NEW YORK – Researchers from the Wellcome Sanger Institute, University of Oslo, and elsewhere have developed a new whole-genome sequencing-based approach that promises to identify multiple common pathogens linked to hospital-associated infections (HAIs) in one assay.
Published in Lancet Microbe last month, the study demonstrates the potential utility of culture-based, pan-pathogen WGS for HAI surveillance and prevention, though more studies are still needed to further validate the method's cost, accuracy, and implementation feasibility.
Traditionally, sequencing-based pathogen surveillance methods rely on analyzing single-colony bacterial isolates. While this approach can offer greater information on the target bacteria strains, it fails to account for the diversity of microorganisms present within a clinical sample, said Harry Thorpe, a researcher from the University of Oslo and the first author of the study.
On the other hand, shotgun metagenomic sequencing allows researchers to analyze all microbial diversity within a sample, Thorpe noted, but it generally cannot offer adequate sequencing depth to achieve strain-level resolution for epidemiology and transmission analysis.
"We wanted to design a method that is intermediate between the single colony method and the metagenomics method," he added.
The method developed by Thorpe and his team starts with sample collection from patients. For the Lancet Microbe paper, they retrospectively collected samples from 256 patients in an Italian hospital between April 7 and May 10, 2020. Of the participants, 169 were men and 87 were women. About half of the patients were hospitalized in the ordinary wards, while the other half were admitted in the ICUs.
From these patients, the researchers collected 1,130 clinical samples, which included 497 rectal swabs, 439 nasal swabs, and 194 respiratory samples to represent bacteria in the gut, upper airways, and lungs. The collected samples were cultured using six different media to select various pathogen species of interest. Instead of only picking one colony as with traditional methods, they collected DNA en masse from each plate with positive growth, Thorpe said.
After DNA extraction and library preparation, the samples then underwent deep whole-genome sequencing using the Illumina NovaSeq 6000 platform with the 150 bp paired-end chemistry. The sequencing data were subsequently aligned against the reference datasets to identify the organisms present in each of the samples and their abundance.
Overall, the analysis identified 52 different bacteria species present in the patient samples. Of these species, 66 percent comprised various strains of the seven most common bacterial infections seen in hospitals: Acinetobacter baumannii, Escherichia coli, Enterococcus faecium, Enterococcus faecalis, Klebsiella pneumoniae, Pseudomonas aeruginosa, and Staphylococcus aureus.
The authors further investigated the association between each pathogen species and the ward type. There, the results indicated that A. baumannii, K. pneumoniae, P. aeruginosa, and E. faecium were "significantly associated" with ICU patients for both nasal and rectal samples, while E. faecalis was associated with ICU nasal samples. In the ordinary wards, E. coli was mostly associated with rectal samples.
To extrapolate possible hospital-associated transmissions, the researchers conducted pairwise bacterial single-nucleotide polymorphism distance analysis and concluded that HAIs were "likely to be a significant mode of acquisition" for each of the seven pathogen species identified in the study. That is especially likely for A. baumannii, K. pneumoniae, P. aeruginosa, and E. faecium, for which the authors believe that more than 75 percent of occurrences can be attributed to potential hospital transmission.
Beyond transmission events, Thorpe's team also leveraged the whole-genome data to look for common antimicrobial-resistant (AMR) genes in the samples. From there, they found that the prevalence of AMR genes was "very high" in ICU samples, with key carbapenemase and extended spectrum ß-lactamase resistance genes detected in over 40 percent of the critical care patients.
"The approach described in the paper is pretty exciting if the methods can hold true and are accurate," said Alexander Sundermann, an infectious disease professor at the University of Pittsburgh who was not involved in the study.
Echoing the study authors' points, Sundermann said one of the potential limitations of current sequencing-based pathogen surveillance methods is that they typically focus on a single colony. As such, researchers might be under-sampling the microbial diversity in the patient and, therefore, cannot fully capture all the transmission events. The newly described method, on the other hand, enables researchers to have a better grasp of the microorganism diversity within a patient, he added.
"I would characterize this method as kind of like a Goldilocks — not too hot, not too cold, just right," agreed Daria Van Tyne, another infectious disease professor at Pitt who was not involved in the study. In collaboration with Sundermann and others, Van Tyne has been leading a real-time WGS surveillance program at Pitt to help identify and curb HAIs.
While the team has been mostly doing single-isolate sequencing, Van Tyne said, she and collaborators are also exploring adopting a similar WGS strategy to capture greater microbial diversity in their studies. As such, Van Tyne said her team is interested in giving "a test drive" of the method described in the paper.
Despite the method's promises, Van Tyne said one limitation of the paper is that the study authors did not go beyond the genomic analysis to investigate if there was a true epidemiologic link among the patients suspected of HAIs. "Genomic analysis can simply generate a hypothesis," she noted. "That hypothesis still would need to be tested before it could be proven."
Additionally, given that the method is a reference-based approach, it could hinder data reproducibility if people are using different reference databases. Moreover, as WGS is still not widely adopted within the healthcare setting for infection surveillance, Van Tyne said a general hurdle for all sequencing-based methods is to prove their feasibility and boost their implementation.
Calling the current study a "proof of concept," Thorpe said that while the method developed by his team demonstrated some potential utility, there is still a ways to go before it can be deployed in hospitals. As such, he noted some of the future directions for the team are to further validate the method's accuracy and to analyze the method's cost and implementation, which the current study did not explore.
"We're definitely not saying that this could be implemented tomorrow," Thorpe said. "I think the challenge is still to get institutions to adopt sequencing-based surveillance in general."