OBJECTIVE To investigate the pathogenic characteristics detected by metagenomic next-generation sequencing (mNGS) and its impact on treatment and prognosis in hematological disease patients with febrile neutropenia.

METHODS A retrospective analysis was conducted on 170 hematological disease patients with febrile neutropenia and were ineffective to empirical anti-infective treatment at the Hematology Department of the First Medical Center of the Chinese PLA General Hospital from Apr. 2019 to Dec. 2023. Both blood culture and blood mNGS were simultaneously performed on these patients, and the effects of mNGS and blood culture on pathogen detection rates, pathogen distribution, consistency of test results, the proportion of treatment adjustments based on etiology, and all-cause mortality were compared and analyzed.

RESULTS The detection rate of pathogens of blood culture was 12.94%, with a contamination rate of 4.12%. The pathogen detection rate of mNGS was 25.88%, with a contamination rate of 0.59%. mNGS showed higher detection rates for pathogens, gram-negative bacteria and fungi than blood culture (all P < 0.05). However, the detection rate for gram-positive bacteria by mNGS was slightly lower than that by blood culture, but there was no significant statistical difference (P=0.455). The most common pathogenic bacteria detected by blood culture was coagulase negative Staphylococcus aureus, followed by Escherichia coli. The most common pathogenic bacteria detected by mNGS are Acinetobacter baumannii, followed by Klebsiella pneumoniae. The positive and negative agreement rates for detecting gram-negative bacteria by blood culture and mNGS methods were 72.70% and 88.10%, respectively. The positive and negative agreement rates for detecting gram-positive bacteria were 10.00% and 96.30%, respectively. No fungi were detected consistently by both methods. The viral detection rate by mNGS was 68.82%, while no virus were detected by blood culture. The 30-day all-cause mortality rate of the 170 patients was 22.35%. The improvement rate after treatment adjusted according to mNGS was significantly higher than that after treatment adjusted according to blood culture (48.00% vs. 20.00%, P= 0.007).

CONCLUSIONS For patients with hematological disease accompanied by febrile neutropenia and are ineffective to empirical treatment, mNGS is significantly superior to blood culture in detecting pathogen (gram-negative bacteria, fungi, and viruses) and guiding treatment. In terms of gram-positive bacterial infections, mNGS and blood culture have their own advantages, and the combination of the two methods has greater clinical value.