|
Pharmacogenetic Challenges For The Health Care System
Pharmacogenetics—the effect of genotype on drug response—holds the promise of safer and more effective drug therapy. Genetic tests would be routinely given to patients prior to prescription of a drug, with therapeutic decisions based on the patient’s drug-response profile. This paper examines the operational changes and the ethical, legal, and policy challenges that pharmacogenetic medicine poses for key actors in the health care system. Adaptation by drug companies, regulatory agencies, physicians, patients, insurers, and public funding agencies will be necessary to integrate pharmacogenetic medicine into health care.
A major problem for health care systems is providing safe and effective drugs to persons who need them. Effective drugs are difficult to develop, vary in their effect on individuals, and are increasingly costly, making fair access for all patients difficult.1 The prospect that genetic science may soon be able to correlate drug response with individual genotype is thus welcome news for patients, providers, and insurers. Where genetic variations in drug response are established, patients would have their drug-response genotype determined before prescription of a drug, so that drugs that would be unsafe or ineffective for them would not be prescribed. Although this practice might raise the cost of individual prescriptions, it could lower overall health costs by reducing the prescription of drugs that are ineffective or harmful to particular individuals. Although this scenario appears optimistic, it is likely that great attention will be paid in coming years to the genetics of drug response. Genomic and genetic data will increasingly become part of drug development and use patterns.2 A shift in the health care system toward genetically based prescribing presents challenges for all affected parties. This paper describes the field of pharmacogenetics and obstacles to its development and then analyzes the issues raised by pharmacogenetic medicine for drug companies, regulatory agencies, physicians, patients, insurers, and public funding agencies.
Although variability in clinical response to a therapeutic intervention may result from many factors, there is increasing evidence that genetic factors cause the differences in drug responsiveness. Pharmacogenetics (PGx) is the study of those differences.3 For example, because of genetic polymorphisms, some persons may not produce the receptors, proteins, or enzymes needed for drugs to enter cells, be metabolized, and excreted in expected ways. As a result, some persons may experience toxicity at dosage levels appropriate for other patients, or process drugs too quickly to have the intended therapeutic effect. There is growing evidence that PGx differences exist widely in the population for many kinds of drugs. Polymorphisms in genes encoding drug receptors, cellular transporters, or metabolizing enzymes appear to affect patients’ responses to a wide range of drugs, including therapies such as pravastatin to lower cholesterol; procainamide and beta-blockers for heart disease; bronchodilators for asthma; mercaptopurine, azathioprine, and fluorouracil for cancers; levodopa, dopamine, and selective serotonin reuptake inhibitors (SSRIs) for mental disorders; and anticoagulants for stroke.4 Genetic variation also may affect the ability to process the metabolites in tobacco and alcohol.5 Potential benefits of PGx research. Understanding the links between genotypic variations and drug response phenotypes has several potential benefits.6 It could lead to more rational drug design and testing, with drugs targeted more precisely to individual variations in cell receptor and transport systems and tested more efficiently in more focused clinical trials. Drug use patterns also could be rationalized, so that medications are prescribed according to an individual’s genotype.7 If a patient has a genetic variation that prevents the drug from being properly metabolized, use of the drug is likely to be ineffective or harmful and would waste health care resources.8 Individualized drug prescriptions would also aid postmarket surveillance of approved drugs. For example, allowing PGx information to be collected in the postmarketing phase would prevent the need to recall drugs that are highly toxic for some patients but benefit many others. A PGx test for responsiveness to alosetron, the first drug approved for irritable bowel syndrome, might have prevented the need to recall the drug because of serious adverse reactions in a small group of patients.9 Scientific and social barriers. Realizing the potential benefits of PGx, however, will require overcoming both scientific and social barriers. Scientific barriers include the lack of broader and deeper knowledge of which genotypic variations affect drug response and why. Because drug responsiveness can vary with environment, diet, health status, and interactions among drugs, teasing out the genetic contribution may be difficult. The development of high-throughput methods for comparing bioinformatic data about genotypes will help to identify populationwide polymorphisms that affect drug responsiveness and to locate their presence in individuals.10 Whether PGx will play a prominent role in clinical medicine will depend partly upon whether there is enough genetic differentiation to make PGx testing and drug development worthwhile, yet not so many subgroup variations that an impractically large number of PGx tests—and a correspondingly large number of genotype-tailored drugs—would be necessary. The pace of integration of PGx into medicine also will depend on medical and social acceptance of a greatly increased use of genetic testing in the health care system. Actors at every level in the system will have to learn to approach many diseases and drugs in PGx terms. Addressing those challenges now will speed the assimilation of pharmacogenetic medicine into medical practice and health care.
At the heart of PGx medicine is the principle that genotypic variations among individuals cause or correlate with their response to drugs and that such variations are frequent enough to justify testing for them before prescribing a drug. If this principle is established, then PGx tests might routinely be given before drugs are prescribed, so that prescribing decisions may take account of test results. On this model, PGx tests would play a gatekeeper role in determining whether a patient would have access to certain drugs at particular dosage levels, such as pravastatin for coronary heart disease or thiopurine methyltransferase for childhood leukemia. As research confirms genotypic differences in responsiveness to drugs, a major issue will be how much clinical weight such PGx tests should carry in clinical decision making. The answer will depend upon the extent of genetic differentiation in drug response, the analytic and clinical validity of tests used to identify those responses, and other aspects of the patient’s situation. All clinical tests vary in their sensitivity and specificity. Given variation in the frequency of certain genotypes in a population, the positive predictive value of the PGx analysis will also need to be established—that is, the likelihood that a person with a positive test result will have the phenotypic condition of concern: in this case, whether a person will be able to benefit from a drug without unacceptable toxicity. Because PGx tests will vary in their sensitivity, specificity, and predictive power, a modified rather than strict gatekeeper model may be more appropriate for much of PGx medicine. Under this model a PGx test may, depending on its positive predictive value and other aspects of the patient’s situation, be strongly indicative of whether or not a drug should be prescribed but should not be the determining factor, trumping other factors that affect clinical judgment. Under either version of the gatekeeper model, it is essential that PGx tests have validated predictive capabilities, so that patients, physicians, and insurers can reasonably rely on them as relevant indicators for clinical decision making. It is also essential that clinicians understand test limitations and use good clinical judgment in making use of them.
The prospect of PGx medicine depends to a large extent on the willingness of public and private investors to support the research and product development necessary to bring PGx testing into medical practice. PGx research is now occurring in many areas, with both private and public funding, and is of great interest to the pharmaceutical industry. Impact on drug revenue. The initial impression has been that PGx testing would be highly advantageous to drug companies, with some firms already advertising its promise.11 In fact, the challenges presented for the pharmaceutical industry are much more complex.12 In the long run, a well-developed PGx medicine could undermine the current focus on blockbuster drugs that bring in large annual revenues for the life of a drug patent. Although this pattern is likely to continue for some time, and segmented markets already exist for many drugs, the emergence of large variation in the genetics of drug response could divide demand for pharmaceuticals into increasingly smaller subgroups, each with its own unique PGx needs. Prescribing drugs according to genotype could reduce demand for particular drugs, thus lessening the stream of revenue and the profits available to invest in research and development (R&D). Unless more new drugs coming on the market offset that reduction, PGx medicine could increase market differentiation, reduce profits, and perhaps affect change in investment patterns and the scope of industry-funded research. If genotypic variation in drug responsiveness is relatively small for particular drugs—for example, affecting only 10–20 percent of the persons who otherwise would be eligible for the drug—the use of PGx tests might not reduce demand for that drug enough to affect overall R&D or investment patterns. Indeed, identifying small populations of patients for whom drugs are unsafe or ineffective could in the long run be beneficial for firms. Although it would reduce the sales of certain drugs, it might permit others to remain on the market by screening out persons likely to have adverse reactions that could lead to postmarketing recall or litigation. Incentives to develop PGx tests. A drug company’s interest in selling drugs may also conflict with patients,’ providers,’ and insurers’ interests in having PGx tests developed and performed to prevent overuse of a drug. Most drug makers are involved in postmarket research on drug safety and efficacy. However, the maker of an already approved drug may have few incentives to develop a genotypic test to determine which patients should receive the drug and might even discourage others from developing such tests.13 On the other hand, incentives to develop and use PGx tests could exist if doing so would provide a marketing advantage over competing drugs. Skillful attention to PGx factors might enable drug companies to get drugs on the market more quickly. Also, if they can show physicians that there are safety and efficacy advantages from administering a drug response test before prescribing the drug, they may capture greater market share than they would if the market were larger but there were no PGx competition. PGx testing might also "rescue" drugs that failed clinical trials because of overall lack of safety or efficacy, or drugs that are threatened with market withdrawal because of a large number of postmarketing adverse events. Using genotypic drug response tests for these purposes would be a great advantage because of the costs already sunk into the development process. Bundling the test with the drug might also increase profits.14 New drug development. Finally, the genetics of drug responsiveness might help drug companies to identify the genetic profile of those patients not responding favorably to treatment, which could help to fuel the new drug development pipeline and meet unmet medical needs.15 With smaller and more efficient clinical trials, the drug approval process may be quicker, thus producing more years of sale under a drug’s patent.16 Drug companies are now gathering genotypic data from subjects in Phase II and Phase III clinical trials so that they can better identify factors affecting safety and efficacy. This will enable some drugs to be approved for patients with particular genotypes, while preventing their prescription for patients who are likely to suffer untoward reactions. Genotypic information may also protect against legal liability by allowing more accurate labeling of drug contraindications, warnings, and precautions.
Given the potential gatekeeping functions of PGx testing, it is essential that the sensitivity, specificity, and positive predictive value of PGx tests be reliably established and accurately communicated to clinicians. In the past some predictive genetic tests were introduced without sufficient attention to their reliability or predictive power or to accurate communication of those facts to clinicians, thus risking mistakes and poor outcomes for patients.17 With many more predictive genetic tests likely to be introduced in the future, including PGx tests, there is a strong need for some regulatory oversight of the validity and utility of genetic tests and laboratories’ performance of them.18 Regulation of PGx tests. The Food and Drug Administration (FDA) has primary responsibility for regulating PGx tests to ensure their validity and utility. PGx tests sold as kits are medical diagnostics, which require FDA premarket review.19 Test kits also may be used as investigational devices. To date, however, most genetic tests have been sold as "clinical services" by the academic centers and laboratories that have developed "home brews"—chemicals and reagents that a lab develops on its own—to test DNA samples sent to them. The FDA has been less diligent with policing "home brews" than test kits, although it now requires the reagents used in "home brews" to be registered.20 Federal standards for laboratory proficiency also would apply to any laboratory providing test results for patients.21 Using PGx test results on drug labels. A second important issue for the FDA is what the label of approved drugs should say about the advisability of undergoing PGx tests or using PGx test results in prescribing a drug. With DNA information now routinely being collected in many Phase II and Phase III trials, drug makers will increasingly be able to make claims of about the safety and efficacy of drugs based on the genotypes of subjects. If their data show that many more subjects with drug response genotype A improved without suffering adverse events than did those with genotype B, a drug maker may establish that a drug is safe and effective for As but not for Bs. This information could affect regulatory decisions—for example, whether the FDA approves the drug or whether the FDA requires that the drug’s label contain information about inappropriate genotypes for receiving the drug. An important decision in the regulatory process will be how the drug label conveys genotypic information about drug response—for example, whether the need to have a PGx test to determine whether a patient has a particular drug-response genotype should be listed as a contraindication, a warning, or a precaution to prescribing a drug. Depending on the drug and available data, a drug company might wish that the need for a PGx test be included as a precaution, rather than as a contraindication or a warning, although concerns about legal liability could push the other way. For example, with Accutane (isotretinoin) for treatment of acne in women of childbearing age, the maker agreed with the FDA to label use in that population as contraindicated unless patients first had two negative pregnancy tests and agreed to mandatory contraception to prevent birth defects in fetuses.22 As the predictive power and reliability of PGx tests increases, it is likely that a drug’s labeling will contain information about what PGx tests are desirable before a drug is prescribed. The FDA also may need more staff or resources to deal with the regulatory issues presented.
The prospect of PGx-based medicine presents a major challenge for physicians and for medical education. If drug prescriptions are increasingly based on the drug-response genotype of patients, physicians will have to learn how genetics affects drug response, know what genotypic tests to order and when to do so, and then be skilled in using test results in making clinical decisions. A major obstacle at present is that physicians generally receive little education in genetics. Many practicing physicians have not had a single hour of instruction in genetics in their formal training.23 Even those who are aware of predictive genetic testing, such as testing for inherited forms of cancer, make errors in advising patients about the significance of results.24 As PGx testing and knowledge become an important part of clinical practice, both undergraduate, postgraduate, and continuing medical educators will have to devote more time to their use in clinical practice. Physicians and other providers also may need access to clinical decision-support tools and the training to use them. To avoid errors, it is essential that PGx tests be easy to use and interpret, but not all tests might meet that standard. The key questions for physicians are (1) when to order a PGx test, and (2) how to use the results in clinical decision making. When to test. The answer to the first question will depend on the availability of PGx tests for the patient’s situation. Physicians should be knowledgeable about whether genotypic tests are available that could affect the safety or efficacy of drugs that they might prescribe. In some cases, tests for general drug responsiveness—such as a mutation affecting cellular transport for any drug—should be ordered. In other cases, the need for a PGx test might be specific to a drug, as indicated on the drug’s label. If the risk of side effects from a drug is great, a physician who failed to order a PGx test indicative of toxicity could be found to have been negligent if a proper test would have prevented that outcome.25 In still other cases, the predictive value of the test may be so low, or other factors present, that a decision to order the test is optional.
How to use.
The answer to the second question will depend on the patient’s condition, therapeutic alternatives, and the positive predictive value of the test. Physicians may not always need to know the precise way in which the genotypic variation affects drug responsiveness—for example, whether it impairs a cell-surface receptor or an intracellular transport mechanism—but they should understand the clinical significance of test results, so that they may prescribe appropriately. In some cases, the test will have sufficient predictive power and relevance to operate as a strict gatekeeper, with the prescribing decision largely determined by it (for example, that tricyclic antidepressants not be given to a particular patient or given only in a certain dosage). In other cases, even highly predictive test results will have to balanced against other clinical facts to determine whether or not use of a particular drug is wise, as now occurs with the use of tacrine for Alzheimer’s patients with the APO As PGx knowledge becomes more firmly established, decisions about when and whom to test and how to use the results will enter into conventional medical practice and, indeed, come to define the standard of care. In the future, lawsuits for failure to give a test or properly interpret it will occur, both reflecting and reinforcing acceptance of PGx testing as an ordinary part of medical practice.27 An important issue overall will be whether a strong or modified gatekeeper approach develops for the clinical use of most PGx tests. A strong gatekeeper model is more easily administered but may insufficiently acknowledge the probabilistic nature of most PGx correlations and the chance that any patient with that genotype might still benefit from a drug. However, because PGx tests will vary in their predictive power and patients vary in their disease, health status, other drugs, and environment, a modified gatekeeping function might be more appropriate for most PGx tests. In rare cases, prescribing a drug against warnings or contraindications on the label might be clinically sound, if the patient has no other therapeutic alternatives and is willing to run the indicated risk. However, health and drug benefit insurers, with their population rather than individual patient perspective, might choose not to pay for drugs that "fail" the PGx test indicated for prescription of that drug.
The chief challenge for patients is that they need to become more aware of PGx medicine and more accepting of how it works. Patients will need to know that the safety or efficacy of drugs may depend on their genotype and thus that testing for those variations prior to prescribing may be necessary for proper clinical decision making. They must also understand that certain drugs may not be prescribed if they have a particular genotype. For some people these changes will require a changed attitude toward "genetics" and "genetic tests," so that they become less frightened of things "genetic" and come to accept knowledge of genotypic variation as a standard part of medical practice.28 Increased public awareness that many existing clinical practices actually employ genetic knowledge, such as family history of heart disease and blood type, should facilitate the acceptance of the routine use of genotypic information in medical care. Privacy concerns. Patients’ acceptance of PGx testing will be more easily achieved if their rights to consent and privacy are fully protected. A key policy question will be how much information to provide patients in requesting DNA samples for PGx testing prior to prescription of a drug. Generally speaking, PGx tests carry lower psychosocial risk than do genetic tests that confirm the diagnosis of or predict genetic diseases or that identify carrier status for genetic diseases. In many cases, a PGx test, like many diagnostic or clinical lab tests that are now routinely administered with at most minimal informed consent, will carry no major risk of psychosocial harm. In some cases, however, PGx tests might carry secondary information about the patient’s genetic or health status that should not be communicated to others without the patient’s consent. For example, a mutation in a gene that indicates the absence of an enzyme essential for processing one drug may indicate that the patient is unlikely to respond to other drugs, as is the case with cytachrome P450 gene mutations.29 Information that a patient may not be easily treated for a range of diseases would be of interest to employers or insurers who want to minimize health costs. Depending upon the nature of the secondary information that is conveyed by the PGx test result and the social, economic, and cultural context in which the test occurs, some PGx tests may carry great risk of psychosocial harm. Consent procedures. Consent procedures for PGx testing should be commensurate with the risk entailed by the test in question.30 In most instances, a simple statement will suffice that the physician wishes "to test a sample of the patient’s DNA to see if a drug will be safe or whether it will work" and assurance that that is the only use to which the sample will be put. In cases where the PGx test also conveys sensitive secondary information, a more extensive informed consent discussion may be appropriate. Practice guidelines may be needed to help clinicians distinguish between "low-risk" PGx tests that require only minimal informed consent and "higher-risk" PGx tests that require fuller informed consent. The process of developing such guidelines should be a cooperative effort, with input from test producers, laboratories, professional organizations, regulatory agencies, and patient groups.31 Secondary information. Patients’ acceptance of PGx testing also will depend on their confidence that DNA samples and test results will be protected from unauthorized disclosure. There is considerable public concern about genetic privacy, in part because no comprehensive system exists for safeguarding the confidentiality of genetic information or guarding against its discriminatory use in employment and insurance decisions. The greater use of "firewalls" between how information is handled and stored and its potential users can reduce the risk of such breaches of confidentiality and reassure the public. The risk of sensitive secondary information is related to the choice of genetic markers used in PGx tests. Research should identify markers that contain the least secondary information, and PGx tests employing these markers should be preferred. In cases where PGx tests will inevitably generate secondary information, trusted third-party intermediaries could hold genotypic information in secure online databases and release it only in certain well-defined circumstances.32 Greater legal protection. Finally, greater legal protection for unauthorized disclosures of all medical information, including genotypic information, may be needed to instill public confidence that PGx information will remain private. The privacy regulations now in effect under the Health Insurance Portability and Accountability Act (HIPAA) of 1996 are an important step forward, but state laws protecting the privacy of DNA samples and test results may also be needed.33
The benefits of PGx medicine will not be realized unless public and private health insurers, managed care organizations, and pharmacy benefit managers recognize their importance and fund both PGx tests and prescriptions based on them. If the principles underlying PGx testing are sound, and testing and drug products exist for implementing them, both public and private health benefit providers should warmly welcome PGx testing. Testing offers the potential for increasing the efficacy and reducing the harm of covered drugs.34 There should be little resistance to covering PGx tests when requested by a physician or to covering a drug that the clinician thinks is justified by the test. Indeed, just as benefit managers work to get doctors to prescribe generics when patents on drugs expire, they will have a strong interest in persuading doctors to use PGx tests where indicated, despite the cost of the genetic test also required.35 A potential ethical issue is whether public or private drug benefit providers may or should condition insurance coverage on the patient’s first having a PGx test to determine whether the drug is likely to be safe or effective.36 Such a requirement would function as a precertification requirement for coverage of certain drugs, such as those that are very expensive or are known to have variable efficacy or serious toxicity for many persons. The case for PGx precertification rests on the strength of PGx correlations (the test’s reliability and predictive value) and the disease and patient at issue. FDA labeling information could be a helpful guide for determining when such tests may be required. Insurers clearly have an interest in valid and reliable PGx tests and should support regulatory efforts to establish their reliability. In addition to requiring that PGx tests be taken, some benefit programs may adopt a strict gatekeeping model and require that the patient "pass" the test for coverage to occur (that is, that the test show that the patient has the appropriate genotype to benefit from the drug without a high risk of a serious adverse effect). In many cases, this will be a rational course of action, for it is not reasonable to prescribe drugs that are not clinically indicated, and few physicians would order them. At the same time, it is important to recognize that PGx tests may not always function well as strict gatekeepers and that a modified gatekeeping role might be more appropriate. Like most clinical lab tests, PGx tests will vary in their reliability and predictive value and will apply to patients in highly variable circumstances. As the predictive power of the test weakens and viable therapeutic alternatives shrink, some physicians may judge that the drug, despite the PGx test result, is the best clinical alternative for the patient. These cases are likely to be rare, particularly if the risk of adverse reactions is great. Drug benefit programs should be prepared for this contingency and provide adequate review procedures for decisions that deny drug coverage on PGx grounds.37
The logic underlying PGx medicine may lead to some patients’ no longer receiving the drugs they would have received prior to PGx testing. If PGx tests and prescribing decisions are based on evidence, they will not deny patients drugs that would have been helpful. Patients do not have a right to receive a drug that reasonably appears unlikely to help or that might harm them, nor do insurers ordinarily have an obligation to fund such drugs. The challenge posed for public funding agencies is whether there is a special duty to support research to find other drugs that will work for this new class of therapeutic orphans. The challenge is all the greater if drug responsiveness also correlates with race or ethnicity, as occurs with some beta-blockers, or else groups that have traditionally received inferior medical care will be excluded from suitable drug therapy as well.38 What should be done for patients whom PGx testing has identified as inappropriate candidates for drugs that they would have received without the testing? In some cases, other therapies or drugs may be available, each with their own array of costs and benefits. In other cases, however, no alternative may exist. One could alleviate some of this problem by adopting a more modified gatekeeper role for PGx testing, particularly as the predictive value of the test weakens. Even then, however, there will be some patients who are not effectively treated. Ensure a test’s accuracy. Note that most of these patients have only apparently been made worse off by PGx medicine. It is true that they will not now receive a drug that they otherwise would have received, but if the principles of PGx testing are correct and the exclusion is evidence-based, most of these patients would not have benefited from the drug, and some of them may actually have been hurt. Only if the test’s predictive power were weak, and they otherwise would have benefited from inclusion, would such patients have been injured by the advent of PGx medicine. Ensuring the accuracy of tests thus is essential to make sure that persons denied treatment as a result would not have benefited. Some such cases, however, will always occur because even the best tests will not be 100 percent accurate. Subsidize new research. Persons for whom there is no effective therapy because of their drug-response genotype present a problem different from the conventional problem of providing health care to the uninsured. The problem for these "therapeutic orphans" is not the lack of health insurance coverage but rather the absence of safe and effective drugs for their condition. Meeting their needs may require research investment directed to their genotypically unique problems. Some PGx-excluded groups may be large enough to induce private firms to develop drugs specifically for their genotype—for example, African American cardiac patients for whom standard beta-blockers do not work.39 In most cases, however, the group may be too small for the market alone to produce solutions, and public or private subsidies may be needed to induce the necessary R&D. Given the competition among disease groups for research funding, it may be hard for smaller, genotypically segmented groups to lobby effectively for direct research subsidies. If the population of such groups is less than 200,000, they would still qualify for the tax credits, clinical testing subsidies, and seven-year sales monopolies now provided under the 1982 federal Orphan Drug Act.40 As the field of PGx medicine develops, it may become necessary to extend or create incentives to promote the research necessary to meet the health care needs of those persons.41 Whether those incentives take the form of subsidies for research into particular diseases or expansion of the current Orphan Drug Law will have to await further developments. The prospect of genotypically based prescriptions is only one of the ways that growing knowledge of genome sequence and function will in coming years affect medical practice.42 Genetics also will be increasingly used for molecular staging of disease and for more rational drug design. In addition, genetic screening will occur to predict late-onset diseases or to prevent susceptibility conditions, and many prospective parents will use genetic information to ensure the health of their offspring. Safe and effective gene therapies are also likely to become available in the next two decades. Successful handling of the challenges presented by pharmacogenetics will help pave the way to the gene-based medicine of the future.
John Robertson holds the Vinson and Elkins Chair at the University of Texas School of Law at Austin. Baruch Brody is the Leon Jaworski Professor of Biomedical Ethics and director of the Center for Ethics at Baylor College of Medicine in Houston. Allen Buchman is professor of law and philosophy at the University of Arizona in Tucson. Jeffrey Kahn is director of the Center for Bioethics and professor in the Department of Medicine at the University of Minnesota School of Medicine. Elizabeth McPherson is the genetics ethics advisor at GlaxoSmithKline in Research Triangle Park, North Carolina. This paper was supported by the Pharmacogenetics Consortium, funded by an unrestricted grant from GlaxoSmithKline, First Genetic Trust, and IBM. The authors also acknowledge helpful comments from Ned McColluch of IBM.
This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||
Top
Pharmacogenetics: Prospects And...
The Logic Of PGx...
4 allele.
