Fen-phen was never approved by the FDA, nor, as things stand to date, did it ever need to be approved. Fen-phen is the combination of two drugs that were individually FDA-approved many years ago for short-term use in support of comprehensive treatment of serious obesity. Their concurrent long-term use is an "off-label" use, falling into the gray area of prescriber discretion. If physicians are willing to take the usually slight chance that something bad will happen, they can prescribe essentially anything in the pharmacy, in any dose, for any indication, to any patient.

Most physicians are caring, cautious, and intelligent. However, give some a double-blind, placebo-controlled trial published in a peer-reviewed journal, present them an imploring, overweight patient who almost surely could benefit from losing weight, throw in the fact that other physicians are using the drugs in question with apparent success (both financial and therapeutic), and the pen is attracted to the prescription pad, FDA-approved use or not. Pharmaceutical manufacturers and clinical researchers know this. It is a fast track to medical acceptance for new treatments.

The standard process for prescription drug approval in the United States is rigorous and quite lengthy, the journey from Erlenmeyer flask to druggist's shelf averaging 8 to 12 years, assuming the FDA finds no problems. This generally laudable caution dates from prior to such 1950s events as the thalidomide disaster, which the United States largely escaped. To this day, the rest of the world serves as a kind of unofficial proving ground for drugs making their way laboriously through the FDA approval process. Which brings us to the story of anorexic drugs.

Drugs that produce anorexia, or loss of appetite, have been available for at least 50 years. Amphetamines were quite popular in the fifties and sixties. Unfortunately, they were often addictive, which contributed to the unsavory reputation of patients and physicians who used them. They also needed to be administered in escalating doses because of the rapid development of tolerance among users, which made them ultimately more dangerous since the therapeutic-to-toxic ratio was not broad. As with most medications, they also had the unfortunate quality of no longer being effective once discontinued; hence, weight regain was the rule. All these features continued to stigmatize the subsequent generations of anorexic drugs, at least up until about 1990.

There were justifications for this new permissive attitude toward obesity drugs. First, over the last 20 years or so, the prevalence of overweight among American adults has increased from about 25 percent to about 35 percent, with no asymptote in sight. [1] Second, the increasing recognition of the strong genetic component to variations in body weight, and recent molecular genetic research advances (e.g., [2]), have led many to conclude that obesity is a problem of biological origin and therefore warrants a biological solution, and that it is a chronic problem therefore warranting chronic treatment. Third, many practitioners and patients had come to see "behavioral" approaches as something tried and failed. Fourth, a widely cited review [3] presented convincing data that even modest weight losses (e.g., 10 percent of initial body weight) could produce meaningful health benefits. Fifth, a widely cited paper [4] seemed to imply that thinner was always better and that even slight degrees of overweight substantially increased mortality rate. For many people, all of these points seemed to lead to the inescapable conclusion that antiobesity drugs were justified for the majority of obese people and for long-term use. Additionally, although we have only anecdotal information in this regard, it appeared that many physicians also considered the drugs appropriate for many non-obese people who wanted to lose weight.

Finally, the public had an unquenched desire for an easier solution to the weight-control struggle than a permanent change of lifestyle. Long-term use of medications that safely and effectively relieved hunger and caused weight loss seemed ideal. Enter fen-phen.

The rationale for combining the two medications is that they have different mechanisms, which might have a greater effect than either alone, or than two drugs of the same class. In addition, their side-effect profiles seemed complementary. Phentermine is heir to the amphetamine-like drugs of the past era and, while apparently nonaddictive, exhibits mildly stimulative properties that interfere with some patients' sleep. Fenfluramine (Pondimin), either as the original racemic (DL) mixture or as the more recently approved dexfenfluramine (Redux) stereoisomer, while also an amphetamine derivative, has seritonergic and central nervous system depressant actions and may cause drowsiness. Taking both was hoped to cancel the side effects, leaving the mind in balance. The investigator responsible for testing this hypothesis is Michael Weintraub, who published the results of a NIH-funded, initially double-blind, placebo-controlled trial in 1992. [5] While it was noticed that there was a substantial proportion of dropouts by the study's end (60 percent after four years), the results were nonetheless impressive. The group randomized to the combination of fenfluramine and phentermine experienced about twice the weight loss of the placebo group, and showed little tendency to regain weight for as long as the medications were continued. Though more subjects dropped out of the treatment group than the placebo group, the side effects were typically mild or transient, and none seemed truly dangerous.

Since no other long-term studies were available, and the results seemed plausible, and the timing was right, the Weintraub study spawned a phenomenal increase in public and professional interest in the use of anorexic drugs, especially the combination of fenfluramine and phentermine. Now, instead of short-term or intermittent use, the logical conclusion was that long-term, perhaps lifelong treatment was needed to maintain the achieved weight loss. Indeed, as had been noted years ago, it was inappropriate to attach a higher standard of efficacy to anorexics than to other medications. We do not say of beta-blockers used to treat hypertension, for example, "Sure they lower blood pressure, but it goes right back up as soon as you stop them; why bother using them in the first place?" Clinics using fen-phen grew in popularity, and new clinics appeared. Occasionally, the combination was even touted as useful in other conditions, including alcohol and drug dependence. Some clinics and diet programs that could not or did not use medications as an aid to weight loss did not enjoy the same increase in popularity, leading some to forge alliances with practitioners who could and did use them.

In the midst of this rebirth of anorexic fervor, dexfenfluramine (Redux) appeared on the FDA's stage. While it was indeed to be, as its sponsors proclaimed, the first drug approved for the treatment of obesity in twenty-odd years, it cannot truly be called a new drug since the recommended daily dose of the old fenfluramine (Pondimin) contains precisely the same amount of dexfenfluramine as the recommended daily dose of the new dexfenfluramine. Despite this fact, there was the possibility that the lower recommended dose of dexfenfluramine would result in fewer or less severe adverse effects on average than the old dexfenfluramine, a worthwhile contribution if it proved true.

There was some controversy regarding toxicity. One potential toxic effect was already known, though its magnitude was unclear: Over the years, the fenfluramines had been associated with cases of a rare, frequently fatal disease called primary pulmonary hypertension (PPH). A 1996 compilation of the European PPH experience found that patients who had used anorexics (mostly fenfluramine) for three or more months were 23 times more likely to suffer from PPH than those who did not use anorexic agents. [6] This still amounted to an incidence of only about one in 30,000 users per year. Unknown was whether the risk would rise with lifelong use. An interesting analysis in an accompanying editorial concluded that the risk-benefit ratio of using anorexics like the fenfluramines was highly favorable. The authors estimated that 280 lives would be saved from weight-loss-related decreases in cardiovascular and other disorders for every 14 killed by treatment-related PPH [7]. Shortly thereafter, the editors of the New England Journal of Medicine issued a clarification: both of the editorial authors had been paid consultants to the manufacturers of Redux. More important than this ad hominem critique of the editorial was the point that there were substantial flaws in the editorialists' reasoning. [8]

There was also evidence that short-term use of dexfenfluramine at high, but not extraordinarily high doses when adjusted for body surface area, caused long-term depletion of brain serotonin in primates and other animals. [9] A group of scientists, including one of us (Lawrence J. Cheskin), expressed concern to the FDA that further studies were needed, as the clinical significance of these findings was not known. At one point in the FDA proceedings, a vote of the outside advisory committee was taken to delay approval until further data was presented regarding this safety issue. A later meeting, at which no new data addressing this question were apparently presented, resulted in a vote of five to four in favor of approval in 1996. The Redux Revolution, as a popular book was titled [10], had been given the green light.

About a year and many prescriptions later, in July 1997, the editors of NEJM allowed the authors of an upcoming article to release the study results to the lay press immediately. The study, from the Mayo Clinic, reported 24 cases of valvular heart disease in formerly healthy patients who had been treated for as little as a month with fen-phen. [11] The valvular problems were often of a specific type that was seen in serotonin-excess conditions like some carcinoids and ergotamine toxicity. Six had led to open-chest valvular surgery. Being an uncontrolled case series, there was no direct evidence that this association was causative or even real. Thus, case gathering continued, while the manufacturers added to the package insert of the fenfluramines a statement noting the possible problem.

Analysis of echocardiograms from five centers soon revealed that a surprisingly high and consistent percentage of patients who had taken either fenfluramine had valvular abnormalities. Though most of these abnormal findings were mild or moderate in severity, 30 percent had either aortic or mitral valve regurgitation. Meanwhile, an additional 58 cases of valvular disease were reported to the FDA of patients who had taken fen-phen, fenfluramine alone, or dexfenfluramine alone, but not phentermine alone. The case of a 29-year-old woman who died of pulmonary hypertension after having taken fen-phen for only 23 days was also published. [12] Such cases led to the state of Florida banning fen-phen. The nation soon followed suit. At the FDA's request, the manufacturer of fenfluramine and dexfenfluramine - but not phentermine - withdrew the drugs from market on September 15.

Are there lessons to be learned from this tale? Perhaps. For one, it exposes the problem of "off-label" use of FDA-approved prescription drugs. A risk-benefit analysis of this common practice might go something like this: Benefits include faster access to new and innovative uses of medications already considered safe. Risks include the danger of unstudied and unknown risks, as exemplified above. Some would parenthetically add the risk of lawsuits. Of indeterminate status might go the dangers of what some pejoratively call "cookbook medicine." Those of us more favorably inclined call it "clinical care guidelines," "care maps," or "managed care." This category of risk is probably more imagined than real, as we are aware of mostly favorable results when the practice of medicine is scrutinized with an eye toward elevating the level of knowledge applied to specific conditions. Overall, we would judge that a case has been made for less tolerance of off-label practice, rather than continued "looking the other way" until there's a problem. This would require closer surveillance of prescribing patterns and perhaps legislation to enhance enforcement.

For all concerned, a lesson is that medications, especially those that have known toxicities like the fenfluramines, should be treated with the utmost respect. Patients must understand that all drugs have side effects, and that we can never claim to know what the adverse effects will be in advance. To demand (or prescribe) a medication for solely cosmetic reasons is potentially a fatal mistake. Practitioners must remind themselves that there is a risk to any treatment recommended; the patient must understand the known and unknown risks. This is true informed consent. If you are treating obesity, make sure the patient is truly medically obese. [13] There is nothing worse than a serious complication in a patient who did not clearly need the treatment. Fen-phen must also give manufacturers pause: conservative prescribing indications are best, as an overmarketed, overprescribed drug may end in disaster.

The truth is that we can never be assured that adverse episodes like the rise and fall of fen-phen will not again occur, even with the most stringent practice guidelines. Science is hypothesis driven, and when no one has a hypothesis, as was the case with the valvular problems prior to some astute clinicians' and pathologists' uncontrolled observations, even what later appears to have been a rather obvious problem can remain invisible. Throw in a high-stakes issue like controlling obesity, and the results are hard to predict. Perhaps the best we can advise is that the FDA remains as independent and rigorous as we can make it, that the post-market surveillance phase be as rigorous and organized as the approval phases when there is any question of toxicities, and that the burden of proof remain on the manufacturers to continually prove that its products are both safe and effective: that it be as difficult to keep a medication on the market as to get it there in the first place. What is required of us as scientists, and as care providers, is to reexamine and refine continually what it means to "do no harm." The public deserves no less from us.

David B. Allison, Ph.D., is associate research scientist at the Obesity Research Center and assistant professor of psychology in psychiatry at the Columbia University College of Physicians and Surgeons, New York City.

Andrzej Krauze is an illustrator, poster maker, cartoonist, and painter who illustrates regularly for HMS Beagle, The Guardian, The Sunday Telegraph, Bookseller, and New Statesman.

Send us your comments and ideas for future articles.

Endlinks

References - includes linked, freely accessible abstracts for most cited papers.

"The New Miracle Drug?" - this September 23, 1996, Time magazine cover story article contrasts the excitement surrounding Redux with the increasing awareness of its dangerous side effects.

Phentermine and Fenfluramine Links - Barbara Hirsch, who uses phen-fen as part of a weight-loss plan, maintains this extensive set of links to fen-phen-related articles. Most of these articles are from the popular media. Part of Barbara's Obesity Meds and Research News site.