Teams from the Paris Public Hospital Network (AP-HP), Université Paris Cité, Université Sorbonne Paris Nord, Inserm, the Institut Pasteur and FHU Prem'IMPACT, coordinated by Professor Laurent Mandelbrot (Université Paris Cité), Head of the Gynecology and Obstetrics Department at Louis Mourier Hospital (AP-HP), conducted a study to assess whether analyzing bacteria from vaginal swab samples taken at delivery could help predict the risk of neonatal sepsis. The results, published on December 18, 2025 in the American Journal of Obstetrics and Gynecology, show that some identifiable microbial signatures in these samples are associated with a risk of neonatal sepsis.
Early-onset neonatal bacterial infection is a major cause of morbidity and mortality, especially in preterm newborns. It is most often caused by ascending infection via the maternal genital tract. Identifying risk situations with greater precision is a major clinical hurdle in the development of more tailored maternal and neonatal treatments that reduce excessive use of antibiotics.
Current diagnostic tools are mainly based on screening for group B streptococcus (GBS) near the end of pregnancy. While this approach has reduced infection, it fails to account for the complexity of the vaginal microbiota and does not provide an overall picture of infectious risk. This has led to widespread and probabilistic use of antibiotics, with an impact on bacterial resistance and the neonatal microbiota.
In this prospective multicenter study sponsored by the Paris Public Hospital Network (AP-HP), scientists showed that vaginal swab samples taken during delivery can be used to identify microbial signatures associated with a risk of neonatal infection.
The study was based on a prospective cohort of more than 2,500 women treated at three maternity departments in the Greater Paris region (Port-Royal, Louis Mourier and Bichat AP-HP), representing 560 cases of premature rupture of membrane (PROM) before 37 weeks gestation (646 newborns in total because there were some twins). The vaginal swab samples were analyzed using two complementary approaches:
- A conventional bacteriology approach, including bacterial culture identification, targeted molecular testing (PCR) and characterization of antibiotic resistance profiles.
- A metagenomic approach for general untargeted characterization of vaginal bacterial communities.
This integrative strategy showed that the risk of early-onset neonatal sepsis is associated with general imbalances in the vaginal microbiota rather than the presence of a single infectious agent. Risk situations were characterized by a reduction in Lactobacillus dominance, a rise in bacterial diversity and the increased presence of potentially pathogenic bacteria, primarily Escherichia coli, found frequently in both mother and newborn.
By combining the analysis of vaginal microbiota composition with the detection of clinically relevant bacteria, the metagenomic approach was more effective for infectious risk stratification than the usual bacteriological methods, raising the prospect of a more targeted risk analysis for patients with premature rupture of membranes.
These results pave the way for the development of a rapid, non-invasive, multiplex molecular test based on innovative technologies currently under development. This test would improve risk stratification for perinatal infections and encourage a more targeted, rational use of antibiotics. The findings represent a step towards a personalized diagnostic approach incorporating clinical data and microbial signatures.
The results were produced by a consortium set up by the Prem'IMPACT Hospital University Federation (FHU Prem'IMPACT) and the InSPIRe (Innovative Strategies for Perinatal Infectious Risk Reduction) project. This project, which seeks innovative strategies to reduce the risk of perinatal infection, is led by AP-HP, Inserm, the Institut Pasteur and BforCure, with the support of Bpifrance
Sources
Predicting neonatal infection in PPROM with vaginal microbiology and metagenomics: a prospective cohort study, American Journal of Obstetrics and Gynecology, 18 décembre 2025
Laurent Mandelbrot1 , Sean Kennedy 2 , Jessica Rousseau3 , François Goffinet4 , Luce Landraud5 , Céline Plainvert6 , Valérie Marcou7 , Luc Desfrère8 , Tiphaine Barral9 , Lahçene Allal10 , Agnès Baud11 , Nathalie Grall12 , Claire Poyart13 , Pierre-Yves Ancel14 , Asmaa Tazi13
1 Assistance Publique-Hôpitaux de Paris, Department of Obstetrics and Gynecology, Hôpital Louis Mourier, Colombes, France; Université Paris Cité and Université Sorbonne Paris Nord, Inserm, IAME F-75018 Paris, France; FHU Prem'Impact, Paris, France. Electronic address: laurent.mandelbrot@aphp.fr.
2 Computational Biology Department, Institut Pasteur, Université Paris Cité, Paris, France. Electronic address: sean.kennedy@pasteur.fr.
3 Assistance Publique- Hôpitaux de Paris, URC-CIC Paris Descartes Necker/Cochin, Paris, France.
4 FHU Prem'Impact, Paris, France; Université Paris Cité and Université Sorbonne Paris Nord, INSERM, INRAE, Centre de Recherche Epidémiologie et StatistiqueS (CRESS), Obstetrical, Perinatal and Pediatric Life Course Epidemiology (OPPALE Team), Paris, France; Assistance Publique-Hôpitaux de Paris, Department of Obstetrics and Gynecology, Hôpital Cochin-Port Royal, Paris, France.
5 Université Paris Cité and Université Sorbonne Paris Nord, Inserm, IAME F-75018 Paris, France; Assistance Publique-Hôpitaux de Paris, Department of Microbiology, Hôpital Louis Mourier, Colombes, France.
6 Assistance Publique-Hôpitaux de Paris, Department of Bacteriology, National Reference Center for Streptococci, Hôpital Cochin, Paris, France.
7 FHU Prem'Impact, Paris, France; Assistance Publique-Hôpitaux de Paris, Department of Neonatal Medicine, Hôpital Cochin-Port Royal, Paris, France.
8 FHU Prem'Impact, Paris, France; Assistance Publique-Hôpitaux de Paris, Department of Neonatal Medicine, Hôpital Louis Mourier, Colombes, France.
9 Assistance Publique-Hôpitaux de Paris, Department of Obstetrics and Gynecology, Hopital Bichat, Paris, France.
10 Assistance Publique-Hôpitaux de Paris, Neonatal Medicine Unit, Hôpital Bichat, Paris, France.
11 Computational Biology Department, Institut Pasteur, Université Paris Cité, Paris, France.
12 Assistance Publique-Hôpitaux de Paris, Department of Microbiology, Hôpital Bichat, Paris, France.
13 FHU Prem'Impact, Paris, France; Assistance Publique-Hôpitaux de Paris, Department of Bacteriology, National Reference Center for Streptococci, Hôpital Cochin, Paris, France; Inserm U1016, CNRS UMR 8104, Université Paris Cité, Institut Cochin, Paris, France.
14 Assistance Publique- Hôpitaux de Paris, URC-CIC Paris Descartes Necker/Cochin, Paris, France; Université Paris Cité and Université Sorbonne Paris Nord, INSERM, INRAE, Centre de Recherche Epidémiologie et StatistiqueS (CRESS), Obstetrical, Perinatal and Pediatric Life Course Epidemiology (OPPALE Team), Paris, France.
