Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Clin Pharmacol Ther. Author manuscript; available in PMC 2010 July 27.
Published in final edited form as:
PMCID: PMC2910521

Addressing The Challenges Of The Clinical Application Of Pharmacogenetic Testing

Ogechi N. Ikediobi, Pharm.D., Ph.D.,1 Jaekyu Shin, Pharm.D., M.S.,1 Robert L. Nussbaum, M.D.,2 Kathryn A. Phillips, Ph.D.,1 the UCSF Center for Translational and Policy Research on Personalized Medicine (TRANSPERS), Judith M. Walsh, M.D., M.P.H,2 Uri Ladabaum, M.D.,2 and Deborah Marshall, Ph.D.3


Pharmacogenomics aims to use molecular genetic markers to predict treatment outcome. Indeed within the past decade, there has been a rapid emergence of pharmacogenetic tests to aid clinicians to predict efficacy or toxicity for some drugs. Despite this major advance in therapeutic drug management there remain challenges to the appropriate use of pharmacogenetic tests. We discuss UGT1A1 pharmacogenetic testing to illustrate the knowledge gaps impeding widespread use of pharmacogenetic tests in the clinical setting.


Pharmacogenetic tests are potentially useful tools for making a therapeutic decision by identifying patients who should or should not receive a particular drug, or guiding individual drug-dosing. Pharmacogenetic test information is currently included in over 200 drug labels among those approved in the United States (1). The information is classified into three categories that guide the clinical use of pharmacogenetic tests for reaching a therapeutic decision (Table) (2): 1) test required, 2) test recommended and 3) information only. Thus far, only four drugs – cetuximab, trastuzumab, maraviroc and dasatinib – require a pharmacogenetic test before they are prescribed. Therefore the majority of drugs with labels containing pharmacogenetic test information do not require pharmacogenetic testing. For example, the label for the anticancer drug irinotecan recommends testing for the presence of a variant of UDP-glucuronosyl transferase 1A1 (UGT1A1) to prevent drug toxicity. However, pharmacogenetic testing for UGT1A1 has many challenges for its appropriate clinical use, similar to most pharmacogenetic tests. In this paper, we use the UGT1A1 pharmacogenetic test as an example to illustrate some of the major challenges to the clinical application of pharmacogenetic tests – incomplete knowledge of the extent of human genetic variation, availability of alternative biomarkers, and the lack of a model of delivery for pharmacogenetic information.

Selected pharmacogenetic tests available for clinical application.

UGT1A1 Pharmacogenetic Testing

UGT1A1 is a hepatic enzyme involved in the glucuronidation of bilirubin and many drugs such as the active metabolite of the anticancer drug, irinotecan. Enzymatic activity of UGT1A1 differs among individuals and can be indirectly affected by environmental, physiologic and epigenetic changes. However, UGT1A1 enzyme function is genetically determined by inherited sequence variation in the coding and non-coding regions, most notably a two base pair insertion (TA)n in its promoter region (3). The most common promoter variants are (TA)5, (TA)6, (TA)7, and (TA)8 (3). The (TA)5 and (TA)6 variants are associated with high UGT enzymatic activity; the (TA)6 allele in particular is referred to as the *1 or “wildtype” allele. The (TA)7 and (TA)8 repeats are associated with low UGT enzymatic activity (3). For the purposes of this paper, we focus on the variant (TA)7, or *28, allele because it has received the most attention.

The *28 allele is associated with inherited forms of unconjugated hypberbilirubinemia such as Gilbert syndrome and Crigler-Najjar syndromes (4). With regard to drug therapy, retrospective and prospective studies have demonstrated an association between the presence of reduced UGT1A1 enzymatic activity (especially with the *28 allele) and increased incidence of toxicity, most notably severe neutropenia in patients treated with irinotecan (5,6). The association between the *28 allele and severe diarrhea is not as statistically significant as is neutropenia. Therefore there may be other yet unknown genetic factors that are more predictive of irinotecan-induced diarrhea. In 2005 the FDA amended the package insert for irinotecan to include a recommendation, but not a requirement, to test for the *28 UGT1A1 variant to predict those at risk for neutropenia (7). Shortly thereafter, the Invader® Molecular assay for UGT1A1 genotyping appeared on the market as an FDA approved test, with an overall correct call rate of 98.8% (8).

A recent meta-analysis reports that the risk of severe neutropenia for patients homozygous for the *28 allele may be a function of the administered dose of irinotecan (9). The study found that risk of toxicity was higher for patients homozygous for *28 allele and on chemotherapeutic regimens, of irinotecan alone or in combination with other myelotoxic drugs, containing medium (125-150 mg/m2 OR= 3.22, 95% CI= 1.52 to 6.81; P=0.008) and high dose irinotecan (> 150 mg/m2 OR=27.8, 95% CI=4.0 to 195; P=0.005) compared to patients heterozygous for *28 and homozygous for *1 (wild type) (9). For this reason the study authors have asserted, contrary to the FDA's initial recommendation of a dose reduction based on genotype alone, that a reduced irinotecan dose or an alternative treatment regimen be used only when the *28 allele is present and a high dose irinotecan regimen is used (9). This recent finding may help refine the FDA labeling of irinotecan as pertains to the interpretation of the UGT1A1 genotype within the context of appropriate dosing of irinotecan. However, recent reports show that a decrease in irinotecan dose for patients homozygous for the *28 allele may result in decreased tumor response, further complicating the appropriate drug-dosing to prevent irinotecan toxicity (8).

In light of the FDA's initial recommendation, oncologists at several academic health centers incorporated the UGT1A1 test into their clinical practice. It is unknown, however, how knowledge of the genotype impacted on their care of the patient (8). On the other hand there are others still who did not heed the FDA's recommendation for UGT1A1 genotyping, due in part to the uncertainty of its clinical utility. What then are the challanges to the use of pharmacogenetic testing for UGT1A1 in the clinic?

The first challenge is that the UGT1A1 test may be limited in its general applicability in diverse populations. There are at least 113 variants in the UGT1A1 gene, most of which are associated with reduced or inactive enzyme activity, a couple associated with increased activity, and still others are of unknown significance (4). However, the most widely used UGT1A1 test (Invader® Molecular assay) only assesses one promoter variant, the *28 allele. Homozygosity for the *28 allele occurs in 10% of the North American population and at a similar frequency in Caucasians and African Americans, but less so in Asians (4). The frequency of the *28 allele has not been thoroughly studied in other ethnic groups, such as Hispanics living in North America. Even if one were to comprehensively test for all the known genotypes of the UGT1A1 promoter polymorphisms, the mere fact that only that portion of the gene is assessed and not the entire gene sequence will mean that other potentially important variants are missed, variants that may impact on clinical care. For example the currently FDA approved UGT1A1 test does not include the reduced activity *6 allele in the coding region (211G>A, G71R), which is less common than *28 but has been shown to be predictive of hyperbilirubinemia, especially in Asians (4).

Second, an inexpensive biomarker, total serum bilirubin, is available as a clinical predictor of liver function, and can serve as a surrogate marker of UGT1A1 enzyme function and severe neutropenia prior to administering irinotecan (6). Some oncologists argue against the use of the UGT1A1 genotype because they think bilirubin levels, with which they are more comfortable, may be as reliable an indicator as the genotype in predicting the appropriate selection of irinotecan dose (anecdotal evidence). However, there are no studies showing how total bilirubin levels may be used to guide irinotecan dose modification or a priori selection. Thus, a second challenge is uncertainty over the clinical utility of UGT1A1 genotype and serum bilirubin in guiding the appropriate dose selection for irinotecan.

Third, there is no model for the delivery and interpretation of the UGT1A1 test. In considering the FDA label change for irinotecan, the FDA recommends a reduced initial dose of irinotecan for patients homozygous for the *28 allele. The FDA recomendations go on to state however that the precise dose reduction in patients homozygous for the *28 allele is unknown and subsequent dose modifications should be tailored according to the individual patient's treatment tolerance (7). Unlike other genetic tests for disease prognosis, the reports for pharmacogenetic tests are currently not interpreted for the ordering physicians. There is neither a detailed explanation of how and to what extent the initial dose should be reduced, nor for how long the patient should remain on the reduced dose. The onus is on the ordering oncologist to consult with other colleagues and the literature to make sense of how to use the test result to manage irinotecan dosing. Without clear instructions on how to interpret and use the test results, oncologists, who are generally not trained in interpreting genetic tests, may be reluctant to order the test in the first place.


Given the challenges for the appropriate use of the UGT1A1 pharmacogenetic test, it is clear that a lot more needs to be done to clarify the role of the UGT1A1 test in the management of irinotecan toxicity. Importantly, these challenges are not unique to the UGT1A1 test and therefore solutions to redress knowledge gaps in the use of UGT1A1 testing can extend to other pharmacogenetic tests. First, pharmacogenetic tests could be more clinically applicable if they included a comprehensive survey of variation in the human genome. Many current pharmacogenetic tests evaluate one or a few candidate genes with a biologically plausible link to drug responses. Although successful for some drugs, this approach may miss important contributions of variation in other genes, therefore reducing the predictive value of the test. It is promising that many recent pharmacogenomic studies employ a genome-wide approach. In addition, many populations such as African-Americans and Hispanics are currently under-represented in pharmacogenomic studies and potentially important variants in those groups are not well known. Although diversity in a study population may complicate analysis, multiple pharmacogenomic studies in different ethnic groups will support the clinical use of the test in broader populations.

Second, it is essential to evaluate the clinical utility of pharmacogenetic testing, including the risks, costs and benefits relative to use of the available alternative biomarkers. Lack of clinical utility data is one of the reasons the Evaluation of Genomic Applications in Practice and Prevention (EGAPP) Working Group cannot recommend for or against the clinical use of some pharmacogenetic tests (9). Such data would be very useful if third party payers are to shoulder the cost of pharmacogenomic testing.

Third, successful integration of pharmacogenetic testing in the clinic requires not only a clinical laboratory that establishes and validates the assay but also a rapid reporting system that provides appropriate guidance as to the interpretation of the test results. Given the complexity of the process, it will likely require the coordinated efforts of multiple healthcare professionals including laboratory medicine specialists, physicians, nurses, and pharmacists.

Finally, clinicians’ knowledge of pharmacogenetic testing may also influence its successful integration into the clinic. A recent systematic review suggests that clinicians are not generally confident in providing genetic services to patients because of lack of training and knowledge (10). Therefore, efforts are needed to improve clinicians’ knowledge of pharmacogenetic tests in order to facilitate their successful integration into clinical practice.


This study was partially supported by a Program Project Grant P01CA130818 from the National Cancer Institute to Dr. Phillips


Drs. Ikediobi and Shin contributed equally to this work.


1. Pollack A. Patient's DNA may be signal to tailor medication. The New York Times. 2008 December 29;
2. Table of valid genomic biomarkers in the context of approved drug labels. U.S. Food and Drug Administration. [January 26, 2009].
3. Beutler E, Gelbart T, Demina A. Racial variability in the UDP-glucuronosyltransferase 1 (UGT1A1) promoter: a balanced polymorphism for regulation of bilirubin metabolism? Proc Natl Acad Sci U S A. 1998;95:8170–8174. [PubMed]
4. Strassburg CP. Pharmacogenetics of Gilbert's syndrome. Pharmacogenomics. 2008;9:703–715. [PubMed]
5. Ando Y, et al. Polymorphisms of UDP-glucuronosyltransferase gene and Irinotecan toxicity: a pharmacogenetic analysis. Cancer Res. 2000;60:6921–6926. [PubMed]
6. Innocenti F, et al. Genetic variants in the UDP-glucuronosyltransferase 1A1 gene predicts the risk of severe neutropenia of Irinotecan. J Clin Oncol. 2004;8:1382–1388. [PubMed]
7. Camptosar Full Prescribing Information. [January 4, 2009]. Available at
8. Evaluation of genomic applications in practice and prevention (EGAPP) working Group Recommendations from the EGAPP Working Group: can UGT1A1 genotyping reduce morbidity and mortality in patients with metastatic colorectal cancer treated with Irinotecan? Genet Med. 2009;11:15–20. [PMC free article] [PubMed]
9. Hoskins JM, Goldberg RM, Qu P, Ibrahim JG, McLeod HL. UGT1A1*28 genotype and Irinotecan-induced neutropenia: dose matters. J Natl Cancer Inst. 2007;99:1290–1295. [PubMed]
10. Scheuner MT, Sieverding P, Shekelle PG. Delivery of genomic medicine for common chronic adult diseases: A systematic review. JAMA. 2008;299:1320–1334. [PubMed]