Legionnaires’ disease – mounting the first investigation

January 2006

When a mysterious pneumonia hit Pennsylvania in 1976, Stephen Thacker, MD, an Epidemic Intelligence Service (EIS) officer with the CDC, was a little wet behind the ears — he and many of his colleagues had just finished their initial EIS training.

“I was assigned to the D.C. Health Department. On my first day on the job, I was going around D.C. with my supervisor, and I came back to the office around 4:00 pm and got a message from my supervisor in Atlanta. He said that they had a little problem with an unusual pneumonia in Pennsylvania and that I needed to drive up there the next morning,” said Thacker, who is still an epidemiologist with the CDC. “My first day,” he said, “I learned a lot.”

Throughout Pennsylvania, war veterans who had attended a July American Legion convention in Philadelphia were suffering a mysterious illness.

Swine flu

Initially, public health officials were concerned that this was swine flu; in February 1976, two Fort Dix, N.J., recruits came down with pneumonia. “They cultured the two, and they got an influenza virus that they started thinking resembled the flu virus of 1918-1919,” said Philip Brachman, MD, who was director of the CDC epidemiology office at the time. “It seemed to have certain characteristics that people thought might be similar to swine flu virus.”

That information made public health officials nervous, so the CDC had a special team waiting for influenza to strike. Against the advice of the vaccine manufacturers, because they did not have enough time to prepare it, the variant strain was added to the new 1976-77 influenza vaccine.

“This was a presidential election year,” said Thacker. “So, this was President [Gerald] Ford’s program.” He said the vaccine manufacturers were reluctant to make vaccine because they felt they didn’t have enough time, and they wanted no liability if something went wrong.

Unfortunately, something did go wrong. In October 1976, during the influenza vaccination season, some of those vaccinated came down with Guillain-Barré syndrome.

“We had a problem, and it shut the vaccine program down. It got the CDC in deep, deep trouble,” he said. “They had to have a scapegoat, and it was Dr. [David] Sencer, and he got fired. He was a very popular director and a very effective director, but he lost his job over this.”

Ford did not win the election.

But that was months away. It was August, and there was a mysterious respiratory illness in Pennsylvania that was killing veterans. “The problem was that when legionnaires’ disease developed, having a lot of publicity about this potential swine flu strain, we were afraid we were going to have a pandemic of swine flu in the winter. Now, we’ve got a respiratory disease in Philadelphia, could it be influenza?” Brachman asked.

The samples, however, came back negative — this was not swine flu.

The investigation then went to the Special Pathogens Branch, and Brachman sent David W. Fraser, MD, a staff epidemiologist at that time, to head it. “I headed the federal field investigation,” said Fraser. “On Monday night, the night the first team went out, one of my EIS officers, Ted Tsai, was on that first team. Sencer that evening called a meeting and asked all the units at the CDC who had some understanding of what was going on to be there. During that meeting, he asked Phil Brachman to send a senior person out to head the field team. Brachman chose me, I think in part, because I wasn’t already engaged in things like swine flu, which were occupying lots of people at that time.”

Challenges to the investigation

At the request of the state health department, the investigation was set up in Harrisburg.

“I went out the second day to Harrisburg, which was where the state had asked that we set up the headquarters of the investigation,” said Fraser. “I assume it was because it was a statewide problem. The state health department was centrally located in Harrisburg; its laboratory was there.”

EIS officers were also dispatched to Pittsburgh and Philadelphia. “In terms of organizing the investigation, with all those EIS officers with three different cities, the location, it was an interesting logistical problem and an interesting management problem,” Fraser said.

That was not the only challenge in this investigation, Fraser said in an article in The Lancet Infectious Diseases. Other factors complicated the investigation: cases were dispersed across the state because the incubation period was two to 10 days, symptoms developed after the convention was over, there was no centralized record of convention attendance and laboratory methods at the time could not identify an etiologic agent.

“The reliance on a clinical case definition and the need to coordinate local, state, and federal jurisdiction both hindered efforts. Intense public scrutiny, anxiety, and promotion of implausible theories, researchers’ memories of prior unsolved major investigations, and the inherent difficulty of documenting airborne spread also had an impact,” Fraser wrote.

The press was demanding almost hourly updates. To allow the EIS officers the freedom and the peace to conduct their investigation, the state health department handled the media and provided twice-daily press briefings.

The investigation

Brachman said it was one of the largest epidemic investigations at that time; there were 25 CDC EIS officers throughout Pennsylvania and in neighboring states.

Fraser agreed. “The number of EIS officers involved increased rapidly, but the first thing we tried to do was get an EIS officer out to see every person who’d been reported to be sick who’d been at that convention because we wanted a physician epidemiologist to go over the medical record, examine the patient, talk to the patient, try to understand what this disease was,” he said.

“We had only sketchy information at that time. We wanted to collect specimens. We wanted to get some preliminary information about what activities people had been involved in,” Fraser explained, “so we could start the systematic epidemiologic investigation that we were trained to do.”

The EIS officers developed a questionnaire about activities that would help differentiate the activities of those who were sick from those who were not. “We had a couple of problems with that: We didn’t know who wasn’t sick because there was no master list of who’d been at the convention,” he said.

Therefore, the CDC asked each American Legion post which of its delegates attended the convention. Then, the CDC distributed questionnaires to the posts. The leadership distributed them to attendees.

“So, they distributed the questionnaires to everybody in the posts who’d been at the convention and then collected the questionnaires and sent them back to Legion headquarters, so we then got them,” Fraser said. “It was a different way to get the denominator from ways that I’m used to, but it seemed to work. We also collected specimens from people who were sick: blood specimens, swabs and autopsy specimens. We split them with the state health department laboratory and shipped half of each of those to the CDC.”

Thacker added, “We had daily meetings of the staff who were working on different aspects of the investigation, including looking for potential environmental and occupational causes, examining case findings, finding out everything about the convention and trying to contact everyone who might know something about it.

“On that following day, several of us went out and interviewed patients, their physicians and their families with a questionnaire we developed the night before,” Thacker said. In addition, they reviewed medical and pathology records.

Thacker was out one day conducting interviews when he received a page to call Fraser. “After my second or third hospital, I got a call to call David [Fraser], and, of course, this was pre-cell phones, so I had to find a phone. David said, ‘Steve, this is getting a lot of press, and the press wants to keep asking the EIS officers questions and to work with one.’ I’m looking for someone who’s articulate and bright and well-spoken, and close – and you’re close.”

By the time Thacker reached the hospital in Chambersburg, chosen members of the press corps were waiting for him. He answered their questions, and then they followed him to see a patient. “I went to visit a patient, and this patient had been quite ill but was, you know, alive, and his chart said he had had a temperature of 108° F, which surprised me. The only patients I had remembered from medical school and residency with that temperature were not doing well,” Thacker said.

“He was very comfortable answering questions. There were physicians and nurses who seemed more nervous than he was. I turned to make a point to the media people, and they weren’t there – they had their heads sticking in the door. They did not come into the room.”

That scene brought home to Thacker how frightened people were with legionnaires’ disease. They didn’t know what caused it, they didn’t know how it was transmitted and they didn’t know how to treat it. They did, however, know it was deadly.

Although the surroundings were chaotic at first, a surveillance system emerged. Fraser separated the EIS officers into three- to five-person teams. Each team reported its findings every day. The state health department contacted the hospitals every day for updates, and the legionnaires provided information about the convention and distributed the questionnaires. EIS officers contacted patients, families, physicians and pathologists and collected serum and other samples.

Thacker was in charge of the line list, Fraser said. In the era before the laptop, all pertinent information was placed in a timeline on a large wall-sized paper to help identify similarities, differences and coincidences.

Case definition

The EIS team established a working case definition: pneumonia diagnosed by a physician or confirmed by x-ray or tissue diagnosis; cough; fever of 102°F or higher; and onset after July 1 in someone who attended the American Legion convention or had been in the Bellevue-Stratford Hotel, the convention’s headquarters.

One hundred eighty-two people met the case definition. Thirty-nine more had pneumonia, but had not been in the hotel. However, those patients had been outside the front doors of the hotel. They were called “Broad Street pneumonias,” explained Brachman. “We started getting cases of pneumonia confirmed, but in people who had not been in the hotel, but had been outside on Broad Street, which is the street in front of the hotel. And so, we said, ‘OK, we’ll put these in a category, and we’ll call them Broad Street pneumonias. We didn’t know what they were, if they were the same thing or not.”

Between both sets of cases, 34 people died.

“We investigated various things,” said Thacker. “We looked at the possibility of psittacosis because there was this pigeon lady who was outside the Bellevue, and Phil Graitcer tracked her down and interviewed her. We looked at cases that were a little different. They weren’t legionnaires, but they fit the criteria. There was this 82-year-old woman who came in the Bellevue Stratford to use the ladies room, was there for 20 minutes and, two weeks later, died.

“It turned out she had been a case,” he said. “There was the airplane pilot who came in at noon and slept six hours and left for his next flight, and he became ill.”

Brachman added, “We cultured the environment, anything you could think of: dead pigeons and rodents; their skeletons were found outside the hotel, pigeon droppings were on the roof, swabs were taken from all over, water was cultured, etc. Everything was negative. We then sent a lot of those specimens down to the CDC where there was more equipment, more people and better virology. Again, after days, weeks, everything was negative. That was very frustrating.”

When it became clear that the outbreak started in Philadelphia, Fraser moved the team from Harrisburg to Philadelphia. He and a few others on the team stayed at the Bellevue Stratford Hotel. “There was no problem getting a room,” Brachman said. “The place was moderately empty.”

None of the EIS officers, however, became ill.

Next month: Narrowing the suspects

For more information:

• Fraser DW. The challenges were legion. Lancet Infect Dis. 2005;5:237-241.
• Thomas G, Morgan-Witts M. Anatomy of an Epidemic. Garden City, NY: Doubleday; 1982.