Introduction The vaginal microbiota is thought to play a protective role against STIs. While urine has long been used for detection of genital STIs, there have been few studies evaluating the use of urine samples in vaginal microbiome studies. We hypothesise that urine samples could serve as a surrogate for vaginal swab collection. We sought to compare mid-vaginal swabs and random catch urine samples.

Methods Mid-vaginal swabs and random catch urine samples were collected in one sitting from 75 reproductive-age women. Microbiota composition was characterised by sequencing the V3-V4 regions of the 16S rRNA gene on the Illumina platform. Vaginal microbiota were targeted for classification using PECAN, a rapid and accurate taxonomic classifier designed for the vaginal environment. Hierarchical clustering was used to assign community state type (CST) to each sample. CST-I, -II, -III, and -V are dominated by L. crispatus, L. gasseri, L. iners and L. jensenni, respectively, while CSTs IV-A and IV-B represent low-Lactobacillus states with an array of strict and facultative anaerobes. Kappa statistics and Jensen-Shannon distances were used to evaluate the concordance of urine and vaginal samples.

Results A 77% concordance and a 0.70 kappa value were observed for CST assignments, indicating substantial agreement in microbiota structure and composition between vaginal and urine samples within a woman. Out of 17 discordant pairs, 10 pairs had one sample assigned to CST-IV and the other to CST-III. These two CSTs are known to be associated with rapidly fluctuating dysbiotic states. When comparing the population structure of all urine and vaginal samples, no statistical differences were observed (PERMANOVA: F1,148=1.0815, p=0.31).

Conclusion Vaginal and random catch urine samples from the same participant showed substantial agreement on bacterial composition. Random catch urine samples could present another sampling option to assess the vaginal and urogenital microbiota.