Chlamydia related bacteria (Chlamydiales) in early pregnancy: community-based cohort study

Abstract

Objectives: Serological case-control studies suggest that certain chlamydia-related bacteria (Chlamydiales) which cause cows to abort may do the same in humans. Chlamydiales include Waddlia chondrophila, Chlamydia abortus and Chlamydia trachomatis. Data on prevalence of Chlamydiales in pregnancy are sparse. Using stored urine samples from a carefully characterised cohort of 847 newly pregnant women recruited from 37 general practices in London, UK, we aimed to investigate the prevalence and types of Chlamydiales infections. We also explored possible associations with miscarriage or spontaneous preterm birth.

Methods: Samples were tested using W. chondrophila and pan-Chlamydiales specific real-time PCRs targeting the 16S rRNA gene. Samples positive on either PCR were subjected to DNA sequencing and C. trachomatis PCR.

Results: The overall prevalence of Chlamydiales was 4.3% (36/847, 95% CI 3.0% to 5.8%). The prevalence of W. chondrophila was 0.6% (n = 5), C. trachomatis 1.7% (n = 14), and other Chlamydiales species 2.0% (n = 17). Infection with C. trachomatis was more common in women aged <25, of black ethnicity or with bacterial vaginosis, but this did not apply to W. chondrophila or other Chlamydiales. Follow up was 99.9% at 16 weeks gestation and 90% at term. No infection was significantly associated with miscarriage at ≤12 weeks (prevalence 10%, 81/827) or preterm birth <37 weeks (prevalence 4%, 23/628). Of 25 samples sequenced, seven (28%) were positive for Chlamydiales bacterium sequences associated with respiratory tract infections in children.

Conclusion: In the first study to use the pan-Chlamydiales assay on female urine samples, 4% of pregnant women tested positive for Chlamydiales, including species known to be pathogenic in mothers and neonates.

Keywords: Chlamydiales; Cohort study; Miscarriage; Pregnancy; Preterm birth; Prevalence.

Fig. 1

Chlamydiales flow chart. *Samples were collected in 1998–2000 and labelled by hand with ID number and date of birth. By 2013–14, after 15 years of storage, some labels were difficult to read, some containers were empty and some were missing. From 1216 participants we analysed 847 samples (70%) that had a clear ID number and date of birth that matched details on the questionnaire.