In June of last year, a 43-year old woman was admitted to the Clinical Center of the National Institutes of Health in Bethesda, Maryland, for a lung disease. Doctors knew she was carrying a highly resistant form of a deadly bacterium known as Klebsiella pneumoniae—although it didn't make her sick—and they placed her in isolation. When the woman was discharged, no one else appeared to have become infected. A few weeks later, however, another patient was found to be carrying the bacterium, and over the next 3 months, 12 more intensive care patients contracted it. Six died as a direct result of the infection.
Doctors could not make sense of the outbreak with the usual methods: A survey of bed locations showed that the first patient had had no direct contact with any of the others and, in theory, Klebsiella might have been introduced into the hospital multiple times. So physicians turned to the bacterium's genome for answers. The approach, known as genomic epidemiology, helped them track the path of the microbe, contain the disease, and save lives, according to a new study.
Genomic epidemiology makes use of the fact that when bacteria divide, they accumulate mutations. As a result, the bacterial genome differs slightly—often by just one or two letters of genetic code, or base pairs—from one patient to the next. By fully sequencing the genomes of patients' bacteria and finding these minute differences, researchers can track microbial movements with unprecedented precision. The technique has already been used to reconstruct the spread of methicillin-resistant Staphylococcus aureus (MRSA)  around the world and to pinpoint the origin of a cholera outbreak in Haiti .
It also helped the doctors at the hospital in Bethesda. Comparing the genomes from all patients showed that the female patient admitted in June had indeed initiated the outbreak; the researchers showed that the bacteria had been transmitted from her to other patients three times independently. Apparently, transmission occurred in ways the researchers didn't understand, says Tara Palmore, an infectious disease physician at the hospital. "When we realized there was more than met the eye, we started testing everyone in the hospital," she says. That helped identify four more infected patients outside the intensive care unit , the scientists report online today in Science Translational Medicine. They were quickly isolated, which Palmore believes prevented further spread.
Just how the microbes were transmitted is still unclear. Palmore assumes that the bacteria mainly traveled on the hands of doctors. But the clinic had stationed a person outside the isolation rooms to make sure everyone who entered followed a hygiene regimen 24/7. That suggests that bacteria might have established colonies on surfaces or medical equipment and spread that way as well. "The conventional wisdom is that Klebsiellas do not really survive in the environment, but we found them in six sink drains and a ventilator," Palmore says.
"This small study demonstrates the potential power of whole genome sequencing for outbreak investigation and surveillance," says Sharon Peacock, a microbiologist at the University of Cambridge in the United Kingdom who was not involved in the work. And infectious disease specialist Dag Harmsen of the University Clinic of Münster in Germany says it is "further proof that the time is ripe for using genomic sequencing of pathogens in a hospital setting." The paper also highlights the dangers of resistant Gram-negative bacteria like Klebsiella p., he adds. In many patients, the bacteria were not susceptible to any available antibiotic; not even to colistin, an old compound used only when all else fails. "This is even more dramatic than MRSA, because you have nothing left to treat the patients with," Harmsen says. Since the outbreak, every patient at the hospital is checked for such dangerous pathogens; one more resistant Klebsiella case—although a different strain—has been found so far.
Genomic epidemiology could make it easier for hospitals to deal with similar outbreaks, Palmore says. "A lot of academic centers have the ability to do this now," she says. The cost is becoming less of an issue; during last year's outbreak, scientists still paid about $2000 per genome sequenced; now that would be closer to $500. But Peacock cautions that it still takes bioinformatics specialists several weeks to interpret the data. "This technology will not be applicable to routine clinical practice until automated interpretation tools become available."