Intravenous antibiotics in preterm infants have a negative effect upon microbiome development throughout preterm life

Abstract

Background: Intestinal dysbiosis is implicated in the origins of necrotising enterocolitis and late-onset sepsis in preterm babies. However, the effect of modulators of bacterial growth (e.g. antibiotics) upon the developing microbiome is not well-characterised. In this prospectively-recruited, retrospectively-classified, case-control study, high-throughput 16S rRNA gene sequencing was combined with contemporaneous clinical data collection, to assess the within-subject relationship between antibiotic administration and microbiome development, in comparison to preterm infants with minimal antibiotic exposure.

Results: During courses of antibiotics, diversity progression fell in comparison to that seen outside periods of antibiotic use (-0.71units/week vs. + 0.63units/week, p < 0.01); Enterobacteriaceae relative abundance progression conversely rose (+ 10.6%/week vs. -8.9%/week, p < 0.01). After antibiotic cessation, diversity progression remained suppressed (+ 0.2units/week, p = 0.02).

Conclusions: Antibiotic use has an acute and longer-lasting impact on the developing preterm intestinal microbiome. This has clinical implications with regard to the contribution of antibiotic use to evolving dysbiosis, and affects the interpretation of existing microbiome studies where this effect modulator is rarely accounted for.

Fig.1

Recruitment Flowchart

Fig. 2

A Composite plot of individual trajectories of longitudinal diversity progression in stable babies minimally-exposed to antibiotics (Group 1) shown in grey. The subject with chorioamnionitis who was excluded from all calculations of mean trajectories is shown in red. Weighted mean trajectory (thick black line)—gradient labelled in green. 95%CI for weighted mean trajectory denoted by dotted lines. B Composite graph demonstrating the three components of assessed Enterobacteriaceae relative abundance progression in stable infants minimally-exposed to antibiotics (Group 1): weighted median peak relative abundance (purple); weighted mean time to peak relative abundance (blue); weighted mean regression coefficient of fall from peak (black, with 95%CI denoted by accompanying dotted lines), with gradient shown in green; individual weighted regression coefficients from which the mean is derived (grey). As in A data from the excluded infant with chorioamnionitis (red) is not included in the calculations. CLongitudinal taxonomic profiles of the majority taxa in stable babies minimally-exposed to antibiotics (Group 1; a-m), and the excluded subject (n), highlighted in red. Horizontal red arrows indicate periods of antibiotic administration; black spots on horizontal access represent samples from which no DNA was extracted. Weeks of life is indicated on the x-axis

Fig. 3

A Composite plot of individual trajectories of longitudinal diversity progression during antibiotic administration (Group 2a) shown in thin grey lines, with weighted mean trajectory (thick black line). For comparative purposes, trajectory of diversity progression in Group 1 overlaid (thick green line). 95%CI for weighted mean trajectories denoted by dotted lines. For clarity and comparative purposes, the intercept of the two trajectories is plotted at the median intercept of the group on antibiotics. B Composite plot of individual trajectories of longitudinal Enterobacteriaceae relative abundance progression during antibiotic administration (Group 2a) shown in thin grey lines, with weighted median trajectory (thick black line). For comparative purposes, median trajectory of Enterobacteriaceae relative abundance progression in Group 1 overlaid (thick green line). 95%CI for weighted median trajectories denoted by dotted lines. For clarity and comparative purposes, the intercept of the two trajectories is plotted at the median intercept of the group on antibiotics. C Longitudinal taxonomic profiles of the majority taxa in babies during periods of receiving intravenous antibiotics (Group 2a). Red arrows indicate the sample taken prior to commencing antibiotics. Each panel represents the entire course of antibiotics. Antibiotic courses containing ceftazidime highlighted in blue. Days on antibiotics is indicated on the x-axis

Fig. 4

A Composite plot of individual trajectories of longitudinal diversity progression following antibiotic cessation (Group 2b) shown in thin grey lines, with weighted median trajectory (thick black line). For comparative purposes, trajectory of diversity progression in stable subjects (Group 1) overlaid (thick green line). 95% CI for weighted median trajectories denoted by dotted lines. For clarity and comparative purposes, the intercept of the two trajectories are both plotted at the median y-intercept of Group 2b. B Composite plot of individual trajectories of Enterobacteriaceae relative abundance progression following antibiotic cessation (Group 2b) shown in thin grey lines, with weighted mean trajectory (thick black line). For comparative purposes, trajectory of Enterobacteriaceae relative abundance progression in stable subjects (Group 1) overlaid (thick green line). 95% CI for weighted mean trajectories denoted by dotted lines. For clarity and comparative purposes, the intercept of the two trajectories are both plotted at the median y-intercept of Group 2b. C Longitudinal taxonomic profiles of the majority taxa in babies during periods after cessation of intravenous antibiotics. Days off antibiotics is indicated on the x-axis