Clinical Infectious Diseases Advance Access published January 16, 2015
MAJOR ARTICLE
Macrolide Resistance and Azithromycin Failure in a Mycoplasma genitalium–Infected Cohort and Response of Azithromycin Failures to Alternative Antibiotic Regimens Downloaded from http://cid.oxfordjournals.org/ at Rudolph Matas Medical Library on January 20, 2015
Melanie Bissessor,1,2 Sepehr N. Tabrizi,3,4 Jimmy Twin,4 Houda Abdo,4 Christopher K. Fairley,1,5 Marcus Y. Chen,1,5 Lenka A. Vodstrcil,1,2,4 Jorgen S. Jensen,6 Jane S. Hocking,2 Suzanne M. Garland,3 and Catriona S. Bradshaw1,2,5 1
Melbourne Sexual Health Centre, Alfred Hospital, 2Melbourne School of Population and Global Health, and 3Department of Obstetrics and Gynaecology, University of Melbourne, 4Department of Microbiology, Infectious Diseases, The Royal Women’s Hospital, Murdoch Childrens Research Institute, and 5 Central Clinical School, Monash University, Melbourne, Victoria, Australia; and 6Statens Serum Institut, Copenhagen, Denmark
Background. Our aim was to determine the efficacy of 1 g azithromycin and alternative antibiotic regimens in a prospective cohort of Mycoplasma genitalium–infected participants, and factors associated with azithromycin failure. Methods. Consecutive eligible M. genitalium–infected men and women attending the Melbourne Sexual Health Centre between July 2012 and June 2013 were treated with 1 g of azithromycin and retested by polymerase chain reaction (PCR) on days 14 and 28. Cure was defined as PCR negative on day 28. Cases failing azithromycin were treated with moxifloxacin, and those failing moxifloxacin were treated with pristinamycin. Pre- and posttreatment samples were assessed for macrolide resistance mutations (MRMs) by high-resolution melt analysis. Mycoplasma genitalium samples from cases failing moxifloxacin were sequenced for fluoroquinolone resistance mutations. Multivariable analysis was used to examine associations with azithromycin failure. Results. Of 155 participants treated with 1 g azithromycin, 95 (61% [95% confidence interval {CI}, 53%–69%]) were cured. Pretreatment MRM was detected in 56 (36% [95% CI, 28%–43%]) participants, and strongly associated with treatment failure (87% [95% CI, 76%–94%]; adjusted odds ratio, 47.0 [95% CI, 17.1–129.0]). All 11 participants who had MRM detected in posttreatment samples failed azithromycin. Moxifloxacin was effective in 53(88% [95% CI, 78%–94%]) of 60 cases failing azithromycin; all failures had gyrA and parC mutations detected in pretreatment samples. Six of 7 patients failing moxifloxacin treatment received pristinamycin, and all were PCR negative 28 days after pristinamycin treatment. Conclusions. We report a high azithromycin failure rate (39%) in an M. genitalium–infected cohort in association with high levels of pretreatment macrolide resistance. Moxifloxacin failure occurred in 12% of patients who received moxifloxacin; all had pretreatment fluoroquinolone mutations detected. Pristinamycin was highly effective in treating macrolide- and quinolone-resistant strains. Keywords.
Mycoplasma genitalium; macrolide resistance; azithromycin; moxifloxacin; pristinamycin.
Received 13 September 2014; accepted 12 December 2014. Correspondence: Melanie Bissessor, FRACGP, FAChSM, 580 Swanston Street, Carlton, VIC 3053, Australia ([email protected]). Clinical Infectious Diseases® © The Author 2014. Published by Oxford University Press on behalf of the Infectious Diseases Society of America. All rights reserved. For Permissions, please e-mail: [email protected]. DOI: 10.1093/cid/ciu1162
Mycoplasma genitalium is an established cause of nongonoccocal urethritis (NGU) and cervicitis, and implicated in pelvic inflammatory disease (PID) [1, 2] and increased transmission of human immunodeficiency virus [3–5]. Widespread use of 1 g azithromycin for M. genitalium has been associated with emergence of macrolide resistance [6–16] and declining cure rates in some settings [8, 9, 14, 15, 17]. Moxifloxacin has been
Macrolide Resistance and Azithromycin Failure in M. genitalium
•
CID
•
1
shown to be effective in M. genitalium infections failing azithromycin [1, 17]; however, reports of failure have emerged in association with mutations in the gyrA and parC genes [9, 11, 17–19]. We conducted a prospective cohort study of M. genitalium– infected men and women attending an urban sexually transmitted infection (STI) clinic with the primary aim of determining the microbiological cure rate of 1 g of azithromycin. Additional aims included determining the (1) contribution of macrolide resistance mutations (MRMs) and organism load to azithromycin failure; (2) optimal timing of test of cure (TOC) following azithromycin; (3) effectiveness of moxifloxacin in cases failing azithromycin; and (4) effectiveness of pristinamycin in cases failing moxifloxacin.
Recruitment and Patient Management
This study was conducted between 1 July 2012 and 30 June 2013 at Melbourne Sexual Health Centre, Victoria, Australia. Routine testing for M. genitalium is performed in patients with NGU, cervicitis, and/or PID, and sexual contacts of infected partners. Clinic attendees aged 18 years and older, diagnosed with M. genitalium by quantitative PCR (qPCR) [20, 21] and treated with 1 g of azithromycin first-line were eligible. Participants were asked to abstain from sexual activity for 14 days and to return 14 and 28 days after azithromycin treatment for repeat testing. Partner notification/treatment was actively followed up. Participants completed detailed questionnaires at day 14 and 28 on antibiotic use, STI symptoms, and interim sexual activity. Cases returning a day 28 TOC within 56 days of treatment were included in analyses to allow for delay in clinic reattendance for TOC. Where reinfection was suspected on history, index patients and contactable partners were recalled and retreated simultaneously with 1 g azithromycin, with repeat day 14 and 28 TOCs following retreatment; only data following retreatment were included in analyses. MRM results and day 14 TOCs were not available to clinicians and did not influence management. Azithromycin cure was defined as a negative TOC 28 days following 1 g azithromycin. Azithromycin failure was defined as either (1) positive day 28 TOC with no reinfection risk (moxifloxacin given at day 28) or (2) persistent symptoms prior to day 28 with no reinfection risk (moxifloxacin given prior to day 28). Moxifloxacin was prescribed as 400 mg daily for 10 days. Patients who remained M. genitalium PCR positive at day 28 following moxifloxacin, and who had no reinfection risk, were given 1 g pristinamycin 4 times daily for 10 days, and retested at day 28. The selection and dose of pristinamycin was based on in vitro data acquired from prior moxifloxacin failures (J. Jensen, personal communication).
2
•
CID
•
Bissessor et al
One milliliter of first-void urine specimen was centrifuged for 10 minutes at 10 000g and the pellet resuspended in 200 µL of phosphate-buffered saline (PBS). Genital swabs were rotated 10 times in 400 µL of PBS. Two hundred microliters of PBS containing urine pellet or swab cells was then extracted using MagNA pure 96 platform (Roche Diagnostic) and eluted in 100 µL of MagNA pure elution buffer. Detection of M. genitalium DNA was performed using qPCR targeting a 517-bp region of the 16S ribosomal RNA (rRNA) gene [20]. Mycoplasma genitalium load was determined by qPCR assay targeting the MgPa gene [20] using genomic DNA from M. genitalium (ATCC 33530) as the quantified comparator. This assay was run on a Roche LC480 real-time PCR instrument using SensiFAST Probe No-ROX chemistry (Bioline). Testing for mutations at positions 2058 and 2059 (Escherichia coli numbering) in the 23S rRNA gene of M. genitalium was performed on all pretreatment and persistently positive posttreatment samples using high-resolution melt analysis [10]. Fluoroquinolone mutations in parC and gyrA were detected using Sanger DNA sequencing in samples limited to cases not responding to moxifloxacin [11, 18–19]. Statistical Methods
A sample size of 155 provided 95% confidence intervals (CIs) of 73%–86% around an estimated treatment efficacy of 1 g azithromycin of 80%. Data were analyzed using SPSS software version 20. The χ2 and Fisher exact tests were used to calculate P values, where indicated. Multivariable logistic regression analysis was conducted to examine factors associated with treatment failure. Due to significant correlation between sex and symptoms, 2 separate multivariable models are presented including only 1 of the correlated variables. Analysis of the association between M. genitalium load (log10 transformed) and treatment failure was limited to urine samples from men as organism load is influenced by sample type, and male urine comprised the majority of samples. Linear regression was used to determine if M. genitalium load differed in pretreatment male urine samples between cases with no MRM, pretreatment MRM, and posttreatment MRM. Ethical approval for this study was granted by the Alfred Hospital Research Ethics Committee (number 150/12). RESULTS Recruitment and Participation
One hundred seventy-two men and women were diagnosed with M. genitalium during the study period (Figure 1). Twelve patients were ineligible as they did not receive 1 g azithromycin (n = 10) or were departing Australia (n = 2). Of the 160 enrolled participants, 155 (97%) completed all aspects of the study; 5 participants did not provide follow-up samples within 56 days.
Downloaded from http://cid.oxfordjournals.org/ at Rudolph Matas Medical Library on January 20, 2015
METHODS
Laboratory Methods
Downloaded from http://cid.oxfordjournals.org/ at Rudolph Matas Medical Library on January 20, 2015
Figure 1.
Recruitment and participation. Abbreviations: MG, Mycoplasma genitalium; PCR, polymerase chain reaction; TOC, test of cure.
Demographic, Clinical, and Laboratory Characteristics
One hundred twelve (72%) participants were men and 43 (28%) were women (Table 1). One hundred seven (95%) M. genitalium–infected men had symptomatic NGU and only 5 (5%) were asymptomatic contacts, whereas the majority of
M. genitalium–infected women presented as asymptomatic contacts of men (n = 25 [58%]). The remaining 18 symptomatic women reported vaginal or urinary symptoms. Importantly, sex and symptoms were highly correlated in this population. Men were predominantly diagnosed using first-void urine (96%),
Macrolide Resistance and Azithromycin Failure in M. genitalium
•
CID
•
3
Table 1. Characteristics of Participants
Characteristic
Male (n = 112), No. (%)
Age, y, median (range) Male sexual partner in last 3 mo
23 (19–55)
22 (20–51)
Macrolide Resistance Mutations and Response to Azithromycin
No Yes Asymptomatic Symptomatic
Female (n = 43), No. (%)
80 (71)
0 (0)
32 (29) 5 (5)
43 (100) 25 (48)
107a (95)
Contacts of Mycoplasma genitalium infection
18b (42)
5 (5)
25 (58)
107 (96) 5c (4)
8 (19) 35d (81)
Specimen type First-void urine Genital swab
asymptomatic at day 14 and 28. Notably, all 60 participants with a positive day 14 TOC had MRM detected in either pretreatment (n = 49) or posttreatment (n = 11) samples; 38 (63%) were symptomatic by day 14.
One hundred seven is the denominator for proportion of men with symptoms or signs.
b
Eighteen is the denominator for proportion of women with symptoms or signs.
c
Four urethral; 1 anal.
d
Eleven cervical; 24 high vaginal swabs.
whereas 81% of women were diagnosed by high vaginal or cervical swab. Treatment Outcomes Following 1 g Azithromycin
Ninety-five cases had a negative day 28 TOC, yielding an azithromycin cure rate of 61% (95% CI, 53%–69%). Sixty patients failed azithromycin (39% [95% CI, 31%–47%]); 22 (37%) were PCR positive at day 28 and received moxifloxacin, and 38 (63%) had interim symptoms and received moxifloxacin prior to day 28. All 60 patients who failed treatment were also PCR positive at day 14. Management of Reinfection
Following treatment, 18 (12%) participants reported interim sexual activity with an untreated partner. All 18 patients and partners were successfully recalled, retreated simultaneously with 1 g azithromycin, and recommenced at day 0, with only TOC data following retreatment included in analyses. Five of the 18 (27%) at risk of reinfection failed azithromycin following retreatment of the index and partner, compared with 55 of the 137 (40%) not at risk of reinfection (P = .31). All 5 remained symptomatic despite simultaneous partner retreatment, and were prescribed moxifloxacin prior to day 28 with their partners.
Associations With Azithromycin Failure by Multivariable Analysis
Male sex, symptoms, and presence of pretreatment MRM were significantly associated with azithromycin failure by univariate analysis (Table 3); no significant association was found between MSM status and treatment failure (odds ratio [OR], 1.7 [95% CI, .8–4.0]; P = .187). Because sex and symptoms were highly correlated, Table 3 presents 2 adjusted analyses including each of the correlated variables. The presence of pretreatment MRM significantly increased the odds of azithromycin failure in both models, and having symptoms was associated with a 5-fold increased risk of failure (model 2), whereas being male was not associated with a significant increase in azithromycin failure (model 1). Association of M. genitalium Load With Treatment Failure Within Men Tested by First-Pass Urine
The association of organism load with azithromycin failure was examined within urine samples from male participants, and not the whole cohort, as sample type influences organism load. Within men tested by first-pass urine, the overall median M. genitalium load was 3.78 log10 load/sample. By univariate
Table 2. Macrolide Resistance Mutations Detected by HighResolution Melt Analysis Pretreatment Mutation (n = 56), No. (%)
Posttreatment Mutation (n = 11b), No. (%)
A2059G/A2058G A2058C/A2059C
50 (89) 4 (7)
9 (82) 0 (0)
A2059T/A2058T
1 (2)
2 (18)
Timing of Test of Cure
Mixed
1 (2)
0 (0)
There was a median of 14 days (interquartile range [IQR], 12– 15 days) between treatment and first TOC, and 28 days (IQR, 27–36 days) between treatment and second TOC. All 95 participants with a negative day 28 TOC were negative at day 14 and
a
4
•
CID
•
Bissessor et al
Mutationa
High-resolution melt analysis does not differentiate between the singlenucleotide polymorphisms at position 2058 and 2059, only the actual nucleotide substitution.
b
Nine posttreatment mutations detected at day 14 and 2 posttreatment mutations detected at day 28.
Downloaded from http://cid.oxfordjournals.org/ at Rudolph Matas Medical Library on January 20, 2015
a
MRMs were detected in 56 (36% [95% CI, 28%–43%]) pretreatment samples (Table 2). The most common SNP was adenine to guanine (A2059G/A2058G), detected in 89% of the 56 samples. Forty-nine (87% [95% CI, 76%–94%]) of the 56 failed azithromycin, but 7 (13% [95% CI, 7%–20%]) had negative PCRs at day 14 and day 28 and were asymptomatic. In the remaining 99 participants without pretreatment MRMs, 11 had MRMs detected exclusively in posttreatment samples; all 11 failed azithromycin (PCR positive at day 14 and 28).
Table 3. Factors Associated With Azithromycin Failure in the Study Population (N = 155)
Factor
Azithromycin Cure, No. (%)
Azithromycin Failure, No. (%)
Sex Female
35 (38)
8 (13)
Male
60 (62)
52 (87)
Symptomatic No
26 (27)
4 (7)
69 (73)
56 (93)
Pretreatment MRM None 88 (93)
11c (20)
Yes
Present
7 (7)
49 (80)
P Value
OR (95% CI) 1.0
.001
3.8 (1.6–8.9)
Model 1 AORa (95% CI)
Model 2 AORb (95% CI)
1.0
1.0
2.9 (.9–9.6)
1.0
.001
...
5.3 (1.8–16.1)
P Value .082
... 1.0
.042
5.0 (1.1–22.9)
1.0
MAJOR ARTICLE
Macrolide Resistance and Azithromycin Failure in a Mycoplasma genitalium–Infected Cohort and Response of Azithromycin Failures to Alternative Antibiotic Regimens Downloaded from http://cid.oxfordjournals.org/ at Rudolph Matas Medical Library on January 20, 2015
Melanie Bissessor,1,2 Sepehr N. Tabrizi,3,4 Jimmy Twin,4 Houda Abdo,4 Christopher K. Fairley,1,5 Marcus Y. Chen,1,5 Lenka A. Vodstrcil,1,2,4 Jorgen S. Jensen,6 Jane S. Hocking,2 Suzanne M. Garland,3 and Catriona S. Bradshaw1,2,5 1
Melbourne Sexual Health Centre, Alfred Hospital, 2Melbourne School of Population and Global Health, and 3Department of Obstetrics and Gynaecology, University of Melbourne, 4Department of Microbiology, Infectious Diseases, The Royal Women’s Hospital, Murdoch Childrens Research Institute, and 5 Central Clinical School, Monash University, Melbourne, Victoria, Australia; and 6Statens Serum Institut, Copenhagen, Denmark
Background. Our aim was to determine the efficacy of 1 g azithromycin and alternative antibiotic regimens in a prospective cohort of Mycoplasma genitalium–infected participants, and factors associated with azithromycin failure. Methods. Consecutive eligible M. genitalium–infected men and women attending the Melbourne Sexual Health Centre between July 2012 and June 2013 were treated with 1 g of azithromycin and retested by polymerase chain reaction (PCR) on days 14 and 28. Cure was defined as PCR negative on day 28. Cases failing azithromycin were treated with moxifloxacin, and those failing moxifloxacin were treated with pristinamycin. Pre- and posttreatment samples were assessed for macrolide resistance mutations (MRMs) by high-resolution melt analysis. Mycoplasma genitalium samples from cases failing moxifloxacin were sequenced for fluoroquinolone resistance mutations. Multivariable analysis was used to examine associations with azithromycin failure. Results. Of 155 participants treated with 1 g azithromycin, 95 (61% [95% confidence interval {CI}, 53%–69%]) were cured. Pretreatment MRM was detected in 56 (36% [95% CI, 28%–43%]) participants, and strongly associated with treatment failure (87% [95% CI, 76%–94%]; adjusted odds ratio, 47.0 [95% CI, 17.1–129.0]). All 11 participants who had MRM detected in posttreatment samples failed azithromycin. Moxifloxacin was effective in 53(88% [95% CI, 78%–94%]) of 60 cases failing azithromycin; all failures had gyrA and parC mutations detected in pretreatment samples. Six of 7 patients failing moxifloxacin treatment received pristinamycin, and all were PCR negative 28 days after pristinamycin treatment. Conclusions. We report a high azithromycin failure rate (39%) in an M. genitalium–infected cohort in association with high levels of pretreatment macrolide resistance. Moxifloxacin failure occurred in 12% of patients who received moxifloxacin; all had pretreatment fluoroquinolone mutations detected. Pristinamycin was highly effective in treating macrolide- and quinolone-resistant strains. Keywords.
Mycoplasma genitalium; macrolide resistance; azithromycin; moxifloxacin; pristinamycin.
Received 13 September 2014; accepted 12 December 2014. Correspondence: Melanie Bissessor, FRACGP, FAChSM, 580 Swanston Street, Carlton, VIC 3053, Australia ([email protected]). Clinical Infectious Diseases® © The Author 2014. Published by Oxford University Press on behalf of the Infectious Diseases Society of America. All rights reserved. For Permissions, please e-mail: [email protected]. DOI: 10.1093/cid/ciu1162
Mycoplasma genitalium is an established cause of nongonoccocal urethritis (NGU) and cervicitis, and implicated in pelvic inflammatory disease (PID) [1, 2] and increased transmission of human immunodeficiency virus [3–5]. Widespread use of 1 g azithromycin for M. genitalium has been associated with emergence of macrolide resistance [6–16] and declining cure rates in some settings [8, 9, 14, 15, 17]. Moxifloxacin has been
Macrolide Resistance and Azithromycin Failure in M. genitalium
•
CID
•
1
shown to be effective in M. genitalium infections failing azithromycin [1, 17]; however, reports of failure have emerged in association with mutations in the gyrA and parC genes [9, 11, 17–19]. We conducted a prospective cohort study of M. genitalium– infected men and women attending an urban sexually transmitted infection (STI) clinic with the primary aim of determining the microbiological cure rate of 1 g of azithromycin. Additional aims included determining the (1) contribution of macrolide resistance mutations (MRMs) and organism load to azithromycin failure; (2) optimal timing of test of cure (TOC) following azithromycin; (3) effectiveness of moxifloxacin in cases failing azithromycin; and (4) effectiveness of pristinamycin in cases failing moxifloxacin.
Recruitment and Patient Management
This study was conducted between 1 July 2012 and 30 June 2013 at Melbourne Sexual Health Centre, Victoria, Australia. Routine testing for M. genitalium is performed in patients with NGU, cervicitis, and/or PID, and sexual contacts of infected partners. Clinic attendees aged 18 years and older, diagnosed with M. genitalium by quantitative PCR (qPCR) [20, 21] and treated with 1 g of azithromycin first-line were eligible. Participants were asked to abstain from sexual activity for 14 days and to return 14 and 28 days after azithromycin treatment for repeat testing. Partner notification/treatment was actively followed up. Participants completed detailed questionnaires at day 14 and 28 on antibiotic use, STI symptoms, and interim sexual activity. Cases returning a day 28 TOC within 56 days of treatment were included in analyses to allow for delay in clinic reattendance for TOC. Where reinfection was suspected on history, index patients and contactable partners were recalled and retreated simultaneously with 1 g azithromycin, with repeat day 14 and 28 TOCs following retreatment; only data following retreatment were included in analyses. MRM results and day 14 TOCs were not available to clinicians and did not influence management. Azithromycin cure was defined as a negative TOC 28 days following 1 g azithromycin. Azithromycin failure was defined as either (1) positive day 28 TOC with no reinfection risk (moxifloxacin given at day 28) or (2) persistent symptoms prior to day 28 with no reinfection risk (moxifloxacin given prior to day 28). Moxifloxacin was prescribed as 400 mg daily for 10 days. Patients who remained M. genitalium PCR positive at day 28 following moxifloxacin, and who had no reinfection risk, were given 1 g pristinamycin 4 times daily for 10 days, and retested at day 28. The selection and dose of pristinamycin was based on in vitro data acquired from prior moxifloxacin failures (J. Jensen, personal communication).
2
•
CID
•
Bissessor et al
One milliliter of first-void urine specimen was centrifuged for 10 minutes at 10 000g and the pellet resuspended in 200 µL of phosphate-buffered saline (PBS). Genital swabs were rotated 10 times in 400 µL of PBS. Two hundred microliters of PBS containing urine pellet or swab cells was then extracted using MagNA pure 96 platform (Roche Diagnostic) and eluted in 100 µL of MagNA pure elution buffer. Detection of M. genitalium DNA was performed using qPCR targeting a 517-bp region of the 16S ribosomal RNA (rRNA) gene [20]. Mycoplasma genitalium load was determined by qPCR assay targeting the MgPa gene [20] using genomic DNA from M. genitalium (ATCC 33530) as the quantified comparator. This assay was run on a Roche LC480 real-time PCR instrument using SensiFAST Probe No-ROX chemistry (Bioline). Testing for mutations at positions 2058 and 2059 (Escherichia coli numbering) in the 23S rRNA gene of M. genitalium was performed on all pretreatment and persistently positive posttreatment samples using high-resolution melt analysis [10]. Fluoroquinolone mutations in parC and gyrA were detected using Sanger DNA sequencing in samples limited to cases not responding to moxifloxacin [11, 18–19]. Statistical Methods
A sample size of 155 provided 95% confidence intervals (CIs) of 73%–86% around an estimated treatment efficacy of 1 g azithromycin of 80%. Data were analyzed using SPSS software version 20. The χ2 and Fisher exact tests were used to calculate P values, where indicated. Multivariable logistic regression analysis was conducted to examine factors associated with treatment failure. Due to significant correlation between sex and symptoms, 2 separate multivariable models are presented including only 1 of the correlated variables. Analysis of the association between M. genitalium load (log10 transformed) and treatment failure was limited to urine samples from men as organism load is influenced by sample type, and male urine comprised the majority of samples. Linear regression was used to determine if M. genitalium load differed in pretreatment male urine samples between cases with no MRM, pretreatment MRM, and posttreatment MRM. Ethical approval for this study was granted by the Alfred Hospital Research Ethics Committee (number 150/12). RESULTS Recruitment and Participation
One hundred seventy-two men and women were diagnosed with M. genitalium during the study period (Figure 1). Twelve patients were ineligible as they did not receive 1 g azithromycin (n = 10) or were departing Australia (n = 2). Of the 160 enrolled participants, 155 (97%) completed all aspects of the study; 5 participants did not provide follow-up samples within 56 days.
Downloaded from http://cid.oxfordjournals.org/ at Rudolph Matas Medical Library on January 20, 2015
METHODS
Laboratory Methods
Downloaded from http://cid.oxfordjournals.org/ at Rudolph Matas Medical Library on January 20, 2015
Figure 1.
Recruitment and participation. Abbreviations: MG, Mycoplasma genitalium; PCR, polymerase chain reaction; TOC, test of cure.
Demographic, Clinical, and Laboratory Characteristics
One hundred twelve (72%) participants were men and 43 (28%) were women (Table 1). One hundred seven (95%) M. genitalium–infected men had symptomatic NGU and only 5 (5%) were asymptomatic contacts, whereas the majority of
M. genitalium–infected women presented as asymptomatic contacts of men (n = 25 [58%]). The remaining 18 symptomatic women reported vaginal or urinary symptoms. Importantly, sex and symptoms were highly correlated in this population. Men were predominantly diagnosed using first-void urine (96%),
Macrolide Resistance and Azithromycin Failure in M. genitalium
•
CID
•
3
Table 1. Characteristics of Participants
Characteristic
Male (n = 112), No. (%)
Age, y, median (range) Male sexual partner in last 3 mo
23 (19–55)
22 (20–51)
Macrolide Resistance Mutations and Response to Azithromycin
No Yes Asymptomatic Symptomatic
Female (n = 43), No. (%)
80 (71)
0 (0)
32 (29) 5 (5)
43 (100) 25 (48)
107a (95)
Contacts of Mycoplasma genitalium infection
18b (42)
5 (5)
25 (58)
107 (96) 5c (4)
8 (19) 35d (81)
Specimen type First-void urine Genital swab
asymptomatic at day 14 and 28. Notably, all 60 participants with a positive day 14 TOC had MRM detected in either pretreatment (n = 49) or posttreatment (n = 11) samples; 38 (63%) were symptomatic by day 14.
One hundred seven is the denominator for proportion of men with symptoms or signs.
b
Eighteen is the denominator for proportion of women with symptoms or signs.
c
Four urethral; 1 anal.
d
Eleven cervical; 24 high vaginal swabs.
whereas 81% of women were diagnosed by high vaginal or cervical swab. Treatment Outcomes Following 1 g Azithromycin
Ninety-five cases had a negative day 28 TOC, yielding an azithromycin cure rate of 61% (95% CI, 53%–69%). Sixty patients failed azithromycin (39% [95% CI, 31%–47%]); 22 (37%) were PCR positive at day 28 and received moxifloxacin, and 38 (63%) had interim symptoms and received moxifloxacin prior to day 28. All 60 patients who failed treatment were also PCR positive at day 14. Management of Reinfection
Following treatment, 18 (12%) participants reported interim sexual activity with an untreated partner. All 18 patients and partners were successfully recalled, retreated simultaneously with 1 g azithromycin, and recommenced at day 0, with only TOC data following retreatment included in analyses. Five of the 18 (27%) at risk of reinfection failed azithromycin following retreatment of the index and partner, compared with 55 of the 137 (40%) not at risk of reinfection (P = .31). All 5 remained symptomatic despite simultaneous partner retreatment, and were prescribed moxifloxacin prior to day 28 with their partners.
Associations With Azithromycin Failure by Multivariable Analysis
Male sex, symptoms, and presence of pretreatment MRM were significantly associated with azithromycin failure by univariate analysis (Table 3); no significant association was found between MSM status and treatment failure (odds ratio [OR], 1.7 [95% CI, .8–4.0]; P = .187). Because sex and symptoms were highly correlated, Table 3 presents 2 adjusted analyses including each of the correlated variables. The presence of pretreatment MRM significantly increased the odds of azithromycin failure in both models, and having symptoms was associated with a 5-fold increased risk of failure (model 2), whereas being male was not associated with a significant increase in azithromycin failure (model 1). Association of M. genitalium Load With Treatment Failure Within Men Tested by First-Pass Urine
The association of organism load with azithromycin failure was examined within urine samples from male participants, and not the whole cohort, as sample type influences organism load. Within men tested by first-pass urine, the overall median M. genitalium load was 3.78 log10 load/sample. By univariate
Table 2. Macrolide Resistance Mutations Detected by HighResolution Melt Analysis Pretreatment Mutation (n = 56), No. (%)
Posttreatment Mutation (n = 11b), No. (%)
A2059G/A2058G A2058C/A2059C
50 (89) 4 (7)
9 (82) 0 (0)
A2059T/A2058T
1 (2)
2 (18)
Timing of Test of Cure
Mixed
1 (2)
0 (0)
There was a median of 14 days (interquartile range [IQR], 12– 15 days) between treatment and first TOC, and 28 days (IQR, 27–36 days) between treatment and second TOC. All 95 participants with a negative day 28 TOC were negative at day 14 and
a
4
•
CID
•
Bissessor et al
Mutationa
High-resolution melt analysis does not differentiate between the singlenucleotide polymorphisms at position 2058 and 2059, only the actual nucleotide substitution.
b
Nine posttreatment mutations detected at day 14 and 2 posttreatment mutations detected at day 28.
Downloaded from http://cid.oxfordjournals.org/ at Rudolph Matas Medical Library on January 20, 2015
a
MRMs were detected in 56 (36% [95% CI, 28%–43%]) pretreatment samples (Table 2). The most common SNP was adenine to guanine (A2059G/A2058G), detected in 89% of the 56 samples. Forty-nine (87% [95% CI, 76%–94%]) of the 56 failed azithromycin, but 7 (13% [95% CI, 7%–20%]) had negative PCRs at day 14 and day 28 and were asymptomatic. In the remaining 99 participants without pretreatment MRMs, 11 had MRMs detected exclusively in posttreatment samples; all 11 failed azithromycin (PCR positive at day 14 and 28).
Table 3. Factors Associated With Azithromycin Failure in the Study Population (N = 155)
Factor
Azithromycin Cure, No. (%)
Azithromycin Failure, No. (%)
Sex Female
35 (38)
8 (13)
Male
60 (62)
52 (87)
Symptomatic No
26 (27)
4 (7)
69 (73)
56 (93)
Pretreatment MRM None 88 (93)
11c (20)
Yes
Present
7 (7)
49 (80)
P Value
OR (95% CI) 1.0
.001
3.8 (1.6–8.9)
Model 1 AORa (95% CI)
Model 2 AORb (95% CI)
1.0
1.0
2.9 (.9–9.6)
1.0
.001
...
5.3 (1.8–16.1)
P Value .082
... 1.0
.042
5.0 (1.1–22.9)
1.0
