Comparison of metagenomic next-generation sequencing and blood culture for diagnosis of bloodstream infections

Abstract

Objectives: This study aimed to evaluate the clinical performance of plasma cell-free DNA (cfDNA) next-generation sequencing (NGS) for pathogen detection in patients with sepsis.

Methods: A total of 43 pairs of blood and plasma samples form 33 blood culture-positive patients were used as testing samples in metagenomic NGS (mNGS) and NGS of 16S ribosomal RNA gene amplicons (16S rRNA NGS). The results of routine tests, including microbial culture, complete blood count, and biochemical tests, were collected from electronic medical records.

Results: Using blood as an mNGS testing sample, the proportion of host DNA was 99.9%, with only three bacteria and no fungi detected. When using plasma in mNGS, the proportion of host DNA was approximately 97%, with 84 bacteria and two fungi detected. Notably, 16S rRNA NGS detected 15 and 16 bacteria in 43 pairs of blood and plasma samples, respectively. Blood culture detected 49 bacteria (23 gram-negative bacilli and 26 gram-positive cocci) and four fungi, with 14 bacteria considered contaminants by clinical microbiologists. For all blood cultures, plasma cfDNA mNGS detected 78.26% (19/23) gram-negative rods, 17% (2/12) gram-positive cocci, and no fungi. Compared to blood cultures, the sensitivity and specificity of plasma cfDNA mNGS for detecting bacteria and fungi were 62.07% and 57.14%, respectively.

Conclusion: Compared to blood, plasma is more suitable for the detection of bloodstream infections using mNGS and is less affected by host DNA. The positive detection rate of plasma cfDNA mNGS for bloodstream infections caused by gram-negative bacteria was higher than that caused by gram-positive cocci.

Keywords: 16S rRNA; bloodstream infections; diagnosis; metagenomic next-generation sequencing; plasma cell-free DNA.

Figure 1

Comparison of the influence of host DNA on detection and identification of pathogens using metagenomic next-generation sequencing (mNGS). The relative abundance and reads of non-host DNA (A, B) and internal controls (C, D).

Figure 2

Comparison of mNGS and 16S rRNA gene next-generation sequencing (16S rRNA NGS) in detection of bloodstream infections using blood and plasma as test samples. Venn diagram analysis of the consistency of the detection results (A: bacteria and fungi; B: viruses) using mNGS and 16S rRNA NGS. Distribution and reads per million (RPM) of microbes (C, D: bacteria and fungi; E, F: virus) detected by mNGS. Distribution (G) and reads (H) of bacteria detected by 16S rRNA NGS. Plasma_mNGS, mNGS using plasma as a test sample. Blood_mNGS, mNGS using blood as a test sample. Plasma_16S, 16S rRNA NGS using plasma as a test sample. Blood_16S, 16S rRNA NGS using blood as a test sample.

Figure 3

Comparison of plasma mNGS and blood culture in the detection of bacteria and fungi. (A) Distribution of blood culture. (B) Composite of monomicrobial and polymicrobial bacteremia after removal of the contaminating strains. (C) The source of bloodstream infections. (D) The influence of the type of pathogen detected by plasma mNGS versus blood culture. (E) Effect of time of positivity of blood culture on plasma mNGS. (F) Number of bottles of blood culture used in plasma mNGS.