Has the Effectiveness of a Single Chlamydia Test in Preventing Pelvic Inflammatory Disease over 12 Months Been Overestimated?

“…147 women need to be screened to prevent one case of PID. If screened through the National Chalmydia Screening Programme (NCSP) this would cost at least UK£4851 per case of PID prevented … Is this level of cost justified?”

Chlamydia trachomatis is found in approximately 30% of women with pelvic inflammatory disease (PID) and is strongly implicated in the complications that ensue: tubal factor infertility, ectopic pregnancy and chronic pelvic pain [1–4]. It is the most common bacterial sexually transmitted infection in the USA [5] and the UK [101]. To reduce the incidence of asymptomatic case finding associated with these complications, screening at population level has been advocated in many countries. Benefit from screening can arise from interrupting chlamydial infection before PID develops, and/or from reducing the population prevalence [6].

The USA Preventive Services Task Force recommends screening for chlamydial infection for all sexually active, nonpregnant young women aged 24 years or younger and older nonpregnant women who are at increased risk, at least annually [7]. There is recognition that additional research is needed in order to determine the most effective intervals for screening nonpregnant women. A national screening program has been established in the UK, using opportunistic screening of men and women aged below 25 years. When it was launched there was anticipation that screening would be an effective intervention to reduce the burden of disease and would benefit both the person screened and the population. This was summarized by the UK Chief Medical Officer's Expert Advisory Group on Chlamydia trachomatis [102]. They concluded that screening for C. trachomatis satisfies the Wilson and Junger criteria [8], a commonly used test to determine whether screening should be advocated as a public health procedure and concluded that, ‘There is now sufficient evidence, in the published scientific literature, to show that control and prevention programs for genital chlamydial infection that are based on screening for infection amongst sections of the sexually active population give rise to significant health gain and are cost beneficial’.

These recommendations are based predominantly on two studies of a single chlamydia test with outcomes assessed 1 year later. The studies have been reviewed elsewhere and inevitably have some limitations [7]. Scholes and colleagues used cluster randomization of healthcare facilities within an health maintenance organization to allocate women to screening or control groups [9]. The authors identified 2607 women at high risk, based on risk factors including age less than 25 years, ethnicity and two or more sexual partners in the previous 12 months to be offered screening. They found a relative risk of PID of 0.44 (0.20–0.90) for the intervention group. As other researchers have observed, this appears to be fortuitous since chlamydia is only identified in 30% women with PID, and more than a third of those in the intervention group were not tested [7]. It is possible that other organisms that can cause PID, such as Neisseria gonorrhoeae or Mycoplasma genitalium, might also be eradicated by the antibiotics prescribed for chlamydia. In Denmark, Ostergaard and colleagues used cluster randomization of high schools to recruit 1700 sexually active female students [10]. They were offered home sampling for chlamydia and routine care from a GP or sexually transmitted infection clinic. Only 63 of 833 (8%) in the control group were tested, compared with 867 of 928 (93%) in the intervention arm. Men in the same schools were also offered screening. Treatment for PID was reported in 2.1% of those in the home sampling group and 4.2% in the usual care group 1 year later. Nearly 50% of the subjects were lost to follow-up. Since it is likely that students in the same schools were in sexual networks it is difficult to know how much impact the screening of male students had.

There is now growing uncertainty about the scientific validity of chlamydia screening. In June 2010 the Journal of Infectious Diseases dedicated a complete supplement to the discussion of this uncertainty. It concluded that many questions remained unanswered. Publication of the Prevention of Pelvic Infection (POPI) study from London has clarified some of the issues [11]. The key question that the POPI study was designed to answer was: does Chlamydia screening reduce the risk of PID? This randomized controlled trial enrolled 2529 sexually active female students with a mean age of 21 years to either: immediate chlamydia screening and treatment if positive (n = 1191); or collecting a screening specimen but not testing it for 1 year (n = 1186). Oakeshott and colleagues obtained outcome data after 1 year from 94% (n = 2377) of the participants. Questionnaires were sent by post and emails linked to an online questionnaire after 1 year. The diagnosis of PID was made by three genitourinary medicine physicians, who were blinded to the randomization and initial chlamydia results, and assessed symptom questionnaires. If the questionnaires indicated the possibility of PID or there was no response from the participant, GP and hospital records were obtained.

“If screening intervals are shortened, more incident cases will be detected and more chlamydial PID could be prevented…”

The baseline prevalence of chlamydia was 5.4% (68/1254) in screened women and 5.9% (75/1265) in controls. The study found no statistically significant protection from PID. The incidence of PID was 1.3% (15/1191) in screened women and 1.9% (23/1186) in controls (relative risk: 0.65; 95% CI: 0.34–1.22).

A total of seven out of 74 control women (9.5%; 95% CI: 4.7–18.3) who tested positive for chlamydial infection at baseline developed PID over 12 months compared with one out of 63 (1.6%) screened women (relative risk: 0.17; 95% CI: 0.03–1.01).

Most episodes of PID occurred in women who tested negative for chlamydia at baseline (79%; 30/38). A relatively high proportion of women, (22%; 527/2377), reported being tested independently for chlamydia during the trial: 15% (n = 38) of the tested control women, and 10% (n = 27) of the tested screened women reported a positive test. A total of 26 of the 38 women diagnosed with PID had a chlamydia test outside of the trial and 16 (61%) of these were positive, although some of these occurred at the time of PID diagnosis.

All study participants were encouraged to test for chlamydia. Women in the control group were blind to their randomization allocation and baseline chlamydia status. Interestingly, those who were positive for chlamydia at baseline were more likely than those who were negative to report having been tested independently (43%; 29/67 vs 24%; 229/968, respectively; p < 0.001). The authors could not determine the date of testing for chlamydia with sufficient reliability to perform a secondary analysis to compensate for such testing outside the study protocol.

The risk of clinical PID over 12 months in women screened for chlamydia was nonsignificantly reduced by 35%. The overall incidence of PID was, however, low (1.6%). In 137 women with chlamydial infection at baseline, 9.5% in the deferred screening control group developed PID compared with only 1.6% in the screened group. Over 90% (67/74) of control women with chlamydial infection at baseline did not develop clinical PID, and most cases (79%; 30/38) of PID, including ten cases of chlamydia-positive PID, occurred in women who were negative for chlamydia at baseline, suggesting these were incident infections.

The authors estimated that it was necessary to screen 147 women for chlamydia or to treat 13 women who tested positive for chlamydia to prevent one case of PID over 12 months. This is considerably higher than Scholes' findings, which suggest that treating four infected women will prevent one case of symptomatic PID [12]. Ostergaard used a different definition of PID, but had similar success: 867 female students were tested, 43 infections detected and nine reported cases of acute PID among the 443 interviewed at follow-up were prevented [10]. Why are the results in POPI so different? The rate of diagnosed PID was lower than in the previous studies. The high proportion of women in the control group who were tested for chlamydia might have reduced the relative efficacy of the intervention, but it is unlikely to account for such a difference with the earlier studies. Differences could be due to different designs and populations. The diagnostic tests are different: Scholes used an ELISA test, which is generally only 60% sensitive compared with the Neucleic Acid Amplification Test (GenProbe transcription mediated amplification) used by Oakeshott and Ostergaard. As immunity to chlamydia develops, the chlamydial shedding from the cervix decreases and women who screen positive with ELISA are more likely to have a new chlamydia infection than those who screen positive with a NAAT. Scholes therefore probably missed chlamydia infections with a low level of shedding from the cervix (old chlamydia), who might be at lower risk of progression to PID than those with higher numbers of organisms.

Whatever the reason, the POPI study, which is stronger methodologically, suggests that a single test for chlamydia is less likely to prevent PID than was previously thought [13]. Most cases of PID over 12 months were not prevented by a single chlamydia screen and occurred in women who were negative for chlamydia at baseline. If PID follows soon after incident infection [14], then the latent period (i.e., the time between infection and symptomatic disease) in chlamydia may be much shorter than previously thought.

If screening intervals are shortened, more incident cases will be detected and more chlamydial PID could be prevented; however costs of screening would increase for the health service and the screening participant. The costs of preventing PID through chlamydia screening are already very high – as previously commented: ‘147 women need to be screened to prevent one case of PID. If screened through the National Chalmydia Screening Programme (NCSP) this would cost at least UK£4851 per case of PID prevented’ [13]. Is this level of cost justified?

Policymakers will have to learn from the history of chlamydia screening. In 1998 experts were certain of the scientific basis for chlamydia screening. We now know that this was premature. However, it would also be premature to stop chlamydia screening. Instead, we could focus on more frequent testing of those at higher risk, such as women with a new sexual partner or a recent history of chlamydial infection, or screening for more than one sexual health problem at the same time. Such a screen could include tests for gonorrhoea, chlamydia and M. genitalium; as well as questions about unmet contraceptive needs. Whatever the programmatic response to the POPI study will be, it should not be pursued without rigorous research and evaluation.

In addition, since culture has been virtually abandoned for diagnosis we have no systematic monitoring for antibiotic resistance. If azithromycin is used routinely, and macrolide resistance arises, the effectiveness of our intervention would be in jeopardy.