6

inhaled steroids in asthma and polymorphisms in the type 1 corticotrophin-releasing hormone receptor (67). This identifies a new predictor of drug response and suggests that this receptor could be a target for new drug development. Similarly, screening 145 polymorphisms in 5 candidate ion-channel genes has identified a noncoding region in a potassium-channel gene as a potential contributor to risk for torsades de pointes (68). Screening 148 polymorphisms in 10 genes involved in cholesterol control identified non-coding variants in the 3-hydroxy-3-methylglutaryl coenzyme A reductase gene-modulating response to statin drugs that interact with this gene product (69). One advantage of this approach is that the number of rationally chosen candidate genes is markedly expanded but remains technologically and computationally tractable. In addition, increasingly sophisticated computational techniques are evolving to analyze how the behavior of a complex biological system changes in response to variable function of an individual component (70, 71). An increasingly sophisticated understanding of the relationship between genetic variants and variability in drug response has identified and highlighted many important obstacles that must be overcome if this approach is to enter and enrich clinical practice. Design and Interpretation of Pharmacogenetic and Pharmacogenomic Studies Identification of an association between a clinical phenotype, such as drug response, and a genetic variant or a set of genetic variants is an increasing theme in the medical literature. However, many such associations were not reproduced in subsequent studies (72, 73); false-positive associations are one common reason. Another reason may be subtle differences in patient study groups, such as ethnicity or definition of end points. Thus, similar to other diagnostic approaches, pharmacogenetic testing will be implemented only when its predictive value is established. Regulatory Issues in Genetic Testing Questions regarding regulatory issues include mechanisms for regulating genotyping tests, the extent to which pharmacogenetic analyses should be incorporated into new drug development before or after large clinical trials, and whether and how pharmacogenetic information can be incorporated into the product labels that inform clinicians and patients. The U.S. Food and Drug Administration has launched an initiative to collect pharmacogenetic information during drug development that may help address some of these issues (74, 75); in addition, as already discussed, some drug labels have been changed to include pharmacogenetic information. Development of New Genomic Technologies Delivery of reliable genetic information to the clinician for decision making requires very rapid turn-around, or a setting in which drug therapy can be reasonably initiated after the results of a genetic test are available (for example, in chronic disorders, such as hypertension). Technology in this area is evolving very rapidly. A third scenario, in which an individual patient's whole genome is sequenced and available for decision making, seems likely within 5 to 10 years. This will require whole genome sequencing for less than several thousand dollars, appropriate information management systems, and new genomic statistical analysis approaches. Ethical Issues Identification of genetic variants associated with altered pathophysiology has raised questions of whether individuals or groups enriched for such genetic variants could be stigmatized (for example, by being denied insurance) (76, 77). Education The human genome is a new resource, and considerable education will be required to apply its promise in practice (78). Even for common and well-defined genetic variants with reproducible and important consequences for disease or drug therapy, wide acceptance by the medical community has been slow. This may reflect failure of education or lack of clear-cut studies showing added clinical value to understanding genetic information before prescribing drugs. Increasing educational efforts for the medical and lay communities will be required to understand the advantages and limitations of new

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18