Pregnant women and neonates were recruited at Shenzhen Nanshan Maternity and Child Health Care Hospital. The study protocols were approved by the ethics committee of Shenzhen Nanshan Maternity and Child Health Care Hospital (No. NSFYEC-KY-2022018) and conducted in accordance with all relevant guidelines and regulations. Written informed consent was obtained from all participants, and for neonates, consent was provided by their parents or legal guardians. Participants who did not consent to participation were excluded from the study. Additionally, participants were excluded if they had taken antibiotics or amicrobial therapy during pregnancy, consumed probiotics within three months before delivery, experienced adverse pregnancy outcomes, had a history of smoking and alcohol consumption during pregnancy, had complicated singleton pregnancies, had any underlying medical diseases prior to pregnancy, or if the newborn had physical abnormalities.

A total of forty-four healthy women with a gestational age (GA) greater than 37 weeks who delivered vaginally at maternity centers were included. Specifically, 17 pregnant women were classified into the VVC group, while the remaining 27 pregnant women were designated as the normal controls (HC group). VVC diagnosis was established through microscopic and culture-based identification of Candida, as previously reported (Ceccarani et al., 2019). All pregnant women were tested at 35-36 weeks of gestation at our hospital, and only asymptomatic, without therapeutic intervention VVC cases recruited to minimize the interference of external factors on study outcomes.

Vaginal swab samples were collected from all study participants in two different phases: upon admission to the hospital for delivery (phase BD, before delivery), and immediately after delivery (phase PD, after delivery). To ensure sufficient DNA concentration, three sterile cotton swab sticks were used to collect vaginal swab samples by trained nurses. Following a previously reported method (Li et al., 2023), the sterile swab was placed carefully on the vaginal sidewall about halfway between the introitus and the cervix, and rolled dorsally-ventrally back and forth four times to coat the swab.

Meconium samples were obtained within 24 hours after delivery and collected in a sterile tube under antiseptic handling by trained nurses according to previously reported (Turunen et al., 2021). All samples were stored at -80°C until DNA extraction.

Microbial DNA was extracted using the QIAamp DNA Mini kit, and its concentration and purity were measured using the NanoDrop One (Thermo Fisher Scientific, MA, USA). The V3-4 variable region of the 16S rRNA gene was amplified using forward primer 338F (5’-ACTCCTACGGGAGGCAGCAG-3’), and the reverse primer 806R (5’-GGACTACHVGGGTWTCTAAT-3’) (Chen et al., 2021a). PCR reactions, containing 25 μL 2× Premix Taq (Takara Biotechnology, Dalian Co. Ltd., China), 2 μL of each 10 mM primer, and 3 μL DNA template in a volume of 50 μL, were amplified under the following thermal profile: 94 °C for 5 minutes, then 30 cycles of 94 °C for 30 seconds, 52 °C for 30 seconds, 72 °C for 30 seconds, followed by 72 °C for 10 minutes. PCR products were then pooled in equimolar and paired-end sequenced on an Illumina MiSeq platform with V3 chemistry.

The 16S rRNA gene sequences were analyzed using the bioinformatics software package QIIME2 (version 2021.8) (Bolyen et al., 2019). Paired-end reads were denoised by QIIME2 using the command “qiime dada2 denoise-paired”, which aimed to merge paired-end reads, perform quality filtering, and exclude chimeric and phiX sequences. Taxonomic assignment was performed against the Greengenes (13_8 revision) database using command “qiime feature-classifier classify-sklearn”. Alpha- and beta-diversity measures were generated using the commands “qiime phylogeny align-to-tree-mafft-fasttree” and “qiime diversity core-metrics-phylogenetic”.

Continuous variables were presented as the mean ± standard deviation (SD), and categorical characteristics were reported as numbers (percentages). All comparisons were performed using R software at a significant level of 0.05, with chi-square and t-tests for categorical and continuous variables, respectively.

Principal coordinate ordination analysis (PCoA) was performed on Bray-Curtis distances and accompanied by permutational multivariate analysis of variance (PERMANOVA, 999 permutations) using the R package “vegan” and MicrobiotaProcess (Xu et al., 2023). Differentially abundant taxa between the two groups were estimated using LEfSe (Linear Discriminate Analysis of Effect size) software, with a significance threshold of P < 0.05 and LDA > 2.0 (Segata et al., 2011).

The flowchart illustrating the subject recruitment and exclusion process was detailed in Figure 1 . Ultimately, forty-four pregnant women (VVG group, n = 17; HC group, n = 27) were included in this study. All participants were Han Chinese residing in Shenzhen city. The characteristics of all participants were summarized in Table 1 , there were no differences in terms of maternal age, gestational age, neonatal birth length and neonatal gender. Due to insufficient DNA in some meconium samples, only 16 neonates in the VVC group and 17 neonates in the HC groups were retained for further analysis. Then in the data analysis process, two steps were undertaken: first, the number of maternal vaginal samples was adjusted to match the number of newborn meconium samples, and second, all maternal vaginal samples were standardized to ensure consistency across groups.

Alpha diversity was assessed using the observed feature number, Pielou index, and Shannon index. The results demonstrated that the mean values of the observed feature number significantly increased in the VVC group compared to the HC group (P = 0.043), while no significant differences were observed in the Pielou (P = 0.19) and Shannon indices (P = 0.056) ( Figure 2A ). Similar findings were observed in the analysis of all 44 study subjects ( Supplementary Figure S1A ). Furthermore, PCoA with PERMANOVA based on Bray-Curtis distance showed a trend toward significant differences in vaginal microbial communities between the two groups (R² = 0.061, P = 0.069, Figure 2B ). Upon expanding the sample size, a significant difference was confirmed (R² = 0.065, P = 0.015, Supplementary Figure S1B ).

The average relative abundance of dominant phyla, including Firmicutes, Actinobacteria, Proteobacteria, Bacteroidetes, and Tenericutes, was shown in Figure 2C . Firmicutes and Actinobacteria together accounted for over 90% of the microbial composition in both groups. At the genus level, the predominant taxa included Lactobacillus, Gardnerella, Delftia, Atopobium, Methylosinus, Veillonella, Streptococcus, Acinetobacter, Methyloversatilis, and Aerococcus ( Figure 2D ). Similar microbial distributions were observed across all 44 study subjects ( Supplementary Figures S1C, D ). LEfSe analysis revealed significant differences at the genus level between the two groups, with Delftia, Burkholderia, Atopobium, Streptococcus, Stenotrophomonas and Pseudomonas significantly enriched in the VVC group, while Methyloversatilis, Methylosinus, and Lactobacillus significantly decreased compared to the HC group ( Figure 2E ). Expanding sample size confirmed a significant increase in Delftia, Burkholderia and Stenotrophomonas, along with a significant decrease in Methylosinus and Lactobacillus were observed in the VVC group compared to the HC group ( Supplementary Figure S1E ).

After delivery, the mean values of the observed feature number, Pielou and Shannon indices showed no significant differences between the VVC and HC groups ( Figure 3A ). Consistent findings were observed in the analysis of all 44 study subjects ( Supplementary Figure S2A ). However, PCoA with PERMANOVA based on Bray-Curtis distance revealed significant differences in the vaginal microbial communities between the two groups (R² = 0.1394, P = 0.0001, Figure 3B ). When the sample size was expanded, this significant difference was further confirmed ( Supplementary Figure S2B ).

The average relative abundance of dominant phyla was shown in Figure 3C , similar to the vaginal microbiome before delivery, including Firmicutes, Proteobacteria, Actinobacteria, Tenericutes, and Bacteroidetes. The predominant genera included Lactobacillus, Delftia, Gardnerella, Ureaplasma, Veillonella, Streptococcus, Corynebacterium, Staphylococcus, Prevotella, Enterococcus, and Bifidobacterium ( Figure 3D ). These findings were confirmed across all study subjects ( Supplementary Figures S2C, D ). LEfSe analysis revealed significant genus-level differences between the two groups, with Delftia, Methylocaldum, Acinetobacter, Pelomonas, Stenotrophomonas, Propionibacterium, Burkholderia, Exiguobacterium, Sediminibacterium, Sphingomonas, and Ralstonia significantly enriched, while Methyloversatilis, Methylosinus, and Lactobacillus significantly decreased in the VVC group compared to the HC group ( Figure 3E ). These differences in genus were also confirmed with the expanded sample size ( Supplementary Figure S2E ).

Consistently, both before delivery and after delivery, Delftia, Burkholderia, and Stenotrophomonas were significantly enriched in the VVC group. All these genera belong to the phylum Proteobacteria. This was reflected in the differences at the phylum level, with Proteobacteria significantly enriched in the VVC group at both BD (LDA = 4.61, P < 0.001) and PD (LDA = 5.33, P < 0.001) phases.

For the neonatal meconium microbiome, the mean values of the observed feature number, Pielou and Shannon indices showed no differences between the VVC and HC groups ( Figure 4A ). However, PCoA with PERMANOVA revealed significant differences in the neonatal meconium microbial communities between the two groups (R² = 0.0518, P = 0.0001). The VVC group exhibited more closely clustered microbial communities, indicating a more homogeneous community structure among the neonates in the VVC group ( Figure 4B ).

The average relative abundance of dominant phyla was shown in Figure 4C , including Proteobacteria, Firmicutes, Actinobacteria, Bacteroidetes, and Acidobacteria. The predominant genus included Veillonella, Streptococcus, Rhizobium, Proteus, Lactobacillus, Enterococcus, Corynebacterium, Comamonas, Bacteroides, Arthrobacter, Allobaculum and Acinetobacter ( Figure 4D ). LEfSe analysis revealed significant differences at the genus level between the two groups, with Brevundimonas, and Delftia significantly enriched, while Stenotrophomonas, Citrobacter, Rubrivivax, Paracoccus, Chryseobacterium, Xanthomonas, Geobacillus, Klebsiella, Bacteroides, Enterococcus, Proteus, Acinetobacter, Comamonas and Rhizobium significantly decreased in the VVC group compared to the HC group ( Figure 4E ).

It is well-established that the female genital tract is predominantly inhabited by Lactobacillus species, which protect against infections through the production of L-lactic acid and H₂O₂. Pathological changes in their profile may render the vagina susceptible to infections. In this study, we identified various Lactobacillus species, including L. iners, L. helveticus, L. reuteri, L. zeae, L. coleohominis, L. mucosae, L. manihotivorans, L. salivarius and L. acidipiscis ( Supplementary Figure S3 ). Among these, L. iners and L. helveticus were the most abundant species in both vaginal and meconium samples ( Supplementary Figure S3 ). Differences in these Lactobacillus species were further investigated using LEfSe analysis. At the BD phase, no significant differences in the relative abundance of these Lactobacillus species were observed. However, the overall abundance of the genus Lactobacillus was significantly decreased in the VVC group compared to the HC group (LDA = 5.24, P = 0.029), a result confirmed across all 44 study subjects (LDA = 5.14, P = 0.027). A similar pattern was observed at the PD phase, with a significant decrease in the genus Lactobacillus in the VVC group (LDA = 5.44, P < 0.01), particularly in species L. helveticus (LDA = 5.20, P = 0.014). Only the significant difference in genus Lactobacillus was confirmed with the expanded sample size (LDA = 5.32, P = 0.001). In contrast, neonatal meconium samples showed no significant differences of the genus Lactobacillus, though L. unclassified (LDA = 2.67, P = 0.021), L. salivarius (LDA = 2.85, P = 0.042), L. helveticus (LDA = 4.47, P = 0.023) were significantly more abundant in the HC group.

In both the VVC and HC groups, the observed feature number and Shannon index significantly increased in the PD phase compared to the BD phase ( Figures 5A, B ), indicating an increase of microbial diversity after delivery. Moreover, PCoA with PERMANOVA based on Bray-Curtis distance revealed significant differences in the vaginal microbial communities between pre- and post- delivery (R² = 0.0429, P = 0.0254 in the HC group, Figure 5C ; R² = 0.0810, P = 0.001 in the VVC group, Figure 5D ). Additionally, after delivery, the relative abundance of Lactobacillus significantly decreased in both the HC and VVC groups (LDA = 4.98, P < 0.001 in the HC group, Figure 5E ; LDA = 5.18, P = 0.026 in the VVC group; Figure 5F ).