December 24, 2004
The Press on Pain
Reading the Wall Street Journal’s editorial page is one of the more productive parts of my day. No, I’m not a liberal masochist. It’s just a quick way of keeping up with the latest bulletins from the frontiers of corporate public relations and right-wing think tank quackery.
Journal editorials supporting the drug industry usually combine the least admirable traits of both these schools of thought: Outright falsehoods combined with a misleading use of statistics. The Journal’s musings this week on “the painkiller panic” employed both those tools.
Now, before I launch into a dissection of their editorial and you head for the delete button, let me quickly say that as a long-time practitioner of the craft of journalism, I understand perfectly well that it never pays to argue with someone who buys ink by the barrel. But I’m making an exception in this case because the flaws of fact and logic in the Journal’s editorial are not limited to the right-wing cranks that dominate that page. They are the same flaws one finds in the pages and broadcasts of most news outlets when they cover medical news. So consider this a quick tutorial in how to read or watch what I call the medical hype and hope industry – brought to you almost every week on the covers of Time or Newsweek or the front page of your daily newspaper.
The issue of the moment is the safety of painkillers. Since Merck pulled Vioxx from the market in late October because it raised the risk of heart attacks, Pfizer’s Celebrex and Bextra and generic naproxen (the active ingredient in Aleve and some other over-the-counter pain relievers) have come under similar clouds. The Food and Drug Administration on Thursday announced its scrutiny will now extend to all pain relievers, not just the so-called Cox-2 inhibitors like Celebrex and Vioxx. What’s the poor consumer to think?
The Journal’s editorial exploited this confusion to the hilt. “All drugs have side effects, especially when taken in large doses and over the long term. We already knew that more than 15,000 people die – yes, die – annually from gastrointestinal bleeding caused by drugs like naproxen and ibuprofen, the side effect newer drugs like Vioxx and Celebrex were designed to avoid.”
The Journal wasn’t alone in making this outlandish claim about the deaths from over-the-counter painkillers. Last weekend’s New York Times front page story put “perhaps 16,500 deaths” in its lead. It would appear that lots of people – more than a third the total that die in traffic accidents every year – are being hauled off to hospitals and dying from bleeding ulcers after taking aspirin, ibuprofen and other generic painkillers.
As I’ve reported before in this space, that’s an outlandish claim that’s pretty easy to double check, which no reporter has yet taken the time to do. The Centers for Disease Control systematically tracks the causes of death for the 2 million-plus Americans who die each year. In 2002, the most recent year for which statistics are available, just 4,081 people died of burst peptic ulcers, a number that has stayed remarkably constant over the past decade after falling sharply in the early 1990s. If one includes all forms of gastrointestinal bleeding deaths (upper and lower G.I. track, whatever), the total is still under 8,000. Presumably, quite a few of those deaths were for reasons other than the side effects of pain pills.
Now any death is serious, but it’s interesting to note where the 16,500 number came from. In 1998, when the Pfizer and Merck’s pre-FDA approval marketing campaign for Cox-2s was at a fever pitch in the medical literature and in the press (Time Magazine hailed them on its cover as super aspirins), a Stanford University physician named Gurkirpal Singh published an article in a drug-industry funded supplement of the American Journal of Medicine, a second-tier medical journal. He created that number by taking the deaths from a sample of about 12,000 patients, most whom had severe rheumatoid arthritis and took fistfuls of painkillers every day, and extending the ratio to the 13 million arthritis sufferers that the Arthritis Foundation (itself substantially funded by the drug industry) says take pain medicine for their symptoms. The same was true for the claim that painkillers caused 107,000 annual hospitalizations.
Dr. Singh, identified in the Times article only as a professor of medicine at Stanford, had a handsome photo in the paper last Sunday. But when I called him this week at his place of employment to discuss his findings, he wasn’t in. In fact, he wasn’t listed in the Stanford Medical School directory.
So I called one of his colleagues, Dr. James Fries, himself an arthritis specialist. He told me that he always thought the 16,500 number was far too high. He also said that Dr. Singh was no longer affiliated with Stanford and if I wanted to learn why, I’d better call the dean because he wasn’t at liberty to discuss it. When I asked him if Dr. Singh had taken money from drug companies for his studies, Dr. Fries replied, “Oh, Dr. Singh took substantial sums from drug companies.”
There’s another way to look at Dr. Singh’s (and the Wall Street Journal’s and the New York Times’) number. The original clinical trial that showed Vioxx reduced gastrointestinal side effects compared it to naproxen, a pain reliever long thought to be the most destructive of the digestive tract. The Merck-funded study followed about half as many patients as Dr. Singh. Yet it found no deaths from gastrointestinal complications -- zero.
However, it did reduce the number of gastrointestinal events that were serious enough to require medical intervention. How many were there? For every 1000 patients taking naproxen during the trial, there were 14 serious gastrointestinal events. Vioxx reduced that down to 6. It was this same trial that alerted some physicians – like Dr. Eric Topol of the Cleveland Clinic -- that this new class of drugs might cause cardiovascular problems. He’s waged a valiant struggle to bring this to public attention ever since.
Now back to the Journal editorial. “One of the most frustrating things about the latest news on painkillers is that almost none of the people reporting it understand the concept of relative risk – i.e., that a doubling of adverse events like heart attacks still doesn’t mean that event is very likely.” The editorial pointed out that in the trial implicating Celebrex, a “mere” 15 per 1,000 suffered heart attacks compared to 7.5 taking nothing at all.
I couldn’t agree more with the Journal's logic. It’s time for all medical reporters to begin looking at absolute risks and benefits, not relative risks and benefits. How many times do we read that Drug A cuts your chance of getting a disease in half (relative benefit) only to find out later that for every 100 people taking Drug A, one will get the disease instead of two (absolute benefit).
Of course, the Journal was selective in its use of this principle. All trials that have looked at anti-inflammatory drugs, whether over-the-counter painkillers like ibuprofen or naproxen or the mislabeled “super aspirins” like Vioxx or Celebrex, show they are remarkably similar for treating pain with about 80 percent effectiveness. So their absolute benefit is the same.
The only difference is the side effects – and their price. Using the data from above-mentioned clinical trials, we find that for every 1,000 people taking a pricey, prescription "super aspirin" instead of an over-the-counter painkiller, our medical system has traded 8 fewer ulcers for 7.5 more heart attacks.
Now I ask you, how many Americans, if given all the facts, would trade the low risk of an ulcer for the low risk of a heart attack for the same pain relief, especially when the over-the-counter drugs are one-tenth the price?