Ethnicity, Culture, and Psychopharmacology

Keh-Ming Lin and Russell E. Poland

One of the most important issues in clinical pharmacology that perhaps has not received sufficient attention in research, as well as in the day-to-day clinical care setting, is the remarkably large interindividual variability in drug responses and side effect profiles. Such variability, which can be 40-fold or more, has been demonstrated with practically all classes of psychotropics, making it difficult to formulate rational guidelines for the dosing and the interpretation of biological parameters (such as the plasma or serum drug concentrations) that might be associated with therapeutic response (28, 37, 49, 62). Although much remains unknown, a number of factors have been demonstrated to be important determinants of such variability. These include not only genetics, disease state, nutritional status, concurrent use of drugs, and other pharmacoactive substances, but also demographic factors such as age, gender, and ethnicity (49, 72, 79).

The primary focus of this chapter will be to briefly review the influence of ethnicity--and, to a lesser extent, culture--on psychotropic responses. It will start with an overview of the various mechanisms that could be responsible for ethnic and cultural differences in psychotropic response. This will be followed by a review of the literature which already provides substantial evidence of ethnic variations in response to major classes of psychotropics. Finally, the clinical and theoretical implications of these findings, as well as future research directions, will be discussed (see also Short and Long-Term Psychopharmacological Treatment Strategies).

As has been described in greater detail in Short and Long-Term Psychopharmacological Treatment Strategies, the effects of pharmacoactive agents are determined by pharmacokinetic and pharmacodynamic factors (see Fig. 1). While pharmacodynamics deals with the interactions of pharmacoactive substances with the target organ (various receptor systems in the central nervous system in the case of psychopharmacology), pharmacokinetic principles describe the disposition and fate of pharmacoactive substances inside the body. The pharmacokinetics of most drugs are determined by four basic processes: absorption, distribution, biotransformation, and excretion (28, 65). Of these, the biotransformation process (metabolism) shows considerable interindividual as well as cross-ethnic variations (36, 49) and has been an exceptionally active and productive area of research. The rate of biotransformation is determined by both genetic and environmental factors. Evidence suggests that ethnicity and culture exert substantial influences on drug response through both mechanisms.

The development of the field of pharmacogenetics as an academic discipline has been closely intertwined with findings of dramatic ethnic differences in drug responses that were found to be genetically determined (34, 36). Through these efforts, the genetic control of a large number of drug-metabolizing enzymes has been established (Table 1). The activities of many of these enzymes also show substantial cross-ethnic differences. Classic examples include:

1. Differential rates of "isoniazid toxicity" between Asians and Caucasians. This led to the finding of slow versus rapid acetylation across ethnic groups, which accounted for the toxic effects of this drug in certain individuals and which, more recently, accounted for the identification of ethnospecific loci of point mutations responsible for slow acetylation (80). Acetylation represents an important metabolic pathway for a large number of medicinal agents, including some frequently encountered psychoactive agents such as caffeine, phenelzine and nitrazepam.

2. "Primaquine hemolysis" found among African-American soldiers fighting in Southeast Asia during World War II (34, 37). This led to the discovery of an inborn deficiency of glucose-6-phosphate dehydrogenase, a condition that could result in severe hemolytic anemia when the afflicted is exposed to a variety of substances, including primaquine, an antimalarial agent.

3. The "flushing response" in Asians exposed to alcohol. Subsequent studies demonstrated that this was mainly due to a genetically determined deficiency of aldehyde dehydrogenase, which is accentuated further in some individuals by an overactivity of alcohol dehydrogenase. The molecular basis of these mechanisms has been elucidated by a series of studies in recent years (3, 85).

The cytochrome P-450 enzyme system represents a major focus of contemporary research in pharmacogenetics (25, 26, 27, 35, 37). Together, isozymes belonging to this system are responsible for the metabolism and detoxification (usually) of the majority of modern chemotherapeutic agents, including practically all psychotropics that require oxidation prior to conjugation and excretion. At least two of these isozymes, the CYP2D6 (debrisoquine hydroxylase) and the CYPmp (mephenytoin hydroxylase), have been found to be bimodally distributed. It has further been demonstrated that the bimodal distribution of the activities of these important enzymes is genetically controlled and can be traced to mutations in the nucleic acid sequence in the DNA, leading consequently to alterations in the amino acid structure and activity of the enzymes. Because of these mutations, a certain proportion of any given population can be classified as poor metabolizers (PMs)--in contrast to extensive metabolizers (EMs), who do not have such deficiencies.

Interestingly, substantial cross-ethnic differences in the frequency of the PM phenotype exist with these enzymes (37, 49). Table 2 summarizes the frequency of PMs of CYP2D6 in studies involving different populations. The rate ranges from less than 1% in some of the Asians studies to as high as 19% among Sans Bushmen, with the majority of the studies demonstrating a consistent and substantial contrast between Asians (0.5-2.4%) and Caucasians (2.9-10%). Recent genotyping studies have further demonstrated that the majority of the Caucasian PMs have a 44-kb gene insertion associated with several types of point mutations that render the enzyme completely inactive. In contrast, the same 44-kb gene insertion is highly prevalent (34%) among Asians, albeit without the associated point mutations. Phenotypically the Asians with 44-kb gene insertion were classified as EMs. However, their CYP2D6 metabolic capacity was significantly lower than that of either Caucasian EMs or Asian EMs without gene insertion. This led to an overall lower CYP2D6 activity in Asians as compared to Caucasians, although not low enough to be considered PMs. In a recent study with African-Americans, similar results were found, with a high prevalence (33%) of gene-insertion-related slower metabolism. Thus, EM and PM are relative terms and are not necessarily quantitatively or qualitatively comparable across ethnic groups (52).

As also shown in Table 2, the frequency of PM of CYPmp also varies substantially across ethnic groups (37, 49). While PMs of CYPmp are relatively rare among Caucasians, they have been found to be quite prevalent in Asian populations, with approximately 20% of Japanese and Chinese being classified as PMs. A recent study on African-Americans also reported a rate significantly higher than that in Caucasians (52).

The remarkable diversity of the genotypes of drug-metabolizing enzymes poses a challenging question for evolutionary biologists (26, 27, 36). The historical survival value of the inborn deficiency of glucose-6-phosphate dehydrogenase is easier to explain, because those possessing such a trait are more resistant to malarial infection. More puzzling is the case for the cytochrome P-450 isozymes. However, these isozymes together represent one of the most important defense systems that evolved in our ancestors through the millennia to protect against potentially harmful xenobiotics to which they were routinely exposed in their habitat (27). It has been argued (36) that, just as genetic variability in susceptibility to infectious diseases has been shown to be conducive to the survival of populations, so does pharmacogenetic variability help to ensure the survival of a population facing an onslaught of toxic chemicals in the environment.

In contrast to the uncertainties surrounding the phylogenetic basis of ethnic diversity of these drug-metabolizing enzymes, their clinical implications have been more clearly delineated. To greater and lesser extents, both enzymes are involved in the metabolism of a large number of psychotropics (Table 3). Recent studies have demonstrated that these pharmacogenetic traits significantly affect the pharmacokinetics of a number of psychotrropics, and are likely clinically important. For example, as compared to EMs, PMs of CYP2D6 exhibit significantly higher serum concentrations of tricyclics when given comparable doses of the medication, and PMs of CYPmp show significant differences in the metabolism of diazepam (37). Ethnic variations in regard to these, as well as other cytochrome P-450 isozymes, might explain, at least in part, some of the ethnic differences in psychotropic response that will be reviewed below.

The activities of many, but probably not all, of the cytochrome P-450 isozymes can be inhibited, as well as induced, by a wide variety of substances (57, 58). These include not only some of the highly specific and often extremely potent drugs that have been utilized for research in this area, but also commonly used pharmaceutical agents (cimetidine and carbamazepine are prime examples), herbal medicines, environmental toxins, steroid and sex hormones, alcohol, caffeine, constituents of tobacco, substances in charcoal-broiled beef, brussels sprouts and cabbage, and dietary compositions (i.e., high-protein versus high-carbohydrate diets). The mechanisms responsible for inhibition include competitive interaction (reversible inhibition), autocatalytic inactivation (irreversible inhibition), and possibly repression of gene transcription. Induction is most likely the result of increased gene transcription, leading to an increased synthesis of P-450 proteins. However, it is possible that stabilization of these proteins, and/or their corresponding mRNAs, also plays a role.

A major advance in recent years in this regard is the discovery of the "Ah (aromatic hydrocarbon) receptor," a cytosolic receptor responsible for recognizing the inducing agent and triggering the induction response by functioning as an enhancer at the 5¢-flanking region of the P-450 genes (58). This results in an increase in gene transcription, in the production of mRNAs, and consequently in the enhancement of the production of the enzymes. As the name "Ah receptor" indicates, most of the substrates of this receptor are aromatic hydrocarbons, many of them carcinogens. Two P-450 isozymes, CYP1A1 and CYP1A2, which are mainly responsible for the metabolism of these substances, are readily induced through this mechanism. Although the Ah receptor is structurally very similar to the receptors for steroid hormones, cDNA and protein sequencing studies suggest that the former might be more ancient than the latter. In fact, Ah receptor has been found in virtually all organisms, presumably has existed ever since life first emerged on earth, and thus likely serves an extremely vital function in protecting the organisms from harmful environmental chemicals. In addition to inductions mediated by the Ah receptor, there are several other major types of inductions that are also physiologically important, including the induction by phenobarbital-like compounds (P-450b and P-450e, equivalent to human CYP2B), by glucocorticoids (P-450p or CYP3A1), and by ethanol (CPY2E1). Although receptors have been speculated for these inductions, their existence has not been proven.

Remarkable cross-strain differences in the Ah-receptor-mediated inducibility have been demonstrated in inbred mice. This has led to speculation that similar heritable differences could be present in human populations, resulting from postulated polymorphisms in the Ah receptor gene(s). However, this hypothesis remains to be tested.

At the clinical level, a number of studies have demonstrated substantial cross-ethnic differences in drug metabolism that appear to be environmentally determined. Branch et al. (6) compared the rate of the biotransformation of antipyrine and found a significantly longer antipyrine half-life among Sudanese living in their home villages as compared to Sudanese residing in Great Britain and to Caucasian British subjects. The latter two groups metabolized antipyrine at similar rates, suggesting that environmental rather than genetic factors were responsible for the pharmacokinetic profiles observed among Sudanese residing in their native land. Similar findings were reported in subsequent studies involving Asian Indians living in India, Asian Indian immigrants residing in Great Britain, and Caucasian British subjects (19). When Asian Indian immigrants were divided into those who continued to follow their traditional vegetarian diet and those who had switched to a British diet, it became evident that diet was the determining factor underlying the change of the pharmacokinetic profiles in the Indians. Similar results have been found in studies utilizing clomipramine as the test drug (1).

Steroids, including sex hormones, also are predominantly metabolized by cytochrome P-450 isozymes (25, 37), which are inducible by some of the same agents listed above with proven efficacy in inducing drug-metabolizing enzymes. Interestingly, dramatic cross-ethnic differences also have been reported in the metabolism and physiological effects of these compounds, both endogenous and exogenous. For example, Asian women were demonstrated to have 30-75% lower plasma estrone and estradiol concentrations as compared to their Caucasian counterparts. Similar findings with plasma testosterone levels in males with different ethnic backgrounds have been reported (16). In a WHO sponsored multinational study of the efficacy of male contraceptive hormones, it was found that the efficacy of these preparations was much greater in subjects living in nonindustrialized countries than in those residing in western industrialized sites (R. Swerdloff, personal communication). Although much remains to be elucidated, it appears that at least part of these differences might be explained by ethnic differences in dietary practice.

The majority of drugs that utilize P-450 enzymes as the first step in their metabolism also subsequently undergo another metabolic step involving conjugation with endogenous substances (glucuronidation and sulfation) which biotransforms these compounds to more polar ones, so as to further increase their water solubility and excretion (12, 75). Contrary to earlier beliefs, emerging recent data indicate that conjugated compounds might play a significant role in determining the clinical or adverse effects of medications such as haloperidol (76). Mechanisms involved in the control of conjugation remain largely unexplored. Little is known regarding ethnic differences in conjugation. However, a recent study did report significantly slower glucuronidation of codeine in Asians, leading to heightened sensitivity to this analgesic in many of the Asians (37, 84).

As one of the major components of the pharmacokinetic processes, the distribution of drugs also has been found to manifest significant cross-ethnic variability. Because most psychotropics are highly lipophilic, they generally have a large volume of distribution, the size of which is a reflection of the body composition, especially the proportion of fat to water. Diversity in body build across ethnic groups is expected to lead to differences in the volume of distribution and thus the pharmacokinetics of drugs that are lipophilic. This in fact has been identified as one of the reasons for the greater effect of diazepam in Asians than in Caucasians (42).

Protein binding represents another important factor that could significantly influence the distribution of many drugs. Among the proteins in plasma that provide binding sites, and thus function as carriers for these pharmacoactive agents, a1-acid glycoproteins (5, 41, 69) and albumins (67) are most important. Variations in the concentration of these drug-binding proteins in plasma can significantly influence the effect of the drug by changing the free fraction and thus the amount of the unbound (free) drug concentrations in the plasma (44). Because (usually) only the free (unbound) fraction of drug is pharmacologically active and capable of crossing the blood-brain barrier, changes in the concentrations of drug-binding proteins might have profound clinical significance (13).

The structures of these plasma proteins are genetically determined. They are polymorphic, and they have been shown to vary across ethnic group in several studies (32). Cross-ethnic studies (13) of plasma protein-binding have thus far focused only on Asian-Caucasian comparisons and have revealed conflicting results. Zhou et al. (88) reported that, as compared to Caucasians, Asians had significantly lower plasma a1-acid glycoprotein but similar albumin concentrations. Kumana et al. (42), however, reported that their Hong Kong Chinese subjects had significantly lower albumin levels than did their Caucasian counterparts. Similarly, some, but not all, of the cross-ethnic protein binding studies reported lower degree of protein binding, leading to higher free drug fraction in Asians than in Caucasians (49).

The distribution of lithium across cellular membranes is controlled by several membrane transport and countertransport mechanisms (7). Among these, the sodium- lithium countertransport system appears to play a particularly pivotal role. The status of this membrane transport system is clearly under genetic control and is strongly associated with the risk for hypertension (81). This system also is significantly less active in African-Americans and African Blacks than in Caucasians, which might contribute to the higher prevalence of hypertension among Blacks (7). More recent studies have shown that, in addition to its hemodynamic implications, ethnic variability in the activity of this system also leads to significant differences in the RBC/serum lithium ratio (59, 60, 77). This ratio is likely correlated with the intracellular concentration of lithium, which might have important meaning not only in terms of the genetic control of cell membrane permeability to lithium, but also in terms of the clinical and side effects of lithium. Thus, the difference between Blacks and other ethnic groups in the RBC/serum lithium ratio might have important clinical significance. Such a possibility has been recently demonstrated by a study conducted by our group; this study revealed that significant differences in the lithium ratio exist between African-American and Caucasian bipolar patients, and it further demonstrated a higher rate of central nervous system (CNS)-related side effects in African-American patients, suggesting that the higher lithium ratio in this group might indeed lead to higher central toxicity (77).

Although, at least in theory, ethnic variations could also exist in terms of absorption, excretion, and the movement of drugs across the blood-brain barrier, little information currently exists to confirm such possibilities.

Similarly, recent studies (37) have indicated that cytochrome P-450 isozymes also exist in the CNS. Although the amount and activity of these enzymes are much smaller than those found in liver, they directly influence the concentration of their substrates at the receptor sites. Thus, they might be responsible for variations in drug response not readily explainable by peripheral pharmacokinetics. Although it is conceivable and likely that some of the ethnic variations in P-450 isozymes (e.g., CYP2D6) would also exist at the CNS level, this hypothesis has not been directly examined thus far.

Contrasting the remarkably rich literature on ethnic variations in the structure and function of various drug-metabolizing enzymes with the relative paucity of information on receptors, Kalow (34, 36) recently suggested that while diversity in the former is evolutionarily adaptive because the substrates of the enzymes are predominantly xenobiotics, organisms can ill afford substantive variability in receptors, because their substrates are endogenous. This persuasive argument not withstanding, some of the recently emerging information does suggest that ethnic variations in receptors (4) and receptor-coupled responses also exist, although their functional implications have not been fully clarified.

At a more global level, ethnic differences in the pharmacodynamics of various medications have long been reported. As far back as the 1920s, researchers have observed dramatic ethnic differences in the mydriatic responses to various classes of drugs including cocaine, ephedrine, atropine, and scopolamine (2, 20). When the same amount of medication was applied locally, Blacks were consistently least responsive, Asians in the middle, and Caucasians on the other extreme. This reduction in the mydriatic effect appeared to be correlated with the degree of pigmentation of the iris, and it was not seen among albino Africans. The fact that these were all applied locally, and that so many compounds with divergent chemical structures were involved, suggested that this phenomenon was not pharmacokinetically mediated. Recent studies involving the intramuscular administration of atropine and scopolamine (20) confirmed this suspicion. Thus, although the mechanism responsible for this phenomenon has not been fully determined, it appears that it is pharmacodynamic in nature.

Another prominent example of how ethnicity and pharmacodynamics interact is found for beta-blockers such as propranolol, which have been found to be relatively ineffective in treating hypertension in African-American patients (55). In contrast, the doses of propranolol required for the effective treatment of hypertension in Asians were substantially smaller than those required for Caucasians (87). Subsequent studies objectively demonstrated that the effects of propranolol on blood pressure and heart rate were most pronounced in Asians, and least prominent in African-Americans, with Caucasians falling in between (14). These differences could not be explained by pharmacokinetic factors (87). Studies which have demonstrated that Blacks have significantly higher concentrations of cyclic AMP, both at baseline and after the administration of propranolol, suggest that Blacks might have a higher degree of b2-adrenoceptor activity (73). This in turn has led to the hypothesis that differences in the sensitivity of adrenoceptors might be the major cause for the differential effects of propranolol and other beta blockers in various ethnic groups (33).

Clozapine-induced agranulocytosis serves as a different type of example of ethnic differences in drug responses (in this case, adverse responses) that are mediated through nonpharmacokinetic mechanisms. In earlier drug trials, it was observed that this potentially life-threatening condition was significantly more prevalent among Azkenazi Jews. This phenomenon led to the finding that a special cluster of human lymphocyte antigen (HLA) typings (45), which is present among Azkenazi Jews with a significantly higher frequency, is associated with substantially increased risk of clozapine-induced agranulocytosis. In a recent report, such an association also has been observed in an American Indian patient (61).

Observations of the ethnic differences in therapeutic concentrations of various psychotropics (29, 78, 82) and their neurohormonal effects have led to speculation about the existence of ethnic differences in pharmacodynamics of these drugs. These will be discussed below in relation to specific classes of psychotropics.

Compared to the relatively rich data-set reviewed above on ethnic diversity in drug responses that are mediated by biological mechanisms, there is practically no information regarding how, and to what extent, cultural and symbolic processes affect drug responses in general, and psychotropic responses in particular (49). This is most unfortunate because culture and symbolic forces exert powerful influences on treatment outcome (40). It is to be expected that the clinical effect of psychotropics, as well as most other drugs, is determined to a large extent by factors such as physicians' biases, patients' beliefs, expectations, placebo effects, and compliance, rather than by their "real" pharmacological properties (49).

Prescription biases are most clearly demonstrated in a series of studies (15, 51, 66, 90) consistently demonstrating that African-American patients are not only far more likely to be assigned a more severe diagnosis such as schizophrenia, but also far more likely to be treated with neuroleptics irrespective of diagnosis. Given the same diagnosis, they are significantly more likely to be placed on depot rather than oral medications, presumably reflecting the clinicians' heightened concern with problems of compliance.

As demonstrated by earlier large-scale drug trials involving multiple ethnic groups or cross-national comparisons, there is some evidence suggesting that non-Caucasians may be more responsive to placebo treatment than Caucasians (17, 24). Several studies also have elegantly demonstrated that the perception and report of side effects are intimately influenced by the patients' culturally determined beliefs and expectations (49). Sporadic reports (15, 16, 23, 38, 51, 66) also have shown that medication compliance might be a particularly serious problem in cross-cultural clinical settings. In addition, level of stress, quality and quantity of social support, and personality styles all have been reported to significantly influence psychotropic response (49). Cultural forces impinge upon all of these factors, although systematic research has not been performed in these interesting and important areas (49).

Reports of ethnic differences in psychotropic response can be traced back to the 1950s, when these potent therapeutic agents were developed and quickly introduced worldwide (56). Throughout the last four decades, reports of such nature, mostly based on clinical impressions and surveys, have been numerous (49). However, it is only in the last decade that researchers started to tackle these issues with vigorous study designs and sophisticated methodologies. Although many controversies remain unresolved, the results of these studies, taken together, clearly demonstrate that ethnicity is an important issue that should be considered in clinical settings for most, if not all, classes of psychotropics. Because of space limitations, only selected studies will be highlighted and/or discussed.

Several carefully designed studies have demonstrated that Asians and Caucasians differ significantly in terms of haloperidol pharmacokinetics and pharmacodynamics. Asian normal volunteers (46) and schizophrenic patients (64) had approximately 50% higher plasma haloperidol concentrations than their Caucasian counterparts when given comparable doses of medication. As depicted in Fig. 2, this ethnic contrast is superimposed upon an already quite remarkable interindividual variability in both ethnic groups. Thus, although trends for each group were clearly different, there was at the same time considerable overlap in the middle ranges.

The mechanisms responsible for such ethnic differences have not been elucidated. A series of recent studies (10, 30, 31 have indicated that Asians had lower reduced haloperidol/haloperidol ratios, suggesting that a lower rate of reduction (a major metabolic pathway for haloperidol) in Asians might be responsible for the slower rate of metabolism, and consequently for the more prominent effects observed when given equivalent doses. Alternatively, ethnic differences in the activities of CYP2D6 might also play a role, although such a possible association has yet to be tested.

Pharmacodynamic differences between Asians and Caucasians were suggested by the larger prolactin responses to haloperidol challenge in Asians in the above-mentioned study with normal volunteers (46), which remained statistically significant after controlling for differences in haloperidol concentrations, suggesting the existence of ethnic differences in receptor-coupled responses. In a subsequent clinical treatment study (48), Asian schizophrenic patients responded optimally to significantly lower plasma haloperidol concentrations as compared to their Caucasian counterparts, again suggesting that pharmacodynamic factors contribute to ethnic differences in response to haloperidol.

In a comparison study including four patient groups treated with therapeutic doses of haloperidol, Jann et al. (31) reported significantly different pharmacokinetic profiles for Chinese and African-Americans as compared to Caucasians and Hispanics. In contrast, however, Midha et al. (53, 54) found no differences between Canadian Blacks and Caucasians in the pharmacokinetics of two phenothiazines (trifluoroperazine and fluphenazine).

In contrast to neuroleptics, studies of ethnic differences in the pharmacokinetics of the TCAs have led to inconclusive results. Among the six previous studies comparing Asians with Caucasians, three (39, 71, 74) revealed that Asians metabolize TCAs significantly slower than their Caucasian counterparts. Although the other three studies (62, 63; Silver and Potkin, unpublished data) showed differences in the same direction, these differences did not reach statistical significance, particularly after controlling for body weight. In a recently completed study, Lin and colleagues (unpublished data) compared the pharmacokinetics of imipramine among Asians, African-Americans, Hispanics, and Caucasians. With the exception of higher desipramine concentrations in the African-American group, they did not find any significant differences among the four comparison groups. These results were in agreement with earlier study demonstrating lack of difference in the pharmacokinetics of nortriptyline between Mexican-Americans and Caucasians (21). The elevation of secondary amine concentrations also has been previously reported among African-American patients (89).

Pharmacodynamic factors have not been formally examined in studies comparing the use of antidepressants across ethnic groups. However, results from two clinical studies in Asia (29, 82) indicated that severely depressed hospitalized Asian patients responded clinically to lower combined concentrations of imipramine and desipramine (130 ng/ml) than had been previously reported in North American and European studies (180-200 ng/ml), thereby suggesting that differential brain receptor responsivity might also play a role in determining ethnic differences in tricyclic dosage requirement. Although clinical reports have suggested that African-Americans were more susceptible to CNS side effects of TCAs (50, 70), the mechanisms that might be responsible for such a phenomenon have not been carefully evaluated.

Several recent cross-national comparison studies have replicated earlier reports from Japanese researchers regarding the need for lower doses of lithium as well as lower therapeutic lithium levels among Asians (49, 77). Yang (83) studied 101 Taiwanese bipolar patients treated over a 2-year period with clinically determined doses of lithium. He found that the plasma lithium level of the majority of good responders ranged between 0.5 and 0.79 mEq/L. More recently, Lee (43) reported that bipolar patients in Hong Kong were stabilized on an average lithium concentration of 0.63 mmol/L. In two studies conducted separately in Shanghai and in Taipei, Chang et al. (8, 9) reported remarkably similar pharmacokinetic profiles and therapeutic lithium concentrations for these two Chinese groups residing in drastically divergent socioeconomic environments. These Chinese patients did not differ from Caucasians in terms of pharmacokinetics. However, these patients responded optimally to mean lithium concentrations of 0.71 and 0.73 mEq/L, respectively. These were significantly lower than the mean level of 0.98 mEq/L for the matched Caucasian-American patients, as well as significantly lower than the 0.8-1.2 Meq/L therapeutic levels generally reported in Europe and North America. Thus, it appears that, compared to their Caucasian counterparts, Asian bipolar patients may require lower doses of lithium because of pharmacodynamic reasons ("increased CNS responsivity").

As mentioned above, higher RBC/plasma lithium ratio in African-Americans and African Blacks may lead to higher central toxicity (77). This suggests that their therapeutic serum or plasma lithium concentrations might have to be lowered as compared to patients from other ethnic backgrounds. However, this remains a hypothesis that needs to be further examined.

Confirming earlier clinical and survey reports (49, 68), four recent studies (22, 42, 47, 86) involving Asians and Caucasians demonstrated significant pharmacokinetic differences between the two ethnic groups. Three of the studies used diazepam as the test drug, and one utilized alprazolam. These studies involved the administration of the test drugs by either oral or intravenous routes, or both. Furthermore, they were conducted in Asians residing in diverse areas of the world, including Los Angeles, St. Louis, Hong Kong, and Beijing. Given the diversity of sites and research methodology, the consistency in these reports of a slower metabolism of benzodiazepines is quite remarkable, suggesting that genetic factors are more important than environmental factors in the control of benzodiazepine metabolism.

In a recent study of the pharmacokinetics and pharmacodynamics of adinazolam, a triazolobenzodiazepine currently being investigated as an anxiolytic and antidepressant, African-Americans were found to have increased clearance of adinazolam, resulting in significantly higher concentrations of N-desmethyladinazolam, a metabolite of adinazolam, and greater drug effects on psychomotor performance (18, 49). Adinazolam is almost exclusively eliminated by hepatic oxidation to N-desmethyladinazolam, so these findings suggested that African-Americans may have a higher metabolic capacity for adinazolam. Because N-desmethyladinazolam is cleared directly by renal excretion in addition to hepatic metabolism, increases in oxidative capacity are expected to have a lesser effect on N-desmethyladinazolam AUC values. N-Desmethyladinazolam has been shown primarily to mediate the benzodiazepine-like side effects, including effects on psychomotor performance, after adinazolam administration (18). This might explain the greater drug effects on African-Americans despite their higher metabolic capacity for adinazolam.

As reviewed above, multiple mechanisms--not only those belonging to the realm of pharmacokinetics and pharmacodynamics, but also various psychosocial factors--could be responsible for differences in psychotropic response. Although remarkable progress has been made in clarifying the mechanisms by which ethnicity and culture influence drug response, much remains unexplored and/or unresolved. This notwithstanding, the available information clearly indicates that ethnicity and culture are important variables that should not be neglected in the practice of psychopharmacology for several important reasons.

Clinically, the importance of culture and ethnicity has been significantly intensified because of the rapid and accelerating population shifts occurring in all metropolitan areas of the world. Furthermore, because of the rapid pace of intercontinental transportation and large-scale migration, most psychiatrists no longer have the luxury of practicing their trades in culturally or ethnically homogeneous settings. Patients seeking help enter the clinic with divergent beliefs, expectations, dietary practices, and genetic constitution. These all have the potential of significantly affecting the outcome of psychopharmacotherapy and should not be ignored. This is part of the reason that the National Institutes of Health and the Food and Drug Administration have started to pay attention to ethnicity as a factor in the research activities that fall under their aegis.

Along with the escalating cost of drug development and marketing, international collaboration becomes increasingly important in such endeavors. However, in order for pharmacokinetic and clinical trial results to be shared cross-nationally, potential ethnic and cultural influences need to be identified. Failure to do so might lead to the inappropriate application of findings derived from one population to another, sometimes leading to unforeseen and potentially disastrous results. Similar arguments could be made with regard to using safety and efficacy data derived from one particular group (in this country, most often "young male Whites") for approval of pharmaceutical agents, which then are used widely in other populations. Not only are these issues beginning to be addressed early on in clinical trials, but some of these issues are currently being addressed even earlier at the molecular level. For example, in order to tailor-make a new pharmaceutical agent, the availability of cDNA-expressed P-450s will allow for in vitro studies to be performed on newly designed drugs before Phase I studies are even undertaken (11).

Finally, in terms of research, it should be again emphasized that throughout the history of the development of the field of pharmacogenetics, as well as many other fields of medicine, ethnic diversity serves as a major source of stimulus for new discoveries. It is to be expected that this will continue, with cross-ethnic research designs serving as a powerful tool for psychopharmacological research in the future.

This work was supported, in part, by the Research Center on the Psychobiology of Ethnicity (grant MH47193) and by NIMH Research Scientist Development Award MH00534 (Dr. Poland). The authors thank Dora Anderson, R.N. for her helpful comments and editorial assistance.

K.-M. Lin and R. E. Poland: Research Center on the Psychobiology of Ethnicity and the Department of Psychiatry, Harbor-UCLA Medical Center, Torrance, California 90509.

published 2000