METHODS

Women attending two metropolitan FPV sexual and reproductive health clinics with contrasting populations in Melbourne were studied. The first was an inner city clinic with a predominately adolescent population (93% female; 80% under 25 years; high proportion from socioeconomically disadvantaged areas). The second was a suburban clinic in a more affluent district with a wider age distribution (95% female; 35% under 25 years). The study was approved by the FPV ethics committee.

All women attending the clinics during a 5 week period in 2001 were asked to participate. Informed consent was obtained. Practitioners used a structured questionnaire to collect demographic and sexual history details from participants. These included age, genitourinary symptoms (vaginal discharge, dysuria, lower abdominal pain, intermenstrual bleeding, post-coital bleeding, and dyspareunia), sexual history details, barrier contraception use, time of last voiding, and reason for attendance.

A first pass urine was stored and transported at 4°C, and tested within 96 hours. Detection of C trachomatis was undertaken by polymerase chain reaction (PCR) using Cobas Amplicor (Roche Molecular Diagnostics) using an internal co-amplified control target. Samples with controls testing negative twice were designated unassessable. Positive tests were confirmed by ligase chain reaction (LCR) using Abbott LCx (Abbott Laboratories).

Women positive for C trachomatis were treated and screened for gonorrhoea, trichomonas, syphilis, hepatitis B, and HIV, and had contacts traced.

Analysis and statistics

Analysis was undertaken using stata v7.0. Associations between each risk factor and infection status were assessed using Pearson’s χ² test. Odds ratios (OR) with 95% confidence intervals (CI) were calculated. Variables that showed any evidence of association with infection on univariate analysis (p value <0.1) were included in a multivariate logistic regression model.

RESULTS

Participants

Of 1107 women attending during the study, 866 (78.2%) participated (376 (82.6%) of 455 in the inner city and 490 (75.2%) of 652 in the suburban clinic). The mean (range) ages of the women participating in the inner city and suburban clinics were 21.3 (16.3–38.3) and 23.0 (13.4–62.3) years respectively. The age distribution at the two clinics is shown in figure 1 (on the STI website) and the reasons for attendance in table 1.

Non-participants in the inner city clinic were more likely to be under 25 years (p=0.18), and less likely to have vaginal discharge (p=0.08), intermenstrual bleeding (p=0.03), dyspareunia (p=0.033), any symptom (p<0.005), use barrier contraception (p=0.03), or have had a recent change of partner (p=0.002). Non-participants in the suburban clinic were less likely to have lower abdominal pain (p=0.005), dyspareunia (p=0.013), any symptom (p=0.008), or use barrier contraception (p=0.034).

Prevalence

Of the 866 women tested, the urine was assessable in 851 (98.3%)—373 (99.2%) in the inner city and 478 (97.6%) in the suburban clinic. C trachomatis was detected in 18 (4.8% (95% CI 2.9 to 7.5)) of 373 women in the inner city and eight (1.7% (95% CI (0.7 to 3.3)) of 478 women in the suburban clinic. Of women under 25 years, 17 (6.2% (95% CI 3.7 to 9.8)) of 273 in the inner city in contrast with three (1.7% (95% CI 0.4 to 5.0)) of 174 in the suburban clinic were infected with C trachomatis. The prevalence in women overall and that in women under 25 years in the two clinics was significantly different (p<0.05).

Symptoms and risk factors

Infection was asymptomatic in five (29%) infected women in the inner city and three (38%) women in the suburban clinic (table 2). In the inner city clinic, age under 25 years (OR 5.4 (95% CI 0.7 to 41.5)), vaginal discharge (OR 4.1 (95% CI 1.5 to 11.1)), and recent change of sexual partner (OR 4.6 (95% CI 1.6 to 12.9)) were associated with C trachomatis (table 2). Multivariate analysis of data from the inner city clinic showed that recent change of partner (OR 4.5 (95% CI 1.5 to 13.8)) was the most strongly associated independent risk factor for infection (table 2). In contrast, in the suburban clinic, only vaginal discharge (OR 3.5 (95% CI 0.9 to 14.3)) and recent change of sexual partner (OR 3.4 (95% CI 0.8 to 15.7)) were identified as risk factors (table 2).

Other sexually transmitted infections

Two (11%) of the 18 women with C trachomatis were co-infected with Neisseria gonorrhoea (positive urine PCR and endocervical swab culture). No other sexually transmitted infections were detected.

DISCUSSION

Prevalence

The prevalence of C trachomatis is above the 2.1% threshold identified for universal testing in the inner city clinic but not the suburban clinic.^14–17 One possible explanation for the lower prevalence in the latter is that women attending the inner city clinic were significantly more likely to be attending for emergency contraception and pregnancy tests and therefore potentially more likely to be having unsafe sex. Against this is the absence of an association in this study between non-barrier contraception and C trachomatis.

Influence of non-participants

It is unclear whether non-participants were more or less likely than participants to have been positive for C trachomatis and consequently their influence on the true prevalence. Even in the unlikely event that all the non-participants were uninfected, the minimum overall prevalence would remain high at 4.0% (18/455). Similarly in the suburban clinic, the prevalence would not alter significantly if all non-participants were uninfected (1.2% (8/652)).

Previous studies in Australia

A study in 1988 in women presenting to FPV for a pelvic examination found a prevalence of 5.1%.¹⁸ However this study is not comparable with the current study as C trachomatis was detected by direct immunofluorescence and culture. In addition, the selected clients in the earlier study were likely to have included a higher proportion of infected women and therefore overestimated the true prevalence. The prevalence in unselected women in the current study is therefore consistent with the increase in C trachomatis notifications in Victoria.⁶

Other studies in Australian women have shown variable prevalences for C trachomatis from 2.8%¹⁹ to more than 5% in remote areas.^20,21 A study in women in an urban sexual health centre in Sydney showed a rise in prevalence from 1.8% to 3.5% between 1994 and 2000.²²

Previous studies overseas

Wide ranging prevalence rates have been documented in other countries, including studies in family planning clinics.^23–27 A recent study in young women in the United Kingdom, using urine PCR, showed a prevalence of 10%.²⁸ Higher rates have been recorded in some populations in the United States including family planning clinics²⁵ and sexually transmitted diseases clinics.²⁹ Rates as high as 27% have been documented in adolescents³⁰ and women in prison.³¹

Age as risk factor

In Victoria, 72% of notified C trachomatis infections are in individuals under 29 years and 48% are in those under 25.⁶ The highest prevalence is in women between 20 and 29 years.¹¹ In this study, in the inner city clinic, women under 25 years were more than five times more likely to be infected, consistent with studies in other populations.¹⁰

In contrast, women under 25 years in the suburban clinic were not at increased risk. In this clinic, five of the eight infected women were over 25 years. There are three main differences in the two clinic populations that may have contributed to this difference: the suburban clinic has a lower prevalence of C trachomatis, a lower proportion of younger women, and a lower proportion of women from socioeconomically disadvantaged areas. In contrast with other sexually transmitted infections, C trachomatis has not been shown to be associated with socioeconomic factors independent of race and ethnicity.³² The lack of an increased risk of infection in young women in the suburban clinic in our study is in accordance with other studies that have shown age to be a poor predictor of infection in populations with high affluence³³ and/or lower prevalence rates.^16,17

Symptoms and partners as risk factors

C trachomatis infection was symptomatic in an unexpectedly high proportion of women. Vaginal discharge was significantly associated with C trachomatis in the inner city clinic where 10 (59%) of the infected women had vaginal discharge. In contrast, in the suburban clinic, only one (13%) of the infected women had this symptom. Previous studies have found between 70 and 90% of C trachomatis infection is asymptomatic.¹ The high proportion of infected women with symptoms in our study may be explained by selection bias as a result of symptomatic women being more likely to attend the clinics in this study. In the inner city clinic, recent change of partner was also found to be significantly associated with C trachomatis infection, consistent with other studies.^34,35

Logistic regression using the three variables most strongly associated with infection in the inner city clinic as predictors showed that recent change of partner was the strongest independent risk factor. In contrast, multiple regression analysis produced a moderate reduction in the adjusted OR for both age under 25 years and vaginal discharge suggesting that at least some of the crude association between these two variables and infection was due to confounding as a result of an association between these risk factors and recent change of partner.

Despite infection being symptomatic in an unexpectedly high proportion of women, no single symptom or combination of symptoms offers sufficient sensitivity and specificity to be useful as criteria for selective screening.

CONCLUSION

This study highlights the importance of undertaking local studies of prevalence and risk factors. It shows that identifiable risk factors for C trachomatis are not transferable between populations even in clinics run by the same organisation in the same city. It also suggests that well established risk factors such as age under 25 years may not be applicable in more affluent populations with lower prevalence rates.