Skip to main content
Premium Trial:

Request an Annual Quote

Study Details Impact of WGS in Resolving Three-year Outbreak of Opportunistic Pathogen


NEW YORK (GenomeWeb) — Researchers from the University of Warwick have published a study describing their use of whole-genome sequencing to analyze and categorize bacterial isolates to track, and eventually resolve, a lingering outbreak of Acinetobacter baumannii in a UK hospital.

According to the authors, the example illustrates how WGS can provide a "powerful adjunct to conventional laboratory and epidemiological investigation." Sequencing, combined with other epidemiological data, allowed the group to identify the sources of several unexplained transmissions of A. baumannii through the hospital and informed more effective infection-control strategies to reduce the risk of further infections and eventually end the outbreak.

In the study, published today in Genome Medicine, researchers led by the University of Warwick's Mark Pallen became involved in helping to track and resolve the outbreak, which began at the Queen Elizabeth Hospital Birmingham in 2011 and finally ended in February 2013.

Queen Elizabeth is a public hospital that often houses military casualties. According to the authors, A. baumannii is known to be introduced to hospitals by military personnel.

The first case in the outbreak was noted in July 2011, in a patient injured by a blast in Afghanistan. Clinical investigation identified the patient's infecting strain as a multi-drug resistant A. baumannii from the pulse-field gel electrophoresis type, or pulsotype 27, a type not previously seen at the hospital.

Over the following 80 weeks, isolates were recovered from an additional 51 patients, including both military casualties and civilians, and were assigned as matching the outbreak strain based on pulsotype.

According to the authors, WGS added several important pieces to this epidemiological puzzle, and was integral in determining transmission patterns and implementing control measures that eventually ended the outbreak.

In their sequencing efforts, Pallen and his colleagues first created a reference genome for the outbreak, starting by doubly sequencing the genome of an early isolate, from the sixth infected patient, using both the 454 FLX and the Illumina MiSeq to create a hybrid assembly.

Studying this reference, the team concluded that the outbreak strain was distinct from all other well-characterized A. baumannii strains, including all those from previous outbreaks in local hospitals.

The team then attempted to sequence the genomes of an additional 114 isolates, and were successful in obtaining draft sequences from 102 of them. Overall, 52 patient isolates, and 10 environmental samples showed sufficient similarity to the reference strain for the group to rule them into the outbreak.

Importantly, sequencing also revealed that another 18 isolates, also obtained from patients hospitalized at Queen Elizabeth during the outbreak period, did not belong to the outbreak. Interestingly, four of these isolates appeared to make up a second distinct small outbreak group.

In attempting to understand the transmission patters of the outbreak, the study team first looked at how considering only conventional epidemiological information would classify the outbreak. This standard approach identified 273 potential transmission events.

When the team added in its WGS data, it was able to narrow this down to just 57 potential transmissions, which linked all but 10 of the outbreak patients.

Overall, the sequencing effort identified 31 SNVs in the outbreak isolates. This data, coupled with the presence or absence of a cryptic plasmid, defined seven distinct outbreak genotypes, which allowed the team to determine routes of transmission.

The epidemiological and genomic data suggested that in the early period of the outbreak, transmissions were largely caused by infections spreading directly between patients located on the same ward at the same time. All the isolates from the first and second, and many from the fourth genotype, were from patients on a single ward of the hospital.

But, as the outbreak progressed, patients outside this one ward also became infected. The group determined that a single operating theater, which specialized in burn treatment, appeared to be the source of much of this continuing transmission and the theater was closed and deep cleaned.

Unfortunately, after six weeks where no new theater-acquired cases were identified, the outbreak appeared to resume with another burn patient infected with an isolate matching the genotype of those from the operating theater.

Standard epidemiological investigations failed to find any direct ward- or theater-based transmission that linked this new patient with earlier outbreak cases, but the group's WGS data indicated that the strain was a match.

Based on this, the researchers looked harder for a way to explain the transmission, and finally discovered that the patient had been linked to earlier cases, having used a special hospital bed previously occupied by another patient, number 50, who carried a strain of A. baumannii with the same genotype.

After this, another dozen patients were infected over nine weeks, and genomic and epidemiologic studies showed that, again, the burn treatment operating theater was responsible. After a second deep cleaning of the space, no further acquisitions of the strain appeared, and the outbreak was formally declared closed after a period of 12 weeks with no new infections.

Although most of the group's sequencing was retrospective, the authors wrote that in the latter part of the outbreak, analyzing isolates in real time, they achieved a turnaround time of less than one week from colony to determining an SNV-based genotype. This allowed the team to rule patients in or out of the outbreak more quickly than standard methods.

Pallen and his colleagues have also been involved in other efforts to explore and demonstrate the potential of next-gen sequencing in infectious disease diagnosis and epidemiology.

Earlier this year, his team published a study in the open access journal PeerJ demonstrating that direct metagenomic sequencing of DNA in sputum samples can accurately detect TB infection.