Relationship between vitamin D status and the vaginal microbiome during pregnancy

Abstract

Objective: Evidence supports an inverse association between vitamin D and bacterial vaginosis (BV) during pregnancy. Furthermore, both the vaginal microbiome and vitamin D status correlate with pregnancy outcome. Women of African ancestry are more likely to experience BV, to be vitamin D deficient, and to have certain pregnancy complications. We investigated the association between vitamin D status and the vaginal microbiome.

Study design: Subjects were assigned to a treatment (4400 IU) or a control group (400 IU vitamin D daily), sampled three times during pregnancy, and vaginal 16S rRNA gene taxonomic profiles and plasma 25-hydroxyvitamin D [25(OH)D] concentrations were examined.

Result: Gestational age and ethnicity were significantly associated with the microbiome. Megasphaera correlated negatively (p = 0.0187) with 25(OH)D among women of African ancestry. Among controls, women of European ancestry exhibited a positive correlation between plasma 25(OH)D and L. crispatus abundance.

Conclusion: Certain vaginal bacteria are associated with plasma 25(OH)D concentration.

Figure 1.. Total circulating 25(OH)D concentrations increase over pregnancy.

Boxplots of 25(OH)D concentration in plasma collected longitudinally over pregnancy. Whiskers extend to the highest/lowest value within 1.5 times the interquartile range and outliers beyond the whiskers are plotted as points. Significant differences in 25(OH)D between visits are indicated for A. all participants, B. women of African ancestry, C. women of European ancestry, and D. Hispanic women. Symbols indicating statistical significance - ns: p>0.05; *: p<=0.05; **: p<=0.01; ***: p<=0.001; ****: p<=0.0001. Figure was prepared using ggpubr R package.

Figure 2.. Vaginal microbial taxa differ between women with plasma 25(OH)D <30 ng/mL or >40 ng/mL.

A. and B. stacked bar plots showing vaginal microbial community profiles from 3 visits from each of the 236 women in the cohort. The profiles are grouped by the most abundant species and samples within each community group are clustered on bray distances using ward method. The microbial profiles are annotated by Nugent score. The distances were calculated using vegan R package and the figure was prepared using ggplot2.

Figure 3.. The association between 25(OH)D and microbiome differs among ethnic groups.

Stacked bar plots showing vaginal microbial community profiles from 3 visits from each of the 236 women in the cohort grouped by ethnicity and 25(OH)D status.

Figure 4.. Statistical association analysis using LEfSe.

Bacterial species with significant differential abundance between women with plasma 25(OH)D concentrations >40 ng/mL or <30 ng/mL were identified using LEfSe. Features with LDA score greater than 3.0 are shown in the figure, with bacterial species associated with plasma 25(OH)D concentrations >40 ng/mL shown in green, while those associated with <30 ng/mL are shown in red.

Figure 5.. G. vaginalis abundance decreases over pregnancy irrespective of 25(OH)D concentration.

Box-plots showing plasma 25(OH)D concentration (left-panel) and G. vaginalis abundance (right-panel) for all subjects with G. vaginalis abundance >1% at visit 1. Data points from same subject at visit 1 and visit 7 are connected by gray line. There is a significant increase in 25(OH)D among the treatment group and significant decrease in G. vaginalis abundance in both control and treatment group. The figure was prepared using ggpubr R package.