Pursuit of a pathogenEpidemiologist races clock, overcomes odds to solve baffling medical mystery
By John Easton
Sliding backward, well after midnight, down an ice-coated highway in the backwoods of New Hampshire, the gas gauge in her rented economy sedan well on its way to empty, and with less than 12 hours to solve a puzzle that had baffled her team for three months, Tina Chang was not thinking, "lead author, New England Journal of Medicine." She wasn't speculating about the Mackel Award. She wasn't even contemplating the three meals she had just missed, or estimating the miles she had to go before she could sleep. She was wondering how to get that Hyundai up the hill. She had to see a man about a dog.
Three weeks earlier, Chang, now a Robert Wood Johnson Clinical Scholar at the University, had been dispatched from the Centers for Disease Control and Prevention in Atlanta to the Dartmouth-Hitchcock Medical Center, to help physicians there get rid of something strange that was infecting their smallest, most fragile patients.
The first case had emerged on Oct. 17, 1993. A premature infant developed an unexplained fever and trouble breathing. Whatever was causing the fever didn't respond to antibiotics, and when the neonatologists tried to culture the bacteria, nothing grew. It wasn't until mid-November that they learned the child was playing host to Malassezia pachydermatis, a rare fungal infection in the bloodstream. Luckily, the child quickly recovered when treated with anti-fungal medications.
The next month, however, there was another case in the same intensive care nursery, then two more. In 1994, there were 11 more cases in two separate clusters for a total of 15 patients. Since human infection with this pathogen had been reported only once before -- and the source of that outbreak was never identified -- and since it kept coming back in cluster after cluster, the call for help went out to the CDC.
"Our mission was to figure out where the germ comes from, who's at risk, how it spreads and how to prevent it," said Chang, who was then in the Epidemiologic Intelligence Service within the Hospital Infections Branch at the CDC. She describes their approach as "shoe-leather epidemiology. We gather every bit of data that we can, try to make sense of it, then go back for more." The easiest part is usually determining who's at risk and how it spreads. The hospital had carefully mapped out the distribution of cases, the first step in detecting a pattern.
The two-person CDC team -- they later added a third and called in several experts as consultants -- went a bit deeper, charting the course of each of the 500 babies who had been in the unit during the period. They found that only those who weighed less than 1,300 grams and spent two weeks or more in the unit -- the smallest, sickest children -- were at risk.
How was it spread? Chang unrolls Exhibit A, a 20-foot-long scroll charting how often every nurse, on each of three shifts, might have touched any child in the unit. The chart reveals another risk factor, exposure to two of the unit's 54 nurses.
Exhibit B is a surreptitious study of hand-washing practices within the unit. While pretending to pore over medical charts, Chang's team covertly noted whether each caregiver washed his or her hands before and after every contact with a patient, a standard but often unheeded infection-control policy. Compliance was good, but far from perfect. Those who cut corners cited emergency situations, overwork, or just dry hands as an excuse.
Meanwhile, Chang's team was culturing everything in the unit. They swabbed babies, equipment and nurses' hands, shipping everything back to Atlanta where microbiologists prepared a special culture medium to grow this unusual pathogen.
They found lots of the fungus throughout the unit, ready sources for new infections. More than one-third of the healthy babies in the nursery produced positive cultures. There was limited prior evidence that these organisms could colonize human skin, but the team found it thriving, at least temporarily, on one nurse's hand. An educational session dramatically improved hand-washing in the unit, which quickly stopped the spread. Soon the pathogen had disappeared from the nursery.
But without learning where the organism originally came from, the team couldn't guarantee it wouldn't return. Malassezia pachydermatis was first described in 1925 in a rhinoceros with exfoliative dermatitis, not a likely source in this setting; none of the health care workers possessed pet pachyderms. So the focus of the investigation began to shift away from the hospital to the library, where the team combed the literature for clues to fungal breeding grounds.
Meanwhile, an obstacle emerged in Washington, D.C. Newt Gingrich, backed by an influx of newly elected supporters, threatened to downsize the federal government by delaying renewal of the operating budget. The CDC wasn't likely to close, but Chang's departmental credit card mysteriously ran dry. Her team was encouraged to wrap up quickly -- within the next 36 hours -- present their results and pull out. Late that night, they stumbled onto the answer in an obscure veterinary infectious disease journal.
Far from being restricted to large, thick-skinned hoofed mammals, Malassezia pachydermatis, they learned, was commonly found in ear infections, and sometimes in the healthy ears of floppy-eared dogs, like cocker spaniels. A quick phone survey uncovered 53 pets -- mostly cats, dogs and horses -- owned by the 81 health care workers in the NICU. Now each of them had to be swabbed and cultured. "The nurses were incredibly helpful," recalls Chang. Many of the caregivers were able to bring their pets to a meeting point where their ears could be conveniently scraped. The team had to go chase down the rest, however, and an ice storm had just rolled in, coating the region's roads.
It took all night, and they did run out of gas, but by 8 a.m. Chang's team had coaxed the Hyundai up every hill, swabbed every known pet, and prepared the samples for the trip to Atlanta. That morning, Chang and her two colleagues presented their hypothesis to the thankful unit and withdrew from the scene. A few weeks later, DNA testing confirmed that one strain of the fungus, found in three of the 12 infected dogs, was identical to the one found on the nurse's hand, which exactly matched that found on the children.
"It wasn't quite the smoking gun we hoped for," said Chang, "with clear proof of transmission from pet to staff and patient to patient." But it was still a convincing explanation of the spread of an emerging pathogen. In April 1994, the research team received the CDC's Mackel Award for the year's most outstanding project combining field epidemiology and laboratory research, and last week the paper appeared in the March 12 issue of the New England Journal of Medicine. Years from now, in intensive-care nurseries around the world, that issue still will be found -- dog-eared, no doubt.