Urethritis is the most common genitourinary complaint amongst men. In women, cervicitis corresponds to the sexually transmitted urethritis in men although the standardised definition of cervicitis has not been as widely adopted.1 2 Chlamydia trachomatis infection is the most common cause of non-gonococcal urethritis (NGU)/cervicitis, although it accounts for less than half of cases.3 There is now good evidence that Mycoplasma genitalium is a sexually transmitted pathogen and it has been shown to be a cause of non-chlamydial NGU in men.4^–8 Unlike association studies in men, few studies have established a firm connection between M genitalium infection in women and signs of cervicitis.9^–11 For example, in Falk et al’s study,11 M genitalium infection was found to be associated with microscopic signs of urethral inflammation when comparing symptomatic patients from a sexually transmitted infections (STIs) clinic and an asymptomatic control group from a cervical screening programme. However, M genitalium infection was not associated with microscopic signs of urethral inflammation when comparing symptomatic patients with asymptomatic control patients from the same STI clinic. Few such studies have been published to date and the reports have been conflicting (see Discussion for further examples). This is probably due to practicalities in clinical settings in obtaining samples from relevant control groups, more complicated female urogenital anatomy, complications with female menstruation cycle as well as obtaining samples from all the relevant organs, such as the urethra, the cervix and the vagina. The aim of this study was to analyse clinical and laboratory data from a large number of female patients attending an STI clinic and to determine if there is an association between infection with M genitalium with objectively quantified microscopic signs of lower genital tract inflammation (LGTI).

MATERIALS AND METHODS

Study population and clinical methods

Data from all female patients that had voluntarily attended a drop-in STI clinic in Oslo between November 2005 and December 2007, who were examined by a physician and tested for both M genitalium and C trachomatis, were analysed. Referral to a physician was based on the following criteria made by a consulting triage nurse: discharge, dysuria, pelvic pain, any other genital symptoms (for example, itch, rash, blisters or ulcers, warts), those who had many partners, women with inconsistent condom use and women who were traced to a partner with a STI. Patients with Neisseria gonorrhoea (n = 13) and follow up visits were excluded from analysis. From 7646 patients, 7445 patients were tested for both microorganisms in both first void urine (FVU) and cervical swabs, 102 were tested for microorganisms in FVU alone and 99 in cervical swab alone. Although records of microscopic examination of urethral smears were available for all 7646 patients, records of microscopic examination of cervical smears were available for only 7472 patients.

In the clinic, a urethral smear was taken with a sterile blunt curette and a cervical specimen with a cotton swab after cleaning the portio uteri. Both were stained with methylene blue and examined microscopically by a physician. In addition, wet smears of vaginal discharge were taken with a 10 μl plastic loop, mixed with saline solution and 10% potassium hydroxide and examined microscopically. Vaginal specimens were also tested for vaginal pH with a paper strip and a whiff test of the potassium hydroxide smear was performed. Overall, 27 clinicians were involved in the examination of these patients. Female patients were diagnosed clinically with non-gonococcal LGTI for treatment if they had at least three of the following four criteria: >4 polymorphic mononuclear leucocytes per high-powered field (PMNLs/HPF) in the absence of typical intracellular diplococci in the urethral smear, >30 PMNLs/HPF in the cervical smear, more leucocytes than epithelial cells in the saline wet smear and clinical symptoms of cervicitis (yellow discharge, fragile portio uteri). For diagnosis of bacterial vaginosis, the Amsel criteria were used.12 For statistical analysis of association in this study, patients were diagnosed with normal, borderline and moderate to severe signs of urethritis according to microscopic examination of urethral smears (0–4, 5–9 and >10 PMNLs/HPF, respectively).

Laboratory methods

For the detection of C trachomatis and M genitalium, DNA was isolated from 200 μl of FVU or 200 μl of medium from the cervical swab (Roche, Indianapolis, Indiana, USA) using MagNA pure LC DNA isolation Kit I (Roche) and eluted into 100 μl of elution buffer. Of the DNA preparation, 25 μl was used for detection of C trachomatis using Cobas Amplicor (Roche) or Cobas Taqman 48 (Roche); 10 μl of the same DNA preparation was used for the detection of M genitalium by real-time PCR in a volume of 25 μl using primers and probes described by Jensen13 and using the 7900HT instrument (Applied Biosystems, Foster City, California, USA). Non-competitive primers and probes to an internal control template that was spiked into PCR master-mixes were used to assess inhibition in PCR reactions and to repeat tests if necessary.

Statistical analysis

Descriptive statistics were used to assess the characterisation of study participants, prevalence of microorganism infection and clinical measurements using SPSS (v. 16) and Microsoft Excel. When calculating prevalence of M genitalium for association analysis, patients that were positive for C trachomatis were excluded from the analysis. Likewise, when calculating prevalence of C trachomatis for association analysis, patients that were positive for M genitalium positive were excluded. Measures of association by odds ratio (OR) between prevalence of microorganisms and patient characteristics and clinical measurements and 95% confidence intervals (CIs) were performed using the χ2 test.

RESULTS

A total of 7646 women were tested for M genitalium and C trachomatis in FVU or in cervical swab or in both. Of these, 304 (4%) were infected with M genitalium, 726 (9.5%) were infected with C trachomatis and 42 patients (0.6%) were infected with both microorganisms. Patients co-infected with M genitalium and C trachomatis were excluded from further analysis. The mean age (SD) of patients positive for M genitalium was 25.5 (SD 5.7), similar to patients positive for C trachomatis (25, SD 5.2) and to patients without any infection (26.8, SD 6.5). Frequency of patients reporting symptoms of LGTI (discharge or dysuria) was the same in patients positive for M genitalium or C trachomatis as well as in patients without microorganism infection (60–64%).

In analysing patients where results from both FVU and cervical swab specimens were available, we found FVU to be the less sensitive specimen in identifying patients positive for M genitalium (FVU 62% vs cervical swabs 86%; table 1). FVU and cervical swabs were both significantly more sensitive in identifying patients positive for C trachomatis compared with M genitalium (88% vs 90%; table 1).

Patients were analysed with low to severe signs of inflammation of the urethra according to microscopic examination of urethral smears (table 2). Microscopy results of PMNLs/HPF from the urethra were divided into three stages: 0–4 (normal), 5–9 (borderline) and ⩾10 (moderate to severe) in at least five consecutive HPFs. Table 2 shows that women with M genitalium detected in the FVU had a higher probability of having moderate to severe inflammation of the urethra compared with women who were free of M genitalium and C trachomatis in FVU (34% vs 22%; p<0.001). This association was less significant with M genitalium detected in cervical swabs (27% vs 22%; p<0.05). Compared with 22% of women without M genitalium and C trachomatis infection, 43% of women who tested positive for C trachomatis in FVU samples (p<0.001) and 42% of women who tested positive for C trachomatis in cervical swabs (p<0.001) had moderate to severe urethritis (table 2).

Altogether, 68% of patients with C trachomatis detected in FVU and/or in cervical swabs had >30 PMNLs/HPF in cervical smears (p<0.001; table 3) compared with 45% of women where microorganisms were not detected,. This difference in prevalence was considerably less pronounced in patients where M genitalium was detected in FVU (55%, p<0.05) and not significant in patients where M genitalium was detected in cervical swabs (51%; p>0.05; table 3).

Malodour was found more frequently in patients infected with C trachomatis or M genitalium compared with patients where microorganisms were not detected and independent of whether the microorganisms were detected in FVU or in cervical swabs (31–34% vs 24%; p<0.05; table 4). Bacterial vaginosis was marginally lower in patients without microorganism infection (23%) compared with patients with C trachomatis (29%; OR 1.4; 95% CI 1.2 to 1.6), and not significantly higher in patients with M genitalium infection (28%; OR 1.3; 95% CI 1.0 to 1.7). As with the association with malodour, association with bacterial vaginosis was independent of detecting microorganisms in FVU or in cervical swabs.

DISCUSSION

Although the results of studies of men with urethritis and M genitalium infection has been reproducible and conclusive across different studies,4 6–8 10 14–16 the role of M genitalium in causing NGU or cervicitis in women has been contradictory. For example, Manhart et al9 have found a significant association between M genitalium infection and mucopurulent cervicitis in 50 patients as determined by cervical and vaginal PMNL counts, although they reported a non-compounding independent factor of 28% coinfection with C trachomatis and/or N gonorrhoea. Uno et al17 have also found an association between detecting M genitalium in cervical swabs in symptomatic non-pregnant women with cervicitis and non-symptomatic pregnant women. Korte et al18 have found an association between detecting M genitalium infection in cervical and vaginal swabs with increased genitourinary symptoms but no association was found with cervicitis. Tosh et al19 found no association between M genitalium and the presence of clinical symptoms or signs of inflammation. Huppert et al20 found no association between M genitalium infection, vaginal symptoms, physical evidence of cervicitis or findings on wet mount or Gram stain. In Anagrius’s study,10 M genitalium was found as frequently as C trachomatis in the study population and more frequently in female patients with symptoms or signs of urethritis or cervicitis. M genitalium infection, but not C trachomatis, was associated with microscopic signs of urethritis and/or cervicitis; however, symptoms regardless of microscopic signs were not associated with M genitalium infection. In a study of young people in the general population in Denmark, urogenital symptoms was not associated with being infected with M genitalium.21 Pepin et al22 report M genitalium to be highly prevalent in West African sex workers with a weak association with symptoms of cervicitis.

Our data highlight some of the potential pitfalls in studies of M genitalium in female patients with genitourinary complaints and points to source(s) of contradiction. First, M genitalium is not as frequently detected as C trachomatis and, therefore, for meaningful statistics one needs to look at a large number of patients. In this study, we have analysed data from 7646 consecutively attending patients. Analysis of a large number of patients should also reduce the potential effect of inter-observer bias in clinical examination of patients. Second, most studies have not included analysis of association of M genitalium infection with signs of inflammation in urethral smears, probably because microscopic examination of urethral smears is not carried out commonly outside of Scandinavia and female dysuria is often referred to as urethral syndrome. Third, we found that M genitalium was most often detected in cervical swabs compared with FVU, although the most significant association of M genitalium infection with urethritis was found when the microorganism was detected in FVU. Therefore, even though cervical swabs may be a better specimen for finding patients positive for M genitalium, it is the results from FVU (presumably reflecting infection of urethral epithelium) that will demonstrate an association with urethritis. In most studies, M genitalium infection has been tested for in vaginal or cervical specimens as they have given higher sensitivity than urine.9–11 17–19 21 23 Difference between detecting M genitalium in urine and cervical swabs may reflect two different pathogen–host interaction mechanisms depending on whether the site of infection is the urethral epithelial cells or if it is the primary cervical epithelial cells. It may, therefore, be informative to follow-up this study with an assessment of M genitalium in urethral swabs. Fourth, clinical diagnosis of cervicitis in women is more difficult than diagnosis of urethritis in men. Depending on the study, the criterion for cervicitis has been variably defined as presence of mucopurulent discharge, friability or high PMNLs/HPF in cervical smears. In our patient group, 43% of patients without symptoms of LGTI had >30 PMNLs/HPF in cervical smears and 47% of patients with symptoms had >30 PMNLs/HPF in cervical smears. The correlation of symptoms to signs of urethritis was a lot more clear when correlated to microscopic examination of urethral smears: 19% of patients without symptoms had ⩾10 PMNLs/HPF while 25% of patients with symptoms had ⩾10 PMNLs/HPF in urethral smears (OR 1.4; 95% CI 1.3 to 1.6), although even this correlation was not striking. Other research studies have also failed to find a correlation between the presence of microscopic signs of urethritis or cervicitis and their corresponding clinical symptoms.10 18 20 24 This supports previous observations that bacterial STIs in women are asymptomatic in a high proportion of cases.24

In our study, M genitalium was tested for in the same DNA preparation as that used for detecting C trachomatis. Therefore, C trachomatis acts as a control for the analysis of the data and associations, although C trachomatis is more prevalent and hence allows for more conclusive results from statistical analysis. With the caveat of detecting C trachomatis more often than M genitalium, clear differences and similarities between laboratory and clinical findings of these two organisms is apparent. First, the sensitivity of detecting C trachomatis in FVU alone or cervix alone is much higher than detecting M genitalium (88% and 91% vs 60% and 85%, respectively; table 1). Second, even though M genitalium is more readily detected in the cervix compared with FVU, M genitalium detection in FVU shows a clearer association with signs of inflammation in urethral smears (⩾10 PMNLs/HPF) and in cervical smears (>30 PMNLs/HPF). In fact, we found that in patients where M genitalium was detected in FVU and not in the cervix (n = 40), a higher percentage had moderate to severe signs of urethritis (32.5%) compared with patients where M genitalium was detected in the cervical swab and not in FVU (17.5%), although the numbers were relatively small and not highly significant (n = 114; OR 2.3; 95% CI 1.0 to 5.0). We also found that malodour and bacterial vaginosis to be more common in patients with C trachomatis and M genitalium compared with patients without microorganisms infection, although the association of bacterial vaginosis and M genitalium was not significant (p = 0.05). Falk et al11 have similarly reported bacterial vaginosis to be more common in patients with C trachomatis compared with patients with M genitalium. We also noted other similarities between C trachomatis and M genitalium in our data. Prevalence of both M genitalium and C trachomatis increased incrementally with the number of partners in the last 6 months. Furthermore, there was a negative association with microorganism infection if the patient was living with her partner or if the patient reported using condoms. There was an absence of association of infection with either microorganism with level of education or employment status.

In conclusion, we found a significant association between detecting M genitalium in FVU with signs of urethritis in women attending an STI clinic. This association was less apparent if M genitalium was associated with detection in cervical swabs rather than FVU, implying a difference in clinical manifestation depending on the site of infection in the genitourinary tract.