ROANOKE TIMES
Copyright (c) 1996, Roanoke Times
DATE: Sunday, December 8, 1996 TAG: 9612090066
SECTION: VIRGINIA PAGE: B-1 EDITION: METRO
SOURCE: MARK CLOTHIER STAFF WRITER
FINDING THE SOURCE of the Legionnaires' disease outbreak in the New River Valley required old-fashioned detective work by health officials.
Of the killer, this much was known: It had a thing for older people, especially men, and was as elusive as the fog in which it could appear.
By the time detectives knew there was a pattern, four people were hospitalized.
The next day, the number of victims doubled. Later that day, reinforcements were called.
One of the state investigators, epidemiologist Betty Rouse, was a 23-year veteran who had tracked countless diseases, including a 1988 outbreak of hepatitis A in Tidewater.
But this was different.
"This was far more intense," she said. "This was Legionnaires' disease. Folks die from this."
By the fourth day, someone had: a 50-year-old Pearisburg man, the outbreak's only fatality. It was Sunday, Oct.27.
By then there were 11 victims, and the investigators, a team of 11 specialists from the Virginia Department of Health and the federal Centers for Disease Control and Prevention in Atlanta, had mustered.
By Nov.14, health officials pinned the source of the Legionnaires' disease outbreak to a display hot tub in the Christiansburg Lowe's, a home improvement store.
How the doctors and investigators traced the source of the disease, how they found the link between 15 sick people and one hot tub, is a detective story - a medical detective story.
* * *
The story began Oct.14. Lung specialist John White was the doctor on call for his Blacksburg practice that night.
White's curiosity was piqued when he admitted eight people with pneumonia to Columbia Montgomery Regional Hospital.
Legionnaires' is an especially severe form of pneumonia, named for the 1976 outbreak among American Legion conventioneers in a Philadelphia hotel that killed 34 people.
"I was thinking Legionnaires' from that point," White said. "I'm always looking for it. It's one of the things I'm trained to do, and that's an unusually high number of pneumonia for one night."
The patients shared similar symptoms. They were mostly older males, with chronic flulike symptoms and severe pneumonia.
The next day, White called Jennifer Brumfield, the infection control nurse at the hospital, and asked her to check if there were any more pneumonia patients.
She pulled chest X-rays and looked at symptoms - fever and blood counts. When she was done, 10 more pneumonia patients were confirmed.
Brumfield checked with other New River Valley hospitals and found a higher-than-usual number of pneumonia patients. Then she called the New River Health District.
As the state Health Department's director in the district, Jody Hershey is trained to see the big picture.
"When we got the call from Jennifer, I was concerned enough to call the state and CDC and to start looking," he said. "I knew there was something going on, but I wasn't quite sure what. I didn't want to jump to conclusions. I wanted to do my homework."
That included taking blood from the pneumonia patients to try to confirm White's Legionnaires' hunch.
Hershey guessed that if there was an increase in pneumonia patients in New River Valley hospitals, there also would be a lot of people showing up in local doctors' offices with lesser, flulike symptoms. But there weren't.
"That told me there's something going on here," he said. "Something wasn't right."
By Oct.23, testing confirmed White's hunch.
The next day, Hershey held a news conference announcing four confirmed cases of Legionnaires' disease. That doubled the next day, and the eight Legionnaires' cases were deemed an outbreak. The state's first.
``There was no evidence of any connection among the cases, but my gut-level feeling was we had a problem," Hershey said.
Saturday, Oct.26, the team started testing patients who had been admitted with pneumonia as far back as Sept.1.
They also started talking to the confirmed Legionnaires' patients, looking for common threads - traffic and shopping patterns, shared job sites and favorite restaurants.
In the case of the New River Valley Legionnaires' disease outbreak, the numbers pointed to Montgomery County - home to 11 of the 16 known patients, Virginia Tech and the New River Valley's highest concentration of shopping.
"You ask yourself where could a large number of people be exposed?" Hershey said. "Where do a lot of people go?"
* * *
In such medical mysteries, initial presumptions become working theories about the most likely sources of disease-causing bacteria. But without statistical confirmation, theories stay hunches.
If, for example, the bulk of the confirmed Legionnaires' patients had fallen ill after visiting the produce section of the same grocery store, that would be enough to formulate a theory.
To turn the theory into a solution, a control group was needed: people who didn't contract the disease but who shared age, medical history and gender with those who did.
If few of the controls visited the hypothetical produce section over the same period, the chance the produce section was the source would increase.
In the case of the New River Valley outbreak, though, it wasn't a produce section, but a hot tub in a home improvement store.
Rouse, the epidemiologist, said clues started pointing to Lowe's during patient interviews the first weekend. Fourteen of the 15 confirmed patients had been there for an average of 78 minutes within 10 days of getting sick. Ten of the 14 patients recalled walking by the hot tub that sat on one of the store's main aisles.
In the meantime, investigators asked Lowe's officials to drain and cover the hot tub and turn off the greenhouse sprinklers and decorative fountains. Legionnaire bacteria are water-based and spread when vaporized.
For the investigators, testing their theories was a matter of assigning a number to the victims' length of exposure to the suspected source, putting the numbers into a few computer programs and hitting "enter." More or less.
Finding the control group proved tougher.
* * *
For each Legionnaires' patient, three controls were needed. But to be safe, six were sought for each on the presumption some people would not participate.
On Halloween, investigators started looking.
Rouse and another epidemiologist, Bob Hackler, spent that afternoon in a doctor's office looking through patient records for six approximate matches for a 70-year-old man with diabetes.
They needed a man, of whom there were several, but few were between the ages of 60 and 80.
Rouse and Hackler also needed someone with a disease similar to diabetes, which is chronic but usually doesn't suppress the immune system.
Some forms of cancer were OK, but only if the potential control patient wasn't on chemotherapy or radiation treatment, which suppress the immune system.
Potential controls with leukemia and lymphoma wouldn't work, either - those cancer forms also suppress the immune system.
It was no easy task.
Rouse looked through 500 files that afternoon and found 40 that matched the patient's age and gender. Of those, one shared a similar medical history.
And there were still five more controls to find.
Two days later, the CDC's Emily McClure had to pick through 1,400 files to find one potential control patient.
Most controls weren't that difficult to find. But the older and sicker the patient, the harder it was. That most of the patients were old and male made it even tougher.
More typical was a ratio of 150 files pulled to land one potential control.
During the control group search, and the 22 days of the overall investigation, workers clocked 12- to 14-hour days, most with no time off.
Rouse's only free time was for Election Day. She drove 335 miles home to Virginia Beach the night before, grabbed some winter clothes, voted, then drove back.
Once the potential controls were found, they had to be called, and if willing, interviewed. All 60 were done by Nov.9.
McClure said most who were contacted were eager to help. Those who weren't were worried about privacy.
"Folks for most part, I think, saw it as a chance to help solve a problem," she said. "I think people saw a chance to really make a contribution, and it made all the difference in the world. It really did."
Once the 45-member control group was together, it was question-asking time.
And the way in which questions were asked was key. Rouse said interviewers used the same tension, voice tone and delivery with each person to avoid interview bias, which could unconsciously sway the subject.
"Maintaining an open mind is one of the tough parts," Rouse said. "You have that gut feeling, but you must leave [your mind] open enough to ask the right questions to implicate or send you in the right direction."
The patient interviews had pointed to Lowe's. If the control person had shopped there, follow-up questions were asked: When? For how long? What departments were you in? How long were you in each?
More often than not, memories needed jogging.
If a person couldn't remember where he'd been, but always paid by check, he was asked to look at old check registers.
A Virginia Tech home football game also helped. People tended to stay away from the Christiansburg U.S. 460 retail corridor when traffic was bad and knew they hadn't visited say, Wal-Mart, on weekends in question.
Two of the control group members even kept journals and could tell investigators their comings and goings with helpful, near-excruciating, detail.
Of the 45 controls, 12 had been to Lowe's over the same period, but for an average visit of less than 30 minutes.
* * *
To further solidify their hot tub theory, investigators needed physical evidence: to place the bug at the scene of the crime.
This was done by matching bacteria specimens from patients with a specimen from the hot tub, which by then had been taken down and sold to a Lowe's employee.
McClure took the dry paper filter from the spa, wrapped it in a white trash bag, taped it, boxed it and shipped it Oct.30 to the CDC in Atlanta for DNA testing.
On Nov.12, CDC tests matched the bugs.
Without the DNA match, Rouse said, investigators would have had to rely on the percentages, the statistical data from the interviews, which would have made the job tougher.
"The DNA confirmation was a real high," Rouse said. "There were times, before we were really in the clear, when you sort of had that 'this-had-better-be-right' feeling. The stakes were higher, and there's an awful lot to watch.
"But when the lab confirmation came back and we knew the filter matched the patient, it was like, 'Yes! Yes! We did it.'''
LENGTH: Long : 202 lines ILLUSTRATION: PHOTO: 1. GENE DALTON\Staff. Nurse Jennifer Brumfield and Dr.by CNBJohn White alerted the state to the unusual number of pneumonia
cases. 2. VICKI CRONIS\Landmark News Service. Betty Rouse, regional
epidemiologist for the state Health Department, said patient
interviews pointed to the Lowe's in Christiansburg as the source of
the outbreak. color.