A Lactobacillus-Deficient Vaginal Microbiota Dominates Postpartum Women in Rural Malawi

Abstract

The bacterial community found in the vagina is an important determinant of a woman's health and disease status. A healthy vaginal microbiota is associated with low species richness and a high proportion of one of a number of different Lactobacillus spp. When disrupted, the resulting abnormal vaginal microbiota is associated with a number of disease states and poor pregnancy outcomes. Studies up until now have concentrated on relatively small numbers of American and European populations that may not capture the full complexity of the community or adequately predict what constitutes a healthy microbiota in all populations. In this study, we sampled and characterized the vaginal microbiota found on vaginal swabs taken postpartum from a cohort of 1,107 women in rural Malawi. We found a population dominated by Gardnerella vaginalis and devoid of the most common vaginal Lactobacillus species, even if the vagina was sampled over a year postpartum. This Lactobacillus-deficient anaerobic community, commonly labeled community state type (CST) 4, could be subdivided into four further communities. A Lactobacillus iners-dominated vaginal microbiota became more common the longer after delivery the vagina was sampled, but G. vaginalis remained the dominant organism. These results outline the difficulty in all-encompassing definitions of what a healthy or abnormal postpartum vaginal microbiota is. Previous identification of community state types and associations among bacterial species, bacterial vaginosis, and adverse birth outcomes may not represent the complex heterogeneity of the microbiota present. (This study has been registered at ClinicalTrials.gov as NCT01239693.)IMPORTANCE A bacterial community in the vaginal tract is dominated by a small number of Lactobacillus species, and when not present there is an increased incidence of inflammatory conditions and adverse birth outcomes. A switch to a vaginal bacterial community lacking in Lactobacillus species is common after pregnancy. In this study, we characterized the postpartum vaginal bacterial community of a large group of women from a resource-poor, undersampled population in rural Malawi. The majority of women were found to have a Lactobacillus-deficient community, and even when sampled a year after delivery the majority of women still did not have Lactobacillus present in their vaginal microbiota. The effect of becoming pregnant again for those who do not revert to a Lactobacillus-dominant community is unknown, and this could suggest that not all Lactobacillus-deficient community structures are adverse. A better understanding of this complex community state type is needed.

Keywords: 16S rRNA gene; Gardnerella vaginalis; Lactobacillus spp.; postpartum; sub-Saharan Africa; vaginal microbiota.

Copyright © 2018 Doyle et al.

Figures

FIG 1

Study participant flow diagram.

FIG 2

The 25 most prevalent OTUs recovered from participants' vaginal swab samples (n = 994).

FIG 3

(A) Principal-coordinate analysis of Bray-Curtis distances comparing presence of OTUs in participants' vaginal swab samples. Participants are colored based on de novo clustering by the shared presence of OTUs (n = 994). (B) Box-and-whisker plot of the distribution of OTU richness. OTU richness was measured among all five CSTs using the Shannon index. Three asterisks (***) above a CST indicate that the difference between it and all other groups was statistically significant (P < 0.0001) as determined by Student's t test (false discovery rate [FDR]-corrected), and ns indicates there was no statistically significant difference between the Shannon index values in CST 4-II and CST-IV.

FIG 4

Heat map of the relative abundances of 30 OTUs ordered by CST. Each individual is annotated both by the given CST and by HIV status where information was available (n = 994).

FIG 5

(A) Histogram of the number of days after delivery that vaginal samples were collected. (B) The proportion of participants assigned to each CST stratified by how long after delivery samples were taken. Participants were binned into 5 groups: 5 to 20 days (n = 809), 21 to 40 days (n = 98), 41 to 100 days (n = 40), 101 to 200 days (n = 21), and 201 to 583 days (n = 26). P values show where there was a significant pairwise comparison between groups as determined by Fisher's exact test (FDR-corrected). (C) Box-and-whisker plot showing the association between relative abundance of L. iners, G. vaginalis, L. amnionii, and P. anaerobius and how long after delivery the vagina was sampled. P values show where there was a significant pairwise comparison between groups as determined by Student's t test (FDR-corrected).