Key Points
Drug summary:
Pharmacogenomic information is contained in the labeling of many drugs; the type of data available determine whether the results of a test are clinically actionable or useful.
Pharmacogenomic tests can be used to select patients for therapy based on their ability to identify responders, predict adverse events, or guide drug dosing.
Most pharmacogenomic labeling has focused on drugs that
Have a narrow therapeutic index
Exhibit highly variable pharmacokinetics or responses
Are used to treat morbid or mortal conditions
Have serious toxicities
Several examples exist where knowledge of a patient’s genotype can significantly influence the benefit-risk profile of a drug product and aid therapeutic decision making. Some examples of this include
Warfarin—CYP2C9 and VKORC1 genotypes predict stable warfarin doses and risk for severe bleeding.
Clopidogrel—CYP2C19 genotype identifies individuals with low active metabolite exposure and diminished responses who may benefit from drugs not metabolized by CYP2C19.
Abacavir—HLA-B*5701 genotype identifies patients at risk for severe hypersensitivity reactions that should not receive the drug.
For a comprehensive list of all drugs with pharmacogenomic information appearing in their labels see .
Introduction
In the last decade, new tools and methods to explore the human genome have tremendously accelerated discovery of genetic markers for disease susceptibility, prognosis, and drug response. Hence, many clinicians are now faced with staying abreast of the enormous amount of genomic information being generated. Numerous resources are available to keep prescribers informed of important information related to drugs, including the FDA-approved drug label. With greater ability to characterize the genetic underpinnings of drug exposure (ie, pharmacokinetics) and response, whether intended effects or toxicities, drug labels will increasingly incorporate pharmacogenomic information and serve as a key resource for genetic biomarker information. In this chapter, we highlight some of the considerations that go into labeling drugs with pharmacogenomic biomarker information, as well as some translational challenges borne out in recent years.
General Labeling Considerations
Prescription drug labeling is intended to provide a summary of the essential scientific information needed for the safe and effective use of a drug, so that clinicians are able to make informed prescribing decisions. The CFR§ Sec. 201.56 states that labels meet the following requirements: (1) The label must contain a summary of the essential scientific information needed for the safe and effective use of the drug. (2) The label must be informative and accurate and must be updated when new information becomes available. (3) The label must be based where possible on data derived from human experience and no claims may be made if there is inadequate evidence of safety or lack of efficacy.
Currently, pharmacogenomic information appears in the labeling of dozens of drugs, covering nearly all therapeutic areas (Fig. 3-1; a comprehensive listing of these labels can be found on the FDA website). To date, most pharmacogenomic labeling has focused on drugs that (1) have a narrow therapeutic index, (2) exhibit highly variable pharmacokinetics or responses, (3) are used to treat morbid or mortal conditions, or (4) have serious toxicities. Drugs with these characteristics tend to be the most difficult to use in the absence of clinical monitoring tools, and are thus amenable to genetic testing as a means of assessing an individual’s probability of benefit or risk.
Figure 3-1
Pharmacogenomic Information in Approved Drug Labels
The figure illustrates the therapeutic areas and nature of pharmacogenomic information in the labels of 74 unique molecules. *Other host factors include errors of inborn metabolism (eg, glucose-6-phosphate dehydrogenase [G6PD deficiency]) and immunologic factors (eg, human leukocyte antigen [HLA] type).
Clinical testing for patient selection or dose adjustment is not necessary for most drugs although numerous examples of clinically relevant pharmacogenomic interactions do exist. As shown in Fig. 3-1, most pharmacogenomic labeling is related to genetic defects in drug metabolism that affect drug concentrations (eg, because of CYP2D6 polymorphisms). Generally, doses are to be reduced in patients that cannot metabolize a particular drug, but genotyping does not always need to be performed prospectively. For instance, tetrabenazine is a drug that exhibits very high concentrations in CYP2D6 poor metabolizers at a certain dose threshold. Therefore, genotyping is recommended if that dose is reached to determine whether the dose should be increased further, so as to avoid toxicities. In the case of prodrugs, on the other hand, inability to activate the drug because of genetic polymorphisms may result in loss of efficacy, as discussed in the clopidogrel example later in this chapter.
Many examples exist where the genetic information is used to characterize the drug target in order to select patients for treatment, particularly for oncology drugs. A classic example is trastuzumab which is indicated for breast cancer patients whose tumors express HER2 since clinical trials for this drug were conducted only in patients whose tumors expressed this marker. More recently, KRAS mutations were found in postmarketing trials to significantly influence the outcome of antiepidermal growth factor receptor (anti-EGFR) therapy in metastatic colorectal cancer patients, which resulted in revisions to the panitumumab and cetuximab labels. Additionally, genetic factors that are major predictors of response to standards of care or disease prognosis are also captured in labeling, such as IL28B genotype for peg- interferon and the direct-acting antiviral drugs, which are used to treat chronic hepatitis C. This information may not be used to select patients per se, but may be used to determine the most appropriate treatment strategy.
Pharmacogenomic information may appear in a variety of forms within the drug label, depending on the implications of the biomarker on appropriate use of the medication. The quantity and quality of pharmacogenomic information depend in part on the strength of the relationship between the biomarker and the outcome of interest, the predictive utility of the biomarker, the severity of the outcome, availability of alternative therapies or feasibility of dose adjustment, and numerous other contextual factors. For example, genetic information may be described in the Clinical Pharmacology section for informational purposes if the drug is metabolized by an enzyme that is known to have genetic variations, but might only be critical to dosage and administration if it is necessary to modify the dosing regimen to prevent toxicities. Understanding the relevance and clinical utility of pharmacogenomic information is not only important for all healthcare practitioners, but also helps to provide targeted therapy to decrease side effects while increasing efficacy. In the following sections, we would review three scenarios where genetic factors were found to be important markers of response, but varied with respect to labeling recommendations and uptake in clinical practice.
