Introduction
Vaginal washing has no known health benefits and has been associated with adverse reproductive health outcomes. This analysis tested the hypothesis that vaginal washing is associated with group B streptococcus (GBS) colonization during the periconception period and the first trimester of pregnancy.

Method
Women planning pregnancies contributed monthly visits during which vaginal fluid specimens were collected and urine pregnancy testing was performed. In women who became pregnant, additional vaginal fluid samples were collected at 9-12 weeks gestation. Broad-range 16S rRNA gene PCR with next-generation sequencing was performed to identify bacterial species. Generalized estimating equations with a log link, Poisson family, and independent correlation structure were used to generate prevalence ratios comparing the prevalence of GBS detection at vaginal washing visits versus non-vaginal washing visits.

Results
The 196 women who became pregnant contributed 506 samples collected at periconception 196 (38.9%), early first trimester 151 (29.8%), and first trimester 159 (31.4%) visits. The prevalence of GBS at the three time periods was 20/196 (10.2%), 11/151 (7.3%), and 2/159 (1.3%) respectively. Recent vaginal washing was reported by 51/196 (26.0%), 27/151 (17.9%), and 32/159 (20.1%) participants at the three-time points. Compared to visits with no vaginal washing, there was no increased prevalence of GBS detection at visits where vaginal washing with water was reported (PR 0.51, 95% CI 0.16 - 1.62), but the prevalence of GBS detection was substantially higher at visits when vaginal washing using water and soap was reported (PR 4.66, 95% CI 1.51, 14.33) (Table 1).

Conclusion
Vaginal washing with soap and water was associated with more than four-fold higher GBS prevalence during the periconception period and first trimester. Our findings underscore the need for strategies to discourage this practice.