Abstract
Metagenomic next-generation sequencing (mNGS) and droplet digital PCR (ddPCR) have recently demonstrated a great potential for pathogen detection. However, few studies have been undertaken to compare these two nucleic acid detection methods for identifying pathogens in patients with bloodstream infections (BSIs). This prospective study was thus conducted to compare these two methods for diagnostic applications in a clinical setting for critically ill patients with suspected BSIs. Upon suspicion of BSIs, whole blood samples were simultaneously drawn for ddPCR covering 20 common isolated pathogens and four antimicrobial resistance (AMR) genes, mNGS, and blood culture. Then, a head-to-head comparison was performed between ddPCR and mNGS. A total of 60 episodes of suspected BSIs were investigated in 45 critically ill patients, and ddPCR was positive in 50 (83.3%), mNGS in 41 (68.3%, not including viruses), and blood culture in 10 (16.7%) episodes. Of the 10 positive blood cultures, nine were concordantly identified by both mNGS and ddPCR methods. The head-to-head comparison showed that ddPCR was more rapid (~4 h vs. ~2 days) and sensitive (88 vs. 53 detectable pathogens) than mNGS within the detection range of ddPCR, while mNGS detected a broader range of pathogens (126 vs. 88 detectable pathogens, including viruses) than ddPCR. In addition, a total of 17 AMR genes, including 14 bla(KPC) and 3 mecA genes, were exclusively identified by ddPCR. Based on their respective limitations and strengths, the ddPCR method is more useful for rapid detection of common isolated pathogens as well as AMR genes in critically ill patients with suspected BSI, whereas mNGS testing is more appropriate for the diagnosis of BSI where classic microbiological or molecular diagnostic approaches fail to identify causative pathogens.