The association between Chlamydia trachomatis and pelvic inflammatory disease : findings from observational studies

• Main Author: Davies, Bethan
• Other Authors: Ward, Helen ; Garnett, Geoffrey
• Published: Imperial College London 2014
• Subjects: 618.1
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.700657

Estimates of the cost-effectiveness of chlamydia control interventions are highly sensitive to the risk of progression from genital chlamydia infection to pelvic inflammatory disease (PID). There is no consensus for the risk of PID following asymptomatic chlamydia infections detected through population-based testing. The aim of this thesis is to generate improved estimates of this risk of PID that can be used to parameterise mathematical models and inform chlamydia control policy. We have determined the risk of PID following a positive chlamydia test in three cohorts: a small historic prospective clinical cohort of sex workers (Praed Street Project (PSP)); a large population-based retrospective cohort from Manitoba, Canada established for this research; and a large nationally representative retrospective cohort from Denmark. The risk of PID was higher in women with a positive chlamydia test compared to women who tested negative (PSP: adjusted hazard ratio (AHR) 2.03 (95%CI 0.75-5.49); Manitoba: 1.55 (95%CI 1.43-1.70); Denmark: 1.42 (95%CI 1.32-1.53)). There was heterogeneity in this risk: 13-23% higher following a repeat infection; up to four-fold higher in younger women (Manitoba: AHR 4.55 (95%CI 3.59-5.78) in 12-15 compared to 30-40 years); two-fold higher following previous gonorrhoea (PSP: 2.28 (95%CI 1.14-4.56)). The increased risk following a positive test lasted considerably longer than the likely duration of infection and fewer than 10% of PID diagnoses within 12 months of a test could be attributed to a positive result. This suggests that there are other important causes of PID. Individual-based risks of progression that capture this heterogeneity may improve the accuracy of estimates from mathematical models and therefore their utility to policy makers. Further research is needed to fully characterise the aetiology of PID to inform the design of chlamydia control interventions. In the meantime, interventions should focus on young women and those at risk of repeat chlamydia infection.