The Promise of Precision Prescriptions – ‘Pharmacogenomics’ Also Raises Issues of Race, Privacy — (Washington Post)

To view original article click here

Washington Post

June 24, 2000

Author: Rick Weiss; Washington Post Staff Writer

Unlike most doctors, Georgetown University physician and pharmacologist David Flockhart enjoys talking about his patients who didn’t get better from the medicines he prescribed, or who got even worse because of side effects.

There was the woman who suffered excruciating pain after surgery no matter how much codeine he prescribed.

The patient whose tendons became so inflamed from taking a common antibiotic years ago that she remains debilitated to this day.

The woman who took Prozac only to find her blood pressure going through the roof.

Flockhart likes these sorry stories because they highlight the great potential of “pharmacogenomics,” a new field of medicine he and others are pioneering that promises to reduce treatment failures by matching medicines to patients’ personal genetic codes.

Among the many vaunted medical benefits promised by the Human Genome Project–the $2 billion effort to spell out the entire human genetic blueprint, whose virtual completion is expected to be announced on Monday–none is as close to broad clinical use as pharmacogenomics.

The new science seeks to solve a simple but long-standing problem: Medicines are made and sold on a “one size fits all” basis, even though people vary substantially in how they respond to those compounds. As a result of this variability, more than 100,000 Americans die every year from side effects of properly prescribed medicines and another 2 million are made seriously ill.

Now, with scientists identifying more and more genes that control individual responses to drugs, some leading-edge medical centers are starting to make prescribing decisions on the basis of patients’ genetic makeup. Five years from now, some experts predict, it will be common for patients to take genetic tests before their doctors decide which drug, or what dose, to prescribe for them. That’s far sooner than anyone expects to see similar success rates in the more widely touted but failure-plagued field of gene therapy, which aims to treat people by giving them new genes.

Drug companies, as well as patients, stand to gain from the pharmacogenomic revolution. By testing new and experimental drugs only in volunteers whose genes preordain a positive response, companies can generate data that will prove irresistible to the Food and Drug Administration, streamlining today’s long and costly drug approval process. Some companies also hope to profit from sales of the genetic tests that will be used to match patients with appropriate drugs.

But like so many advances in genetics, pharmacogenomics raises difficult questions, too. As the first attempt to meld the contentious field of genetic testing with the everyday practice of medicine, it will serve as a test case of how society will deal with issues of genetic privacy and discrimination.

Many people are already nervous about their genetic profiles falling into the wrong hands, for example, but so far few people have had a reason to take such tests. How will people react when a genetic test is required just to get a proper prescription?

Of equal concern, some of the gene patterns relevant to pharmacogenomics are ethnically linked, raising issues of racial stereotyping and access to care. One fear is that profit-conscious pharmaceutical companies will use pharmacogenomics to aim their drug development efforts toward genetic subgroups of people who can best afford to pay for them, further marginalizing already underserved minorities.

“What happens when the patient comes in and says, ‘I hear there’s a great new drug for asthma,’ and the doctor says . . . ‘Yeah, but it’s only for whites?’ ” asked Mark Rothstein of the University of Houston’s Health Law & Policy Institute.

Pharmacogenomics alters medicine’s legal landscape, too. At what point should doctors and drugmakers be liable for not taking genetic information into account? Already a group of patients has sued the maker of a Lyme disease vaccine, claiming that people with a particular genetic signature should not have been given the vaccine because they are especially prone to getting serious side effects from the shot.

All told, experts said, pharmacogenomics presents a challenge to society for precisely the same reason it offers so much promise: because it focuses on people’s differences.

“We believe strongly that this is a new era and the dawning of a golden age of personalized medicine,” said Elliott Sigal, a senior vice president at Bristol-Myers Squibb, one of several pharmaceutical companies that have launched pharmacogenomics programs. “It does raise important issues for society in terms of privacy, discrimination and insurance issues,” Sigal said. “I think the government will have to deal with these issues.”

Custom Treatment

Flockhart’s surgical patient presents a classic example of the need for pharmacogenomics. After undergoing surgery on her ovaries, she was given codeine, the most commonly prescribed painkiller in America.

“She wanted more and more of the stuff,” Flockhart recently recalled. Before long, he said, “she was labeled a drug abuser.”

The truth was far simpler. Tests ultimately showed she was among the 7 percent of Caucasians in this country who harbor an inactive form of a gene called CYP 2D6, which helps break down many common medicines, including codeine. Unable to metabolize the codeine into its desired breakdown product, morphine, the woman got no relief.

Flockhart’s patient on Prozac had the same genetic variation, but in her case it led to an effective overdose instead of undermedication. That’s because Prozac, like codeine, is also broken down with the help of the 2D6 gene. Since she could not metabolize the antidepressant, the woman suffered an overaccumulation of the drug, with high blood pressure and other side effects.

The patient with tendinitis carries a different gene variant that triggers painful inflammatory reactions in response to so-called quinalone antibiotics.

Scientists know of six common CYP drug-metabolizing genes, all of which operate in the liver and are responsible for breaking down about a dozen different drugs each. They have also begun to find other genes in other parts of the body, which people can inherit in various forms and which can affect how efficiently these people absorb, transport, use and excrete various medicines.

With the recent advent of simple tests that can reveal a person’s genetic subtype from a drop of blood or a cheek swab, Flockhart said, “we’re beautifully set up to look at variations in these things from person to person. It’s a real opportunity to help a lot of people.”

In a few settings, genetic testing is already helping doctors inform their prescribing decisions and even save lives.

At the Mayo Clinic in Rochester, Minn., for example, Richard M. Weinshilboum and his colleagues have been using genetic tests to individualize treatments for children with acute lymphoblastic leukemia.

The childhood cancer was universally fatal a few decades ago, but with the advent of a drug called 6-mercaptopurine, more than 80 percent of children can now be permanently cured. In the years since that drug’s introduction, though, two problems have plagued doctors: Why doesn’t the drug work in every child, and why does it cause serious and even fatal side effects in some?

The answers finally arose with the tools of pharmacogenomics.

Mercaptopurine, it turns out, is broken down in the body by an enzyme abbreviated TPMT. But 10 percent of Caucasians and blacks carry a variant of the TPMT gene that renders the crucial enzyme relatively ineffective, and an additional one in 300 of these children lack the gene and the enzyme altogether. (The variant is almost unknown in Asians.) Unable to metabolize the drug, these children essentially overdose on even small doses, and often die from the immune system suppression that ensues.

At the Mayo Clinic and at St. Jude Children’s Research Hospital in Memphis, doctors now routinely test for TPMT activity in leukemic children before treating them. Children with especially high levels of the enzyme–who are, in essence, breaking down the drug before it has a chance to kill their cancers–are given doses up to 50 percent higher than normal in order to get a therapeutic effect. And those who are effectively overdosing on the drug because of their relatively inactive TPMTgenes are getting doses as low as one-fifteenth normal, providing all the anti-cancer efficacy of a normal dose but without the extra side effects they once endured.

One result of the new focus on pharmacogenomics is that old definitions of diseases are breaking down. Cancers, for instance, which have traditionally been classified by their location in the body, are increasingly being typed further by their genetic characteristics and drug sensitivities, and medicines are being developed to take aim at those hallmarks. Experts predict that in the next few years a bevy of new genetic tests will show that many other diseases also deserve to be parsed and targeted according to their genetic signatures.

“We’re going to learn it’s not just ‘MS,’ ” or multiple sclerosis, said Kathleen Giacomini, a University of California at San Francisco researcher who is identifying genes that affect drug absorption and transport in the body. “It’s going to be MS a, b, c and d. And we can develop new drugs for each of these types.”

Philosophical Shift

The driving force behind the emergence of pharmacogenomics is a little-known follow-up project to the widely hailed Human Genome Project. Unlike the genome project, which sought to spell out a generic “consensus” version of the human genetic sequence, the follow-up project seeks to define the subtle differences in that code from person to person.

The goal is to identify hundreds of thousands of “single nucleotide polymorphisms,” or SNPs (pronounced “snips”), which are the tiny spelling variations that occur about once in every 500 to 1,000 letters within the 3 billion-letter human genetic code.

The recent shift by federally funded and privately financed gene researchers away from the genome project and over to the search for SNPs represents more than a scientific change of course. It brings with it a philosophical change of perspective that will focus not on people’s similarities but on their differences.

That shift is necessary if personalized gene-based medicine is to take off. But it also brings new possibilities for genetic bias and discrimination.

“The great promise of pharmacogenomics is that it will target drugs and therapies to individuals,” said Morris Foster, a University of Nebraska anthropologist who is studying the field’s potential impact on society. “But the way it may work out is it will just end up emphasizing social identifiers like race and ethnicity . . . and exacerbate social inequities.”

Not all of the genetic variations that affect how a person will respond to various drugs track along racial or ethnic lines. Indeed, doctors could easily run into trouble if they try to use race as a substitute for genetic testing: Think how many people might suppose that Tiger Woods is African American or Caribbean, one geneticist said, when in fact his mother is Thai.

But many relevant genes do correlate with race. While only 7 percent of Caucasians have the 2D6 variant that caused problems for some of Flockhart’s patients, for example, more than one in four Asians carry a similar variant.

Those realities present interesting economic and ethical quandaries.

On the economic side, pharmaceutical companies could save millions of dollars by preselecting participants in a clinical trial so that virtually everyone benefited and no one got side effects.

“Efficacy could be proven in small cohorts of a few hundred patients compared to 3,000 or 5,000 as we do now,” said Gualberto Ruano, chief executive officer of Genaissance Pharmaceuticals, a Science Park, Conn., company that is developing genetic tests to predict people’s responses to various drugs. “This is going to change the economics of drug development radically.”

It would also reverse the current federal regulatory push to include more diverse groupings of people in clinical trials–a push that arose a few years ago in response to concerns that minorities were being left out of the drug approval process. Under the new rubric of pharmacogenomics, though, expert reviewers might deem it unethical to test a new drug on people whose genes suggest they won’t respond well to the drug, since those participants’ sole purpose would be to demonstrate sideeffects.

At the same time, any drug that gained approval through such a process would have proven its mettle only in people with a narrow genetic or racial grouping, and would probably be approved by the FDA only for use in those people.

“That could be a problem,” said Larry Palmer, a professor at the Cornell University Law School who is studying the legal implications of pharmacogenomics. “You know the drug companies are going to market it as best they can. And the average practitioner may use it without the deep understanding of the science needed to use it appropriately on the right individuals.”

Moreover, given the chance to focus their efforts this way, would companies even bother developing drugs for people with rare genes? It’s too soon to say, but there is some evidence that some might not.

About five years ago, for example, pharmaceutical giant Merck & Co. was supporting a small Seattle company called Ostex International to develop a test that could identify women at high risk of osteoporosis. The hope, said Gilbert S. Omenn, executive vice president for medical affairs at the University of Michigan, was to use the test as a marketing tool for Merck’s new bone-building drug, Fosamax, which Merck hoped to sell widely to post-menopausal women.

But when early studies with the test indicated that only half of all postmenopausal women might actually need the drug, Omenn said, Merck pulled out of the partnership and left the little company in a state of near financial ruin, from which it is now recovering.

“Merck wanted to say that as soon as women get menopause they should take these drugs,” said Thomas O. Bologna, Ostex’s CEO. Rather than use an objective test to measure real need, Bologna said, “they decided to rely on salespeople to convince physicians to convince patients to use these drugs.”

A Merck spokeswoman said the company’s withdrawal from Ostex was not related to marketing concerns but came out of a decision to focus solely on the treatment of osteoporosis, not on methods for diagnosing it.

Others said that even if the big drug companies do end up focusing only on the most common genetic subpopulations, there are plenty of small biotechnology companies that would be happy to find niches that cover even 5 or 10 percent of the population.

Another hurdle that pharmacogenomics must clear is public wariness of genetic tests–especially among minorities, who have traditionally been more mistrustful of such tests.

Proponents say gene tests that predict a person’s response to a medicine should not be as controversial as other kinds of gene tests that predict, for example, a person’s long-term odds of getting a fatal disease.

“It’s not the same as telling someone that they’re going to get Alzheimer’s disease someday, which freaks people out,” said Kathleen Giacomini, the University of California at San Francisco researcher. “This is a more people-friendly kind of genetic test.”

A handful of companies are counting on that attitude, and are already developing kits that people will be able to use at home to see which versions of several key drug-metabolizing genes they’ve inherited. At least one company, PPGx Inc. near Research Triangle Park, N.C., hopes to offer at-home test kits by the end of this year, said the company’s CEO, Josh Baker.

If genetic tests prove as essential to the art and science of prescribing as proponents believe, it may come to pass that for some drugs, at least, anything less rigorous than individualized dosing will be considered malpractice.

Last winter, several patients brought a class action suit against SmithKline Beecham, the maker of a Lyme disease vaccine, claiming that the company should have warned people about evidence that the shot can cause a severe form of arthritis in certain people.

Federal records show that the FDA advisory committee that recommended the vaccine’s approval spent considerable time looking at preliminary evidence that people with a genetic hallmark called HLA-DR4 might be especially likely to experience the severe reaction.

Ultimately, however, the researchers who tested the vaccine for SmithKline convinced the officials that the data were too weak to believe. Now, in what may be the first intersection of pharmacogenomics and the law, they will have to defend that conclusion in court.


Scientists are starting to identify the genes that vary from person to person and that influence how different people respond to various medicines. They are developing genetic tests that will tell doctors which people should get which drugs, to maximize drug effectiveness and minimize side effects in each patient.

Activation: Some people have genes that are more responsive to some medications than others, explaining in part why some antidepressants, for example, work better in some people than in others.

Transport: Some people have genes that make blood-borne proteins that bind tightly to medicines, making them unavailable where they’re needed.

Metabolism: Some people have genes that break drugs down very quickly or unusually slowly, which can complicate efforts to achieve proper dosing.

Elimination: Some people have genes that speed up or slow the elimination of drugs from the blood, which can affect dosing schedules or the choice of drug.

Absorption: Some people have genes that result in inefficient absorption into the blood, so they may need higher doses or more easily absorbed drugs.

Medicine and Race

Race influences which people are genetically predisposed to lack various enzymes needed to break down medications. Without those enzymes, the medication can have either a heightened or lessened effect.

Enzyme: 2D6

Note: Also reduced in 5% of Caucasians, 12% of blacks and 25% of Asians.

Absent in: 7% Caucasians

2% Blacks

1% Asians

Some drugs affected: Codeine

Metoprolol (Lopressor)

Paroxetine (Paxil)

Enzyme: 2C19

Absent in: 3% Caucasians

3% Blacks

20-25% Asians

Some drugs affected: Diazepam (Valium)

Phenytoin (Dilantin)

Nelfinavir (Viracept)

Enzyme: 2C9

Absent in: 2% Caucasians

2% Blacks

<1% Asians

Some drugs affected: Warfarin (Coumadin)

Ibuprofen (Motrin)

Naproxen (Naprosyn)

SOURCE: David Flockhart, Georgetown University; staff reporting

About This Series

Scientists are on the brink of deciphering the entire human genetic blueprint. The accomplishment will launch a new era in biology and medicine. The long-term ramifications are profound: new medical treatments; a fundamental new understanding of the human organism; novel ways to diagnose disease, before birth or even before conception; the potential to manipulate traits such as looks or intelligence. In many ways, society is not yet prepared for the full implications. The Washington Post will continue to explore these issues in depth in a series of occasional reports over the next several months.
Copyright 2000 The Washington Post
Record Number:  062400LA01TH1402