Hospital Surveillance Sequencing Helps Link Cases to Nationwide Contaminated Eyedrop Outbreak

NEW YORK – A whole-genome sequencing (WGS)-based hospital outbreak surveillance program has helped researchers at the University of Pittsburgh to identify two drug-resistant Pseudomonas aeruginosa infections associated with a national outbreak of contaminated eyedrops.

While the cases were ruled out as in-hospital transmissions, the findings illustrate the potential value of routine hospital WGS surveillance for outbreak detection linked to healthcare-associated infections (HAIs) beyond hospitals.

The two P. aeruginosa outbreak isolates were initially picked up as a potential outbreak cluster in October of last year using a real-time WGS- and machine learning-based healthcare pathogen surveillance program named Enhanced Detection System for Healthcare-Associated Transmission (EDS-HAT), said Alexander Sundermann, an infectious disease expert at the University of Pittsburgh whose team described its findings in a preprint posted on MedRxiv last month.

First implemented in late 2021, the program was developed by Sundermann and his colleagues to help monitor and curb major bacterial pathogens — including P. aeruginosa — commonly linked to HAIs within the University of Pittsburgh Medical Center and its affiliated facilities.

Twice every week, potentially healthcare-associated clinical cultures are collected, and the samples undergo weekly sequencing on the Illumina NextSeq platform as part of EDS-HAT, which is primarily funded by an NIH research project (R01) grant. Sundermann said the program typically can investigate around 1,600 patient infections a year, largely replacing the traditional approach of identifying hospital-associated outbreaks, which can often include chart reviews, staff interviews, and a series of other procedures.

Despite the matching genetic sequences of the two isolates, Sundermann said, an initial investigation found no evidence indicating in-hospital transmission between the two patients. The first patient presented with complaints of an ear infection as an outpatient in early 2022, while the second patient was admitted to the intensive care unit in October 2022 due to multiple medical issues.

“Nothing connected the dots” at that time, Sundermann noted. “We just put it aside and said we would monitor for more.”

In February of this year, then, the Centers for Disease Control and Prevention released a health advisory for a multistate outbreak of Verona Integron-encoded Metallo-β-lactamase (VIM) and Guiana Extended-Spectrum β-Lactamase (GES)-producing carbapenem-resistant P. aeruginosa (VIM-GES-CRPA) associated with the use of artificial tears. The agency subsequently released the sequence type, resistance genes, and reference genome for the outbreak strain.

Upon checking the patient isolates against the reference genome, Sundermann said, the samples in question were “genetically essentially identical” to the outbreak strain, pinning them to the national outbreak of contaminated eyedrops.

As a result, the hospital infection prevention and control department and local public health authorities were immediately notified, triggering further chart reviews and “appropriate interventions aimed at halting transmission” for these patients. According to the paper, an additional chart review for Patient 1 revealed that the patient had been using eyedrops purchased from an online retailer before the presentation of the initial complaint. Meanwhile, further review of Patient 2’s records did not reveal any apparent over-the-counter eye drop use. 

“If we had not collected the sequences and later compared them to a reference genome that they shared, we would have never been able to report those infections,” said Sundermann.

“It is great that [these researchers] are doing this sequencing and that this data is available,” Alexander Greninger, assistant director of the University of Washington Medicine’s clinical virology laboratory who was not involved in the study, wrote in an email. “Most hospitals are not doing this, and they should be commended.”

Greninger said the University of Pittsburgh team’s experience illustrates “the awesome power of genomics” for HAI surveillance, which allows researchers to quickly connect the dots of suspicious samples using informatics tools when a vector, such as eyedrops, is determined. 

At the same time, he said, the study also demonstrates some of the remaining challenges of using the WGS-based system for outbreak detection, given that the researchers were only able to solve the puzzle after the CDC had released the reference genome, which was months after their initial interrogation.

“To truly take advantage of genomics, it would be better if these pathogen genomes were being deposited into public databases in real time,” Greninger pointed out. “If the data was available publicly, the CDC or others could have connected these dots in real time as they were conducting their epidemiological investigation.”

The authors of the preprint also acknowledged that one limitation of the study was that they only solved the cases retrospectively once WGS data from representative outbreak isolates were available. In addition, they could have missed outbreak cases from patients who did not have recent healthcare at the hospital, and they did not study transmission to asymptomatic individuals.

Echoing Greninger’s point, Sundermann said that if systems such as EDS-HAT are implemented at a national scale with open data sharing, they could potentially play an important role in curbing public health outbreaks. “Imagine that multiple hospitals were doing what we are doing, and you had some type of data sharing sources, you could potentially detect these types of outbreaks really early on,” he said.

However, Sundermann said implementing a large network of WGS hospital outbreak surveillance systems with data sharing schemes won’t be “a snap of the fingers” and there will be “a whole hurdle of issues” to overcome. For instance, appropriately sharing genomic data, standardizing analysis pipelines, and sharing protected patient information for outbreak investigations are all logistics to be sorted out, he said.

As the University of Pittsburgh is the only hospital that, to Sundermann’s knowledge, currently employs such a prospective WGS surveillance program for HAIs, his team is keen to demonstrate the cost-effectiveness of EDS-HAT, hoping that professional healthcare societies will implement recommendations and guidelines for this type of program.

“I think what we’re trying to build up as the evidence base is that real-time sequencing surveillance is feasible,” Sundermann said. “It is able to guide your infection prevention interventions, and it detects all these outbreaks that otherwise go undetected.”