Vaginal Lactobacillus crispatus in early pregnancy associates with favorable gestational outcomes in a Japanese maternal-neonatal microbiome cohort

Abstract

The maternal microbiome during pregnancy and the peripartum period plays a critical role in maternal health outcomes and establishing the neonatal gut microbiome, with long-term implications for offspring health. However, a healthy microbiome during these key periods has not been definitively characterized. This longitudinal study examines maternal and neonatal microbiomes using 16S rRNA amplicon sequencing in a Japanese cohort throughout pregnancy and the postpartum period. Forty-two mothers and their forty-five offspring participate in the study. The maternal vaginal microbiome remains relatively stable during pregnancy but significantly changes in the postpartum period. Among Lactobacillus species, the Lactobacillus crispatus group is predominant. A higher abundance of Lactobacillus early in pregnancy is associated with a favorable gestational period. The maternal gut microbiome is associated with the vaginal microbiome throughout pregnancy. The neonatal gut microbiome substantially changes in early life, with bacterial composition influenced by delivery mode. Over time, bacteria shared with the maternal gut microbiome become dominant in the neonatal gut. This study provides insights into microbiome dynamics in Japanese mothers and their offspring during pregnancy and the postpartum period. Identification of common patterns across diverse populations may help define keystone microbes essential for human health and inform the development of microbiome-based interventions.

Fig. 1. Analysis of the maternal vaginal microbiome during pregnancy and the postpartum period.

a Study design outlining the collection of maternal samples. b Chao1 and Shannon diversity indices in the vaginal microbiome (Week 12: n = 42, Week 20: n = 40, Week 30: n = 41, Week 36: n = 39, Birth: n = 41, Postpartum: n = 36). Each p-value is shown in Supplementary Table 1. c Principal coordinates analysis (PCoA) plots based on unweighted and weighted UniFrac distances of bacterial compositions in the maternal vaginal microbiome. Each p-value is shown in Supplementary Table 2. Data are presented as means ± standard errors of the mean. *p < 0.05, **p < 0.01, ****p < 0.0001, Pairwise comparisons (two-sided) were performed using estimated marginal means, accounting for repeated measures, and adjusted for multiple comparisons using the Benjamini-Hochberg method in panel (b). Bacterial compositions were compared using permutational analysis of variance with permutations restricted within each participant, and adjusted for multiple comparisons using the Benjamini-Hochberg method in panel (c).

Fig. 2. Dynamics of Lactobacillus in the maternal vaginal microbiome during pregnancy and the postpartum period, and its potential role in preterm birth.

a Comparison of the relative abundance of bacterial genera in the vaginal microbiome between pregnancy and the postpartum period. b Relative abundances of the top 10 bacterial genera in the vaginal microbiome during pregnancy. c Chao1 and Shannon diversity indices in samples where Lactobacillus was dominant (n = 129) compared with those in which other genera were dominant (n = 33) (Chao1 index: p = 0.00000005, Shannon diversity index: p = 0.0000004). d Longitudinal shifts in the relative abundance of Lactobacillus: sustained high levels from early to late pregnancy (blue), low levels in early pregnancy followed by high levels in late pregnancy (green), and sustained low levels from early to late pregnancy (red). e Comparison of pregnancy continuation beyond 38 weeks between mothers in whom Lactobacillus was dominant during early pregnancy and those in whom other genera were dominant at that time (bar chart; p = 0.0164). Kaplan–Meier curves illustrate pregnancy continuation up to 38 weeks (p = 0.0025). f Comparison of pregnancy continuation beyond 38 weeks between mothers in whom Lactobacillus was dominant during late pregnancy and those in whom other genera were dominant at that time (bar chart; p = 1.0000). Kaplan–Meier curves illustrate pregnancy continuation up to 38 weeks (p = 1.0000). Data are presented as means ± standard errors of the mean. *p < 0.05, ***p < 0.001, Mann–Whitney U test (two-sided) for comparisons of two groups in panels (c), (e), and (f); log-rank test for comparisons of Kaplan–Meier curves in panels (e) and (f).

Fig. 3. Lactobacillus subgenera in the maternal vaginal microbiome during pregnancy.

a Heatmap and dot plots showing the relative abundances of Lactobacillus subgenera at each time point. In the heatmap, each column represents an individual mother (Week 12: n = 42, Week 20: n = 40, Week 30: n = 41, Week 36: n = 39). b Alluvial plot depicting the dominant Lactobacillus subgenera in vaginal samples over time. c Comparison of Chao1 and Shannon diversity indices between samples in which the L. crispatus group was dominant (n = 104) and those in which other Lactobacillus subgenera were dominant (n = 25) during pregnancy (Chao1 index: p = 0.0125, Shannon diversity index: p = 0.0294). Data are presented as means ± standard errors of the mean. *p < 0.05, Mann–Whitney U test (two-sided) for comparisons of two groups in panel (c).

Fig. 4. Analysis of the maternal gut microbiome during pregnancy and the postpartum period.

a Chao1 and Shannon diversity indices in the maternal gut microbiome during pregnancy and the postpartum period (Week 12: n = 40, Week 20: n = 40, Week 30: n = 41, Week 36: n = 39, Birth: n = 21, Postpartum: n = 38). Each p-value is shown in Supplementary Table 3. b Principal coordinates analysis (PCoA) plots based on unweighted and weighted UniFrac distances of bacterial compositions in the maternal gut microbiome during pregnancy and the postpartum period. Each p-value is shown in Supplementary Table 4. c Comparison of bacterial genera in the maternal gut microbiome between pregnancy and the postpartum period. d Comparison of bacterial genera in the maternal gut microbiome between early pregnancy (Week 12) and late pregnancy (Week 36). e Multivariable association analysis of the relative abundance of Lactobacillus in the maternal vagina and the relative abundances of bacterial genera in the maternal gut. f Multivariable association analysis of the relative abundances of Lactobacillus subgenera in the maternal vagina and bacterial genera in the maternal gut. Data are presented as means ± standard errors of the mean. Pairwise comparisons (two-sided) were performed using estimated marginal means, accounting for repeated measures, and adjusted for multiple comparisons using the Benjamini-Hochberg method in panel (a). Bacterial compositions were compared using permutational analysis of variance with permutations restricted within each participant, and adjusted for multiple comparisons using the Benjamini-Hochberg method in panel (b).

Fig. 5. Dynamics of the neonatal gut microbiome in early life.

a Time points for neonatal sample collection. b Chao1 and Shannon diversity indices in the neonatal gut microbiome (Day 0: n = 27, Day 1: n = 28, Day 4: n = 43, Month 1: n = 39). Each p-value is shown in Supplementary Table 5. c Principal coordinates analysis (PCoA) plots based on unweighted and weighted UniFrac distances of bacterial compositions in the neonatal gut microbiome. Each p-value is shown in Supplementary Table 6. d Comparison of the relative abundances of bacterial genera in the neonatal gut microbiome between Day 1 and Day 4. e Comparison of the relative abundances of bacterial genera in the neonatal gut microbiome between Day 4 and Month 1. f PCoA plots based on unweighted and weighted UniFrac distances at each time point, comparing bacterial composition in the neonatal gut microbiome between vaginally delivered and cesarean-section-delivered offspring. g Multivariable association analysis of bacterial genera in the neonatal gut microbiome, comparing vaginally delivered and cesarean-section-delivered offspring at each time point. Data are presented as means ± standard errors of the mean. *p < 0.05, **p < 0.01, ****p < 0.0001. Pairwise comparisons (two-sided) were performed using estimated marginal means, accounting for repeated measures, and adjusted for multiple comparisons using the Benjamini-Hochberg method in panel (b). Bacterial compositions were compared using permutational analysis of variance with permutations restricted within each participant, and adjusted for multiple comparisons using the Benjamini-Hochberg method in panels (c) and (f).

Fig. 6. Shared microbes between the maternal microbiome and the neonatal gut microbiome.

a, b Proportion and relative abundance of ASV types in the neonatal gut microbiome that were shared with the maternal vaginal or gut microbiome in vaginally delivered offspring (Vagina Day 0: n = 16, Day 1: n = 17, Day 4: n = 20, Month 1: n = 18; Gut Day 0: n = 16, Day 1: n = 17, Day 4: n = 19, Month 1: n = 17). c, d Proportion and relative abundance of ASV types in the neonatal gut microbiome that were shared with the maternal vaginal, gut, or skin microbiome in cesarean-section-delivered offspring (Vagina Day 0: n = 10, Day 1: n = 13, Day 4: n = 22, Month 1: n = 20; Gut Day 0: n = 8, Day 1: n = 12, Day 4: n = 22, Month 1: n = 20; Skin Day 0: n = 6, Day 1: n = 10, Day 4: n = 18, Month 1: n = 17). e Comparison of the proportion and relative abundance of ASV types shared with the maternal gut microbiome between the gut microbiomes of vaginally delivered (Day 0: n = 16, Day 1: n = 17, Day 4: n = 19, Month 1: n = 17) and cesarean-section-delivered offspring (Day 0: n = 8, Day 1: n = 12, Day 4: n = 22, Month 1: n = 20; Skin Day 0: n = 6, Day 1: n = 10, Day 4: n = 18, Month 1: n = 17). f Comparison of the proportion and relative abundance of ASV types shared with the maternal vaginal microbiome between the gut microbiomes of vaginally delivered (Day 0: n = 16, Day 1: n = 17, Day 4: n = 20, Month 1: n = 18) and cesarean-section-delivered offspring (Vagina Day 0: n = 10, Day 1: n = 13, Day 4: n = 22, 96Month 1: n = 20). Each p-value in all Figures is shown in Supplementary Table 7. Data are presented as means ± standard errors of the mean. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, Mann–Whitney U test (two-sided) for comparisons of two groups in panels (a), (b), (e), and (f). Kruskal–Wallis test followed by Dunn’s test (two-sided) for multiple comparisons in panels (c) and (d).