Association of specific microbiota taxa in the amniotic fluid at birth with severe acute and longer-term outcomes of very preterm infants: a prospective observational study

Abstract

Background: Dysbiotic microbial colonization predisposes to severe outcomes of prematurity, including mortality and severe morbidities like necrotizing enterocolitis (NEC), late-onset infection (LOI) and bronchopulmonary dysplasia (BPD). Here, we studied the variations in the bacterial signatures in the amniotic fluid (AF) of very preterm deliveries < 32 weeks with severe acute and longer-term outcomes within a prospective cohort study.

Methods: One hundred twenty-six AF samples were available for 16S rRNA gene metabarcoding to describe bacterial community structure and diversity in connection to intraventricular haemorrhage (IVH), LOI, focal intestinal perforation (FIP), NEC, retinopathy of prematurity (ROP) and the 2-year cognitive (MDI) and motor (PDI) outcome.

Results: Diversity and overall bacterial community composition did not differ between the studied outcomes. But disparities in sequences assigned to single bacterial taxa were observed for the acute outcomes LOI and ROP and the longer-term impairments of MDI and PDI. Enrichments associated with a poor acute outcome were particularly detected in the Escherichia-Shigella cluster, while the predominance of Ureaplasma and Enterococcus species was associated with unrestricted acute and longer-term outcomes. Analysis for FIP did not reach any significance. IVH and NEC constituted rare events, prohibiting the analyses.

Conclusions: Our data provide evidence that microbiota patterns at birth might allow the early identification of infants at risk for the severe outcomes of prematurity and argue against morbidity-specific associations. The data support the early origins hypothesis and relevant contribution of prenatal factors. The partly existing disparities between acute and longer-term outcomes might be traced back to the relevant impact of the diverse longitudinal exposures and socioeconomic factors.

Keywords: 16S rRNA; Amniotic fluid; Bronchopulmonary dysplasia; Intraventricular haemorrhage; Late-onset infection; Microbiome; Preterm infant; Psychomotor outcome; Respiratory distress; Retinopathy of prematurity.

Fig. 1

Study population flow chart. Flow chart of inclusion and exclusion of infants into the study population. Exclusion criteria included AF samples with low read bacterial signal, death before the outcome estimates and the severe morbidities of prematurity with high impact on the outcomes studied. ROP, retinopathy of prematurity; IVH, intraventricular haemorrhage; LOI, late onset infection (>72 hours after birth); FIP, focal intestinal perforation; MDI, mental developmental index; PDI, psychomotor developmental index

Fig. 2

Differences in 16S rRNA gene microbial abundance in AF samples from preterm deliveries and ROP. a Alpha diversity measured as Shannon diversity index shows no significant (p < 0.05) difference between AF samples of preterm deliveries with no (blue), non-severe (orange) and severe (red) retinopathy of prematurity (ROP). Statistical analysis was performed using Wilcoxon Rank-Sum test. b NMDS plot of weighted Unifrac distances shows no significant (p < 0.05) altered bacterial 16S rRNA gene community composition for AF samples of preterm deliveries from Figure 2a. Statistical analysis was performed using PERMANOVA with Benjamini-Hochberg correction for multiple comparisons. c Heat tree of log-fold changes calculated with edgeR including top 100 genera (accounting for 97% of all reads in median). The labelled tree on the left shows the taxonomic information (domain to genus) and is the key for the unlabelled smaller trees. Smaller trees represent a comparison between no, non-severe and severe ROP. Coloured taxa are more abundant in the samples indicated by the coloured subtitle. Significant changes (p < 0.05 in both edgeR and generalized linear model) are marked with green asterisks

Fig. 3

Differences in 16S rRNA gene microbial abundance in AF samples from preterm deliveries and IVH/LOI/FIP. a Alpha diversity (Shannon index) index and beta diversity (NMDS plot of weighted Unifrac distances) shows no significant (p < 0.05) difference in AF samples of preterm deliveries with (red) and without (blue) intraventricular haemorrhage (IVH). Statistical analysis was performed using Wilcoxon Rank-Sum test and PERMANOVA with Benjamini-Hochberg correction for multiple comparisons, respectively. b Alpha diversity (Shannon index) index and beta diversity (NMDS plot of weighted Unifrac distances) shows no significant (p < 0.05) difference in AF samples of preterm deliveries with (red) and without (blue) late onset infection (LOI). Statistical analysis was performed using Wilcoxon Rank-Sum test and PERMANOVA with Benjamini-Hochberg correction for multiple comparisons, respectively. c Alpha diversity (Shannon index) index and beta diversity (NMDS plot of weighted Unifrac distances) shows no significant (p < 0.05) difference in AF samples of preterm deliveries with (red) and without (blue) focal intestinal perforation (FIP). Statistical analysis was performed using Wilcoxon Rank-Sum test and PERMANOVA with Benjamini-Hochberg correction for multiple comparisons, respectively. d Heat tree of log-fold changes calculated with edgeR including top 100 genera (accounting for 97% of all reads in median). The labelled upper tree shows the taxonomic information (domain to genus) and is the key for the unlabelled smaller trees. Smaller trees represent a comparison between the patient groups from Figure 4b and 4c, respectively (blue: increased in 0, red: increased in 1) . Significant changes (p < 0.05 in both edgeR and generalized linear model) are marked with green asterisks

Fig. 4

Differences in 16S rRNA gene microbial abundance in AF samples from preterm deliveries and neurological development. a Alpha diversity (Shannon index) and beta diversity (NMDS plot of weighted Unifrac distances) shows no significant (p < 0.05) difference in AF samples of preterm deliveries with mental developmental index (MDI) ≥85 (0, blue) and <85 (1, red). Statistical analysis was performed using Wilcoxon Rank-Sum test and PERMANOVA with Benjamini-Hochberg correction for multiple comparisons, respectively. b Alpha diversity (Shannon index) and beta diversity (NMDS plot of weighted Unifrac distances) shows no significant (p < 0.05) difference in AF samples of preterm deliveries with psychomotor developmental index (PDI) ≥85 (0, blue) and <85 (1, red). Statistical analysis was performed using Wilcoxon Rank-Sum test and PERMANOVA with Benjamini-Hochberg correction for multiple comparisons, respectively. c Heat tree of log-fold changes calculated with edgeR including top 100 genera (accounting for 97% of all reads in median). The labelled tree on the left shows the taxonomic information (domain to genus) and is the key for the unlabelled smaller trees. Smaller trees represent a comparison between the patient groups from Figure 4a and 4b, respectively (blue: increased in patients with index ≥85, red: increased in patients with index <85). The upper coloured trees include all patients (without those with missing outcome), the lower coloured trees only patients with ROP, left trees results for the MDI and right trees those for the PDI. Significant changes (p < 0.05 in both edgeR and generalized linear model) are marked with green asterisks