Monday, September 26, 2016

Response to the American Academy of Pediatrics statement on codeine: let's consider CYP2D6

The American Academy of Pediatrics (AAP) issued a statement last Tuesday that codeine should not be used in children. As discussed in the article, pharmacogenetic variation can affect response to codeine, which is metabolized into its active form by the hepatic enzyme CYP2D6. Thus poor metabolizers are unlikely to benefit from codeine, and ultrarapid metabolizers can experience high concentrations of morphine that may result in respiratory depression or sleep apnea even at normal codeine doses.

While we agree that the variability in patient response to codeine is important to consider, we feel that the AAP did not give enough credit to the power of CYP2D6 genotyping prior to treatment. The article devotes only two sentences to the possibility of pre-emptive genotyping, saying “although CYP2D6 genotyping that could identify patients at higher risk is available (although currently expensive), patients with normal metabolism are also at theoretical risk of high morphine levels. Therefore, further investigation is required to determine the value of such testing, which will depend on the population in whom it is applied.”

It is unclear what is intended by the claim that “patients with normal metabolism are also at theoretical risk of high morphine levels”, as no citation is provided. Patients who are CYP2D6 normal metabolizers are expected to have normal morphine formation. Furthermore, codeine is still an appropriate and effective drug choice in the majority of cases, regardless of the population. As the AAP statement points out, ~29% of African/Ethiopian, ~21% of Saudi Arabian and other Middle Eastern, and ~3.4 - 6.5% of African-American and Caucasian patients are expected to be ultrarapid metabolizers, with even lower frequencies of poor metabolizers, meaning that the majority of patients will receive the intended benefit from codeine treatment, contrary to what one doctor was quoted as saying in at least two media outlets (another article here). 

While pharmacogenetics is still in the early stages of clinical adoption, it can be an effective tool for improving patient therapy, particularly when drug alternatives are not simple. In the case of codeine, the official AAP statement urges caution in using alternative analgesics, including oxycodone, hydrocodone, and tramadol, because they also are metabolized to some degree by CYP2D6 (though more than one public media article presented comments by one of the statement authors suggesting that hydrocodone or tramadol be used instead of codeine, without any caveats). Instead, AAP suggests NSAIDs for mild to moderate pain, and an IV drip for greater pain, though these therapies also require clinician education to avoid postoperative bleeding. If further education is required to deliver care, why does this curriculum not consider CYP2D6 pharmacogenetic information, which could make the effects of codeine more predictable and restore its value as an analgesic? As for an alternative to the antitussive properties of codeine, CBS and CBS local New York suggested that popsicles or honey be used.

We are concerned about the misinformation and creative interpretations related to the AAP statement that have been communicated in public media. Chicago Tonight reported that “children are among those considered to be ‘ultra-rapid metabolizers’ of codeine”, which implies to readers that all children are CYP2D6 ultra-rapid metabolizers; they are not. Similarly, a Huffington Post article stated that “children are specifically less capable of metabolizing codeine than adults because the enzyme changes as the body matures into adulthood.” While functional CYP2D6 activity is not appreciably expressed in fetal liver, it increases rapidly after birth, and CYP2D6 genotype is expected to be equally reliable for inferring phenotype in children as it is in adults.

CPIC has published a freely downloadable guideline for adjusting codeine therapy based on CYP2D6 genotype.  The guideline recommends using alternative therapies -- specifically not hydrocodone or tramadol, which are also metabolized by CYP2D6 to some degree -- in patients with genotypes predicting poor and ultra-rapid metabolizer status. The “black box” warning on the FDA label for codeine can also be viewed at PharmGKB.

In the meantime, please pass the medical-grade popsicles.

Thursday, September 22, 2016

Clinical implementation of PGx discussed in JAMA article

A recent article in the Journal of the American Medical Association (JAMA) discusses the current state of pharmacogenetic (PGx) implementation in the clinic. The article notes that clinical uptake of pharmacogenetics has been slow, despite data from the American Medical Association showing that 98% of the more than 10,000 physicians surveyed agree that drug responses may be influenced by genetic variations. This may be due to a lack of implementation guidance and clinical outcome data. Dr. Mary Relling, co-chair of the Clinical Pharmacogenetics Implementation Consortium (CPIC) and chair of the pharmaceutical sciences department at St Jude Children's Research Hospital, is interviewed, and discusses the lack of guidance on FDA labels that contain PGx information, which PharmGKB has previously blogged about. The article also provides an overview of CPIC and its role in providing implementation guidance to clinicians. Dr. Julie Johnson, dean and distinguished professor at the College of Pharmacy at the University of Florida, is interviewed about her Implementing Genomics in Practice (IGNITE) network project, which tracks pharmacogenetic implementation metrics and clinical outcomes, among other areas of research. The article also includes a discussion with Dr. Muin J. Khoury, director of the Office of Public Health Genomics (OPHG) at the Centers for Disease Control and Prevention, on the need for more clinical utility and outcome data for PGx gene-drug pairs. Background on the field of pharmacogenetics is provided, as well as interviews with multiple other leaders in the PGx space, including Dr. Peter H. O'Donnell, associate director for clinical implementation at the University of Chicago's Center for Personalized Therapeutics.

Wednesday, September 21, 2016

Ivacaftor pathway, pharmacokinetics/pharmacodynamics published in Pharmacogenetics and Genomics

The PharmGKB summary of the pharmacokinetics and pharmacodynamics of ivacaftor has been published in Pharmacogenetics and Genomics. Ivacaftor is approved for use in patients who carry particular variants in the CFTR gene. By repotentiating the CFTR ion channel, ivacaftor targets the underlying cause of some forms of cystic fibrosis. Ivacaftor is metabolized by CYP3A4 and CYP3A5 and inhibits ABCB1. Its metabolites are excreted through the transporter SLCO1B1. 

A stylized illustration of the PK/PD pathways of ivacaftor accompanies the publication, and can also be viewed on the PharmGKB website. The illustration and text also includes lumacaftor, a drug that is sometimes used in conjunction with ivacaftor to treat patients with CFTR variants that result in localization defects. The full text and pathway image can be found on the pharmgkb website at

Tuesday, September 20, 2016

NIH Request for Information: Metrics to Assess Value of Biomedical Digital Repositories

NIH has issued a Request for Information to gather input about metrics to assess the impact and value of biomedical digital data repositories, including deposition repositories and knowledgebases.  NIH is interested in collecting information about the usage, quality of data, quality of service and other aspects of existing repositories.  Community use of repositories, in addition to individual use, is emphasized.  Responses should include information such as how specific resources are used, and how community use or importance of a resource can be measured.

A full description of the RFI is available at:

Responses are due by September 30, 2016.  

Wednesday, September 14, 2016

CPIC and PGx discussed in recent GEN article

A recent article in Genetic Engineering & Biotechnology News (GEN) discusses the future of pharmacogenetics in the context of clinical practice. The article notes that personalized medicine will likely be enacted through the work of intermediary groups, such as the Clinical Pharmacogenetics Implementation Consortium (CPIC), who can help clinicians keep up with the latest genetic information. The article interviews Dr. Kelly Caudle, CPIC coordinator, and Dr. Robert Freimuth, co-chair of the CPIC Informatics Working Group, on how the consortium is working to enable the translation of genetic information into actionable clinical results. Ongoing work of Dr. Stuart Scott at the Icahn School of Medicine at Mount Sinai and several European institutes is also discussed.

Friday, September 9, 2016

Nature Outlook: Precision Medicine on Pharmacogenetics highlights its status, promise, and burdens

An outlook on precision medicine in Nature discusses the status, promise, and burdens of the implementation of pharmacogenetic knowledge into clinical care. It highlights the PGEN4Kids program at St Jude, which is implementing CPIC recommendations in a preemptive genetic testing program for pediatric patients. As discussed in the article, CPIC has published recommendations for genotype-guided therapy of 33 drugs, with more guidelines being released each year. CPIC guidelines can be read and downloaded here:

The article reports successful implementation of using TPMT genotypes to adjust thiopurine treatment and of screening for HLA-B genotype to avoid hypersensitivity reactions to abacavir in treatment for HIV. Moving the field forward, we agree with Drew’s emphasis on the increased value of preemptive genetic testing compared to reactionary testing, on the need for sustainable program funding, and on the importance of developing strong infrastructure to support clinical decision making in response to test results.

The article points out the struggles of the pharmacogenetics field, citing the failure of clinical trials to show improved warfarin dosing when using a genotype-guided dosing algorithm compared to a clinical algorithm. However, as Daneshjou, et al. pointed out in NEJM in their 2014 letter to the editor, the finding that the genotype-guided dosing algorithm is inferior to a clinical-dosing algorithm may be biased by different important variants and different frequencies of the variants included in the algorithm in different populations globally. For example, the apparent ineffectiveness of a genotype-guided dosing algorithm for warfarin for people of African descent may be affected by low population frequencies of the CYP2C9*2 and *3 variants that are included in this algorithm. As a result, “the authors’ ability to draw appropriate conclusions about the usefulness of genetics when determining dosages of warfarin for patients of African descent is thus very limited, and the benefits for this population have not been adequately tested.” Furthermore, if a minority of patients receive altered drug treatment as a result of genetic test results in a clinical trial, these benefits may not appear in overall summary statistics. However, the improved personalized treatment for this subset of patients should not be undervalued.

Integrating pharmacogenetics in a conservative medical system seems subject to a genetic exceptionalism resulting in a high burden of proof. Drugs thought to be affected by pharmacogenetic variability comprise 18% of the US drug market, suggesting far-reaching benefits of preemptive testing programs. We see great power in implementing the information we have now, and recognize the need for additional research to expand knowledge of variation globally and of the impacts of variants on drug response so that the benefits of pharmacogenetic testing can continue to grow.