Benderly (2007) - Experimental Drugs on Trial (Scientific American)

Benderly (2007) - Experimental Drugs on Trial (Scientific American)

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: MEDICINE CREDIT WILL THE MEDICINE CABINET of the future contain experimental drugs? 92 SCIENTIFIC AMERICAN © 20 07 SCIENTIFIC AMERIC AN, INC. October 2007 EXPERIMENTAL DRUGS ON TRIAL A controversial lawsuit challenges the FDA’s system of controlling access to experimental drugs and, some say, the scientific basis of drug approval By Beryl Lieff Benderly STEVEN PUETZER Iconica/Getty Images A bigail Burroughs was only 21 when she died. If her father and his supporters get t heir wish, however, she will attain a kind of immortality, joining Brown, Griswold, Roe and Miranda in the band of ordinary citizens whose personal travails have permanently changed the way Americans live. A lawsuit, Abigail Alliance for Better Access to Developmental Drugs v. Andrew von Eschenbach, contends that government regulations kept Burroughs from obtaining potentially lifesaving experimental cancer medicines that her doctor recommended, violating her constitutional right to defend her life. The U.S. Court of Appeals in the District of Columbia ruled against this claim in August, and the plaintiffs plan to appeal to the U.S. Supreme Court [see “Case Study,” on next page]. The ultimate outcome could become one of the most important court decisions ever to affect medical science, climaxing a conflict that has simmered since the early days of the AIDS epidemic. On the one hand are the heartrending emotional pleas of dying patients who are eager to try unproved experimental drugs. On the other are scientists determined to preserve the “gold standard” method that for decades has established which medications w w w. S c i A m . c o m work— a series of clinical trials that in total can take 10 years to complete. Both sides want to protect patients’ lives, but they differ on the best strategies toward that noble end. The other name on this potential landmark suit is commissioner of the U.S. Food and Drug Administration, the agency that regulates the sale of medicines in the U.S. and the trials that determine which are safe and effective and thus eligible to be marketed. The agency, which is at the same time facing intense pressure to speed drug approvals and also severe criticism for not being careful enough with decisions that could affect public health, declined to comment on the case. It is clear, however, that an Alliance win could vastly broaden access to medications that have shown only limited evidence of safety and, often, none at all of efficacy. This change would devastate the clinical trial system that has fueled unparalleled medical progress over 45 years, argues the Society for Clinical Trials, one of many opponents. If patients could get experimental d rugs outside of rigorous clinical trials, they would be reluctant to bother with such trials, and the surest way to determine whether a drug works and is safe would go out the window. Hope of being helped motivates many people © 20 07 SCIENTIFIC AMERIC AN, INC. KEY CONCEPTS ■ ■ ■ Large clinical trials are the gold standard for testing whether experimental drugs work. But they take time and money; meanwhile people who have run out of treatment options may die waiting for results. A recent lawsuit challenged the FDA’s rigorous require ments for drug testing, with the aim of making experimental drugs much more available. But if anyone can gain access at will to experimental drugs — most of which prove useless or even dangerous — will clinical trials be unable to attract sub jects? And will scientists end up with no good way to distinguish the truly helpful drugs from the chaff? —The Editors SCIENTIFIC AMERICAN 93 The quest to protect patients against dangerous drugs is at the root of the modern clinical trial system, which goes back decades. The Food, Drug, and Cosmetic Act of 1938 first required drugs sold in the U.S. to be proved safe after 107 p eople died from the then new wonder drug Elixir Sulfanilamide, which contained the solvent diethylene glycol, a poison used in antifreeze. In 1962 amendments added the requirement that drugs also be proved effective after t he worldwide thalidomide furor. The experi- ence with thalidomide was a catastrophe in m any countries but a triumph for the F DA . Widely used abroad for insomnia and morning sickness, thalidomide caused severe malformations in 10,000 newborns. Because an F DA examiner who was skeptical of the safety data had blocked approval, however, few of those victims were American. Then, 25 years later, after years of pressure from AIDS activists, the F DA began permitting patients with life-threatening illnesses but no approved treatment to take the risk of using experimental drugs outside of trials in limited circumstances. When the alternative is certain death, the FDA acknowledged, the balance of risk and benefit changes. A merican law has never yet recognized a constitutional right to unapproved drugs, however. Getting a drug developed and approved takes about eight and a half years, the FDA estimates, and only a small minority of those who start the process finish. The Pharmaceutical Research and Manufacturers of America estimates that five of 5,000 compounds that enter preclinical testing make it to human trials— and only one in five of those may reach the market. The “gold standard” of drug testing, the double-blind controlled clinical trial, compares an experimental drug against t he best standard treatment or, sometimes, against an inactive placebo. Neither the subjects nor the health care professionals in contact with them are supposed to know who receives what March 2007: Full nine-member Circuit Court hears the case reargued. ABIGAIL BURROUGHS, who was 21 when she died ERBITUX ( above), marketed by Bristol-Myers August 2007: Circuit Court votes, 8 –2, against the plaintiffs. of cancer in 2001, had sought to join clinical trials of the then experimental drugs Erbitux (cetuximab) and Iressa (gefitinib). Squibb, received approval in 2004; Iressa ( not shown), made by AstraZenica, was approved in 2003 for treatment of certain cancers. C ASE STUDY This fall the U.S. Supreme Court could decide whether to hear arguments about granting dying patients greater access to experimental drugs. THE CASE Abigail Alliance for Better Access to Developmental Drugs v. Andrew von Eschenbach (the current FDA commissioner) THE CONTENTION FDA regulations kept Abigail Burroughs from obtaining potentially lifesaving experimental cancer medicines that her doctor recommended, violating her constitutional right to defend her life. to enter trials, although the studies’ entirely different purpose — answering scientific questions about safety and efficacy rather than providing t herapy for individual participants — is sup posed to be explained before volunteers give their informed consent to participate. So are the statistical unlikelihood of benefit and the real r isks of harm from untested chemicals. Re search shows, however, that most subjects nonetheless hold what bioethicists call the “therapeutic misconception” that trials aim to cure them and offer a good chance of helping. Even m any physicians misinterpret their patients’ prospects in trials. Entering a trial really resembles “going to Las Vegas,” however, says bio ethicist and biochemist Adil Shamoo of the University of Maryland. “Everybody thinks they’re going to win in Las Vegas.” Most don’t. L ong Road to Rx THE HIGHLIGHTS July 2003: Abigail Alliance and the Washington Legal Foundation file suit. May 2006: U.S. Circuit Court of Appeals for the District of Columbia panel rules, 2–1, that a “mentally competent, terminally ill adult patient” has a constitutional right to use, on advice of a doctor, any drug that has passed the first step of the FDA’s threephase approval process [see “Safety in Numbers,” on page 97]. 94 SCIENTIFIC AMERICAN © 20 07 SCIENTIFIC AMERIC AN, INC. October 2007 COURTESY OF FRANK BURROUGHS ( left); IMCLONE SYSTEMS, INC. ( right) August 2004: U.S. District Court throws out the case, in part because the “court is not persuaded that the plaintiffs seek a recognized [constitutional] right.” while the experiment is under way. Participants must agree to be randomly assigned to either an experimental arm that receives the as yet unproved medication or to a control arm that does not receive it. Clinical trials, which occur in three phases, are designed to make statistically valid comparisons of the study arms, not to provide care or to cure subjects. “Clinical research ... is not a therapeutic activity devoted to the personal care of patients” but rather an experiment “answering a scientific question, with the aim of producing ‘generalizable knowledge,’” write bioethicists Franklin G. Miller of the National Institutes of Health and Howard Brody of the University of Texas Medical Branch in an article in Hastings C enter Report . Trials “promot[e] the medical good of future patients [through] scientific knowledge derived from experimentation with present patients— a frankly utilitarian purpose.” Phase I introduces the drug into 20 to 80 individuals who are sometimes healthy volunteers. In cases involving highly toxic substances such as cancer drugs, subjects are generally patients with advanced, untreatable disease. This phase seeks to learn about safe dosage levels, side effects, and the drug’s effect on metabolism and its mechanism of action in humans. Only occasionally does phase I produce evidence of effectiveness. Drugs that appear sufficiently safe get p ermission to proceed to the next phase, although that safety finding is not definitive. Phase II compares the experimental drug and the control using several hundred people who have the condition. It seeks evidence of efficacy, risks and short-term side effects. If the results and the study design appear satisfactory, phase I II next tests the drug against a control in up to several thousand people with the condition, seeking much more extensive data about effective ness and safety. Based on these results, the F DA determines whether the drug merits approval. COURTESY OF FDA HISTORY OFFICE Too Long to Wait Current procedures do permit some patients with serious or fatal diseases but no approved treatment options to obtain, outside of clinical trials, drugs that have passed the second phase. Burroughs’s inability to obtain such a promising drug inspired her father, Frank Burroughs, to start the Alliance in November 2001 after her death earlier that year. During her final months, her Johns Hopkins University oncologist thought t hat Erbitux (cetuximab) or Iressa (gefitinib) could help fight her squamous cell head and neck w w w. S c i A m . c o m * RECALLED * E LIXIR SULFANILAMIDE, which contained a poison found in antifreeze, killed 107 people in 1937. The tragedy spurred the passage of the 1938 Food, Drug, and Cosmetic Act, which required that manufacturers prove drug safety to the FDA before receiving marketing approval. cancer because both counter epidermal growth factor receptors (EGFR) plentiful in her tumor. Neither had yet won approval. To be permitted to experiment on human beings in the first place, sponsors of candidate drugs must submit data from preclinical studies showing a reasonable likelihood they will prove safe and effective. The sponsor must also recruit the experimental subjects, who, in the case of serious or fatal conditions, are often individuals such as Burroughs, with advanced disease and no approved treatment options. A clinical trial offers them a shot at getting an experimental drug, but how good a shot depends on the phase of the trial they participate in and the study’s research design. The sponsor provides the test drug free to participants but is under no obligation to accommodate others out of its often small supply. That, Scott Gottlieb believes, is why an Alliance win may increase access less than many people expect. Gottlieb is a former F DA deputy commissioner for medical and scientific affairs and now a physician in private practice and a resident fellow at the American Enterprise Insti- © 20 07 SCIENTIFIC AMERIC AN, INC. Most subjects hold what bioethicists call the “therapeutic misconception” that trials aim to cure them and offer a good chance of helping. SCIENTIFIC AMERICAN 95 T HALOMID (THALIDOMIDE), manu- factured by Celgene, received approval in 1998 to treat an inflammatory condition of Hansen’s disease, or leprosy. In the early 1960s a skeptical FDA examiner declined to approve thalidomide, sparing Americans from the number of drug-related birth defects experienced in countries where the drug was used to combat morning sickness. * DENIED * L AETRILE, sold in other countries as a cancer treatment, is not approved by the FDA, because it lacks proof of efficacy and produces side effects resembling cyanide poisoning. 96 SCIENTIFIC AMERICAN tute. “The biggest impediment [to early access ing drugs that could compromise the ability to outside of trials] is the unwillingness of some conduct very formal and rigorous clinical studcompanies to offer the drug,” he says. A group ies or through more rigorous evaluations that of patients with Parkinson’s disease, for exam- might forestall early access but enable the develple, lost a suit to force Amgen Corporation to opment of rigorous clinical data that can guide continue providing them with an experimental future decision making,” states Gottlieb in an drug they believed to be very effective that they essay in a February 2007 white paper published received as subjects in a trial the company chose by the Food and Drug Law Institute, a nonprofit to terminate. Sponsors may stop trials at any educational organization in Washington, D.C. time. The F DA also stops trials when the pattern Instead of sound science, Shamoo believes, of “adverse events” indicates undue risk or when open access to experimental drugs would create the evidence of benefit is so strong that main- “large, uncontrolled clinical trials” that would taining a control group becomes unethical. endanger patients by exposing them to untested The Iressa trial had no room for Burroughs, chemicals and would slow research by removing and the Erbitux trial was studying only colon subjects’ incentives to enter the controlled clinicancer. Neither sponsoring company would sup- cal trials capable of producing valid results. What ply her under the rules permitting compassionate is more, people would pay for the drugs they reuse of experimental drugs outside of trials. She ceived outside of trials, giving sponsors incendied soon after finally securing a spot in a trial of tives to “sell the unapproved drug today” rather a third EGFR inhibitor. Both drugs she sought than financing the “scientifically rigorous and subsequently gained approval; Erbitux is now expensive clinical trial process,” the FDA argued widely used for her cancer. Abigail Alliance’s co- in its brief for the Appeals Court en banc hearing. founder and senior adviser, professional geolo- L etting everyone choose any drug could even gist Steven Walker, has seen the benefits of using harm promising medicines, critics believe, be such drugs. Walker’s wife took Erbitux in a trial cause very sick people using inappropriate chemi n September 2002 and was “resurrected,” icals might suffer adverse reactions serious changing from “a bedridden, dying person into enough to convince the FDA to reject drugs that someone who was skiing, hiking and working could otherwise help carefully chosen patients. every day,” he says. After “marginal progression Walker, however, says that open access will [of the disease] in her liver,” however, she was no “end the abandonment” of the great majority of longer given the drug and died in June 2003. patients who cannot get into trials. He foresees no damage to recruitment because, under the A Better System? Alliance’s proposal, anyone seeking access to an The dispute ultimately involves a “fundamental experimental drug would first have to apply for trade-off” over “how ... patients are best served a clinical trial. Nor would patients opt to try in the long run: through earlier access to promis- dangerous or ineffective chemicals, Walker in- © 20 07 SCIENTIFIC AMERIC AN, INC. October 2007 COURTESY OF FDA HISTORY OFFICE * RECONSIDERED * sists, because “patients and physicians do not pursue investigational drugs in any kind of significant number unless that drug has in fact shown some evidence of effectiveness and at least enough safety . . . that the patient’s physician” recommends trying it. Perry Cohen, founder and director of the Parkinson’s Pipeline Project, a patient group working to “accelerate innovation” in treatment, also asserts that patients can make good choices. Cohen, who was forced by the disease to retire as a health consultant, now serves as an FDA patient representative for Parkinson’s. “Even if you’re not going to die,” he says, people with “a serious illness leading to a worse fate” deserve the right “to make the risk-benefit trade-off individually, with proper counseling from expert medical professionals” and to decide whether to try experimental drugs. Critics reject those contentions. Many side effects are unknown before large-scale testing, they argue. Once a right is established for some, they add, it inexorably expands to others. A llowing people to choose among unap proved chemicals would not mean that “the patient would suddenly have access to the one drug out there” right for him and his condition, says Colin Begg, chair of epidemiology and bio - * APPROVED * * CAVEAT EMPTOR * P ENICILLIN, the “wonder drug,” was tested by the FDA for wartime use and later certified for public consumption. Cooper’s New Discovery and Mrs. Winslow’s Soothing Syrup, pre-1906 “patent” medicines that predate the FDA, were typical in their extravagant claims. SAFETY IN N UMBERS New drugs must pass three clinical trial milestones before they can be marketed to the public. A small fraction of candidates complete the development and approval process, which takes about eight and a half years. COURTESY OF FDA HISTORY OFFICE Phase I: Tests for safe dosage levels and side effects in 20 to 80 volunteers. Also probes the drug’s effect on metabolism and its mechanism of action. Phase II: Several hundred subjects receive the drug in a probe of its efficacy, risks and short-term side effects. arketed by Eli Lilly, * APPROVED * HUMULIN (HUMAN INSULIN), mby the FDA in 1982. became the first genetically engineered drug approved w w w. S c i A m . c o m © 20 07 SCIENTIFIC AMERIC AN, INC. Phase III: More extensive data about effectiveness and safety come from testing the drug in up to several thousand people. SCIENTIFIC AMERICAN 97 [THE AUTHOR] Beryl Lieff Benderly is a prizewinning health and science journalist in Washington, D.C. Her books include I n Her Own Right: The Institute of Medicine’s Guide to Women’s Health Issues ( National Academies Press, 1997). 98 SCIENTIFIC AMERICAN SHORTENING THE CRITICAL PATH A lthough clinical trials are still the FDA’s gold standard for assessing the safety and effectiveness of drugs and medical devices, the agency acknowledges plenty of room for improvement in testing methods for prospective therapies. Its Critical Path Initiative, launched in 2004, is an attempt to modernize the process at every stage on the path from promising compound to approved product, in the hope of improving data quality and speeding the best drugs to market. A list of 76 pri ority projects emphasizes applying the latest scientific tools and techniques to ruling out useless or toxic compounds early and assessing more quickly whether candidate drugs have the desired effect on disease. Causing liver injury, for instance, leads many drugs to fail in early human trials, and even ap proved medicines must sometimes be pulled from the market for the same reason. To avoid these late and costly realizations, the FDA is collaborating with a computer-modeling firm to design a “virtual liver” that can help predict whether a compound may be toxic in people, alone or in combination with other drugs. Many Critical Path projects are also exploring the use of biomarkers, w hich are various forms of evidence of biological activity inside the body — such as levels of certain proteins, gene activity or internal imaging data — to gauge whether a drug is having the predicted effect without having to await results of a large human trial. This year, for example, the FDA permitted the first “phase 0” human trials to look for biomarker evidence of a candidate drug’s action on its target in the body. In the earliest of these, National Cancer Institute researchers gave tiny doses of a compound called ABT-888 to six ad vanced cancer patients and saw that their levels of a significant protein dropped by more than 9 0 percent— a sign that the FDA and the drug’s developer can use to accelerate the compound’s progress to formal clinical testing. Making those trials shorter and smarter is an other agency goal, and approaches under consid eration include adaptive design, which allows pro tocol changes in response to preliminary data while a trial is still under way. Testing drugs in multiple combinations, as they are likely to be used in reality, could also speed approvals. Plug ging data from early trials into computer simula tions can help identify the most effective drug doses for later, larger trials. The FDA is also encouraging pharmaceutical companies to pool their data from every stage of drug testing and use bioinformatic techniques to extract the maximum amount of in formation from that raw material, so that prob lems with a whole class of compounds are flagged sooner, as are promising leads to valuable new medicines. — Christine Soares, staff editor © 2007 SCIENTIFIC AMERIC AN, INC. October 2007 COURTESY OF FDA HISTORY OFFICE Proponents of greater availability for new drugs want more emphasis on earlier approvals followed by extensive postmarket monitoring. statistics at Memorial Sloan-Kettering Cancer and Bayesian statistics can make this feasible, he Center in New York City and of the Society for believes. (Ironically, the FDA is meanwhile facing Clinical Trials committee that drafted a posi- criticism for giving overly speedy approval to tion paper on the Abigail Alliance proposals. drugs such as Vioxx, for pain, and Rezulin, for The patient “would have access to an entire cup - diabetes, which have been withdrawn from the board full of drugs, most of [them] useless or market as unsafe.) toxic, [without] any idea of which . . . to take,” Bayesian statistics do not require “perfect exbecause physicians not expert in the drug would periments,” Gottlieb explains, but rather use have no evidence to guide the decision. “real-world data, ‘dirty data’ . .. to derive concluThe trade-off, however, need not be as stark sions about probabilities from nonrandomized as the most extreme positions imply, Gottlieb be- data sets.” Adaptive trial design permits changes lieves. Improvements in trial design could get an- to be made while a trial is under way based on swers sooner, shortening everyone’s wait for new the data that emerge during the trial. Researchdrugs and reducing the number of participants ers, for example, could modify randomization to needed for control arms, he says. He sees “a lot give more subjects a treatment apparently pro of opportunity to develop better scientific tools ducing a better response or to overenroll categofor measuring drug response and [for designing] ries of patients who, because of their individual clinical trials using more adaptive approaches.” characteristics or responses, appear likeliest to The “gold standard is crumbling,” Cohen adds. benefit from the experimental drug. Gottlieb, “Most of the acute conditions that fit that model Walker and Cohen also all favor increased use of are [already] done.” Like Walker, Cohen wants surrogate end points, which substitute physio more emphasis on earlier approval followed by logical markers believed to predict outcome for extensive postmarket monitoring. Modern in- evidence of actual outcomes. Such approaches formation technology, adaptive research designs can “speed drug development and … make clin- * WITHDRAWN * V IOXX (ROFECOXIB), an anti- inflammatory approved in 1999 that soon became widely prescribed, was in 2004 withdrawn from the market by developer Merck & Company in the wake of concerns about the risk of heart attack and stroke in patients. * A LERT * JOSH SHER ( Vioxx box) AND JOHN KAPRIELIAN ( Vioxx pack) Photo Researchers, Inc. ( left); JUSTIN SULLIVAN Getty Images ( right) AVANDIA (ROSIGLITAZONE), made ical trials more targeted to patients who are likelier to benefit,” yielding “much more reliable information about who’s going to ultimately benefit from using the drug,” Gottlieb says. The FDA is currently evaluating these techniques [see box on opposite page]. But “there’s no promised land” in that direction, Begg warns. He believes that Bayesian methods are relatively minor adaptations of the way patients are randomized and the way trials are designed. “They do not get the answer faster; they result in a longer trial,” he says. Surrogate markers do not always reliably predict clinical outcome. Nor will these techniques resolve the assertions of the Abigail Alliance, Begg continues: “It’s just a different way of designing clinical trials. You still have to do the efficacy testing [on] patients to figure out which drugs work and which drugs don’t.” “The long history of medical research is replete with treatments that initially seemed promising to patients, doctors, and especially to their own inventors, but which careful study revealed to be worthless or harmful,” states the Society for Clinical Trials position paper. In the cancer field alone, these failures include laetrile, initially impressive in animal studies, and high-dose chemo- More on the Web To see more historic drugs, go to w w w w. S c i A m . c o m therapy followed by autologous bone marrow or stem cell transplant for breast cancer. Thousands of women underwent, and some died from, this excruciating and costly experimental procedure after a lawsuit forced insurers to pay but before clinical trials finally proved it no more effective than standard therapy. Other Implications The courts will decide not on issues of experimental design and risk-benefit analysis but on issues of legal precedent and interpretation. An Alliance win, suggests John A. Robertson of the University of Texas at Austin School of Law, could, for example, overturn rules or laws against creating embryonic stem cells. “Under the principle recognized in Abigail Alliance,” he writes in an article in the Hastings Center Report, courts would have to weigh “whether the state could protect embryos at the expense of the patient’s life and liberty.” Even if the FDA prevails, some change appears likely. The agency is working on new rules it says will make early access outside of trials less onerous. “Scientists have to realize that there’s a crisis of confidence” in clinical trials among the public, Cohen says, and “holding up this standard as immutable is part of” the reason. Whatever the Abigail Alliance case brings, the health of the clinical trial system, and thus of the American people, depends on restoring that confidence. g © 20 07 SCIENTIFIC AMERIC AN, INC. by GlaxoSmithKline, was approved in 1999 to treat type 2 diabetes; in May the FDA issued a safety alert about its potential to raise the risk of heart attack, but it has not yet confirmed the clinical significance of related findings. A drug for type 2 diabetes in the same class, Rezulin (troglitazone), was withdrawn voluntarily in 2000 by the ParkeDavis division of WarnerLambert at the FDA’s request because of concerns about its risk of liver toxicity. ➥ MORE TO E XPLORE Access before Approval — A Right to Take Experimental Drugs? Susan Okie in New England Journal of Medicine, Vol. 355, No. 5, pages 437–440; August 3, 2006. The Right to a Trial. Jerome Groopman in the New Yorker; December 18, 2006. w content/articles/061218fa_fact CDER Handbook, a description of the drug development and approval process provided by the Center for Drug Evaluation and Research of the Food and Drug Administration: w index.htm FDA press release on the proposed new regulations: w NEWS/2006/NEW01520.html SCIENTIFIC AMERICAN 99 ...
View Full Document

This note was uploaded on 11/07/2011 for the course MAP V55.0310.0 taught by Professor Tracejordan during the Spring '11 term at NYU.

Ask a homework question - tutors are online