Performance of mNGS for Bloodstream Infections in Perioperative Critically Ill Patients

Monday, March 23, 2026

11:30 AM - 12:30 PM Central Time

Location: Connections Central - RST 05

Qiao Qin, N/A (she/her/hers)

Beijing, CHN

Disclosure(s): No relevant financial relationship(s) to disclose.
YN

Ya-Chan Ning, N/A

Associate Attending Physician, Department of Critical Care Medicine
Xuanwu Hospital, Capital Medical University

Disclosure information not submitted.
SZ

Sai-Nan Zhu, n/a

Associate Professor, Department of Biostatistics
Peking University First Hospital

Disclosure information not submitted.
JM

Jia-hui Ma, N/A

Associate Researcher, Department of Anesthesiology
Peking University First Hospital

Disclosure information not submitted.
WT

Wei Tian, n/a

Senior Attending Physician, Department of Geriatrics
Beijing Jishuitan Hospital, Capital Medical University

Disclosure information not submitted.
WC

Wei Chen, n/a

Senior Attending Physician, Department of Critical Care Unit
Beijing Shijitan Hospital, Capital Medical University

Disclosure information not submitted.
CW

Chun-Mei Wang, n/a

Senior Attending Physician ,Department of Intensive Care Medicine
Xuanwu Hospital, Capital Medical University

Disclosure information not submitted.
YW

Ying-Feng Wu, n/a

Senior Attending Physician, Department of Critical Care Medicine
Xuanwu Hospital, Capital Medical University

Disclosure information not submitted.
SL

Shuang-Ling Li, n/a

Senior Attending Physician, Department of Critical Care Medicine
Peking University First Hospital

Disclosure information not submitted.

Introduction: Bloodstream infections (BSI) in intensive care unit (ICU) patients are associated with high morbidity and mortality, necessitating rapid and accurate pathogen identification to guide early antimicrobial therapy. However, traditional blood culture (BC) is limited by the long turnaround time and low sensitivity. Metagenomic next-generation sequencing (mNGS) has been applied in infectious disease diagnostics, but its clinical utility for perioperative ICU patients with BSI requires further evaluation.

Methods: This prospective, multi-center diagnostic study enrolled perioperative ICU patients suspected of BSIs from July 2020 to June 2023. Blood samples were collected for mNGS and BC detection simultaneously. The study compared pathogen detection differences between the two methods, and evaluated the diagnostic value of mNGS for clinical BSI based on mNGS-assisted clinical diagnostic criteria. Additionally, the impact of mNGS findings on clinical antimicrobial management was assessed.

Results: mNGS demonstrated a higher overall pathogen detection rate than BC in the 219 enrolled patients (25.6% vs. 9.6%, p < 0.001), with significant advantages in detecting Gram-negative bacteria (13.2% vs. 5.9%, p = 0.009), anaerobes (3.6% vs. 0.5%, p = 0.018), and fungi (6.4% vs. 0.9%, p = 0.002). Mixed-pathogen infections were identified in 20% of mNGS-positive clinical BSI cases, whereas BC-positive cases exclusively had single-pathogen infections. Ultimately, 64 patients (29.2%) were diagnosed with clinical BSIs. The sensitivity and specificity of the mNGS were 85.9% (95% CI: 74.5%–93.0%), and 80.6% (95% CI: 73.4%–86.4%), respectively, and the area under the receiver operating characteristic curve was 0.833 (95% CI: 0.772–0.894). Additionally, mNGS led to a positive impact in 56 patients (25.6%), manifested by the identification of new pathogens and guidance for targeted therapy, a negative impact in 11 patients (5.0%), and no clinical impact in 152 patients (69.4%).

Conclusions: For perioperative ICU patients, mNGS demonstrated superior pathogen detection rates, broader microbial spectrum coverage, and enhanced polymicrobial infection detection capability versus BC. mNGS exhibited high diagnostic value for clinical BSI, with the potential to facilitate targeted antimicrobial therapy adjustments.