Article Text


Epidemiology poster session 3: Burden of disease: PID
P1-S3.09 Estimating the incidence of pid following chlamydia infection in sex workers
  1. B Davies1,
  2. S Day2,
  3. H Ward1
  1. 1Imperial College London, London, UK
  2. 2Goldsmith's College, University of London, UK


Background There is a lack of consensus on the true incidence of PID following chlamydia infection with few published prospective studies. We have used data from an old prospective cohort of sex workers to investigate the association between Chlamydia trachomatis and subsequent pelvic inflammatory disease (PID).

Methods 598 sex workers were recruited between 1985 and 1993 in London. Self-reported exposure to chlamydia and gonorrhoea at enrolment, and diagnoses of chlamydia, gonorrhoea, trichomoniasis, bacterial vaginosis (BV), candida and PID during the study were recorded. Chlamydia was diagnosed by direct immunofluorescence. Incident infections and PID diagnoses were determined in women who attended clinic more than once and did not have PID at enrolment. Crude incidence rates of PID were calculated. Cox proportional hazards regression was used to investigate the association between chlamydia, gonorrhoea and PID.

Results 275 women with 387 person years (py) of follow-up were diagnosed with 85 cases of chlamydia, 34 cases of gonorrhoea, 25 cases of trichomoniasis, 253 cases of BV and 148 cases of candida. Thirty-eight women had at least one episode of PID and of these, 24 had chlamydia concurrently or preceding their PID. Nine women presented with both chlamydia and PID, 1 had concurrent gonorrhoea and PID, and 3 had gonorrhoea during the 8 weeks preceding their PID. The crude incidence rate was 16.03 cases per 100 py (95% CI 12.50 to 20.56) for chlamydia and 6.72 cases per 100 py (95% CI 4.58 to 9.87) for gonorrhoea. The crude rate and crude and adjusted HRs of PID were calculated, with women censored after their first episode of PID, for all women and for women classified according to their exposure to chlamydia and gonorrhoea (Abstract P1-S3.09 table 1).

Conclusions In this cohort chlamydia was a risk factor for PID. After controlling for other factors, exposure to chlamydia at enrolment was a borderline predictor of PID while incident infection was not associated, possibly due to the small number of incident cases or to the reduced risk because of prompt treatment. The absolute risk of PID after an incident case was higher than recent estimates which may be due to the greater sensitivity of current chlamydia tests. This emphasises the importance of using contemporary parameters when modelling the cost-effectiveness of chlamydia screening.

Abstract P1-S3.09 Table 1

Crude rate and crude and adjusted HRs of PID for all women and classified by exposure to chlamydia and gonorrhoea

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.