The clinical importance of metagenomic next-generation sequencing in detecting disease-causing microorganisms in cases of sepsis acquired in the community or hospital setting

Abstract

Objectives: Although metagenomic next-generation sequencing (mNGS) is commonly used for diagnosing infectious diseases, clinicians face limited options due to the high costs that are not covered by basic medical insurance. The goal of this research is to challenge this bias through a thorough examination and evaluation of the clinical importance of mNGS in precisely identifying pathogenic microorganisms in cases of sepsis acquired in the community or in hospitals.

Methods: A retrospective observational study took place at a tertiary teaching hospital in China from January to December 2021. Data on 308 sepsis patients were collected, and the performance of etiological examination was compared between mNGS and traditional culture method.

Results: Two hundred twenty-nine cases were observed in the community-acquired sepsis (CAS) group and 79 cases in the hospital-acquired sepsis (HAS) group. In comparison with conventional culture, mNGS showed a significantly higher rate of positivity in both the CAS group (88.21% vs. 25.76%, adj.P < 0.001) and the HAS group (87.34% vs. 44.30%, adj.P < 0.001), particularly across various infection sites and specimens, which were not influenced by factors like antibiotic exposure or the timing and frequency of mNGS technology. Sepsis pathogens detected by mNGS were broad, especially viruses, Mycobacterium tuberculosis, and atypical pathogens, with mixed pathogens being common, particularly bacterial-viral co-detection. Based on the optimization of antimicrobial therapy using mNGS, 58 patients underwent antibiotic de-escalation, two patients were switched to antiviral therapy, and 14 patients initiated treatment for tuberculosis, resulting in a reduction in antibiotic overuse but without significant impact on sepsis prognosis. The HAS group exhibited a critical condition, poor prognosis, high medical expenses, and variations in etiology, yet the mNGS results did not result in increased medical costs for either group.

Conclusions: mNGS demonstrates efficacy in identifying multiple pathogens responsible for sepsis, with mixed pathogens of bacteria and viruses being prevalent. Variability in microbiological profiles among different infection setting underscores the importance of clinical vigilance. Therefore, the adoption of mNGS for microbiological diagnosis of sepsis warrants acknowledgment and promotion.

Keywords: community-acquired sepsis; hospital-acquired sepsis; mNGS; medical expenses; microorganisms; optimizing antimicrobial therapy.

Figure 1

Comparison of positivity rates between metagenomic next-generation sequencing (mNGS) and traditional culture methods. (A) The positive rates of mNGS outweigh the culture method in all cases, samples, HAS, and CAS (all adj.P < 0.001). (B) The detection rate of mNGS in different infection locations was higher compared to culture in all patients, particularly in cases of lower respiratory infections, bloodstream infections, central nervous system infections, and abdominal infections (all adj.P < 0.01). (C) In the CAS group, mNGS showed higher positivity rates than culture in lower respiratory infections, bloodstream infections, and central nervous system infection sites were higher than in culture (all adj.P < 0.001). (D) The detection rate of mNGS in the lower respiratory tract was higher than that of culture in the HAS group (adj.P < 0.001). Significance levels: ^**P < 0.01; ^***P < 0.001; ns, no statistically significant variation.

Figure 2

The distribution of infection in CAS group and HAS group. (A) Distribution of infection sites in CAS group. (B) Distribution of infection sites in HAS group. Lower respiratory tract infections were found to be the predominant site of infection, regardless of whether the patients were in the CAS or HAS group.

Figure 3

Comparison of the positive rate of mNGS and traditional culture in different samples. (A–C) In every patient, the detection rate of mNGS in each specimen was notably higher compared to culture, particularly in blood, sputum, BALF, CSF (all adj.P < 0.01). (D–F) Within the CAS cohort, the detection rate of mNGS was significantly higher than that of traditional culture for various sample types including blood, sputum, BALF, and CSF (all adj.P < 0.01). (G–I) In the HAS group, mNGS results showed a higher positive rate in blood and BALF (all adj.P < 0.01). CSF, cerebrospinal fluid; BALF, bronchoalveolar lavage fluid. Significance levels: ^**P < 0.01; ^***P < 0.001; ns, no statistically significant variation.

Figure 4

Comparison of pathogen traits identified by mNGS and conventional culture methods. (A) Comparison of pathogen types and positive rate in all patients (all adj.P < 0.001). (B) Comparison of pathogen types and positive rate in CAS group (all adj.P < 0.001). (C) Comparison of pathogen types and positive rates in the HAS group (all adj.P < 0.001, except for atypical pathogens). (D) mNGS was more common for mixed pathogen infections, while traditional culture was more frequent for single pathogen detection. With the exception of identifying a single pathogen in the HAS group, the positive rate of mNGS significantly surpassed that of traditional culture (all adj.P < 0.05). Significance levels: ^*P < 0.05; ^***P < 0.001; ns, no statistically significant variation.

Figure 5

Distribution of pathogen species detected by mNGS and culture. Tested pathogens are represented on the X-axis by their counts.