How a New Policy Led to Seven Deadly Drugs

For most of its history, the United States Food and Drug Administration approved new prescription medicines at a grudging pace, paying daily homage to the physician’s creed, “First, do no harm.”

Then in the early 1990s, the demand for AIDS drugs changed the political climate. Congress told the FDA to work closely with pharmaceutical firms in getting new medicines to market more swiftly. President Clinton urged FDA leaders to trust industry as “partners, not adversaries.”

The FDA achieved its new goals, but now the human cost is becoming clear.

Seven drugs approved since 1993 have been withdrawn after reports of deaths and other severe side effects. A two-year Los Angeles Times investigation has found that the FDA approved each of those drugs while disregarding danger signs or blunt warnings from its own specialists. Then, after receiving reports of significant harm to patients, the agency was slow to seek withdrawals.

According to “adverse-event” reports filed with the FDA, the seven drugs were cited as suspects in 1,002 deaths. Because the deaths are reported by doctors, hospitals and others on a voluntary basis, the true number of fatalities could be far higher, according to epidemiologists.

An adverse-event report does not prove that a drug caused a death; other factors, such as preexisting disease, could play a role. But the reports are regarded by public health officials as the most reliable early warnings of danger.

The FDA’s performance was tracked through an examination of thousands of pages of government documents, other data obtained under the Freedom of Information Act and interviews with more than 60 present and former agency officials.

The seven drugs were not needed to save lives. One was for heartburn. Another was a diet pill. A third was a painkiller. All told, six of the medicines were never proven to offer lifesaving benefits, and the seventh, an antibiotic, was ultimately judged unnecessary because other, safer antibiotics were available.

The seven are among hundreds of new drugs approved since 1993, a period during which the FDA has become known more for its speed than its caution. In 1988, only 4% of new drugs introduced into the world market were approved first by the FDA. In 1998, the FDA’s first-in-the-world approvals spiked to 66%.

The drug companies’ batting average in getting new drugs approved also climbed. By the end of the 1990s, the FDA was approving more than 80% of the industry’s applications for new products, compared with about 60% at the beginning of the decade.

And the companies have prospered: The seven unsuccessful drugs alone generated U.S. sales exceeding $5 billion before they were withdrawn.

Once the world’s unrivaled safety leader, the FDA was the last to withdraw several new drugs in the late 1990s that were banned by health authorities in Europe.

“This track record is totally unacceptable,” said Dr. Curt D. Furberg, a professor of public health sciences at Wake Forest University. “The patients are the ones paying the price. They’re the ones developing all the side effects, fatal and non-fatal. Someone has to speak up for them.”

The FDA’s faster and more lenient approach helped supply pharmacy shelves with scores of new remedies. But it has also yielded these fatal missteps, according to the documents and interviews:

* Only 10 months ago, FDA administrators dismissed one of its medical officer’s emphatic warnings and approved Lotronex, a drug for treating irritable bowel syndrome. Lotronex has been linked to five deaths, the removal of a patient’s colon and other bowel surgeries. It was pulled off the market on Nov. 28.

* The diet pill Redux, approved in April 1996 despite an advisory committee’s vote against it, was withdrawn in September 1997 after heart-valve damage was detected in patients put on the drug. The FDA later received reports identifying Redux as a suspect in 123 deaths.

* The antibiotic Raxar was approved in November 1997 in the face of evidence that it may have caused several fatal heart-rhythm disruptions in clinical studies. FDA officials chose to exclude any mention of the deaths from the drug’s label. The maker of the pill withdrew it in October 1999. Raxar was cited as a suspect in the deaths of 13 patients.

* The blood pressure medication Posicor was approved in June 1997 despite findings by FDA specialists that it might fatally disrupt heart rhythm and interact with certain other drugs, posing potentially severe risk. Posicor was withdrawn one year later; reports cited it as a suspect in 100 deaths.

* The painkiller Duract was approved in July 1997 after FDA medical officers warned repeatedly of the drug’s liver toxicity. Senior officials sided with the manufacturer in softening the label’s warning of the liver threat. The drug was withdrawn 11 months later. By late 1998, the FDA had received voluntary reports citing Duract as a suspect in 68 deaths, including 17 that involved liver failure.

* The diabetes drug Rezulin was approved in January 1997 over a medical officer’s detailed opposition and was withdrawn this March after the agency had linked 91 liver failures to the pill. Reports cite Rezulin as a suspect in 391 deaths.

* The nighttime heartburn drug Propulsid was approved in 1993 despite evidence that it caused heart-rhythm disorders. The officials who approved the drug failed to consult the agency’s own cardiac specialists about the signs of danger. The drug was taken out of pharmacies in July after scores of confirmed heart-rhythm deaths. Overall, Propulsid has been cited as a suspect in 302 deaths.

The FDA’s handling of Propulsid put children at risk.

The agency never warned doctors not to administer the drug to infants or other children even though eight youngsters given Propulsid in clinical studies had died. Pediatricians prescribed it widely for infants afflicted with gastric reflux, a common digestive disorder.

Parents and their doctors had no way of knowing that the FDA, in August 1996, had found Propulsid to be “not approvable” for children.

“We never knew that,” said Jeffrey A. Englebrick, a heavy-equipment welder in Shawnee, Kan., whose 3-month-old son, Scott, died on Oct. 28, 1997, after taking Propulsid. “To me, that means they took my kid as a guinea pig to see if it would work.”

By the time the drug was pulled, the FDA had received reports of 24 deaths of children under age 6 who were given Propulsid. By then the drug had generated U.S. sales of $2.5 billion for Johnson & Johnson Co.

Questions also surround the recent approvals of other compounds that remain on the market, including a new flu drug called Relenza. In February of 1999, an FDA advisory committee concluded that Relenza had not been proved safe and effective. The agency nevertheless approved it. Following the deaths of seven patients, the FDA in January issued a “public health advisory” to doctors.

A ‘Lost Compass’

A total of 10 drugs have been pulled from the market in just the past three years for safety reasons, including three pills that were approved before the shift that took hold in 1993. Never before has the FDA overseen the withdrawals of so many drugs in such a short time. More than 22 million Americans--about 10% of the nation’s adult population--took those drugs.

With many of the drugs, the FDA used tiny-print warnings or recommendations in package labeling as a way to justify approvals or stave off withdrawals. In other instances, the agency has withheld safety information from labels that physicians say would call into question the use of the product.

Present and former FDA specialists said the regulatory decisions of senior officials have clashed with the agency’s central obligation, under law, to “protect the public health by ensuring . . . that drugs are safe and effective.”

“They’ve lost their compass and they forget who it is that they are ultimately serving,” said Dr. Lemuel A. Moye, a University of Texas School of Public Health physician who served from 1995 to 1999 on an FDA advisory committee. “Unfortunately the public pays for this, because the public believes that the FDA is watching the door, that they are the sentry.”

The FDA’s shift is felt directly in the private practice of medicine, said Dr. William L. Isley, a Kansas City, Mo., diabetes specialist. He implored the agency to reassess Rezulin three years ago after a patient he treated suffered liver failure taking the pill.

“FDA used to serve a purpose,” Isley said. “A doctor could feel sure that a drug he was prescribing was as safe as possible. Now you wonder what kind of evaluation has been done, and what’s been swept under the rug.”

FDA officials said that they have tried conscientiously to weigh benefits versus risks in deciding whether to approve new drugs. They noted that many doctors and patients complain when a drug is withdrawn.

“All drugs have risks; most of them have serious risks,” said Dr. Janet Woodcock, director of the FDA’s drug review center. She added that some of the withdrawn drugs were “very valuable, even if not lifesaving, and their removal from the market represents a loss, even if a necessary one.”

Once a drug is proven effective and safe, Woodcock said, the FDA depends on doctors “to take into account the risks, to read the label. . . . We have to rely on the practitioner community to be the learned intermediary. That’s why drugs are prescription drugs.”

In a May 12, 1999, article co-authored with FDA colleagues and published by the Journal of the American Medical Assn., Woodcock said, “The FDA and the community are willing to take greater safety risks due to the serious nature of the [illnesses] being treated.”

Compared to the volume of new drugs approved, they wrote, the number of recent withdrawals “is particularly reassuring.”

However, agency specialists point out that both approvals and withdrawals are controlled by Woodcock and her administrators. When they consider a withdrawal, they face the unpleasant prospect of repudiating their original decision to approve.

Woodcock, 52, received her medical degree at Northwestern University and is a board-certified internist. She alluded in a recent interview to the difficulty she feels in rejecting a proposed drug that might cost a company $150 million or more to develop. She also acknowledged the commercial pressures in a March 1997 article.

“Consumer protection advocates want to have drugs worked up well and thoroughly evaluated for safety and efficacy before getting on the market,” Woodcock wrote in the Food and Drug Law Journal. “On the other hand, there are economic pressures to get drugs on the market as soon as possible, and these are highly valid.”

But this summer--following the eighth and ninth drug withdrawals--Woodcock said the FDA cannot rely on labeling precautions, alone, to resolve safety concerns.

“As medical practice has changed . . . it’s just much more difficult for [doctors] to manage” the expanded drug supply, Woodcock said in an interview. “They rely upon us much more to make sure the drugs are safe.”

Another FDA administrator, Dr. Florence Houn, voiced similar concern in remarks six months ago to industry officials: “I think the lessons learned from the drug withdrawals make us leery.”

Yet the imperative to move swiftly, cooperatively, remains.

“We are now making decisions more quickly and more predictably while maintaining the same high standards for product safety and efficacy,” FDA Commissioner Jane E. Henney said in a National Press Club speech on Dec. 12.

Motivated by AIDS

The impetus for change at the FDA emerged in 1988, when AIDS activists paralyzed operations for a day at the agency’s 18-story headquarters in Rockville, Md. They demanded immediate approval of experimental drugs that offered at least a ray of hope to those otherwise facing death.

The FDA often was taking more than two years to review new drug applications. The pharmaceutical industry saw a chance to loosen the regulatory brakes and expedite an array of new products to market. The companies and their Capitol Hill lobbyists pressed for advantage: If unshackled, they said, the companies could invent and develop more remedies faster.

The political pressure mounted, and the FDA began to bow. By 1991, agency officials told Congress they were making significant progress in speeding the approval process.

The emboldened companies pushed for more. They proposed that drugs intended for either life-threatening or “serious” disorders receive a quicker review.

“The pharmaceutical companies came back and lobbied the agency and the Hill for that word, ‘serious,’ ” recalled Jeffrey A. Nesbit, who in 1991 was chief of staff to FDA Commissioner David A. Kessler. “Their argument was, ‘Well, OK, there’s AIDS and cancer. But there are drugs [being developed] for Alzheimer’s. And that’s a serious illness.’ They started naming other diseases. They began to push that envelope.”

The wielding of this single, flexible adjective--”serious”--swung wide the regulatory door knocked ajar by the AIDS crisis.

New Order Takes Hold

In 1992, Kessler issued regulations giving the FDA discretion to “accelerate approval of certain new drugs” for serious or life-threatening conditions. That same year a Democrat-controlled Congress approved and President Bush signed the Prescription Drug User Fee Act. It established goals that call for the FDA to review drugs within six months or a year; the pharmaceutical companies pay a user fee to the FDA, now $309,647, with the filing of each new drug application.

The newly elected Clinton administration climbed aboard with its “reinventing government” project. Headed by Vice President Al Gore, the project called for the FDA, by January 2000, to reduce “by an average of one year the time required to bring important new drugs to the American public.” As Clinton put it in a speech on March 16, 1995, the objective was to “get rid of yesterday’s government.”

For the FDA’s medical reviewers--the physicians, pharmacologists, chemists and biostatisticians who scrutinize the safety and effectiveness of emerging drugs--a new order had taken hold.

The reviewers work out of public view in secure office buildings clustered along Maryland’s Route 355. At the jet-black headquarters building, the decor is institutional, the corridors and third-floor cafeteria without windows. The reviewers examine truckloads of scientific documents. They are well-educated; some are highly motivated to do their best for a nation of patients who unknowingly count on their expertise.

One of these reviewers was Michael Elashoff, a biostatistician who arrived at the FDA in 1995 after earning degrees from UC Berkeley and the Harvard School of Public Health.

“From the first drug I reviewed, I really got the sense that I was doing something worthwhile. I saw what a difference a single reviewer can make,” said Elashoff, the son and grandson of statisticians.

Last year he was assigned to review Relenza, the new flu drug developed by Glaxo Wellcome. He recommended against approval.

“The drug has no proven efficacy for the treatment of influenza in the U.S. population, no proven effect on reducing person-to-person transmissibility, and no proven impact on preventing influenza,” Elashoff wrote, adding that many patients would be exposed to risks “while deriving no benefit.”

An agency advisory committee agreed and on Feb. 24 voted 13 to 4 against approving Relenza.

After the vote, senior FDA officials upbraided Elashoff. They stripped him of his review of another flu drug. They told him he would no longer make presentations to the advisory committee. And they approved Relenza as a safe and effective flu drug.

Lost Faith in the System

Elashoff and other FDA reviewers discern a powerful message.

“People are aware that turning something down is going to cause problems with [officials] higher up in FDA, maybe more problems than it’s worth,” he said. “Before I came to the FDA I guess I always assumed things were done properly. I’ve lost a lot of faith in taking a prescription medicine.”

Elashoff left the FDA four months ago.

“Either you play games or you’re going to be put off limits . . . a pariah,” said Dr. John L. Gueriguian, a 19-year FDA medical officer who opposed the approval of Rezulin, the ill-fated diabetes drug. “The people in charge don’t say, ‘Should we approve this drug?’ They say, ‘Hey, how can we get this drug approved?’ ”

Said Dr. Rudolph M. Widmark, who retired in 1997 after 11 years as a medical officer: “If you raise concern about a drug, it triggers a whole internal process that is difficult and painful. You have to defend why you are holding up the drug to your bosses. . . . You cannot imagine how much pressure is put on the reviewers.”

The pressure is such that when a union representative negotiated a new employment contract for the reviewers last year, one of his top priorities was to defend what he called the “scientific integrity” of their work.

“People feel swamped. People are pressured to go along with what the agency wants,” said Dr. Robert S.K. Young, an FDA medical officer who in 1998 formed a union chapter to represent the reviewers. “You’re paying for these highly educated, trained people, and they’re not being allowed to do their job.”

Each new drug application is accompanied by voluminous medical data, enough at times to fill 1,000 or more phone books. The reviewers must master this material in less than six months or a year, while juggling other tasks.

“The devil is in the details, and detail is something we no longer have the time to go into,” said Gurston D. Turner, a veteran pharmacologist with the FDA’s scientific investigations division who retired this year. “If you know you must have your report done by a certain date, you get something done. That’s what they [top FDA officials] count, that’s all they count. And that is really, to me, a worrisome thing.”

The FDA did spur reviewers to move at record speed.

In 1994, the FDA’s goal was to finish 55% of its new drug reviews on time; the agency achieved 95%. In 1995, the goal was 70%; the FDA achieved 98%. In 1996, the goal was 80%; the FDA achieved 100%. In both 1997 and 1998, the goal was 90% and the FDA achieved 100%.

From 1993-99 the agency approved 232 drugs regarded as “new molecular entities,” compared with 163 during the previous seven years, a 42% increase.

The time-limit goals quickly were treated as deadlines within the FDA--imposing relentless pressure on reviewers and their bosses to quickly conclude their work and approve the drugs.

“The goals were to be taken seriously. I don’t think anybody expected the agency to make them all,” said William B. Schultz, a deputy FDA commissioner from 1995 to 1999.

Schultz, who helped craft the 1992 user-fee act as a congressional staff lawyer, added: “You can meet the goal by either approving the drug or denying the approval. But there are some who argue that what Congress really wanted was not just decisions, but approvals. That is what really gets dangerous.”

Indeed, the FDA drug center’s 1999 annual report referred to the review goals as “the law’s deadlines.” And, Dr. Woodcock, the center director, elaborated in a subsequent agency newsletter:

“In exchange [for the user fees], FDA makes a commitment to meet certain goals for review times. [The agency] has exceeded almost all of the goals, and it expects to continue to exceed them. Basically, the number of new approved drugs has doubled, and the review times have been cut in half.”

The user fees have enabled the FDA to hire more medical reviewers. Last year, 236 medical officers examined new drugs compared with 162 officers on duty in 1992, the year before the user fees took effect.

Even so, Woodcock acknowledged in an FDA publication this fall that the workloads and tight performance goals “create a sweatshop environment that’s causing high staffing turnover.”

An FDA progress report in 1998, describing the work of agency chemists, said that “too many reviews are coming ‘down to the wire’ against the goal date. . . . This suggests a system in stress.”

Said Nesbit, the former aide to Commissioner Kessler: “The clock is always running, whereas before the clock was never running. And that changes people’s behavior.”

Dozens of officials interviewed by The Times made similar observations.

“The pressure to meet deadlines is enormous,” said Dr. Solomon Sobel, 65, director of the FDA’s metabolic and endocrine drugs division throughout the 1990s. And the pressure is not merely to complete the reviews, he said. “The basic message is to approve.”

Over the last seven years, “there has been a huge shift,” said Kathleen Holcombe, a former FDA legislative affairs staffer and congressional aide who now is a drug industry consultant. “FDA, historically, had an approach of, ‘Regulate, be tough, enforce the law [and] don’t let one thing go wrong,’ ” Holcombe said, adding that now, “the FDA sees itself much more in a cooperative role.”

The perception of coziness with drug makers is perpetuated by potential conflicts of interest within the FDA’s 18 advisory committees, the influential panels that recommend which drugs deserve approval or should remain on the market. The FDA allows some appointees to double as consultants or researchers for the same companies whose products they are evaluating on the public’s behalf. Such was the case during committee appraisals of several of the recently withdrawn drugs, including Lotronex and Posicor, The Times found.

Few doubt the $100-billion pharmaceutical industry’s clout. Over the last decade, the drug companies have steered $44 million in contributions to the major political parties and to candidates for the White House and both houses of Congress.

The FDA reviewers said they and their bosses fear that unless the new drugs are approved, companies will erupt and Congress will retaliate by refusing to renew the user fees. This would cripple FDA operations--and jeopardize jobs.

The companies’ money now covers about 50% of the FDA’s costs for reviewing proposed drugs--and agency officials say that persuading Congress to renew the user fees into 2007 is now a top priority.

Yet even if the user fees remain, the FDA is prohibited from spending the revenue for anything other than reviewing new drugs. So while the budget for pre-approval reviews has soared, the agency has gotten no similar increase of resources to evaluate the safety of the drugs after they are prescribed.

“It’s shocking,” said Dr. Brian L. Strom, chairman of epidemiology at the University of Pennsylvania. “How can you say, ‘Release drugs to the market sooner,’ and not know if they’re killing people? . . . It really is a dramatic statement of public priorities.”

More than 250,000 side effects linked to prescription drugs, including injuries and deaths, are reported each year. And those “adverse-event” reports by doctors and others are only filed voluntarily. Experts, including Strom, believe the reports represent as few as 1% to 10% of all such events.

“There’s no incentive at all for a physician to report [an adverse drug reaction],” said Strom, who has documented the phenomenon. “The underreporting is vast.”

Even when deaths are reported, records and interviews show that companies consistently dispute that their product has caused a given death by pointing to other factors, including preexisting disease or use of another medicine.

To be sure, a chain of events affects the safe use of a prescription drug: The companies’ conduct of clinical studies; the FDA’s regulatory actions; the doctor’s decision to prescribe; the pharmacist’s filling of a handwritten prescription; the patient’s ability to take the drug as directed. A lapse at any link could prove fatal.

And once a pill is approved by the FDA, the manufacturer often spends heavily on promotion to seize the largest possible market share. This can exacerbate the risk to public health, according to experts.

“Aggressive promotion increases exposure--and doesn’t give you the time to find the problem before patients get hurt,” said Dr. Raymond L. Woosley, pharmacology department chairman at Georgetown University and a former FDA advisory committee member.

When serious side effects emerge, the FDA officials have championed using package labeling as a way to, in their words, “manage” risks. Yet the agency typically has no way to know if the labeling precautions--dense, lengthy and in tiny print--are read or followed by doctors and their patients.

The FDA often addresses unresolved safety questions by asking companies to conduct studies after the product is approved. But the research frequently is not performed--prompting the inspector general of the Department of Health and Human Services to say in 1996 that “FDA can move to withdraw drugs from the market if the post-marketing studies are not completed with due diligence.”

Since that report was issued, the FDA has not withdrawn any drug due to a company’s failure to complete a post-approval safety study. Officials conceded this week that they still do not know how often the studies are performed.

One consequence is that greater risk is shifted to doctors and patients.

For example, Woodcock and her senior aides allowed Rezulin to remain on the U.S. market nearly 2 1/2 years after it was withdrawn in Britain in December 1997. The FDA recommended frequent laboratory testing of patients using the drug but had no scientific assurance that the tests would prevent Rezulin-induced liver failure.

“They kept increasing the number of liver-function tests you should have,” noted Dr. Alastair J.J. Wood, a former FDA advisory committee member who is a professor of medicine at Vanderbilt University. “That was clearly designed to protect the FDA, to protect the manufacturer, and to dump the responsibility on the patient and the physician. If the patient developed liver disease and he hadn’t had his [tests] done, somebody was to blame and it wasn’t the manufacturer and it wasn’t the FDA.”

Industry Assurances

Leading industry officials say Americans have nothing to fear from the wave of drug approvals.

“Do unsafe drugs enter and remain in the marketplace? Absolutely not,” said Dr. Bert A. Spilker, senior vice president for scientific and regulatory affairs for the Pharmaceutical Research and Manufacturers of America, in remarks last year to industry and FDA scientists.

But during interviews over the last two years, current and former FDA specialists cited repeated instances when drugs were approved with less than compelling evidence of safety or effectiveness. They also said that important information has been excluded from the labels on some medications.

Elashoff, for instance, was surprised at the labeling for a drug called Prograf, approved in 1997 to prevent rejection of transplanted kidneys. The drug first had been approved in 1994 for use among liver-transplant patients.

The new label notes that Prograf was proved effective in a study of 412 U.S. kidney transplant patients. But no mention is made of the company’s 448-patient European study, in

which 7% of the patients who took Prograf died--double the 3.5% death rate among those who received a different anti-rejection drug, documents show.

An auditor from the FDA’s scientific investigations unit, Antoine El-Hage, examined the European study results and concluded the “data are reliable.” Elashoff agreed in his review.

Yet the only way for doctors or patients to find that data is to search the medical literature or seek the FDA’s review documents.

Excluding the European study from the Prograf label, Elashoff said, “was just a total whitewash. . . . I think any rational person would reconsider taking this drug if they knew what happened in Europe.”

A spokesman for the manufacturer of Prograf said the company had no objection to including the European study results in the labeling. William E. Fitzsimmons, a vice president of drug development for Fujisawa Healthcare Inc., said the decision to exclude the results was entirely the FDA’s.

“We submitted that data,” he said. “It came down to what the FDA was comfortable putting in the label. . . . We certainly have no interest in trying to hide that information. We presented it at major meetings on transplantation. . . . We’re comfortable with that information being out in the public domain.”

But if the FDA had included the European results in the label, it would have impugned the agency’s basis for approving the new, expanded use for Prograf, according to Elashoff and others.

Asked why the agency excluded the information, Woodcock said the European results were “unreliable . . . and could be potentially misleading to doctors and patients in the U.S. if these were included in the label.”

CASE STUDIES

Drug After Drug, Warnings Ignored

Danger signs were present as each of the prescription medicines on the following pages were considered for the FDA’s approval. Even so, top administrators moved ahead, often leaving doctors to assume the risks listed in fine-print labels. Seven were eventually withdrawn, but only after reports of deaths.

PROPULSID

A Heartburn Drug, Now Linked to Children’s Deaths

Once evidence of harm emerged, FDA took years to withdraw approval.

*

“Those of us here at the FDA who are aware of your loss wish to again extend our deepest sympathy and sincere condolences to you and your family.”

--FDA Administrator Florence Houn, writing July 27, 2000, to the mother of 9-month-old Gage E. Stevens, who died on Thanksgiving 1999.

*

In mid-1993, FDA officials prepared to approve Propulsid, a drug that eased nighttime heartburn. But a sign of danger loomed.

FDA medical officer Andre Dubois noted that 48 of 1,993, or 2.4%, of the patients who took Propulsid in U.S. studies experienced “heart rate and rhythm disorders.” In addition, eight children age 6 or younger who were given Propulsid had died.

Dubois found that the drug’s chemical makeup could disturb cardiac function. But he agreed with drug maker Janssen Pharmaceutica, a Johnson & Johnson Co. subsidiary, that the deaths in the studies were attributable to other causes.

He recommended approval along with disclosure in the label of potential cardiac effects. “The risk seems very low,” he said.

Dubois, however, worked in a division that focuses on drugs for the gastrointestinal tract.

No one at the FDA consulted with the agency’s division of cardiac specialists before approving Propulsid on July 29, 1993, according to physicians familiar with the matter. By not tapping their expertise, FDA officials failed to notice what should have been another warning flag: Electrocardiograms showed that Propulsid prolonged patients’ “QT interval,” the time during which the heart’s main pumping chambers contract and then relax.

If the QT interval--typically about 4/10 of a second--is extended even slightly, it can trigger a disruption or cessation of the heartbeat. Called an arrhythmia, it can result in sudden death.

FDA officials outside the gastrointestinal division had already warned publicly--on June 11, 1990--that two allergy drugs, Seldane and Hismanal, prolonged the QT interval and therefore posed lethal risk. Both drugs were later withdrawn.

Indeed, the danger had been stressed for several years by Dr. Raymond J. Lipicky, director of the agency’s cardiology division. Lipicky, writing in the August 1993 issue of the American Journal of Cardiology, said if a drug that prolonged the QT interval had a benefit that was “less than lifesaving . . . any risk of death would likely be considered unacceptable.”

In approving Propulsid, the FDA agreed to labeling that advised doctors of “rare cases” of increased heartbeats. The labeling said Propulsid’s role in the events “was not clear.”

In response to written questions, Dr. Janet Woodcock, director of the FDA’s drug review center, said the danger associated with non-cardiac drugs that prolonged the QT interval “was not well appreciated” at the time Propulsid was approved. Consequently, she said, this “was not identified as a concern” by the gastrointestinal division.

By early 1995, Propulsid’s danger to the heart was certainly identified as a concern within the gastrointestinal division, agency records show.

On Jan. 25, 1995, a senior FDA medical officer, Dr. Stephen B. Fredd, told Janssen executives that recent adverse-reaction reports showed their drug was prolonging the QT interval, perhaps resulting in deaths.

According to the meeting summary, “It was the firm’s position that the cases cited by Dr. Fredd were not ‘clean’ cases, thus making it difficult to attribute the effect to [Propulsid].” Fredd responded that “unequivocal evidence” of Propulsid’s culpability was unlikely to be captured outside of a controlled clinical study.

But within a month, the FDA and the company agreed to the first of five safety-labeling changes that would help keep the drug on the market over the next five years.

Meanwhile, a significant market for Propulsid emerged in the treatment of children.

Propulsid was never proved effective or safe for infants, yet it became the drug of choice for many pediatricians in treating gastric reflux, a common disorder that is usually outgrown by age 1. Reflux can impede infants’ digestion and, due to their crying, disrupt their parents’ sleep. As with almost all drugs, doctors could lawfully prescribe Propulsid for any use, or “indication,” they chose.

On Aug. 15, 1996, the FDA informed the Johnson & Johnson subsidiary that Propulsid was “not approvable” for children, interviews and documents obtained by The Times show. The rejection, in keeping with FDA practice, was not made public.

In private correspondence a year later, on Aug. 19, 1997, Dr. Lilia Talarico, FDA’s gastrointestinal drugs division director, cited “at least” three recently reported deaths among child patients. She told a company official the agency was considering altering the label of Propulsid to “contraindicate,” or to warn against its use in infants.

Asked why the FDA did not immediately inform doctors and patients of the deaths, Woodcock told The Times: “Labeling changes [advising of infant deaths] were requested by FDA in August of 1997 but were not agreed to by the company until June of 1998.”

That revised label did acknowledge “several pediatric deaths” but left physicians guessing whether Propulsid was the culprit, saying, “Causality has not been established.”

Parents of children who died after taking Propulsid said in interviews that they had no inkling of danger.

“If I had known that this drug caused cardiac arrhythmias, I would never have given it to him,” said Tina Englebrick, the mother of 3-month-old Scott, who died in October 1997. The Kansas health department identified Scott’s cause of death as sudden infant death syndrome.

Had the parents of Gage Stevens, the deceased 9-month-old, “been informed of a risk of sudden death, they would not have administered the medication to their son,” according to a lawsuit they filed in a Pennsylvania court on Sept. 10 against the manufacturer and the doctor and hospital who treated him. Gage, who had reflux, was given Propulsid within a pediatric study that was approved by the FDA and performed by researchers at the University of Pittsburgh.

He died at 6:30 a.m. on Thanksgiving 1999. The county coroner concluded that the death was “directly related” to Propulsid and one other drug administered to the child. The coroner said Gage “most probably” had died after suffering a cardiac arrhythmia.

Said Dr. Robert R. Fenichel, who retired this year as deputy director of the FDA’s cardiac drugs division: “It was scandalous that all of these kids were being treated with [Propulsid]” in the absence of proven safety and effectiveness.

On March 23, 2000, the FDA announced that Propulsid would be taken off the market as of July as a normally prescribed drug because of scores of confirmed heart-rhythm deaths. Overall, Propulsid has been cited as a suspect in 302 deaths.

FDA administrators now concede that the agency failed to contain Propulsid’s fatal risk.

“We’ve had a seven-year history with this drug where it’s a very rich opportunity for us to learn,” the FDA’s Dr. Florence Houn told drug industry officials in a Webcast on June 22. “One of the things we have learned is the approved indication for a drug really needs to [justify] the serious and life-threatening” side effects.

In comments the same month to an FDA advisory committee, Houn added, “The labeling probably was not effective.”

Why did the agency wait so long to seek the withdrawal of this drug for nighttime heartburn in adults?

“We simply tried a variety of measures,” Woodcock said in an interview. “We have to sort of walk that line: Where do we inform and where do we intervene by removing a drug from the market? That is a very draconian step. . . . And so, we do try to avoid that.”

Six specialists involved with the FDA’s decisions concerning Propulsid said the volume of prescriptions for reflux in infants helped keep the drug on the market.

One specialist who sought earlier withdrawal of Propulsid said, “If it were just the nocturnal heartburn indication we were considering . . . it’s a pretty easy decision” to pull it off the market. Many alternative therapies existed, including over-the-counter products like Tums and Maalox and Zantac.

Woodcock, who was appointed to her position 10 months after Propulsid was approved, said the FDA did not formally weigh the off-label use while deciding to keep the drug on the market. She acknowledged that it was prescribed widely for children but said she relied on pediatricians to make prudent decisions.

“They’re aware of the QT-prolongation issue,” Woodcock said. “This isn’t as if it’s some mystery. . . . They evaluated this and came to their own conclusions about the risks.”

A spokesman for the Johnson & Johnson subsidiary, Greg Panico, said the company did not promote Propulsid for use by children. However, he acknowledged that it did make two “educational grants” to the North American Society for Pediatric Gastroenterology and Nutrition. The society’s literature advised doctors that Propulsid could be used safely and effectively in children.

Panico declined to say how much money the company provided; according to the society’s Web site, the group has been “generously supported” by the Johnson & Johnson subsidiary. The society held a symposium on the use of Propulsid at an October 1998 conference in Orlando, Fla. A spokeswoman for the pediatric society said the company’s grants came with “no strings attached.”

The removal that Woodcock and her aides negotiated this year allows the continued sale of Propulsid under a “limited access plan.” This authorizes doctors to administer the drug to patients of all ages who have not benefited from other treatments and who would be closely monitored.

In September, the British Medicines Control Agency rejected continued sales of Propulsid there under such conditions, saying, “Restricted-access schemes . . . are not adequate to protect public health.” The British have warned since 1998 against any use of Propulsid in infants and cautioned against prescribing it to children up to age 12.

For her part, Woodcock said she remains “concerned” about the drug’s use among children. A recent agency review found that, while “no clear evidence” implicated Propulsid as the primary cause of eight children’s deaths before the July 1993 approval, neither was there enough data to exclude a “role” for the drug in several of those cases.

As for adult patients who died, Woodcock said, “It’s a terrible thing to happen to somebody who is just taking the drug for heartburn.”

Panico said there remains a place for Propulsid.

“When we made the decision to limit access to the drug, we had pleas from families of children who are taking this drug to make sure that these kids can have continued access to it,” he said. “So, it’s a balancing act.”

RAXAR: Warning on Label Omits Deaths

Heart problems were mentioned in fine print, but not key dosage data.

When the antibiotic Raxar was approved on Nov. 6, 1997, FDA officials knew that it too might cause irregular rhythm and stop a patient’s heart.

An agency medical officer, Dr. Andrea N. Meyerhoff, suspected that two of four patients who died after taking Raxar in clinical trials possibly suffered heart-rhythm disturbances caused by the drug.

Meyerhoff noted in her review that the drug manufacturer, Glaxo Wellcome Inc., said Raxar played no role in the deaths. But Meyerhoff wrote that the two cases posed an open question. Each patient who died had taken 600-milligram doses.

Regarding one of those patients, a 68-year-old man who died a week after completing the clinical trial, Meyerhoff wrote: “This patient may have been at higher risk for [fatal] arrhythmia due to QT interval prolongation from grepafloxacin,” the chemical name of Raxar. The second patient died five days after withdrawing from the clinical trial.

She added in her review, dated November 1997: “Again it is not clear that this event is unrelated to [Raxar]. Sudden death in a patient with no prior cardiac history is suggestive of an arrhythmia. . . . The label will need to have an adequate warning regarding the possibility of QT prolongation.” Overall, she found a “significantly higher” rate of adverse events among patients who had taken 600 milligrams compared with lower doses.

With Meyerhoff’s assent, the FDA approved Raxar for treating bronchitis, pneumonia, urinary tract infections and gonorrhea. The drug’s label stated that “prolongation of the QT interval has been observed in healthy volunteers receiving Raxar.”

But the label did not disclose the fatalities described in Meyerhoff’s review. It said that “there were no deaths or permanent disabilities” among those who took Raxar in 400-milligram doses. The statement was true, if incomplete: All four of the study patients who died took Raxar in 600-milligram doses. And Glaxo marketed the drug at doses of 200 milligrams, 400 milligrams and 600 milligrams. A total of 925 patients took the 600 milligrams dose in the clinical studies.

On Oct. 27, 1999, Glaxo pulled Raxar off the market.

In a subsequent letter to doctors, Glaxo said that because of Raxar’s effect on “QT interval prolongation” the drug was unacceptably risky. In a separate statement, the company said it “is no longer convinced that the benefits of Raxar outweigh the potential risk to patients, given the availability of alternative antibiotics.”

Records filed with the FDA show that Raxar was cited as a suspect in the voluntarily reported deaths of 13 patients. They ranged in age from 42 to 86; most of them were under 70.

“[Raxar] goes on the market, kills people and has to come off,” said Dr. Raymond L. Woosley, the pharmacology department chairman at Georgetown University who served on an FDA advisory committee in the 1980s. “It had been proven, over and over, that this QT prolongation predicts terrible events.”

By the time of the withdrawal, Raxar had generated $23.5 million in U.S. sales. Securities analysts had predicted it could be a $1-billion drug.

With so many other antibiotics on the market, why did the FDA expose patients to the risk of Raxar?

In a written response to questions, Woodcock indicated that the FDA sought to address the drug’s cardiac risk through precautionary language in its labeling.

Asked why that labeling did not acknowledge the deaths of patients who took doses of 600 milligrams, Woodcock wrote that none of the fatalities “was shown to be attributable to Raxar.”

In an interview over the summer, Woodcock said the FDA’s patience was gone for new drugs that prolong the QT interval. “We’re encouraging people, if there’s QT prolongation, don’t develop it,” she said.

This would mark a turnabout.

Just last December--less than two months after the withdrawal of Raxar--the FDA approved a new antibiotic, called Avelox, despite the drug’s well-documented propensity in clinical studies to prolong the QT interval.

Avelox was approved for treating sinus infections, bronchitis and pneumonia.

On the 267th line of the Avelox label, doctors are warned in bold type that it “has been shown to prolong the QT interval.”

So far, Avelox, made by Bayer Corp., has been prescribed for more than 300,000 patients in the U.S. The drug has been cited as a suspect in 18 deaths here and abroad. A Bayer spokesman, Robert Kloppenburg, said that the company does not believe any of the fatalities were “attributable” to Avelox and that most of the patients had serious preexisting conditions.

Avelox, he said, holds an advantage over many antibiotics because it need only be taken once daily for five days to be effective against bronchial infections. Securities analysts predicted in February that Avelox would generate sales topping $1 billion within three years.

Woodcock said the FDA approved Avelox because “the extent of QT prolongation . . . was too small to pose a significant risk in the face of the benefits.” She noted that an agency advisory committee recommended approval and said that “a conservative approach was taken in the label.”

REDUX: Unheeded Warnings on Lethal Diet Pill

Heart damage causes billions of dollars in potential legal liability.

Before coming to the FDA as a medical officer in 1989, Dr. Leo Lutwak had specialized in the fields of obesity and osteoporosis as a Cornell University professor, as a drug company consultant and as a practicing physician. He said he hired on at the FDA because he relished the scientific challenge of new drugs and the call of public service.

In 1995, Lutwak was the lead FDA medical officer reviewing the diet drug Redux, which in one pill approximated half of the now-infamous slimming cocktail known as fen-phen.

Both Lutwak and his boss, Dr. Solomon Sobel, told The Times that they resisted the approval of Redux.

“I, as the primary reviewer, felt that the drug had low effectiveness and very high risk for neurotoxicity and pulmonary hypertension,” a disorder that damages the respiratory system, Lutwak said.

“I was insisting on a black box,” he added, referring to the bold border at the top of a prescription label that alerts doctors and patients to severe life-threatening risk. “But the management accepted the company’s arguments against the black box. And I don’t know why.”

Sobel, director of the FDA’s endocrine and metabolic drugs division throughout the 1990s and who remains at the agency, was concerned that Redux did not work. He said he refused to sign the agency’s formal letter of approval.

“Well let me tell you,” Sobel said. “I was supposed to sign off on that letter. . . . I told [an FDA administrator, Dr. James] Bilstad that I would not sign on it. If he wanted to approve it, he should sign on it. And the record shows, he’s the one who signed on it.”

How Redux came to be approved in April 1996 remains a curiosity.

After an FDA advisory committee voted, 5 to 3, that evidence of Redux’s safety was “not sufficient to warrant approval,” Bilstad took the unusual step of scheduling a second meeting, just two months later. At that meeting, in November 1995, the committee voted, 6 to 5, to recommend approval.

Lutwak said he was “shocked” by the scheduling of the second meeting.

Much was riding on Redux. Analysts at one securities firm, Rodman & Renshaw, estimated the drug would gross $1.8 billion within four years.

But Redux was withdrawn on Sept. 15, 1997, after heart valve damage was detected in patients put on the drug. Civil lawsuits that are now pending also allege that Redux caused the potentially fatal respiratory disorder that had worried Lutwak.

American Home Products Corp., which marketed Redux and Pondimin, a diet pill that was used widely in formulations of fen-phen, agreed last fall to pay or set aside $4.75 billion to settle lawsuits related to the drugs’ potential to cause heart valve damage. The company more recently has set aside up to an additional $4.75 billion to pay other patients who have suffered from the respiratory condition or heart valve damage.

“We are aggressively settling as many cases as we can,” said Douglas Petkus, a spokesman for Wyeth-Ayerst Laboratories Inc., a subsidiary of American Home.

In its one year on the market, Redux generated sales of $255.3 million. The FDA received reports before and after the withdrawal that cited Redux as a suspect in 123 deaths.

Bilstad, who left the FDA in January, declined to be interviewed this fall when reached at his home.

In a written statement, Woodcock acknowledged that “the possibility of including a black box warning” on Redux’s label was discussed with Wyeth-Ayerst. But, she said, FDA officials decided “it was not warranted.” She said that the drug’s potential respiratory risk was noted within the labeling in bold type. Before Redux went on the market, Woodcock said, “there was no hint” that it would cause heart valve damage.

Lutwak, now 72, said he regrets the approval of Redux--and the agency’s failure to insist on a black box warning.

“It might have saved lives,” he said.

DURACT: Painkiller Posed Risk of Damage to Liver

Drugmaker’s advocacy won fine print instead of prominent warning.

The FDA medical officers who reviewed a proposed painkiller called Duract saw the problem from the outset: Too many patients who took the pill in clinical trials suffered liver injury.

“It seems imprudent to open the doors to extensive use when there have been early warning signs,” an FDA medical officer, Dr. John E. Hyde, wrote on July 31, 1996. He said the specialists reviewing Duract “were concerned about the frequency and severity” of the injuries reflected in patients’ blood tests.

Hyde and a colleague, Dr. Rudolph M. Widmark, concluded in another report: “The [liver] toxicity is a significant concern with this drug.”

Believing that the risk increased the longer a patient remained on Duract, they sought to rid the label of any reference to long-term use. They also proposed a prominent black box warning regarding Duract’s liver toxicity.

This was not what the manufacturer, Wyeth-Ayerst Laboratories, had in mind.

“They were unhappy with my review,” Widmark said in an interview.

In a market already stocked with more than 20 prescription and over-the-counter painkillers, a black box warning could turn off doctors and cripple sales.

Wyeth-Ayerst took its case to Widmark’s superiors. Widmark responded, in a memo dated Nov. 14, 1996, to the FDA drug center’s No. 2 administrator, Dr. Murray M. “Mac” Lumpkin:

“The company would like a label that actually puts the onus on the prescribing physician because if severe and maybe fatal liver toxicity [occurs], the physician will be sued and will be found liable if he/she did not ‘monitor’ for liver damage. Wyeth-Ayerst will be in the clear, because ‘it is in the label.’ ”

Widmark added, “I hope that this short memo will help you to make the right decision in this dispute.”

When the company rolled out Duract following the FDA’s approval on July 15, 1997, there was no black box on the label. Securities analysts predicted that in four years Duract could yield annual sales topping $500 million.

Beginning on the 135th line, the label’s fine print informed doctors that Duract was recommended for “generally less than 10 days.” The label also advised that, “if a physician chooses to administer Duract for a longer duration,” patients’ liver functions should be checked after a month.

Seven months after Duract’s market launch, the FDA and Wyeth-Ayerst responded to reports of severe liver damage: A black box was added.

The revised labeling also flatly warned doctors for the first time “not” to prescribe the drug for longer than 10 days.

“Patients using Duract for more than 10 days have developed jaundice, fulminant hepatitis and liver failure requiring transplants,” the FDA said, announcing the label change.

By the time Wyeth-Ayerst announced Duract’s withdrawal on June 22, 1998, the FDA had received 13 voluntarily filed reports of liver failure. The agency said that “almost all” of the cases occurred among patients who took the drug longer than 10 days.

Widmark, an Austrian immigrant, said he believes that lives would have been saved if FDA administrators had stood behind his original recommendation for a black box warning.

“I personally think yes,” Widmark said. “They were more impressed with the company’s consultants than they had confidence in their own reviewers. . . .

“Something is wrong and something should be done to avoid this in the future.”

Now 75, Widmark retired in December 1997 after spending 11 years with the FDA. He still works as a consultant to the pharmaceutical industry.

The spokesman for Wyeth-Ayerst, Petkus, said the company’s consultants “made a case that there was no need for a black box,” believing the recommendation to use Duract generally less than 10 days was sufficient. The FDA’s management, he said, agreed.

In their May 1999 medical journal article, Woodcock and Lumpkin said the problems that emerged with Duract were “unexpected,” adding: “Given the availability of other analgesics with a wider margin of safety than [Duract], the FDA believed that the risk from this product outweighed its benefits.”

In a written response to questions, Woodcock said, that if used short term, “it was felt that Duract would not cause liver damage more often than” certain other painkillers. She said the findings of potential danger, identified in advance by the agency’s two medical officers, involved tests that “do not always signal clinically important [liver] toxicity.”

By late 1998, the FDA had received voluntary reports citing Duract as a suspect in 68 deaths, including 17 that involved liver failure. During its one year on the market, Duract generated sales totaling $89.7 million for Wyeth-Ayerst.

POSICOR: 143 Sudden Deaths Did Not Stop Approval

With study results kept secret, nation got another blood-pressure drug.

Senior FDA officials with the power to approve new drugs were warned in advance about the dangers of Posicor, a pill for high blood pressure and symptomatic chest pain.

The clinical studies of Posicor “cast a shadow of potential risk for serious arrhythmias,” FDA medical team leader Dr. Shaw T. Chen wrote on Dec. 18, 1996. The data in hand also showed Posicor would interact with certain other drugs, posing potentially severe risk.

A 70-year-old man suffered “sudden death” in one study of Posicor’s effect on chest pain. The senior FDA officials also were told of sudden deaths in 142 other patients who took either Posicor or a placebo in an ongoing study focused on congestive heart failure. However, details from the 2,400-patient study remained sealed because the manufacturer opposed breaking the experiment’s confidentiality until it was finished.

This left the FDA officials a choice: Wait a year or more, or approve Posicor without knowing the details.

“I sure don’t feel good about what I’ve seen,” said Dr. Lemuel A. Moye, a member of the FDA’s Cardiovascular and Renal Drugs Advisory Committee that met on Feb. 28, 1997. Moye, a physician and biostatistician at the University of Texas, suggested it would be prudent to delay judgment until the study’s results were unsealed. “I’m afraid that we are rushing into this.”

According to a transcript of the meeting, Moye voiced concern about Posicor’s effect on heart rhythm and its potential to interact with other compounds. “Patients will be taking this in fairly uncontrolled situations in combinations of drugs which have ramifications yet unknown,” he said.

Another committee member, Dr. Robert Califf, professor of medicine at Duke University, said: “If this [drug] was really something that was dramatically different, better than anything else in the way of relieving symptoms, then I would look at it differently. But given the fact there are a lot of other effective therapies out there, why not be safe with the public?”

Indeed, scores of other drugs for treating high blood pressure were already on the market, and Posicor was not proved to offer lifesaving benefit.

The drug’s manufacturer, New Jersey-based Hoffman-La Roche Inc., saw no need to delay.

“There is no signal that there is arrhythmic or potentially arrhythmic risk with the drug,” said Roche’s Dr. Isaac Kobrin, terming the sudden deaths of four patients in another study “a chance finding.”

The committee voted, 5 to 3, to recommend approval of Posicor, with Califf and Moye in the minority.

After presiding over the five-hour discussion, the committee chairman, Dr. Barry M. Massie of San Francisco, abstained from voting amid a financial conflict: Massie was a co-investigator in Roche’s ongoing study of Posicor. After that meeting, Roche hired him as a speaker for the drug, Massie acknowledges.

On June 20, 1997, the FDA approved Posicor.

Four days later, a Roche news release quoted Massie to buttress the company’s claim that “the incidence of side effects was low” during clinical studies of Posicor.

Asked about this sequence of events, Massie said, “You do wonder how the world would perceive it. I’m glad I didn’t vote, let’s put it that way.”

Doctors were cautioned in speck-sized type--beginning on the 278th line of the drug’s label and again on the 365th about prescribing Posicor in combination with various medications, including allergy pills, tranquilizers, a sleeping pill and the heartburn drug Propulsid.

Authorities in Sweden in mid-1997 saw sufficient danger to keep Posicor off the market. But the U.S. approval spurred high hopes for Roche. Analysts at one brokerage firm, Salomon Smith Barney, projected sales of $2.9 billion within four years.

Six months after approving Posicor, the FDA advised doctors of the pill’s “life-threatening” danger. The agency announced that it had “received reports of dangerously lowered heart rates in about 20 patients.” Roche agreed to a label change--advising that Posicor should not be taken in combination with cholesterol-lowering drugs. This brought to 26 the number of drugs that doctors were warned not to prescribe with Posicor.

On June 8, 1998, Roche announced Posicor’s withdrawal, citing “evolving information concerning the potential for drug interactions” and “preliminary results” from the ongoing heart failure study that had drawn the attention of the advisory committee.

The study, Roche said, showed that the patients gained “no overall” benefit from Posicor.

According to those familiar with the matter, the study also found that the patients given Posicor died at a rate about 10% greater than those who took a comparator. “It definitely did not look good,” recalled Fenichel, who was then the cardiac drug division’s deputy director.

Apart from the clinical research, records filed with the FDA show that doctors and others reported Posicor as a suspect in the deaths of 100 patients.

“Posicor should not have been approved,” Moye said. “Therefore, any death that was attributable to Posicor was an unnecessary death.”

The FDA’s Woodcock and Lumpkin wrote in their May 1999 medical journal article that the problems that sunk Posicor also were “unexpected.” Asked recently about this, Woodcock retreated slightly; she said that serious adverse reactions resulting from Posicor’s interaction with one drug “perhaps could have been anticipated.”

Overall, she said, the agency had hoped for better compliance with advice in the label to avoid the concomitant use of Posicor and the 26 other drugs it interacted with.

A spokesman for Roche, Martin Hirsch, declined to answer questions but said the company “demonstrated sound and responsible judgment in the way [it] developed, launched and marketed and withdrew Posicor. And we have cooperated with the FDA throughout.”

LOTRONEX: Officer Foresaw Deadly Effects

‘Irritable bowel’ remedy pulled after reports of serious injuries.

Agency officials agreed in July 1999 to conduct a fast-track medical review of Lotronex, a pill from Glaxo Wellcome Inc. intended to treat irritable bowel syndrome in women. To justify such accelerated review, the FDA must find that the targeted disease is “serious.”

Irritable bowel syndrome can result in abdominal pain and frequent trips to the bathroom. But it neither maims nor kills people.

An FDA medical officer, Dr. John R. Senior, discovered during the review that four Lotronex patients in clinical studies suffered a potentially life-threatening complication called ischemic colitis, which results from inadequate blood flow to the colon.

Senior, a former pharmaceutical industry executive and a gastrointestinal specialist, knew the rarity of ischemic colitis: Some physicians can practice for decades without treating a single case.

While some cases would be mild and reversible, Senior wrote, ischemic colitis “can be catastrophic.”

Senior found other troubling results. He warned that 27% of the patients who took Lotronex in Glaxo’s studies experienced constipation. He noted that not a single patient who took a placebo pill developed ischemic colitis and that only 5% of the placebo patients got constipated.

Glaxo representatives denied that Lotronex had caused the cases of ischemic colitis and said any risks could be adequately managed. But Senior warned of the potential for Lotronex patients to suffer debilitating bowel injuries or death.

If these were the risks, what were the potential benefits?

FDA reviewers found that Lotronex improved symptoms in only 10% to 20% of the patients. Still, an FDA advisory committee, whose participants included a paid consultant to Glaxo, unanimously recommended approval. (The yes vote voiced by the Glaxo consultant, Dr. Arnold Wald of Pittsburgh, was invalid, agency officials say, because of his status as a temporary appointee.)

The FDA had a choice: Withhold approval of Lotronex until Glaxo undertook a major safety study to assess the drug’s link to ischemic colitis or approve the drug conditioned on a pledge by Glaxo to perform the study in the following year.

Top FDA officials chose not wait. They approved Lotronex on Feb. 9, 2000. The original labeling said that ischemic colitis had occurred “infrequently” in the clinical studies and that there was no way to predict which patient was at highest risk.

It was Lotronex’s first approval worldwide. Securities analysts estimated it would generate sales of up to $2 billion within five years.

A spate of bowel injuries emerged quickly--consistent with Senior’s fears.

In June, the FDA’s Woodcock embraced the crafting of a “medication guide” aimed at advising patients of Lotronex’s risks. But the leaflets were not delivered to pharmacies until late September. Meanwhile, Woodcock’s staff proposed a black box warning for Lotronex’s label but retreated when Glaxo publicly opposed the idea.

By October, 49 cases of ischemic colitis in Lotronex patients--including five deaths--had been reported to the FDA. Records show that no fewer than 91 patients were hospitalized, many with severe constipation. Several bowel surgeries, including the removal of a patient’s colon, were performed.

FDA officials who had backed the approval of Lotronex maintained their support for the drug into November, but staff epidemiologists pointed to the surgeries and deaths and the likelihood that those voluntarily reported events were a small fraction of the true scope of harm.

They urged withdrawal.

Glaxo and the FDA announced on Nov. 28 that Lotronex would be pulled from the U.S. market. At that point, Glaxo’s promised study of the drug’s link to ischemic colitis still had not enrolled a single patient.

Asked why the FDA approved Lotronex, given the ischemic colitis risk, Woodcock indicated that her aides had believed Glaxo’s view of the risk more than Senior’s.

“At the [November 1999] advisory co