Use of metagenomic next-generation sequencing for diagnosis of peritonitis in end-stage liver disease

Background: Conventional methods are low in positive rates and time-consuming for ascites pathogen detection in patients with end-stage liver disease (ESLD). With many advantages, metagenomic next-generation sequencing (mNGS) may be a good alternative method. However, the related studies are still lacking. Methods: Ascites from 50 ESLD patients were sampled for pathogen detection using mNGS and conventional methods (culture and polymorphonuclear neutrophils detection) in this prospective observational study. Results: Forty-two samples were detected positive using mNGS. 29 strains of bacteria, 11 strains of fungi, and 9 strains of viruses were detected. 46% of patients were detected to be co-infected with 2 or more pathogens by mNGS. Moreover, mNGS showed similar and high positive rates in ESLD patients with different clinical characteristics. Compared to conventional methods, mNGS had higher positivity rates (84% vs. 20%, P<0.001), sensitivity (45.2% vs. 23.8%, P=0.039), broader pathogen spectrum, shorter detection time (24 hours vs. 3-7 days), but lower specificity (25% vs 100%, P = 0.010). Furthermore, compared to conventional methods, mNGS showed similar consistence with final diagnosis (42% vs. 36%, P=0.539). Conclusions: mNGS may be a good supplement for conventional methods and helpful to early etiological diagnosis of peritonitis, and thus improve ESLD patients' survival.