<b><i>Introduction:</i></b> Antibiotic treatment in premature infants is often empirically prescribed, and practice varies widely among otherwise comparable neonatal intensive care units. Unnecessary and prolonged antibiotic treatment is documented in numerous studies. Recent research shows serious side effects and suggests long-term adverse health effects in prematurely born infants exposed to antibiotics in early life. One preventive measure to reduce unnecessary antibiotic exposure is implementation of antibiotic stewardship programs. Our objective was to review the literature on implemented antibiotic stewardship programs including premature infants with gestational age ≤34 weeks. <b><i>Methods:</i></b> Six academic databases (PubMed [Medline], McMaster PLUS, Cochrane Database of Systematic Reviews, UpToDate, Cochrane Central Register of Controlled Trials, and National Institute for Health and Care Excellence) were systematically searched. PRISMA guidelines were applied. <b><i>Results:</i></b> The search retrieved 1,212 titles of which 12 fitted inclusion criteria (11 observational studies and 1 randomized clinical trial). Included articles were critically appraised. We grouped the articles according to common area of implemented stewardship actions: (1) focus on reducing initiation of antibiotic therapy, (2) focus on shortening duration of antibiotic therapy, (3) various organizational stewardship implementations. The heterogeneity of cohort composition, of implemented actions and of outcome measures made meta-analysis inappropriate. We provide an overview of the reduction in antibiotic use achieved. <b><i>Conclusion:</i></b> Antibiotic stewardship programs can be effective for premature newborns especially when multifactorial and tailored to this population, focusing on reducing initiation or on shortening the duration of antibiotic therapy. Programs without specific measures were less effective.

IntroductionLow microbial biomass and high human DNA content in nasopharyngeal aspirate samples hinder comprehensive characterization of microbiota and resistome. We obtained samples from premature infants, a group with increased risk of developing respiratory disorders and infections, and consequently frequent exposure to antibiotics. Our aim was to devise an optimal protocol for handling nasopharyngeal aspirate samples from premature infants, focusing on host DNA depletion and microbiome and resistome characterization.MethodsThree depletion and three DNA extraction protocols were compared, using RT-PCR and whole metagenome sequencing to determine the efficiency of human DNA removal, taxonomic profiling and assignment of antibiotic resistance genes. Protocols were tested using mock communities, as well as pooled and individual patient samples.ResultsThe only extraction protocol to retrieve the expected DNA yield from mock community samples was based on a lytic method to improve Gram positive recovery (MasterPure™). Host DNA content in non-depleted aliquots from pooled patient samples was 99%. Only samples depleted with MolYsis™ showed satisfactory, but varied reduction in host DNA content, in both pooled and individual patient samples, allowing for microbiome and resistome characterisation (host DNA content from 15% to 98%). Other depletion protocols either retrieved too low total DNA yields, preventing further analysis, or failed to reduce host DNA content. By using Mol_MasterPure protocol on aliquots from pooled patient samples, we increased the number of bacterial reads by 7.6 to 1,725.8-fold compared to non-depleted reference samples. PCR results were indicative of achieved microbial enrichment. Individual patient samples processed with Mol_MasterPure protocol varied greatly in total DNA yield, host DNA content (from 40% to 98%), species and antibiotic resistance gene richness.DiscussionDespite high human DNA and low microbial biomass content in nasopharynx aspirates of preterm infants, we were able to reduce host DNA content to levels compatible with downstream shotgun metagenomic analysis, including bacterial species identification and coverage of antibiotic resistance genes. Whole metagenomic sequencing of microbes colonizing the nasopharynx may contribute to explaining the possible role of airway microbiota in respiratory conditions and reveal carriage of antibiotic resistance genes.

Cell free DNA (cfDNA) in plasma has been described as a potential diagnostic indicator for a variety of clinical conditions, including neonatal hypoxia. Neonatal hypoxia or perinatal asphyxia is a severe medical condition caused by a temporary interruption in oxygen availability during birth. Previously, we have reported temporal changes of cfDNA detected in blood in a newborn piglet model of perinatal asphyxia. However, cfDNA can also be found in other body liquids, opening for a less invasive diagnostic prospective. The objective of this study was to test and establish a reliable method for the isolation and quantification of cfDNA from urine and to explore changes in the quantities of cfDNA using a newborn piglet model of asphyxia. Animals were exposed to hypoxia-reoxygenation (n = 6), hypoxia-reoxygenation + hypothermia (n = 6) or were part of the sham-operated control group (n = 6) and urine samples (n = 18) were collected at 570 minutes post-intervention. Two alternative applications of cfDNA measurement were tested, an indirect method comprising a centrifugation step together with DNA extraction with magnetic beads versus a direct assessment based on two centrifugation steps. CfDNA concentrations were determined by a fluorescent assay using PicoGreen and by qRT-PCR. Genomic (gDNA) and mitochondrial DNA (mtDNA) cfDNA were determined in parallel, taking into account potential differences in the rates of damages caused by oxidative stress. In contrast to previous publications, our results indicate that the direct method is insufficient. Application of the indirect method obtained with the fluorescence assay revealed mean cfDNA levels (SD) of 1.23 (1.76) ng/ml for the hypoxia samples, 4.47 (6.15) ng/ml for the samples exposed to hypoxia + hypothermia and 2.75 (3.62) ng/ml for the control animals. The mean cfDNA levels in piglets exposed to hypoxia + hypothermia revealed significantly higher cfDNA amounts compared to mean cfDNA levels in the samples purely exposed to hypoxia (p < 0.05); however, no significant difference could be determined when compared to the control group (p = 0.09). Application of the indirect method by qRT-PCR revealed mean cfDNA levels of mtDNA and gDNA at the detection limit of the technique and thus no reliable statistics could be performed between the observed cfDNA levels in the investigated groups. The methodology for detection and monitoring of cfDNA in urine has to be further optimized before it can be applied in a clinical setting in the future.