Bitter Pill.

Rezulin may cause acute liver failure, but the FDA still won't take it off the market

THE FIRST SYMPTOMS OF IMPEDING liver failure typically are dark-colored urine, fatigue, nausea, loss of appetite--not unlike a touch of the flu. And that's precisely the initial misdiagnosis made by doctors across the country, doctors who long were clueless about the risks posed by Rezulin, a new drug for people with diabetes. The diagnosis can soon enough get easier--though by then it may be too late to save the patient. The skin or eyes turn yellow with jaundice. Fluid sacs, called ascites, can form in the abdominal cavity.

Within weeks, other organs can start tipping toward failure. The body's own capacities compromised, the patient is left poisoned in her own toxins. And when it comes to Rezulin, gender matters: Approximately 70 percent of those known to have suffered liver failure from this drug are women, according to data studied by the FDA. I have read the all-too-repetitive daily hospital charts of Rezulin patients who lost their lives to the drug. I have spent days interviewing the patients' loved ones and doctors. Ask a hospital RN: Liver failure is a wretched way to die.

By January 1997, the Food and Drug Administration's files were loaded with enough warning signs to have kept Rezulin out of the hands of unsuspecting diabetic patients. Laboratory studies suggested that Rezulin played havoc with the livers and hearts of rodents. In rats, scientists found the livers were enlarged, with areas of dead tissue. And the rat hearts were distended, dilated, and discolored. "These changes were drug-related, and were responsible for the early mortality in both sexes," FDA pharmacologists concluded in a January 1997 internal report.

Clinical studies of Rezulin's effects on humans provided further cause for alarm: About 1 of every 50 people who took this drug in the original clinical studies developed some degree of liver injury. More than one fourth of those patients with liver injuries experienced toxicities that spiked to potentially life-threatening levels. Overall, nearly four times the rate of liver injury was observed in Rezulin patients than in those who took a placebo in the controlled clinical trials. In the parlance of drug-development scientists, this amounts to a "signal" of a dangerous problem. And a clear one at that.

With any drug, the risks have to be weighed against the potential benefits. Rezulin was developed to combat adult-onset, "Type 2" diabetes, which afflicts 15 million Americans, and can result in complications that include blindness and amputations. (The condition is distinguished from juvenile-onset, Type 1 diabetes, in which patients are incapable of producing their own insulin and would die without daily infusions. Although I am aware of doctors who have done so, Rezulin certainly should not be prescribed to Type 1 diabetics.) The question that faced the FDA: All things considered, was Rezulin safe enough?

The call was not a close one for Dr. John L. Gueriguian, the veteran agency medical officer who was assigned to evaluate the drug. In a detailed review dated Oct. 9, 1996, Gueriguian, citing Rezulin's potential to harm the liver and the heart, recommended the compound be rejected. "[I]t is unwise," Gueriguian wrote, "to force the FDA to hastily introduce a drug into the marketplace with such potential for worrisome toxicity."

Initially, Gueriguian was not alone in questioning Rezulin's safety. Another FDA medical officer, Dr. Robert I. Misbin, took specific note of Gueriguian's concerns regarding potential liver and heart toxicities. In an Oct. 18, 1996 agency memorandum Misbin wrote: "My primary concern about troglitazone [the chemical name for Rezulin] is related to its potential for cardiac toxicity.... [I] do not believe that the sponsor has made a convincing argument that failure to make troglitazone rapidly available would put a significant number of diabetic patients at risk. I believe the greater danger may be to make troglitazone available to patients who do not really need it before we have been convinced of its safety." Newly obtained records show that a study of Rezulin's effect on human hearts ended without definitive results.

No matter. On Jan. 29, 1997 the FDA (with Dr. Misbin's assent) pronounced Rezulin fit for sale, a safe and effective prescription drug. It was the FDA's fastest-ever approval of a diabetes pill.

There was, however, at least one subtext of considerable intrigue: In November 1996, senior FDA officials had stripped Gueriguian of any further involvement with the review of Rezulin, following his use of intemperate language in a meeting with a drug company executive. According to Gueriguian, he tired of the sponsoring company's tactics and told the executive: "You can't shine shit with words." His forced removal robbed the FDA of its best source of institutional memory at a time when the agency was racing to complete a six-month, "fast-track" review of the drug.

Fast forward just over two years, to March 26, 1999, to a Holiday Inn in Bethesda, Md., its upstairs conference room brimming with cell-phone-toting surrogates of Wall Street. The FDA by this time was in the midst of an extraordinary re-evaluation of Rezulin, undertaken in response to a series of articles that I had authored for the Los Angeles Times. Among other things, the series documented the steady stream of liver-failure deaths and transplants associated with Rezulin. Now, the FDA's Endocrinologic and Metabolic Drugs Advisory Committee was being asked to weigh in.

This same panel had given Rezulin a unanimous endorsement in December 1996, paying scant attention to the issue of liver toxicity. And on this day, the front rows were packed with no small number of practicing doctors who also happened to be paid consultants to Warner-Lambert Co., the maker of Rezulin. One by one, they told the committee that they hadn't personally seen much or any evidence of liver toxicity. I wondered how many of their patients knew about the fees they had been taking from the drug company.

Enter into this commercially steeped milieu one Dr. David J. Graham, the FDA's top specialist in epidemiology, the discipline that examines the prevalence and means of controlling drug-related deaths and injuries. Graham, 45, was trained at Johns Hopkins and Yale medical schools. Unlike several of the participating advisory committee members, he had no financial ties to Warner-Lambert. His sole employer was the United States government; his client, the public interest. With the annoyed, and at moments palpably irritated, indulgence of the advisory committee's chairman--a physician who presided with his own freshly issued FDA conflict-of-interest "waiver" in hand--Graham presented a grisly report card to the advisory committee: One of every 1,800 Rezulin patients can be expected to develop liver failure, most of whom will die or, if they're lucky, undergo organ transplant. He or she is at least 1,200 times more likely to suffer liver failure by going on Rezulin than by staying off it. Graham estimated that Rezulin use had resulted in approximately 430 cases of acute liver failure.

Ahem. These numbers, safe to say, clashed with what both the FDA and Warner-Lambert had long been cooing to doctors, patients, and news reporters. The FDA, at times unable to glean up-to-date statistics from its own computer system, has consistently understated the number of reported, Rezulin-associated deaths. As for Warner-Lambert, the company's spokesman, Stephen J. Mock, told me last December that if Rezulin patients complied with liver-monitoring guidelines recommended by the company and the FDA, "at most, one in several hundred thousand" patients would die or require a transplant. Variations of this claim formed the essence of the company's Rezulin reassurance message: Not to worry, doctor. Look at our numerator and denominator--the number of confirmed dead vs. the one million or more patients who have taken the drug. It amounts to an acceptable risk. Odds are it's going to be someone else's patient who dies.

Of course a central flaw to this argument is that adverse events related to prescription drugs are reported only voluntarily--and rarely--in the United States. Independent experts have studied the issue and concluded that, routinely, just one to 10 percent of these events are made known to pharmaceutical companies or to the FDA. Doctors and nurses say privately that they are too busy to be bothered with the additional paperwork. And who wants to raise a red flag that might prompt a victim or survivors to sue?

Yet when Dr. Graham told the advisory committee that, based on his research, as few as 10 percent of Rezulin-related adverse events had been reported, members of the panel would have none of it. The committee's reluctance to embrace a more excoriating view of Rezulin was perhaps best explained by another panelist, Dr. Robert A. Kreisberg, an endocrinologist from Birmingham, Alabama. After hearing hours of conflicting presentations from the FDA's Graham and from Warner-Lambert, Kreisberg summed up his bottom line.

Gesturing toward the front rows of the audience, where sat 10 physicians who had urged the advisory panel to stand behind Rezulin, Kreisberg said: "I'm going to rely very heavily on valued colleagues who practice and take care of patients with diabetes on a day in and day out basis.... I think this information is as valuable or more valuable than the theoretic issues that have been brought up here [by Graham]." What Kreisberg didn't say (or know) was that nine out of ten of the doctors he was pointing to were paid consultants to Warner-Lambert or one of its affiliates.

More later about the curious handling of conflict of interest at today's FDA. First, back to those vaunted liver-monitoring guidelines, which Warner-Lambert held up in December 1998 as a safeguard against liver damage from Rezulin. In fact, the guidelines didn't even emerge until Rezulin had been on the market for eight months, and only after the first liver-failure victims were dead and buried. Despite Dr. Gueriguian's protests and Dr. Misbin's initial reservations, no one at the FDA had suggested that monthly liver testing be a condition for approving Rezulin. Such a condition would certainly have undermined the carefully crafted luster Warner-Lambert gave the drug on its March 1997 launch. Recommended liver testing also would have been, well, inconsistent, with the company's magazine-ad claim: "Side effects comparable to placebo."

Yet how tenaciously would these liver-monitoring guidelines come to be embraced by top FDA executives, as a basis--some would say a pretext--for keeping Rezulin on the market. In a statement issued Nov. 3, 1997, the FDA endorsed liver-function testing "within the first one to two months [and] every three months thereafter" And the agency offered this added assurance: "Few, if any, of these patients will go on to develop permanent liver damage if the drug is stopped."

Just one month later, on Dec. 1, 1997, the FDA reacted to the news of additional liver-failure deaths from Japan by recommending more liver tests, "every month for the first six months, every other month for the next six months." And again, the federal agency offered an added assurance: "The deaths in Japan occurred in patients treated before the stronger label warning and recommendation for liver [tests] took effect there." Translation: If the Japanese patients had followed the regimen of liver testing endorsed by the FDA as of November 1997, they might have survived. In Britain, Rezulin was withdrawn from the market for safety reasons just prior to the FDA's Dec. 1997 statement. The drug had been sold there for just three months.

The FDA endorsed a third safety-related change for Rezulin in July 1998, agreeing with Warner-Lambert to recommend an extension of the liver testing, from six monthly procedures to eight. Officials said that they remained convinced Rezulin was safe, provided that patients complied with the recommended liver testing.

This FDA mantra persisted even after December of 1998, when our first Rezulin work was published in the Los Angeles Times. Indeed, when Dr. Janet Woodcock, one of the FDA's top executives and a supporter of fast-track drug review, was asked what it would take for the agency to pull Rezulin off the market by "Nightline's" Chris Wallace on Dec. 23, 1998, she replied: "If the monitoring program is not successful and we have evidence that, in fact, the same rate of liver disease is occurring without any warning or ability to control that, then we'll have to take another look at this drug."

Four months later, Woodcock was in the packed audience at the Bethesda Holiday Inn, listening as Dr. Graham told the advisory committee what he had ascertained about the "monitoring" of Rezulin patients. Citing the most up-to-date scientific data available, Graham said that more than 99 percent of the patients taking the drug were found not to be complying with the liver-monitoring guidelines after just four months of use. After studying the clinical courses of patients who suffered liver failure and died, Graham concluded that 75 percent of them fell sick too fast for monthly monitoring to have worked. "We can't point to monitoring and say that this intervention has had any impact on the incidence of acute liver failure," Graham told the committee. Rezulin can so quickly shock the liver, Graham explained, that a person whose test results were normal at the beginning of the month could be in mortal jeopardy just a week or so later. Compared to the predictability and severity of side effects to be expected, generally, from all medications, Graham said, "[Rezulin] looks very different than most other drugs."

As the meeting broke up, I approached Dr. Woodcock. Was she, at last, persuaded that Rezulin was too dangerous to remain on the market? Hardly. It was clear that she was retreating from Grahams presentation as far and as fast as physics would allow. Here's a transcript of my brief interview with her that afternoon.

Question: "The top FDA epidemiologist said here today that the expected rate of acute liver failure is one in 1,800 patients. Does this cause you to reassess the drug?"

Woodcock: "David Graham felt that one in 1,800 was his best estimate. But it's based on a variety of judgments and assumptions."

Question: "Do you have reason to doubt his judgments and assumptions?"

Woodcock: "We aren't doubting his judgments and assumptions. We're saying that there is a very broad range, as Dr. Bone [the advisory committee chairman] said, of best guesses"

Would that be the same Dr. Bone who had been granted an FDA conflict-of-interest "waiver"? Yes indeed. But the nature of Dr. Bone's interest was not disclosed to the press or public. At today's FDA, officials say they no longer are required to disclose the facts and circumstances that

underlie any such waivers. In a bit of devolution that might best be called Transparency Trumped, it's up to the office of FDA Commissioner Jane E. Henney to decide whether an advisory committee member's medical expertise outweighs the conflict. And the telling details are no longer the public's business. For his part, Dr. Bone refused after the March 26 meeting to discuss the nature of his potential conflict of interest ("My statement was read by the secretary and I have nothing to add to that.") Ditto for two of Bone's other conflicted colleagues on the panel.

The financial interests of part-time appointees to local planning commissions and transit boards are typically wide open to view in this country. But at the FDA, the "independent" experts--with material influence over which drugs get approved and/or whether they stay on the market--operate in a confidential cocoon. All that's at stake are the lives of millions of Americans.

Money Pill

Whatever the financial interests of the FDA officials, there's no question that Rezulin's maker, Warner-Lambert, was banking on it to become a "blockbuster." Beginning in the early '90s, the company's financial statements had tabbed Rezulin as a star-in-waiting. By mid-decade, with its pharmaceutical division profits slumping, the fate of Rezulin loomed ever larger. The company's president and chief operating officer said in a 1995 address that Rezulin "has the potential to redefine the diabetes market." In its annual report issued in March 1996, Warner-Lambert termed Rezulin, "A Breakthrough Drug," that "may delay the onset of diabetes." (It has not been proven that Rezulin can prevent either Type 1 or Type 2 diabetes) Although the drug had at this point not yet been approved by the FDA, the Warner-Lambert annual report trumpeted Rezulin's potential. "The company believes this breakthrough therapy, once approved, could become one of the largest-selling pharmaceuticals it has ever marketed."

Warner-Lambert assured investors in spring 1996 that it was "working aggressively to accelerate" the FDA's approval of Rezulin. The work appeared to pay off: The FDA quickly assured Warner-Lambert that it would grant a six-month, fast-track review. The timing was ideal. By July 1996, when the FDA began reviewing the company's "New Drug Application" for Rezulin, forces hostile to strong pharmaceutical-industry regulation were in control of both the House and the Senate. Legislation with bipartisan support had moved in both houses, threatening to severely limit the FDA's drug review powers. The bills ultimately stalled at the end of the 104th Congress. But the seed of fear was sown at the FDA, undermining vigorous scrutiny of newly proposed drugs, according to government physicians. And Warner-Lambert, not unlike other drug companies, was poised to capitalize.

But perhaps the most timely, government-conferred cachet gained by Warner-Lambert came just a few miles south of the FDA's offices, at the National Institutes of Health. Although the drug had not yet been approved by the FDA, NIH officials and affiliated researchers in late 1995 and early 1996 selected Rezulin for use in a $150 million diabetes-prevention study. Warner-Lambert touted the NIH decision to investors; the selection created a momentum that made it even more unlikely that officials at the FDA would block or reject the drug.

Then came May 17, 1998, the day the NIH's selection of Rezulin boomeranged, tragically. This was the day when a participant in the NIH study, a 55-year-old high school teacher from East St. Louis, Ill., named Audrey LaRue Jones, died. Doctors who investigated said she had suffered Rezulin-induced liver failure. Within a month, NIH officials banished Rezulin from the nationwide clinical trial, citing safety concerns.

But again, a government institution's handling of Rezulin was attended by the specter of conflict of interest. It turned out that the senior official with overall responsibility for the NIH study had, at the same time, fashioned his own financial relationship with Warner-Lambert. The NIH official, actually the United States government's top diabetes researcher, Dr. Richard C. Eastman, went on Warner-Lambert's payroll as a consultant in November 1995. During the next two years, he served on the "faculty" of the sales-promoting "Rezulin National Speakers Bureau;" Eastman, whose government salary exceeds $144,000, accepted nearly $80,000 in compensation from Warner-Lambert and one of its affiliates.

Following our disclosures in the Los Angeles Times, the inspector general of the Department of Health and Human Services opened an investigation of Eastman's conduct. Eastman quietly informed NIH officials that he would no longer accept income from drug companies or other outside employers. (No doubt to the relief of NIH Director Harold E. Varmus, the affair has remained unmentioned in the nation's other major newspapers.)

In March of this year, the Medicines Control Agency of Britain refused to allow re-introduction of Rezulin. And on June 16, the FDA announced the fourth safety-labeling change for Rezulin.

The FDA, citing "new safety information," advised doctors in June to have patients submit to monthly liver tests for a full year. The agency also declared that Rezulin was no longer appropriate for use as a front-line, stand-alone diabetes pill. Barely two years on the U.S. market--and four safety-related labeling changes. Officials acknowledge that, in the history of the FDA, no other drug has met such a fate.

The agency's handling of the drug was mocked this summer at a conference sponsored by Georgetown University Medical Center. Dr. Alastair Wood, a Vanderbilt University professor who is also drug therapy editor of the New England Journal of Medicine, analogized the FDA's label changes to managing the risk posed by a dangerous mountain cliff. "The point was, you don't put up more and more signs if people continue to fall off the cliff," Wood recalled. "You try to do something more definitive, like try to prevent them from falling off. You put up a fence" At the FDA, however, officials have persisted with incremental changes to Rezulin's labeling and have eschewed ordering its withdrawal. Dr. Wood's analogy elicited much laughter from an audience that included representatives from academia, industry, and the FDA. Yet his point was entirely serious: The label changes have been made, and the liver-failure deaths persist. (When the FDA approved Rezulin in Jan. 1997, at least nine other medicines were already on the market for treatment of Type 2 diabetes.)

Some private physicians who once spoke on Rezulin's behalf have told me that they hope the drug withers in the face of competition from two newly introduced, chemically similar compounds. Unlike Rezulin, these two drugs have demonstrated virtually no liver toxicity in clinical studies. Not surprisingly, the companies that make these drugs are conducting negative marketing campaigns at Rezulin's expense. Why, they ask doctors, prescribe a drug that kills livers when there is a safer alternative?

Rezulin may eventually recede from the pharmaceutical landscape, its demise not compelled by the federal government, but by the not-so-invisible hand of market forces. This would please at least a few officials, perhaps including one thoughtful member of the advisory committee that met in March, Dr. Glenn Braunstein, an endocrinologist from Cedars-Sinai Medical Center in Los Angeles.

Along with a majority of his colleagues on the FDA committee, Braunstein said on March 26 that he favored leaving Rezulin on the market, but he added: "I know that there are a number of other drugs in the pipeline in the same [chemical] class, and I'm hopeful that one or more of those will prove not to be hepatotoxic. And then we'll let the marketplace tell us which is the best drug." But the unaddressed question was and remains: What about the well-being of all those patients who, in the interim, begin taking or stay on Rezulin?

This much is clear: By now Rezulin has exceeded the original expectations of Warner-Lambert, whose chief executive officer told investors in May 1997 that Rezulin could become a "billion-dollar blockbuster." As of this spring, the drug had generated sales exceeding $1.4 billion. It remains to be seen to what extent recently-filed product-liability and wrongful-death lawsuits will erode the gross.

Death In The Fast Lane

As disturbing as the human toll is, equally unnerving is the fact that the saga of Rezulin is surely not an anomaly. The forces of commerce that long derided the FDA as a glacially paced dinosaur now rule. The new paradigm of prescription-drug regulation has arrived at FDA headquarters in Rockville, Md. by way of a Republican-controlled Congress, with scant dissent from Democrats. The controlling culture at today's FDA, officials say, holds rejecting most new drugs to be passe. If a drug has dangers, goes this algorithm, they usually can be "managed" via some speck-sized language buried within a label.

"Nobody wants to slow up," one veteran FDA specialist involved with the original, fast-track review of Rezulin told me. "The only thing our management is truly concerned about is getting it done in a timely fashion. It's the speed, and the `Hurry up,' and the rush. Nothing else means anything. That's what they want; that's what they're getting. I'm waiting for a disaster to happen--maybe this is it. And then they'll start giving us a little extra time to get a job done."

It was three long decades ago that an FDA physician, Dr. Francis O. Kelsey, delayed--and withheld--the agency's approval of a proposed sedative called thalidomide. Nearly two years later, when thalidomide was proven to have caused the disfiguring of thousands of newborns worldwide, Dr. Kelsey was honored by members of Congress and by President Kennedy for keeping the compound away from expectant mothers in America. "She wouldn't have had that latitude today," said the veteran FDA specialist.

DAVID WILLMAN is a reporter in the Washington Bureau of the Los Angeles Times, for which he wrote a series of articles on Rezulin in December 1998.