Article Text

Download PDFPDF
The POPI trial: what does it mean for chlamydia control now?
  1. Nicola Low1,
  2. Jane Hocking2
  1. 1Institute of Social and Preventive Medicine, University of Bern, Bern, Switzerland
  2. 2Centre for Women's Health, Gender and Society, University of Melbourne, Melbourne, Australia
  1. Correspondence to Professor Nicola Low, Institute of Social and Preventive Medicine, University of Bern, Finkenhubelweg 11, Bern CH-3012, Switzerland; low{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

The results of the Prevention Of Pelvic Infection (POPI) trial,1 published soon after criticism about the value for money of the chlamydia screening programme in England,2 should make us think even harder about how to investigate and control chlamydia and its associated morbidity. Should we be doing more of the same, but better? Should we be developing new approaches? Or do we need both?

The innovative trial by Oakeshott et al in female college students in London provides new information about the effects on pelvic inflammatory disease (PID) of a single round of screening for Chlamydia trachomatis.1 By collecting specimens from all participants at baseline, Oakeshott and colleagues have provided the first data comparing the incidence of PID according to chlamydia infection status in women receiving immediate testing and treatment with those for whom testing and treatment were deferred by 12 months. The ethical decision to allow deferred testing is justified by the uncertainty about the incidence of PID in the community before the start of the trial, the low observed incidence in practice and the advice for all study participants about the availability of free chlamydia testing. The investigators randomised and followed 2529 women (mean age 20.9 years) and collected outcome data on 2377 (94%) from the women or their GPs. They obtained information from medical records of suspected PID cases and two specialists (with a third as arbiter) conducted a blinded assessment against predefined criteria.

The results of the primary trial outcome were inconclusive because the study was underpowered. There were 38 cases of probable or possible PID, 1.3% (15/1191) in the intervention group and 1.9% (23/1186) in the control group (RR 0.65, 95% CI 0.34 to 1.22). This uncertainty could have been avoided. The originally planned sample size was based on an expected PID incidence rate of 2% in the control group,3 close to what was observed. Owing to poor recruitment, the investigators revised the sample size calculation based on a higher expected incidence of PID in the control group (3%). This was, however, estimated from a study of women with a much higher risk of PID because of high levels of baseline chlamydia and previous PID.4

Interpreting the new data about the incidence of PID is not straightforward, partly because of the small numbers of events. Understanding patient- and physician-related factors is particularly important because of the difficulty of standardising the diagnosis. Although fewer women in the POPI trial were assessed as having PID in the immediate group than in the deferred screening group, slightly higher percentages of women in the immediate screening group had records evaluated (17.3% vs 14.0%, p=0.031) or reported any symptom (34.0% vs 31.3%, p=0.193). It would therefore be useful to know the level of agreement between assessors, whether the blinded assessors based their diagnoses on additional information about sexual behaviour in addition to symptoms, and the distribution of chlamydia infection at baseline among women who reported symptoms and whose records were assessed.

The 35% reduction in the overall incidence of PID in the POPI trial is probably the maximum benefit of this type of intervention in which 100% of a population has a one-off screen and the vast majority of positive cases are treated. This result is not out of line with the results of the two previous trials,5 6 especially if the smaller effect estimate reflects the more rigorous methodology. The reduction in PID among women with chlamydia at baseline (1/63 in the intervention group vs 7/74 in the control group, RR 0.17, 95% CI 0.03 to 1.01) is also encouraging. Nevertheless, the 10 reported cases of chlamydia-positive PID among women who were chlamydia-negative at baseline could not have been prevented by the screening intervention. These cases highlight ongoing chlamydia transmission as the underlying problem; infections acquired in new relationships and reinfections within ongoing partnerships that result from failed partner notification still need to be tackled.7 Obtaining specimens at follow-up for chlamydia testing and genotyping would have provided valuable information about these issues.

Given the difficulties of using PID as the primary end point, two new trials that involve multiple screening rounds and use chlamydia prevalence as an end point are going to provide essential additional information about the effectiveness of chlamydia screening.8 In three areas of The Netherlands, 16% of all invited individuals aged 16–29 years have been tested in the first round of the Chlamydia Screening Implementation (CSI) project.9 The Australian government has funded the Australian Chlamydia Control Effectiveness Pilot (ACCEPt, trial registration number ACTRN12610000297022), a pragmatic cluster randomised trial.8 The aim is to determine whether annual recall for 16–29-year-old women and men attending general practices can increase repeated chlamydia screening to levels that are high enough to reduce its prevalence in the population.10 Australia is well placed to conduct such a trial as current screening levels are still quite low (<9% of under 25-year-olds). The intervention will be cluster randomised at the postcode (town) level so that the intervention can be delivered to people within social or sexual networks. All primary care clinics in each town will be invited to participate and general practitioners in the intervention areas will receive a multifaceted package to facilitate screening, annual recall and case management. The ACCEPt study aims to enrol around 54 towns and to compare the prevalence of chlamydia before and after the intervention in around 3500 people. Chlamydia prevalence will be estimated from consecutive samples of patients attending intervention and control clinics for any reason. The incidence of PID will be a secondary outcome. Enrolment of study sites is beginning in May 2010.

Controlling chlamydia transmission is proving much more difficult to deal with than gonorrhoea in the 1970s.11 The characteristics of C trachomatis, including the long duration of asymptomatic infection and prolonged life cycle, contribute to sustaining transmission and reinfection which drive the progression to PID. The results of the POPI trial suggest that current levels of chlamydia screening uptake in England are unlikely to have a measurable impact on the overall incidence of PID. The interventions being evaluated now should prioritise both screening and prevention of reinfection to maximise the impact on chlamydia transmission.



  • Funding Australian Government Department of Health and Ageing.

  • Competing interests JH is principal investigator on the ACCEPt trial. NL is a co-investigator on the ACCEPt trial and has given advice on the design of the CSI project. The ACCEPt trial is funded by the Australian government through the Department of Health and Ageing. However, the views expressed in the editorial are those of the authors.

  • Provenance and peer review Commissioned; not externally peer reviewed.