Mycoplasma genitalium (MG) has been associated with cervicitis, pelvic inflammatory disease (PID), spontaneous abortion and preterm delivery in a recent meta-analysis.1,2 However, there are limited published data on the contribution of MG to PID. More data are needed to establish the attributable risk of MG for female genital tract infections.2 International guidelines for treatment of PID vary, however predominantly recommend presumptive use of antimicrobials, such as tetracyclines, beta-lactams, and nitro-imidazoles, before detection of the causative organism. These regimens aim to treat Chlamydia trachomatis (CT), Neisseria gonorrhea (NG) and anaerobes, and do not contain highly effective antimicrobials against MG.3,4 When PID was treated in accordance with US Centers for Disease Control and Prevention (CDC) guidelines with cefoxitin and doxycycline, 41% of women with PID remained MG positive after 30 days.5 Inadequately treated PID increases the risk of chronic pelvic pain, ectopic pregnancy and infertility.6
Mycoplasma genitalium has marked propensity to develop antimicrobial resistance, with macrolide resistance seen in 40% to 60% of infections in many countries.7,8 Moxifloxacin has been highly effective against macrolide-resistant MG,9 is active against CT and NG,10,11 and was as effective as ofloxacin and metronidazole for all cause PID in a randomized controlled trial (RCT).11 Although only 3 patients in this trial had MG-PID, all were cured.11 The US and UK guidelines now recommend 14 days of moxifloxacin for MG-PID,2,12 although efficacy data are limited,11 and a recent meta-analysis has shown a decline in cure to 89% (95% CI, 82–94%) since 2010.13
With recent regulatory approval of MG assays in many countries, testing for MG in presumptive PID may become increasingly common. However, little has been published on clinical characteristics of MG-PID and whether they differ from those associated with CT-PID. In our study, we compared women who met CDC criteria for presumptive treatment of PID with MG to women with CT-PID to investigate whether MG-PID is associated with distinct clinical and laboratory characteristics. In addition, we report the clinical and microbiological outcomes following moxifloxacin and standard antimicrobial therapy.
MATERIALS AND METHODS
Study Design and Participants
We undertook a retrospective study of women with MG and presumptive PID (henceforth MG-PID) between February 2006 and March 2017, at Melbourne Sexual Health Centre (MSHC), Victoria, Australia. Women were included if they were 18 years or older, PID presumptively diagnosed using CDC criteria, and MG was the sole pathogen (ie, negative for CT and NG). Testing for MG in women with presumptive PID was at clinician discretion from 2006 to 2010, and from 2011 recommended in MSHC protocols. MG-PID cases were women with presumptive PID at presentation (87%), who underwent sexually transmitted infection (STI) testing and were commenced on recommended therapy for PID (henceforth standard treatment). In a minority of cases (13%), women developed symptoms and signs after testing and were diagnosed with presumptive PID at follow up.
We compared characteristics of MG-PID to those of CT-PID (chlamydia sole pathogen), and treatment outcomes (clinical and microbial cure) after standard PID therapy and moxifloxacin. All cases diagnosed with MG-PID between February 2006 and March 2017 were extracted from the clinic database. Cases of CT-PID were selected over the same period from the clinic database.
Definitions and Data Collection
Presumptive PID was diagnosed using CDC criteria, which recommends treatment be initiated in sexually-active women at risk of STIs who are experiencing pelvic or lower abdominal pain and have one of the following: uterine, cervical motion, or adnexal tenderness.3 These criteria are designed to maximize sensitivity for PID diagnosis and err on the side of over rather than undertreatment of PID. Only approximately half of women meeting these criteria have been shown to have laparoscopic evidence of salpingitis or plasma cell endometritis,14,15 although the positive predictive value for these criteria are higher in a STI clinic compared with a community setting.3 All women included in the analysis were sexually active and presented with abdominal pain or pelvic pain, and had either lower abdominal/pelvic tenderness or cervical/adnexal motion tenderness on examination. All reference to MG-PID and CT-PID refers to presumptive PID diagnosis using CDC criteria. Mycoplasma genitalium was detected on vaginal or cervical swab, or first pass urine, using an in-house PCR assay targeting the 16S rRNA gene of MG.16 Chlamydia trachomatis was detected using strand displacement amplification (Becton Dickinson) before 2015, and Aptima transcription-mediated amplification assay from 2015 (Hologic, Marlborough). Bacterial vaginosis (BV) was diagnosed using Amsel's criteria.17 Trichomoniasis was diagnosed by wet preparation and culture. Vaginal and cervical polymorphonuclear leukocyte counts (PMN) on Gram stain were grouped in the following categories: vaginal PMN/high power field (hpf): less than 1, 1 to 4, and 5 or greater; and cervical PMN/hpf: less than 5, 5 to 8, greater than 8.
Epidemiological, clinical, and laboratory data were extracted from paper-based and electronic records for women with MG-PID and CT-PID. Clinical information recorded included: (i) symptoms: lower abdominal pain, dyspareunia, vaginal discharge, fever, intermenstrual or postcoital bleeding; (ii) signs: lower abdominal tenderness, cervical motion or adnexal tenderness, mucopurulent cervicitis, cervical-contact bleeding, abnormal vaginal discharge; (iii) laboratory and microbiological results: vaginal and cervical PMN count, detection of CT, NG, Trichomonas vaginalis, and BV; and (iv) antimicrobial regimens prescribed including duration, adherence and adverse effects.
Patients were excluded if they did not fulfill presumptive PID criteria, were coinfected with another STI, or if symptoms were subsequently attributed to another condition.
Women with presumptive PID were tested and treated with a standard regimen containing metronidazole 400 mg twice daily and doxycycline 100 mg twice daily for 14 days +/− azithromycin 1 g.18 The MG test results were available 48 hours after testing. Beginning in 2011 all MG-positive women were recalled and the standard regimen replaced with moxifloxacin 400 mg once daily for 14 days, in accordance with MSHC guidelines. Before 2011, patients often received standard PID regimens only. Treatment outcomes in all women receiving moxifloxacin and standard regimens were analyzed.
Microbiological cure was defined as a negative test of cure (TOC) between 14 and 120 days. Patients were excluded if: (i) TOC was not performed or outside 14 to 120 days after treatment, (ii) they received less than 7 days of moxifloxacin, (iii) TOC was performed after more than 2 courses of antibiotics.
Clinical cure was defined as resolution of all PID-related symptoms. Antibiotic adherence, partner treatment, reinfection risk, and microbiological and clinical outcomes were recorded, where available. Reinfection risk was defined as none (no sex), possible (sex with a new or treated partner), or probable (sex with a regular partner who had not been tested or treated).
The Alfred Hospital Research Ethics Committee approved this study (project number 304/15).
Data were analyzed using STATA (v14). Univariate and multivariable logistic regression analysis were performed to determine factors associated with MG-PID compared with CT-PID. Variables were included in multivariable models if the P value was 0.05 or less; if correlated, the variable most strongly associated with the outcome was used. Models were built in a forward-stepwise fashion, using the likelihood ratio test to determine the significance of the contribution of each variable to avoid over-fitting the model. Ninety-five percent CIs were calculated for proportions. As vaginal and cervical PMN counts were correlated and vaginal PMN count is a minor CDC criterion, only vaginal PMN was included. Logistic regression was used to assess differences in outcomes by treatment group for those with MG-PID.
During the study period (2006–2017), 149 records of women with MG-PID were identified, and MG accounted for 5.5% of PID cases at MSHC during this timeframe. Ninety-two women fulfilled the criteria for presumptive PID and had MG with no other pathogen detected; 57 women were excluded for reasons listed in Figure 1. Ninety-two CT-PID cases were randomly selected, as described in Materials and Methods. Eight-seven percent of patients had PID diagnosed at their initial consultation, and 13% on return to clinic.
Demographic and Behavioral Characteristics
There was no significant difference between median ages of women with MG-PID and CT-PID (25 years, interquartile range [IQR], 21–29; and 24 years, IQR, 21–28, respectively; Table 1). In unadjusted analyses, compared to CT-PID, women with MG-PID were more likely to be sex workers (odds ratio [OR] 3.07; 95% CI, 1.33–7.08), and to have fewer recent male partners (excludes sex work clients) (MSP > 1; OR, 0.54; 95% CI, 0.30–0.97).
Univariate analyses showed women with MG-PID had similar clinical characteristics to women with CT-PID, although were less likely to report postcoital bleeding (OR, 0.42; 95% CI, 0.18–0.98; Table 1). On clinical examination, MG-PID was more likely to be associated with lower abdominal tenderness (OR, 2.36; 95% CI, 1.29–4.28).
Women were tested for MG by cervical swab (64%), vaginal swab (20%), and first-pass urine (16%). On unadjusted analyses, women with MG-PID were less likely than women with CT-PID to have elevated vaginal or cervical PMN counts (Table 1).
Associations With MG-PID Compared With CT-PID by Multivariable Analyses
Women with MG-PID were more likely to have lower abdominal tenderness (adjusted OR, 2.29; 95% CI, 1.14–4.60), but less likely to have a modest elevation in vaginal PMN counts, compared to women with CT-PID (Table 2).
Treatment Outcomes in MG-PID
Of the 92 women with MG-PID, 54 (59%) received moxifloxacin and 38 (41%) a standard regimen only (Fig. 1). Thirty-seven of 54 women who received moxifloxacin returned for TOC at 14 to 120 days (Fig. 1). Of these women, 10 of 37 (27%) received moxifloxacin only and 27 of 37 (73%)started a standard regimen before being recalled for moxifloxacin. Premoxifloxacin, standard therapy in 23 (85%) of 27 women was azithromycin followed by doxycycline and metronidazole and in 4 (15%) of 27 women was doxycycline and metronidazole without azithromycin. Ceftriaxone, which has no effect on MG, was only dispensed presumptively for individuals considered at risk of gonorrhea. Median duration of standard therapy premoxifloxacin was 7 days (range, 2–14 days). Median time to TOC was 32 days (IQR, 24–41 days). A higher proportion of women receiving moxifloxacin returned for TOC at 14 to 120 days (37/54 [69%]) in contrast to standard treatment (20/38 [53%]) (Fig. 1).
Of the 37 women with MG-PID who had moxifloxacin and a 14 to 120 day TOC, 35/37 [95% (95% CI 82–99)] were microbiologically cured (Table 2, Fig. 1). There was no significant difference in microbiological cure between the moxifloxacin group and the standard treatment group (P = 0.948). Of note, both of the moxifloxacin failures received standard PID treatment before moxifloxacin. Clinical cure, defined as resolution of all pelvic symptoms, was significantly higher in women treated with moxifloxacin (33/37 [89%]; 95% CI, 75–97) compared with standard treatment (10/19 [53%]; 95% CI, 29–76; P = 0.004; Table 2). Of the 4 patients in the moxifloxacin group with persistent symptoms, all had abdominal pain and 2 reported dyspareunia. Of the 9 patients in the standard treatment group with ongoing symptoms, 7 had persistent abdominal pain, and 3 had persistent dyspareunia.
Adherence to Antimicrobial Therapy and Reinfection Risk
Adherence to therapy was documented in 32 of 37 (86%) patients who had moxifloxacin and returned for a TOC visit, with 27 of 32 (84%) (95% CI, 67–95) reporting 100% adherence and 5 of 32 reporting <100% adherence (all took >7 days) (Table 2). Of the 2 microbiological failures following moxifloxacin, one had sex with an untreated partner before TOC and was at high risk of reinfection; the other did not have reinfection risk or adherence documented. Adherence to standard treatment was documented in 16 of 20 patients returning for TOC with 12 of 16 (75%) (95% CI, 48–93) reporting 100% adherence. There was no difference in adherence between the 2 groups (P = 0.436).
Adverse effects were significantly more common among those who took moxifloxacin (P = 0.026), with 15/37 (41%) women who received moxifloxacin reporting side effects (Table 2). The most common side effects were nausea (7 of 15), diarrhea (3 of 15), and candidiasis (2 of 15); 1 patient reported tendon pain. No serious adverse effects were reported but 4 of 5 patients who ceased moxifloxacin early experienced side effects. Of the 20 women who received standard treatment, only 2 (10%) reported any adverse effects, the nature of which was not recorded.
This study of women meeting the criteria for presumptive treatment of PID associated with Mycoplasma genitalium, found a similar clinical presentation to that of CT-PID. Among 37 evaluable women, moxifloxacin microbiologically cured 95% of MG infections but this did not differ to standard treatment. It is interesting that moxifloxacin was associated with significantly higher resolution of clinical symptoms (89%) compared to standard treatment (53%), although side effects were common with moxifloxacin.
There are limited published data examining the association between MG and PID. While cases included in this dataset fulfilled the CDC criteria for presumptive treatment for PID, past studies found that up to half of all cases of presumptive PID did not have histological endometritis or salpingitis.14 Mild abdominal pain has previously been reported in MG-associated acute endometritis.19 Investigators in 1 study found gonococcal-PID to be more severe than MG-PID, but found no difference in the clinical presentation of CT-PID and MG-PID.20 Although we found MG-PID was more likely to be associated with abdominal tenderness than CT-PID, the significance of this finding is unclear. Universal testing for MG in PID at MSHC was not recommended until 2011, and although the case definition of presumptive PID required the presence of abdominal or pelvic pain and lower abdominal/pelvic tenderness or cervical/adnexal motion tenderness on examination, it is possible that only more severe cases were tested for MG before 2011. In contrast to the clinical findings, MG-PID was associated with modest reduction in cervicovaginal PMN response compared to CT-PID, supporting the role of CT as an established cause of PID and morbidity. An important limitation of this study is that outpatient sexual health services are likely to see milder PID. Studies in hospitals attended by women with more severe PID may have different findings. While records were reviewed by the same researcher to control for differences in interpretation of records, missing data and clinician variability in documentation are unavoidable limitations of a retrospective case review. Despite these limitations, this remains the largest study to date of MG-PID examining clinical features.
In this study, both regimens achieved high levels of microbial cure, although clinical cure was significantly higher with moxifloxacin, and only 53% of women receiving standard regimens experienced complete resolution of PID-symptoms. This raises the possibility of persistent low load infection, as this pattern occurs in MG-urethritis.21 However, a direct comparison of treatment outcomes should be interpreted with care due to substantial differences between those who received standard treatment, and those given moxifloxacin. Firstly, treatment regimens were not randomized and were at clinician discretion, with temporal differences between the 2 treatment groups. Women receiving standard treatment only were predominantly treated earlier in the study (60% treated before 2013), while the majority of those in the moxifloxacin group were treated later in the study (85% after 2013). MSHC has experienced an extraordinary rise in macrolide-resistant MG and azithromycin-failure from less than 10% in 2006 to greater than 40% in 2016.21–23 The group who received standard treatment only predominantly reflected the period when macrolide resistance was uncommon and this regimen may be less effective as high levels of macrolide resistance (>50%) are now seen in 2018.21 These data suggest that regimens containing azithromycin may be effective in regions with low levels of macrolide resistance. It is possible that cure is also enhanced by the presence of doxycycline, which we have recently shown has a substantial effect on MG load and selection of macrolide resistance.21 Published data suggest that it is the inclusion of azithromycin that is likely to have affected cure, as when MG-PID was treated with cefoxitin and doxycycline, the microbiological cure rate was only 59%.5 Further data regarding the efficacy of doxycycline-azithromycin inclusive regimens, particularly with respect to clinical cure, would be of value considering the cost and lower tolerability of moxifloxacin.
Adverse effects were significantly more common in women treated with moxifloxacin than standard therapy. Although this may be directly attributable to moxifloxacin, 73% of women receiving moxifloxacin had also been exposed to a standard PID regimen. Gastrointestinal side effects are common with quinolones; however, rare serious adverse effects, including arrhythmias, neuropathy, and tendon rupture, have also been reported. Risk of adverse effects needs to be balanced with risk of serious sequelae from delaying antimicrobial therapy in PID.6,24,25 This study was conducted at a time of rising macrolide resistance in Australia and during the emergence of quinolone resistance in the region. A recent meta-analysis reported the efficacy of moxifloxacin for MG to be declining, from 100% (95% CI, 99–100) in studies before 2010, to 89% (95% CI, 82–94) since.13 Due to the small numbers of treatment failures in this study, temporal trends in efficacy could not be assessed. Of note, a significant proportion of women treated with moxifloxacin had exposure to standard PID regimens before moxifloxacin, which may have improved moxifloxacin cure. Although this may have impacted on the reported efficacy of moxifloxacin, this is also reflection of real world practice in the absence of point of care tests. Moxifloxacin, although costly and associated with more adverse effects, currently remains the only available therapeutic option in most clinical settings for patients with macrolide-resistant MG.
This study reports a case series of women with presumptive PID and MG detected as the sole pathogen. It did not find clinically meaningful differences between women with MG-PID and CT-PID and does not provide prospective data to inform clinicians of the likelihood that MG will lead to PID. The CDC recommended standard PID regimen includes doxycycline but not azithromycin, and has been shown to have low microbiological and clinical cure for MG-PID.5 Macrolide resistance is becoming increasingly common worldwide and in the majority of high-income countries exceeds 40%, so azithromycin-based regimens can be expected to have a declining efficacy. If MG is identified in a woman presenting with the clinical features of PID and no other pathogen is detected, then this study provides data showing moxifloxacin is highly effective in achieving microbiological cure of MG, but that 2 in 5 women will experience predominately mild adverse effects. The future will see point-of-care assays for MG that incorporate resistance markers, and this will assist researchers in determining the efficacy of specific regimens for MG-associated syndromes and clinicians in selecting appropriate antimicrobials to ensure high-level clinical and microbial cure.
1. Lis R, Rowhani-Rahbar A, Manhart LE. Mycoplasma genitalium
infection and female reproductive tract disease: A meta-analysis. Clin Infect Dis 2015; 61:418–426.
2. Wiesenfeld HC, Manhart LE. Mycoplasma genitalium
in women: Current knowledge and research priorities for this recently emerged pathogen. J Infect Dis 2017; 216(suppl 2):S389–S395.
3. Workowski KA, Bolan GA, et alCenters for Disease C. Sexually transmitted diseases treatment guidelines, 2015. MMWR Recomm Rep 2015; 64:1–137.
4. Ross JD, McCarthy G. UK national guideline for the management of pelvic inflammatory disease. British Association for Sexual Health and HIV 2011.
5. Haggerty CL, Totten PA, Astete SG, et al. Failure of cefoxitin and doxycycline to eradicate endometrial Mycoplasma genitalium
and the consequence for clinical cure of pelvic inflammatory disease. Sex Transm Infect 2008; 84:338–342.
6. Ness RB, Soper DE, Holley RL, et al. Effectiveness of inpatient and outpatient treatment strategies for women with pelvic inflammatory disease: Results from the pelvic inflammatory disease evaluation and clinical health (peach) randomized trial. Am J Obstet Gynecol 2002; 186:929–937.
7. Gesink D, Racey CS, Seah C, et al. Mycoplasma genitalium
in Toronto, Ont: Estimates of prevalence and macrolide resistance. Can Fam Physician 2016; 62(2):e96–e101.
8. Dumke R, Thurmer A, Jacobs E. Emergence of Mycoplasma genitalium
strains showing mutations associated with macrolide and fluoroquinolone resistance in the region Dresden, Germany. Diagn Microbiol Infect Dis 2016; 86:221–223.
9. Bradshaw CS, Jensen JS, Tabrizi SN, et al. Azithromycin failure in Mycoplasma genitalium
urethritis. Emerg Infect Dis 2006; 12:1149–1152.
10. Boothby M, Page J, Pryor R, et al. A comparison of treatment outcomes for moxifloxacin versus ofloxacin/metronidazole for first-line treatment of uncomplicated non-gonococcal pelvic inflammatory disease. Int J STD AIDS 2010; 21(3):195–197.
11. Ross JD, Cronje HS, Paszkowski T, et al. Moxifloxacin versus ofloxacin plus metronidazole in uncomplicated pelvic inflammatory disease: Results of a multicentre, double blind, randomised trial. Sex Transm Infect 2006; 82:446–451.
12. Ross J. 2017 United Kingdom National Guideline for the Management of Pelvic Inflammatory Disease. BASHH, 2017.
14. Jacobson L, Westrom L. Objectivized diagnosis of acute pelvic inflammatory disease. Diagnostic and prognostic value of routine laparoscopy. Am J Obstet Gynecol 1969; 105:1088–1098.
15. Sellors J, Mahony J, Goldsmith C, et al. The accuracy of clinical findings and laparoscopy in pelvic inflammatory disease. Am J Obstet Gynecol 1991; 164:113–120.
16. Yoshida T, Maeda S, Deguchi T, et al. Phylogeny-based rapid identification of mycoplasmas and ureaplasmas from urethritis patients. J Clin Microbiol 2002; 40(1):105–110.
17. Money D. The laboratory diagnosis of bacterial vaginosis. Can J Infect Dis Med Microbiol 2005; 16:77–79.
19. Cohen CR, Manhart LE, Bukusi EA, et al. Association between Mycoplasma genitalium
and acute endometritis. Lancet 2002; 359:765–766.
20. Short VL, Totten PA, Ness RB, et al. Clinical presentation of Mycoplasma genitalium
infection versus Neisseria gonorrhoeae
infection among women with pelvic inflammatory disease. Clin Infect Dis 2009; 48:41–47.
21. Read TRH, Fairley CK, Murray GL, et al. Outcomes of resistance-guided sequential treatment of Mycoplasma genitalium
infections: A prospective evaluation. Clin Infect Dis 2018; doi: 10.1093/cid/ciy477.
22. Jensen JS, Bradshaw CS, Tabrizi SN, et al. Azithromycin treatment failure in Mycoplasma genitalium
-positive patients with nongonococcal urethritis is associated with induced macrolide resistance. Clin Infect Dis 2008; 47:1546–1553.
23. Bissessor M, Tabrizi SN, Twin J, et al. Macrolide resistance and azithromycin failure in a Mycoplasma genitalium
-infected cohort and response of azithromycin failures to alternative antibiotic regimens. Clin Infect Dis 2015; 60:1228–1236.
24. Ross JD. Pelvic inflammatory disease. Medicine 2014; 42:333–337.
25. Hillis SD, Joesoef R, Marchbanks PA, et al. Delayed care of pelvic inflammatory disease as a risk factor for impaired fertility. Am J Obstet Gynecol 1993; 168(5):1503–1509.