The Klebsiella pneumoniae outbreak at the US National Institutes of Health's Clinical Center began with one patient in June of 2011. By the following January, it had infected 17 other patients, killing six of them.
With whole-genome sequencing and epidemiological techniques, NIH researchers led by Julie Segre were able to determine the route of K. pneumoniae transmission in real time, helping to contain the outbreak, as they published in Science Translational Medicine in August.
"If [you] get information to hospitals in real time, they are able to use that information to generate hypotheses and test them," Segre says. "It can affect the outcome of other, future patients."
Whole-genome sequencing is increasingly being put to use to study and monitor infectious disease outbreaks. A 2007 outbreak at NIH of Acineobacter baumannii infected 40 patients, and, even after it was under control, investigators there weren't sure what the source — or sources — of the outbreak was. Retrospectively, using whole-genome sequencing, Segre and her team determined that there were two separate introductions of the bacteria into the clinic, a finding they published in PNAS in 2011.
Also last year, University of Münster researchers reported their sequencing of an Escherichia coli strain behind an outbreak in northern Germany — showing that real-time sequencing was feasible — and they then characterized a K. pneumoniae strain that infected patients in a Dutch hospital, information that was used to develop a multiplex PCR-based assay for the disease.
For this latest outbreak, Segre and her colleagues sequenced isolates from the 18 infected patients using a 454/Roche machine. From their Newbler assembly and subsequent analysis, they found that isolates from patient 1 and patient 2 differed by two SNPs out of the 6 megabase K. pneumoniae genome, indicating they were infected with the same strain.
Then the researchers determined that the isolate from patient 3 differed from that of patient 1 by only one SNP, suggesting that patient 1 transmitted the disease to patient 3, who then gave it to patient 2.
Segre recalls telling her colleague in infectious disease control, Tara Palmore, about her suspicions. "[Palmore] said, that's exactly how I think it happened," Segre says. "That was a very freaky moment." Palmore told her that that fit with the epidemiological data: Patient 1 and patient 2 did not overlap in the ICU, but patient 1 and patient 3 did. Then patient 3 and patient 2 overlapped in the ICU.
The researchers similarly tracked down transmission events between other patients, providing information to try to help halt the spread of the disease. Patient 4 was infected by patient 1, though from a separate colonization site than where patient 3 was infected from, a finding indicated by a different SNP pattern.
In the hospital, patient 4 never overlapped with patient 1, but the researchers determined that there were five patients who did overlap with patient 1 and then with patient 4. "When you get that information in real time, we have to step up the contact isolation for those five patients because we now suspect, based on genetic information and epidemiologic data put together, that one of those five patients might be possibly colonized below the level of detection," Segre says.
As the cost of sequencing is dropping, Segre says that many major medical centers could soon implement whole-genome sequencing to monitor disease exposure of patients undergoing extensive procedures like transplants and who are at risk of infection. "If the analysis is that you are spending $100,000 to have an organ transplant or go through some very intensive cancer treatment that involves a few nights' stay in the ICU," she says, "at that point, putting up $500 to sequence an organism that may be affecting another patient in the ICU … I could imagine in the future that that becomes part of the economic model."