Abstract
Many diverse pathogens have been discovered from reproductive-tract infections, but the relationship between the presence and abundance of particular pathogen species and disease manifestations is poorly defined. The present work examined the association of multiple common pathogens causing sexually transmitted infections (STIs) with cervicitis and vaginitis. The presence and abundance of 15 STI pathogens and the genotypes of human papillomavirus were determined in a cohort of 944 women that included 159 cervicitis patients, 207 vaginitis patients, and 578 healthy controls. Logistic regression and random forest models were constructed and validated in a separate cohort of 420 women comprising 52 cervicitis patients, 109 vaginitis patients, and 259 healthy controls. The frequency of individual STI pathogen species varied among the symptomatic patients and healthy controls. Abundance determination was necessary for most pathogens that were associated with the studied diseases. STI pathogens were more commonly associated with cervicitis than with vaginitis. Pathogen identification- and quantification-based diagnosis was observed for cervicitis with high sensitivity and specificity, but for vaginitis, the assay results would need to be combined with results of other diagnostic tests to firmly establish the pathogen-disease correlation. Integrated qualitative and quantitative detection of a selected panel of common STI pathogens can reveal their association with cervicitis and vaginitis. STI pathogen identification and quantification can be used to diagnose cervicitis and also help improve correct diagnosis of vaginitis. IMPORTANCE Scarce information exists with regard to whether STI pathogens can be defined as valid microbiological predictive markers for the diagnosis of cervicitis and vaginitis. We therefore conducted this study to assess the presence and abundance of a wide range of STI pathogens among patients having these two diseases and healthy controls as well. High sensitivity and specificity were observed for cervicitis by pathogen identification- and quantification-based diagnosis. In contrast, the assay results obtained for vaginitis would need to be combined with test results obtained by other diagnostic methods to decisively establish the pathogen-disease correlation. Simultaneous qualitative and quantitative detection of a selected panel of common STI pathogens and further coupling with machine learning models is worthwhile for establishing pathogen-based diagnosis of gynecological inflammations, which could be of great value in guiding the rational use of antimicrobials to control the spread of STIs.