Cervicitis is defined by the presence of mucopurulent discharge and/or cervical friability at the endocervical os (1, 2). Cervicitis is associated with HIV transmission (3) and is a potential precursor of upper genital tract infection, resulting in pelvic inflammatory disease (PID) and obstetric sequelae (4, 5). Identification and effective treatment of cervicitis is important to relieve symptoms and prevent sequelae.

Chlamydia trachomatis and Neisseria gonorrhoeae have traditionally been considered the most common causes of cervicitis. Other etiologic pathogens less commonly attributed to cervicitis include Trichomonas vaginalis and Herpes Simplex Virus (HSV) (2, 6). In the absence of these pathogens, Mycoplasma genitalium and the common vaginal dysbiosis, bacterial vaginosis (BV), are emerging as other infections associated with cervicitis. M. genitalium is associated with cervicitis by meta-analysis [pooled odds ratio (OR) = 1.66, 95%CI: 1.35–2.04] (7), and in a recent case-control study of >1,300 women in Melbourne, Australia, M. genitalium was associated with mucopurulent cervicitis on examination (adjusted OR = 4.38, 95% CI: 1.69–11.33, p = 0.002) (8). Limited studies have also shown that BV is associated with cervicitis (9–11). A secondary analysis of the same Australian case-control study, which utilised samples from 65 STI-negative cases with cervicitis and 128 STI-negative asymptomatic women, found that STI-negative cervicitis cases were five times more likely to have BV compared to STI-negative asymptomatic controls (12). Cases were also more likely than controls to have a Lactobacillus-deficient non-optimal microbiota and an increased abundance of four BV-associated bacteria (Gardnerella, Fannyhessea vaginae, Prevotella bivia, Dialister micraerophilus) (12).

Despite this evidence, the relative contribution of M. genitalium and BV to the burden of cervicitis is poorly understood, leading to inconsistency in testing recommendations across international guidelines (2, 13–15). The majority of guidelines list chlamydia and gonorrhoea as the main causes of cervicitis and recommend routine testing, but these infections are detected in <30%–40% of cases (2, 16–18), leaving a high proportion of cases with no clear cause (Supplementary Table 1).

We aimed to describe the proportion of cervicitis cases with C. trachomatis, N. gonorrhoeae, M. genitalium and/or BV, and to determine how commonly M. genitalium and BV were detected among women with non-chlamydial, non-gonococcal cervicitis in order to inform contemporary testing and management practices.

This retrospective case-series incorporated women with cervicitis who attended Melbourne Sexual Health Centre (MSHC) from 1st March 2011 to 1st March 2021. Ethical approval was provided by the Alfred Hospital Ethics Committee (ID: 718/20).

Women presenting with vaginal, pelvic and/or abdominal symptoms underwent a speculum and bimanual examination, STI testing, and microscopy of vaginal secretions. Cervicitis was defined as the presence of mucopurulent discharge at the cervical os and/or cervical ectopy or a friable cervix with easily induced bleeding (13). Clinical signs, symptoms, diagnoses, and test results were documented.

A total of 840 cervicitis diagnoses were recorded over the study period; 27 cases were excluded as individuals were ≤18 years of age (n = 11) or they had a repeat cervicitis diagnosis within 6 months (n = 16). The remaining 813 cases were from 805 attendees with at least one diagnosis of cervicitis >6 months apart. None of the identified cases were women living with HIV.

This retrospective study examined infections detected among women with cervicitis attending the largest public sexual health centre in Australia over a decade. This study was not designed to determine causality but to establish the proportion of cervicitis cases with C. trachomatis, N. gonorrhoeae, M. genitalium and/or BV, and to estimate how commonly M. genitalium and BV were detected among women with non-chlamydial, non-gonococcal cervicitis. As expected, a substantial proportion of cervicitis cases tested positive for C. trachomatis, with no significant change in the proportion of cases attributable to C. trachomatis over time. In contrast, N. gonorrhoeae was rarely detected among cervicitis cases. Collectively these two infections were only detected in 22% of cases tested for all four genital infections. In women with non-chlamydial non-gonococcal cervicitis, BV and M. genitalium were diagnosed in a third of cases, with BV the most common sole diagnosis made. Furthermore, despite no increase in testing, the number of BV diagnoses among cervicitis cases significantly increased throughout the study period. M. genitalium was the third most common infection found in cervicitis cases, and M. genitalium detection among cases and the clinic population both significantly increased over time from 3%–7% to 6%–15%, respectively. Although these data cannot establish causation, this study supports recent findings on the association between both BV (and BV-associated bacteria) and M. genitalium in case-control studies and informs contemporary clinical practice.

There are a limited number of studies that investigate the relationship between BV/BV-associated bacteria and cervicitis (9–12, 27, 28). The most recent evidence is from a case-control study of 65 STI-negative cases with cervicitis and 128 STI-negative asymptomatic women attending for STI-screening at MSHC. Cases with STI-negative cervicitis were 5 times more likely to have BV than STI-negative asymptomatic controls and in vaginal microbiota analyses, there was an increased abundance of specific BV-associated bacteria in cases compared to controls (12). This evidence builds on prior studies showing an association between BV (and BV-associated bacteria) and cervicitis. One US study reported that 67/423 (15%) BV cases had cervicitis, and of these, most (87%) had no other pathogen detected (11). In a small Seattle cohort, of which 14/210 had cervicitis, the BV-associated species, Mageeibacillus indolicus, was significantly associated with cervicitis, although this finding was not replicated among the comparator Kenyan cohort (28). Among a Chinese cohort, using qPCR Prevotella was more abundant in BV-positive cervicitis cases vs. people without cervicitis or BV (27). Emerging evidence also suggests that BV-associated bacteria account for a significant proportion of PID cases. One large prospective study of 2,956 women found BV was associated with incident PID, after adjusting for concurrent STIs (29). Two recent reviews reported that <50% of women with PID had chlamydial or gonococcal infections, while M. genitalium and detection of BV/BV-associated bacteria (Sneathia spp., Atopobium vaginae, Megasphaera spp.) accounted for a substantial fraction of PID cases (30, 31). In support of this, the addition of metronidazole to empiric PID treatment improved clinical outcomes for people with PID (32). Furthermore, the addition of metronidazole to cervicitis treatment improved cervicitis resolution, and people with BV were more likely to experience resolution (10). Clearly, there is a need for prospective research to improve our understanding of causal relationships between BV or specific BV-associated bacteria and the genital syndromes of cervicitis and PID, although current evidence supports consideration of BV in diagnostic recommendations.

As the third most common infection detected, M. genitalium positivity among cervicitis cases was 6% overall and increased over time, although this may reflect increased testing. Throughout the study period two different PCR assays were used with similar sensitivity, but lower sensitivity than a transcription-mediated amplification (TMA) assay (33). As M. genitalium is often a low load infection it is possible that positivity may have been under-reported, but even based on an additional 25%–30% of infections being detected by TMA, this would not increase the overall proportion of cases to more than 8%. A recent study at our service showed M. genitalium to be associated with a four-fold increased odds of cervicitis (8), in line with other publications (7, 34). Importantly however, M. genitalium was often co-detected with other infections, especially BV. BV may enhance an individual's risk of M. genitalium infection (35), although the importance of co-infections to an individual's susceptibility to cervicitis remains unknown.

Over the study period, C. trachomatis remained a significant contributor to cervicitis; it was detected in 1 in 5 cervicitis cases, which is within the previously reported range (11%–50%) (1, 36, 37), but greater than a similar Sydney-based population (6%) (38). Although the number of attendees to MSHC increased over time, the proportion of cervicitis cases with C. trachomatis was stable. This finding may be explained by a relative decrease in the contribution of C. trachomatis to cervicitis in comparison to other emerging infections, which has also been reported in relation to PID (30, 31). The low prevalence of N. gonorrhoeae was expected and is in line with prevalence in urban Australian populations (39). It is possible that increased screening for chlamydia and gonorrhoea, and practices such as expedited treatment, have reduced the risk of women developing more severe symptoms/syndromes including cervicitis. It is important to note that in some populations, N. gonorrhoeae positivity is higher and may have a greater contribution to cervicitis, highlighting the need for contemporary local data to inform regional guidelines.

We detected at least one of the four bacterial infections in half of cervicitis cases, which aligns with previous studies (1, 11, 36). While it is possible that a small number of the cervicitis cases without an infection identified had T. vaginalis or HSV, our limited data supports their selective testing based on clinical indications for our population. However, guidelines should consider regional risk data as these infections have higher prevalence estimates in other populations. Recent PID studies have reported three broad groups of aetiological organisms; established STIs (N. gonorrhoeae, C. trachomatis, M. genitalium, T. vaginalis), BV-associated bacteria (A. vaginae, Sneathia, Megasphaera), and organisms usually associated with the gastrointestinal or respiratory tracts (i.e., Bacteroides, Escherichia coli, Streptococcus, Haemophilus influenzae) (31). While PID is a distinct syndrome involving endometrial and upper genital tract infection, cervical infection commonly precedes PID, and it is logical to conclude that the spectrum of organisms involved is similar. The recent microbiota studies of women with cervicitis (12) and PID (30, 31) provide insights into likely infectious aetiologies, and prospective studies using whole metagenome sequencing will provide deeper and broader knowledge of the spectrum of infectious causes of cervicitis.

This study was conducted at the largest public Australian sexual health centre and accessed comprehensive clinical and meta-data. As this retrospective audit was of symptomatic cervicitis cases, these findings may not be generalisable to asymptomatic cases. However, the clinical importance of asymptomatic cervicitis is not clear. Longitudinal prospective studies that include asymptomatic women are challenging but may be needed to further interrogate causality. Additionally, our findings may not apply to other sites with differing population prevalence of infections.

In this study, BV and/or M. genitalium were detected in a third of non-chlamydial and non-gonococcal cervicitis cases, and BV was often the sole diagnosis made. M. genitalium was detected as the sole infection in 3% of cervicitis cases, and two-thirds of all cases with M. genitalum were co-infected with another infection. How co-infections/polymicrobial infections contribute to cervicitis requires further investigation. These findings build on recent case-control studies showing an association between BV (and BV-associated bacteria) and cervicitis and M. genitalium and cervicitis. Though prospective studies are needed to inform causality, these are challenging to undertake and unlikely to be rapidly forthcoming, as they require substantial funding and time. In the meantime, although chlamydia and gonorrhoea account for <30% of infections in cervicitis, testing remains universally recommended by global guidelines. In contrast, despite published data supporting an association, recommendations for testing for M. genitalium, and particularly BV, are inconsistent and uncommon outside of specialist sexual health services, potentially leading to delays in appropriate treatment and persistence of infection. Given that >60% of cases of cervicitis are not caused by C. trachomatis or N. gonorrhoeae in the majority of studies it is time to consider assessing women for BV and M. genitalium on presentation with cervicitis, as current empiric treatment may be inadequate.