American Association for Clinical Chemistry
Better health through laboratory medicine
Question and Answer Session

June 6, 2006 Presentation: Implementing Parmacogenetics Into a Molecular Diagnostics Reference Lab

Welcome to AACC’s Expert Access Live Online program

This month's experts are Elaine Lyon, PhD, and Gwen McMillin, PhD, from ARUP Laboratories and the University of Utah, Salt Lake City, UT. View the presentation and direct your questions to our online experts. AACC would like to thank Bayer HealthCare Diagnostics for making this program possible.

would it be possible for you to update the online slides to make certain tables are legible (very faint b/w text)? Thanks for the great presentation!

AACC: Our apologies. The PDF version of these slides will be reposted in the near future, however the resolution of slides with embedded images will be limited by the source file. If you need immediate clarification of any slide please open the "online" version of the presentation, which has limited printability but crystal-clear graphics and text. Thank you.

If you can, discuss adverse drug reactions (ADRs)
Washington, DC

Gwendolyn A. McMillin, PhD: Adverse drug reactions could encompass any undesirable effect of a drug, including no effect. If an ADR is dose-related, it will most likely correlate with the circulating concentration of active drug, and may require multiple doses to develop. ADRs are not always dose-related and pharmacogenetic testing cannot predict all ADRs.

would it be possible for you to update the online slides to make certain tables are legible (very faint b/w text)? Thanks for the great presentation!

Elaine Lyon, PhD: We will look into this. Several versions of the presentation are available (i.e. .pdf version).

the insertion/deletion polymorphisim of Angiotensin Converting Enzyme (ACE) effects the plasma levels of ACE and can be used to guide theraphy with ACE inhibitors. Yet some studies show that plasma levels of the enzyme also is effected by gender. Does drug dosing and/or CYP polymorphisims show variation with gender Thank you Bilgen Dolek, PhD
Istanbul, Turkey

Gwendolyn A. McMillin, PhD: CYP polymorphisms are not sex-linked genetically, so are not genetically affected by gender. However, optimal drug dosing must consider many factors, including gender (often associated with hormonal differences between genders), age, weight, protein status, renal function, concomitant medications, diet and lifestyle factors (such as smoking), and genetics.

We are currently completing the validation process for the Roche AmpliChip CYP 450 with plans to launch the test by the end of the month. Physician education will be critical for the successful launch of this test. What recommendations do you have for the scope of the education and how do we choose a target physician population?
Grand Rapids, Michigan

Elaine Lyon, PhD: We need to be very pro-active with physician education. We’ve found that many physicians know about pharmacogenetics, but are uncertain how to incorporate this information into their clinical practice. One difficulty is that dosing guidelines are not always available. Offering to present at Grand Rounds or staff meetings usually receives a positive response. Direct your presentation to specific drugs used by physicians’ specialty, the test available and whenever possible, dosage guidelines based on genotype. A written sheet such as a technical bulletin with your contact information for questions is also helpful.

When will be available the guidelines for Pharmacogenetics? Are there rules that include the inicial dose to start a drug according to the CYP profile?
São Paulo, Brazil

Elaine Lyon, PhD: Several societies are working towards guidelines in pharmacogenetics, but are not yet published. One from the National Academy of Clinical Biochemistry (NACB) focuses on laboratory guidelines to ensure accurate testing. Another multidisciplinary effort will focus specifically on warfarin sensitivity. There are a number of published papers suggesting dosage changes by genotype, but these studies have limited study participants and don’t address combinations of drugs an individual may be taking. We recommend a consultation with a clinical pharmacist with our laboratory reports. I would foresee more drug-specific guidelines in the future.

I work in Roche Diagnostics Spain being responsible for AmpliChip CYP450. I have some questions: Do you think AmpliChip CYP450 could be cost-effective in Psychiatric considering its price ($500-$600) Have you considered 2C19 genotyping in schlerodermia patients?. We have read some anticonvulsant drugs (mephenitoin, phenitoin, dapsone) could generate schlerodermia because their against-collagenase effect. This ADR's is supposed to be even more severe in CYP2C19 poor metabolizers. Do you have any experience in this field?. Thanks a lot

Gwendolyn A. McMillin, PhD: Cost-utility is quite institution-specific. That said, there are many opportunities for CYP testing to be cost-effective. Besides the obvious savings to be realized by preventing costs associated with managing ADRs, CYP genotyping may reduce costs associated with drug and dose selection. Initial selection of a drug that is compatible with the patient’s CYP status may improve patient compliance and reduce the time to clinical response, due to less trial-and-error. CYP genotyping has also been used to triage less expensive, but traditionally more toxic drugs, such as the TCAs, to patients without known genetic variants in CYPs because these patients are at lower risk for experiencing the dose-related toxicities of the TCAs. Some references for recent studies related to CYP genotyping in psychiatric populations include Clin Chem 2004 50(9):1623-33, and Clin Chem 2005 51(2):376-85. I do not have experience in correlating CYP genotypes with schlerodermia.

Is there a guideline to ajust the Azathioprine dose based on the Azathioprine metabolites?
São Paulo, Brazil

Gwendolyn A. McMillin, PhD: I am not familiar with specific dose recommendations based on azathioprine metabolite concentrations, partially because laboratory testing for these metabolites is not widely available. However, therapeutic drug monitoring and determination of metabolic ratios in general is an excellent means of phenotyping, optimizing drug dose, identification of drug-drug interactions, evaluating patient compliance and clinical status (eg, renal function), and is therefore recommended.

In setting up the menu for our pharmacogen lab, should we be doing some HIS data mining to determine the ADRs most prevalent in our catchment area, or are there common PGx tests that everyone should be reporting (e.g., the examples in your presentation: Amitriptyline, Venlafaxine, imipramine/desipramine)?
Boston, MA

Gwendolyn A. McMillin, PhD: Prioritizing the laboratory test menu should definitely consider the population to be served by the testing. It would be very useful to mine the HIS for “hot spots” that could impact the current status of patient care at your institution. As polypharmacy is so common in patient care today, such ADRs may not be clearly associated with a single drug. Consideration of common denominators between drugs, ADRs, clinical service areas, and probable pharmacogenetic markers would likely identify many opportunities for improving patient care at your institution.

Concerning IP, do you know, (or is there a list) of genes/mutations which have exclusive licensing which limit testing to just one or two providers?
Akron, Ohio

Elaine Lyon, PhD: It’s hard to comment on exclusive licenses, because although they exist, the licensor may negotiate terms with other entities. Many of the pharmacogenetic tests have some type of license, although not particularly exclusive. If you are using reagents (ASR or FDA) from vendors, check if they include the licensing with the product. If you have your own laboratory developed test, you will need to check and possibly negotiate the license yourself.

What physical barriers and/or distance is required between pre- and post-PCR stations for clinical diagnostic work?
North Kingstown, RI

Elaine Lyon, PhD: The pre-PCR and post PCR processes should be physically separated in a clinical laboratory. We have a sample preparation area, dedicated to DNA extraction, a “clean area” for PCR master-mix set-up and innoculation of the patient sample, and an area (a separate room) dedicated to PCR amplification and any post PCR sample manipulation, such as running gels or reading fluorescence levels. Other considerations would be a positive pressure air control system or laminar flow hoods. There should also be a uni-directional workflow from the pre to the post areas, to prevent contaminating the clean area with PCR amplicon.

Since any pharmaceutical compound may have dozens of genes that are required to produce a desired therapeutic result, it would seem that any single pharmacogenetic test might not offer a complete story of a patient’s response phenotype. Are SNP or multi-gene “panels” in the offing, and would this type of technology be practical for clinical use?
Sacramento, CA

Gwendolyn A. McMillin, PhD: Excellent point. In theory, every protein that interacts with a drug, and is derived from a polymorphic gene, could represent a pharmacogenetic target for that drug. As all possible proteins important to drug handling are not known for most (if not all) drugs, it is not possible that a genetic panel could accurately identify the “fit” between a drug and a person 100% of the time. For this reason, phenotyping and monitoring clinical response is not in any way replaced by pharmacogenetic testing. However, I do believe that multi-gene, application-specific pharmacogenetic panels will be available in the future to predict both pharmacokinetics and pharmacodynamics of specific drugs. The best example today is that of the widely used anticoagulant warfarin, for which prospective genotyping of CYP2C9 and VKORC1 can be used to reduce up to 50% of the variability in assigining initial dosing of that drug.

nice presentation. can you please comment on how the test is reimburst and the volume of the test you have in your lab? Thanks.
berkeley, ca

Elaine Lyon, PhD: The test volumes are still rather low, but have been increasing over the last several months. We anticipate that as guidelines are published and as education and awareness increase, we will see an increase in sample volume. We don’t have experience regarding reimbursement to be able to comment.

What then do you recommend as components of the interpretive report? Should dosing information be included or do you refer to the literature?
Tampa, Fl

Gwendolyn A. McMillin, PhD: The content of the report among clinical laboratories that offer CYP testing is somewhat controversial. We provide the analytical method used, common allele nomenclature, specific SNP detected, predicted phenotype (if known), and any background information that may be useful to the clinician, such as allele frequency. When the predicted phenotype is expected to affect drug and/or dose selection, we recommend consultation with a clinical pharmacy professional. We do not provide dosing guidelines, because there are so many other (non-genetic) factors that affect drug selection and dose optimization.

When will be available the guidelines for Pharmacogenetics? Are there rules that include the inicial dose to start a drug according to the CYP profile?
São Paulo, Brazil

AACC: To add to Dr. Lyon's response, NACB's "Guidelines and Recommendations for Laboratory Analysis and Application of Pharmacogenetics to Clinical Practice" can be found on the Internet at

Please elaborate on the role of the lab in working with clinical pharmacists and physicians. I find that my consults with the lab often bring valuable information relating to patient care. Does ARUP make you or colleagues available for discussions with client physicians?
Detroit, MI

Elaine Lyon, PhD: It is critical for the laboratory to work with clinical pharmacists and physicians. At ARUP, Dr. McMillin is available for pharmacology questions, while I’m available for more technical questions regarding the testing. We also need input from a clinician’s point of view. Hopefully we can learn from each other. Pre-test communication between the clinician and laboratorian can ensure proper ordering for tests. This could include discussions such as whether additional genes in a pathway need to be considered for a specific drug. Post-test communication is necessary if any aspect of the laboratory report is not well understood. It is useful for us to receive any follow-up that could help us improve our service. One idea that we’ve discussed is working with clinical pharmacists to have clinical consultations available.

While much of the excitement related to PGx testing involves the use of biomarkers tied to particular therapeutics, many of the short-term applications involve genotyping of cytochrome p450 genes, as you describe. Some payers are suggesting that additional data need to be collected regarding the economic benefits conferred by these tests before they will be covered. How do you envision these studies being performed and funded? Is ARUP involved in such studies?
San Diego, CA

Gwendolyn A. McMillin, PhD: As with any laboratory test, appropriate economic evaluation will impact reimbursement potential and overall utilization. As you suggest, the companion tests, wherein the result of a test is linked to the prescribing of a drug, have well-defined utility and quite good economic justification. This approach is likely to be common in the future. CYP genotyping is much more challenging to justify from a payers perspective. Some studies, which were referenced in a previous question, have attempted to address this issue, but future studies are clearly needed. ARUP is involved in designing and pursuing such studies, and would encourage others to also conduct and publish findings related to the pharmacoeconomics of CYP genotyping. Possible funding opportunities may lie within professional medical associations that offer grants, industry (pharma and diagnostic), government, and private institutions.

What are the current barriers for implementation of molecular diagnositics in mainstream healthcare? Or in essence, once there is reasonable evidence that a test is important what are the key issues in regards to implementation by a clinical lab and acceptance in the medical community (Insurance, cost of the test, lack of education regarding the new test etc...)?
Silver Spring, MD

Elaine Lyon, PhD: Once the clinical validity of a molecular diagnostic test is established, we also need to ensure clinical utility (in other words, that this test is truly useful to the clinician). Most of the time, both validity and utility are established by peer-reviewed studies. Physicians usually quickly adopt tests that can help them diagnose or manage their patients better. Continuing education from medical societies can help distribute this information. However, some clinicians may not be familiar with issues involved with molecular assays. We would encourage clinicians to contact laboratories if they have any question regarding ordering or test results and interpretation. For the clinical laboratory, questions to address are the cost and complexity of implementing a molecular test. We also ask ourselves how many tests can be performed on the same platform, since instruments are a major cost for the laboratory. For example, a single SNP may be easy to implement, but if full gene sequencing is needed due to the number of variants in a specific gene, labs may choose to refer testing to a high-throughput sequencing laboratory. Insurance coverage is an important issue, but if the clinical utility and validity is strong, insurances are likely agree, especially if it can save in other health care dollars.

How can the clinical laboratory community best facilitate the education of physicians and pharmacists required for widespread adoption of pharmacogenetic testing? How can industry facilitate the process?
Washington, DC

Gwendolyn A. McMillin, PhD: Education of clinicians is of utmost importance for incorporating pharmacogenetics into standard patient care. The warfarin story is a good opportunity for pharmacogenetic testing to infiltrate patient care, but will require clinician education to implement (see previous question on physician education). Widespread adoption of pharmacogenetic testing will, in my opinion, require a combined and unified message, which includes actions to be taken based on test results, from the industries that provide the drug and the IVD. Support from regulatory agencies, such as the FDA, particularly related to drug labeling, is another driving force. As such, widespread adoption will require coordination of drug and test clearance (in the US; ie, FDA), and availability, and will be most successful with new drugs. Thus, the initial education of clinicians about the drug product will include education of the test, as a “package.” This practice will improve awareness among clinicians and therefore make it easier for clinician to accept and implement pharmacogenetic testing for existing drugs. CYP genotyping in general will be slow to penetrate widespread patient care completely until drug applications/actions are better defined.

What is the role for genetic counselors in the reporting results? Katherine Berry, MS
Helena, MT

Elaine Lyon, PhD: Pharmacogenetic testing is different than traditional genetic testing in that the variants tested are not associated with disease and are not apparent until after an environmental exposure to a drug. Genetic counselors could discuss the implications for other family members and the need to share information to others that may be at risk for ADR’s. However, most of the counseling will need to be done by clinical pharmacists, directed to the patient and the medications they are taking.

How do you calculate the correct dose of a drug for a patient if three CYPs metabolize the drug of interest and polymorphisms in two or three CYPs influence the clinical outcome?

Gwendolyn A. McMillin, PhD: Interaction of drugs with more than one CYP is common, so your question is very practical. It is, for this reason, important to evaluate/genotype all polymorphic CYPs known to interact with a drug. The easiest way to deal with drug dosing in a patient with genetic variants in more than one CYP is to avoid drugs that require those CYPs. If such alternative drug choices are not available, there are some dosing guidelines proposed for multiple CYP gene variants, and specific drugs. For example, Kirchheiner J et al. Mol Psych 2004;9:442-73 and Hung C et al. Ther Drug Monit 2004, 26(5):534-40. I am optimistic that additional studies addressing the involvement and dosing related to multiple CYP variants will be published. Alternatively, I would recommend frequent therapeutic drug monitoring of active drug and metabolites to optimize dosing.

What is your source of quality control materials? How is QC done differently in pharmacogenetics from traditional chemistry QCs?
Milwaukee, WI

Elaine Lyon, PhD: Acquiring samples for pharmacogenetic test validation is a challenge. To validate the assay, samples need to be genotyped by two different methods or by two separate laboratories. We initially validated our cytochrome p450 assays using DNA obtained from Coriell Cell Repository and compared results to another laboratory that had access to the same samples from Coriell. Results for CYP 2D6, 2C9 and 2C19 panels with these cell lines are now being collected by the CDC and are available on their website. The CDC is active to collect other pharmacogenetic variants. Also, many laboratories are willing to share samples for test validations. In many ways, genotype test validations are similar to other laboratory validations, requiring accuracy and reproducibility. Genetic tests are usually qualitative, with the testing for either the presence or absence of a genetic variant.