], a drug used to prevent breast cancer recurrence. Here’s how. To work, tamoxifen must first be converted to its active form—endoxifen—by CYP2D6. Women with an inherited deficiency in the CYP2D6 gene cannot activate the drug well, so they get minimal benefit from treatment. In one study, the cancer recurrence rate in these poor metabolizers was 9.5 times higher than in good metabolizers. Who are the poor metabolizers? Between 8% and 10% of women of European ancestry have gene variants that prevent them from metabolizing tamoxifen to endoxifen. At this time, the FDA neither requires nor recommends testing for variants in the CYP2D6 gene. However, a test kit is available.
• Variants of the gene that codes for CYP2C19 can greatly reduce the benefits of clopidogrel [Plavix], a drug that prevents platelet aggregation. Like tamoxifen, clopidogrel is a prodrug that must undergo conversion to an active form. With clopidogrel, the conversion is catalyzed by CYP2C19. Unfortunately, about 25% of patients produce a variant form of the enzyme—CYP2C19*2. As a result, these people experience a weak antiplatelet response, which places them at increased risk for stroke, myocardial infarction, and other events. People with this genetic variation should use a different antiplatelet drug.
• Among Americans of European heritage, about 52% metabolize isoniazid (a drug for tuberculosis) slowly and 48% metabolize it rapidly. Why? Because, owing to genetic differences, these people produce two different forms of N-acetyltransferase-2, the enzyme that metabolizes isoniazid. If dosage is not adjusted for these differences, the rapid metabolizers may experience treatment failure and the slow metabolizers may experience toxicity.
• About 1 in 14 people of European heritage have a form of CYP2D6 that is unable to convert codeine into morphine, the active form of codeine. As a result, codeine cannot relieve pain in these people.
The following examples show how a genetically determined variation in drug metabolism can increase drug toxicity.
• Variants in the gene that codes for CYP2C9 can increase the risk for toxicity (bleeding) from warfarin [Coumadin], an anticoagulant with a narrow TI. Bleeding occurs because (1) warfarin is inactivated by CYP2D9 and (2) patients with altered CYP2D9 genes produce a form of the enzyme that metabolizes warfarin slowly, allowing it to accumulate to dangerous levels. To reduce bleeding risk, the FDA now recommends that patients be tested for variants of the CYP2C9 gene. It should be noted, however, that in this case outcomes using expensive genetic tests are no better than outcomes using cheaper traditional tests, which directly measure the effect of warfarin on coagulation.
• Variants in the gene that codes for thiopurine methyltransferase (TPMT) can reduce TPMT activity and can thereby delay the metabolic inactivation of two thiopurine anticancer drugs: thioguanine [generic only] and mercaptopurine [Purinethol]. As a result, in patients with inherited TPMT deficiency, standard doses of thiopurine or mercaptopurine can accumulate to high levels, posing a risk for potentially fatal bone marrow damage. To reduce risk, the FDA recommends testing for TPMT variants before using either drug. Patients who are found to be TPMT deficient should be given these drugs in reduced dosage.
• In the United States, about 1% of the population produces a form of dihydropyrimidine dehydrogenase that does a poor job of metabolizing fluorouracil, a drug used to treat cancer. Several people with this inherited difference, while receiving standard doses of fluorouracil, have died from central nervous system injury owing to accumulation of the drug to toxic levels.
Genetic Variants That Alter Drug Targets
Genetic variations can alter the structure of drug receptors and other target molecules and can thereby influence drug responses. These variants have been documented in normal cells and in cancer cells and viruses.
Genetic variants that affect drug targets on normal cells are illustrated by these two examples.
• Variants in the genes that code for the beta1-adrenergic receptor (ADRB1) produce receptors that are hyperresponsive to activation, which can be a mixed blessing. The bad news is that, in people with hypertension, activation of these receptors may produce an exaggerated increase in blood pressure. The good news is that, in people with hypertension, blockade of these receptors will therefore produce an exaggerated decrease in blood pressure. Population studies indicate that variant ADRB1 receptors occur more often in people of European ancestry than in people of African ancestry, which may explain why beta blockers work better, on average, against hypertension in people with light skin than in people with dark skin.
• The anticoagulant warfarin works by inhibiting vitamin K epoxide reductase complex 1 (VKORC1). Variant genes that code for VKORC1 produce a form of the enzyme that can be easily inhibited, and hence anticoagulation can be achieved with low warfarin doses. If normal doses are given, anticoagulation will be excessive, and bleeding could result. To reduce risk, the FDA recommends testing for variants in the VKORC1 gene before warfarin is used.
Genetic variants that affect drug targets on cancer cells and viruses are illustrated by these three examples.
• Trastuzumab [Herceptin], used for breast cancer, only works against tumors that overexpress human epidermal growth factor receptor type 2 (HER2). The HER2 protein, which serves as a receptor for hormones that stimulate tumor growth, is overexpressed in about 25% of breast cancer patients. Overexpression of HER2 is associated with a poor prognosis but also predicts a better response to trastuzumab. Accordingly, the FDA requires a positive test result for HER2 overexpression before trastuzumab is used.
• Cetuximab [Erbitux], used mainly for metastatic colorectal cancer, only works against tumors that express the epidermal growth factor receptor (EGFR). All other tumors are unresponsive. Accordingly, the FDA requires evidence of EGFR expression if the drug is to be used.
• Maraviroc [Selzentry], a drug for HIV infection, works by binding with a viral surface protein known as chemokine receptor 5 (CCR5), which certain strains of HIV require for entry into immune cells. HIV strains that use CCR5 are known as being CCR5 tropic. If maraviroc is to be of benefit, patients must be infected with one of these strains. Accordingly, before maraviroc is used, the FDA requires that testing be done to confirm that the infecting strain is indeed CCR5 tropic.
Genetic Variants That Alter Immune Responses to Drugs
Genetic variants that affect the immune system can increase the risk for severe hypersensitivity reactions to certain drugs. Two examples follow.
• Carbamazepine [Tegretol, Carbatrol], used for epilepsy and bipolar disorder, can cause life-threatening skin reactions in some patients—specifically, patients of Asian ancestry who carry genes that code for an unusual human leukocyte antigen (HLA) known as HLA-B*1502. (HLA molecules are essential elements of the immune system.) Although the mechanism underlying toxicity is unclear, a good guess is that interaction between HLA-B*1502 molecules and carbamazepine (or a metabolite) may trigger a cellular immune response. To reduce risk, the FDA recommends that patients of Asian descent be screened for the HLA-B*1502 gene before carbamazepine is used. If the test is positive, carbamazepine should be avoided.
• Abacavir [Ziagen], used for HIV infection, can cause potentially fatal hypersensitivity reactions in patients who have a variant gene that codes for HLA-B*5701. Accordingly, the FDA recommends screening for the variant gene before using this drug. If the test is positive, abacavir should be avoided.
In the future, pharmacogenomic analysis of each patient may allow us to engage in revolutionary personalized medicine that addresses the individual patient’s genotype. For the present, however, although many advances have been made in pharmacogenomic knowledge, the science is still relatively new (as science goes). Nevertheless, the rapid expanse of knowledge in this area is astonishing. See Table 6.2 for resources to help you keep abreast of changes in this field.
TABLE 6.2
Pharmacogenomic Resources for Health Care Providers
| Organization | Resource | Website |
| Clinical Pharmacogenetics Implementation Consortium (CPIC) | Guidelines to assist providers in using genetic testing to optimize drug therapy |
https://www.pharmgkb.org/page/cpic https://www.pharmgkb.org/view/dosing-guidelines.do?source=CPIC# |
| U.S. Food and Drug Administration (FDA) | Table of Pharmacogenomic Biomarkers in Drug Labeling | http://www.fda.gov/drugs/scienceresearch/researchareas/pharmacogenetics/ucm083378.htm |
| Genetics/Genomics Competency Center (G2C2) | Genetics and genomics resource-specific search engine | http://g-2-c-2.org |
| Genetics in Primary Care Institute (GPCI) | Multiple resources for application into primary practice | https://geneticsinprimarycare.aap.org |
| Personalized Medicine Coalition | Variety of resources including a table that links drugs, biomarkers, and indications | |
| Pharmacogenomics Knowledgebase (PharmGKB) | A wealth of information including a listing of drugs having labels with genetic information approved by the FDA and Health Canada (Santé Canada) |
Gender- and Race-Related Variations
Gender- and race-related differences in drug responses are, ultimately, genetically based. Our discussion of pharmacogenomics continues with a focus on these important topics.
Gender
Men and women can respond differently to the same drug. A drug may be more effective in men than in women, or vice versa. Likewise, adverse effects may be more intense in men than in women, or vice versa. Unfortunately, for most drugs, we do not have adequate knowledge about gender-related differences because, before 1997 when the FDA pressured drug companies to include women in trials of new drugs, essentially all drug research was done in men. Since that time, research has demonstrated that significant gender-related differences really do exist. Here are four examples.
• When used to treat heart failure, digoxin may increase mortality in women while having no effect on mortality in men.
• Alcohol is metabolized more slowly by women than by men. As a result, a woman who drinks the same amount as a man (on a weight-adjusted basis) will become more intoxicated.
• Certain opioid analgesics (e.g., pentazocine, nalbuphine) are much more effective in women than in men. As a result, pain relief can be achieved at lower doses in women.
• Quinidine causes greater QT interval prolongation in women than in men. As a result, women given the drug are more likely to develop torsades de pointes, a potentially fatal cardiac dysrhythmia.
Although there is still a lack of adequate data related to drug effects in women, information generated by these drug trials, coupled with current and future trials, will permit drug therapy in women to be more rational than is possible today. In the meantime, clinicians must keep in mind that the information currently available may fail to accurately predict responses in female patients. Accordingly, clinicians should remain alert for treatment failures and unexpected adverse effects.
Race
In general, race is not very helpful as a basis for predicting individual variation in drug responses. To start with, race is nearly impossible to define. Do we define it by skin color and other superficial characteristics? Or do we define it by group genetics? If we define race by skin color, how dark must skin be, for example, to define a patient as “black?” On the other hand, if we define race by group genetics, how many ancestors of African heritage must a patient have to be considered genetically “black?” And what about most people, whose ancestry is ethnically heterogeneous? Latinos, for example, represent a mix of ethnic backgrounds from three continents.
What we really care about is not race per se, but rather the specific genetic and psychosocial factors—shared by many members of an ethnic group—that influence drug responses. Armed with this knowledge, we can identify group members who share those genetic or psychosocial factors and tailor drug therapy accordingly. Perhaps more importantly, application of this knowledge is not limited to members of the ethnic group from which the knowledge arose: we can use it in the management of all patients, regardless of ethnic background. How can this be? Owing to ethnic heterogeneity, these factors are not limited to members of any one race. Hence, when we know about a factor (e.g., a specific genetic variation), we can screen all patients for it, and, if it’s present, adjust drug therapy as indicated.
This discussion of race-based therapy would be incomplete without mentioning BiDil, a fixed-dose combination of two vasodilators: isosorbide dinitrate (ISDN) and hydralazine, both of which have been available separately for years. In 2005 BiDil became the first drug product approved by the FDA for treating members of just one race, specifically, African Americans. Approval was based on results of the African-American Heart Failure Trial (A-HeFT), which showed that, in self-described black patients, adding ISDN plus hydralazine to standard therapy of heart failure reduced 1-year mortality by 43%—a very impressive and welcome result. Does BiDil benefit African Americans more than other Americans? We do not know; only patients of African ancestry were enrolled in A-HeFT, so the comparison cannot be made. The bottom line? Even though BiDil is approved for treating a specific racial group, there is no proof that it would not work just as well (or even better) in some other group.
Comorbidities and Drug Interactions
Individuals often have two or more medical conditions or disease processes. When this occurs, drugs taken to manage one condition may complicate management of the other condition. As an example, if a person who has both asthma and hypertension is prescribed a nonselective beta-adrenergic antagonist (beta blocker) to control blood pressure, this may worsen the patient’s asthma symptoms if the dose is sufficient to cause airway constriction. This illustrates the necessity for the provider to consider the whole patient, not only the disease being treated, when selecting drug therapy.
Because patients with comorbidities often take multiple medications, there is the increased likelihood of drug interactions. Drug interactions can be an important source of variability. The mechanisms by which one drug can alter the effects of another and the clinical consequences of drug interactions are discussed at length in Chapter 4.
