Metagenomic analysis of microbial cell-free DNA from whole blood specimens of patients with suspected infections: performance and therapeutic impact in clinical routine. University Hospital Erlangen - DISQVER

Objectives: The sensitivity of blood cultures (BC) is limited especially when antimicrobial therapy has been administered or when non-culturable pathogens are causing the infection. Metagenomic next-generation sequencing (mNGS) of cell-free DNA (cfDNA) from whole blood has the potential to compensate for the disadvantages of BC. Methods: We conducted a prospective study in adult and pediatric patients with suspected infections over a period of 3 months. cfDNA of whole blood was analysed by mNGS (Illumina NextSeq, 25 million reads per sample, read length 75 base pairs) and sequences were analysed with DISQVER®, a CE-IVD-labelled software algorithm and curated database. Simultaneously taken BC and other microbiological results (+/- 7 days) were used for comparison. Results: DISQVER® analysis was performed on 190 samples from 147 patients (23 pediatric, 124 adult). The median time to result was two days (range 2-8). DISQVER® detected 158 pathogens (103 bacteria, 49 viruses, four fungi, two parasites) in 80 whole blood samples (42.1%). The median number of pathogens per positive sample was one (range 1-10). The most common bacteria were Enterobacterales (30.1%), anaerobic bacteria (18.4%) and Enterococcus spp. (15.5%) and the most common viruses were Epstein-Barr virus (28.6%), human herpesvirus 6B (18.4%) and human cytomegalovirus (18.4%). Mycobacterium avium, Legionella pneumophila, Tropheryma whipplei, Rhizomucor pusillus, and Leishmania infantum were detected in one sample each. Simultaneous BC were positive in 11.6%, however, 62.2% of patients received antibiotic therapy at the time of BC sampling. DISQVER® analysis led to a change in treatment in 20 patients (13.6%, ten escalations, eight deescalations, two replacements of central venous catheters). Conclusions: DISQVER® significantly increased the detection rate of pathogens, led to the diagnosis of some serious infections that would otherwise have been missed, and potentially improved treatment in over 10% of patients. A prerequisite for this is that the results must be interpreted carefully.