Impact of metagenomics next-generation sequencing on etiological diagnosis and early outcomes in sepsis.

Summary

In a 19-site prospective cohort (n=859), mNGS on infected site samples had higher positive agreement than conventional microbiological testing (92.0% vs 51.1%) but lower negative agreement. It identified causal microbes in 74% and led to antibiotic adjustments in 29.2%. Early mortality (7-day) was reduced (HR 0.44), while 28-day mortality was unchanged.

Key Findings

• Positive percent agreement of mNGS vs CMT: 92.0% vs 51.1% (p<0.001); negative agreement lower for mNGS (39.6% vs 69.2%).
• Causal microbes identified in 74.0% with mNGS; antibiotic changes in 29.2%.
• 7-day mortality reduced (HR 0.44), but no difference in 28-day mortality (HR 0.82).

Clinical Implications

Adjunctive mNGS of infected site samples can accelerate etiological diagnosis and guide targeted therapy, potentially reducing early mortality. Programs should address false negatives/positives, turnaround time, and stewardship oversight.

Limitations

• Nonrandomized, patient-choice design risks residual confounding despite IPTW
• Lower negative agreement and potential contamination limit specificity; conducted in one country

Future Directions

Randomized or pragmatic trials to test mNGS-guided care pathways on 28-day mortality, time-to-appropriate therapy, and costs; harmonization of interpretation frameworks and contamination control.