Coughing cases mystify specialists

It seemed straightforward enough at first: A 19-month-old boy being treated at Children's Hospital Boston harbored classic symptoms of whooping cough. And a laboratory test confirmed he had the disease.

Then, from late September through early November, three dozen hospital employees in wards where the child had been treated also developed coughs and other symptoms suggesting that they, too, were stricken with the disease. And, just as for the toddler, initial testing confirmed they carried the bacterium.

But now, more than three months after the child fell ill, disease investigators are no longer certain that the respiratory illnesses that struck workers and one other patient at Children's Hospital were caused by whooping cough, also known as pertussis . Additional testing, using approaches different from the first round, could find little evidence of the highly contagious germ.

Still, no one can say for sure what made the workers sick -- they all recovered -- and disease specialists can't definitively rule out pertussis as the cause.

The cluster of illnesses is proving so intriguing that federal disease sleuths and even a lab in Canada have joined specialists from Boston and the state who are trying to crack the mystery.

The episode at Children's Hospital is a cautionary tale about the limits of science, particularly the ability of laboratories to produce absolute, unimpeachable results when disease strikes.

"People think laboratory tests are 100 percent accurate: 'If they're positive, they're positive; if they're negative, they're negative,' " said David Warshauer , a top official at the Wisconsin State Laboratory of Hygiene , which oversaw testing during a pertussis outbreak in 2003 and 2004 that spawned more than 5,000 cases. "But that's not the case with any laboratory test. And when you're looking for an infectious agent like pertussis, there's a lot of variables."

The mystery began Sept. 18, when a toddler came down with symptoms of whooping cough, typically heralded by a runny nose, sneezing, slight fever, and mild cough. The disease can develop far more severe manifestations, most notably spasmodic coughing that can last for months and be ferocious enough to shatter ribs.

While symptoms can provide strong clues about whether a patient has the disease, laboratory tests provide far more robust proof -- especially for pertussis, with early symptoms similar to the common cold. To collect a specimen for testing, a cotton swab- like probe is threaded through the nose and into the back of the throat.

"People generally don't like it much when you do that," said Dr. Alfred DeMaria , director of communicable disease control for the state. "But you can't get to that area through the mouth."

The sample is then used two ways. One is a laboratory classic called culturing: Material from the patient is placed in a dish with a substance that will allow the bacterium to reproduce if it's present. The other approach is more modern: Polymerase chain reaction, or PCR, involves looking for the genetic fingerprint of the disease.

PCR has emerged in the past decade as a favorite testing technology in state labs across the nation, including the Massachusetts State Laboratory Institute in Jamaica Plain, in large part because it's easier and faster than culturing .

It was that lab that tested most of the samples collected from Children's Hospital, initially using PCR. There were hundreds of tests performed, and 38 -- two patients and 36 healthcare workers -- came back positive for pertussis.

"Children's, much like we do at the local health department, really relies on laboratory tests to guide us on what the diagnosis is, especially illnesses that can look like a lot of different things," said Dr. Anita Barry , top disease tracker at the Boston Public Health Commission . "Having accurate test results early on, particularly when they're consistent with the clinical symptoms, really launches us into control steps."

And that is what happened at Children's Hospital , with measures designed to hem in the bacterium and prevent it from colonizing more victims.

"You think you know what's going on," said Dr. Manisha Patel , a disease investigator at the US Centers for Disease Control and Prevention . "And you want to make sure it doesn't spread to the most vulnerable patients in the hospital."

But even as those steps appeared to be reining in the outbreak, something curious began happening at the state lab.

Specialists rarely rely on a single testing method to confirm a diagnosis if multiple methods exist. So state lab workers, in addition to performing PCR testing, proceeded with the laborious task of culturing samples. They also tested blood samples from the hospital workers.

The results of those additional tests: They were almost uniformly negative for pertussis, DeMaria said.

"What was going through my head at the time was, what might explain this?" Barry said. "Is there some type of problem with one of the tests? Could it be something closely related to pertussis?"

Samples were sent to the federal Centers for Disease Control in the quest for answers. Researchers in Atlanta performed a more elaborate, two-tier PCR test.

"The results were inconclusive," said Dr. Amanda Cohn , a medical epidemiologist for the federal agency .

It turned out that one part of the PCR test suggested the presence of the pertussis bacterium while the other part didn't. The state lab, which initially performed just one phase of the PCR test, decided to conduct the second phase and, much like the federal lab, came back with negative results on that second phase.

"One of our group's highest priorities at [Centers for Disease Control] is working out better laboratory diagnostics for pertussis," Cohn said. "The laboratory diagnostics for pertussis are difficult. It's a priority for us to find reliable, rapid, standardized testing for pertussis."

The federal agency and Massachusetts health authorities continue to work with Children's Hospital to ascertain what made the 36 adults and two children ill. There are competing theories, ranging from a cold virus to a bacterial relative of pertussis to the virus that causes the condition commonly known as walking pneumonia.

It's unlikely that the causes of all the respiratory illnesses will ever be fully known.

"What I can say is that whatever it was, it went away," Barry said. "And that's the good news."

Stephen Smith can be reached at stsmith@globe.com.