Development and proof-of-concept demonstration of a clinical metagenomics method for the rapid detection of bloodstream infection.

BACKGROUND: The timely and accurate diagnosis of bloodstream infection (BSI) is critical for patient management. With longstanding challenges for routine blood culture, metagenomics is a promising approach to rapidly provide sequence-based detection and characterisation of bloodborne bacteria. Long-read sequencing technologies have successfully supported the use of clinical metagenomics for syndromes such as respiratory illness, and modified approaches may address two requisite factors for metagenomics to be used as a BSI diagnostic: depletion of the high level of host DNA to then detect the low abundance of microbes in blood. METHODS: Blood samples from healthy donors were spiked with different concentrations of four prevalent causative species of BSI. All samples were then subjected to a modified saponin-based host DNA depletion protocol and optimised DNA extraction, whole genome amplification and debranching steps in preparation for sequencing, followed by bioinformatical analyses. Two related variants of the protocol are presented: 1mL of blood processed without bacterial enrichment, and 5mL of blood processed following a rapid bacterial enrichment protocol-SepsiPURE. RESULTS: After first identifying that a large proportion of host mitochondrial DNA remained, the host depletion process was optimised by increasing saponin concentration to 3% and scaling the reaction to allow more sample volume. Compared to non-depleted controls, the 3% saponin-based depletion protocol reduced the presence of host chromosomal and mitochondrial DNA < 106 and < 103 fold respectively. When the modified depletion method was further combined with a rapid bacterial enrichment method (SepsiPURE; with 5mL blood samples) the depletion of mitochondrial DNA improved by a further > 10X while also increasing detectable bacteria by > 10X. Parameters during DNA extraction, whole genome amplification and long-read sequencing were also adjusted, and subsequently amplicons were detected for each input bacterial species at each of the spiked concentrations, ranging from 50-100 colony forming units (CFU)/mL to 1-5 CFU/mL. CONCLUSION: In this proof-of-concept study, four prevalent BSI causative species were detected in under 12 h to species level (with antimicrobial resistance determinants) at concentrations relevant to clinical blood samples. The use of a rapid and precise metagenomic protocols has the potential to advance the diagnosis of BSI.

Rapid Rule Out of Culture-Negative Bloodstream Infections by Use of a Novel Approach to Universal Detection of Bacteria and Fungi.

BACKGROUND: Currently it can take up to 5 days to rule out bloodstream infection. With the low yield of blood cultures (approximately 10%), a significant number of patients are potentially exposed to inappropriate therapy that can lead to adverse events. More rapid rule out can accelerate deescalation or cessation of antimicrobial therapy, improving patient outcomes. METHODS: A method is described, termed enzymatic template generation and amplification (ETGA), that universally and sensitively detects DNA polymerase activity liberated from viable bacteria and fungi isolated from blood culture samples as a measure of bloodstream infection. ETGA was applied in a diagnostic test format to identify negative blood cultures after an overnight incubation. Performance data for a prototype (Cognitor) and automated (Magnitor) version of the test are presented. RESULTS: The Cognitor manual assay displayed analytical reactivity for a panel of the 20 most prevalent causes of bloodstream infection, with a detection range of 28-9050 CFU/mL. Validation with 1457 clinical blood cultures showed a negative predictive value of 99.0% compared to blood culture incubation for 5 days. Magnitor showed an improved detection range of 1-67 CFU/mL, allowing for detection of bacteria-supplemented blood cultures after 2-8 h incubation, and Candida albicans-supplemented blood cultures at 16-22 h, 5-15 h faster than blood culture. Removing an aliquot from a blood culture bottle and replacing the bottle into the incubator was shown not to result in contaminating organisms being introduced. CONCLUSIONS: The described method displays excellent breadth and detection for microbial cells and demonstrates the capability of confirming negative blood cultures after an overnight incubation in a blood culture instrument.