Metagenomic Next-Generation Sequencing Versus Conventional Microbiological Tests for Pathogen Identification and Prognostic Evaluation in Pediatric Patients with Post-Cardiac Surgery Infections: A Retrospective Cohort Study

Object: Postoperative infection is a severe complication after pediatric cardiac surgery, which is closely associated with sepsis, multiple organ dysfunction, prolonged mechanical ventilation, extended ICU stay and increased mortality. Conventional microbiological tests (CMT) are limited by low sensitivity, long turnaround time, and poor capacity for detecting viruses and polymicrobial infections. This study aimed to compare the diagnostic efficacy of mNGS with that of CMT, and to explore the impact of polymicrobial infection on clinical outcomes in this high-risk pediatric population. Methods: A retrospective cohort study was conducted on 4889 pediatric patients admitted to the PICU after cardiac surgery from January 2025 to March 2026. A total of 510 patients were diagnosed with postoperative infections, including 879 CMT specimens and 86 mNGS specimens enrolled for analysis. Pathogen detection rates, pathogen spectrum and antimicrobial resistance profiles were compared between the two detection methods. Clinical prognostic indicators including mechanical ventilation duration, PICU length of stay and the requirement for continuous renal replacement therapy (CRRT) were further compared between patients with polymicrobial infection and monomicrobial infection. Results: Respiratory tract infection accounted for 87.8% of all postoperative infections, and Gram-negative bacteria were the predominant pathogens, accounting for 65.9%. The overall pathogen detection rate of mNGS was significantly higher than that of CMT (79.1% vs. 56.5%, p < 0.001). Notably, mNGS exhibited significantly better performance in detecting viruses (37.2% vs. 5.8%, p < 0.001), anaerobic pathogens and polymicrobial infections (38.2% vs. 5.4%, p < 0.001). Patients with polymicrobial infections had significantly longer mechanical ventilation time, longer PICU stay, and higher CRRT utilization rate (all p < 0.05), indicating a poorer clinical prognosis. Gram-negative bacteria showed high resistance to penicillins and early-generation cephalosporins, but remained susceptible to carbapenems and β-lactamase inhibitor combination agents. Gram-positive bacteria showed a high resistance rate to penicillin, while maintaining 100% susceptibility to vancomycin and linezolid. Conclusions: mNGS serves as a more sensitive and comprehensive tool for pathogen detection in children with post-cardiac surgery infections, especially for viral and polymicrobial infections. Polymicrobial infection is an independent risk factor for adverse clinical outcomes. Routine application of mNGS in critically ill children may help guide targeted antimicrobial therapy and improve prognosis.