Pharmacogenomics: From Bench to Electronic Health Record

Jharana (Tina) Patel, PharmD MBA
Chief Pharmacy Information Officer
Department of Clinical Research Informatics
National Institutes of Health Clinical Center
Bethesda, MD

Barry R. Goldspiel, PharmD BCOP BCPS FHOPA
Deputy Chief of Pharmacy
Pharmacy Department
National Institutes of Health Clinical Center
Bethesda, MD

Pharmacogenomics (PG), a form of personalized medicine, has many applications to oncology.1 Pharmacogenomics, sometimes used interchangeably with the term pharmacogenetics, has great potential to optimize medication therapy outcomes for both primary cancer treatment and supportive care by improving efficacy and safety when it is translated to the clinical setting.2,3 Although barriers to implantation have been identified, some background information and practical tips will help those who work in oncology begin the process of routinely integrating PG information into patient care.4

The Science
Gene variation can be inherited (germline) or acquired (somatic). Germline gene variations are inherited and may be associated with developing cancer (e.g., BRCA1 and breast or ovarian cancer) or producing variation in drug-metabolizing, enzyme, and transporter genes, which determines medication efficacy and toxicity.5 (Table 1 – See PDF) provides a summary of germline gene variations for medications commonly used in cancer patients. For medications in which metabolism leads to inactive products, it is important to note that variation in at least one of the two copies of a drug-metabolizing gene is sufficient to produce reduced metabolic capacity, with increased systemic exposure often requiring a preemptive dose reduction. This principle was demonstrated by a St. Jude research group for thiopurine methyltransferase and mercaptopurine and has been carried forward for many medications.6 Patients with two reduced-function variations usually require significant dose reductions or substitution of an alternate therapy. Likewise, for medications that require metabolism for the pharmacologic effect, reduced gene activity can lead to ineffective therapy. It is also important to consider the supportive care medications used for pain, depression, neuropathy, nausea and vomiting, and infections; in these cases PG information can improve both safety and efficacy.3

Somatic gene variations are acquired variants, mostly within the tumor, that may predict response to a medication. (Table 2 – see PDF) provides a summary of somatic gene testing for currently approved medications.

The concept that genetic expression profiles for a tumor can define the biology was first demonstrated for mixed lineage leukemia (MLL) and has evolved into use of the tumor’s genetic characteristics to research and define therapy.7 Pembrolizumab is the first medication to be approved for use in cancer patients on the basis of demonstration of genetic variation, either microsatellite instability-high (MSI-H) or deficient mismatch repair (dMMR) solid tumors, and not a specific tumor type.8 Using somatic variations has also paved the way for precision medicine trials such as the National Cancer Institute (NCI) MATCH or umbrella trials, in which the therapy is determined by genetic variation(s) and not tumor histology.

From Bench to Clinical Recommendations
Clinical translation of pharmacogenetic information is available from several sources in addition to the primary literature. The U.S. Food and Drug Administration (FDA) maintains a website for pharmacogenetic information in product labeling.9,10 Of the 269 PG entries for more than 100 medications, the majority (102) are for oncology agents (Table 3 – see PDF).

The Pharmacogenomics Knowledgebase (PharmGKB) is a National Institutes of Health (NIH)–funded international comprehensive resource that provides curated information about the relationship between genetic variation and medication response.11 The information includes pharmacogenomic-guided dosing guidelines, pharmacogenomic information included in product labels for medications from several countries, medication metabolic pathways, clinical annotations (summaries of evidence for relationships between gene variations and medications), variant annotations (summaries of single-gene variation and drug response), and very important pharmacogene (VIP) summaries.

The Clinical Pharmacogenetics Implementation Consortium (CPIC) has developed a rigorous procedure to develop clinical guidelines for how actionable pharmacogenetic information can be used to optimize medication therapy.12 The guidelines provide a translation from genotype (e.g., diplotype, *1/*3) to phenotype (e.g., intermediate metabolizer) to a concrete clinical recommendation (e.g., reduce starting dose by 20%). In addition to the basic supporting information for the recommendation, the CPIC guidelines include clinical decision support algorithms. The guidelines are updated periodically. (Table 4 – See PDF) lists guidelines for the medications used in cancer patients.13,14 A comparative summary of available PharmGKB and CPIC guidelines is available on the PharmGKB website (

Precision medicine services, which involve a pharmacist as a critical team member, have also evolved for cases in which treatment, usually for patients with refractory or rare tumors, is determined by genetics.16-19 These publications highlight the critical role that pharmacists can play in translating pharmacogenomic information into clinical care.

From Clinical Recommendations to Electronic Health Record (EHR)
To successfully implement a PG program in your institution, determine the model that best suits the institution’s needs.20,21 As part of this assessment, identify the physical space requirements and financial, technical, and human resources that will be required. In addition, the cost of the PG testing and the opportunities for reimbursement should be reviewed. Many models may be implemented. Specialized clinics or PG consultation services may require additional staff members. Implementation of an electronic program may require fewer clinician staff additions but may also require a short-term increase in developer resources if the EHR can support customization.

Executive leadership support for establishing a PG program is essential. Guidelines should then be established for implementing the clinical program. A multidisciplinary subcommittee of the pharmacy and therapeutics committee that includes physicians, pharmacists, laboratory medicine personnel, nurses, and information technology (IT) representatives should be formed to establish and maintain the PG program within the EHR.

The multidisciplinary committee should establish the criteria for medication-gene pairs to be included in the program, the expected behavior of the clinician, and other restrictions or requirements. Involving laboratory medicine personnel is critical in establishing procedures for genomics testing, for reviewing the turnaround times for results, and for establishing a process for notification of results. Collaboration between the laboratory medicine and IT departments determines how the results will be stored and whether the results can be used for electronic clinical decision support.

Deciding which Clinical Laboratory Improvement Amendments (CLIA)–certified pharmacogenetic test will be used, and whether to use a single test or a multigene array, is crucial to implementing the program; the connection from the lab to the EHR may ultimately determine which test is implemented.22,23 Haga and colleagues surveyed clinical testing laboratories in the United States.23 Seventy-six of 111 labs offered PG testing. Thirty-one labs offered only single-gene testing, 30 offered multigene testing, and 15 offered both single- and multigene testing. The multigene array covered 295 genes.

Rules should be established for clinical decision support (CDS) according to best practices.24 CDS will be based on retrieving and evaluating a pharmacogenomic lab result that automatically provides a recommendation to the clinician at the time of medication ordering. However, many institutions that have implemented a PG program use manual interpretation of the result before a recommendation is available for the clinician’s review. Ideally CDS should be able to accommodate results for multiple drug-gene associations. For example, in patients under consideration to receive phenytoin, the presence of the HLA-B allele predicts for serious dermatologic reactions, and having one or more reduced-function alleles of CYP2C9 requires dose adjustment.14

Before activation of a PG program, prescribers, pharmacists, nurses, laboratory staff, and patients should be educated about the program. Education material should be created for the patients and be readily available to the provider. Genetic counselors should also be available.

The program’s effectiveness should be monitored, reviewed, and reported regularly to the pharmacy and therapeutics committee. A process should be implemented for periodic review of established guidelines and for creation of new guidelines.

As an example of how effective CDS can facilitate using PG information, O’Donnell and colleagues analyzed 2,279 outpatient encounters in which PG information was provided at the point of care. Medication orders with high pharmacogenomics risk (odds ratio = 26.2 [9 – 75.3]; p < .0001) or cautionary pharmacogenomics risk (odds ratio = 2.4 [1.7 – 3.5]; p <.001) were changed more frequently when PG information was provided.25

Broad implementation of PG requires preemptive testing because the current PG information is most useful when initiating a new therapy. Germline gene variations, many of which are associated with medication therapy, do not change during a patient’s lifetime. Therefore, if the patient is tested early in life, the information can be used to guide medication use across the patient’s lifespan. Although some institutions have adopted this approach, several barriers to wide distribution of this information exist.4,26-28 Klein and colleagues identified 229 publications that included information about PG implementation, barriers, and solutions.4 The major common barriers identified, especially for countries outside the Untied States, are (1) a secure and suitable information technology platform, (2) integrated clinical decision support, (3) regulations, (4) reimbursement, and (5) PG education.

Suggestions for overcoming these barriers include (1) working with EHR vendors to improve the receiving, storing, and displaying of PG information; (2) establishing a standard for PG CDS; (3) implementing preemptive PG programs; (4) collecting evidence to show cost-effectiveness; (5) conducting focus groups to define educational needs and garner stakeholder buy-in; (6) using pharmacists to bridge the communication of genotype information to the provider and patient; and (7) developing a framework for dealing with regulatory issues.

Applying PG information routinely in patient care, especially for oncology patients who are susceptible to adverse effects, should become a standard of care. Pharmacists should be the focal point and a resource for personalized medicine teams and should drive the translation of PG information into patient care through the EHR.


  1. Relling MV, Evans WE. Pharmacogenomics in the clinic. Nature. 2015;526(7573):343-350.
  2. Dey N, Williams C, Leyland-Jones B, De P. Promise of precision medicine in cancers: lessons learned. S D Med. 2017;70:486-490.
  3. Patel JN, Wiebe LA, Dunnenberger HM, McLeod HL. Value of supportive care pharmacogenomics in oncology practice. Oncologist. 2018 Apr 5. pii: theoncologist.2017-0599. [Epub ahead of print]
  4. Klein ME, Parvez MM, Shin JG. Clinical implementation of pharmacogenomics for personalized precision medicine: barriers and solutions. J Pharm Sci. 2017;106(9):2368-2379.
  5. Wellmann R, Borden BA, Danahey K, et al. Analyzing the clinical actionability of germline pharmacogenomic findings in oncology. Cancer. 2018 May 9. doi: 10.1002/cncr.31382. [Epub ahead of print] PubMed PMID: 29742281.
  6. Relling MV, Hancock ML, Rivera GK, et al. Mercaptopurine therapy intolerance and heterozygosity at the thiopurine S-methyltransferase gene locus. J Natl Cancer Inst. 1999;91(23):2001-2008.
  7. Armstrong SA, Staunton JE, Silverman LB, et al. MLL translocations specify a distinct gene expression profile that distinguishes a unique leukemia. Nat Genet. 2002;30(1):41-47.
  8. Boyiadzis MM, Kirkwood JM, Marshall JL, Pritchard CC, Azad NS, Gulley JL. Significance and implications of FDA approval of pembrolizumab for biomarker-defined disease. J Immunother Cancer. 2018;6(1):35.
  9. Table of Pharmacogenomic Biomarkers in Drug Labeling. Available from: Accessed on May 19, 2018.
  10. Green DJ, Mummaneni P, Kim IW, Oh JM, Pacanowski M, Burckart GJ. Pharmacogenomic information in FDA-approved drug labels: application to pediatric patients. Clin Pharmacol Ther. 2016;99(6):622-632.
  11. The PharmGKB website. Accessed on May 19, 2018.
  12. Caudle KE, Klein TE, Hoffman JM, et al. Incorporation of pharmacogenomics into routine clinical practice: the clinical pharmacogenetics implementation consortium (CPIC) guideline development process. Curr Drug Metab. 2014;15(2):209-217.
  13. Whirl-Carrillo M, McDonagh EM, Hebert JM, et al. Pharmacogenomics knowledge for personalized medicine. Clin Pharmacol Ther. 2012;92(4):414–417.
  14. Clinical Pharmacogenetics Implementation Consortium (CPIC) Guidelines. Available from Accessed May 19, 2018.
  15. Dosing Guidelines. Available from: Accessed on May 18, 2018.
  16. Arnall JR, Petro R, Patel JN, Kennedy L. A clinical pharmacy pilot within a precision medicine program for cancer patients and review of related pharmacist clinical practice. J Oncol Pharm Pract. 2017 Jan 1:1078155217738324. doi:10.1177/1078155217738324. [Epub ahead of print] PubMed PMID: 29078708.
  17. Dunnenberger HM, Biszewski M, Bell GC, et al. Implementation of a multidisciplinary pharmacogenomics clinic in a community health system. Am J Health Syst Pharm. 2016;73(23):1956-1966.
  18. Knepper TC, Bell GC, Hicks JK, et al. Key lessons learned from Moffitt’s Molecular Tumor Board: the clinical genomics action committee experience. Oncologist. 2017;22(2):144-151.
  19. Cavallari LH, Lee CR, Duarte JD, et al. Implementation of inpatient models of pharmacogenetics programs. Am J Health Syst Pharm. 2016;73:1944-1954.
  20. Hicks JK, Dunnenberger HM, Gumpper KF, Haidar CE, Hoffman JM. Integrating pharmacogenomics into electronic health records with clinical decision support. Am J Health Syst Pharm. 2016;73:1967-1976.
  21. Caudle KE, Gammal RS, Whirl-Carrillo M, Hoffman JM, Relling MV, Klein TE. Evidence and resources to implement pharmacogenetics knowledge for precision medicine. Am J Health Syst Pharm. 2016;73:1977-1985.
  22. Vo TT, Bell GC, Owusu Obeng A, Hicks JK, Dunnenberger HM. Pharmacogenomics implementation: considerations for selecting a reference laboratory. Pharmacotherapy. 2017;37(9):1014-1022.
  23. Haga SB, Kantor A. Horizon scan of clinical laboratories offering pharmacogenetic testing. Health Aff (Millwood). 2018 May;37(5):717-723.
  24. Danahey K, Borden BA, Furner B, et al. Simplifying the use of pharmacogenomics in clinical practice: building the genomic prescribing system. J Biomed Inform. 2017;75:110-121.
  25. O’Donnell PH, Wadhwa N, Danahey K, et al. Pharmacogenomics based point-of-care clinical decision support significantly alters drug prescribing. Clin Pharmacol Ther. 2017;102:859-869.
  26. Dunnenberger HM, Crews KR, Hoffman JM, et al. Preemptive clinical pharmacogenetics implementation: current programs in five US medical centers. Annu Rev Pharmacol Toxicol. 2015;55:89-106.
  27. Sissung TM, McKeeby JW, Patel J, et al. Pharmacogenomic implementation at the National Institutes of Health Clinical Center. J Clin Pharmacol. 2017;57(suppl. 10):S67-S77.
  28. Volpi S, Bult CJ, Chisholm RL, et al. Research directions in the clinical implementation of pharmacogenomics: an overview of US programs and projects. Clin Pharmacol Ther. 2018;103(5):778-786.