June 10, 2005
Merck's Singh Sings the Blues
NPR last night ran a revealing report about Merck’s efforts to silence potential critics of its blockbuster painkiller Vioxx. It introduced the story with this damning statistic: “At least 38,000 Americans are believed to have died from taking the pain pill.”
Relying on internal Merck documents turned up by trial lawyers representing the victims’ families, reporter Snighda Prakash revealed how marketing officials and a Merck senior scientist sought to silence one of their own consultants, who had become upset by signals from a clinical trial that Vioxx might be causing heart problems among its millions of users.
Who was that consultant? It was none other than Stanford researcher Gurkirpal Singh, who bears at least some responsibility for the fact that millions of Americans were taking Cox-2 inhibitors in the first place. The new class of pain relief medicines was designed to lessen the gastrointestinal side effects of generic pain pills like ibuprofen, naproxen and aspirin, the problem that was the focus of Singh’s own research. Instead, it wound up killing tens of thousands of Americans.
Prakash told an important story about the power that drug companies exercise over academic medicine, and for that reason alone the transcript is worth reading. But she stumbled badly in her portrayal of Singh as, in her words, an “independent scientific expert.”
Singh’s original study showed that traditional painkillers caused 16,500 deaths and 100,000 hospitalizations each year. The study was paid for by the drug companies that made Cox-2 inhibitors. After publication in a minor journal, the statistics were trumpeted in company press releases and appeared in mainstream media stories about the new “super-aspirins,” Pfizer’s Celebrex and Merck’s Vioxx, which were no more effective at halting pain than the older generation of drugs.
How did Singh arrive at his numbers? He got his death and injury ratios from a limited database made up of people suffering largely from serious rheumatoid arthritis, who take high doses of painkillers. Singh then extrapolated those numbers to the larger population – 17 million people – who popped painkillers for the more common (and less painful) osteo-arthritis. Dr. James Fries, Singh’s colleague at Stanford and also an arthritis specialist, recognized Singh’s estimates were way too high. No matter. Merck was willing to pay him tens of thousands of dollars a year to trumpet those numbers on the caviar-and-golf continuing medical education circuit.
When signals appeared in clinical trials that his sponsor’s drug might be causing heart problems, Singh, to his credit, began raising questions about the drug. But ever the industry sycophant, he went about it in his own way. The telling fact slipped past quickly in Prakash’s report, so it is worth highlighting:
“Singh's allegiance was shifting. He was now promoting Vioxx's rival, Celebrex. He was being paid by Pfizer, and he was telling his audiences that Merck had refused to answer his questions about Vioxx's safety.”
Mind you, this was about the time that the Cleveland Clinic’s Eric Topol and other physicians were looking through the same published trial on Vioxx that showed the heart attack risk. Letters of concern began appearing in medical journals. Topol prepared a paper for the Journal of the American Medical Association. But that wasn’t Singh’s way. He went to work for Merck’s competitor.
Singh’s science or ethics haven’t improved much over the years. A few months back, I called Singh after he once again repeated his “16,500 deaths” statistic for a story that appeared on the front page of the New York Times. He got back to me too late for my story (see GoozNews: The Press on Pain). But he stuck by his numbers. In fact, he sent me an abstract of his latest study, which hadn’t yet been peer reviewed or published.
It reads like a press release for Pfizer, the last remaining Cox-2 maker. Using a national sample drawn from community hospitals and extrapolating for the entire U.S. population, he estimated that total hospitalizations for gastrointestinal ulcers fell from 166,725 in 1988 to 139,597 in 2001. There were two periods of sharp declines over that 13-year period, according to his study, 11 percent in 1995 and 8 percent in 1999. He attributed the former to the 1995 National Institutes Health consensus conference recommending eradication of H. pylori, the bacteria that causes ulcers; and the latter to the introduction of Cox-2 inhibitors.
I pulled out my trusty calculator. The total decline over the period was 16.3 percent. Singh’s numbers were mathematically impossible. Moreover, a decline of 12,000 hospitalizations a year – perhaps saving the health care system $250 million annually – pales beside the $5 billion a year spent on Cox-2s.
Merck’s efforts to silence academic critics of its blockbuster drug, as documented in Prakash’s report, is unconscionable. But it’s hard to defend someone’s academic freedom when it’s already been sold to the drug industry.