Detection of prespecified CE-associated pathogen targets by quantitative polymerase chain reaction is associated with CD138-positive endometrium: a retrospective cohort study

INTRODUCTION: We examined whether endometrial quantitative polymerase chain reaction (qPCR) detection of DNA from prespecified pathogen targets classically implicated in chronic endometritis (CE), from other reproductive-tract bacterial targets, or from Lactobacillus is associated with CD138-defined CE, independent of incidental endometrial polyps. METHODS: We included 126 infertile women who underwent concurrent CD138 immunostaining and endometrial qPCR and who had also undergone hysteroscopy. Women with antibiotic use within 1 month before sampling, Lactobacillus probiotic use, insufficient samples for CD138 assessment, or no hysteroscopy were excluded. Robustness was evaluated by two prespecified polyp-exclusion analyses, by exclusion of women with endometriosis or adenomyosis, and by multivariable logistic regression. RESULTS: CD138-positive endometrium was detected in 57 of 126 women (45.2%). Detection of CE-associated pathogen targets was associated with CD138 positivity (odds ratio 8.01, 95% confidence interval 1.70-37.85, P = 0.003; exact conditional odds ratio 7.89, 1.61-76.56), with specificity 97.1%, positive predictive value 84.6%, and sensitivity 19.3%, but this rested on only 13 target-positive women, of whom 11 were CD138-positive, so the estimate is imprecise. Other reproductive-tract bacterial targets were not associated with CD138 positivity (odds ratio 0.55, 0.25-1.22, P = 0.171). Lactobacillus detection was not significantly associated with CD138 positivity (odds ratio 0.54, 0.27-1.11, P = 0.107), although several analyses were directionally consistent with lower reported Lactobacillus values in CD138-positive women; this was suggestive rather than confirmatory. Conversely, 80.7% (46/57) of CD138-positive women had no detectable pathogen target. After polyp exclusion, the association was maintained (odds ratios 17.44 and 18.79), whereas the other categories remained nonsignificant. The association persisted after exclusion of women with endometriosis or adenomyosis (odds ratio 8.32, 1.75-39.60) and in the parsimonious multivariable model (adjusted odds ratio 8.94, 1.85-43.24; Firth penalized adjusted odds ratio 7.40, 1.67-32.90). DISCUSSION: Detection of prespecified CE-associated pathogen targets, but not detection of other reproductive-tract bacterial targets, was associated with CD138-positive CE, independent of incidental polyps; the Lactobacillus findings were suggestive but not confirmatory. Because the panel provides no universal bacterial quantification, these data cannot evaluate non-Lactobacillus-dominant microbiota as a community state. qPCR may help identify a high-specificity, DNA-defined pathogen-associated subset, while CD138-positive CE appears etiologically heterogeneous and still requires histologic diagnosis.