Importance: Patients in neonatal intensive care units (NICU) are at risk of transmission events (TE) by bacteria with multidrug-resistance or with epidemic potential, such as Enterobacterales, P. aeruginosa, S. marcescens, A. baumannii, MSSA and MRSA, which may precede invasive infections ("multidrug-resistant organisms plus", MDRO+). Resolution of MDRO+ transmission clusters (TC) is important to invoke timely and thus effective infection prevention control (IPC) measures. Objectives: Exploration of the potential of timely whole genome sequencing (WGS) in resolving putative MDRO+ transmission chains on a NICU, based on screening-isolates. Exploration of patient- and environmental-specific risk factors of becoming part of a TC. Design, Setting, Participants: Prospective monocentric cohort study at a level III NICU of the University Medical Center Freiburg, Germany. Inclusion of 434 of 551 patients (preterm and term) that were on the NICU for at least 48h and screened at least once in the time between February 15 2019 and November 16 2020. Exposures: Integration of (1) routine culture-based screening for MDRO+ (at admission, then once a week; in accordance with national guidelines), (2) genetic typing with amplified fragment length polymorphism (AFLP) and WGS and (3) granular clinical and staffing data. Statistical analysis of time-dependent risk factors based on multivariate model analysis, using logistic regression and moving averages (MA). Main Outcomes: Primary outcome: Identification of bacterial transmission events (TE) with pathogens of the same species in different patients, which were indistinguishable by AFLP or WGS. Secondary outcomes: MDRO+ colonization rates; identification of TE influencing factors; blood stream infection (BSI) rates. Results: Among 434 NICU patients (27.9% [95% CI, 23.9%-32.3%] with birth weight < 1,500 g), 51.8 % (95% CI, 47.1%-56.5%) were colonized with at least one MDRO+ species; 32.5% (95%CI, 28.3%-37.0%) were colonized as part of a TE as revealed by WGS. With 38 unique TC, E. coli was the most common cluster-forming MDRO+, whereas K. oxytoca formed the largest cluster involving up to 19 patients. Out of ten bloodstream infections, four originated from TE. Multivariate model analysis revealed three key factors for the risk of becoming part of a TC: Increased nurse staffing levels and antibiotic administration lowered the risk of being part of a bacterial transmission cluster, while vascular catheter usage increased it. Conclusions and Relevance: Prospective WGS of routine screening isolates from newborn infants in intensive care is a powerful tool for resolving MDRO+ transmission chains, exceeding AFLP in precision. Despite the associated costs, systematic strain identification by WGS seems justified in high-risk neonates. Both delayed TE identification and "false" TE/TC, which inevitably occur in conventional microbiological screening, have grave organizational consequences. Within model boundaries, we observe factors influencing the risk of becoming part of a bacterial transmission cluster.
