Metagenomic Next-Generation Sequencing Improves Pathogen Detection and Rational Use in Patients with Infectious Fever: A Retrospective Study

Objective : To compare the diagnostic performance of metagenomic next‑generation sequencing (mNGS) and conventional microbiological tests (CMTs) in infectious fever, and to explore optimal specimen types, predictive biomarkers, and clinical scenarios for rational mNGS utilization. Methods: This study retrospectively included 125 patients with confirmed infectious fever who underwent mNGS and CMTs between January 2021 and July 2025. Clinically relevant pathogen detection rates, specimen-specific performance, pathogen spectrum, and clinical impact were analyzed. Procalcitonin (PCT) was evaluated as a predictive marker using ROC analysis. Results: mNGS demonstrated a significantly higher clinically relevant pathogen detection rate than CMTs (56.80% vs. 44.00%, P < 0.05). Stratified by specimen types, mNGS achieved the highest diagnostic yield in pus (95.00%), and was also superior in blood and bronchoalveolar lavage fluid (BALF) (all P <0.05). mNGS indentified a broader pathogen spectrum, covering not only common pathogenic microorganisms but also rare/atypical organisms ( Leptospira interrogans , Coxiella burnetii and Orientia tsutsugamushi ) which were not identified by CMTs. Higher PCT levels were associated with increased mNGS diagnostic yield, and ROC analysis showed moderate predictive performance (AUC = 0.669, 95% CI: 0.572–0.765, P <0.05). Prior antibiotic exposure did not reduce mNGS detection rates. mNGS results were associated with clinical management changes in 55.2% of patients. Conclusion : mNGS significantly improves pathogen detection in infectious fever, especially for atypical/rare pathogens. Pus, blood, and BALF are optimal specimens for testing. mNGS exhibits distinct diagnostic advantages in populations with higher PCT levels and a history of prior antibiotic treatment, promotes rational antibiotic use.