Metagenomic Next-Generation Sequencing Versus Traditional Laboratory Methods for the Diagnosis and Treatment of Infection in Liver Transplantation

BACKGROUND: Metagenomic next-generation sequencing (mNGS) has emerged as an effective method for the noninvasive and precise detection of infectious pathogens. However, data are lacking on whether mNGS analyses could be used for the diagnosis and treatment of infection during the perioperative period in patients undergoing liver transplantation (LT). METHODS: From February 2018 to October 2018, we conducted an exploratory study using mNGS and traditional laboratory methods (TMs), including culture, serologic assays, and nucleic acid testing, for pathogen detection in 42 pairs of cadaveric liver donors and their corresponding recipients. Method performance in determining the presence of perioperative infection and guiding subsequent clinical decisions was compared between mNGS and TMs. RESULTS: The percentage of liver donors with mNGS-positive pathogen results (64.3%, 27/42) was significantly higher than that using TMs (28.6%, 12/42; P<0.05). The percentage of co-infection detected by mNGS in liver donors was 23.8% (10/42) significantly higher than 0.0% (0/42) by TMs (P<0.01). Forty-three pathogens were detected using mNGS, while only 12 pathogens were identified using TMs. The results of the mNGS analyses were consistent with results of the TM analyses in 91.7% (11/12) of donor samples at the species level, while mNGS could be used to detect pathogens in 66.7% (20/30) of donors deemed pathogen-negative using TMs. Identical pathogens were detected in 6 cases of donors and recipients by mNGS, among which 4 cases were finally confirmed as donor-derived infections (DDIs). For TMs, identical pathogens were detected in only 2 cases. Furthermore, 8 recipients developed early symptoms of infection (<7 days) after LT; we adjusted the type of antibiotics and/or discontinued immunosuppressants according to the mNGS results. Of the 8 patients with infections, 7 recipients recovered, and 1 patient died of severe sepsis. CONCLUSIONS: Our preliminary results show that mNGS analyses can provide rapid and precise pathogen detection compared with TMs in a variety of clinical samples from patients undergoing LT. Combined with symptoms of clinical infection, mNGS showed superior advantages over TMs for the early identification and assistance in clinical decision-making for DDIs. mNGS results were critical for the management of perioperative infection in patients undergoing LT.