Then, as abruptly as it began, the outbreak stopped. On Dec. 14, a 37-year-old woman with sickle cell disease was the last patient for the next seven months to test positive for the KPC originating from Patient One.
By then, the hospital had treated 18 patients with KPC. Six had died from a bloodstream infection caused by KPC; five had died from their underlying disease.
By January 2012, the hospital was finally out of crisis mode. That’s when Palmore, Segre and Snitkin were able to combine the DNA technology with the epidemiological information concerning patient overlap to pinpoint the exact transmission of the outbreak.
By sequencing each patient’s bacteria, Segre was able to establish a “fingerprint” of the germ unique to each patient. She could specify from which part of the body the germ spread, which proved important in understanding the outbreak. Bacteria reproduce by cloning themselves. A whole generation of bacteria reproduce about every hour. As they reproduce, there is one change a week in the lettering of the DNA, providing scientists with a sort of time stamp.
When the New York patient entered the hospital, samples of bacteria were taken from her throat, groin and lung. Each strain had a unique genetic code that enabled researchers to pinpoint the location in the body from which it spread. The bacteria from each patient with KPC thereafter were compared to Patient One’s strains.
Snitkin wrote a program to match the details about each patient’s bacteria with the person’s whereabouts during his or her stay.
“It was really hardcore computer science,” says Snitkin, who grew up outside New York City in Westchester, N.Y., and graduated in 2002 from the State University of New York at Binghamton, where he majored in biology and computer science.
The data showed that the transmission didn’t follow a traditional epidemiological path, where the first patient gives it to the second patient, who gets sick and gives it to the third patient.
Instead, the New York woman, Patient One, had infected Patient Three, whose stay in the ICU overlapped with Patient One. Patient Three gave it to Patient Five, who gave it to Patient Two, but Patient Two got worse faster than Patients Three or Five, highlighting the fact that the bacteria can exist in a patient longer than anticipated before the person shows signs of illness.
“Figuring this out was like playing Sudoku or a classic math puzzle,” Segre says.
The bacteria had spread in three clusters. KPC from Patient One’s throat spread to Patient Three, who passed it to Patient Five and then to Patient Two. In another cluster, bacteria from Patient One’s lung spread to Patient Four, who then spread it to every other patient except one. In a third case, bacteria from Patient One’s lung spread to Patient Eight, but went no farther.
The original data also showed that Patient One gave it to Patient Four even though they were in different wards and were never in the ICU at the same time. That likely meant a silent carrier was the link between Patients One and Four.
Back in late September 2011, after Palmore learned that Patients One and Four were connected, she had looked for patients who could have been silent carriers. But none of the potential candidates tested positive in surveillance tests.
“Seeing detailed information like this [faster] can help you target infection control,” Snitkin says.
Still, there were gaps in understanding the transmission routes. Take the case of Patient Eight. The 71-year-old male lymphoma patient was never in the ICU or on the same ward as Patient One or the others, and his bacterial strain didn’t match that of any other patient except One’s. He tested positive for KPC almost two months after the New York patient left the hospital, though he was never sickened by it. Palmore and Segre were unable to solve the mystery.
Then, to everyone’s shock, the bug returned.
A bone marrow transplant patient in his 20s tested positive for KPC in July 2012. He failed to respond to colistin and died in September.
DNA sequencing revealed that the bacteria was of the same strain as Patient One’s, suggesting that one of the survivors of the original outbreak had returned to the hospital and that there had been a breach in infection control, says the Clinical Center’s Gallin. An environmental swab of the handrail outside the bone marrow transplant patient’s room showed traces of the bug, but no one knows if the patient put it there or if someone else did.
For the hospital staff, it was a terrifying turn of events, like being in a horror movie where the slain monster rises up to claim one last life before disappearing into the ether.
Despite the few unresolved mysteries, the hospital staff says it learned from the experience. Two other patients admitted in 2012 brought KPC with them from other hospitals. But this time, the bacteria was contained.
The hospital now maintains an isolated ICU for incoming patients with dangerous bugs. A hall monitor remains on staff to keep hand hygiene near 100 percent. And the hospital continues to check patients for any return of KPC in the building.
As for the seven patients infected with the strain who survived, including Patient One, they’ll continue to live with the bug. Doctors don’t yet know how to permanently rid them of it. They remain at risk of getting sick, and when they return to NIH for follow-up exams or research, the hospital treats them as if they were contagious by placing them in isolation.
“There was one patient who thought someone should apologize [for the outbreak],” Palmore says. “So on behalf of the hospital, I did.”
She’s all too aware that if her team had known the exact transmission route of the KPC earlier, the hospital might have found silent carriers sooner and saved lives. That’s why Palmore and Segre believe the precise data revealed by sequencing the KPC genome can be such a significant tool in future patient-safety efforts. By matching up the data, they can get a good idea of how bacteria spreads and work to stop it.
Last August, they produced a paper on their experience in the scientific journal Science Translational Medicine. It was the first the public heard of the outbreak. It’s rare for a hospital to discuss an outbreak publicly because of fear of lawsuits, but “the [culture] of NIH is a learning organization,” Segre says.
The response to the paper was unexpected. It set off a firestorm of criticism among consumer advocates and local government officials, including Montgomery County Council President Roger Berliner and former Rep. Connie Morella, R-Md., who thought the public should have known about the outbreak in 2011.
“It’s disturbing this information came out a year later,” says Lisa McGiffert, director of Consumer Union’s Safe Patient Project, an Austin, Texas-based campaign to raise public awareness about patient safety issues.
Gallin says patients admitted to the hospital were informed of the outbreak and had a choice not to come. “We run this phenomenally delicate balance of educating the community, but not inappropriately scaring them,” he says.
An infectious disease physician who has been director of the hospital since 1994, Gallin says his “big fear is that people who need modern health care at the NIH are going to get so scared they won’t come here. They won’t realize the risk of getting these infections is very small.”
Nonetheless, the NIH responded to the criticisms by announcing in late November that it would henceforth notify county, as well as state, officials in the event of any outbreak.
Since the story came out, Palmore and Segre have become known as disease detectives and have been quoted by dozens of news organizations about their story and issues of patient safety.
Segre hopes their story bolsters the authority and resources of infection-control leaders. She doesn’t think hospitals need to start sequencing all bacteria, but she does think the Clinical Center’s experience demonstrates that hospitals should be doing more aggressive surveillance testing, periodically checking on whether MRSA, KPC or other bugs are present in the environment or in ICU patients.
Meanwhile, Gallin sees the story as an urgent call to pharmaceutical companies to develop new antibiotics. The NIH is spending about $300 million on antibiotic development, “but we need to be more aggressive,” he says.
“The public needs to understand that these infections are not going away.”
Bara Vaida is a longtime health and public policy reporter who lives in Washington, D.C.