Identification of the specific microbial community compositions in saliva associated with periodontitis during pregnancy.

Abstract

To identify the specific microbial community compositions in saliva associated with periodontitis during pregnancy. Unstimulated saliva samples were collected from 53 pregnant women during weeks 24-28 of gestation, and the V3-V4 regions of the 16S rRNA gene were amplified from isolated saliva DNA and sequenced. Phylum-, genus-, and species-level taxonomic compositions were separately compared between subjects with (n = 12) and without (n = 41) periodontitis. Taxa were selected using the random forest algorithm to distinguish subjects with periodontitis at each taxonomic level, and principal component biplots were constructed to visualize the composition of selected taxa in each subject. The genus-level biplot indicated that 44 subjects clustered around the origin. The prevalence of periodontitis was significantly higher among subjects outside the cluster compared with subjects inside the cluster (6/9 [67%] vs. 6/44 [14%], respectively; p = 0.002). Subjects outside the cluster also had significantly decreased abundance of Neisseria and increased abundances of several putative periodontopathic genera. Phylum- and species-level biplots failed to discriminate subjects with periodontitis more efficiently than the genus-level biplot. The specific taxonomic composition of the saliva microbiota in pregnant women with periodontitis could be clearly identified at the genus level. The formula developed based on the present findings, (%Treponema + %Tannerella + %Filifactor + %Anaeroglobus)/%Neisseria, can be used to predict periodontitis during pregnancy with sensitivity and specificity values of 0.67 (8/12) and 0.95 (39/41), respectively.