The current culture-based method for diagnosing neonatal sepsis is slow (2–5 days) and often fails to detect anaerobic, fastidious, and non-culturable pathogens, leading to diagnostic inaccuracies. To address this, the study proposes a novel Next-Generation Sequencing (NGS) approach for real-time, long-read sequencing to directly identify bacterial and fungal pathogens and antimicrobial resistance genes (ARGs) from blood samples. This method aims to enhance diagnostic speed and accuracy. 

The study has three objectives: optimizing DNA extraction from small blood samples with low pathogen loads, comparing NGS-based pathogen detection with culture methods, and validating NGS accuracy using culture-positive samples. It will progress in three phases: laboratory standardization, a pilot study with culture-positive samples, and a multicentric diagnostic accuracy study in neonates with suspected sepsis.

Assuming an 18% prevalence of culture-positive sepsis (DeNIS study), a sample size of 1,046 clinically suspected cases is required to detect a sensitivity of 98% ± 2% and a specificity of 98% ± 1%, including 189 culture-positive cases. Over 2.5 years, four centers will recruit ~105 subjects annually, performing ~3,118 NGS and culture tests in total. The expected outcome is a real-time, NGS-based diagnostic method that directly detects bacterial and fungal pathogens and assesses ARGs, offering significant improvements in speed and accuracy over traditional culture-based methods for neonatal sepsis.