Personalized medicine: The promise, the reality

Genetic tests to guide warfarin dosing could avert 85,000 serious bleeding events and 17,000 strokes annually, according to a report from the AEI-Brookings Joint Center for Regulatory Studies, a Washington, DC, think tank. The report further suggests that by integrating genetic testing into warfarin therapy, American health care spending could be reduced by $1.1 billion annually.¹ Unfortunately, the promise of using genetic testing to guide such pharmacological treatment has largely gone unfulfilled.²

Case in point: Genetic testing can tell us whether a patient is likely to be an ultra-rapid metabolizer of warfarin (and need larger doses) or a poor metabolizer (and need lower doses), but there are no guidelines to tell us how to dose accordingly. International normalized ratios (INRs) still need to be ordered and the patient will likely have to pick up the tab for the genetic test ($250), since Medicare and private insurers don’t cover the cost. (See “Warfarin: An ideal, but far from ready, candidate”)

Hints that change may be on the horizon. The government—specifically the Department of Health and Human Services—created the Secretary’s Advisory Committee on Genetics, Health, and Society (SACGHS) to assess how genetic and genomic technologies are being integrated into health care and to identify opportunities and gaps in research. To that end, SACGHS issued a draft report earlier this year that notes that genetic-based treatment has “the potential to yield significant gains in personal health, population health, and cost-effective resource allocation.” Among its many recommendations, SACGHS calls for greater collaboration between the public and private sectors to expand our knowledge of the clinical validity and utility of using genetics to guide treatment.³

A standard of care, potentially. Readying ourselves for the ways that genetics is likely to shape the way we prescribe such drugs as anticoagulants, antidepressants, and antiarrhythmics requires that we step back and assess the progress made so far, and the work that still needs to be done before genetic testing becomes a common occurrence, and perhaps even a standard of care.

The goal: Avert adverse events

The wide variation in the way different people respond to the same dose of medications is a major contributor to the problem of adverse drug reactions. Lazarou and colleagues estimated that 6.7% of hospitalized patients—over 2 million patients in the US—experienced an adverse drug reaction and 0.32% (106,000) had a fatal adverse drug reaction.⁴

Individual response to medications is determined by a host of factors including age, environment, other medications being taken, and genetic differences in drug absorption and metabolism. These genetic differences have spawned the fields of pharmacogenomics and pharmacogenetics.

Pharmacogenomics is the biotechnological science that combines the techniques of medicine, pharmacology, and genomics and is concerned with developing drug therapies to compensate for genetic differences in patients, which cause varied responses to a single therapeutic regimen.

• A good example of pharmacogenomics at work is the use of trastuzumab in addition to chemotherapy for breast cancer patients who are positive for the human epidermal growth factor receptor 2 (HER2) oncogene.⁵

Pharmacogenetics is the branch of pharmacology that examines the relation of genetic factors to variations in response to drugs.

The use of pharmacogenetics to predict individualized responses to medications and to prevent adverse drug reactions through individualized dosing regimens or avoidance of certain medications hinges on our knowledge of genetic polymorphisms, that is, gene-based differences in drug absorption, distribution, metabolism, or excretion.

Polymorphisms of the cytochrome P450 family of drug metabolizing enzymes have been the most extensively studied. The names of these enzymes are abbreviated by using CYP and then a series of letters and numbers to describe individual enzymes. The 4 most extensively studied CYP enzymes are CYP2A6, CYP2C9, CYP2C19, and CYP2D6. These 4 metabolize an estimated 40% of all drugs; the distributions of polymorphisms of each vary considerably by race/ethnicity.^6-8

A sampling of medications and classes of medications where polymorphisms play a significant role in drug metabolism is listed in the TABLE. Three with significant potential for prevention of adverse drug reactions are the antipsychotics, because of the severity of a specific drug adverse reaction, tardive dyskinesia;^6,9 warfarin, because of the risk of bleeding complications and its narrow therapeutic index;^7,10 and chemotherapeutic agents because of the serious nature of the disease and the potential for tailoring individualized therapies to maximize tumor response to medication and to minimize adverse reactions of very toxic drugs.¹¹

TABLE
Polymorphisms to these enzymes affect drug metabolism

Clinical resources reflect an information gap

In spite of the potential for improved patient care, there remains very little clinical application of pharmacogenetic information in primary care practice. Zineh and colleagues reviewed prescribing information in the electronic version of the Physicians’ Desk Reference (PDR) in 2004 and found that only 76 package inserts out of 3382 contained pharmacogenetic information.¹² In only 25 was there enough information to affect treatment decisions. Just 5 inserts mentioned that the chance of successful response to treatment could be predicted by genetic testing, and only one insert mentioned that a specific genetic subgroup should not take a drug.

The authors concluded that, generally, the pharmacogenetic information was inadequate to guide drug therapy and the majority of information was available for drugs that are not commonly prescribed. The FDA is addressing this issue by requiring the inclusion of pharmacogenetic information in package inserts more frequently.

Consider, too, the 2005 edition of Applied Therapeutics, a commonly used medical textbook.¹³ In it there is no mention of the use of pharmacogenetics in managing pharmacological therapies, which tells us that very little teaching on this topic is going on in medical schools and residencies.

But why? Why are clinicians and the tools we rely on so out of sync with the recommendations and expectations of personalized medicine advocates?

There are 5 likely culprits:

1. A lack of clinically useful pharmacogenetic tests.
2. A lack of test standardization and availability.
3. A lack of coverage by third-party payers.
4. A low level of physician knowledge about genetic testing.
5. A lack of evidence of improved outcomes.

How useful are pharmacogenetic tests?

For a laboratory test to be clinically useful, it should provide information that will influence a therapeutic decision. Decisions that could be influenced by this information include the dose of a particular drug and the potential use of an alternative because of a contraindication or likelihood of a poor outcome based on a particular genetic polymorphism.

The use of pharmacogenetic laboratory information can place a patient into one of several groups:

• ultra-rapid metabolizers, who need a larger dose of medication
• normal metabolizers, often called extensive metabolizers, who do not need dose modifications
• poor metabolizers, who need lower doses.⁹

Most medications have a wide therapeutic margin of effectiveness and safety. This means that the medication works within a wide range of serum drug levels and is safe at these different levels, making refinement in dosing based on genetic information unnecessary. In medications with narrower therapeutic windows for effectiveness or adverse reactions, there are frequently alternative means of drug level monitoring.

In many instances, the genetic test predictability of a patient’s actual metabolic responses—and resulting drug levels—is poor, leading to the need to monitor drug levels anyway.^14-16 This occurs because there is often a great deal of overlap in the response to a medication dose among the different metabolism classifications.^9,14,16-18 All of these realities have limited the clinical usefulness of pharmacogenetics up to this point.

Test standardization: Poised to improve?

Genotyping, the determination of the actual genetic makeup of the patient is not always predictive of an individual patient’s metabolic response. In other words, genotype does not always equate to a phenotype. Further complicating matters is the fact that many of the pharmacogenetic studies have been performed at research laboratories and have used tests that are not standardized, or widely available.^19-21

The recent commercial availability of pharmacogenetic tests by well-established and reputable laboratories will probably improve both standardization and availability.

An example is the AmpliChip CYP450 test by Roche Diagnostics.²² This microassay-based test identifies 29 CYP-2D6 polymorphisms and 2 CYP-2C19 polymorphisms. These genes affect the metabolism of 25% of currently prescribed drugs. Using this test, patients can be classified as poor, intermediate, extensive, or ultra rapid metabolizers of CYP-2D6 affected drugs, including antidepressants, antiarrhythmics, and antipsychotics. They can also be classified as poor or extensive metabolizers of CYP-2C19 affected drugs, including phenytoin and proton pump inhibitors.

How costly?

The cost of genetic testing will also affect availability. Tests will be widely available only if covered by third-party payers. The AmpliChip test costs between $300 and $500. Clearly, then, both cost and insurance coverage are issues that impede the adoption of pharmacogenetic testing, though there is little written about this in the medical literature.^19,23

AAFP explores ways to teach genomics

Physicians who are currently in practice received little or no training in the clinical use of pharmacogenetics or other genetic tests, such as genetic testing for the prediction of cancer risk.^24,25 The main resources of pharmacological information for practicing physicians do not contain much, if any, useful genomic information. A recent continuing education monograph for family physicians on clinical genetics mentioned pharmacogenetics only as a promising future technology.²⁶

For its part, the American Academy of Family Physicians has formed a genomics work group and is exploring how to educate family physicians on clinically useful genomic topics.

Evidence-based outcomes are needed

To date there has not been a head-to-head comparison of the outcomes of using clinical pharmacogenetics with those obtained from standard drug level monitoring practices. The CDC has formed a committee modeled after the USPSTF, the Evaluations of Genomics Applications in Practice and Prevention (EGAPP), which will evaluate the effectiveness of genomic clinical tests and make recommendations to physicians on their use.²⁷ The group’s first report, on the use of CYP450 testing in depression, concluded that there is a paucity of good quality data that addresses whether testing for CYP450 polymorphisms in adults entering SSRI treatment leads to improved outcomes.²⁸

In addition, SACGHS, the Department of Health and Human Services’ Committee, recommends in its draft report that HHS “provide resources to identify and address evidentiary gaps in the analytic validity, clinical validity, clinical utility, and cost effectiveness of pharmacogenomics.”³

Outcomes data will undoubtedly be key. With it, pharmacogenetic testing has the potential to grow by leaps and bounds—perhaps even becoming a standard of care in guiding pharmacological therapy. Without it, such testing will remain a promising, but as yet unrealized advance in personalized medicine.

Correspondence
Doug Campos-Outcalt, MD, MPA, 550 E. Van Buren, Phoenix, AZ 85004; dougco@u.arizona.edu.

Genetic tests to guide warfarin dosing could avert 85,000 serious bleeding events and 17,000 strokes annually, according to a report from the AEI-Brookings Joint Center for Regulatory Studies, a Washington, DC, think tank. The report further suggests that by integrating genetic testing into warfarin therapy, American health care spending could be reduced by $1.1 billion annually.¹ Unfortunately, the promise of using genetic testing to guide such pharmacological treatment has largely gone unfulfilled.²

Case in point: Genetic testing can tell us whether a patient is likely to be an ultra-rapid metabolizer of warfarin (and need larger doses) or a poor metabolizer (and need lower doses), but there are no guidelines to tell us how to dose accordingly. International normalized ratios (INRs) still need to be ordered and the patient will likely have to pick up the tab for the genetic test ($250), since Medicare and private insurers don’t cover the cost. (See “Warfarin: An ideal, but far from ready, candidate”)

Hints that change may be on the horizon. The government—specifically the Department of Health and Human Services—created the Secretary’s Advisory Committee on Genetics, Health, and Society (SACGHS) to assess how genetic and genomic technologies are being integrated into health care and to identify opportunities and gaps in research. To that end, SACGHS issued a draft report earlier this year that notes that genetic-based treatment has “the potential to yield significant gains in personal health, population health, and cost-effective resource allocation.” Among its many recommendations, SACGHS calls for greater collaboration between the public and private sectors to expand our knowledge of the clinical validity and utility of using genetics to guide treatment.³

A standard of care, potentially. Readying ourselves for the ways that genetics is likely to shape the way we prescribe such drugs as anticoagulants, antidepressants, and antiarrhythmics requires that we step back and assess the progress made so far, and the work that still needs to be done before genetic testing becomes a common occurrence, and perhaps even a standard of care.

The goal: Avert adverse events

The wide variation in the way different people respond to the same dose of medications is a major contributor to the problem of adverse drug reactions. Lazarou and colleagues estimated that 6.7% of hospitalized patients—over 2 million patients in the US—experienced an adverse drug reaction and 0.32% (106,000) had a fatal adverse drug reaction.⁴

Individual response to medications is determined by a host of factors including age, environment, other medications being taken, and genetic differences in drug absorption and metabolism. These genetic differences have spawned the fields of pharmacogenomics and pharmacogenetics.

Pharmacogenomics is the biotechnological science that combines the techniques of medicine, pharmacology, and genomics and is concerned with developing drug therapies to compensate for genetic differences in patients, which cause varied responses to a single therapeutic regimen.

• A good example of pharmacogenomics at work is the use of trastuzumab in addition to chemotherapy for breast cancer patients who are positive for the human epidermal growth factor receptor 2 (HER2) oncogene.⁵

Pharmacogenetics is the branch of pharmacology that examines the relation of genetic factors to variations in response to drugs.

The use of pharmacogenetics to predict individualized responses to medications and to prevent adverse drug reactions through individualized dosing regimens or avoidance of certain medications hinges on our knowledge of genetic polymorphisms, that is, gene-based differences in drug absorption, distribution, metabolism, or excretion.

Polymorphisms of the cytochrome P450 family of drug metabolizing enzymes have been the most extensively studied. The names of these enzymes are abbreviated by using CYP and then a series of letters and numbers to describe individual enzymes. The 4 most extensively studied CYP enzymes are CYP2A6, CYP2C9, CYP2C19, and CYP2D6. These 4 metabolize an estimated 40% of all drugs; the distributions of polymorphisms of each vary considerably by race/ethnicity.^6-8

A sampling of medications and classes of medications where polymorphisms play a significant role in drug metabolism is listed in the TABLE. Three with significant potential for prevention of adverse drug reactions are the antipsychotics, because of the severity of a specific drug adverse reaction, tardive dyskinesia;^6,9 warfarin, because of the risk of bleeding complications and its narrow therapeutic index;^7,10 and chemotherapeutic agents because of the serious nature of the disease and the potential for tailoring individualized therapies to maximize tumor response to medication and to minimize adverse reactions of very toxic drugs.¹¹

TABLE
Polymorphisms to these enzymes affect drug metabolism

Clinical resources reflect an information gap

In spite of the potential for improved patient care, there remains very little clinical application of pharmacogenetic information in primary care practice. Zineh and colleagues reviewed prescribing information in the electronic version of the Physicians’ Desk Reference (PDR) in 2004 and found that only 76 package inserts out of 3382 contained pharmacogenetic information.¹² In only 25 was there enough information to affect treatment decisions. Just 5 inserts mentioned that the chance of successful response to treatment could be predicted by genetic testing, and only one insert mentioned that a specific genetic subgroup should not take a drug.

The authors concluded that, generally, the pharmacogenetic information was inadequate to guide drug therapy and the majority of information was available for drugs that are not commonly prescribed. The FDA is addressing this issue by requiring the inclusion of pharmacogenetic information in package inserts more frequently.

Consider, too, the 2005 edition of Applied Therapeutics, a commonly used medical textbook.¹³ In it there is no mention of the use of pharmacogenetics in managing pharmacological therapies, which tells us that very little teaching on this topic is going on in medical schools and residencies.

But why? Why are clinicians and the tools we rely on so out of sync with the recommendations and expectations of personalized medicine advocates?

There are 5 likely culprits:

1. A lack of clinically useful pharmacogenetic tests.
2. A lack of test standardization and availability.
3. A lack of coverage by third-party payers.
4. A low level of physician knowledge about genetic testing.
5. A lack of evidence of improved outcomes.

How useful are pharmacogenetic tests?

For a laboratory test to be clinically useful, it should provide information that will influence a therapeutic decision. Decisions that could be influenced by this information include the dose of a particular drug and the potential use of an alternative because of a contraindication or likelihood of a poor outcome based on a particular genetic polymorphism.

The use of pharmacogenetic laboratory information can place a patient into one of several groups:

• ultra-rapid metabolizers, who need a larger dose of medication
• normal metabolizers, often called extensive metabolizers, who do not need dose modifications
• poor metabolizers, who need lower doses.⁹

Most medications have a wide therapeutic margin of effectiveness and safety. This means that the medication works within a wide range of serum drug levels and is safe at these different levels, making refinement in dosing based on genetic information unnecessary. In medications with narrower therapeutic windows for effectiveness or adverse reactions, there are frequently alternative means of drug level monitoring.

In many instances, the genetic test predictability of a patient’s actual metabolic responses—and resulting drug levels—is poor, leading to the need to monitor drug levels anyway.^14-16 This occurs because there is often a great deal of overlap in the response to a medication dose among the different metabolism classifications.^9,14,16-18 All of these realities have limited the clinical usefulness of pharmacogenetics up to this point.

Test standardization: Poised to improve?

Genotyping, the determination of the actual genetic makeup of the patient is not always predictive of an individual patient’s metabolic response. In other words, genotype does not always equate to a phenotype. Further complicating matters is the fact that many of the pharmacogenetic studies have been performed at research laboratories and have used tests that are not standardized, or widely available.^19-21

The recent commercial availability of pharmacogenetic tests by well-established and reputable laboratories will probably improve both standardization and availability.

An example is the AmpliChip CYP450 test by Roche Diagnostics.²² This microassay-based test identifies 29 CYP-2D6 polymorphisms and 2 CYP-2C19 polymorphisms. These genes affect the metabolism of 25% of currently prescribed drugs. Using this test, patients can be classified as poor, intermediate, extensive, or ultra rapid metabolizers of CYP-2D6 affected drugs, including antidepressants, antiarrhythmics, and antipsychotics. They can also be classified as poor or extensive metabolizers of CYP-2C19 affected drugs, including phenytoin and proton pump inhibitors.

How costly?

The cost of genetic testing will also affect availability. Tests will be widely available only if covered by third-party payers. The AmpliChip test costs between $300 and $500. Clearly, then, both cost and insurance coverage are issues that impede the adoption of pharmacogenetic testing, though there is little written about this in the medical literature.^19,23

AAFP explores ways to teach genomics

Physicians who are currently in practice received little or no training in the clinical use of pharmacogenetics or other genetic tests, such as genetic testing for the prediction of cancer risk.^24,25 The main resources of pharmacological information for practicing physicians do not contain much, if any, useful genomic information. A recent continuing education monograph for family physicians on clinical genetics mentioned pharmacogenetics only as a promising future technology.²⁶

For its part, the American Academy of Family Physicians has formed a genomics work group and is exploring how to educate family physicians on clinically useful genomic topics.

Evidence-based outcomes are needed

To date there has not been a head-to-head comparison of the outcomes of using clinical pharmacogenetics with those obtained from standard drug level monitoring practices. The CDC has formed a committee modeled after the USPSTF, the Evaluations of Genomics Applications in Practice and Prevention (EGAPP), which will evaluate the effectiveness of genomic clinical tests and make recommendations to physicians on their use.²⁷ The group’s first report, on the use of CYP450 testing in depression, concluded that there is a paucity of good quality data that addresses whether testing for CYP450 polymorphisms in adults entering SSRI treatment leads to improved outcomes.²⁸

In addition, SACGHS, the Department of Health and Human Services’ Committee, recommends in its draft report that HHS “provide resources to identify and address evidentiary gaps in the analytic validity, clinical validity, clinical utility, and cost effectiveness of pharmacogenomics.”³

Outcomes data will undoubtedly be key. With it, pharmacogenetic testing has the potential to grow by leaps and bounds—perhaps even becoming a standard of care in guiding pharmacological therapy. Without it, such testing will remain a promising, but as yet unrealized advance in personalized medicine.

Correspondence
Doug Campos-Outcalt, MD, MPA, 550 E. Van Buren, Phoenix, AZ 85004; dougco@u.arizona.edu.

Genetic tests to guide warfarin dosing could avert 85,000 serious bleeding events and 17,000 strokes annually, according to a report from the AEI-Brookings Joint Center for Regulatory Studies, a Washington, DC, think tank. The report further suggests that by integrating genetic testing into warfarin therapy, American health care spending could be reduced by $1.1 billion annually.¹ Unfortunately, the promise of using genetic testing to guide such pharmacological treatment has largely gone unfulfilled.²

Case in point: Genetic testing can tell us whether a patient is likely to be an ultra-rapid metabolizer of warfarin (and need larger doses) or a poor metabolizer (and need lower doses), but there are no guidelines to tell us how to dose accordingly. International normalized ratios (INRs) still need to be ordered and the patient will likely have to pick up the tab for the genetic test ($250), since Medicare and private insurers don’t cover the cost. (See “Warfarin: An ideal, but far from ready, candidate”)

Hints that change may be on the horizon. The government—specifically the Department of Health and Human Services—created the Secretary’s Advisory Committee on Genetics, Health, and Society (SACGHS) to assess how genetic and genomic technologies are being integrated into health care and to identify opportunities and gaps in research. To that end, SACGHS issued a draft report earlier this year that notes that genetic-based treatment has “the potential to yield significant gains in personal health, population health, and cost-effective resource allocation.” Among its many recommendations, SACGHS calls for greater collaboration between the public and private sectors to expand our knowledge of the clinical validity and utility of using genetics to guide treatment.³

A standard of care, potentially. Readying ourselves for the ways that genetics is likely to shape the way we prescribe such drugs as anticoagulants, antidepressants, and antiarrhythmics requires that we step back and assess the progress made so far, and the work that still needs to be done before genetic testing becomes a common occurrence, and perhaps even a standard of care.

The goal: Avert adverse events

The wide variation in the way different people respond to the same dose of medications is a major contributor to the problem of adverse drug reactions. Lazarou and colleagues estimated that 6.7% of hospitalized patients—over 2 million patients in the US—experienced an adverse drug reaction and 0.32% (106,000) had a fatal adverse drug reaction.⁴

Individual response to medications is determined by a host of factors including age, environment, other medications being taken, and genetic differences in drug absorption and metabolism. These genetic differences have spawned the fields of pharmacogenomics and pharmacogenetics.

Pharmacogenomics is the biotechnological science that combines the techniques of medicine, pharmacology, and genomics and is concerned with developing drug therapies to compensate for genetic differences in patients, which cause varied responses to a single therapeutic regimen.

• A good example of pharmacogenomics at work is the use of trastuzumab in addition to chemotherapy for breast cancer patients who are positive for the human epidermal growth factor receptor 2 (HER2) oncogene.⁵

Pharmacogenetics is the branch of pharmacology that examines the relation of genetic factors to variations in response to drugs.

The use of pharmacogenetics to predict individualized responses to medications and to prevent adverse drug reactions through individualized dosing regimens or avoidance of certain medications hinges on our knowledge of genetic polymorphisms, that is, gene-based differences in drug absorption, distribution, metabolism, or excretion.

Polymorphisms of the cytochrome P450 family of drug metabolizing enzymes have been the most extensively studied. The names of these enzymes are abbreviated by using CYP and then a series of letters and numbers to describe individual enzymes. The 4 most extensively studied CYP enzymes are CYP2A6, CYP2C9, CYP2C19, and CYP2D6. These 4 metabolize an estimated 40% of all drugs; the distributions of polymorphisms of each vary considerably by race/ethnicity.^6-8

A sampling of medications and classes of medications where polymorphisms play a significant role in drug metabolism is listed in the TABLE. Three with significant potential for prevention of adverse drug reactions are the antipsychotics, because of the severity of a specific drug adverse reaction, tardive dyskinesia;^6,9 warfarin, because of the risk of bleeding complications and its narrow therapeutic index;^7,10 and chemotherapeutic agents because of the serious nature of the disease and the potential for tailoring individualized therapies to maximize tumor response to medication and to minimize adverse reactions of very toxic drugs.¹¹

TABLE
Polymorphisms to these enzymes affect drug metabolism

Clinical resources reflect an information gap

In spite of the potential for improved patient care, there remains very little clinical application of pharmacogenetic information in primary care practice. Zineh and colleagues reviewed prescribing information in the electronic version of the Physicians’ Desk Reference (PDR) in 2004 and found that only 76 package inserts out of 3382 contained pharmacogenetic information.¹² In only 25 was there enough information to affect treatment decisions. Just 5 inserts mentioned that the chance of successful response to treatment could be predicted by genetic testing, and only one insert mentioned that a specific genetic subgroup should not take a drug.

The authors concluded that, generally, the pharmacogenetic information was inadequate to guide drug therapy and the majority of information was available for drugs that are not commonly prescribed. The FDA is addressing this issue by requiring the inclusion of pharmacogenetic information in package inserts more frequently.

Consider, too, the 2005 edition of Applied Therapeutics, a commonly used medical textbook.¹³ In it there is no mention of the use of pharmacogenetics in managing pharmacological therapies, which tells us that very little teaching on this topic is going on in medical schools and residencies.

But why? Why are clinicians and the tools we rely on so out of sync with the recommendations and expectations of personalized medicine advocates?

There are 5 likely culprits:

1. A lack of clinically useful pharmacogenetic tests.
2. A lack of test standardization and availability.
3. A lack of coverage by third-party payers.
4. A low level of physician knowledge about genetic testing.
5. A lack of evidence of improved outcomes.

How useful are pharmacogenetic tests?

For a laboratory test to be clinically useful, it should provide information that will influence a therapeutic decision. Decisions that could be influenced by this information include the dose of a particular drug and the potential use of an alternative because of a contraindication or likelihood of a poor outcome based on a particular genetic polymorphism.

The use of pharmacogenetic laboratory information can place a patient into one of several groups:

• ultra-rapid metabolizers, who need a larger dose of medication
• normal metabolizers, often called extensive metabolizers, who do not need dose modifications
• poor metabolizers, who need lower doses.⁹

Most medications have a wide therapeutic margin of effectiveness and safety. This means that the medication works within a wide range of serum drug levels and is safe at these different levels, making refinement in dosing based on genetic information unnecessary. In medications with narrower therapeutic windows for effectiveness or adverse reactions, there are frequently alternative means of drug level monitoring.

In many instances, the genetic test predictability of a patient’s actual metabolic responses—and resulting drug levels—is poor, leading to the need to monitor drug levels anyway.^14-16 This occurs because there is often a great deal of overlap in the response to a medication dose among the different metabolism classifications.^9,14,16-18 All of these realities have limited the clinical usefulness of pharmacogenetics up to this point.

Test standardization: Poised to improve?

Genotyping, the determination of the actual genetic makeup of the patient is not always predictive of an individual patient’s metabolic response. In other words, genotype does not always equate to a phenotype. Further complicating matters is the fact that many of the pharmacogenetic studies have been performed at research laboratories and have used tests that are not standardized, or widely available.^19-21

The recent commercial availability of pharmacogenetic tests by well-established and reputable laboratories will probably improve both standardization and availability.

An example is the AmpliChip CYP450 test by Roche Diagnostics.²² This microassay-based test identifies 29 CYP-2D6 polymorphisms and 2 CYP-2C19 polymorphisms. These genes affect the metabolism of 25% of currently prescribed drugs. Using this test, patients can be classified as poor, intermediate, extensive, or ultra rapid metabolizers of CYP-2D6 affected drugs, including antidepressants, antiarrhythmics, and antipsychotics. They can also be classified as poor or extensive metabolizers of CYP-2C19 affected drugs, including phenytoin and proton pump inhibitors.

How costly?

The cost of genetic testing will also affect availability. Tests will be widely available only if covered by third-party payers. The AmpliChip test costs between $300 and $500. Clearly, then, both cost and insurance coverage are issues that impede the adoption of pharmacogenetic testing, though there is little written about this in the medical literature.^19,23

AAFP explores ways to teach genomics

Physicians who are currently in practice received little or no training in the clinical use of pharmacogenetics or other genetic tests, such as genetic testing for the prediction of cancer risk.^24,25 The main resources of pharmacological information for practicing physicians do not contain much, if any, useful genomic information. A recent continuing education monograph for family physicians on clinical genetics mentioned pharmacogenetics only as a promising future technology.²⁶

For its part, the American Academy of Family Physicians has formed a genomics work group and is exploring how to educate family physicians on clinically useful genomic topics.

Evidence-based outcomes are needed

To date there has not been a head-to-head comparison of the outcomes of using clinical pharmacogenetics with those obtained from standard drug level monitoring practices. The CDC has formed a committee modeled after the USPSTF, the Evaluations of Genomics Applications in Practice and Prevention (EGAPP), which will evaluate the effectiveness of genomic clinical tests and make recommendations to physicians on their use.²⁷ The group’s first report, on the use of CYP450 testing in depression, concluded that there is a paucity of good quality data that addresses whether testing for CYP450 polymorphisms in adults entering SSRI treatment leads to improved outcomes.²⁸

In addition, SACGHS, the Department of Health and Human Services’ Committee, recommends in its draft report that HHS “provide resources to identify and address evidentiary gaps in the analytic validity, clinical validity, clinical utility, and cost effectiveness of pharmacogenomics.”³

Outcomes data will undoubtedly be key. With it, pharmacogenetic testing has the potential to grow by leaps and bounds—perhaps even becoming a standard of care in guiding pharmacological therapy. Without it, such testing will remain a promising, but as yet unrealized advance in personalized medicine.

Correspondence
Doug Campos-Outcalt, MD, MPA, 550 E. Van Buren, Phoenix, AZ 85004; dougco@u.arizona.edu.