Flu: Grim stats

A diagnosis of cancer, diabetes — even cirrhosis or Parkinson’s disease — will send shivers down the spines of most people. But flu? We tend to view it as little more than an occasional feverish nuisance that sends us to bed with joint pain, congestion and nausea (or worse). What so many of us fail to appreciate is that flu can kill.

Having filed a news story, yesterday, on concerns about the potential for development of resistance to the leading flu-fighting drug (Tamiflu), I was sensitized to influenza incidence data. From the World Health Organization to the U.S. government, websites abound with notices on what to do when anticipated flu emergencies begin erupting internationally within the next few months. I thought I got it; lots of people are projected to get sick.

Then I glanced at a newspaper story this morning that cited flu as the eighth leading cause of death in the United States. That’s a wow stat. It was unattributed, however, so I decided to investigate whether there were sound data to support it.

Bottom line: There are.
 
I’d heard from some academics this week that flu was estimated to claim 40,000 to 50,000 lives a year in the United States alone. But I had no idea how many people die each year, so there was no denominator against which to compare those figures — those numerators which also appear on a National Institute of Allergy and Infectious Diseases website.

Alas, NIAID doesn’t record flu-mortality rankings, so I was directed to the Centers for Disease Control and Prevention, where officials recommended I check out the agency’s National Center for Health Statistics. And it’s there that I finally hit paydirt — a death-data motherlode.

On April 17, the center issued its annual report on what Americans die from (113 different choices are available) and at what age. Ethnicity data are also sometimes available. And this 135-page document confirms on page one that ranking in eighth place on the grim reaper’s scorecard is the deadly duo: influenza and pneumonia. A few pages deeper into the report we learn that together they claimed 58,326 of the 2.4-plus million lives lost in the United States during 2006, the most recent year for which data were available. That amounts to 18.8 people out of every 100,000.

But not quite a third of the way into this tome is a table reporting that only about one in every 300,000 U.S. residents had actually died from “flu.” So back to CDC. There I learned that most flu victims die from complications of their infection, which is why the death stats pair flu and pneumonia — pneumonia being far and away the most frequent complication.

This is something that Jonathan Dushoff (then at Princeton, and now at McMaster University in Hamilton, Ontario) and his colleagues wrestled with several years ago when trying to gauge influenza’s direct and indirect death toll.

“Every year a lot of people get flu, then recover from the virus,” Dushoff says. Later, they suffer heart attacks, strokes, pneumonia or some other lethal event. Because this occurs a day to maybe a month later, he notes, “flu never shows up on their death certificates — even though there’s a lot of evidence that many of these deaths were basically triggered by the flu.”

The trick is to figure out which of these deaths from other causes had been catapulted onto some lethal trajectory by flu.

Several research teams over the years have applied different statistical methods to gauge this. His team’s, reported in the American Journal of Epidemiology, took a two-pronged approach. First, it compared summer versus winter — non-flu versus flu season — death rates. Then the researchers compared those numbers against varying winter flu incidence throughout a span of 22 years, ending in 2001.

“What we found,” he says, “was that the more [seasonal] influenza circulating — particularly the more H3N2 there was — the more deaths there were.” Overall, it suggested a rough average death rate due to flu from all causes of about 41,400 in the United States over that period. It was a number quite similar to what emerged from analyses by others using different statistical approaches. “And that’s a reassuring sign of the robustness of these estimates — that 40,000 to 50,000 number,” he says.

My undergraduate sees no need for flu shots (though I arm-twisted her into getting one this year anyway). And after reviewing the statistics, this afternoon, I understand her complacency. Between the ages of 1 and 44, the incidence of flu mortality is only between 0.1 and 1.9 per 100,000. Afterwards, this window of seeming invulnerability starts climbing rapidly. By ages 45 to 54, the incidence reaches about 4.6 per 100,000. It doubles again in the next 10 year span, then triples during each of the successive two decades. By the time people reach 85, some 500 Americans per 100,000 succumb annually to flu and/or pneumonia.

So how do we keep from getting sick? My family subscribes to flu shots as a first line of defense. If and when H1N1 swine flu — or H5N1 bird flu — reaches locally epidemic proportions, we’ll pull out the masks that we’ve stockpiled.

One piece of somewhat hopeful news emerged today in a paper released early by the Journal of the American Medical Association. It concludes that ordinary surgical masks confer substantial protection against flu infection.

Mark Loeb, an infectious diseases physician at McMaster University in Ontario, Canada, and his colleagues found that “Surgical masks appear to be no worse than, and nearly as effective as N95 respirators in preventing influenza.” So no imperative to buy those super-pricey N95 anti-viral masks.

The trial included 446 nurses in eight Canadian hospitals. All had been caring for patients with fevers and respiratory illness (presumed flu). The nurses were assigned one type of mask or the other, and roughly one-quarter in each group of nurses went on to develop flu.

Both the CDC and the Institute of Medicine recommend use of N95 respirators around people infected with the H1N1 virus, note Arjun Srinivasan at the CDC and Trish Perl of the Johns Hopkins School of Medicine and School of Public Health. Writing in an editorial that accompanies the new study, this pair acknowledges that the Canadian findings are encouraging. Still, Srinivasan and Perl aren’t thrilled about the body of data available for setting mask-selection policy. “That this study is, to our knowledge, the first and only published randomized trial assessing respiratory protection for preventing influenza transmission is a sad commentary on the state of research in this area,” they observe.