A Microbial World: Could Metagenomic Next-Generation Sequencing Be Involved in Acute Respiratory Failure?

Abstract

Background: The usefulness of metagenomic next-generation sequencing (mNGS) in identifying pathogens is being investigated. We aimed to compare the power of microbial identification between mNGS and various methods in patients with acute respiratory failure.

Methods: We reviewed 130 patients with respiratory failure, and 184 specimens including blood, bronchoalveolar lavage fluid (BALF), sputum, pleural effusion, ascitic fluid, and urine were tested by mNGS and conventional methods (culture, PCR). We also enrolled 13 patients to evaluate the power of mNGS and pathogen targets NGS (ptNGS) in microbial identifications. Clinical features and microbes detected were analyzed.

Results: mNGS outperformed the conventional method in the positive detection rate of Mycobacterium tuberculosis (MTB) (OR, ∞; 95% CI, 1-∞; P < 0.05), bacteria (OR, 3.7; 95% CI, 2.4-5.8; P < 0.0001), fungi (OR, 4.37; 95% CI, 2.7-7.2; P < 0.0001), mycoplasma (OR, 10.5; 95% CI, 31.8-115; P = 0.005), and virus (OR, ∞; 95% CI, 180.7-∞; P < 0.0001). We showed that 20 patients (28 samples) were detected with Pneumocystis jirovecii (P. jirovecii) by mNGS, but not by the conventional method, and most of those patients were immunocompromised. Read numbers of Klebsiella pneumoniae (K. pneumoniae), Acinetobacter baumannii (A. baumannii), Pseudomonas aeruginosa (P. aeruginosa), P. jirovecii, cytomegalovirus (CMV), and Herpes simplex virus 1 (HSV1) in BALF were higher than those in other sample types, and the read number of Candida albicans (C. albicans) in blood was higher than that in BALF. We found that orotracheal intubation and type 2 diabetes mellitus (T2DM) were associated with a higher detection rate of bacteria and virus by mNGS, immunosuppression was associated with a higher detection rate of fungi and virus by mNGS, and inflammatory markers were associated with mNGS-positive detection rate of bacteria. In addition, we observed preliminary results of ptNGS.

Conclusion: mNGS outperformed the conventional method in the detection of MTB, bacteria, fungi, mycoplasma, and virus. Orotracheal intubation, T2DM, immunosuppression, and inflammatory markers were associated with a higher detection rate of bacteria, fungi, and virus by mNGS. In addition, ptNGS results were consistent with the detection of abundant bacteria, fungi, and mycoplasma in our specimens.

Keywords: acute respiratory failure; conventional methods; mNGS; microbial detection; ptNGS.

Figure 1

Microbes detected by metagenomic next-generation sequencing (mNGS) method in different sample types. (A) Microbes detected in BALF, blood, sputum, pleural effusion, ascitic fluid, and urine samples. (B–D) Comparison of reads number of specific bacteria, fungi, and viruses in BALF, blood, and sputum.

Figure 2

Comparison of positive results by mNGS and conventional methods. (A) Comparison of MTB and bacterial positivity by mNGS and conventional methods. (B) Comparison of G+ and G− bacterial positivity detected by mNGS and conventional methods. (C–E) Comparison of fungal, atypical pathogen, and viral positivity detected by mNGS and conventional methods. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. OR, odds ratio.

Figure 3

Implications of mNGS and conventional methods in different infection types. (A) Percentage of patients with different infection types. (B) The percentage of patients with modified treatment regimes.

Figure 4

Positive detection rate of mNGS associated with clinical characteristics. *P < 0.05, **P < 0.01, ***P < 0.0001.