One For All? Treatment Choices in Ethnicity

An increasing number of study results is uncovering differences among ethnic population groups in the metabolism, effectiveness, and side effect profiles of many important drugs.

Eight percent of new drugs approved between 1995 and 1998 contained a statement in their labeling about ethnic or ancestral differences in effectiveness. Although ethnic background can be a marker only for an “average” patient within a group, cultural and biological differences must be taken into account in prescribing medications, especially for minority individuals who may be at risk for suboptimal care. It is important to point out, however, that the reported differences in drug response across different ethnic groups do not imply “racial” differences. In fact, most geneticists and anthropologists currently do not believe in the concept of “race.” Most studies showing population differences in response to medicines depend on self-reported identity (eg, African American, Asian, or Hispanic) that is based on social or cultural affiliation. Although some drug-response differences are correlated with gene variations across populations from different geographical origins (eg, Africa, Asia, or Europe), genetic differences between these populations are not large or coherent enough to constitute distinct racial groups.

But grouping people into ethnic or ancestral categories can be useful when discussing trends among populations and in increasing the clinician’s awareness of possible differences in drug response and of considering the possibility of alternative treatment modalities. But these groupings should not be used to make firm decisions about medications or for limiting options for an individual patient. To do so would be to stereotype in ways that could contribute to existing healthcare disparities.

Hypertension and Heart Failure

Hypertension and heart failure are two important examples in which ethnic differences in response to medications have been reported. Some of these differences may be related to differences in the underlying etiology of these conditions. There is increasing recognition that presentation, response to therapy, and clinical outcome differ according to ethnicity for patients with hypertension. For example, African Americans have a higher prevalence of hypertension, are more likely to be salt-sensitive, and are more prone to left ventricular hypertrophy than Caucasian Americans. Genetic differences, environmental factors, or maladaptive kidney, endocrine, and autonomic responses may account for the interracial differences.

These factors may underlie reports of differential effects of various classes of hypertension medications in African Americans.

Some studies have shown that beta-blockers, ACE inhibitors-and angiotensin receptor-blocking agents may not control blood pressure in African Americans with the same degree of effectiveness as in Caucasians. For example, fewer African Americans than Caucasians have a good antihypertensive response to beta-blockers. Conversely, because African American patients retain

more salt and have a higher incidence of salt-sensitive high blood pressure, the use of diuretics may be more effective in treating high blood pressure in African Americans than in Caucasians.

With regard to heart failure, ACE inhibitors and angiotensin 2 receptor antagonists also appear to be less effective in African American populations for its prevention and treatment of this disease. African Americans have a higher rate of heart failure than the American population as a whole, and heart disease in African Americans tends to be linked more to hypertension than to arterial blockages that are a major cause in Caucasian heart patients. African Americans also tend not to respond to some heart failure drugs as well as other populations.

Conversely, since African Americans are relatively more likely to have a deficiency of nitric oxide, which plays a role in dilating blood vessels and controlling blood pressure, they may respond better to drugs that enhance nitric oxide activity. A regimen consisting of two such drugs, hydralazine plus isosorbide dinitrate, appears to be especially beneficial in African American patients.

The Role of Genes in Response to Medications

Variability in drug response among individual patients is determined by a combination of physiology (eg, age and body weight), environmental factors (diet, alcohol, caffeine, etc.) and genetics. Since some variations in genes governing the sensitivity to drugs are either more or less common among certain population groups, response to drug therapy can also differ. One study found significant differences in frequencies between people of European vs. African ancestry for more than two-thirds of gene variants associated with drug response reported in previous studies.

Variations in genes encoding drug-receptor proteins can alter the sensitivity to a drug’s effect, and some differences in drug response among ethnic groups appear to be related to such genetic variations. Studies of twins and blood relatives have shown that genetic factors are important determinants of the normal variability of drug effects. Gene-linked variability across ethnic groups in response to medicines used in the treatment of cancer, hepatitis, psychiatric disorders, and asthma has been reported.

Cancer

Only a minority of lung cancer patients respond to gefitinib (Iressa). It has recently been discovered that the tumors of patients who do respond to gefitinib carry mutations in a gene encoding a receptor for epidermal growth factor (EGF), whereas non-responders to gefitinib do not carry these mutations. EGF receptor mutations are more frequent in Japanese patients with lung cancer than in US patients, and Japanese are correspondingly more likely to respond to gefitinib.

Hepatitis C

African Americans may not respond as well as Caucasians to anti-viral treatments for hepatitis C; a drug combination of peginterferon alpha-2b ribavirin was far less effective among African Americans compared with Caucasians. In this study, 52% of non-Hispanic Caucasians vs. only 19% of African Americans showed an absence of the hepatitis virus in their blood six months after treatment with this combination. The difference in response rate was not associated with viral strain, or with

socio-demographic or clinical characteristics. The reason the treatment was less effective in African Americans is currently

unknown.

Depression and Psychosis

Hispanics have been reported to require lower doses and are more prone to side effects at normal doses of tricyclic and SSRI antidepressants. Hispanics also tend to require lower doses of antipsychotic medications. In one study, the average effective dose of antipsychotic medication for Hispanics was half the dose in Caucasians and African Americans. Genetic factors

are believed to contribute to this differential response to medications.

Asthma

Genetic variations may underlie differences in the effectiveness of the beta-agonist albuterol used in the treatment of asthma.

Groups of genetic variations are called haplotypes, and different haplotypes are associated with patients’ varying response

to albuterol. Haplotype 4 is associated with depressed responsiveness and haplotype 2 with increased responsiveness. Some of the haplotypes appear to be either more or less common in different ethnic groups. Moderate responses to albuterol may occur

less frequently, and low responses may occur more frequently in Hispanics compared with non-Hispanic Caucasians. Responsiveness to albuterol also differs across Hispanic sub-populations; Puerto Ricans appear to be less responsive to albuterol than Mexicans, and this is believed to have a genetic basis.

It is important to note that although a drug can on average have a greater or lesser effect on persons of a given ethnic group, the range of responses within each group is often wide, and there is usually a considerable overlap in ranges. Thus, despite average differences in treatment outcomes, it would never be justifiable to withhold any drug from a patient on the basis of ancestry

alone. Ethnic background can be a marker only for an “average” patient within a group and should be considered in the initial choice of therapy.

Genes and Drug MetabolismIn addition to their influence on drug-receptor proteins, with consequential effects on sensitivity to drug action, genes also code for drug-metabolizing enzymes, thereby influencing the rate of drug metabolism. Variations in these genes can result in slower or faster drug metabolism. A person receives one copy of a gene from each parent, but each of those copies may have a change, or mutation.

Mutations can result in the production of an enzyme with reduced function, or no enzyme may be produced at all. Depending on whether a person has 0, 1, or 2 normal copies of a gene, he or she may be a poor, intermediate, or rapid metabolizer. The fewer normal copies a person has, the poorer the metabolism and the higher blood levels of the drug. This, in turn, may result in greater effectiveness or, more likely, greater side effects.

Different Dosages Are Required, Often Due to Differences in Drug Metabolism

The drug dosages used in clinical trials with East Asian subjects (conducted in both Asia and the US) are typically lower than dosages in trials in which the subjects are westerners. For instance, East Asians respond to lower doses of antipsychotic drugs and are more sensitive to side effects. In Asian countries, optimal treatment with lithium is achievable with serum concentrations below 0.8 mEq/mL, as opposed to the range of 0.8-1.2 mEq/mL generally regarded as therapeutic in western countries. Similarly, the beta-blocker propranolol is more effective in reducing blood pressure and heart rate in Chinese populations than in Caucasians.

The differences in drug response between East Asians and westerners seem to persist in East Asian emigrants to western countries. The observed higher maximum blood concentration of the antipsychotic drug haloperidol persists in US-born compared with foreign-born Chinese, Filipino, Japanese, and Korean subjects. Similarly, the percentage of individuals with a slow clearance rate of the anti-tuberculosis drug isoniazid is similar among Japanese and Chinese populations in western countries and in their countries of origin.

Failure to consider ethnic differences in drug metabolism rates can lead to sub-optimal treatment. For example, since African Americans may metabolize psychiatric medications more slowly than Caucasians, they may experience more severe side effects; as a result, they may stop taking medications at a greater rate than Caucasians with similar diagnoses.

Variation in Drug-Metabolism Genes Can Alter Blood Levels of Drugs and Clinical Outcome

Variations in the CYP2C9 gene are clinically relevant to treatment with the antidiabetic glipizide and the anticoagulant warfarin. The presence of only one inactive copy of this gene (out of two) is sufficient to interfere with the metabolism of warfarin, a drug with a narrow window between effective blood levels and toxic blood levels. Individuals with one inactive copy of the gene are at risk of excessive anticoagulation. Since fewer East Asians than Caucasians (2-5% vs. 25-30%, respectively) carry an inactive

variant of this gene, East Asians are at relatively less risk for excessive anticoagulation.

In the case of the CYP2C19 gene, however, the converse situation pertains; here, East Asians are more likely than Caucasians to have gene variations that alter the effectiveness of drugs whose metabolism is controlled by this gene. More East Asians than Caucasians (15—25% vs. 2%) carry two inactive copies of the gene. This pattern has been shown to be predictive of poor metabolism of the ulcer drug omeprazole (Prilosec). Thus, East Asians are more likely than Caucasians to have gene variations that result in higher blood levels of omeprazole. Exposure to omeprazole and, consequently the therapeutic effect of suppression of gastric acidity, is greater in these individuals.

Dozens of drug types are metabolized by the CYP2D6 gene, including cardiovascular agents, antipsychotics, antidepressants, and analgesics. The possession of two inactive variants of the CYP2D6 gene results in completely inactive enzymes, poor metabolism, and higher blood levels of drugs. This condition occurs in only about 0.5% of East Asians vs. 3-10% of Caucasians. Conversely, possession of a double copy of a particular variant of this gene (the *10 variant) is rare among Caucasians but occurs in up to 25% of East Asians. This “*10/*10” gene pattern results in a still-active enzyme, with residual but reduced activity. This produces an “intermediate” metabolizer individual with a reduced rate of metabolism of many drugs, which can lead to somewhat higher blood levels. For example, exposure to the effects (and side effects) of the antidepressant

nortriptyline is increased in *10/*10 individuals. By contrast, codeine, which must be metabolized to morphine by CYP2D6 to become active, is less effective in *10/*10 individuals, who may find that a normal dose of codeine is inadequate.16

Implications of Drug Metabolism Differences for Therapeutic Substitution

Cost management programs that may require therapeutic substitution within drug classes in patients stabilized on a given agent of the class may create the need for substantial dosage adjustments when the patient is switched to an alternate agent from that class. Two drugs within the same therapeutic class may be metabolized by different enzymes, which are controlled by different genes. Clinicians should be especially alert for the potential need for dosage adjustment if the two drugs utilize different metabolic pathways, where one pathway is subject to genetic variation and the other is not.

Two examples of drug classes whose members use different metabolic pathways are the proton pump inhibitor (PPI) ulcer drugs and the tricyclic antidepressants. Among the PPIs, the metabolism of omeprazole (Prilosec) is dependent on the CYP2C19 genetic variations, whereas the metabolism of rabeprazole (Aciphex) is not. Similarly, among the antidepressants, CYP2C19 is important in the metabolism of clomipramine, but not of nortriptyline.18 Genetic testing may eventually be useful

in identifying situations in which therapeutic switching is appropriate. Studies that measure the clinical effects of substituting agents whose drug metabolism pathways differ are generally not available at this time.

The Future: Genetic Testing May Predict Likelihood of Drug Response or Altered Drug MetabolismThere are many current examples, mostly in cancer drugs, in which genetic factors that influence drug response are known, detected, and used to predetermine if a drug will work for an individual patient. Genetic tests are also available to detect variants in some of the genes commonly involved with drug metabolism, including several of the CYP series genes. Testing

for genetic variations and metabolic status prior to Rx treatment would allow accurate selection of medications and dosages, thereby reducing costly drug interactions and toxic responses. Among 27 drugs frequently cited in adverse drug reaction studies, 59% are metabolized by at least one enzyme with a gene variant known to cause poor metabolism.

However, the sensitivity and positive predictive value of genetic screening for drug metabolism defects isn’t perfect, because diet, alcohol consumption, and physiological differences also affect drug metabolism. In addition, variations in other genes (eg, drug receptor genes) may also play an important role in a drug’s effectiveness. The use of DNA chips—solid surfaces dotted with thousands of gene snippets that can be screened simultaneously— to detect variations in drug-metabolizing genes may eventually make genetic screening easier and faster.

However, as there is not a perfect, one-to-one correspondence between a genetic test result and poor metabolism, not all poor metabolizers will be detected; and not all subjects with a positive test result will actually have poor metabolism. Thus, it is unlikely that genetic testing will ever be completely predictive of an individual’s drugmetabolism capacity.

Implications for Physicians

The numerous ethnic variations in drug metabolism and drug sensitivity underscore the need for individualized prescribing and dosing and should be important considerations in determining the range of available therapeutic choices on drug formularies. Differences in metabolic pathways among drugs of a class, coupled with ethnic differences in frequency of gene variation in these pathways, call for caution when implementing drug substitution policies. Until individualized genetic profiles of patients become a clinical and economic reality, ethnic background, like other categories such as age or gender, can provide useful information for drug selection and the potential need for dosage adjustment.

Richard Levy, PhD, is a Senior Research Advisor for the National Pharmaceutical Council.