Flu threats are tough to pin down
Predicting range, lethality of a virus always challenges public health officials
The number, big and scary, seeped into headlines, onto newscasts - and into the public consciousness: As many as 90,000 Americans could perish this fall from swine flu infections.
It appeared prominently in a White House report drafted by some of the nation’s brightest scientific minds, including researchers from Boston. But it was never intended as an actual forecast, and all the evidence so far - a lot of people have reported swine flu symptoms but relatively few have died - suggests the germ won’t claim anywhere near 90,000 lives.
Instead, that number represented one of several scenarios government agencies could use in planning for a resurgent wave of swine flu.
The truth is that disease trackers cannot predict precisely how lethal any flu strain will turn out to be.
Flu viruses are capable of making subtle alterations in their genetic machinery that transform them into more - or less - fearsome pathogens. Timing is critical, too, and that’s especially true with swine flu: A vaccine isn’t expected to be widely available until mid-October, and the germ is already circulating among a population with little natural protection against it.
Even in a regular season, chronicling the death toll from flu is notoriously tricky. Flu symptoms mimic other ailments, office tests are hardly foolproof, and the virus proves most lethal to patients already burdened with chronic illnesses.
“People die of heart attacks and pneumonias and so on who wouldn’t have died if they hadn’t been infected with influenza,’’ said Dr. David Ozonoff, a flu specialist at the Boston University School of Public Health. “But if you look at many of those death certificates, they won’t have influenza as the immediate cause or even the contributing cause.’’
Marc Lipsitch, a Harvard School of Public Health epidemiologist, speculated that the official count of swine flu deaths - it stands below 1,600, as of the middle of last week - “is certainly an underestimate of the number of deaths. It may be a bad underestimate or it may be a modest underestimate, but it is certainly an underestimate.’’
For example, research from New York City, particularly hard hit by the virus in the spring, showed that almost 20 percent of the people killed by swine flu hadn’t been hospitalized, Lipsitch said, making it less likely such deaths would be linked to the disease. Multiplied across the country, that phenomenon could contribute to making the flu strain appear less lethal than it is.
Assessing swine flu’s impact was further muddied by a decision in the past month to change what counts as an H1N1 death.
Until the end of August, only cases confirmed by sophisticated labs were included in the national tally of deaths, which, by then, had climbed to 593. But state health departments, overwhelmed by the number of people infected, have mostly stopped doing the sophisticated and expensive tests to confirm swine flu.
So federal disease specialists expanded the definition to include deaths that doctors attribute to flu and pneumonia based on symptoms, and both the swine type and the seasonal strain count. Since that change was made, the count has grown by an additional 936.
Flu stands in stark contrast to other diseases that are far easier to monitor because they cause fewer cases with more distinctive symptoms. “With flu, we can have 60 million people potentially infected,’’ said Dr. Lyn Finelli, flu surveillance chief at the US Centers for Disease Control and Prevention. “We could never count all those cases, and we could never count all those deaths.’’
Harvard’s Lipsitch had a central role in developing the swine flu planning scenario authored by the President’s Council of Advisors on Science and Technology. That report - which said that in a “plausible scenario,’’ H1N1 could kill 30,000 to 90,000 - emphasizes “this is a planning scenario, not a prediction.’’
A cochairman of the council, Eric Lander, said it would have been irresponsible not to include dire scenarios. Failing to do so, he said, would be akin to neglecting the possibility that a coastal region could be walloped by the most powerful of hurricanes.
“We’d better be prepared for a hurricane devastating New Orleans,’’ said Lander, director of the Broad Institute in Cambridge. “But that’s not the same thing as predicting a hurricane is going to devastate New Orleans.’’
To develop their models, the scientists reviewed what happened during the two most recent global flu epidemics, in 1957 and 1968. They examined the course of the H1N1 virus since it emerged in Mexico, traveled to the United States, and then hopscotched the globe. And they considered what might happen if infections are rampant before a vaccine is available.
But all of those considerations involve significant - if educated - guesswork. Consider even the widely cited figure that, in an average flu season, 36,000 people are killed by flu viruses in the United States. It was distilled by federal specialists from more than two decades’ worth of flu surveillance.
“Just because the CDC or some other agency puts out a number as an official health statistic, that doesn’t mean it’s a true and accurate number. It’s more of an estimate,’’ said Al Ozonoff of BU’s public health school, who specializes in the study of influenza.
A recent study found that people under 30 have little existing protection against the swine flu virus, while older adults harbor disease-fighting cells against the germ. It’s believed the older adults’ immune systems carry the memory of encounters with ancestors of the virus.
The reality of the flu lands at Dr. Benjamin Kruskal’soffice, where physicians are steeling themselves for a flood of patients.
“I think we’re likely to see quite a few more cases than we typically see during flu season,’’ said Kruskal, director of infection control for Harvard Vanguard Medical Associates.
Stephen Smith can be reached at stsmith@globe.com. 