw CS, Tabrizi SN, Fairley CK, Hamasuna R. Azithromycin
treatment failure in Mycoplasma genitalium-positive patients with nongonococcal
urethritis is associated with induced macrolide resistance. Clin Infect Dis 2008;
47(12): 1546–53.
5. Lau A, Bradshaw CS, Lewis D, et al. The efficacy of azithromycin for the treatment
of genital Mycoplasma genitalium: a systematic review and meta-analysis. Clin
Infect Dis 2015; 61(9):1389–99.
6. Bissessor M, Tabrizi SN, Twin J, et al. Macrolide resistance and azithromycin fail-
ure in a Mycoplasma genitalium-infected cohort and response of azithromycin
failures to alternative antibiotic regimens. Clin Infect Dis 2015; 60(8): 1228–36.
7. Jensen JS, Bradshaw C. Management of Mycoplasma genitalium infections—can
we hit a moving target? BMC Infect Dis 2015; 15: 343.
8. Horner PJ, Blee K, Falk L, van der Meijden W, Moi H. 2016 European guide-
line on the management of non-gonococcal urethritis. Int J STD AIDS 2016;
27(11):928–37.
9. Anagrius C, Loré B, Jensen JS. Treatment of Mycoplasma genitalium. Observations
from a Swedish STD clinic. PLoS One 2013; 8(4):e61481.
10. Bjornelius E, Anagrius C, Bojs G, et  al. Antibiotic treatment of symptomatic
Mycoplasma genitalium infection in Scandinavia: a controlled clinical trial. Sex
Transm Infect 2008; 84(1): 72–6.
11. Falk L, Enger M, Jensen JS. Time to eradication of Mycoplasma genitalium
after antibiotic treatment in men and women. J Antimicrob Chemother 2015;
70(11):3134–40.
12. Twin J, Taylor N, Garland SM, et al. Comparison of two Mycoplasma genitalium
real-time PCR detection methodologies. J Clin Microbiol 2011; 49(3):1140–2.
13. Gundevia Z, Foster R, Jamil MS, McNulty A. Positivity at test of cure following
first-line treatment for genital Mycoplasma genitalium: follow-up of a clinical
cohort. Sexually Transmitted Infections 2015; 91(1): 11–3.
14. Jernberg E, Moghaddam A, Moi H. Azithromycin and moxifloxacin for microbi-
ological cure of Mycoplasma genitalium infection: an open study. Int J STD AIDS
2008; 19(10): 676–9.
15. Chrisment D, Charron A, Cazanave C, Pereyre S, Bébéar C. Detection of mac-
rolide resistance in Mycoplasma genitalium in France. J Antimicrob Chemother
2012; 67(11):2598–601.
16. Unemo M, Endre KM, Moi H. Five-day azithromycin treatment regimen for
Mycoplasma genitalium infection also effectively eradicates Chlamydia trachoma-
tis. Acta Derm Venereol 2015; 95(6):730–2.
17. Manhart LE, Jensen JS, Bradshaw CS, Golden MR, Martin DH. Efficacy of anti-
microbial therapy for Mycoplasma genitalium infections. Clin Infect Dis 2015; 61
Suppl 8: S802–17.
18. Guschin A, Ryzhikh P, Rumyantseva T, Gomberg M, Unemo M. Treatment effi-
cacy, treatment failures and selection of macrolide resistance in patients with high
load of Mycoplasma genitalium during treatment of male urethritis with josamy-
cin. BMC Infect Dis 2015; 15:40.
19. Chow EP, Tomnay J, Fehler G, et al. Substantial increases in chlamydia and gon-
orrhea positivity unexplained by changes in individual-level sexual behaviors
among men who have sex with men in an Australian sexual health service from
2007 to 2013. Sexually Transmitted Diseases 2015; 42(2): 81–7.
20. Kirkcaldy RD, Zaidi A, Hook EW 3rd, et al. Neisseria gonorrhoeae antimicrobial
resistance among men who have sex with men and men who have sex exclusively
with women: the Gonococcal Isolate Surveillance Project, 2005–2010. Ann Intern
Med 2013; 158(5 Pt 1): 321–8.
21. Lewis DA. The role of core groups in the emergence and dissemination of antimi-
crobial-resistant N. gonorrhoeae. Sexually transmitted infections 2013; 89 Suppl
4: iv47–51.