Phil Brewer thought he knew exactly what to do when the ambulance crew wheeled a well-dressed man in his late sixties into the emergency department. What he didn’t know: He was about to be involved in a series of events that would kill his patient. Brewer, then an assistant professor of emergency medicine at Yale School of Medicine, had been alerted by the crew that the man, Sanders Tenant (a pseudonym), had suddenly begun to talk gibberish while dining out with his family. Then his right arm and leg had gone weak.
Brewer suspected an acute stroke, but first he had to rule out conditions that can masquerade as a stroke, such as low blood sugar, a seizure, a brain tumor, and migraine headache. He had only minutes to make the correct diagnosis. Then the gathering medical team would decide whether to use a new stroke treatment that had recently been approved, a clot-buster known as tPA. Brewer called in a neurologist and the stroke team. After a CAT scan of the patient’s brain showed no sign of bleeding (something that would prevent the use of a clot-buster), the decision was made: Yes, use tPA. Despite following each step of the established protocol for this new treatment, Brewer experienced the unthinkable—his patient’s death. Tenant suffered a massive brain hemorrhage and died, not from his stroke but from effects of tPA, the drug that was meant to save him.
When we go to the doctor, we assume that the drugs he or she prescribes have been carefully tested to make sure they are both safe and effective. Most times they are. Yet sometimes the drugs cause more problems than they solve. Adverse drug reactions kill tens of thousands of people annually; one widely cited study published in the Journal of the American Medical Association (JAMA) in 1998 puts the number at more than 100,000. Recently a series of drug recalls have pulled back the curtain to show how the media, the public, and some doctors can misinterpret medical studies or take them out of context in ways that make medical treatments look safer and more effective than they actually are.
To a greater degree than ever before, powerful forces in the marketplace are impacting the quality, use, and safety of prescription medications. Drug manufacturers are spending more to promote their products while being subjected to tighter regulation and greater pressure for financial returns. The media are talking up each new “miracle cure” in headlines and television segments. Doctors have to navigate a tangle of administrative and medical concerns, one physician noting that “if you have a patient in your office, you can’t say, ‘Oh, I’m going to look at the drug company’s online database about Zyprexa.’ Most doctors don’t even know the databases exist. But even if they did, the next thing you know, three or four hours have gone by and you’ve missed all the patients waiting to see you.” Insurance companies and even the stock market play a role too. And consumers, increasingly subject to pharmaceutical advertising, are routinely urged to demand the best and the newest for their health. All together, this is a perfect storm for prescription drug problems.
How often do today’s medical “breakthroughs” become tomorrow’s discredited science? John P. A. Ioannidis, an epidemiologist at Tufts University School of Medicine in Boston and the University of Ioannina School of Medicine in Greece, studied the question. He examined the most-cited clinical studies published in the top three medical journals between 1990 and 2000 to see how well researchers’ initial claims held up against subsequent research. His findings, published in JAMA, show that the key claims of nearly one-third (14 out of 49) of the original research studies he examined were either false or exaggerated. Small study size, design flaws, publication bias (failure to publish negative results or duplication of positive results), drug-industry influence, and the play of chance were among the problems Ioannidis found that caused false or exaggerated claims.
Studies can be designed and interpreted in ways that make even ineffective drugs seem like lifesavers, says Curt Furberg, a well-known cardiovascular epidemiologist and former chief of the clinical trials branch at the National Heart, Lung, and Blood Institute in Bethesda, Maryland. Furberg, a tall, square-faced man with a Swedish accent, wants more objectivity in medical research. “We need more publicly funded studies,” he says, adding that manufacturer-sponsored research tends to minimize risks and exaggerate benefits.
A score of studies support his opinion. Among them is a 2003 analysis by Cary P. Gross, an associate professor of medicine at Yale School of Medicine, that was published in JAMA. In his survey, one study found that industry-sponsored research was positive 87 percent of the time compared with 65 percent positive for research that was not industry sponsored. According to Gross, the evidence was overwhelming that “industry sponsorship was likely to yield pro-industry results.” A 2006 analysis published in the American Journal of Psychiatry found that 90 percent of manufacturer-sponsored studies of antipsychotic drugs led to claims that the study drug was as good as, or superior to, every other drug in its class. Shannon Brownlee, an award-winning medical writer based in Washington, D.C., ascribed this to the “Lake Wobegon effect,” which renders every drug “above average.”
Furberg’s efforts to debunk overly enthusiastic interpretations of medical studies have led to occasional clashes with his colleagues. In 2004 the U.S. Food and Drug Administration (FDA) was preparing to hold hearings on the safety of painkillers known as COX-2 inhibitors, including Vioxx, which David Graham, an official in the FDA’s Office of Drug Safety, said may have caused an estimated 39,000 to 60,000 heart-attack deaths in just five years. At the time, Furberg was a member of the FDA Advisory Committee on Drug Safety and Risk Management. But after he told The New York Times that the COX-2 inhibitor Bextra also caused heart attacks, the agency made a surprising move: It removed Furberg from the advisory panel. Sandra Kweder, acting director of the Office of New Drugs, Center for Drug Evaluation and Research at the FDA, told a reporter that Furberg’s comments showed he could not be objective. Furberg now asks, “If bias was a concern, why did they allow 10 advisory members with ties to the manufacturers to be seated?” He was reinstated to the panel two days later and vindicated when the FDA announced that it had asked Pfizer to voluntarily withdraw Bextra from the market.
Part of the difficulty in detecting drug side effects, Furberg says, has to do with study size. Drugs go through a regimen of tests prior to receiving approval from the FDA. During the first two stages, called Phase I and II trials, an experimental drug is tested on just a few hundred volunteers to look for side effects. If no serious problems are detected, the drug is tested for efficacy in a Phase III trial. But efficacy trials often involve only several hundred to a few thousand patients. And while a study of 200 to 300 arthritis patients is large enough to show whether a new drug relieves pain, just one such study isn’t large enough to pick up less-common—but potentially deadly—side effects. Furberg says, “If only one in a thousand patients will die from a heart attack, an efficacy study of 200 or even 2,000 patients is simply too small to get a reliable answer about rare side effects.” Seemingly rare side effects can take tens of thousands of lives when millions of prescriptions are written.
Such limitations in a study’s design can escape detection even by top peer reviewers and medical editors. Marcia Angell, the former editor-in-chief of The New England Journal of Medicine (NEJM), says that most doctors are ill equipped to critically assess the conclusions of researchers. A trim woman with a warm smile, Angell leans forward in her seat at her home in Cambridge, Massachusetts, and says, “Let me tell you the dirty secret of medical journals: It is very hard to find enough articles to publish. With a rejection rate of 90 percent for original research, we were hard pressed to find 10 percent that were worth publishing. So you end up publishing weak studies because there is so much bad work out there.” Doctors, Angell says, are not skeptical enough about what they read in top journals. “They should say, ‘I don’t believe this; prove it to me.’”
The rest of the media don’t get any better marks. Gary Schwitzer, director of graduate studies in health journalism at the University of Minnesota School of Journalism, examined 400 medical news stories that were carried by 57 of the top print and broadcast media. “The majority failed to adequately discuss costs, quantify harms and benefits, and examine the quality of the studies,” Schwitzer says. Many quoted a sole source and failed to report potential financial conflicts. Schwitzer concluded that media reports give “a kid-in-the-candy-store portrayal, where everything is made to look amazing, harmless, and without a price tag.” Patients themselves “should not escape notice as willing collaborators, wishing for magic potions and taking drug company money to support consumer organizations,” Schwitzer continues.
Phil Brewer, the doctor whose stroke patient died after being treated with tPA, says media portrayals of new medicines are often “irrationally exuberant.” He points to a May 2007 article in The New York Times that he says typifies the problem. The article, about stroke victims, said that the clot-buster “tPA was shown in 1996 to save lives.” Yet in 2001, the American Heart Association (AHA) had withdrawn the claim that the drug “saves lives” from its promotion of tPA for stroke after the group was challenged to provide scientific evidence to support that claim. The AHA was also the subject of scrutiny when it was revealed that in the decade prior to its recommendation that doctors use tPA for stroke victims, the heart association had received $11 million from Genentech, tPA’s manufacturer.
The same Times article quoted a number of doctors saying that too few stroke patients were receiving tPA, yet failed to mention that many of these same doctors had received funding from Genentech. Nor did the article give a hint of the ferocious battle among doctors about the safety and efficacy of tPA: While a number of professional associations endorsed the drug, many others, such as the American Academy of Emergency Medicine, said it should not be considered the standard of care for acute stroke.
Asked about this reporting, Barbara Strauch, health editor of The Times, responded, “While some researchers had said in interviews that they believed the drug saved lives, our article incorrectly stated that the study had made that conclusion.” Strauch said the paper would publish a correction—which it did this past April, nearly a year later. (This was done in response to DISCOVER’s inquiries.) She added, “It is also true that some researchers quoted in the article, like many stroke researchers and many who study other diseases, are funded by and receive honoraria from the pharmaceutical industry. However, the main sources for our article were researchers at the National Institute of Neurological Disorders and Stroke (NINDS), who, the National Institutes of Health says, do not take money from drug manufacturers.”
“What you read in the media are these stories of a stroke patient getting tPA and miraculously improving within minutes,” Brewer says. “But we’ve all seen that happen in the ER, even before tPA was ever invented.” After his patient died, Brewer, wary of drawing conclusions on the basis of a single case, wanted to make sense of the data about tPA. Even though a landmark 1995 study conducted by NINDS showed that 12 to 13 of every 100 treated patients had less disability, Brewer says that “it was hard to put the conflicting [study] results together and determine whether the benefits really did outweigh the risks.” So, like many physicians, he turned to the articles and analyses of Jerome Hoffman, a professor of medicine and emergency medicine at UCLA. “Dr. Hoffman has a brilliant mind,” Brewer says. “He is listened to and trusted by more emergency physicians than anyone I know.” An authority on medical studies, Hoffman, it turns out, was also the lone dissenting member of the AHA panel that recommended tPA for stroke.
Tall, white-haired, and wearing thick glasses, Hoffman looks the part of an elder statesman of medicine. He says he became interested in the interpretation of medical literature when he was just starting out as a resident physician at UCLA. He read studies and their interpretations voraciously, and eventually other physicians began coming to his talks to residents and medical students on how to interpret the medical literature. “Some studies just didn’t make sense to me,” he says. “I was reading all these things that came to opposite conclusions. They couldn’t all be right.” Besides, Hoffman says, “there were studies that didn’t represent what I was seeing in clinical practice.”
When the NINDS study was published in December 1995, Hoffman paid attention. “It was a big deal,” he says. “If tPA worked, it would be a real advance over what we could offer patients during an acute stroke.” But, he says, “you [should] never believe one study—especially of a drug that has only a small benefit, and especially a study that is contradicted by other studies, as was the case here.” Hoffman is also critical of a study that Genentech says supports the NINDS trial findings. He contends that this study, known as SITS-MOST, shows “how study design and spin can inflate perceived benefit.” This is because “no patient with a severe stroke was allowed into SITS-MOST—by design—so the patients who did get included were virtually certain to do well as a group, no matter what treatment they did or didn’t get. Comparing them to the much sicker patients in the big trials isn’t like comparing apples with oranges—it’s comparing apples with elephants.”
Hoffman says that the truth in any drug study can be camouflaged by how it is reported. “One way that’s been done—for many treatments, and not just [clot-busters]—is to use combination end points.” Here’s how it works: A single drug can be tested for a variety of outcomes; for example, a cholesterol-lowering drug can be tested for its effect on cholesterol level, blood pressure, and/or rates of heart failure, heart attack, or death. By combining two or more of these outcomes to create a single category, you can say it helped “A and B” even if it only helped A and not B. For example, although there was no statistically significant effect from tPA in the NINDS trial on the number of patients who died, there was a small decrease in disability for those who survived. With the two factors combined, there was technically a decrease in the combination end point of “death and disability.” From there, it’s a short step to the incorrect assumption that death and disability were each decreased—an assumption made by many physicians and patients.
David L. Brown, chief of the division of cardiovascular medicine at the State University of New York, Stony Brook School of Medicine, calls the use of combination end points a “brilliant marketing tool.” Brown says that while combination end points have a legitimate purpose when researchers are testing drugs for a rare or infrequent outcome, far too often researchers use the information in ways that “mislead both doctors and the general public.”
Another way to make drugs look better and safer than they are is to report or cite only successful studies while ignoring those with bad outcomes. The problem of cherry-picking studies is a very real one, especially for antidepressants, says Erick Turner, a former FDA reviewer, now an assistant professor of psychiatry at the Oregon Health & Science University. Turner recently published research in NEJM showing that “when studies of antidepressants were negative, they were reported as negative only 8 percent of the time—but when studies were positive, they were reported as positive 97 percent of the time.”
Genentech acknowledges that no controlled study has ever shown—or been conducted to show—that tPA “saves lives” in cases of acute stroke. What the NINDS study showed, Genentech spokesperson Krysta Pellegrino says, is that “patients were at least 30 percent more likely to have a decrease in stroke-related disability three months after treatment compared with placebo.” Although Genentech admits that high-risk patients were excluded from analysis in SITS-MOST, Pellegrino says the company believes the data from that study “add to the body of evidence that supports the conclusion that tPA is safe and effective for the treatment of acute stroke.”
Ken Johnson, senior vice president of Pharmaceutical Research and Manufacturers of America (PhRMA), says that FDA drug approval procedures are “the gold standard of the world” and that the United States has “one of the strongest drug safety records.” He acknowledges, though, that “adverse reactions are sometimes not detected until a medicine has been approved and made available to an entire population,” adding, “That’s why the postmarket surveillance system is so important.” He notes that under the Food and Drug Administration Amendments Act of 2007, the agency has new authority to “require additional postmarket studies and make faster changes to product labeling.” The intent is to reduce the time it takes to detect adverse drug reactions and protect the public.
An adverse reaction is exactly what Duane Graveline suspects happened to him. Graveline, a former NASA astronaut and flight surgeon, suffered a bizarre episode in 1999 shortly after he was prescribed the popular statin drug Lipitor for his elevated cholesterol. Just six weeks after he began taking the drug, the normally very active and healthy Graveline plunged down the rabbit hole when he abruptly lost his memory. His wife rushed him to the hospital, where doctors examined him carefully but could find no medical or psychiatric problem. His brain scan showed no sign of a stroke or brain disorder. Then, almost as quickly as his memory had disappeared, it came back after just six hours, without any treatment. Doctors termed the strange episode transient global amnesia, or TGA. The cause? Unknown.
Graveline’s case was certainly unusual. Most people with sudden memory loss have suffered a blow to the head, a stroke, or some other medical problem. But Graveline had no history of medical or psychiatric troubles. This made him wonder: Could the episode have been a side effect of Lipitor? Deciding not to test fate, Graveline stopped taking the drug. For the next year he was fine. But when his annual astronaut’s physical exam rolled around, he was told that his cholesterol level had crept up again, and his doctor urged him to restart the Lipitor.
Graveline, still uncertain about the connection between the drug and TGA, complied. Within 10 weeks he suffered another, even more severe episode of amnesia. His wife found him wandering outside their home, unable to recognize her and unaware even that he was a physician. That episode also resolved without treatment. “That’s when I decided never to take statins again,” he said. That’s also when Graveline began scouring the medical literature for an explanation of what had happened. What he found troubled him: There were few studies examining statins’ side effects on memory, even though cholesterol, he says, plays an important role in brain function. Graveline worried: What if he had had an episode while he was driving? What if a pilot developed TGA during flight?
Graveline’s research eventually brought him into contact with Beatrice Golomb, an associate professor of medicine at the University of California, San Diego School of Medicine. Golomb, a dark-haired whiz kid who graduated from the University of Southern California with highest honors at age 19, is an M.D. with a Ph.D. in biology and has been studying cholesterol and statins for more than a decade under research grants given by the Robert Wood Johnson Foundation and the Harry Frank Guggenheim Foundation.
Golomb says that Graveline’s episodes of TGA, and other cases like his, raise questions about the ways statins can affect memory. Her research, she says, shows that although statins can reduce the risk of heart attack, they may also have serious side effects. In Golomb’s opinion, the potential benefits of statins may not outweigh their risks except among middle-aged men who have heart disease—or who are at high risk for it. The only way to weigh risks against benefits, she says, is to evaluate all-cause morbidity (sickness) and all-cause mortality (death).
Christopher Loder, a spokesperson for Pfizer, the maker of Lipitor, says that studies of the drug were “not designed nor powered to look specifically at all-cause mortality.” The studies, he says, “were powered and designed to look at a composite end point consisting of heart attack or death from coronary causes.” In addition, Loder says, “There is overwhelming clinical evidence to support the benefit of Lipitor. All statins have been shown to reduce LDL cholesterol.”
Golomb says one reason many doctors overlook risks and believe statins to be safe is that most controlled studies of statins wind up excluding people who originally begin to participate in a study but stop taking the drug because they experience problems from it; these test participants are then dropped from the study as “noncompliant.” Confusion arises, Golomb says, “because the absence of evidence that statins cause harm—having excluded those who would have permitted detection of harm—is interpreted wrongly as evidence of absence of harm. And the treatment is generalized to a larger population with a very different risk-to-benefit profile.”
John Abramson, a clinical instructor at Harvard Medical School and author of Overdo$ed America: The Broken Promise of American Medicine, says he grew concerned when he learned that the authors of professional guidelines recommending an expanded use of statins had ties to the drugs’ manufacturers. So, Abramson, a tall, dark-haired man with owlish glasses, decided to review the study data. What he found stunned him. Statins could reduce heart attacks and strokes—but only in a small fraction of the people taking the drugs. “Doctors give statins in one of two ways,” Abramson explains. “The first way is to give the drugs to people with elevated cholesterol as primary prevention—to prevent a heart attack, stroke, or other serious cardiovascular event. [These are] people who have never suffered any of those events. The other way to give statins is as secondary prevention, after people have had one of those events or develop diabetes.”
Despite broad recommendations in the National Cholesterol Education Program guidelines, Abramson found that there were no studies that showed statins were beneficial for primary prevention for women of any age or men over 65. Yet more than three-quarters of people taking statins take them for primary prevention—meaning that many patients stand to gain no benefit at all. Abramson, who with a colleague published his findings in the British medical journal The Lancet, says that even when statins are used for men at the highest risk, “you have to treat about 238 men for one year to prevent one heart attack.”
Another problem with statin studies, according to Abramson, is that many do not measure clinically and critically important outcomes like heart attacks, serious adverse events, or all-cause mortality. Instead they measure surrogate markers—outcomes that are associated with a risk of disease—but not a bad outcome itself. In the case of statins, the surrogate marker most commonly used is cholesterol levels. If a drug reduces cholesterol, it is said to be “effective.” But lowering cholesterol doesn’t necessarily mean a drug will reduce the bad outcomes people are worried about—such as death or heart attack.
This was the issue in last winter’s congressional investigation into the nonstatin cholesterol-lowering drug ezetimibe, sold as Zetia and contained in Vytorin. Hearings in January revealed that the release of negative results of a clinical trial of ezetimibe had been delayed. The drug, while lowering cholesterol effectively, failed to slow the progression of carotid artery plaque. While manufacturers Merck and Schering-Plough delayed the negative study’s release for more than 18 months, ezetimibe had turned into a blockbuster drug, even though it had never been shown to reduce heart attacks or deaths.
“You can lower cholesterol levels with a drug, yet provide no health benefits whatsoever,” Abramson says. “And dying with a corrected cholesterol level is not a successful outcome in my book.” Suddenly Abramson, who had taken many hits for his critiques of cholesterol-lowering drugs, was joined by physicians calling for more openness in research and more careful examination of the evidence before drugs are put on the market.
Jerome Hoffman of UCLA agrees, saying it is a shame that in the face of so many medical and pharmacological advances there is such an exposure to risk. “It’s ironic that one of the unintended consequences of the publication of so many untenable claims, based on poorly done research and spin, is that it can obscure true advances when they do occur,” he says. Citing the number of great successes medical research has brought us—lifesaving drugs from penicillin to insulin, along with invaluable treatments and medical devices—he adds, “It’s one more reason why we have to be appropriately skeptical, unafraid to speak out about misleading claims, and insistent upon holding clinical research to the standards of science.”