Helsinki Discord? A Controversial Declaration

Washington—A long-awaited fifth revision of the Declaration of Helsinki, widely acknowledged as the foundation of medical research ethics, is receiving mixed reactions for discouraging placebos in clinical trials and mandating researchers to provide "best proven" therapy to participants after a trial.

Patient protection advocates, who lobbied hard for the new language, hail the revision as a major step toward abolishing what they see as unethical research, especially in developing nations. Delon Human, MD, secretary general of the World Medical Association (WMA), which put forth the declaration, is equally enthusiastic. He says the revision is an ideal—not a "manual of protocols and procedures"—but that it sets an ethical standard "which I don't think can be set high enough."

Or can it be? Some researchers and bioethicists are calling the recommendations on placebo use unwarranted and unnecessary, a possible bane to future research. And the US agencies responsible for protecting patients in medical studies, namely the Food and Drug Administration (FDA) and the Department of Health and Human Services (DHHS), are noncommittal about whether they will subscribe to the revision, which provides no means of enforcement.

"I think it's scientifically and ethically incorrect," said Robert Temple, MD, director of the FDA's Office of Drug Evaluation. "We'll have to see if the Declaration of Helsinki remains the ethical standard for the world." Temple cautioned that he was speaking only for himself, not for the FDA, but he later added that he would try to persuade others in the agency to his point of view.

Greg Koski, MD, PhD, director of the DHHS's Office for Human Research Protections, simply said, "This is an area of active discussion in the department. There will be more information forthcoming as we try to offer some guidance to researchers."

Both the DHHS and FDA are reviewing the new declaration with one eye on their own policies and another on parallel ethics statements being prepared by the National Bioethics Advisory Commission and the Commission for International Organizations of Medical Sciences, part of the World Health Organization. If these groups issue statements straying from Helsinki's policy—a very real possibility, according to several sources—researchers will have to play "pick the doctrine" when planning new studies.

Helsinki's paragraph 29 is clear: new treatments should be tested against the "best current" treatment, not a placebo, unless no proven treatment exists. But the interpretation and practical implications of such a sweeping statement are murky. "This will be where we have our biggest debate in the years to come," said Human.

Nancy Dickey, MD, former president of the American Medical Association and chair of the WMA work group that hammered out the revision, takes a pragmatic stance. "We are asking scientists to give a lot more consideration to when they use placebos and why they're using placebos," she said. "I think the declaration makes it very clear that it is not acceptable to take a new treatment and test it only against placebo. An adequate research protocol will test proposed new treatments against the best proven treatment."

But Dickey, who is now interim dean of the College of Medicine at Texas A & M University, said it may be justified, in cases where currently commonly used treatments have not been proven or have been proven in substantially different populations than the study group, to test a new treatment against both a placebo and the best current treatment. Such three-armed studies—while prohibited by a strict interpretation of the revised declaration—are quite useful, according to study design experts. "They're an excellent design," said Temple, who has recently published on this topic (Ann Intern Med. 2000;133:455-463). They provide the maximum amount of information about the existing drug and the one being tested, something Temple and Dickey agree on.

In contrast, comparing a new treatment with only the established treatment, the study design mandated by a literal interpretation of Helsinki , is less optimal, according to Temple. If no difference in effect emerges between the standard drug and the new one, "you don't know if either of them worked."

Conditions for which there are no proven treatments are clearly excepted from the placebo restriction. Studies of nondebilitating, short-term illnesses like headaches and colds also should not be subject to the requirement, said Peter Lurie, MD, MPH, who pushed hard for the antiplacebo language as deputy director of Public Citizen, a consumer rights group in Washington, DC.

But this gray area has led to confusion. "I wish they would have clarified their intent," said Temple. He interprets Helsinki literally, and believes the framers meant the placebo restriction to apply to all clinical trials, no matter how trifling or transient the condition under study. He was surprised to learn that Dickey supports three-armed studies.

As an alternative model, Temple points to the standards of the International Conference on Harmonisation (ICH), an international consortium of drug regulatory agencies. The ICH posits that "whether or not it's OK to use a placebo depends on what happens to the person who isn't getting the usual treatment," said Temple. In other words, if no harm comes to the patient, a placebo is allowable.

The placebo debate aside, the term "best current" swings open another wide door of interpretation, especially when considering trials in developing nations. For researchers conducting trials outside the United States, does it mean the best care available in the United States, or the best available locally?

Human acknowledged that it is "extremely difficult" to define the term but that, in most cases, it means the standard of care in the developed world. "The intent was to give the patient access to the best potential or possible or current treatment of the day." That interpretation has the drug industry and some academic researchers on edge.

"The worry in anything like this declaration are the exceptions where we have a standard of care in the United States and you're doing an overseas study and that standard of care is not available. What do you do then?" asked Gillian Woollett, PhD, senior vice president of the Pharmaceutical Research and Manufacturers of America (PhRMA), a Washington, DC, lobbying group.

For patients' rights activists like Lurie, the answer is clear: "The most simple reading is that you must provide the best proven universal therapy. It's a scientific standard, not an economic one."

But economics certainly factor into the equation. The most contentious debate—and, according to many, the very reason Helsinki was revisited—involves HIV/AIDS studies in Africa. Human says that triple-drug therapy is the standard of care and that anyone who wants to do an HIV/AIDS study in Africa or anywhere else should provide such treatment to everyone in the study. But researchers who have spent time in the field say this is not only expensive but practically impossible.

Thomas Quinn, MD, a Johns Hopkins and National Institute of Allergy and Infectious Diseases researcher whose studies in Thailand and Uganda have been attacked as unethical—despite approval by multiple institutional review boards, including those in the study countries—said many places lack the resources to provide basic needs, let alone the close medical supervision HIV/AIDS therapy requires. "They don't have running water and they don't have watches, they use the sun to tell time," said Quinn, referring to the strict schedule called for by the complex drug regimens. In addition, physicians and clinics may not be available to deal with severe adverse effects if they occur. "It's an infrastructure problem," Quinn said, a very distressing situation, but one too large for medical researchers alone to solve.

Public Citizen's Lurie dismisses these considerations as "ethical relativism," saying, "We object to studies that would be unethical in rich countries, but are deemed OK in the developing world because they have no medical care." Quinn and others warn that this stance could slow the pace of research.

First adopted in 1964 and revised four times previously, the latest revision of Helsinki was unanimously ratified by the WMA's General Assembly in October. But it took 3 years of wrangling about language and semantics, a divisive first draft that many said weakened the statement, and constant prodding by Lurie and others to reach the accord.

Despite all this, several researchers charged the WMA with not being open enough during the process. Dickey and Human countered that the WMA canvassed a wide range of viewpoints during the ordeal. Dickey said the group "took pains to make the process inclusive" by soliciting comments via the Internet and presenting drafts at international meetings. Human said he "aggressively" consulted with other ethics groups, including the National Bioethics Advisory Commission and the Commission for International Organizations of Medical Sciences.

These efforts apparently did not reach the drug industry. Woollett said she was not aware of any efforts by the WMA to contact drug companies: "After all, it's much easier to recommend that somebody else volunteer something."

The "something" she referred to—namely, supplies of drug provided to study participants after a trial—could be expensive for drug makers. Paragraph 30 of the revision addresses this issue for the first time, stating that all participants in a clinical trial should have access to the best therapy after the study. Making drug companies bear this burden could slow new trials, argues Woollett. Asking the governments of developing nations to pay for the treatments is another option, but developing nations generally spend woefully few dollars on health care as it is.

Research participants always subject themselves to risks, asserted Dickey, "So what's the benefit? If the benefit is that we've now proven it for the sake of the rich Americans . . . and we're not going to give it to you, that's a problem." She added that the details of who pays for what should be negotiated before the study begins.

Quinn points to his Ugandan study of sexually transmitted diseases (STDs) as an example of effective posttrial treatment (Lancet.1999;353:525-535). While the study did not support its main hypothesis—that treating STDs would reduce the rate of HIV transmission—all participants eventually received antibiotic therapy, significantly reducing the prevalence among them of syphilis and other diseases.

While answers to the "who should pay" question remain unclear, it is clear that with the weight of a revised Declaration of Helsinki behind them, developing nations now have the leverage to demand more consideration from medical researchers.

Whether or not they actually get it, despite the Declaration of Helsinki's good intentions, remains to be seen.

The revised declaration makes several other recommendations that affect publication of research. These points include the following:

Although the recommendation regarding the publication of research articles that do not follow the declaration's principles is not new, journal editors might have problems with it because of the placebo policy. According to Catherine DeAngelis, MD, editor of JAMA, "In some cases good science dictates studies with placebo arms. These might include new studies of populations different from those previously studied, including differences in age group (children vs adults), sex, or racial/ethnic origin."

The full text of the Declaration of Helsinki is available online at http://www.wma.net/e/policy/17-c_e.html.

New Orleans—Quick and aggressive treatment of coronary events enjoyed preference over conservative approaches as results from late-breaking trials were announced at the American Heart Association's annual scientific sessions last month.

Attendees in New Orleans heard data suggesting that use of the statin class of cholesterol-lowering drugs immediately after a myocardial infarction (MI) or unstable angina significantly reduces recurrent cardiac events and strokes. They also learned that patients with such conditions fared better if treated immediately with platelet blockers, aspirin, or heparin, and underwent early catherization and revascularization rather than conservative treatment with only medical therapy.

The Myocardial Ischemia Reduction with Aggressive Cholesterol Lowering (MIRACL) trial randomized 3086 people, 65% male and 85% white, with an average age of 65 years and total cholesterol level of 270 mg/dL, into two double-blinded groups. Within 4 days after the acute event (non–Q-wave MI or unstable angina), and continuing for 16 weeks, patients were given 80 mg of atorvastatin daily or placebo. Both groups were given dietary counseling.

The primary end points were death, cardiac arrest, MI, or worsening angina requiring emergency rehospitalization. The patients were enrolled between June 1997 and January 2000 and came from 122 hospitals in 19 countries; 33% of the patients were in North America. The trial was funded by Parke-Davis Pharmaceuticals, which manufactures Lipitor, a brand of atorvastatin.

The lead researchers, Anders G. Olsson, MD, PhD, of the University of Linköping, Sweden, and Gregory G. Schwartz, MD, PhD, of the Denver Veterans Affairs Medical Center in Colorado, found that at least one of the conditions defined as an end point occurred in 17.4% of patients in the placebo group, while 14.8% in the atorvastatin group experienced one of these end points—a 16% relative risk reduction. The researchers found a relative risk reduction for all four end points for patients receiving atorvastatin, with the greatest benefit, a 26% reduction vs the placebo group, accruing in those who had presented with chest pain. The researchers also found a "surprising" 50% reduction in stroke incidence in the atorvastatin group, although this was not a primary end point.

"We conclude from our study that patients with a threatened, or mild, heart attack benefit from immediate and intensive lipid-lowering therapy with atorvastatin, and that this treatment produces a significant reduction in both recurrent cardiac events and strokes," Schwartz said. "Furthermore, we believe that treatment with atorvastatin should be initiated in these patients in the hospital and treatment should be considered irrespective of the baseline cholesterol levels at the time of the acute event."

The lead researcher of another trial, Christopher P. Cannon, MD, of Brigham and Women's Hospital and Harvard Medical School, noted that 1.5 million patients in the United States are hospitalized each year with an MI or unstable angina. "The question faced is whether they should go on for a cardiac catherization and have angioplasty or whether they should have a wait-and-see approach to see how the medical therapy works," Cannon said. He thinks he has a conclusive answer.

The TIMI 18 trial (Treat Angina with Aggrastat and Determine the Cost of Therapy with Invasive or Conservative Strategies [TACTICS]) involved 2220 patients with MI or unstable angina from 169 hospitals in nine countries. Participants in the trial were given aspirin, heparin, β-blockers, cholesterol-lowering medication (as needed), and tirofiban, a glycoprotein IIb/IIIa inhibitor. All were randomized within 4 days of hospitalization to receive either an invasive procedure (angioplasty and stenting or a coronary artery bypass graft [CABG] operation) or medical therapy only. The end points were death, MI, or rehospitalization for worsening chest pains within 6 months. The study was funded by Merck & Co, which manufactures Aggrastat, the brand of tirofiban used in the trial.

Cannon described the results as "clear" and "significant." For the primary end points, 19.4% of patients in the conservatively treated group had died, had an MI, or were rehospitalized. In the group that had an invasive procedure, only 15.9% experienced an end point condition. If only death or MI were considered, 9.5% of patients who received only medical therapy reached either of these end points vs 6.3% in the intervention group.

"These results mean there will be a large change in the practice of medicine for those patients with worsening chest pain and unstable angina with a broader use of glycoprotein IIb/IIIa inhibitors and an early invasive catheterization strategy," Cannon said.

He made his statement in light of previous studies comparing invasive vs conservative strategies without use of glycoprotein IIB/IIIa inhibitors whose results were equal (Circulation.1994;89:1545-1556).

However, not all the news was good for tirofiban. In the TARGET trial (Do Tirofiban and ReoPro Give Similar Efficacy Outcomes?), researchers concluded that abciximab (ReoPro, Merck & Co), was better than tirofiban when used during angioplasty and stenting.

The trial, funded by Merck, enrolled 4800 patients from 150 hospitals in 18 countries who underwent percutaneous coronary intervention with intracoronary stent placement. To assess the 30-day efficacy of tirofiban and abciximab, patients were randomized in a double-blind, double-dummy fashion to receive one of the two drugs. The primary end points were death, MI, or emergency revascularization.

At the 30-day mark, 7.5% of the patients receiving tirofiban had died, experienced an MI, or required emergency revascularization, while only 6.1% who were receiving abciximab experienced those outcomes—a 20% relative risk ratio in favor of abciximab. The results for the individual components of the end points were similar, but were statistically significant only for MI (P = .04). Of the patients receiving tirofiban, 0.5% died vs 0.4% receiving abciximab; 6.9% receiving tirofiban had an MI vs 5.4% receiving abciximab; and 0.8% receiving tirofiban needed emergency revascularization vs 0.7% receiving abciximab.

"The effect was seen early," said Eric J. Topol, MD, of the Cleveland Clinic Foundation in Ohio. "It was manifest even by the first enzyme draw at 8 hours. As for MI, the data show that the larger the heart attack, the greater the effect of the abciximab advantage."

The trial had other implications, Topol said. Researchers have wondered if the differences in receptor affinity mattered in clinical outcomes. Abciximab binds not only to the glycoprotein IIb/IIIa receptor but also other receptors such as the vitronectin and Mac-1. Tirofiban binds specifically to the glycoprotein IIb/IIIa receptor. Also, there is a cost factor. The typical cost of an abciximab bolus and a 12-hour infusion for an 80-kg patient is $1350; for a 10-µg/kg bolus and 18- to 24-hour infusion with tirofiban, the cost is $689, according to a journal article (JAMA. 2000;284:1549-1558).

Editor's Note: Miscellanea Medica appears in the Medical News & Perspectives section occasionally. Items submitted for consideration should be directed to the attention of Marsha F. Goldsmith, Editor, Medical News & Perspectives.