Using a new bioinformatics tool that can rapidly identify pathogens in a patient's gut microbiome, researchers have found evidence that a patient's own microbiome may be the source of hospital-acquired bloodstream infections (BSIs).
Fiona B. Tamburini, a graduate student, and Tessa Andermann, MD, MPH, an infectious disease fellow, both in the laboratory of Ami S. Bhatt, MD, PhD, at Stanford University, developed the new tool, called StrainSifter.
They then used it to demonstrate that the gut microbiome may serve as a reservoir for BSI, including infections by bacterial species traditionally considered nonenteric. Typically, enteric microorganisms enter the bloodstream through damaged gut lining, and environmental or nonenteric bacteria enter through skin breaks or through intravenous lines.
"With these powerful genomic tools, we anticipate that precision source identification and strain tracking will lead us to a new, sharpened model of infectious disease," the researchers write in an article published in Nature Medicine.
StrainSifter compares variable parts of genome sequences from bacterial isolates from the bloodstream with samples from microbiomes, such as the community living in the human gut. Tracking these single nucleotide variants (SNVs) distinguishes strains of a bacterial species — hence the name StrainSifter. Genomic epidemiologists compare strains to trace outbreaks in healthcare facilities.
In a retrospective cohort study, the researchers collected weekly stool samples from 30 patients undergoing hematopoietic cell transplant (autologous or allogeneic) at Stanford University Hospital between October 5, 2015, and June 9, 2017. Participants had provided a stool sample within 30 days prior to developing a BSI and at least one sample after an infection met standard BSI criteria.
First, the investigators sequenced the genome of the infectious agent and the genomes of all of the bacteria in the stool samples from before and after onset of the BSI, obtaining a snapshot of which bacteria in a patient's bloodstream were also in the gut.
Of 32 infectious agents sequenced, 15 were also found in the patient's gut microbiome (47%), at an abundance threshold equal to or greater than 0.1% of the total microbiota in the gut. Ten of these 15 are traditionally classified as enteric (eight intestinal and two oral).
Not all of the BSI species were abundant in the stool prior to diagnosis. Escherichia coli and Enterococcus faecium were the dominant species in two cases; Klebsiella pneumoniae and Enterobacter cloacae were present at 2.8% and 0.6% relative abundance, respectively.
Yet for some patients, potential pathogens abundant in their stool did not match the bacterial species in their bloodstreams. For example, one patient had a Klebsiella pneumoniae BSI, but a stool sample 9 days before infection revealed E coli at 64% relative abundance, and 19 days after infection E faecium at 82% relative abundance.
StrainSifter enabled the investigators to drill down to the strain level, yielding information that may provide more clues to the crosstalk between the gut microbiome and bacteria in the bloodstream. The approach showed that strains of the same microbial species within a patient were more alike than strains between patients.
None of the 30 patients had enough Staphylococcus aureus in their stool to register with StrainSifter. However, strains of two nonenteric species, Staphylococcus epidermidis and Pseudomonas aeruginosa, were each identified in the gut and the bloodstream samples from a patient.
The team then used genomic analysis to try to determine whether a patient's infection came from their microbiome or the environment. They found that the S epidermidis strains differed by just one SNV in one patient, and in another patient, P aeruginosa was identical in stool and blood specimens; both cases suggest the BSI stemmed from the patient's own microbiome.
In addition, the finding of S epidermidis in the gut suggests that bloodstream infections with this pathogen aren't only from placement of intravenous lines but could come from the patient's gut microbiome, the researchers write.
"Our findings underscore the importance of looking past the conventional assumptions of infectious source. The conventional source is not necessarily always the correct one," Tamburini said.
The potential route of infection from gut to bloodstream that the study reveals also suggests "the compelling possibility that altering the gut microbiome may prevent BSIs in hospitalized patients. We may need to start thinking more about how antibiotics and other medications affect the complex microbial ecosystem that lives inside us," Andermann said.
Although an analysis can take just hours and the software is easily accessed (https://github.com/bhattlab/strainsifter), the StrainSifter approach must still fit into the clinical scenario. "When a patient presents with neutropenic fever, empiric antibiotics are started immediately. We can generate sequencing data very quickly, but we're not at the point yet where the timescale is clinically relevant," said Tamburini.
Added Bhatt, "As advances in microbial measurement technologies, including sequencing, take place, software tools like StrainSifter may be increasingly useful in clinical settings." In the meantime, before altering infection prevention or treatment strategies, larger studies using the tool with many more patients are necessary, the researchers told Medscape Medical News.
A limitation of StrainSifter is that it can only identify the dominant strain of a bacterial species in a community. However, it can be used on any metagenome source.
The researchers have disclosed no relevant financial relationships.
Nat Med. Published online October 15, 2018. Abstract
Medscape Medical News © 2018
Cite this: Gut Microbiome May Seed Sepsis - Medscape - Nov 12, 2018.