The prevalence of Chlamydia trachomatis infection in Australia: a systematic review and meta-analysis

Chlamydia trachomatis (here after referred to as chlamydia) is the most commonly diagnosed bacterial sexually transmitted infection (STI) in Australia [1, 2]. In women, chlamydia can lead to serious and costly health consequences, particularly if the infection ascends from the endocervix to the upper genital tract and causes pelvic inflammatory disease (PID), which may result in fallopian tube scarring, ectopic pregnancy, tubal infertility and chronic pelvic pain [3‐6]. Chlamydia also causes epididymo-orchitis in men [7, 8] and can act as a co-factor in increasing the risk of HIV transmission in both men and women [9].

Genital chlamydial infection became a notifiable disease in 1991 in all Australian States and Territories except for NSW, which introduced mandatory notification in 1997 [1]. Notification rates have been steadily rising over the past decade [2], with rates highest among young people. Chlamydia notification rates are highest in the Northern Territory [2], which records high rates among Indigenous Australians [10]. Given that chlamydia is asymptomatic in up to 90% of infections, testing rates remain low (less than 10% in the younger age groups) [11], and re-infections are common, notification data greatly underestimate the true burden of infection. It is therefore important to gather high quality, region-specific epidemiological data to estimate the prevalence of chlamydia in Australian populations. A review published in 2005 reported an overall prevalence of chlamydia of 4.6% (95%CI: 4.4%, 4.8%) [12]; however, there has been considerable further chlamydia epidemiological research conducted in Australia since then. This review examines the available data on the prevalence of chlamydia across Australia and provides an up-to-date picture of the burden of chlamydia in Australian communities. This information will help to inform future clinical practice and screening policies.

A total of 76 studies, described in 78 published articles and six conference abstracts, from 129 articles and 48 abstracts fit the inclusion criteria (Figure 1; Table 1; Additional files 1, 2, 3, 4, and 5). The majority of studies (49 of 76) estimated chlamydia prevalence through cross-sectional surveys, with sample sizes of between 44 and 2817 participants (median 346). There were also 19 clinical audits (median sample size 505; range 80–26,097), three cohort studies (median 457; range 456–1642), three sentinel surveillance reports (median 30,516; range 3551–69,927), and two randomised controlled trials (RCTs; mean 1218; range 843–1593).

Where both male and female participants were included (30 studies), a substantial portion (16 studies) did not report these data separately, or failed to report by sex for some categories. These data were excluded from further meta-analysis.

Where reported, the most commonly utilised sample was first-void urine for both men and women; however, other samples included urethral, rectal and pharyngeal swabs for men; and endocervical swabs, self-obtained lower vaginal swabs, self-inserted tampons, and fallopian tube washings for women.

We found that chlamydia prevalence estimates were highly variable, with rates generally higher among young Australians, Indigenous Australians, and MSM. Other high-risk populations, including youth clinic attendees, pregnant teenagers and prison inmates, were also found to have higher prevalence estimates. This review builds considerably on the earlier review by Vajdic et al. [12], and a key strength is that studies were limited to those that used the more sensitive and specific NAAT testing methods.

Measuring the true prevalence of chlamydia in a community is a challenging task, with non-population-based studies frequently suffering from sampling bias and low participation rates [98, 99]. Studies utilising population-based data on chlamydia prevalence are rare, both in Australia and the UK [12, 100], and this review identified just one survey that recruited participants from the general population. This study, of women recruited from the telephone directory, was limited by its small sample size and low participation rate (43%) [16]. Both community-based and clinic-based sampling is subject to participation bias: community-based sampling often relies on convenience sampling, where participants are able to self-select; and clinic-based sampling is likely to bias towards symptomatic and higher risk participants, and towards more health care–seeking, and therefore well-educated, Australian-born participants [34]. This can also be true of population-based studies [16]. Sexual health centre studies can be particularly susceptible to bias towards symptomatic and high-risk groups, although as a research setting, they can provide large sample sizes at low costs, especially through clinic audits.

Community-based studies are important for estimating chlamydia prevalence as notification rates are known to under-estimate infection levels, as they are highly influenced by testing rates. The only population-based estimate identified in this review reported a prevalence of 3.1% among all 18 to 24 year old women in 2003–2004 [16]; no similar study has been conducted in men. This would translate into over 10,500 cases of chlamydia among Victorian women aged 15–24 years in 2004 (and 10,900 cases in men of the same age), which is considerably higher than the 7694 cases notified among both men and women of any age in Victoria that year [2], highlighting that notification data considerably underestimate of the true number of individuals infected.

A key factor that limits the conclusions that can be drawn from the studies included in the review is the high level of heterogeneity both within and between the populations studied. In most key populations reviewed, pooled prevalence estimates could not be calculated, even after stratifying data by age and sex of individuals tested. This was particularly true for female data. Although we calculated pooled prevalence estimates within the different population sub-groups, it is important that these be interpreted with caution and only indicative of the true prevalence.

The high level of variability between studies also prevented prevalence trends over time to be assessed, despite the review period spanning almost 15 years. Only seven individual studies determined time trends [18, 36, 43, 44, 48, 69], three finding a statistically significant increase in prevalence over time [36, 43, 48]; and two reporting no change over time [18, 44]. One study in an Indigenous community found a decrease over time, however, this was associated with an STI intervention to increase testing rates [69]. Despite the variability, the studies by Vodstrcil et al. [48] and O’Rourke et al. [43] – both conducted in sexual health clinics – provide evidence to support an increasing chlamydia prevalence among heterosexual men and women in Australia, because these studies had large sample sizes and adjusted for changes in sexual behaviour over time. Both Australian and international data show that sexual risk behaviour has changed over the last decade with increasing numbers of sex partners reported by young adults. The National Survey of Sexual Attitudes and Lifestyles, a sexual behaviour survey conducted in the United Kingdom in 1990 and again in 2000 [101] showed that the number of heterosexual partners in the preceding five years increased significantly for both sexes. Australian data show that age at first sexual intercourse has decreased, with women aged 16–24 years reporting a median age of 16 at first sexual intercourse compared with 19 for women aged 50–59 years [102].

Similar to previous systematic reviews in both the UK and Australia [12, 100], we found that the study setting influenced the prevalence estimates reported; however, significant heterogeneity again hampered comparisons in most cases. For men, prevalence was higher in sexual health and family planning clinics compared with GP and community-based settings. This was similar for women aged <25 years; however, this difference was not statistically significant. GP-based studies, similar to those conducted by Vodstrcil et al. [48] and O’Rourke et al. [43] in sexual health centres, would greatly enhance our ability to compare between these clinical settings. Between-setting comparisons are especially fraught for men; only one study conducted in general practice reported prevalence among men alone [32], and no studies reported chlamydia prevalence in heterosexual men from a population-based sample.

Where studies reported age-based estimates, younger participants had higher prevalence estimates than older participants. These data are consistent with the national notification data that show that notification rates are highest among those <25 years [2] and are also consistent with sexual behaviour data which show that numbers of sexual partners are highest in these younger age groups [103]. We also found high rates among disadvantaged youth and young people attending youth clinics. These findings of increased chlamydia among young men and women echo those findings in overseas prevalence studies [98, 99].

Chlamydia notification rates over the past 15 years have been consistently higher in women compared with men; however, this did not emerge as a robust trend in this review. Higher female notification rates can probably be attributed to differences in chlamydia testing rates. In Australia, recent Medicare data indicates that about 12–13% of sexually active young women and 3–4% of young men are tested for chlamydia each year [11]. As chlamydia testing rates increase in Australia, notification data will be able to provide a better estimate of the population prevalence of chlamydia.

We found that prevalence estimates were comparable among heterosexual men and women; however, the picture is neither complete nor consistent. In the general practice setting, no studies directly compare prevalence between men and women; and in sexual health clinics, prevalence tended to be higher among men. This is probably because men are more likely to attend a sexual health centre due to the presence of urethral symptoms [104]. Curiously, fifteen studies were identified that did not report male and female data separately, thereby excluding the data from calculations of pooled prevalence estimates. A number of recent studies reporting chlamydia prevalence in men attending sporting clubs [21, 22, 24, 25] and general practices [32] have started to address the predominance of female studies, which has been previously noted [12]; however, there remains a need for additional studies that directly compare men and women in community and clinical settings.

Similar discrepancies between notification data and population-based prevalence surveys have also been observed in the USA. In 2010, the notification rate reported to the Centers for Disease Control and Prevention was 2.6 times higher for women (610.6 per 100,000 population) than for men (233.7 per 100,000) [105]; rates in 2002 differed by almost four-fold between the sexes. By comparison, prevalence estimates reported in the National Health and Nutrition Examination Survey (NHANES) conducted between 1999–2002 were similar in women (2.5%; CI: 1.8, 3.4) and men (2.0%; CI: 1.6–2.5) [106]. Although the prevalence was twice as high in women aged 14–19 years (4.6%) compared to men (2.3%), the trend was reversed in the 20–29 year age group, where more men were infected (3.2%) than women (1.9%) [106]. These data underscore the fact that at low testing rates, notification data do not provide a full picture of the prevalence of chlamydia infection in the community.

In contrast to heterosexual men, several studies explored chlamydia prevalence among MSM, with most providing estimates from multiple anatomical sites (Additional file 4). In line with data from the UK and USA [107, 108], prevalence was highest in rectal swabs compared with urethral samples, and lowest in pharyngeal swabs. This highlights the importance of rectal chlamydia screening in MSM and the need to include both urethral and rectal sampling when conducting chlamydia prevalence surveys in this population group as recommended in national guidelines [109]. Unfortunately, Australian national notification data do not include site of infection nor sexual orientation, thereby reducing our ability to monitor trends in this population group over time.

The key gaps identified by Vajdic et al. [12] still remain today, including a need for population-based data for young men and women and systematically collected serial sentinel data with which trends in chlamydia prevalence over time can be monitored, particularly within the nationally identified target risk groups (young men and women, MSM and Indigenous Australians) [110]. The further advances in information technology including improvements in medical records software, the development of data extraction software [111] and data linkage [112], will facilitate the collection of standardised and detailed socio-demographic, behavioural and clinical data (including presence or absence of chlamydia-related symptoms) from sentinel sites. This will allow trends to be evaluated over time within different risk groups, adjusting for any changes in behavioural data and clinical presentation. The Australian Collaboration for Chlamydia Enhanced Sentinel Surveillance (ACCESS; http://​www.​access-study.​org) [44] and the Victorian Primary Care Network for Sentinel Surveillance (VPCNSS) [45] are both important sentinel surveillance projects that collect detailed demographic, clinical and behavioural data with which trends in risk groups can be monitored over time. It is vital that such surveillance systems continue to be funded. Further, the Australian Chlamydia Control Effectiveness Pilot (ACCEPt; http://​www.​accept.​org.​au), a randomised controlled trial of chlamydia testing in general practice, is collecting chlamydia testing data from about 250 GP clinics across Victoria, New South Wales, Queensland and South Australia and will provide trends in chlamydia positivity over time among young men and women.

This comprehensive systematic review identified 76 studies reporting prevalence data for individuals tested for anogenital or pharyngeal chlamydia and provides an up-to-date summary of the underlying burden of chlamydia in Australian populations. The review highlights that the burden of chlamydia in Australia is greatest among young adults, Indigenous populations and MSM and identifies important gaps in the surveillance and monitoring of chlamydia infection in Australia. Given that that the Australian Government is currently pilot testing chlamydia screening as a national program and State Governments continue to fund chlamydia control activities, it is vital that good sentinel surveillance systems continue.