Comparison of Blood-Based Shotgun and Targeted Metagenomic Sequencing for Microbiological Diagnosis of Infective Endocarditis

Background Shotgun and targeted metagenomic sequencing have been shown in separate studies to be potentially useful for culture-free pathogen identification in blood and/or plasma of patients with infective endocarditis (IE). However, the 2 approaches have not been directly compared. The aim of this study was to compare shotgun metagenomic sequencing with targeted metagenomic sequencing (tMGS) for organism identification in blood or plasma of patients with IE.Methods Patients with possible or definite IE were prospectively enrolled from October 2020 to July 2021. Shotgun metagenomic sequencing was performed with the Karius test, which uses microbial cell-free DNA (mcfDNA) sequencing to detect, identify, and quantitate DNA-based pathogens in plasma. tMGS was performed using a 16S ribosomal RNA (rRNA) polymerase chain reaction assay targeting the V1 to V3 regions of the 16S rRNA gene. Results were compared using the McNemar test of paired proportions.Results Samples from 34 patients were investigated. The Karius test was positive in 24/34 (71%), including 3/6 (50%) with blood culture-negative endocarditis (BCNE), which was not significantly different from the positivity rate of tMGS (P = .41). Results of the Karius test were concordant with tMGS in 75% of cases. The Karius test detected 2 cases of methicillin-resistant Staphylococcus aureus among the 7 S. aureus detections, in accordance with results of phenotypic susceptibility testing. The combination of blood cultures, the Karius test, and tMGS found a potential causative pathogen in 33/34 (97%), including 5/6 with BCNE.Conclusions The Karius test and tMGS yielded comparable detection rates; however, beyond organism identification, the Karius test generated potentially useful antibiotic resistance data. The Karius test (sMGS) and tMGS had similar positivity rate in blood samples from 34 patients with IE. Overall, positivity of both tests and blood cultures combined was 97%. Either method has potential in routine practice for culture-free diagnostic.